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

* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (311 commits) perf tools: Add mode to build without newt support perf symbols: symbol inconsistency message should be done only at verbose=1 perf tui: Add explicit -lslang option perf options: Type check all the remaining OPT_ variants perf options: Type check OPT_BOOLEAN and fix the offenders perf options: Check v type in OPT_U?INTEGER perf options: Introduce OPT_UINTEGER perf tui: Add workaround for slang < 2.1.4 perf record: Fix bug mismatch with -c option definition perf options: Introduce OPT_U64 perf tui: Add help window to show key associations perf tui: Make <- exit menus too perf newt: Add single key shortcuts for zoom into DSO and threads perf newt: Exit browser unconditionally when CTRL+C, q or Q is pressed perf newt: Fix the 'A'/'a' shortcut for annotate perf newt: Make <- exit the ui_browser x86, perf: P4 PMU - fix counters management logic perf newt: Make <- zoom out filters perf report: Report number of events, not samples perf hist: Clarify events_stats fields usage ... Fix up trivial conflicts in kernel/fork.c and tools/perf/builtin-record.c
author: Linus Torvalds <torvalds@linux-foundation.org> 2010-05-18 08:19:03 -0700
committer: Linus Torvalds <torvalds@linux-foundation.org> 2010-05-18 08:19:03 -0700
commit: 4d7b4ac22fbec1a03206c6cde353f2fd6942f828 (patch)
tree: 2d96a9e9c28cf6fa628a278decc00ad55a8b043b /arch/x86/kernel
parent: 3aaf51ace5975050ab43c7d4d7e439e0ae7d13d7 (diff)
parent: 94f3ca95787ada3d64339a4ecb2754236ab563f6 (diff)
21 files changed, 2394 insertions, 2965 deletions
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 4c58352209e..e77b2208372 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -47,8 +47,6 @@ obj-$(CONFIG_X86_TRAMPOLINE)	+= trampoline.o
 obj-y				+= process.o
 obj-y				+= i387.o xsave.o
 obj-y				+= ptrace.o
-obj-$(CONFIG_X86_DS)		+= ds.o
-obj-$(CONFIG_X86_DS_SELFTEST)		+= ds_selftest.o
 obj-$(CONFIG_X86_32)		+= tls.o
 obj-$(CONFIG_IA32_EMULATION)	+= tls.o
 obj-y				+= step.o
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 1366c7cfd48..f5e5390d345 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -12,7 +12,6 @@
 #include <asm/processor.h>
 #include <asm/pgtable.h>
 #include <asm/msr.h>
-#include <asm/ds.h>
 #include <asm/bugs.h>
 #include <asm/cpu.h>
 
@@ -388,7 +387,6 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c)
 			set_cpu_cap(c, X86_FEATURE_BTS);
 		if (!(l1 & (1<<12)))
 			set_cpu_cap(c, X86_FEATURE_PEBS);
-		ds_init_intel(c);
 	}
 
 	if (c->x86 == 6 && c->x86_model == 29 && cpu_has_clflush)
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index db5bdc8addf..fd4db0db370 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -31,46 +31,51 @@
 #include <asm/nmi.h>
 #include <asm/compat.h>
 
-static u64 perf_event_mask __read_mostly;
+#if 0
+#undef wrmsrl
+#define wrmsrl(msr, val) 					\
+do {								\
+	trace_printk("wrmsrl(%lx, %lx)\n", (unsigned long)(msr),\
+			(unsigned long)(val));			\
+	native_write_msr((msr), (u32)((u64)(val)), 		\
+			(u32)((u64)(val) >> 32));		\
+} while (0)
+#endif
 
-/* The maximal number of PEBS events: */
-#define MAX_PEBS_EVENTS	4
+/*
+ * best effort, GUP based copy_from_user() that assumes IRQ or NMI context
+ */
+static unsigned long
+copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
+{
+	unsigned long offset, addr = (unsigned long)from;
+	int type = in_nmi() ? KM_NMI : KM_IRQ0;
+	unsigned long size, len = 0;
+	struct page *page;
+	void *map;
+	int ret;
 
-/* The size of a BTS record in bytes: */
-#define BTS_RECORD_SIZE		24
+	do {
+		ret = __get_user_pages_fast(addr, 1, 0, &page);
+		if (!ret)
+			break;
 
-/* The size of a per-cpu BTS buffer in bytes: */
-#define BTS_BUFFER_SIZE		(BTS_RECORD_SIZE * 2048)
+		offset = addr & (PAGE_SIZE - 1);
+		size = min(PAGE_SIZE - offset, n - len);
 
-/* The BTS overflow threshold in bytes from the end of the buffer: */
-#define BTS_OVFL_TH		(BTS_RECORD_SIZE * 128)
+		map = kmap_atomic(page, type);
+		memcpy(to, map+offset, size);
+		kunmap_atomic(map, type);
+		put_page(page);
 
+		len  += size;
+		to   += size;
+		addr += size;
 
-/*
- * Bits in the debugctlmsr controlling branch tracing.
- */
-#define X86_DEBUGCTL_TR			(1 << 6)
-#define X86_DEBUGCTL_BTS		(1 << 7)
-#define X86_DEBUGCTL_BTINT		(1 << 8)
-#define X86_DEBUGCTL_BTS_OFF_OS		(1 << 9)
-#define X86_DEBUGCTL_BTS_OFF_USR	(1 << 10)
+	} while (len < n);
 
-/*
- * A debug store configuration.
- *
- * We only support architectures that use 64bit fields.
- */
-struct debug_store {
-	u64	bts_buffer_base;
-	u64	bts_index;
-	u64	bts_absolute_maximum;
-	u64	bts_interrupt_threshold;
-	u64	pebs_buffer_base;
-	u64	pebs_index;
-	u64	pebs_absolute_maximum;
-	u64	pebs_interrupt_threshold;
-	u64	pebs_event_reset[MAX_PEBS_EVENTS];
-};
+	return len;
+}
 
 struct event_constraint {
 	union {
@@ -89,18 +94,41 @@ struct amd_nb {
 	struct event_constraint event_constraints[X86_PMC_IDX_MAX];
 };
 
+#define MAX_LBR_ENTRIES		16
+
 struct cpu_hw_events {
+	/*
+	 * Generic x86 PMC bits
+	 */
 	struct perf_event	*events[X86_PMC_IDX_MAX]; /* in counter order */
 	unsigned long		active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
-	unsigned long		interrupts;
 	int			enabled;
-	struct debug_store	*ds;
 
 	int			n_events;
 	int			n_added;
 	int			assign[X86_PMC_IDX_MAX]; /* event to counter assignment */
 	u64			tags[X86_PMC_IDX_MAX];
 	struct perf_event	*event_list[X86_PMC_IDX_MAX]; /* in enabled order */
+
+	unsigned int		group_flag;
+
+	/*
+	 * Intel DebugStore bits
+	 */
+	struct debug_store	*ds;
+	u64			pebs_enabled;
+
+	/*
+	 * Intel LBR bits
+	 */
+	int				lbr_users;
+	void				*lbr_context;
+	struct perf_branch_stack	lbr_stack;
+	struct perf_branch_entry	lbr_entries[MAX_LBR_ENTRIES];
+
+	/*
+	 * AMD specific bits
+	 */
 	struct amd_nb		*amd_nb;
 };
 
@@ -114,44 +142,75 @@ struct cpu_hw_events {
 #define EVENT_CONSTRAINT(c, n, m)	\
 	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n))
 
+/*
+ * Constraint on the Event code.
+ */
 #define INTEL_EVENT_CONSTRAINT(c, n)	\
-	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK)
+	EVENT_CONSTRAINT(c, n, ARCH_PERFMON_EVENTSEL_EVENT)
 
+/*
+ * Constraint on the Event code + UMask + fixed-mask
+ *
+ * filter mask to validate fixed counter events.
+ * the following filters disqualify for fixed counters:
+ *  - inv
+ *  - edge
+ *  - cnt-mask
+ *  The other filters are supported by fixed counters.
+ *  The any-thread option is supported starting with v3.
+ */
 #define FIXED_EVENT_CONSTRAINT(c, n)	\
-	EVENT_CONSTRAINT(c, (1ULL << (32+n)), INTEL_ARCH_FIXED_MASK)
+	EVENT_CONSTRAINT(c, (1ULL << (32+n)), X86_RAW_EVENT_MASK)
+
+/*
+ * Constraint on the Event code + UMask
+ */
+#define PEBS_EVENT_CONSTRAINT(c, n)	\
+	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)
 
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)
 
 #define for_each_event_constraint(e, c)	\
-	for ((e) = (c); (e)->cmask; (e)++)
+	for ((e) = (c); (e)->weight; (e)++)
+
+union perf_capabilities {
+	struct {
+		u64	lbr_format    : 6;
+		u64	pebs_trap     : 1;
+		u64	pebs_arch_reg : 1;
+		u64	pebs_format   : 4;
+		u64	smm_freeze    : 1;
+	};
+	u64	capabilities;
+};
 
