Merge branch 'perfcounters-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'perfcounters-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (574 commits)
  perf_counter: Turn off by default
  perf_counter: Add counter->id to the throttle event
  perf_counter: Better align code
  perf_counter: Rename L2 to LL cache
  perf_counter: Standardize event names
  perf_counter: Rename enums
  perf_counter tools: Clean up u64 usage
  perf_counter: Rename perf_counter_limit sysctl
  perf_counter: More paranoia settings
  perf_counter: powerpc: Implement generalized cache events for POWER processors
  perf_counters: powerpc: Add support for POWER7 processors
  perf_counter: Accurate period data
  perf_counter: Introduce struct for sample data
  perf_counter tools: Normalize data using per sample period data
  perf_counter: Annotate exit ctx recursion
  perf_counter tools: Propagate signals properly
  perf_counter tools: Small frequency related fixes
  perf_counter: More aggressive frequency adjustment
  perf_counter/x86: Fix the model number of Intel Core2 processors
  perf_counter, x86: Correct some event and umask values for Intel processors
  ...

27 files changed, 2130 insertions, 70 deletions

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index aafae3b140d..68f5578fe38 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -739,6 +739,7 @@ config X86_UP_IOAPIC
 config X86_LOCAL_APIC
 	def_bool y
 	depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_APIC
+	select HAVE_PERF_COUNTERS if (!M386 && !M486)
 
 config X86_IO_APIC
 	def_bool y
diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index dcef387ddc3..e590261ba05 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -825,10 +825,11 @@ ia32_sys_call_table:
 	.quad compat_sys_signalfd4
 	.quad sys_eventfd2
 	.quad sys_epoll_create1
-	.quad sys_dup3			/* 330 */
+	.quad sys_dup3				/* 330 */
 	.quad sys_pipe2
 	.quad sys_inotify_init1
 	.quad compat_sys_preadv
 	.quad compat_sys_pwritev
 	.quad compat_sys_rt_tgsigqueueinfo	/* 335 */
+	.quad sys_perf_counter_open
 ia32_syscall_end:
diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h
index 85b46fba422..aff9f1fcdcd 100644
--- a/arch/x86/include/asm/atomic_32.h
+++ b/arch/x86/include/asm/atomic_32.h
@@ -247,5 +247,241 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 #define smp_mb__before_atomic_inc()	barrier()
 #define smp_mb__after_atomic_inc()	barrier()
 
+/* An 64bit atomic type */
+
+typedef struct {
+	unsigned long long counter;
+} atomic64_t;
+
+#define ATOMIC64_INIT(val)	{ (val) }
+
+/**
+ * atomic64_read - read atomic64 variable
+ * @v: pointer of type atomic64_t
+ *
+ * Atomically reads the value of @v.
+ * Doesn't imply a read memory barrier.
+ */
+#define __atomic64_read(ptr)		((ptr)->counter)
+
+static inline unsigned long long
+cmpxchg8b(unsigned long long *ptr, unsigned long long old, unsigned long long new)
+{
+	asm volatile(
+
+		LOCK_PREFIX "cmpxchg8b (%[ptr])\n"
+
+		     :		"=A" (old)
+
+		     : [ptr]	"D" (ptr),
+				"A" (old),
+				"b" (ll_low(new)),
+				"c" (ll_high(new))
+
+		     : "memory");
+
+	return old;
+}
+
+static inline unsigned long long
+atomic64_cmpxchg(atomic64_t *ptr, unsigned long long old_val,
+		 unsigned long long new_val)
+{
+	return cmpxchg8b(&ptr->counter, old_val, new_val);
+}
+
+/**
+ * atomic64_xchg - xchg atomic64 variable
+ * @ptr:      pointer to type atomic64_t
+ * @new_val:  value to assign
+ * @old_val:  old value that was there
+ *
+ * Atomically xchgs the value of @ptr to @new_val and returns
+ * the old value.
+ */
+
+static inline unsigned long long
+atomic64_xchg(atomic64_t *ptr, unsigned long long new_val)
+{
+	unsigned long long old_val;
+
+	do {
+		old_val = atomic_read(ptr);
+	} while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val);
+
+	return old_val;
+}
+
+/**
+ * atomic64_set - set atomic64 variable
+ * @ptr:      pointer to type atomic64_t
+ * @new_val:  value to assign
+ *
+ * Atomically sets the value of @ptr to @new_val.
+ */
+static inline void atomic64_set(atomic64_t *ptr, unsigned long long new_val)
+{
+	atomic64_xchg(ptr, new_val);
+}
+
+/**
+ * atomic64_read - read atomic64 variable
+ * @ptr:      pointer to type atomic64_t
+ *
+ * Atomically reads the value of @ptr and returns it.
+ */
+static inline unsigned long long atomic64_read(atomic64_t *ptr)
+{
+	unsigned long long curr_val;
+
+	do {
+		curr_val = __atomic64_read(ptr);
+	} while (atomic64_cmpxchg(ptr, curr_val, curr_val) != curr_val);
+
+	return curr_val;
+}
+
+/**
+ * atomic64_add_return - add and return
+ * @delta: integer value to add
+ * @ptr:   pointer to type atomic64_t
+ *
+ * Atomically adds @delta to @ptr and returns @delta + *@ptr
+ */
+static inline unsigned long long
+atomic64_add_return(unsigned long long delta, atomic64_t *ptr)
+{
+	unsigned long long old_val, new_val;
+
+	do {
+		old_val = atomic_read(ptr);
+		new_val = old_val + delta;
+
+	} while (atomic64_cmpxchg(ptr, old_val, new_val) != old_val);
+
+	return new_val;
+}
+
+static inline long atomic64_sub_return(unsigned long long delta, atomic64_t *ptr)
+{
+	return atomic64_add_return(-delta, ptr);
+}
+
+static inline long atomic64_inc_return(atomic64_t *ptr)
+{
+	return atomic64_add_return(1, ptr);
+}
+
+static inline long atomic64_dec_return(atomic64_t *ptr)
+{
+	return atomic64_sub_return(1, ptr);
+}
+
+/**
+ * atomic64_add - add integer to atomic64 variable
+ * @delta: integer value to add
+ * @ptr:   pointer to type atomic64_t
+ *
+ * Atomically adds @delta to @ptr.
+ */
+static inline void atomic64_add(unsigned long long delta, atomic64_t *ptr)
+{
+	atomic64_add_return(delta, ptr);
+}
+
+/**
+ * atomic64_sub - subtract the atomic64 variable
+ * @delta: integer value to subtract
+ * @ptr:   pointer to type atomic64_t
+ *
+ * Atomically subtracts @delta from @ptr.
+ */
+static inline void atomic64_sub(unsigned long long delta, atomic64_t *ptr)
+{
+	atomic64_add(-delta, ptr);
+}
+
+/**
+ * atomic64_sub_and_test - subtract value from variable and test result
+ * @delta: integer value to subtract
+ * @ptr:   pointer to type atomic64_t
+ *
+ * Atomically subtracts @delta from @ptr and returns
+ * true if the result is zero, or false for all
+ * other cases.
+ */
+static inline int
+atomic64_sub_and_test(unsigned long long delta, atomic64_t *ptr)
+{
+	unsigned long long old_val = atomic64_sub_return(delta, ptr);
+
+	return old_val == 0;
+}
+
+/**
+ * atomic64_inc - increment atomic64 variable
+ * @ptr: pointer to type atomic64_t
+ *
+ * Atomically increments @ptr by 1.
+ */
+static inline void atomic64_inc(atomic64_t *ptr)
+{
+	atomic64_add(1, ptr);
+}
+
+/**
+ * atomic64_dec - decrement atomic64 variable
+ * @ptr: pointer to type atomic64_t
+ *
+ * Atomically decrements @ptr by 1.
+ */
+static inline void atomic64_dec(atomic64_t *ptr)
+{
+	atomic64_sub(1, ptr);
+}
+
+/**
+ * atomic64_dec_and_test - decrement and test
+ * @ptr: pointer to type atomic64_t
+ *
+ * Atomically decrements @ptr by 1 and
+ * returns true if the result is 0, or false for all other
+ * cases.
+ */
+static inline int atomic64_dec_and_test(atomic64_t *ptr)
+{
+	return atomic64_sub_and_test(1, ptr);
+}
+
+/**
+ * atomic64_inc_and_test - increment and test
+ * @ptr: pointer to type atomic64_t
+ *
+ * Atomically increments @ptr by 1
+ * and returns true if the result is zero, or false for all
+ * other cases.
+ */
+static inline int atomic64_inc_and_test(atomic64_t *ptr)
+{
+	return atomic64_sub_and_test(-1, ptr);
+}
+
+/**
+ * atomic64_add_negative - add and test if negative
+ * @delta: integer value to add
+ * @ptr:   pointer to type atomic64_t
+ *
+ * Atomically adds @delta to @ptr and returns true
+ * if the result is negative, or false when
+ * result is greater than or equal to zero.
+ */
+static inline int
+atomic64_add_negative(unsigned long long delta, atomic64_t *ptr)
+{
+	long long old_val = atomic64_add_return(delta, ptr);
+
+	return old_val < 0;
+}
+
 #include <asm-generic/atomic.h>
 #endif /* _ASM_X86_ATOMIC_32_H */
diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h
index c2e6bedaf25..d750a10ccad 100644
--- a/arch/x86/include/asm/entry_arch.h
+++ b/arch/x86/include/asm/entry_arch.h
@@ -49,7 +49,7 @@ BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR)
 BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR)
 
 #ifdef CONFIG_PERF_COUNTERS
-BUILD_INTERRUPT(perf_counter_interrupt, LOCAL_PERF_VECTOR)
+BUILD_INTERRUPT(perf_pending_interrupt, LOCAL_PENDING_VECTOR)
 #endif
 