 /*
  * struct x86_pmu - generic x86 pmu
  */
 struct x86_pmu {
+	/*
+	 * Generic x86 PMC bits
+	 */
 	const char	*name;
 	int		version;
 	int		(*handle_irq)(struct pt_regs *);
 	void		(*disable_all)(void);
-	void		(*enable_all)(void);
+	void		(*enable_all)(int added);
 	void		(*enable)(struct perf_event *);
 	void		(*disable)(struct perf_event *);
+	int		(*hw_config)(struct perf_event *event);
+	int		(*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign);
 	unsigned	eventsel;
 	unsigned	perfctr;
 	u64		(*event_map)(int);
-	u64		(*raw_event)(u64);
 	int		max_events;
-	int		num_events;
-	int		num_events_fixed;
-	int		event_bits;
-	u64		event_mask;
+	int		num_counters;
+	int		num_counters_fixed;
+	int		cntval_bits;
+	u64		cntval_mask;
 	int		apic;
 	u64		max_period;
-	u64		intel_ctrl;
-	void		(*enable_bts)(u64 config);
-	void		(*disable_bts)(void);
-
 	struct event_constraint *
 			(*get_event_constraints)(struct cpu_hw_events *cpuc,
 						 struct perf_event *event);
@@ -159,11 +218,32 @@ struct x86_pmu {
 	void		(*put_event_constraints)(struct cpu_hw_events *cpuc,
 						 struct perf_event *event);
 	struct event_constraint *event_constraints;
+	void		(*quirks)(void);
 
 	int		(*cpu_prepare)(int cpu);
 	void		(*cpu_starting)(int cpu);
 	void		(*cpu_dying)(int cpu);
 	void		(*cpu_dead)(int cpu);
+
+	/*
+	 * Intel Arch Perfmon v2+
+	 */
+	u64			intel_ctrl;
+	union perf_capabilities intel_cap;
+
+	/*
+	 * Intel DebugStore bits
+	 */
+	int		bts, pebs;
+	int		pebs_record_size;
+	void		(*drain_pebs)(struct pt_regs *regs);
+	struct event_constraint *pebs_constraints;
+
+	/*
+	 * Intel LBR
+	 */
+	unsigned long	lbr_tos, lbr_from, lbr_to; /* MSR base regs       */
+	int		lbr_nr;			   /* hardware stack size */
 };
 
 static struct x86_pmu x86_pmu __read_mostly;
@@ -198,7 +278,7 @@ static u64
 x86_perf_event_update(struct perf_event *event)
 {
 	struct hw_perf_event *hwc = &event->hw;
-	int shift = 64 - x86_pmu.event_bits;
+	int shift = 64 - x86_pmu.cntval_bits;
 	u64 prev_raw_count, new_raw_count;
 	int idx = hwc->idx;
 	s64 delta;
@@ -241,33 +321,32 @@ again:
 static atomic_t active_events;
 static DEFINE_MUTEX(pmc_reserve_mutex);
 
+#ifdef CONFIG_X86_LOCAL_APIC
+
 static bool reserve_pmc_hardware(void)
 {
-#ifdef CONFIG_X86_LOCAL_APIC
 	int i;
 
 	if (nmi_watchdog == NMI_LOCAL_APIC)
 		disable_lapic_nmi_watchdog();
 
-	for (i = 0; i < x86_pmu.num_events; i++) {
+	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_events; i++) {
+	for (i = 0; i < x86_pmu.num_counters; i++) {
 		if (!reserve_evntsel_nmi(x86_pmu.eventsel + i))
 			goto eventsel_fail;
 	}
-#endif
 
 	return true;
 
-#ifdef CONFIG_X86_LOCAL_APIC
 eventsel_fail:
 	for (i--; i >= 0; i--)
 		release_evntsel_nmi(x86_pmu.eventsel + i);
 
-	i = x86_pmu.num_events;
+	i = x86_pmu.num_counters;
 
 perfctr_fail:
 	for (i--; i >= 0; i--)
@@ -277,128 +356,36 @@ perfctr_fail:
 		enable_lapic_nmi_watchdog();
 
 	return false;
-#endif
 }
 
 static void release_pmc_hardware(void)
 {
-#ifdef CONFIG_X86_LOCAL_APIC
 	int i;
 
-	for (i = 0; i < x86_pmu.num_events; 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();
-#endif
-}
-
-static inline bool bts_available(void)
-{
-	return x86_pmu.enable_bts != NULL;
 }
 
-static void init_debug_store_on_cpu(int cpu)
-{
-	struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
-
-	if (!ds)
-		return;
-
-	wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA,
-		     (u32)((u64)(unsigned long)ds),
-		     (u32)((u64)(unsigned long)ds >> 32));
-}
-
-static void fini_debug_store_on_cpu(int cpu)
-{
-	if (!per_cpu(cpu_hw_events, cpu).ds)
-		return;
-
-	wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, 0, 0);
-}
-
-static void release_bts_hardware(void)
-{
-	int cpu;
-
-	if (!bts_available())
-		return;
-
-	get_online_cpus();
-
-	for_each_online_cpu(cpu)
-		fini_debug_store_on_cpu(cpu);
-
-	for_each_possible_cpu(cpu) {
-		struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
-
-		if (!ds)
-			continue;
-
-		per_cpu(cpu_hw_events, cpu).ds = NULL;
-
-		kfree((void *)(unsigned long)ds->bts_buffer_base);
-		kfree(ds);
-	}
-
-	put_online_cpus();
-}
-
-static int reserve_bts_hardware(void)
-{
-	int cpu, err = 0;
-
-	if (!bts_available())
-		return 0;
-
-	get_online_cpus();
-
-	for_each_possible_cpu(cpu) {
-		struct debug_store *ds;
-		void *buffer;
-
-		err = -ENOMEM;
-		buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL);
-		if (unlikely(!buffer))
-			break;
-
-		ds = kzalloc(sizeof(*ds), GFP_KERNEL);
-		if (unlikely(!ds)) {
-			kfree(buffer);
-			break;
-		}
-
-		ds->bts_buffer_base = (u64)(unsigned long)buffer;
-		ds->bts_index = ds->bts_buffer_base;
-		ds->bts_absolute_maximum =
-			ds->bts_buffer_base + BTS_BUFFER_SIZE;
-		ds->bts_interrupt_threshold =
-			ds->bts_absolute_maximum - BTS_OVFL_TH;
-
-		per_cpu(cpu_hw_events, cpu).ds = ds;
-		err = 0;
-	}
+#else
 
-	if (err)
-		release_bts_hardware();
-	else {
-		for_each_online_cpu(cpu)
-			init_debug_store_on_cpu(cpu);
-	}
+static bool reserve_pmc_hardware(void) { return true; }
+static void release_pmc_hardware(void) {}
 
-	put_online_cpus();
+#endif
 
-	return err;
-}
+static int reserve_ds_buffers(void);
+static void release_ds_buffers(void);
 
 static void hw_perf_event_destroy(struct perf_event *event)
 {
 	if (atomic_dec_and_mutex_lock(&active_events, &pmc_reserve_mutex)) {
 		release_pmc_hardware();
-		release_bts_hardware();
+		release_ds_buffers();
 		mutex_unlock(&pmc_reserve_mutex);
 	}
 }
@@ -441,54 +428,11 @@ set_ext_hw_attr(struct hw_perf_event *hwc, struct perf_event_attr *attr)
 	return 0;
 }
 
-/*
- * Setup the hardware configuration for a given attr_type
- */
-static int __hw_perf_event_init(struct perf_event *event)
+static int x86_setup_perfctr(struct perf_event *event)
 {
 	struct perf_event_attr *attr = &event->attr;
 	struct hw_perf_event *hwc = &event->hw;
 	u64 config;
-	int err;
-
-	if (!x86_pmu_initialized())
-		return -ENODEV;
-
-	err = 0;
-	if (!atomic_inc_not_zero(&active_events)) {
-		mutex_lock(&pmc_reserve_mutex);
-		if (atomic_read(&active_events) == 0) {
-			if (!reserve_pmc_hardware())
-				err = -EBUSY;
-			else
-				err = reserve_bts_hardware();
-		}
-		if (!err)
-			atomic_inc(&active_events);
-		mutex_unlock(&pmc_reserve_mutex);
-	}
-	if (err)
-		return err;
-
-	event->destroy = hw_perf_event_destroy;
-
-	/*
-	 * Generate PMC IRQs:
-	 * (keep 'enabled' bit clear for now)
-	 */
-	hwc->config = ARCH_PERFMON_EVENTSEL_INT;
-
-	hwc->idx = -1;
-	hwc->last_cpu = -1;
-	hwc->last_tag = ~0ULL;
-
-	/*
-	 * 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;
@@ -505,16 +449,8 @@ static int __hw_perf_event_init(struct perf_event *event)
 			return -EOPNOTSUPP;
 	}
 
-	/*
-	 * Raw hw_event type provide the config in the hw_event structure
-	 */
-	if (attr->type == PERF_TYPE_RAW) {
-		hwc->config |= x86_pmu.raw_event(attr->config);
-		if ((hwc->config & ARCH_PERFMON_EVENTSEL_ANY) &&
-		    perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN))
-			return -EACCES;
+	if (attr->type == PERF_TYPE_RAW)
 		return 0;
-	}
 
 	if (attr->type == PERF_TYPE_HW_CACHE)
 		return set_ext_hw_attr(hwc, attr);
@@ -539,11 +475,11 @@ static int __hw_perf_event_init(struct perf_event *event)
 	if ((attr->config == PERF_COUNT_HW_BRANCH_INSTRUCTIONS) &&
 	    (hwc->sample_period == 1)) {
 		/* BTS is not supported by this architecture. */
-		if (!bts_available())
+		if (!x86_pmu.bts)
 			return -EOPNOTSUPP;
 
 		/* BTS is currently only allowed for user-mode. */
-		if (hwc->config & ARCH_PERFMON_EVENTSEL_OS)
+		if (!attr->exclude_kernel)
 			return -EOPNOTSUPP;
 	}
 
@@ -552,12 +488,87 @@ static int __hw_perf_event_init(struct perf_event *event)
 	return 0;
 }
 
+static int x86_pmu_hw_config(struct perf_event *event)
+{
+	if (event->attr.precise_ip) {
+		int precise = 0;
+
+		/* Support for constant skid */
+		if (x86_pmu.pebs)
+			precise++;
+
+		/* Support for IP fixup */
+		if (x86_pmu.lbr_nr)
+			precise++;
+
+		if (event->attr.precise_ip > precise)
+			return -EOPNOTSUPP;
+	}
+
+	/*
+	 * Generate PMC IRQs:
+	 * (keep 'enabled' bit clear for now)
+	 */
+	event->hw.config = ARCH_PERFMON_EVENTSEL_INT;
+
+	/*
+	 * Count user and OS events unless requested not to
+	 */
+	if (!event->attr.exclude_user)
+		event->hw.config |= ARCH_PERFMON_EVENTSEL_USR;
+	if (!event->attr.exclude_kernel)
+		event->hw.config |= ARCH_PERFMON_EVENTSEL_OS;
+
+	if (event->attr.type == PERF_TYPE_RAW)
+		event->hw.config |= event->attr.config & X86_RAW_EVENT_MASK;
+
+	return x86_setup_perfctr(event);
+}
+
+/*
+ * Setup the hardware configuration for a given attr_type
+ */
+static int __hw_perf_event_init(struct perf_event *event)
+{
+	int err;
+
+	if (!x86_pmu_initialized())
+		return -ENODEV;
+
+	err = 0;
+	if (!atomic_inc_not_zero(&active_events)) {
+		mutex_lock(&pmc_reserve_mutex);
+		if (atomic_read(&active_events) == 0) {
+			if (!reserve_pmc_hardware())
+				err = -EBUSY;
+			else {
+				err = reserve_ds_buffers();
+				if (err)
+					release_pmc_hardware();
+			}
+		}
+		if (!err)
+			atomic_inc(&active_events);
+		mutex_unlock(&pmc_reserve_mutex);
+	}
+	if (err)
+		return err;
+
+	event->destroy = hw_perf_event_destroy;
+
+	event->hw.idx = -1;
+	event->hw.last_cpu = -1;
+	event->hw.last_tag = ~0ULL;
+
+	return x86_pmu.hw_config(event);
+}
+
 static void x86_pmu_disable_all(void)
 {
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 	int idx;
 