 #ifdef CONFIG_X86_MCE_P4THERMAL
diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h
index 37555e52f98..9ebc5c25503 100644
--- a/arch/x86/include/asm/hardirq.h
+++ b/arch/x86/include/asm/hardirq.h
@@ -13,6 +13,8 @@ typedef struct {
 	unsigned int irq_spurious_count;
 #endif
 	unsigned int generic_irqs;	/* arch dependent */
+	unsigned int apic_perf_irqs;
+	unsigned int apic_pending_irqs;
 #ifdef CONFIG_SMP
 	unsigned int irq_resched_count;
 	unsigned int irq_call_count;
diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h
index 3bd1777a4c8..6df45f63966 100644
--- a/arch/x86/include/asm/hw_irq.h
+++ b/arch/x86/include/asm/hw_irq.h
@@ -29,6 +29,8 @@
 extern void apic_timer_interrupt(void);
 extern void generic_interrupt(void);
 extern void error_interrupt(void);
+extern void perf_pending_interrupt(void);
+
 extern void spurious_interrupt(void);
 extern void thermal_interrupt(void);
 extern void reschedule_interrupt(void);
diff --git a/arch/x86/include/asm/intel_arch_perfmon.h b/arch/x86/include/asm/intel_arch_perfmon.h
deleted file mode 100644
index fa0fd068bc2..00000000000
--- a/arch/x86/include/asm/intel_arch_perfmon.h
+++ /dev/null
@@ -1,31 +0,0 @@
-#ifndef _ASM_X86_INTEL_ARCH_PERFMON_H
-#define _ASM_X86_INTEL_ARCH_PERFMON_H
-
-#define MSR_ARCH_PERFMON_PERFCTR0		0xc1
-#define MSR_ARCH_PERFMON_PERFCTR1		0xc2
-
-#define MSR_ARCH_PERFMON_EVENTSEL0		0x186
-#define MSR_ARCH_PERFMON_EVENTSEL1		0x187
-
-#define ARCH_PERFMON_EVENTSEL0_ENABLE	(1 << 22)
-#define ARCH_PERFMON_EVENTSEL_INT	(1 << 20)
-#define ARCH_PERFMON_EVENTSEL_OS	(1 << 17)
-#define ARCH_PERFMON_EVENTSEL_USR	(1 << 16)
-
-#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL	(0x3c)
-#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK	(0x00 << 8)
-#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 /* _ASM_X86_INTEL_ARCH_PERFMON_H */
diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h
index 910b5a3d675..e997be98c9b 100644
--- a/arch/x86/include/asm/irq_vectors.h
+++ b/arch/x86/include/asm/irq_vectors.h
@@ -108,14 +108,14 @@
 #define LOCAL_TIMER_VECTOR		0xef
 
 /*
- * Performance monitoring interrupt vector:
+ * Generic system vector for platform specific use
  */
-#define LOCAL_PERF_VECTOR		0xee
+#define GENERIC_INTERRUPT_VECTOR	0xed
 
 /*
- * Generic system vector for platform specific use
+ * Performance monitoring pending work vector:
  */
-#define GENERIC_INTERRUPT_VECTOR	0xed
+#define LOCAL_PENDING_VECTOR		0xec
 
 /*
  * First APIC vector available to drivers: (vectors 0x30-0xee) we
diff --git a/arch/x86/include/asm/perf_counter.h b/arch/x86/include/asm/perf_counter.h
new file mode 100644
index 00000000000..876ed97147b
--- /dev/null
+++ b/arch/x86/include/asm/perf_counter.h
@@ -0,0 +1,100 @@
+#ifndef _ASM_X86_PERF_COUNTER_H
+#define _ASM_X86_PERF_COUNTER_H
+
+/*
+ * Performance counter hw details:
+ */
+
+#define X86_PMC_MAX_GENERIC					8
+#define X86_PMC_MAX_FIXED					3
+
+#define X86_PMC_IDX_GENERIC				        0
+#define X86_PMC_IDX_FIXED				       32
+#define X86_PMC_IDX_MAX					       64
+
+#define MSR_ARCH_PERFMON_PERFCTR0			      0xc1
+#define MSR_ARCH_PERFMON_PERFCTR1			      0xc2
+
+#define MSR_ARCH_PERFMON_EVENTSEL0			     0x186
+#define MSR_ARCH_PERFMON_EVENTSEL1			     0x187
+
+#define ARCH_PERFMON_EVENTSEL0_ENABLE			  (1 << 22)
+#define ARCH_PERFMON_EVENTSEL_INT			  (1 << 20)
+#define ARCH_PERFMON_EVENTSEL_OS			  (1 << 17)
+#define ARCH_PERFMON_EVENTSEL_USR			  (1 << 16)
+
+/*
+ * Includes eventsel and unit mask as well:
+ */
+#define ARCH_PERFMON_EVENT_MASK				    0xffff
+
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL		      0x3c
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK		(0x00 << 8)
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX 		 0
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT \
+		(1 << (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX))
+
+#define ARCH_PERFMON_BRANCH_MISSES_RETIRED			 6
+
+/*
+ * Intel "Architectural Performance Monitoring" CPUID
+ * detection/enumeration details:
+ */
+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;
+};
+
+union cpuid10_edx {
+	struct {
+		unsigned int num_counters_fixed:4;
+		unsigned int reserved:28;
+	} split;
+	unsigned int full;
+};
+
+
+/*
+ * Fixed-purpose performance counters:
+ */
+
+/*
+ * All 3 fixed-mode PMCs are configured via this single MSR:
+ */
+#define MSR_ARCH_PERFMON_FIXED_CTR_CTRL			0x38d
+
+/*
+ * The counts are available in three separate MSRs:
+ */
+
+/* Instr_Retired.Any: */
+#define MSR_ARCH_PERFMON_FIXED_CTR0			0x309
+#define X86_PMC_IDX_FIXED_INSTRUCTIONS			(X86_PMC_IDX_FIXED + 0)
+
+/* CPU_CLK_Unhalted.Core: */
+#define MSR_ARCH_PERFMON_FIXED_CTR1			0x30a
+#define X86_PMC_IDX_FIXED_CPU_CYCLES			(X86_PMC_IDX_FIXED + 1)
+
+/* CPU_CLK_Unhalted.Ref: */
+#define MSR_ARCH_PERFMON_FIXED_CTR2			0x30b
+#define X86_PMC_IDX_FIXED_BUS_CYCLES			(X86_PMC_IDX_FIXED + 2)
+
+extern void set_perf_counter_pending(void);
+
+#define clear_perf_counter_pending()	do { } while (0)
+#define test_perf_counter_pending()	(0)
+
+#ifdef CONFIG_PERF_COUNTERS
+extern void init_hw_perf_counters(void);
+extern void perf_counters_lapic_init(void);
+#else
+static inline void init_hw_perf_counters(void)		{ }
+static inline void perf_counters_lapic_init(void)	{ }
+#endif
+
+#endif /* _ASM_X86_PERF_COUNTER_H */
diff --git a/arch/x86/include/asm/unistd_32.h b/arch/x86/include/asm/unistd_32.h
index 708dae61262..732a3070615 100644
--- a/arch/x86/include/asm/unistd_32.h
+++ b/arch/x86/include/asm/unistd_32.h
@@ -341,6 +341,7 @@
 #define __NR_preadv		333
 #define __NR_pwritev		334
 #define __NR_rt_tgsigqueueinfo	335
+#define __NR_perf_counter_open	336
 
 #ifdef __KERNEL__
 
diff --git a/arch/x86/include/asm/unistd_64.h b/arch/x86/include/asm/unistd_64.h
index 4e2b0540440..900e1617e67 100644
--- a/arch/x86/include/asm/unistd_64.h
+++ b/arch/x86/include/asm/unistd_64.h
@@ -659,7 +659,8 @@ __SYSCALL(__NR_preadv, sys_preadv)
 __SYSCALL(__NR_pwritev, sys_pwritev)
 #define __NR_rt_tgsigqueueinfo			297
 __SYSCALL(__NR_rt_tgsigqueueinfo, sys_rt_tgsigqueueinfo)
-
+#define __NR_perf_counter_open			298
+__SYSCALL(__NR_perf_counter_open, sys_perf_counter_open)
 
 #ifndef __NO_STUBS
 #define __ARCH_WANT_OLD_READDIR
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index a4c9cf0bf70..076d3881f3d 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -14,6 +14,7 @@
  *	Mikael Pettersson	:	PM converted to driver model.
  */
 
+#include <linux/perf_counter.h>
 #include <linux/kernel_stat.h>
 #include <linux/mc146818rtc.h>
 #include <linux/acpi_pmtmr.h>
@@ -34,6 +35,7 @@
 #include <linux/smp.h>
 #include <linux/mm.h>
 
+#include <asm/perf_counter.h>
 #include <asm/pgalloc.h>
 #include <asm/atomic.h>
 #include <asm/mpspec.h>
@@ -1187,6 +1189,7 @@ void __cpuinit setup_local_APIC(void)
 		apic_write(APIC_ESR, 0);
 	}
 #endif
+	perf_counters_lapic_init();
 
 	preempt_disable();
 
diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile
index 4e242f9a06e..3efcb2b96a1 100644
--- a/arch/x86/kernel/cpu/Makefile
+++ b/arch/x86/kernel/cpu/Makefile
@@ -1,5 +1,5 @@
 #
-# Makefile for x86-compatible CPU details and quirks
+# Makefile for x86-compatible CPU details, features and quirks
 #
 
 # Don't trace early stages of a secondary CPU boot
@@ -23,11 +23,13 @@ obj-$(CONFIG_CPU_SUP_CENTAUR)		+= centaur.o
 obj-$(CONFIG_CPU_SUP_TRANSMETA_32)	+= transmeta.o
 obj-$(CONFIG_CPU_SUP_UMC_32)		+= umc.o
 
-obj-$(CONFIG_X86_MCE)	+= mcheck/
-obj-$(CONFIG_MTRR)	+= mtrr/
-obj-$(CONFIG_CPU_FREQ)	+= cpufreq/
+obj-$(CONFIG_PERF_COUNTERS)		+= perf_counter.o
 