-	for (idx = 0; idx < x86_pmu.num_events; idx++) {
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
 		u64 val;
 
 		if (!test_bit(idx, cpuc->active_mask))
@@ -587,12 +598,12 @@ void hw_perf_disable(void)
 	x86_pmu.disable_all();
 }
 
-static void x86_pmu_enable_all(void)
+static void x86_pmu_enable_all(int added)
 {
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 	int idx;
 
-	for (idx = 0; idx < x86_pmu.num_events; idx++) {
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
 		struct perf_event *event = cpuc->events[idx];
 		u64 val;
 
@@ -667,14 +678,14 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
 	 * assign events to counters starting with most
 	 * constrained events.
 	 */
-	wmax = x86_pmu.num_events;
+	wmax = x86_pmu.num_counters;
 
 	/*
 	 * when fixed event counters are present,
 	 * wmax is incremented by 1 to account
 	 * for one more choice
 	 */
-	if (x86_pmu.num_events_fixed)
+	if (x86_pmu.num_counters_fixed)
 		wmax++;
 
 	for (w = 1, num = n; num && w <= wmax; w++) {
@@ -724,7 +735,7 @@ static int collect_events(struct cpu_hw_events *cpuc, struct perf_event *leader,
 	struct perf_event *event;
 	int n, max_count;
 
-	max_count = x86_pmu.num_events + x86_pmu.num_events_fixed;
+	max_count = x86_pmu.num_counters + x86_pmu.num_counters_fixed;
 
 	/* current number of events already accepted */
 	n = cpuc->n_events;
@@ -795,7 +806,7 @@ void hw_perf_enable(void)
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 	struct perf_event *event;
 	struct hw_perf_event *hwc;
-	int i;
+	int i, added = cpuc->n_added;
 
 	if (!x86_pmu_initialized())
 		return;
@@ -847,19 +858,20 @@ void hw_perf_enable(void)
 	cpuc->enabled = 1;
 	barrier();
 
-	x86_pmu.enable_all();
+	x86_pmu.enable_all(added);
 }
 
-static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc)
+static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc,
+					  u64 enable_mask)
 {
-	(void)checking_wrmsrl(hwc->config_base + hwc->idx,
-			      hwc->config | ARCH_PERFMON_EVENTSEL_ENABLE);
+	wrmsrl(hwc->config_base + hwc->idx, hwc->config | enable_mask);
 }
 
 static inline void x86_pmu_disable_event(struct perf_event *event)
 {
 	struct hw_perf_event *hwc = &event->hw;
-	(void)checking_wrmsrl(hwc->config_base + hwc->idx, hwc->config);
+
+	wrmsrl(hwc->config_base + hwc->idx, hwc->config);
 }
 
 static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left);
@@ -874,7 +886,7 @@ x86_perf_event_set_period(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 	s64 left = atomic64_read(&hwc->period_left);
 	s64 period = hwc->sample_period;
-	int err, ret = 0, idx = hwc->idx;
+	int ret = 0, idx = hwc->idx;
 
 	if (idx == X86_PMC_IDX_FIXED_BTS)
 		return 0;
@@ -912,8 +924,8 @@ x86_perf_event_set_period(struct perf_event *event)
 	 */
 	atomic64_set(&hwc->prev_count, (u64)-left);
 
-	err = checking_wrmsrl(hwc->event_base + idx,
-			     (u64)(-left) & x86_pmu.event_mask);
+	wrmsrl(hwc->event_base + idx,
+			(u64)(-left) & x86_pmu.cntval_mask);
 
 	perf_event_update_userpage(event);
 
@@ -924,7 +936,8 @@ static void x86_pmu_enable_event(struct perf_event *event)
 {
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 	if (cpuc->enabled)
-		__x86_pmu_enable_event(&event->hw);
+		__x86_pmu_enable_event(&event->hw,
+				       ARCH_PERFMON_EVENTSEL_ENABLE);
 }
 
 /*
@@ -950,7 +963,15 @@ static int x86_pmu_enable(struct perf_event *event)
 	if (n < 0)
 		return n;
 
-	ret = x86_schedule_events(cpuc, n, assign);
+	/*
+	 * If group events scheduling transaction was started,
+	 * skip the schedulability test here, it will be peformed
+	 * at commit time(->commit_txn) as a whole
+	 */
+	if (cpuc->group_flag & PERF_EVENT_TXN_STARTED)
+		goto out;
+
+	ret = x86_pmu.schedule_events(cpuc, n, assign);
 	if (ret)
 		return ret;
 	/*
@@ -959,6 +980,7 @@ static int x86_pmu_enable(struct perf_event *event)
 	 */
 	memcpy(cpuc->assign, assign, n*sizeof(int));
 
+out:
 	cpuc->n_events = n;
 	cpuc->n_added += n - n0;
 
@@ -991,11 +1013,12 @@ static void x86_pmu_unthrottle(struct perf_event *event)
 void perf_event_print_debug(void)
 {
 	u64 ctrl, status, overflow, pmc_ctrl, pmc_count, prev_left, fixed;
+	u64 pebs;
 	struct cpu_hw_events *cpuc;
 	unsigned long flags;
 	int cpu, idx;
 
-	if (!x86_pmu.num_events)
+	if (!x86_pmu.num_counters)
 		return;
 
 	local_irq_save(flags);
@@ -1008,16 +1031,18 @@ void perf_event_print_debug(void)
 		rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status);
 		rdmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, overflow);
 		rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR_CTRL, fixed);
+		rdmsrl(MSR_IA32_PEBS_ENABLE, pebs);
 
 		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: pebs:       %016llx\n", cpu, pebs);
 	}
-	pr_info("CPU#%d: active:       %016llx\n", cpu, *(u64 *)cpuc->active_mask);
+	pr_info("CPU#%d: active:     %016llx\n", cpu, *(u64 *)cpuc->active_mask);
 
-	for (idx = 0; idx < x86_pmu.num_events; idx++) {
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
 		rdmsrl(x86_pmu.eventsel + idx, pmc_ctrl);
 		rdmsrl(x86_pmu.perfctr  + idx, pmc_count);
 
@@ -1030,7 +1055,7 @@ void perf_event_print_debug(void)
 		pr_info("CPU#%d:   gen-PMC%d left:  %016llx\n",
 			cpu, idx, prev_left);
 	}
-	for (idx = 0; idx < x86_pmu.num_events_fixed; idx++) {
+	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",
@@ -1095,7 +1120,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
 
 	cpuc = &__get_cpu_var(cpu_hw_events);
 
-	for (idx = 0; idx < x86_pmu.num_events; idx++) {
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
 		if (!test_bit(idx, cpuc->active_mask))
 			continue;
 
@@ -1103,7 +1128,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
 		hwc = &event->hw;
 
 		val = x86_perf_event_update(event);
-		if (val & (1ULL << (x86_pmu.event_bits - 1)))
+		if (val & (1ULL << (x86_pmu.cntval_bits - 1)))
 			continue;
 
 		/*
@@ -1146,7 +1171,6 @@ void set_perf_event_pending(void)
 
 void perf_events_lapic_init(void)
 {
-#ifdef CONFIG_X86_LOCAL_APIC
 	if (!x86_pmu.apic || !x86_pmu_initialized())
 		return;
 
@@ -1154,7 +1178,6 @@ void perf_events_lapic_init(void)
 	 * Always use NMI for PMU
 	 */
 	apic_write(APIC_LVTPC, APIC_DM_NMI);
-#endif
 }
 
 static int __kprobes
@@ -1178,9 +1201,7 @@ perf_event_nmi_handler(struct notifier_block *self,
 
 	regs = args->regs;
 
-#ifdef CONFIG_X86_LOCAL_APIC
 	apic_write(APIC_LVTPC, APIC_DM_NMI);
-#endif
 	/*
 	 * Can't rely on the handled return value to say it was our NMI, two
 	 * events could trigger 'simultaneously' raising two back-to-back NMIs.
@@ -1217,118 +1238,11 @@ x86_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
 	return &unconstrained;
 }
 