-obj-$(CONFIG_X86_LOCAL_APIC) += perfctr-watchdog.o
+obj-$(CONFIG_X86_MCE)			+= mcheck/
+obj-$(CONFIG_MTRR)			+= mtrr/
+obj-$(CONFIG_CPU_FREQ)			+= cpufreq/
+
+obj-$(CONFIG_X86_LOCAL_APIC)		+= perfctr-watchdog.o
 
 quiet_cmd_mkcapflags = MKCAP   $@
       cmd_mkcapflags = $(PERL) $(srctree)/$(src)/mkcapflags.pl $< $@
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index b0517aa2bd3..3ffdcfa9abd 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -13,6 +13,7 @@
 #include <linux/io.h>
 
 #include <asm/stackprotector.h>
+#include <asm/perf_counter.h>
 #include <asm/mmu_context.h>
 #include <asm/hypervisor.h>
 #include <asm/processor.h>
@@ -874,6 +875,7 @@ void __init identify_boot_cpu(void)
 #else
 	vgetcpu_set_mode();
 #endif
+	init_hw_perf_counters();
 }
 
 void __cpuinit identify_secondary_cpu(struct cpuinfo_x86 *c)
diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
new file mode 100644
index 00000000000..895c82e7845
--- /dev/null
+++ b/arch/x86/kernel/cpu/perf_counter.c
@@ -0,0 +1,1704 @@
+/*
+ * Performance counter x86 architecture code
+ *
+ *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
+ *  Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar
+ *  Copyright (C) 2009 Jaswinder Singh Rajput
+ *  Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter
+ *  Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ *
+ *  For licencing details see kernel-base/COPYING
+ */
+
+#include <linux/perf_counter.h>
+#include <linux/capability.h>
+#include <linux/notifier.h>
+#include <linux/hardirq.h>
+#include <linux/kprobes.h>
+#include <linux/module.h>
+#include <linux/kdebug.h>
+#include <linux/sched.h>
+#include <linux/uaccess.h>
+
+#include <asm/apic.h>
+#include <asm/stacktrace.h>
+#include <asm/nmi.h>
+
+static u64 perf_counter_mask __read_mostly;
+
+struct cpu_hw_counters {
+	struct perf_counter	*counters[X86_PMC_IDX_MAX];
+	unsigned long		used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
+	unsigned long		active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
+	unsigned long		interrupts;
+	int			enabled;
+};
+
+/*
+ * struct x86_pmu - generic x86 pmu
+ */
+struct x86_pmu {
+	const char	*name;
+	int		version;
+	int		(*handle_irq)(struct pt_regs *);
+	void		(*disable_all)(void);
+	void		(*enable_all)(void);
+	void		(*enable)(struct hw_perf_counter *, int);
+	void		(*disable)(struct hw_perf_counter *, int);
+	unsigned	eventsel;
+	unsigned	perfctr;
+	u64		(*event_map)(int);
+	u64		(*raw_event)(u64);
+	int		max_events;
+	int		num_counters;
+	int		num_counters_fixed;
+	int		counter_bits;
+	u64		counter_mask;
+	u64		max_period;
+	u64		intel_ctrl;
+};
+
+static struct x86_pmu x86_pmu __read_mostly;
+
+static DEFINE_PER_CPU(struct cpu_hw_counters, cpu_hw_counters) = {
+	.enabled = 1,
+};
+
+/*
+ * Intel PerfMon v3. Used on Core2 and later.
+ */
+static const u64 intel_perfmon_event_map[] =
+{
+  [PERF_COUNT_HW_CPU_CYCLES]		= 0x003c,
+  [PERF_COUNT_HW_INSTRUCTIONS]		= 0x00c0,
+  [PERF_COUNT_HW_CACHE_REFERENCES]	= 0x4f2e,
+  [PERF_COUNT_HW_CACHE_MISSES]		= 0x412e,
+  [PERF_COUNT_HW_BRANCH_INSTRUCTIONS]	= 0x00c4,
+  [PERF_COUNT_HW_BRANCH_MISSES]		= 0x00c5,
+  [PERF_COUNT_HW_BUS_CYCLES]		= 0x013c,
+};
+
+static u64 intel_pmu_event_map(int event)
+{
+	return intel_perfmon_event_map[event];
+}
+
+/*
+ * Generalized hw caching related event table, filled
+ * in on a per model basis. A value of 0 means
+ * 'not supported', -1 means 'event makes no sense on
+ * this CPU', any other value means the raw event
+ * ID.
+ */
+
+#define C(x) PERF_COUNT_HW_CACHE_##x
+
+static u64 __read_mostly hw_cache_event_ids
+				[PERF_COUNT_HW_CACHE_MAX]
+				[PERF_COUNT_HW_CACHE_OP_MAX]
+				[PERF_COUNT_HW_CACHE_RESULT_MAX];
+
+static const u64 nehalem_hw_cache_event_ids
+				[PERF_COUNT_HW_CACHE_MAX]
+				[PERF_COUNT_HW_CACHE_OP_MAX]
+				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI            */
+		[ C(RESULT_MISS)   ] = 0x0140, /* L1D_CACHE_LD.I_STATE         */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI            */
+		[ C(RESULT_MISS)   ] = 0x0141, /* L1D_CACHE_ST.I_STATE         */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x014e, /* L1D_PREFETCH.REQUESTS        */
+		[ C(RESULT_MISS)   ] = 0x024e, /* L1D_PREFETCH.MISS            */
+	},
+ },
+ [ C(L1I ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS                    */
+		[ C(RESULT_MISS)   ] = 0x0280, /* L1I.MISSES                   */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = 0x0,
+	},
+ },
+ [ C(LL  ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS               */
+		[ C(RESULT_MISS)   ] = 0x0224, /* L2_RQSTS.LD_MISS             */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS                */
+		[ C(RESULT_MISS)   ] = 0x0824, /* L2_RQSTS.RFO_MISS            */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference                */
+		[ C(RESULT_MISS)   ] = 0x412e, /* LLC Misses                   */
+	},
+ },
+ [ C(DTLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI   (alias)  */
+		[ C(RESULT_MISS)   ] = 0x0108, /* DTLB_LOAD_MISSES.ANY         */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI   (alias)  */
+		[ C(RESULT_MISS)   ] = 0x010c, /* MEM_STORE_RETIRED.DTLB_MISS  */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = 0x0,
+	},
+ },
+ [ C(ITLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01c0, /* INST_RETIRED.ANY_P           */
+		[ C(RESULT_MISS)   ] = 0x20c8, /* ITLB_MISS_RETIRED            */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+ [ C(BPU ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ALL_BRANCHES */
+		[ C(RESULT_MISS)   ] = 0x03e8, /* BPU_CLEARS.ANY               */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+};
+
+static const u64 core2_hw_cache_event_ids
+				[PERF_COUNT_HW_CACHE_MAX]
+				[PERF_COUNT_HW_CACHE_OP_MAX]
+				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI          */
+		[ C(RESULT_MISS)   ] = 0x0140, /* L1D_CACHE_LD.I_STATE       */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI          */
+		[ C(RESULT_MISS)   ] = 0x0141, /* L1D_CACHE_ST.I_STATE       */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x104e, /* L1D_PREFETCH.REQUESTS      */
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(L1I ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0080, /* L1I.READS                  */
+		[ C(RESULT_MISS)   ] = 0x0081, /* L1I.MISSES                 */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(LL  ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI                 */
+		[ C(RESULT_MISS)   ] = 0x4129, /* L2_LD.ISTATE               */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI                 */
+		[ C(RESULT_MISS)   ] = 0x412A, /* L2_ST.ISTATE               */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(DTLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI  (alias) */
+		[ C(RESULT_MISS)   ] = 0x0208, /* DTLB_MISSES.MISS_LD        */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI  (alias) */
+		[ C(RESULT_MISS)   ] = 0x0808, /* DTLB_MISSES.MISS_ST        */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(ITLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P         */
+		[ C(RESULT_MISS)   ] = 0x1282, /* ITLBMISSES                 */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+ [ C(BPU ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY        */
+		[ C(RESULT_MISS)   ] = 0x00c5, /* BP_INST_RETIRED.MISPRED    */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+};
+
+static const u64 atom_hw_cache_event_ids
+				[PERF_COUNT_HW_CACHE_MAX]
+				[PERF_COUNT_HW_CACHE_OP_MAX]
+				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE.LD               */
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE.ST               */
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(L1I ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS                  */
+		[ C(RESULT_MISS)   ] = 0x0280, /* L1I.MISSES                 */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(LL  ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI                 */
+		[ C(RESULT_MISS)   ] = 0x4129, /* L2_LD.ISTATE               */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI                 */
+		[ C(RESULT_MISS)   ] = 0x412A, /* L2_ST.ISTATE               */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(DTLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE_LD.MESI  (alias) */
+		[ C(RESULT_MISS)   ] = 0x0508, /* DTLB_MISSES.MISS_LD        */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE_ST.MESI  (alias) */
+		[ C(RESULT_MISS)   ] = 0x0608, /* DTLB_MISSES.MISS_ST        */
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(ITLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P         */
+		[ C(RESULT_MISS)   ] = 0x0282, /* ITLB.MISSES                */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+ [ C(BPU ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY        */
+		[ C(RESULT_MISS)   ] = 0x00c5, /* BP_INST_RETIRED.MISPRED    */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+};
+
+static u64 intel_pmu_raw_event(u64 event)
+{
+#define CORE_EVNTSEL_EVENT_MASK		0x000000FFULL
+#define CORE_EVNTSEL_UNIT_MASK		0x0000FF00ULL
+#define CORE_EVNTSEL_EDGE_MASK		0x00040000ULL
+#define CORE_EVNTSEL_INV_MASK		0x00800000ULL
+#define CORE_EVNTSEL_COUNTER_MASK	0xFF000000ULL
+
+#define CORE_EVNTSEL_MASK		\
+	(CORE_EVNTSEL_EVENT_MASK |	\
+	 CORE_EVNTSEL_UNIT_MASK  |	\
+	 CORE_EVNTSEL_EDGE_MASK  |	\
+	 CORE_EVNTSEL_INV_MASK  |	\
+	 CORE_EVNTSEL_COUNTER_MASK)
+
+	return event & CORE_EVNTSEL_MASK;
+}
+
+static const u64 amd_0f_hw_cache_event_ids
+				[PERF_COUNT_HW_CACHE_MAX]
+				[PERF_COUNT_HW_CACHE_OP_MAX]
+				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+ [ C(L1D) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(L1I ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0080, /* Instruction cache fetches  */