-static int x86_event_sched_in(struct perf_event *event,
-			  struct perf_cpu_context *cpuctx)
-{
-	int ret = 0;
-
-	event->state = PERF_EVENT_STATE_ACTIVE;
-	event->oncpu = smp_processor_id();
-	event->tstamp_running += event->ctx->time - event->tstamp_stopped;
-
-	if (!is_x86_event(event))
-		ret = event->pmu->enable(event);
-
-	if (!ret && !is_software_event(event))
-		cpuctx->active_oncpu++;
-
-	if (!ret && event->attr.exclusive)
-		cpuctx->exclusive = 1;
-
-	return ret;
-}
-
-static void x86_event_sched_out(struct perf_event *event,
-			    struct perf_cpu_context *cpuctx)
-{
-	event->state = PERF_EVENT_STATE_INACTIVE;
-	event->oncpu = -1;
-
-	if (!is_x86_event(event))
-		event->pmu->disable(event);
-
-	event->tstamp_running -= event->ctx->time - event->tstamp_stopped;
-
-	if (!is_software_event(event))
-		cpuctx->active_oncpu--;
-
-	if (event->attr.exclusive || !cpuctx->active_oncpu)
-		cpuctx->exclusive = 0;
-}
-
-/*
- * Called to enable a whole group of events.
- * Returns 1 if the group was enabled, or -EAGAIN if it could not be.
- * Assumes the caller has disabled interrupts and has
- * frozen the PMU with hw_perf_save_disable.
- *
- * called with PMU disabled. If successful and return value 1,
- * then guaranteed to call perf_enable() and hw_perf_enable()
- */
-int hw_perf_group_sched_in(struct perf_event *leader,
-	       struct perf_cpu_context *cpuctx,
-	       struct perf_event_context *ctx)
-{
-	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
-	struct perf_event *sub;
-	int assign[X86_PMC_IDX_MAX];
-	int n0, n1, ret;
-
-	/* n0 = total number of events */
-	n0 = collect_events(cpuc, leader, true);
-	if (n0 < 0)
-		return n0;
-
-	ret = x86_schedule_events(cpuc, n0, assign);
-	if (ret)
-		return ret;
-
-	ret = x86_event_sched_in(leader, cpuctx);
-	if (ret)
-		return ret;
-
-	n1 = 1;
-	list_for_each_entry(sub, &leader->sibling_list, group_entry) {
-		if (sub->state > PERF_EVENT_STATE_OFF) {
-			ret = x86_event_sched_in(sub, cpuctx);
-			if (ret)
-				goto undo;
-			++n1;
-		}
-	}
-	/*
-	 * copy new assignment, now we know it is possible
-	 * will be used by hw_perf_enable()
-	 */
-	memcpy(cpuc->assign, assign, n0*sizeof(int));
-
-	cpuc->n_events  = n0;
-	cpuc->n_added  += n1;
-	ctx->nr_active += n1;
-
-	/*
-	 * 1 means successful and events are active
-	 * This is not quite true because we defer
-	 * actual activation until hw_perf_enable() but
-	 * this way we* ensure caller won't try to enable
-	 * individual events
-	 */
-	return 1;
-undo:
-	x86_event_sched_out(leader, cpuctx);
-	n0  = 1;
-	list_for_each_entry(sub, &leader->sibling_list, group_entry) {
-		if (sub->state == PERF_EVENT_STATE_ACTIVE) {
-			x86_event_sched_out(sub, cpuctx);
-			if (++n0 == n1)
-				break;
-		}
-	}
-	return ret;
-}
-
 #include "perf_event_amd.c"
 #include "perf_event_p6.c"
+#include "perf_event_p4.c"
+#include "perf_event_intel_lbr.c"
+#include "perf_event_intel_ds.c"
 #include "perf_event_intel.c"
 
 static int __cpuinit
@@ -1402,48 +1316,50 @@ void __init init_hw_perf_events(void)
 
 	pr_cont("%s PMU driver.\n", x86_pmu.name);
 
-	if (x86_pmu.num_events > X86_PMC_MAX_GENERIC) {
+	if (x86_pmu.quirks)
+		x86_pmu.quirks();
+
+	if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) {
 		WARN(1, KERN_ERR "hw perf events %d > max(%d), clipping!",
-		     x86_pmu.num_events, X86_PMC_MAX_GENERIC);
-		x86_pmu.num_events = X86_PMC_MAX_GENERIC;
+		     x86_pmu.num_counters, X86_PMC_MAX_GENERIC);
+		x86_pmu.num_counters = X86_PMC_MAX_GENERIC;
 	}
-	perf_event_mask = (1 << x86_pmu.num_events) - 1;
-	perf_max_events = x86_pmu.num_events;
+	x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1;
+	perf_max_events = x86_pmu.num_counters;
 
-	if (x86_pmu.num_events_fixed > X86_PMC_MAX_FIXED) {
+	if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) {
 		WARN(1, KERN_ERR "hw perf events fixed %d > max(%d), clipping!",
-		     x86_pmu.num_events_fixed, X86_PMC_MAX_FIXED);
-		x86_pmu.num_events_fixed = X86_PMC_MAX_FIXED;
+		     x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED);
+		x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED;
 	}
 
-	perf_event_mask |=
-		((1LL << x86_pmu.num_events_fixed)-1) << X86_PMC_IDX_FIXED;
-	x86_pmu.intel_ctrl = perf_event_mask;
+	x86_pmu.intel_ctrl |=
+		((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED;
 
 	perf_events_lapic_init();
 	register_die_notifier(&perf_event_nmi_notifier);
 
 	unconstrained = (struct event_constraint)
-		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_events) - 1,
-				   0, x86_pmu.num_events);
+		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
+				   0, x86_pmu.num_counters);
 
 	if (x86_pmu.event_constraints) {
 		for_each_event_constraint(c, x86_pmu.event_constraints) {
-			if (c->cmask != INTEL_ARCH_FIXED_MASK)
+			if (c->cmask != X86_RAW_EVENT_MASK)
 				continue;
 
-			c->idxmsk64 |= (1ULL << x86_pmu.num_events) - 1;
-			c->weight += x86_pmu.num_events;
+			c->idxmsk64 |= (1ULL << x86_pmu.num_counters) - 1;
+			c->weight += x86_pmu.num_counters;
 		}
 	}
 
 	pr_info("... version:                %d\n",     x86_pmu.version);
-	pr_info("... bit width:              %d\n",     x86_pmu.event_bits);
-	pr_info("... generic registers:      %d\n",     x86_pmu.num_events);
-	pr_info("... value mask:             %016Lx\n", x86_pmu.event_mask);
+	pr_info("... bit width:              %d\n",     x86_pmu.cntval_bits);
+	pr_info("... generic registers:      %d\n",     x86_pmu.num_counters);
+	pr_info("... value mask:             %016Lx\n", x86_pmu.cntval_mask);
 	pr_info("... max period:             %016Lx\n", x86_pmu.max_period);
-	pr_info("... fixed-purpose events:   %d\n",     x86_pmu.num_events_fixed);
-	pr_info("... event mask:             %016Lx\n", perf_event_mask);
+	pr_info("... fixed-purpose events:   %d\n",     x86_pmu.num_counters_fixed);
+	pr_info("... event mask:             %016Lx\n", x86_pmu.intel_ctrl);
 
 	perf_cpu_notifier(x86_pmu_notifier);
 }
@@ -1453,6 +1369,59 @@ static inline void x86_pmu_read(struct perf_event *event)
 	x86_perf_event_update(event);
 }
 
+/*
+ * Start group events scheduling transaction
+ * Set the flag to make pmu::enable() not perform the
+ * schedulability test, it will be performed at commit time
+ */
+static void x86_pmu_start_txn(const struct pmu *pmu)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	cpuc->group_flag |= PERF_EVENT_TXN_STARTED;
+}
+
+/*
+ * Stop group events scheduling transaction
+ * Clear the flag and pmu::enable() will perform the
+ * schedulability test.
+ */
+static void x86_pmu_cancel_txn(const struct pmu *pmu)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	cpuc->group_flag &= ~PERF_EVENT_TXN_STARTED;
+}
+
+/*
+ * Commit group events scheduling transaction
+ * Perform the group schedulability test as a whole
+ * Return 0 if success
+ */
+static int x86_pmu_commit_txn(const struct pmu *pmu)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	int assign[X86_PMC_IDX_MAX];
+	int n, ret;
+
+	n = cpuc->n_events;
+
+	if (!x86_pmu_initialized())
+		return -EAGAIN;
+
+	ret = x86_pmu.schedule_events(cpuc, n, assign);
+	if (ret)
+		return ret;
+
+	/*
+	 * copy new assignment, now we know it is possible
+	 * will be used by hw_perf_enable()
+	 */
+	memcpy(cpuc->assign, assign, n*sizeof(int));
+
+	return 0;
+}
+
 static const struct pmu pmu = {
 	.enable		= x86_pmu_enable,
 	.disable	= x86_pmu_disable,
@@ -1460,9 +1429,38 @@ static const struct pmu pmu = {
 	.stop		= x86_pmu_stop,
 	.read		= x86_pmu_read,
 	.unthrottle	= x86_pmu_unthrottle,
+	.start_txn	= x86_pmu_start_txn,
+	.cancel_txn	= x86_pmu_cancel_txn,
+	.commit_txn	= x86_pmu_commit_txn,
 };
 
 /*
+ * validate that we can schedule this event
+ */
+static int validate_event(struct perf_event *event)
+{
+	struct cpu_hw_events *fake_cpuc;
+	struct event_constraint *c;
+	int ret = 0;
+
+	fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO);
+	if (!fake_cpuc)
+		return -ENOMEM;
+
+	c = x86_pmu.get_event_constraints(fake_cpuc, event);
+
+	if (!c || !c->weight)
+		ret = -ENOSPC;
+
+	if (x86_pmu.put_event_constraints)
+		x86_pmu.put_event_constraints(fake_cpuc, event);
+
+	kfree(fake_cpuc);
+
+	return ret;
+}
+
+/*
  * validate a single event group
  *
  * validation include:
@@ -1502,7 +1500,7 @@ static int validate_group(struct perf_event *event)
 
 	fake_cpuc->n_events = n;
 
-	ret = x86_schedule_events(fake_cpuc, n, NULL);
+	ret = x86_pmu.schedule_events(fake_cpuc, n, NULL);
 
 out_free:
 	kfree(fake_cpuc);
@@ -1527,6 +1525,8 @@ const struct pmu *hw_perf_event_init(struct perf_event *event)
 
 		if (event->group_leader != event)
 			err = validate_group(event);
+		else
+			err = validate_event(event);
 
 		event->pmu = tmp;
 	}
@@ -1574,8 +1574,7 @@ static void backtrace_address(void *data, unsigned long addr, int reliable)
 {
 	struct perf_callchain_entry *entry = data;
 
-	if (reliable)
-		callchain_store(entry, addr);
+	callchain_store(entry, addr);
 }
 
 static const struct stacktrace_ops backtrace_ops = {
@@ -1597,41 +1596,6 @@ perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry)
 	dump_trace(NULL, regs, NULL, regs->bp, &backtrace_ops, entry);
 }
 
-/*
- * best effort, GUP based copy_from_user() that assumes IRQ or NMI context
- */
-static unsigned long
-copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
-{
-	unsigned long offset, addr = (unsigned long)from;
-	int type = in_nmi() ? KM_NMI : KM_IRQ0;
-	unsigned long size, len = 0;
-	struct page *page;
-	void *map;
-	int ret;
-
-	do {
-		ret = __get_user_pages_fast(addr, 1, 0, &page);
-		if (!ret)
-			break;
-
-		offset = addr & (PAGE_SIZE - 1);
-		size = min(PAGE_SIZE - offset, n - len);
-
-		map = kmap_atomic(page, type);
-		memcpy(to, map+offset, size);
-		kunmap_atomic(map, type);
-		put_page(page);
-
-		len  += size;
-		to   += size;
-		addr += size;
-
-	} while (len < n);
-
-	return len;
-}
-
 #ifdef CONFIG_COMPAT
 static inline int
 perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry)
@@ -1727,6 +1691,11 @@ struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
 {
 	struct perf_callchain_entry *entry;
 
+	if (perf_guest_cbs && perf_guest_cbs->is_in_guest()) {
+		/* TODO: We don't support guest os callchain now */
+		return NULL;
+	}
+
 	if (in_nmi())
 		entry = &__get_cpu_var(pmc_nmi_entry);
 	else
@@ -1750,3 +1719,37 @@ void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int ski
 	regs->cs = __KERNEL_CS;
 	local_save_flags(regs->flags);
 }
+
+unsigned long perf_instruction_pointer(struct pt_regs *regs)
+{
+	unsigned long ip;
+
+	if (perf_guest_cbs && perf_guest_cbs->is_in_guest())
+		ip = perf_guest_cbs->get_guest_ip();
+	else
+		ip = instruction_pointer(regs);
+
+	return ip;
+}
+
+unsigned long perf_misc_flags(struct pt_regs *regs)
+{
+	int misc = 0;
+
+	if (perf_guest_cbs && perf_guest_cbs->is_in_guest()) {
+		if (perf_guest_cbs->is_user_mode())
+			misc |= PERF_RECORD_MISC_GUEST_USER;
+		else
+			misc |= PERF_RECORD_MISC_GUEST_KERNEL;
+	} else {
+		if (user_mode(regs))
+			misc |= PERF_RECORD_MISC_USER;
+		else
+			misc |= PERF_RECORD_MISC_KERNEL;
+	}
+
+	if (regs->flags & PERF_EFLAGS_EXACT)
+		misc |= PERF_RECORD_MISC_EXACT_IP;
+
+	return misc;
+}
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index db6f7d4056e..611df11ba15 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -2,7 +2,7 @@
 
 static DEFINE_RAW_SPINLOCK(amd_nb_lock);
 
-static __initconst u64 amd_hw_cache_event_ids
+static __initconst const u64 amd_hw_cache_event_ids
 				[PERF_COUNT_HW_CACHE_MAX]
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
@@ -111,22 +111,19 @@ static u64 amd_pmu_event_map(int hw_event)
 	return amd_perfmon_event_map[hw_event];
 }
 
-static u64 amd_pmu_raw_event(u64 hw_event)
+static int amd_pmu_hw_config(struct perf_event *event)
 {
-#define K7_EVNTSEL_EVENT_MASK	0xF000000FFULL
-#define K7_EVNTSEL_UNIT_MASK	0x00000FF00ULL
-#define K7_EVNTSEL_EDGE_MASK	0x000040000ULL
-#define K7_EVNTSEL_INV_MASK	0x000800000ULL
-#define K7_EVNTSEL_REG_MASK	0x0FF000000ULL
-
-#define K7_EVNTSEL_MASK			\
-	(K7_EVNTSEL_EVENT_MASK |	\
-	 K7_EVNTSEL_UNIT_MASK  |	\
-	 K7_EVNTSEL_EDGE_MASK  |	\
-	 K7_EVNTSEL_INV_MASK   |	\
-	 K7_EVNTSEL_REG_MASK)
-
-	return hw_event & K7_EVNTSEL_MASK;
+	int ret = x86_pmu_hw_config(event);
+
+	if (ret)
+		return ret;
+
+	if (event->attr.type != PERF_TYPE_RAW)
+		return 0;
+
+	event->hw.config |= event->attr.config & AMD64_RAW_EVENT_MASK;
+
+	return 0;
 }
 
 /*
@@ -165,7 +162,7 @@ static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
 	 * be removed on one CPU at a time AND PMU is disabled
 	 * when we come here
 	 */
-	for (i = 0; i < x86_pmu.num_events; i++) {
+	for (i = 0; i < x86_pmu.num_counters; i++) {
 		if (nb->owners[i] == event) {
 			cmpxchg(nb->owners+i, event, NULL);
 			break;
@@ -215,7 +212,7 @@ amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 	struct amd_nb *nb = cpuc->amd_nb;
 	struct perf_event *old = NULL;
-	int max = x86_pmu.num_events;
+	int max = x86_pmu.num_counters;
 	int i, j, k = -1;
 
 	/*
@@ -293,7 +290,7 @@ static struct amd_nb *amd_alloc_nb(int cpu, int nb_id)
 	/*
 	 * initialize all possible NB constraints
 	 */
-	for (i = 0; i < x86_pmu.num_events; i++) {
+	for (i = 0; i < x86_pmu.num_counters; i++) {
 		__set_bit(i, nb->event_constraints[i].idxmsk);
 		nb->event_constraints[i].weight = 1;
 	}
@@ -371,21 +368,22 @@ static void amd_pmu_cpu_dead(int cpu)
 	raw_spin_unlock(&amd_nb_lock);
 }
 
-static __initconst struct x86_pmu amd_pmu = {
+static __initconst const struct x86_pmu amd_pmu = {
 	.name			= "AMD",
 	.handle_irq		= x86_pmu_handle_irq,
 	.disable_all		= x86_pmu_disable_all,
 	.enable_all		= x86_pmu_enable_all,
 	.enable			= x86_pmu_enable_event,
 	.disable		= x86_pmu_disable_event,
+	.hw_config		= amd_pmu_hw_config,
+	.schedule_events	= x86_schedule_events,
 	.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_events		= 4,
-	.event_bits		= 48,
-	.event_mask		= (1ULL << 48) - 1,
+	.num_counters		= 4,
+	.cntval_bits		= 48,
+	.cntval_mask		= (1ULL << 48) - 1,
 	.apic			= 1,
 	/* use highest bit to detect overflow */
 	.max_period		= (1ULL << 47) - 1,
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 9c794ac8783..fdbc652d3fe 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -88,7 +88,7 @@ static u64 intel_pmu_event_map(int hw_event)
 	return intel_perfmon_event_map[hw_event];
 }
 
-static __initconst u64 westmere_hw_cache_event_ids
+static __initconst const u64 westmere_hw_cache_event_ids
 				[PERF_COUNT_HW_CACHE_MAX]
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
@@ -179,7 +179,7 @@ static __initconst u64 westmere_hw_cache_event_ids
  },
 };
 
-static __initconst u64 nehalem_hw_cache_event_ids
+static __initconst const u64 nehalem_hw_cache_event_ids
 				[PERF_COUNT_HW_CACHE_MAX]
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
@@ -270,7 +270,7 @@ static __initconst u64 nehalem_hw_cache_event_ids
  },
 };
 
-static __initconst u64 core2_hw_cache_event_ids
+static __initconst const u64 core2_hw_cache_event_ids
 				[PERF_COUNT_HW_CACHE_MAX]
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
@@ -361,7 +361,7 @@ static __initconst u64 core2_hw_cache_event_ids
  },
 };
 
-static __initconst u64 atom_hw_cache_event_ids
+static __initconst const u64 atom_hw_cache_event_ids
 				[PERF_COUNT_HW_CACHE_MAX]
 				[PERF_COUNT_HW_CACHE_OP_MAX]
 				[PERF_COUNT_HW_CACHE_RESULT_MAX] =
@@ -452,60 +452,6 @@ static __initconst u64 atom_hw_cache_event_ids
  },
 };
 
-static u64 intel_pmu_raw_event(u64 hw_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_REG_MASK		0xFF000000ULL
-
-#define CORE_EVNTSEL_MASK		\
-	(INTEL_ARCH_EVTSEL_MASK |	\
-	 INTEL_ARCH_UNIT_MASK   |	\
-	 INTEL_ARCH_EDGE_MASK   |	\
-	 INTEL_ARCH_INV_MASK    |	\
-	 INTEL_ARCH_CNT_MASK)
-
-	return hw_event & CORE_EVNTSEL_MASK;
-}
-
-static void intel_pmu_enable_bts(u64 config)
-{
-	unsigned long debugctlmsr;
-
-	debugctlmsr = get_debugctlmsr();
-
-	debugctlmsr |= X86_DEBUGCTL_TR;
-	debugctlmsr |= X86_DEBUGCTL_BTS;
-	debugctlmsr |= X86_DEBUGCTL_BTINT;
-
-	if (!(config & ARCH_PERFMON_EVENTSEL_OS))
-		debugctlmsr |= X86_DEBUGCTL_BTS_OFF_OS;
-
-	if (!(config & ARCH_PERFMON_EVENTSEL_USR))
-		debugctlmsr |= X86_DEBUGCTL_BTS_OFF_USR;
-
-	update_debugctlmsr(debugctlmsr);
-}
-
-static void intel_pmu_disable_bts(void)
-{
-	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
-	unsigned long debugctlmsr;
-
-	if (!cpuc->ds)
-		return;
-
-	debugctlmsr = get_debugctlmsr();
-
-	debugctlmsr &=
-		~(X86_DEBUGCTL_TR | X86_DEBUGCTL_BTS | X86_DEBUGCTL_BTINT |
-		  X86_DEBUGCTL_BTS_OFF_OS | X86_DEBUGCTL_BTS_OFF_USR);
-
-	update_debugctlmsr(debugctlmsr);
-}
-
 static void intel_pmu_disable_all(void)
 {
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
@@ -514,12 +460,17 @@ static void intel_pmu_disable_all(void)
 
 	if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask))
 		intel_pmu_disable_bts();
+
+	intel_pmu_pebs_disable_all();
+	intel_pmu_lbr_disable_all();
 }
 
-static void intel_pmu_enable_all(void)
+static void intel_pmu_enable_all(int added)
 {
 	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 
+	intel_pmu_pebs_enable_all();
+	intel_pmu_lbr_enable_all();
 	wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl);
 
 	if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) {
@@ -533,6 +484,42 @@ static void intel_pmu_enable_all(void)
 	}
 }
 