+		[ C(RESULT_MISS)   ] = 0x0081, /* Instruction cache misses   */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(LL  ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(DTLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0,
+		[ C(RESULT_MISS)   ] = 0,
+	},
+ },
+ [ C(ITLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0080, /* Instruction fecthes        */
+		[ C(RESULT_MISS)   ] = 0x0085, /* Instr. fetch ITLB misses   */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+ [ C(BPU ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x00c2, /* Retired Branch Instr.      */
+		[ C(RESULT_MISS)   ] = 0x00c3, /* Retired Mispredicted BI    */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+};
+
+/*
+ * AMD Performance Monitor K7 and later.
+ */
+static const u64 amd_perfmon_event_map[] =
+{
+  [PERF_COUNT_HW_CPU_CYCLES]		= 0x0076,
+  [PERF_COUNT_HW_INSTRUCTIONS]		= 0x00c0,
+  [PERF_COUNT_HW_CACHE_REFERENCES]	= 0x0080,
+  [PERF_COUNT_HW_CACHE_MISSES]		= 0x0081,
+  [PERF_COUNT_HW_BRANCH_INSTRUCTIONS]	= 0x00c4,
+  [PERF_COUNT_HW_BRANCH_MISSES]		= 0x00c5,
+};
+
+static u64 amd_pmu_event_map(int event)
+{
+	return amd_perfmon_event_map[event];
+}
+
+static u64 amd_pmu_raw_event(u64 event)
+{
+#define K7_EVNTSEL_EVENT_MASK	0x7000000FFULL
+#define K7_EVNTSEL_UNIT_MASK	0x00000FF00ULL
+#define K7_EVNTSEL_EDGE_MASK	0x000040000ULL
+#define K7_EVNTSEL_INV_MASK	0x000800000ULL
+#define K7_EVNTSEL_COUNTER_MASK	0x0FF000000ULL
+
+#define K7_EVNTSEL_MASK			\
+	(K7_EVNTSEL_EVENT_MASK |	\
+	 K7_EVNTSEL_UNIT_MASK  |	\
+	 K7_EVNTSEL_EDGE_MASK  |	\
+	 K7_EVNTSEL_INV_MASK   |	\
+	 K7_EVNTSEL_COUNTER_MASK)
+
+	return event & K7_EVNTSEL_MASK;
+}
+
+/*
+ * Propagate counter elapsed time into the generic counter.
+ * Can only be executed on the CPU where the counter is active.
+ * Returns the delta events processed.
+ */
+static u64
+x86_perf_counter_update(struct perf_counter *counter,
+			struct hw_perf_counter *hwc, int idx)
+{
+	int shift = 64 - x86_pmu.counter_bits;
+	u64 prev_raw_count, new_raw_count;
+	s64 delta;
+
+	/*
+	 * Careful: an NMI might modify the previous counter value.
+	 *
+	 * Our tactic to handle this is to first atomically read and
+	 * exchange a new raw count - then add that new-prev delta
+	 * count to the generic counter atomically:
+	 */
+again:
+	prev_raw_count = atomic64_read(&hwc->prev_count);
+	rdmsrl(hwc->counter_base + idx, new_raw_count);
+
+	if (atomic64_cmpxchg(&hwc->prev_count, prev_raw_count,
+					new_raw_count) != prev_raw_count)
+		goto again;
+
+	/*
+	 * Now we have the new raw value and have updated the prev
+	 * timestamp already. We can now calculate the elapsed delta
+	 * (counter-)time and add that to the generic counter.
+	 *
+	 * Careful, not all hw sign-extends above the physical width
+	 * of the count.
+	 */
+	delta = (new_raw_count << shift) - (prev_raw_count << shift);
+	delta >>= shift;
+
+	atomic64_add(delta, &counter->count);
+	atomic64_sub(delta, &hwc->period_left);
+
+	return new_raw_count;
+}
+
+static atomic_t active_counters;
+static DEFINE_MUTEX(pmc_reserve_mutex);
+
+static bool reserve_pmc_hardware(void)
+{
+	int i;
+
+	if (nmi_watchdog == NMI_LOCAL_APIC)
+		disable_lapic_nmi_watchdog();
+
+	for (i = 0; i < x86_pmu.num_counters; i++) {
+		if (!reserve_perfctr_nmi(x86_pmu.perfctr + i))
+			goto perfctr_fail;
+	}
+
+	for (i = 0; i < x86_pmu.num_counters; i++) {
+		if (!reserve_evntsel_nmi(x86_pmu.eventsel + i))
+			goto eventsel_fail;
+	}
+
+	return true;
+
+eventsel_fail:
+	for (i--; i >= 0; i--)
+		release_evntsel_nmi(x86_pmu.eventsel + i);
+
+	i = x86_pmu.num_counters;
+
+perfctr_fail:
+	for (i--; i >= 0; i--)
+		release_perfctr_nmi(x86_pmu.perfctr + i);
+
+	if (nmi_watchdog == NMI_LOCAL_APIC)
+		enable_lapic_nmi_watchdog();
+
+	return false;
+}
+
+static void release_pmc_hardware(void)
+{
+	int i;
+
+	for (i = 0; i < x86_pmu.num_counters; i++) {
+		release_perfctr_nmi(x86_pmu.perfctr + i);
+		release_evntsel_nmi(x86_pmu.eventsel + i);
+	}
+
+	if (nmi_watchdog == NMI_LOCAL_APIC)
+		enable_lapic_nmi_watchdog();
+}
+
+static void hw_perf_counter_destroy(struct perf_counter *counter)
+{
+	if (atomic_dec_and_mutex_lock(&active_counters, &pmc_reserve_mutex)) {
+		release_pmc_hardware();
+		mutex_unlock(&pmc_reserve_mutex);
+	}
+}
+
+static inline int x86_pmu_initialized(void)
+{
+	return x86_pmu.handle_irq != NULL;
+}
+
+static inline int
+set_ext_hw_attr(struct hw_perf_counter *hwc, struct perf_counter_attr *attr)
+{
+	unsigned int cache_type, cache_op, cache_result;
+	u64 config, val;
+
+	config = attr->config;
+
+	cache_type = (config >>  0) & 0xff;
+	if (cache_type >= PERF_COUNT_HW_CACHE_MAX)
+		return -EINVAL;
+
+	cache_op = (config >>  8) & 0xff;
+	if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX)
+		return -EINVAL;
+
+	cache_result = (config >> 16) & 0xff;
+	if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
+		return -EINVAL;
+
+	val = hw_cache_event_ids[cache_type][cache_op][cache_result];
+
+	if (val == 0)
+		return -ENOENT;
+
+	if (val == -1)
+		return -EINVAL;
+
+	hwc->config |= val;
+
+	return 0;
+}
+
+/*
+ * Setup the hardware configuration for a given attr_type
+ */
+static int __hw_perf_counter_init(struct perf_counter *counter)
+{
+	struct perf_counter_attr *attr = &counter->attr;
+	struct hw_perf_counter *hwc = &counter->hw;
+	int err;
+
+	if (!x86_pmu_initialized())
+		return -ENODEV;
+
+	err = 0;
+	if (!atomic_inc_not_zero(&active_counters)) {
+		mutex_lock(&pmc_reserve_mutex);
+		if (atomic_read(&active_counters) == 0 && !reserve_pmc_hardware())
+			err = -EBUSY;
+		else
+			atomic_inc(&active_counters);
+		mutex_unlock(&pmc_reserve_mutex);
+	}
+	if (err)
+		return err;
+
+	/*
+	 * Generate PMC IRQs:
+	 * (keep 'enabled' bit clear for now)
+	 */
+	hwc->config = ARCH_PERFMON_EVENTSEL_INT;
+
+	/*
+	 * Count user and OS events unless requested not to.
+	 */
+	if (!attr->exclude_user)
+		hwc->config |= ARCH_PERFMON_EVENTSEL_USR;
+	if (!attr->exclude_kernel)
+		hwc->config |= ARCH_PERFMON_EVENTSEL_OS;
+
+	if (!hwc->sample_period) {
+		hwc->sample_period = x86_pmu.max_period;
+		hwc->last_period = hwc->sample_period;
+		atomic64_set(&hwc->period_left, hwc->sample_period);
+	}
+
+	counter->destroy = hw_perf_counter_destroy;
+
+	/*
+	 * Raw event type provide the config in the event structure
+	 */
+	if (attr->type == PERF_TYPE_RAW) {
+		hwc->config |= x86_pmu.raw_event(attr->config);
+		return 0;
+	}
+
+	if (attr->type == PERF_TYPE_HW_CACHE)
+		return set_ext_hw_attr(hwc, attr);
+
+	if (attr->config >= x86_pmu.max_events)
+		return -EINVAL;
+	/*
+	 * The generic map:
+	 */
+	hwc->config |= x86_pmu.event_map(attr->config);
+
+	return 0;
+}
+
+static void intel_pmu_disable_all(void)
+{
+	wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0);
+}
+
+static void amd_pmu_disable_all(void)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+	int idx;
+
+	if (!cpuc->enabled)
+		return;
+
+	cpuc->enabled = 0;
+	/*
+	 * ensure we write the disable before we start disabling the
+	 * counters proper, so that amd_pmu_enable_counter() does the
+	 * right thing.
+	 */
+	barrier();
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		u64 val;
+
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+		rdmsrl(MSR_K7_EVNTSEL0 + idx, val);
+		if (!(val & ARCH_PERFMON_EVENTSEL0_ENABLE))
+			continue;
+		val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE;
+		wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
+	}
+}
+
+void hw_perf_disable(void)
+{
+	if (!x86_pmu_initialized())
+		return;
+	return x86_pmu.disable_all();
+}
+
+static void intel_pmu_enable_all(void)
+{
+	wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl);
+}
+
+static void amd_pmu_enable_all(void)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+	int idx;
+
+	if (cpuc->enabled)
+		return;
+
+	cpuc->enabled = 1;
+	barrier();
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		u64 val;
+
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+		rdmsrl(MSR_K7_EVNTSEL0 + idx, val);
+		if (val & ARCH_PERFMON_EVENTSEL0_ENABLE)
+			continue;
+		val |= ARCH_PERFMON_EVENTSEL0_ENABLE;
+		wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
+	}
+}
+
+void hw_perf_enable(void)
+{
+	if (!x86_pmu_initialized())
+		return;
+	x86_pmu.enable_all();
+}
+
+static inline u64 intel_pmu_get_status(void)
+{
+	u64 status;
+
+	rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status);
+
+	return status;
+}
+
+static inline void intel_pmu_ack_status(u64 ack)
+{
+	wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, ack);
+}
+
+static inline void x86_pmu_enable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	int err;
+	err = checking_wrmsrl(hwc->config_base + idx,
+			      hwc->config | ARCH_PERFMON_EVENTSEL0_ENABLE);
+}
+
+static inline void x86_pmu_disable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	int err;
+	err = checking_wrmsrl(hwc->config_base + idx,
+			      hwc->config);
+}
+
+static inline void
+intel_pmu_disable_fixed(struct hw_perf_counter *hwc, int __idx)
+{
+	int idx = __idx - X86_PMC_IDX_FIXED;
+	u64 ctrl_val, mask;
+	int err;
+
+	mask = 0xfULL << (idx * 4);
+
+	rdmsrl(hwc->config_base, ctrl_val);
+	ctrl_val &= ~mask;
+	err = checking_wrmsrl(hwc->config_base, ctrl_val);
+}
+
+static inline void
+intel_pmu_disable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
+		intel_pmu_disable_fixed(hwc, idx);
+		return;
+	}
+
+	x86_pmu_disable_counter(hwc, idx);
+}
+
+static inline void
+amd_pmu_disable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	x86_pmu_disable_counter(hwc, idx);
+}
+
+static DEFINE_PER_CPU(u64, prev_left[X86_PMC_IDX_MAX]);
+
+/*
+ * Set the next IRQ period, based on the hwc->period_left value.
+ * To be called with the counter disabled in hw:
+ */
+static int
+x86_perf_counter_set_period(struct perf_counter *counter,
+			     struct hw_perf_counter *hwc, int idx)
+{
+	s64 left = atomic64_read(&hwc->period_left);
+	s64 period = hwc->sample_period;
+	int err, ret = 0;
+
+	/*
+	 * If we are way outside a reasoable range then just skip forward:
+	 */
+	if (unlikely(left <= -period)) {
+		left = period;
+		atomic64_set(&hwc->period_left, left);
+		hwc->last_period = period;
+		ret = 1;
+	}
+
+	if (unlikely(left <= 0)) {
+		left += period;
+		atomic64_set(&hwc->period_left, left);
+		hwc->last_period = period;
+		ret = 1;
+	}
+	/*
+	 * Quirk: certain CPUs dont like it if just 1 event is left:
+	 */
+	if (unlikely(left < 2))
+		left = 2;
+
+	if (left > x86_pmu.max_period)
+		left = x86_pmu.max_period;
+
+	per_cpu(prev_left[idx], smp_processor_id()) = left;
+
+	/*
+	 * The hw counter starts counting from this counter offset,
+	 * mark it to be able to extra future deltas:
+	 */
+	atomic64_set(&hwc->prev_count, (u64)-left);
+
+	err = checking_wrmsrl(hwc->counter_base + idx,
+			     (u64)(-left) & x86_pmu.counter_mask);
+
+	return ret;
+}
+
+static inline void
+intel_pmu_enable_fixed(struct hw_perf_counter *hwc, int __idx)
+{
+	int idx = __idx - X86_PMC_IDX_FIXED;
+	u64 ctrl_val, bits, mask;
+	int err;
+
+	/*
+	 * Enable IRQ generation (0x8),
+	 * and enable ring-3 counting (0x2) and ring-0 counting (0x1)
+	 * if requested:
+	 */
+	bits = 0x8ULL;
+	if (hwc->config & ARCH_PERFMON_EVENTSEL_USR)
+		bits |= 0x2;
+	if (hwc->config & ARCH_PERFMON_EVENTSEL_OS)
+		bits |= 0x1;
+	bits <<= (idx * 4);
+	mask = 0xfULL << (idx * 4);
+
+	rdmsrl(hwc->config_base, ctrl_val);
+	ctrl_val &= ~mask;
+	ctrl_val |= bits;
+	err = checking_wrmsrl(hwc->config_base, ctrl_val);
+}
+
+static void intel_pmu_enable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
+		intel_pmu_enable_fixed(hwc, idx);
+		return;
+	}
+
+	x86_pmu_enable_counter(hwc, idx);
+}
+
+static void amd_pmu_enable_counter(struct hw_perf_counter *hwc, int idx)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+
+	if (cpuc->enabled)
+		x86_pmu_enable_counter(hwc, idx);
+	else
+		x86_pmu_disable_counter(hwc, idx);
+}
+
+static int
+fixed_mode_idx(struct perf_counter *counter, struct hw_perf_counter *hwc)
+{
+	unsigned int event;
+
+	if (!x86_pmu.num_counters_fixed)
+		return -1;
+
+	event = hwc->config & ARCH_PERFMON_EVENT_MASK;
+
+	if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_INSTRUCTIONS)))
+		return X86_PMC_IDX_FIXED_INSTRUCTIONS;
+	if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_CPU_CYCLES)))
+		return X86_PMC_IDX_FIXED_CPU_CYCLES;
+	if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_BUS_CYCLES)))
+		return X86_PMC_IDX_FIXED_BUS_CYCLES;
+
+	return -1;
+}
+
+/*
+ * Find a PMC slot for the freshly enabled / scheduled in counter:
+ */
+static int x86_pmu_enable(struct perf_counter *counter)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+	struct hw_perf_counter *hwc = &counter->hw;
+	int idx;
+
+	idx = fixed_mode_idx(counter, hwc);
+	if (idx >= 0) {
+		/*
+		 * Try to get the fixed counter, if that is already taken
+		 * then try to get a generic counter:
+		 */
+		if (test_and_set_bit(idx, cpuc->used_mask))
+			goto try_generic;
+
+		hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL;
+		/*
+		 * We set it so that counter_base + idx in wrmsr/rdmsr maps to
+		 * MSR_ARCH_PERFMON_FIXED_CTR0 ... CTR2:
+		 */
+		hwc->counter_base =
+			MSR_ARCH_PERFMON_FIXED_CTR0 - X86_PMC_IDX_FIXED;
+		hwc->idx = idx;
+	} else {
+		idx = hwc->idx;
+		/* Try to get the previous generic counter again */
+		if (test_and_set_bit(idx, cpuc->used_mask)) {
+try_generic:
+			idx = find_first_zero_bit(cpuc->used_mask,
+						  x86_pmu.num_counters);
+			if (idx == x86_pmu.num_counters)
+				return -EAGAIN;
+
+			set_bit(idx, cpuc->used_mask);
+			hwc->idx = idx;
+		}
+		hwc->config_base  = x86_pmu.eventsel;
+		hwc->counter_base = x86_pmu.perfctr;
+	}
+
+	perf_counters_lapic_init();
+
+	x86_pmu.disable(hwc, idx);
+
+	cpuc->counters[idx] = counter;
+	set_bit(idx, cpuc->active_mask);
+
+	x86_perf_counter_set_period(counter, hwc, idx);
+	x86_pmu.enable(hwc, idx);
+
+	return 0;
+}
+
+static void x86_pmu_unthrottle(struct perf_counter *counter)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+	struct hw_perf_counter *hwc = &counter->hw;
+
+	if (WARN_ON_ONCE(hwc->idx >= X86_PMC_IDX_MAX ||
+				cpuc->counters[hwc->idx] != counter))
+		return;
+
+	x86_pmu.enable(hwc, hwc->idx);
+}
+
+void perf_counter_print_debug(void)
+{
+	u64 ctrl, status, overflow, pmc_ctrl, pmc_count, prev_left, fixed;
+	struct cpu_hw_counters *cpuc;
+	unsigned long flags;
+	int cpu, idx;
+
+	if (!x86_pmu.num_counters)
+		return;
+
+	local_irq_save(flags);
+
+	cpu = smp_processor_id();
+	cpuc = &per_cpu(cpu_hw_counters, cpu);
+
+	if (x86_pmu.version >= 2) {
+		rdmsrl(MSR_CORE_PERF_GLOBAL_CTRL, ctrl);
+		rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status);
+		rdmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, overflow);
+		rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR_CTRL, fixed);
+
+		pr_info("\n");
+		pr_info("CPU#%d: ctrl:       %016llx\n", cpu, ctrl);
+		pr_info("CPU#%d: status:     %016llx\n", cpu, status);
+		pr_info("CPU#%d: overflow:   %016llx\n", cpu, overflow);
+		pr_info("CPU#%d: fixed:      %016llx\n", cpu, fixed);
+	}
+	pr_info("CPU#%d: used:       %016llx\n", cpu, *(u64 *)cpuc->used_mask);
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		rdmsrl(x86_pmu.eventsel + idx, pmc_ctrl);
+		rdmsrl(x86_pmu.perfctr  + idx, pmc_count);
+
+		prev_left = per_cpu(prev_left[idx], cpu);
+
+		pr_info("CPU#%d:   gen-PMC%d ctrl:  %016llx\n",
+			cpu, idx, pmc_ctrl);
+		pr_info("CPU#%d:   gen-PMC%d count: %016llx\n",
+			cpu, idx, pmc_count);
+		pr_info("CPU#%d:   gen-PMC%d left:  %016llx\n",
+			cpu, idx, prev_left);
+	}
+	for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++) {
+		rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, pmc_count);
+
+		pr_info("CPU#%d: fixed-PMC%d count: %016llx\n",
+			cpu, idx, pmc_count);
+	}
+	local_irq_restore(flags);
+}
+
+static void x86_pmu_disable(struct perf_counter *counter)
+{
+	struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters);
+	struct hw_perf_counter *hwc = &counter->hw;
+	int idx = hwc->idx;
+
+	/*
+	 * Must be done before we disable, otherwise the nmi handler
+	 * could reenable again:
+	 */
+	clear_bit(idx, cpuc->active_mask);
+	x86_pmu.disable(hwc, idx);
+
+	/*
+	 * Make sure the cleared pointer becomes visible before we
+	 * (potentially) free the counter:
+	 */
+	barrier();
+
+	/*
+	 * Drain the remaining delta count out of a counter
+	 * that we are disabling:
+	 */
+	x86_perf_counter_update(counter, hwc, idx);
+	cpuc->counters[idx] = NULL;
+	clear_bit(idx, cpuc->used_mask);
+}
+
+/*
+ * Save and restart an expired counter. Called by NMI contexts,
+ * so it has to be careful about preempting normal counter ops:
+ */
+static int intel_pmu_save_and_restart(struct perf_counter *counter)
+{
+	struct hw_perf_counter *hwc = &counter->hw;
+	int idx = hwc->idx;
+	int ret;
+
+	x86_perf_counter_update(counter, hwc, idx);
+	ret = x86_perf_counter_set_period(counter, hwc, idx);
+
+	if (counter->state == PERF_COUNTER_STATE_ACTIVE)
+		intel_pmu_enable_counter(hwc, idx);
+
+	return ret;
+}
+
+static void intel_pmu_reset(void)
+{
+	unsigned long flags;
+	int idx;
+
+	if (!x86_pmu.num_counters)
+		return;
+
+	local_irq_save(flags);
+
+	printk("clearing PMU state on CPU#%d\n", smp_processor_id());
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		checking_wrmsrl(x86_pmu.eventsel + idx, 0ull);
+		checking_wrmsrl(x86_pmu.perfctr  + idx, 0ull);
+	}
+	for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++) {
+		checking_wrmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, 0ull);
+	}
+
+	local_irq_restore(flags);
+}
+
+
+/*
+ * This handler is triggered by the local APIC, so the APIC IRQ handling
+ * rules apply:
+ */
+static int intel_pmu_handle_irq(struct pt_regs *regs)
+{
+	struct perf_sample_data data;
+	struct cpu_hw_counters *cpuc;
+	int bit, cpu, loops;
+	u64 ack, status;
+
+	data.regs = regs;
+	data.addr = 0;
+
+	cpu = smp_processor_id();
+	cpuc = &per_cpu(cpu_hw_counters, cpu);
+
+	perf_disable();
+	status = intel_pmu_get_status();
+	if (!status) {
+		perf_enable();
+		return 0;
+	}
+
+	loops = 0;
+again:
+	if (++loops > 100) {
+		WARN_ONCE(1, "perfcounters: irq loop stuck!\n");
+		perf_counter_print_debug();
+		intel_pmu_reset();
+		perf_enable();
+		return 1;
+	}
+
+	inc_irq_stat(apic_perf_irqs);
+	ack = status;
+	for_each_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) {
+		struct perf_counter *counter = cpuc->counters[bit];
+
+		clear_bit(bit, (unsigned long *) &status);
+		if (!test_bit(bit, cpuc->active_mask))
+			continue;
+
+		if (!intel_pmu_save_and_restart(counter))
+			continue;
+
+		if (perf_counter_overflow(counter, 1, &data))
+			intel_pmu_disable_counter(&counter->hw, bit);
+	}
+
+	intel_pmu_ack_status(ack);
+
+	/*
+	 * Repeat if there is more work to be done:
+	 */
+	status = intel_pmu_get_status();
+	if (status)
+		goto again;
+
+	perf_enable();
+
+	return 1;
+}
+
+static int amd_pmu_handle_irq(struct pt_regs *regs)
+{
+	struct perf_sample_data data;
+	struct cpu_hw_counters *cpuc;
+	struct perf_counter *counter;
+	struct hw_perf_counter *hwc;
+	int cpu, idx, handled = 0;
+	u64 val;
+
+	data.regs = regs;
+	data.addr = 0;
+
+	cpu = smp_processor_id();
+	cpuc = &per_cpu(cpu_hw_counters, cpu);
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+
+		counter = cpuc->counters[idx];
+		hwc = &counter->hw;
+
+		val = x86_perf_counter_update(counter, hwc, idx);
+		if (val & (1ULL << (x86_pmu.counter_bits - 1)))
+			continue;
+
+		/*
+		 * counter overflow
+		 */
+		handled		= 1;
+		data.period	= counter->hw.last_period;
+
+		if (!x86_perf_counter_set_period(counter, hwc, idx))
+			continue;
+
+		if (perf_counter_overflow(counter, 1, &data))
+			amd_pmu_disable_counter(hwc, idx);
+	}
+
+	if (handled)
+		inc_irq_stat(apic_perf_irqs);
+
+	return handled;
+}
+
+void smp_perf_pending_interrupt(struct pt_regs *regs)
+{
+	irq_enter();
+	ack_APIC_irq();
+	inc_irq_stat(apic_pending_irqs);
+	perf_counter_do_pending();
+	irq_exit();
+}
+
+void set_perf_counter_pending(void)
+{
+	apic->send_IPI_self(LOCAL_PENDING_VECTOR);
+}
+
+void perf_counters_lapic_init(void)
+{
+	if (!x86_pmu_initialized())
+		return;