+/*
+ * Workaround for:
+ *   Intel Errata AAK100 (model 26)
+ *   Intel Errata AAP53  (model 30)
+ *   Intel Errata BD53   (model 44)
+ *
+ * These chips need to be 'reset' when adding counters by programming
+ * the magic three (non counting) events 0x4300D2, 0x4300B1 and 0x4300B5
+ * either in sequence on the same PMC or on different PMCs.
+ */
+static void intel_pmu_nhm_enable_all(int added)
+{
+	if (added) {
+		struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+		int i;
+
+		wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 0, 0x4300D2);
+		wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 1, 0x4300B1);
+		wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 2, 0x4300B5);
+
+		wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0x3);
+		wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0x0);
+
+		for (i = 0; i < 3; i++) {
+			struct perf_event *event = cpuc->events[i];
+
+			if (!event)
+				continue;
+
+			__x86_pmu_enable_event(&event->hw,
+					       ARCH_PERFMON_EVENTSEL_ENABLE);
+		}
+	}
+	intel_pmu_enable_all(added);
+}
+
 static inline u64 intel_pmu_get_status(void)
 {
 	u64 status;
@@ -547,8 +534,7 @@ static inline void intel_pmu_ack_status(u64 ack)
 	wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, ack);
 }
 
-static inline void
-intel_pmu_disable_fixed(struct hw_perf_event *hwc)
+static void intel_pmu_disable_fixed(struct hw_perf_event *hwc)
 {
 	int idx = hwc->idx - X86_PMC_IDX_FIXED;
 	u64 ctrl_val, mask;
@@ -557,71 +543,10 @@ intel_pmu_disable_fixed(struct hw_perf_event *hwc)
 
 	rdmsrl(hwc->config_base, ctrl_val);
 	ctrl_val &= ~mask;
-	(void)checking_wrmsrl(hwc->config_base, ctrl_val);
-}
-
-static void intel_pmu_drain_bts_buffer(void)
-{
-	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
-	struct debug_store *ds = cpuc->ds;
-	struct bts_record {
-		u64	from;
-		u64	to;
-		u64	flags;
-	};
-	struct perf_event *event = cpuc->events[X86_PMC_IDX_FIXED_BTS];
-	struct bts_record *at, *top;
-	struct perf_output_handle handle;
-	struct perf_event_header header;
-	struct perf_sample_data data;
-	struct pt_regs regs;
-
-	if (!event)
-		return;
-
-	if (!ds)
-		return;
-
-	at  = (struct bts_record *)(unsigned long)ds->bts_buffer_base;
-	top = (struct bts_record *)(unsigned long)ds->bts_index;
-
-	if (top <= at)
-		return;
-
-	ds->bts_index = ds->bts_buffer_base;
-
-	perf_sample_data_init(&data, 0);
-
-	data.period	= event->hw.last_period;
-	regs.ip		= 0;
-
-	/*
-	 * Prepare a generic sample, i.e. fill in the invariant fields.
-	 * We will overwrite the from and to address before we output
-	 * the sample.
-	 */
-	perf_prepare_sample(&header, &data, event, &regs);
-
-	if (perf_output_begin(&handle, event,
-			      header.size * (top - at), 1, 1))
-		return;
-
-	for (; at < top; at++) {
-		data.ip		= at->from;
-		data.addr	= at->to;
-
-		perf_output_sample(&handle, &header, &data, event);
-	}
-
-	perf_output_end(&handle);
-
-	/* There's new data available. */
-	event->hw.interrupts++;
-	event->pending_kill = POLL_IN;
+	wrmsrl(hwc->config_base, ctrl_val);
 }
 
-static inline void
-intel_pmu_disable_event(struct perf_event *event)
+static void intel_pmu_disable_event(struct perf_event *event)
 {
 	struct hw_perf_event *hwc = &event->hw;
 
@@ -637,14 +562,15 @@ intel_pmu_disable_event(struct perf_event *event)
 	}
 
 	x86_pmu_disable_event(event);
+
+	if (unlikely(event->attr.precise_ip))
+		intel_pmu_pebs_disable(event);
 }
 
-static inline void
-intel_pmu_enable_fixed(struct hw_perf_event *hwc)
+static void intel_pmu_enable_fixed(struct hw_perf_event *hwc)
 {
 	int idx = hwc->idx - X86_PMC_IDX_FIXED;
 	u64 ctrl_val, bits, mask;
-	int err;
 
 	/*
 	 * Enable IRQ generation (0x8),
@@ -669,7 +595,7 @@ intel_pmu_enable_fixed(struct hw_perf_event *hwc)
 	rdmsrl(hwc->config_base, ctrl_val);
 	ctrl_val &= ~mask;
 	ctrl_val |= bits;
-	err = checking_wrmsrl(hwc->config_base, ctrl_val);
+	wrmsrl(hwc->config_base, ctrl_val);
 }
 
 static void intel_pmu_enable_event(struct perf_event *event)
@@ -689,7 +615,10 @@ static void intel_pmu_enable_event(struct perf_event *event)
 		return;
 	}
 
-	__x86_pmu_enable_event(hwc);
+	if (unlikely(event->attr.precise_ip))
+		intel_pmu_pebs_enable(event);
+
+	__x86_pmu_enable_event(hwc, ARCH_PERFMON_EVENTSEL_ENABLE);
 }
 
 /*
@@ -708,20 +637,20 @@ static void intel_pmu_reset(void)
 	unsigned long flags;
 	int idx;
 
-	if (!x86_pmu.num_events)
+	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_events; idx++) {
+	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_events_fixed; idx++) {
+	for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++)
 		checking_wrmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, 0ull);
-	}
+
 	if (ds)
 		ds->bts_index = ds->bts_buffer_base;
 
@@ -747,7 +676,7 @@ static int intel_pmu_handle_irq(struct pt_regs *regs)
 	intel_pmu_drain_bts_buffer();
 	status = intel_pmu_get_status();
 	if (!status) {
-		intel_pmu_enable_all();
+		intel_pmu_enable_all(0);
 		return 0;
 	}
 
@@ -762,6 +691,15 @@ again:
 
 	inc_irq_stat(apic_perf_irqs);
 	ack = status;
+
+	intel_pmu_lbr_read();
+
+	/*
+	 * PEBS overflow sets bit 62 in the global status register
+	 */
+	if (__test_and_clear_bit(62, (unsigned long *)&status))
+		x86_pmu.drain_pebs(regs);
+
 	for_each_set_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) {
 		struct perf_event *event = cpuc->events[bit];
 
@@ -787,26 +725,22 @@ again:
 		goto again;
 
 done:
-	intel_pmu_enable_all();
+	intel_pmu_enable_all(0);
 	return 1;
 }
 
-static struct event_constraint bts_constraint =
-	EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0);
-
 static struct event_constraint *
-intel_special_constraints(struct perf_event *event)
+intel_bts_constraints(struct perf_event *event)
 {
-	unsigned int hw_event;
-
-	hw_event = event->hw.config & INTEL_ARCH_EVENT_MASK;
+	struct hw_perf_event *hwc = &event->hw;
+	unsigned int hw_event, bts_event;
 
-	if (unlikely((hw_event ==
-		      x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) &&
-		     (event->hw.sample_period == 1))) {
+	hw_event = hwc->config & INTEL_ARCH_EVENT_MASK;
+	bts_event = x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS);
 
+	if (unlikely(hw_event == bts_event && hwc->sample_period == 1))
 		return &bts_constraint;
-	}
+
 	return NULL;
 }
 
@@ -815,24 +749,53 @@ intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event
 {
 	struct event_constraint *c;
 
-	c = intel_special_constraints(event);
+	c = intel_bts_constraints(event);
+	if (c)
+		return c;
+
+	c = intel_pebs_constraints(event);
 	if (c)
 		return c;
 
 	return x86_get_event_constraints(cpuc, event);
 }
 
-static __initconst struct x86_pmu core_pmu = {
+static int intel_pmu_hw_config(struct perf_event *event)
+{
+	int ret = x86_pmu_hw_config(event);
+
+	if (ret)
+		return ret;
+
+	if (event->attr.type != PERF_TYPE_RAW)
+		return 0;
+
+	if (!(event->attr.config & ARCH_PERFMON_EVENTSEL_ANY))
+		return 0;
+
+	if (x86_pmu.version < 3)
+		return -EINVAL;
+
+	if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
+	event->hw.config |= ARCH_PERFMON_EVENTSEL_ANY;
+
+	return 0;
+}
+
+static __initconst const struct x86_pmu core_pmu = {
 	.name			= "core",
 	.handle_irq		= x86_pmu_handle_irq,
 	.disable_all		= x86_pmu_disable_all,
 	.enable_all		= x86_pmu_enable_all,
 	.enable			= x86_pmu_enable_event,
 	.disable		= x86_pmu_disable_event,
+	.hw_config		= x86_pmu_hw_config,
+	.schedule_events	= x86_schedule_events,
 	.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),
 	.apic			= 1,
 	/*
@@ -845,17 +808,32 @@ static __initconst struct x86_pmu core_pmu = {
 	.event_constraints	= intel_core_event_constraints,
 };
 
-static __initconst struct x86_pmu intel_pmu = {
+static void intel_pmu_cpu_starting(int cpu)
+{
+	init_debug_store_on_cpu(cpu);
+	/*
+	 * Deal with CPUs that don't clear their LBRs on power-up.
+	 */
+	intel_pmu_lbr_reset();
+}
+
+static void intel_pmu_cpu_dying(int cpu)
+{
+	fini_debug_store_on_cpu(cpu);
+}
+
+static __initconst const 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_event,
 	.disable		= intel_pmu_disable_event,
+	.hw_config		= intel_pmu_hw_config,
+	.schedule_events	= x86_schedule_events,
 	.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),
 	.apic			= 1,
 	/*
@@ -864,14 +842,38 @@ static __initconst struct x86_pmu intel_pmu = {
 	 * the generic event period:
 	 */
 	.max_period		= (1ULL << 31) - 1,
-	.enable_bts		= intel_pmu_enable_bts,
-	.disable_bts		= intel_pmu_disable_bts,
 	.get_event_constraints	= intel_get_event_constraints,
 
-	.cpu_starting		= init_debug_store_on_cpu,
-	.cpu_dying		= fini_debug_store_on_cpu,
+	.cpu_starting		= intel_pmu_cpu_starting,
+	.cpu_dying		= intel_pmu_cpu_dying,
 };
 