+
+	/*
+	 * Always use NMI for PMU
+	 */
+	apic_write(APIC_LVTPC, APIC_DM_NMI);
+}
+
+static int __kprobes
+perf_counter_nmi_handler(struct notifier_block *self,
+			 unsigned long cmd, void *__args)
+{
+	struct die_args *args = __args;
+	struct pt_regs *regs;
+
+	if (!atomic_read(&active_counters))
+		return NOTIFY_DONE;
+
+	switch (cmd) {
+	case DIE_NMI:
+	case DIE_NMI_IPI:
+		break;
+
+	default:
+		return NOTIFY_DONE;
+	}
+
+	regs = args->regs;
+
+	apic_write(APIC_LVTPC, APIC_DM_NMI);
+	/*
+	 * Can't rely on the handled return value to say it was our NMI, two
+	 * counters could trigger 'simultaneously' raising two back-to-back NMIs.
+	 *
+	 * If the first NMI handles both, the latter will be empty and daze
+	 * the CPU.
+	 */
+	x86_pmu.handle_irq(regs);
+
+	return NOTIFY_STOP;
+}
+
+static __read_mostly struct notifier_block perf_counter_nmi_notifier = {
+	.notifier_call		= perf_counter_nmi_handler,
+	.next			= NULL,
+	.priority		= 1
+};
+
+static struct x86_pmu intel_pmu = {
+	.name			= "Intel",
+	.handle_irq		= intel_pmu_handle_irq,
+	.disable_all		= intel_pmu_disable_all,
+	.enable_all		= intel_pmu_enable_all,
+	.enable			= intel_pmu_enable_counter,
+	.disable		= intel_pmu_disable_counter,
+	.eventsel		= MSR_ARCH_PERFMON_EVENTSEL0,
+	.perfctr		= MSR_ARCH_PERFMON_PERFCTR0,
+	.event_map		= intel_pmu_event_map,
+	.raw_event		= intel_pmu_raw_event,
+	.max_events		= ARRAY_SIZE(intel_perfmon_event_map),
+	/*
+	 * Intel PMCs cannot be accessed sanely above 32 bit width,
+	 * so we install an artificial 1<<31 period regardless of
+	 * the generic counter period:
+	 */
+	.max_period		= (1ULL << 31) - 1,
+};
+
+static struct x86_pmu amd_pmu = {
+	.name			= "AMD",
+	.handle_irq		= amd_pmu_handle_irq,
+	.disable_all		= amd_pmu_disable_all,
+	.enable_all		= amd_pmu_enable_all,
+	.enable			= amd_pmu_enable_counter,
+	.disable		= amd_pmu_disable_counter,
+	.eventsel		= MSR_K7_EVNTSEL0,
+	.perfctr		= MSR_K7_PERFCTR0,
+	.event_map		= amd_pmu_event_map,
+	.raw_event		= amd_pmu_raw_event,
+	.max_events		= ARRAY_SIZE(amd_perfmon_event_map),
+	.num_counters		= 4,
+	.counter_bits		= 48,
+	.counter_mask		= (1ULL << 48) - 1,
+	/* use highest bit to detect overflow */
+	.max_period		= (1ULL << 47) - 1,
+};
+
+static int intel_pmu_init(void)
+{
+	union cpuid10_edx edx;
+	union cpuid10_eax eax;
+	unsigned int unused;
+	unsigned int ebx;
+	int version;
+
+	if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON))
+		return -ENODEV;
+
+	/*
+	 * Check whether the Architectural PerfMon supports
+	 * Branch Misses Retired Event or not.
+	 */
+	cpuid(10, &eax.full, &ebx, &unused, &edx.full);
+	if (eax.split.mask_length <= ARCH_PERFMON_BRANCH_MISSES_RETIRED)
+		return -ENODEV;
+
+	version = eax.split.version_id;
+	if (version < 2)
+		return -ENODEV;
+
+	x86_pmu				= intel_pmu;
+	x86_pmu.version			= version;
+	x86_pmu.num_counters		= eax.split.num_counters;
+	x86_pmu.counter_bits		= eax.split.bit_width;
+	x86_pmu.counter_mask		= (1ULL << eax.split.bit_width) - 1;
+
+	/*
+	 * Quirk: v2 perfmon does not report fixed-purpose counters, so
+	 * assume at least 3 counters:
+	 */
+	x86_pmu.num_counters_fixed	= max((int)edx.split.num_counters_fixed, 3);
+
+	rdmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl);
+
+	/*
+	 * Install the hw-cache-events table:
+	 */
+	switch (boot_cpu_data.x86_model) {
+	case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */
+	case 22: /* single-core 65 nm celeron/core2solo "Merom-L"/"Conroe-L" */
+	case 23: /* current 45 nm celeron/core2/xeon "Penryn"/"Wolfdale" */
+	case 29: /* six-core 45 nm xeon "Dunnington" */
+		memcpy(hw_cache_event_ids, core2_hw_cache_event_ids,
+		       sizeof(hw_cache_event_ids));
+
+		pr_cont("Core2 events, ");
+		break;
+	default:
+	case 26:
+		memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
+		       sizeof(hw_cache_event_ids));
+
+		pr_cont("Nehalem/Corei7 events, ");
+		break;
+	case 28:
+		memcpy(hw_cache_event_ids, atom_hw_cache_event_ids,
+		       sizeof(hw_cache_event_ids));
+
+		pr_cont("Atom events, ");
+		break;
+	}
+	return 0;
+}
+
+static int amd_pmu_init(void)
+{
+	x86_pmu = amd_pmu;
+
+	switch (boot_cpu_data.x86) {
+	case 0x0f:
+	case 0x10:
+	case 0x11:
+		memcpy(hw_cache_event_ids, amd_0f_hw_cache_event_ids,
+		       sizeof(hw_cache_event_ids));
+
+		pr_cont("AMD Family 0f/10/11 events, ");
+		break;
+	}
+	return 0;
+}
+
+void __init init_hw_perf_counters(void)
+{
+	int err;
+
+	pr_info("Performance Counters: ");
+
+	switch (boot_cpu_data.x86_vendor) {
+	case X86_VENDOR_INTEL:
+		err = intel_pmu_init();
+		break;
+	case X86_VENDOR_AMD:
+		err = amd_pmu_init();
+		break;
+	default:
+		return;
+	}
+	if (err != 0) {
+		pr_cont("no PMU driver, software counters only.\n");
+		return;
+	}
+
+	pr_cont("%s PMU driver.\n", x86_pmu.name);
+
+	if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) {
+		x86_pmu.num_counters = X86_PMC_MAX_GENERIC;
+		WARN(1, KERN_ERR "hw perf counters %d > max(%d), clipping!",
+		     x86_pmu.num_counters, X86_PMC_MAX_GENERIC);
+	}
+	perf_counter_mask = (1 << x86_pmu.num_counters) - 1;
+	perf_max_counters = x86_pmu.num_counters;
+
+	if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) {
+		x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED;
+		WARN(1, KERN_ERR "hw perf counters fixed %d > max(%d), clipping!",
+		     x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED);
+	}
+
+	perf_counter_mask |=
+		((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED;
+
+	perf_counters_lapic_init();
+	register_die_notifier(&perf_counter_nmi_notifier);
+
+	pr_info("... version:                 %d\n",     x86_pmu.version);
+	pr_info("... bit width:               %d\n",     x86_pmu.counter_bits);
+	pr_info("... generic counters:        %d\n",     x86_pmu.num_counters);
+	pr_info("... value mask:              %016Lx\n", x86_pmu.counter_mask);
+	pr_info("... max period:              %016Lx\n", x86_pmu.max_period);
+	pr_info("... fixed-purpose counters:  %d\n",     x86_pmu.num_counters_fixed);
+	pr_info("... counter mask:            %016Lx\n", perf_counter_mask);
+}
+
+static inline void x86_pmu_read(struct perf_counter *counter)
+{
+	x86_perf_counter_update(counter, &counter->hw, counter->hw.idx);
+}
+
+static const struct pmu pmu = {
+	.enable		= x86_pmu_enable,
+	.disable	= x86_pmu_disable,
+	.read		= x86_pmu_read,
+	.unthrottle	= x86_pmu_unthrottle,
+};
+
+const struct pmu *hw_perf_counter_init(struct perf_counter *counter)
+{
+	int err;
+
+	err = __hw_perf_counter_init(counter);
+	if (err)
+		return ERR_PTR(err);
+
+	return &pmu;
+}
+
+/*
+ * callchain support
+ */
+
+static inline
+void callchain_store(struct perf_callchain_entry *entry, unsigned long ip)
+{
+	if (entry->nr < MAX_STACK_DEPTH)
+		entry->ip[entry->nr++] = ip;
+}
+
+static DEFINE_PER_CPU(struct perf_callchain_entry, irq_entry);
+static DEFINE_PER_CPU(struct perf_callchain_entry, nmi_entry);
+
+
+static void
+backtrace_warning_symbol(void *data, char *msg, unsigned long symbol)
+{
+	/* Ignore warnings */
+}
+
+static void backtrace_warning(void *data, char *msg)
+{
+	/* Ignore warnings */
+}
+
+static int backtrace_stack(void *data, char *name)
+{
+	/* Don't bother with IRQ stacks for now */
+	return -1;
+}
+
+static void backtrace_address(void *data, unsigned long addr, int reliable)
+{
+	struct perf_callchain_entry *entry = data;
+
+	if (reliable)
+		callchain_store(entry, addr);
+}
+
+static const struct stacktrace_ops backtrace_ops = {
+	.warning		= backtrace_warning,
+	.warning_symbol		= backtrace_warning_symbol,
+	.stack			= backtrace_stack,
+	.address		= backtrace_address,
+};
+
+static void
+perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry)
+{
+	unsigned long bp;
+	char *stack;
+	int nr = entry->nr;
+
+	callchain_store(entry, instruction_pointer(regs));
+
+	stack = ((char *)regs + sizeof(struct pt_regs));
+#ifdef CONFIG_FRAME_POINTER
+	bp = frame_pointer(regs);
+#else
+	bp = 0;
+#endif
+
+	dump_trace(NULL, regs, (void *)stack, bp, &backtrace_ops, entry);
+
+	entry->kernel = entry->nr - nr;
+}
+
+
+struct stack_frame {
+	const void __user	*next_fp;
+	unsigned long		return_address;
+};
+
+static int copy_stack_frame(const void __user *fp, struct stack_frame *frame)
+{
+	int ret;
+
+	if (!access_ok(VERIFY_READ, fp, sizeof(*frame)))
+		return 0;
+
+	ret = 1;
+	pagefault_disable();
+	if (__copy_from_user_inatomic(frame, fp, sizeof(*frame)))
+		ret = 0;
+	pagefault_enable();
+
+	return ret;
+}
+
+static void
+perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
+{
+	struct stack_frame frame;
+	const void __user *fp;
+	int nr = entry->nr;
+
+	regs = (struct pt_regs *)current->thread.sp0 - 1;
+	fp   = (void __user *)regs->bp;
+
+	callchain_store(entry, regs->ip);
+
+	while (entry->nr < MAX_STACK_DEPTH) {
+		frame.next_fp	     = NULL;
+		frame.return_address = 0;
+
+		if (!copy_stack_frame(fp, &frame))
+			break;
+
+		if ((unsigned long)fp < user_stack_pointer(regs))
+			break;
+
+		callchain_store(entry, frame.return_address);
+		fp = frame.next_fp;
+	}
+
+	entry->user = entry->nr - nr;
+}
+
+static void
+perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry)
+{
+	int is_user;
+
+	if (!regs)
+		return;
+
+	is_user = user_mode(regs);
+
+	if (!current || current->pid == 0)
+		return;
+
+	if (is_user && current->state != TASK_RUNNING)
+		return;
+
+	if (!is_user)
+		perf_callchain_kernel(regs, entry);
+
+	if (current->mm)
+		perf_callchain_user(regs, entry);
+}
+
+struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
+{
+	struct perf_callchain_entry *entry;
+
+	if (in_nmi())
+		entry = &__get_cpu_var(nmi_entry);
+	else
+		entry = &__get_cpu_var(irq_entry);
+
+	entry->nr = 0;
+	entry->hv = 0;
+	entry->kernel = 0;
+	entry->user = 0;
+
+	perf_do_callchain(regs, entry);
+
+	return entry;
+}
diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c
index f6c70a164e3..d6f5b9fbde3 100644
--- a/arch/x86/kernel/cpu/perfctr-watchdog.c
+++ b/arch/x86/kernel/cpu/perfctr-watchdog.c
@@ -19,8 +19,8 @@
 #include <linux/nmi.h>
 #include <linux/kprobes.h>
 