+static void intel_clovertown_quirks(void)
+{
+	/*
+	 * PEBS is unreliable due to:
+	 *
+	 *   AJ67  - PEBS may experience CPL leaks
+	 *   AJ68  - PEBS PMI may be delayed by one event
+	 *   AJ69  - GLOBAL_STATUS[62] will only be set when DEBUGCTL[12]
+	 *   AJ106 - FREEZE_LBRS_ON_PMI doesn't work in combination with PEBS
+	 *
+	 * AJ67 could be worked around by restricting the OS/USR flags.
+	 * AJ69 could be worked around by setting PMU_FREEZE_ON_PMI.
+	 *
+	 * AJ106 could possibly be worked around by not allowing LBR
+	 *       usage from PEBS, including the fixup.
+	 * AJ68  could possibly be worked around by always programming
+	 * 	 a pebs_event_reset[0] value and coping with the lost events.
+	 *
+	 * But taken together it might just make sense to not enable PEBS on
+	 * these chips.
+	 */
+	printk(KERN_WARNING "PEBS disabled due to CPU errata.\n");
+	x86_pmu.pebs = 0;
+	x86_pmu.pebs_constraints = NULL;
+}
+
 static __init int intel_pmu_init(void)
 {
 	union cpuid10_edx edx;
@@ -881,12 +883,13 @@ static __init int intel_pmu_init(void)
 	int version;
 
 	if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
-		/* check for P6 processor family */
-	   if (boot_cpu_data.x86 == 6) {
-		return p6_pmu_init();
-	   } else {
+		switch (boot_cpu_data.x86) {
+		case 0x6:
+			return p6_pmu_init();
+		case 0xf:
+			return p4_pmu_init();
+		}
 		return -ENODEV;
-	   }
 	}
 
 	/*
@@ -904,16 +907,28 @@ static __init int intel_pmu_init(void)
 		x86_pmu = intel_pmu;
 
 	x86_pmu.version			= version;
-	x86_pmu.num_events		= eax.split.num_events;
-	x86_pmu.event_bits		= eax.split.bit_width;
-	x86_pmu.event_mask		= (1ULL << eax.split.bit_width) - 1;
+	x86_pmu.num_counters		= eax.split.num_counters;
+	x86_pmu.cntval_bits		= eax.split.bit_width;
+	x86_pmu.cntval_mask		= (1ULL << eax.split.bit_width) - 1;
 
 	/*
 	 * Quirk: v2 perfmon does not report fixed-purpose events, so
 	 * assume at least 3 events:
 	 */
 	if (version > 1)
-		x86_pmu.num_events_fixed = max((int)edx.split.num_events_fixed, 3);
+		x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3);
+
+	/*
+	 * v2 and above have a perf capabilities MSR
+	 */
+	if (version > 1) {
+		u64 capabilities;
+
+		rdmsrl(MSR_IA32_PERF_CAPABILITIES, capabilities);
+		x86_pmu.intel_cap.capabilities = capabilities;
+	}
+
+	intel_ds_init();
 
 	/*
 	 * Install the hw-cache-events table:
@@ -924,12 +939,15 @@ static __init int intel_pmu_init(void)
 		break;
 
 	case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */
+		x86_pmu.quirks = intel_clovertown_quirks;
 	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));
 
+		intel_pmu_lbr_init_core();
+
 		x86_pmu.event_constraints = intel_core2_event_constraints;
 		pr_cont("Core2 events, ");
 		break;
@@ -940,13 +958,19 @@ static __init int intel_pmu_init(void)
 		memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
 
+		intel_pmu_lbr_init_nhm();
+
 		x86_pmu.event_constraints = intel_nehalem_event_constraints;
-		pr_cont("Nehalem/Corei7 events, ");
+		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
+		pr_cont("Nehalem events, ");
 		break;
+
 	case 28: /* Atom */
 		memcpy(hw_cache_event_ids, atom_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
 
+		intel_pmu_lbr_init_atom();
+
 		x86_pmu.event_constraints = intel_gen_event_constraints;
 		pr_cont("Atom events, ");
 		break;
@@ -956,7 +980,10 @@ static __init int intel_pmu_init(void)
 		memcpy(hw_cache_event_ids, westmere_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
 