-#include <asm/genapic.h>
-#include <asm/intel_arch_perfmon.h>
+#include <asm/apic.h>
+#include <asm/perf_counter.h>
 
 struct nmi_watchdog_ctlblk {
 	unsigned int cccr_msr;
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 1c17d7c751a..a4742a340d8 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -1012,6 +1012,11 @@ apicinterrupt ERROR_APIC_VECTOR \
 apicinterrupt SPURIOUS_APIC_VECTOR \
 	spurious_interrupt smp_spurious_interrupt
 
+#ifdef CONFIG_PERF_COUNTERS
+apicinterrupt LOCAL_PENDING_VECTOR \
+	perf_pending_interrupt smp_perf_pending_interrupt
+#endif
+
 /*
  * Exception entry points.
  */
diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index 9a391bbb8ba..38287b5f116 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -62,6 +62,14 @@ static int show_other_interrupts(struct seq_file *p, int prec)
 	for_each_online_cpu(j)
 		seq_printf(p, "%10u ", irq_stats(j)->irq_spurious_count);
 	seq_printf(p, "  Spurious interrupts\n");
+	seq_printf(p, "%*s: ", prec, "CNT");
+	for_each_online_cpu(j)
+		seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs);
+	seq_printf(p, "  Performance counter interrupts\n");
+	seq_printf(p, "%*s: ", prec, "PND");
+	for_each_online_cpu(j)
+		seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs);
+	seq_printf(p, "  Performance pending work\n");
 #endif
 	if (generic_interrupt_extension) {
 		seq_printf(p, "%*s: ", prec, "PLT");
@@ -165,6 +173,8 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
 #ifdef CONFIG_X86_LOCAL_APIC
 	sum += irq_stats(cpu)->apic_timer_irqs;
 	sum += irq_stats(cpu)->irq_spurious_count;
+	sum += irq_stats(cpu)->apic_perf_irqs;
+	sum += irq_stats(cpu)->apic_pending_irqs;
 #endif
 	if (generic_interrupt_extension)
 		sum += irq_stats(cpu)->generic_irqs;
diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c
index 2e08b10ad51..267c6624c77 100644
--- a/arch/x86/kernel/irqinit.c
+++ b/arch/x86/kernel/irqinit.c
@@ -181,10 +181,15 @@ static void __init apic_intr_init(void)
 {
 	smp_intr_init();
 
-#ifdef CONFIG_X86_64
+#ifdef CONFIG_X86_THERMAL_VECTOR
 	alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt);
+#endif
+#ifdef CONFIG_X86_THRESHOLD
 	alloc_intr_gate(THRESHOLD_APIC_VECTOR, threshold_interrupt);
 #endif
+#if defined(CONFIG_X86_NEW_MCE) && defined(CONFIG_X86_LOCAL_APIC)
+	alloc_intr_gate(MCE_SELF_VECTOR, mce_self_interrupt);
+#endif
 
 #if defined(CONFIG_X86_64) || defined(CONFIG_X86_LOCAL_APIC)
 	/* self generated IPI for local APIC timer */
@@ -199,18 +204,10 @@ static void __init apic_intr_init(void)
 
 	/* Performance monitoring interrupts: */
 # ifdef CONFIG_PERF_COUNTERS
-	alloc_intr_gate(LOCAL_PERF_VECTOR, perf_counter_interrupt);
 	alloc_intr_gate(LOCAL_PENDING_VECTOR, perf_pending_interrupt);
 # endif
 
 #endif
-
-#ifdef CONFIG_X86_32
-#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86_MCE_P4THERMAL)
-	/* thermal monitor LVT interrupt */
-	alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt);
-#endif
-#endif
 }
 
 /**
diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c
index 14425166b8e..0a813b17b17 100644
--- a/arch/x86/kernel/signal.c
+++ b/arch/x86/kernel/signal.c
@@ -6,7 +6,6 @@
  *  2000-06-20  Pentium III FXSR, SSE support by Gareth Hughes
  *  2000-2002   x86-64 support by Andi Kleen
  */
-
 #include <linux/sched.h>
 #include <linux/mm.h>
 #include <linux/smp.h>
diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S
index 734f92c02dd..d51321ddafd 100644
--- a/arch/x86/kernel/syscall_table_32.S
+++ b/arch/x86/kernel/syscall_table_32.S
@@ -335,3 +335,4 @@ ENTRY(sys_call_table)
 	.long sys_preadv
 	.long sys_pwritev
 	.long sys_rt_tgsigqueueinfo	/* 335 */
+	.long sys_perf_counter_open
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index ede024531f8..07d60c870ce 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -942,8 +942,13 @@ void __init trap_init(void)
 #endif
 	set_intr_gate(19, &simd_coprocessor_error);
 
+	/* Reserve all the builtin and the syscall vector: */
+	for (i = 0; i < FIRST_EXTERNAL_VECTOR; i++)
+		set_bit(i, used_vectors);
+
 #ifdef CONFIG_IA32_EMULATION
 	set_system_intr_gate(IA32_SYSCALL_VECTOR, ia32_syscall);
+	set_bit(IA32_SYSCALL_VECTOR, used_vectors);
 #endif
 
 #ifdef CONFIG_X86_32
@@ -960,14 +965,9 @@ void __init trap_init(void)
 	}
 
 	set_system_trap_gate(SYSCALL_VECTOR, &system_call);
+	set_bit(SYSCALL_VECTOR, used_vectors);
 #endif
 
-	/* Reserve all the builtin and the syscall vector: */
-	for (i = 0; i < FIRST_EXTERNAL_VECTOR; i++)
-		set_bit(i, used_vectors);
-
-	set_bit(IA32_SYSCALL_VECTOR, used_vectors);
-
 	/*
 	 * Should be a barrier for any external CPU state:
 	 */
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 5ec7ae36661..c6acc632637 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -10,6 +10,7 @@
 #include <linux/bootmem.h>		/* max_low_pfn			*/
 #include <linux/kprobes.h>		/* __kprobes, ...		*/
 #include <linux/mmiotrace.h>		/* kmmio_handler, ...		*/
+#include <linux/perf_counter.h>		/* perf_swcounter_event		*/
 
 #include <asm/traps.h>			/* dotraplinkage, ...		*/
 #include <asm/pgalloc.h>		/* pgd_*(), ...			*/
@@ -1013,6 +1014,8 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	if (unlikely(error_code & PF_RSVD))
 		pgtable_bad(regs, error_code, address);
 
+	perf_swcounter_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+
 	/*
 	 * If we're in an interrupt, have no user context or are running
 	 * in an atomic region then we must not take the fault:
@@ -1106,10 +1109,15 @@ good_area:
 		return;
 	}
 
-	if (fault & VM_FAULT_MAJOR)
+	if (fault & VM_FAULT_MAJOR) {
 		tsk->maj_flt++;
-	else
+		perf_swcounter_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+				     regs, address);
+	} else {
 		tsk->min_flt++;
+		perf_swcounter_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+				     regs, address);
+	}
 
 	check_v8086_mode(regs, address, tsk);
 
diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c
index 3b285e656e2..b07dd8d0b32 100644
--- a/arch/x86/oprofile/nmi_int.c
+++ b/arch/x86/oprofile/nmi_int.c
@@ -40,8 +40,9 @@ static int profile_exceptions_notify(struct notifier_block *self,
 
 	switch (val) {
 	case DIE_NMI:
-		if (model->check_ctrs(args->regs, &per_cpu(cpu_msrs, cpu)))
-			ret = NOTIFY_STOP;
+	case DIE_NMI_IPI:
+		model->check_ctrs(args->regs, &per_cpu(cpu_msrs, cpu));
+		ret = NOTIFY_STOP;
 		break;
 	default:
 		break;
@@ -134,7 +135,7 @@ static void nmi_cpu_setup(void *dummy)
 static struct notifier_block profile_exceptions_nb = {
 	.notifier_call = profile_exceptions_notify,
 	.next = NULL,
-	.priority = 0
+	.priority = 2
 };
 
 static int nmi_setup(void)
diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c
index 10131fbdaad..4da7230b3d1 100644
--- a/arch/x86/oprofile/op_model_ppro.c
+++ b/arch/x86/oprofile/op_model_ppro.c
@@ -18,7 +18,7 @@
 #include <asm/msr.h>
 #include <asm/apic.h>
 #include <asm/nmi.h>
-#include <asm/intel_arch_perfmon.h>
+#include <asm/perf_counter.h>
 
 #include "op_x86_model.h"
 #include "op_counter.h"
@@ -136,6 +136,13 @@ static int ppro_check_ctrs(struct pt_regs * const regs,
 	u64 val;
 	int i;
 
+	/*
+	 * This can happen if perf counters are in use when
+	 * we steal the die notifier NMI.
+	 */
+	if (unlikely(!reset_value))
+		goto out;
+
 	for (i = 0 ; i < num_counters; ++i) {
 		if (!reset_value[i])
 			continue;
@@ -146,6 +153,7 @@ static int ppro_check_ctrs(struct pt_regs * const regs,
 		}
 	}
 
+out:
 	/* Only P6 based Pentium M need to re-unmask the apic vector but it
 	 * doesn't hurt other P6 variant */
 	apic_write(APIC_LVTPC, apic_read(APIC_LVTPC) & ~APIC_LVT_MASKED);
diff --git a/arch/x86/vdso/vdso32-setup.c b/arch/x86/vdso/vdso32-setup.c
index 1241f118ab5..58bc00f68b1 100644
--- a/arch/x86/vdso/vdso32-setup.c
+++ b/arch/x86/vdso/vdso32-setup.c
@@ -338,6 +338,8 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 		}
 	}
 
+	current->mm->context.vdso = (void *)addr;
+
 	if (compat_uses_vma || !compat) {
 		/*
 		 * MAYWRITE to allow gdb to COW and set breakpoints
@@ -358,11 +360,13 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 			goto up_fail;
 	}
 
-	current->mm->context.vdso = (void *)addr;
 	current_thread_info()->sysenter_return =
 		VDSO32_SYMBOL(addr, SYSENTER_RETURN);
 
   up_fail:
+	if (ret)
+		current->mm->context.vdso = NULL;
+
 	up_write(&mm->mmap_sem);
 
 	return ret;
diff --git a/arch/x86/vdso/vma.c b/arch/x86/vdso/vma.c
index cac083386e0..21e1aeb9f3e 100644
--- a/arch/x86/vdso/vma.c
+++ b/arch/x86/vdso/vma.c
@@ -116,15 +116,18 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 		goto up_fail;
 	}
 
+	current->mm->context.vdso = (void *)addr;
+
 	ret = install_special_mapping(mm, addr, vdso_size,
 				      VM_READ|VM_EXEC|
 				      VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC|
 				      VM_ALWAYSDUMP,
 				      vdso_pages);
-	if (ret)
+	if (ret) {
+		current->mm->context.vdso = NULL;
 		goto up_fail;
+	}
 
-	current->mm->context.vdso = (void *)addr;
 up_fail:
 	up_write(&mm->mmap_sem);
 	return ret;