+		intel_pmu_lbr_init_nhm();
+
 		x86_pmu.event_constraints = intel_westmere_event_constraints;
+		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		pr_cont("Westmere events, ");
 		break;
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
new file mode 100644
index 00000000000..18018d1311c
--- /dev/null
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -0,0 +1,641 @@
+#ifdef CONFIG_CPU_SUP_INTEL
+
+/* The maximal number of PEBS events: */
+#define MAX_PEBS_EVENTS		4
+
+/* The size of a BTS record in bytes: */
+#define BTS_RECORD_SIZE		24
+
+#define BTS_BUFFER_SIZE		(PAGE_SIZE << 4)
+#define PEBS_BUFFER_SIZE	PAGE_SIZE
+
+/*
+ * pebs_record_32 for p4 and core not supported
+
+struct pebs_record_32 {
+	u32 flags, ip;
+	u32 ax, bc, cx, dx;
+	u32 si, di, bp, sp;
+};
+
+ */
+
+struct pebs_record_core {
+	u64 flags, ip;
+	u64 ax, bx, cx, dx;
+	u64 si, di, bp, sp;
+	u64 r8,  r9,  r10, r11;
+	u64 r12, r13, r14, r15;
+};
+
+struct pebs_record_nhm {
+	u64 flags, ip;
+	u64 ax, bx, cx, dx;
+	u64 si, di, bp, sp;
+	u64 r8,  r9,  r10, r11;
+	u64 r12, r13, r14, r15;
+	u64 status, dla, dse, lat;
+};
+
+/*
+ * A debug store configuration.
+ *
+ * We only support architectures that use 64bit fields.
+ */
+struct debug_store {
+	u64	bts_buffer_base;
+	u64	bts_index;
+	u64	bts_absolute_maximum;
+	u64	bts_interrupt_threshold;
+	u64	pebs_buffer_base;
+	u64	pebs_index;
+	u64	pebs_absolute_maximum;
+	u64	pebs_interrupt_threshold;
+	u64	pebs_event_reset[MAX_PEBS_EVENTS];
+};
+
+static void init_debug_store_on_cpu(int cpu)
+{
+	struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
+
+	if (!ds)
+		return;
+
+	wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA,
+		     (u32)((u64)(unsigned long)ds),
+		     (u32)((u64)(unsigned long)ds >> 32));
+}
+
+static void fini_debug_store_on_cpu(int cpu)
+{
+	if (!per_cpu(cpu_hw_events, cpu).ds)
+		return;
+
+	wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, 0, 0);
+}
+
+static void release_ds_buffers(void)
+{
+	int cpu;
+
+	if (!x86_pmu.bts && !x86_pmu.pebs)
+		return;
+
+	get_online_cpus();
+
+	for_each_online_cpu(cpu)
+		fini_debug_store_on_cpu(cpu);
+
+	for_each_possible_cpu(cpu) {
+		struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
+
+		if (!ds)
+			continue;
+
+		per_cpu(cpu_hw_events, cpu).ds = NULL;
+
+		kfree((void *)(unsigned long)ds->pebs_buffer_base);
+		kfree((void *)(unsigned long)ds->bts_buffer_base);
+		kfree(ds);
+	}
+
+	put_online_cpus();
+}
+
+static int reserve_ds_buffers(void)
+{
+	int cpu, err = 0;
+
+	if (!x86_pmu.bts && !x86_pmu.pebs)
+		return 0;
+
+	get_online_cpus();
+
+	for_each_possible_cpu(cpu) {
+		struct debug_store *ds;
+		void *buffer;
+		int max, thresh;
+
+		err = -ENOMEM;
+		ds = kzalloc(sizeof(*ds), GFP_KERNEL);
+		if (unlikely(!ds))
+			break;
+		per_cpu(cpu_hw_events, cpu).ds = ds;
+
+		if (x86_pmu.bts) {
+			buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL);
+			if (unlikely(!buffer))
+				break;
+
+			max = BTS_BUFFER_SIZE / BTS_RECORD_SIZE;
+			thresh = max / 16;
+
+			ds->bts_buffer_base = (u64)(unsigned long)buffer;
+			ds->bts_index = ds->bts_buffer_base;
+			ds->bts_absolute_maximum = ds->bts_buffer_base +
+				max * BTS_RECORD_SIZE;
+			ds->bts_interrupt_threshold = ds->bts_absolute_maximum -
+				thresh * BTS_RECORD_SIZE;
+		}
+
+		if (x86_pmu.pebs) {
+			buffer = kzalloc(PEBS_BUFFER_SIZE, GFP_KERNEL);
+			if (unlikely(!buffer))
+				break;
+
+			max = PEBS_BUFFER_SIZE / x86_pmu.pebs_record_size;
+
+			ds->pebs_buffer_base = (u64)(unsigned long)buffer;
+			ds->pebs_index = ds->pebs_buffer_base;
+			ds->pebs_absolute_maximum = ds->pebs_buffer_base +
+				max * x86_pmu.pebs_record_size;
+			/*
+			 * Always use single record PEBS
+			 */
+			ds->pebs_interrupt_threshold = ds->pebs_buffer_base +
+				x86_pmu.pebs_record_size;
+		}
+
+		err = 0;
+	}
+
+	if (err)
+		release_ds_buffers();
+	else {
+		for_each_online_cpu(cpu)
+			init_debug_store_on_cpu(cpu);
+	}
+
+	put_online_cpus();
+
+	return err;
+}
+
+/*
+ * BTS
+ */
+
+static struct event_constraint bts_constraint =
+	EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0);
+
+static void intel_pmu_enable_bts(u64 config)
+{
+	unsigned long debugctlmsr;
+
+	debugctlmsr = get_debugctlmsr();
+
+	debugctlmsr |= DEBUGCTLMSR_TR;
+	debugctlmsr |= DEBUGCTLMSR_BTS;
+	debugctlmsr |= DEBUGCTLMSR_BTINT;
+
+	if (!(config & ARCH_PERFMON_EVENTSEL_OS))
+		debugctlmsr |= DEBUGCTLMSR_BTS_OFF_OS;
+
+	if (!(config & ARCH_PERFMON_EVENTSEL_USR))
+		debugctlmsr |= DEBUGCTLMSR_BTS_OFF_USR;
+
+	update_debugctlmsr(debugctlmsr);
+}
+
+static void intel_pmu_disable_bts(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	unsigned long debugctlmsr;
+
+	if (!cpuc->ds)
+		return;
+
+	debugctlmsr = get_debugctlmsr();
+
+	debugctlmsr &=
+		~(DEBUGCTLMSR_TR | DEBUGCTLMSR_BTS | DEBUGCTLMSR_BTINT |
+		  DEBUGCTLMSR_BTS_OFF_OS | DEBUGCTLMSR_BTS_OFF_USR);
+
+	update_debugctlmsr(debugctlmsr);
+}
+
+static void intel_pmu_drain_bts_buffer(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	struct debug_store *ds = cpuc->ds;
+	struct bts_record {
+		u64	from;
+		u64	to;
+		u64	flags;
+	};
+	struct perf_event *event = cpuc->events[X86_PMC_IDX_FIXED_BTS];
+	struct bts_record *at, *top;
+	struct perf_output_handle handle;
+	struct perf_event_header header;
+	struct perf_sample_data data;
+	struct pt_regs regs;
+
+	if (!event)
+		return;
+
+	if (!ds)
+		return;
+
+	at  = (struct bts_record *)(unsigned long)ds->bts_buffer_base;
+	top = (struct bts_record *)(unsigned long)ds->bts_index;
+
+	if (top <= at)
+		return;
+
+	ds->bts_index = ds->bts_buffer_base;
+
+	perf_sample_data_init(&data, 0);
+	data.period = event->hw.last_period;
+	regs.ip     = 0;
+
+	/*
+	 * Prepare a generic sample, i.e. fill in the invariant fields.
+	 * We will overwrite the from and to address before we output
+	 * the sample.
+	 */
+	perf_prepare_sample(&header, &data, event, &regs);
+
+	if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1))
+		return;
+
+	for (; at < top; at++) {
+		data.ip		= at->from;
+		data.addr	= at->to;
+
+		perf_output_sample(&handle, &header, &data, event);
+	}
+
+	perf_output_end(&handle);
+
+	/* There's new data available. */
+	event->hw.interrupts++;
+	event->pending_kill = POLL_IN;
+}
+
+/*
+ * PEBS
+ */
+
+static struct event_constraint intel_core_pebs_events[] = {
+	PEBS_EVENT_CONSTRAINT(0x00c0, 0x1), /* INSTR_RETIRED.ANY */
+	PEBS_EVENT_CONSTRAINT(0xfec1, 0x1), /* X87_OPS_RETIRED.ANY */
+	PEBS_EVENT_CONSTRAINT(0x00c5, 0x1), /* BR_INST_RETIRED.MISPRED */
+	PEBS_EVENT_CONSTRAINT(0x1fc7, 0x1), /* SIMD_INST_RETURED.ANY */
+	PEBS_EVENT_CONSTRAINT(0x01cb, 0x1), /* MEM_LOAD_RETIRED.L1D_MISS */
+	PEBS_EVENT_CONSTRAINT(0x02cb, 0x1), /* MEM_LOAD_RETIRED.L1D_LINE_MISS */
+	PEBS_EVENT_CONSTRAINT(0x04cb, 0x1), /* MEM_LOAD_RETIRED.L2_MISS */
+	PEBS_EVENT_CONSTRAINT(0x08cb, 0x1), /* MEM_LOAD_RETIRED.L2_LINE_MISS */
+	PEBS_EVENT_CONSTRAINT(0x10cb, 0x1), /* MEM_LOAD_RETIRED.DTLB_MISS */
+	EVENT_CONSTRAINT_END
+};
+
+static struct event_constraint intel_nehalem_pebs_events[] = {
+	PEBS_EVENT_CONSTRAINT(0x00c0, 0xf), /* INSTR_RETIRED.ANY */
+	PEBS_EVENT_CONSTRAINT(0xfec1, 0xf), /* X87_OPS_RETIRED.ANY */
+	PEBS_EVENT_CONSTRAINT(0x00c5, 0xf), /* BR_INST_RETIRED.MISPRED */
+	PEBS_EVENT_CONSTRAINT(0x1fc7, 0xf), /* SIMD_INST_RETURED.ANY */
+	PEBS_EVENT_CONSTRAINT(0x01cb, 0xf), /* MEM_LOAD_RETIRED.L1D_MISS */
+	PEBS_EVENT_CONSTRAINT(0x02cb, 0xf), /* MEM_LOAD_RETIRED.L1D_LINE_MISS */
+	PEBS_EVENT_CONSTRAINT(0x04cb, 0xf), /* MEM_LOAD_RETIRED.L2_MISS */
+	PEBS_EVENT_CONSTRAINT(0x08cb, 0xf), /* MEM_LOAD_RETIRED.L2_LINE_MISS */
+	PEBS_EVENT_CONSTRAINT(0x10cb, 0xf), /* MEM_LOAD_RETIRED.DTLB_MISS */
+	EVENT_CONSTRAINT_END
+};
+
+static struct event_constraint *
+intel_pebs_constraints(struct perf_event *event)
+{
+	struct event_constraint *c;
+
+	if (!event->attr.precise_ip)
+		return NULL;
+
+	if (x86_pmu.pebs_constraints) {
+		for_each_event_constraint(c, x86_pmu.pebs_constraints) {
+			if ((event->hw.config & c->cmask) == c->code)
+				return c;
+		}
+	}
+
+	return &emptyconstraint;
+}
+
+static void intel_pmu_pebs_enable(struct perf_event *event)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	struct hw_perf_event *hwc = &event->hw;
+
+	hwc->config &= ~ARCH_PERFMON_EVENTSEL_INT;
+
+	cpuc->pebs_enabled |= 1ULL << hwc->idx;
+	WARN_ON_ONCE(cpuc->enabled);
+
+	if (x86_pmu.intel_cap.pebs_trap && event->attr.precise_ip > 1)
+		intel_pmu_lbr_enable(event);
+}
+
+static void intel_pmu_pebs_disable(struct perf_event *event)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	struct hw_perf_event *hwc = &event->hw;
+
+	cpuc->pebs_enabled &= ~(1ULL << hwc->idx);
+	if (cpuc->enabled)
+		wrmsrl(MSR_IA32_PEBS_ENABLE, cpuc->pebs_enabled);
+
+	hwc->config |= ARCH_PERFMON_EVENTSEL_INT;
+
+	if (x86_pmu.intel_cap.pebs_trap && event->attr.precise_ip > 1)
+		intel_pmu_lbr_disable(event);
+}
+
+static void intel_pmu_pebs_enable_all(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (cpuc->pebs_enabled)
+		wrmsrl(MSR_IA32_PEBS_ENABLE, cpuc->pebs_enabled);
+}
+
+static void intel_pmu_pebs_disable_all(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (cpuc->pebs_enabled)
+		wrmsrl(MSR_IA32_PEBS_ENABLE, 0);
+}
+
+#include <asm/insn.h>
+
+static inline bool kernel_ip(unsigned long ip)
+{
+#ifdef CONFIG_X86_32
+	return ip > PAGE_OFFSET;
+#else
+	return (long)ip < 0;
+#endif
+}
+
+static int intel_pmu_pebs_fixup_ip(struct pt_regs *regs)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	unsigned long from = cpuc->lbr_entries[0].from;
+	unsigned long old_to, to = cpuc->lbr_entries[0].to;
+	unsigned long ip = regs->ip;
+
+	/*
+	 * We don't need to fixup if the PEBS assist is fault like
+	 */
+	if (!x86_pmu.intel_cap.pebs_trap)
+		return 1;
+
+	/*
+	 * No LBR entry, no basic block, no rewinding
+	 */
+	if (!cpuc->lbr_stack.nr || !from || !to)
+		return 0;
+
+	/*
+	 * Basic blocks should never cross user/kernel boundaries
+	 */
+	if (kernel_ip(ip) != kernel_ip(to))
+		return 0;
+
+	/*
+	 * unsigned math, either ip is before the start (impossible) or
+	 * the basic block is larger than 1 page (sanity)
+	 */
+	if ((ip - to) > PAGE_SIZE)
+		return 0;
+
+	/*
+	 * We sampled a branch insn, rewind using the LBR stack
+	 */
+	if (ip == to) {
+		regs->ip = from;
+		return 1;
+	}
+
+	do {
+		struct insn insn;
+		u8 buf[MAX_INSN_SIZE];
+		void *kaddr;
+
+		old_to = to;
+		if (!kernel_ip(ip)) {
+			int bytes, size = MAX_INSN_SIZE;
+
+			bytes = copy_from_user_nmi(buf, (void __user *)to, size);
+			if (bytes != size)
+				return 0;
+
+			kaddr = buf;
+		} else
+			kaddr = (void *)to;
+
+		kernel_insn_init(&insn, kaddr);
+		insn_get_length(&insn);
+		to += insn.length;
+	} while (to < ip);
+
+	if (to == ip) {
+		regs->ip = old_to;
+		return 1;
+	}
+
+	/*
+	 * Even though we decoded the basic block, the instruction stream
+	 * never matched the given IP, either the TO or the IP got corrupted.
+	 */
+	return 0;
+}
+
+static int intel_pmu_save_and_restart(struct perf_event *event);
+
+static void __intel_pmu_pebs_event(struct perf_event *event,
+				   struct pt_regs *iregs, void *__pebs)
+{
+	/*
+	 * We cast to pebs_record_core since that is a subset of
+	 * both formats and we don't use the other fields in this
+	 * routine.
+	 */
+	struct pebs_record_core *pebs = __pebs;
+	struct perf_sample_data data;
+	struct pt_regs regs;
+
+	if (!intel_pmu_save_and_restart(event))
+		return;
+
+	perf_sample_data_init(&data, 0);
+	data.period = event->hw.last_period;
+
+	/*
+	 * We use the interrupt regs as a base because the PEBS record
+	 * does not contain a full regs set, specifically it seems to
+	 * lack segment descriptors, which get used by things like
+	 * user_mode().
+	 *
+	 * In the simple case fix up only the IP and BP,SP regs, for
+	 * PERF_SAMPLE_IP and PERF_SAMPLE_CALLCHAIN to function properly.
+	 * A possible PERF_SAMPLE_REGS will have to transfer all regs.
+	 */
+	regs = *iregs;
+	regs.ip = pebs->ip;
+	regs.bp = pebs->bp;
+	regs.sp = pebs->sp;
+
+	if (event->attr.precise_ip > 1 && intel_pmu_pebs_fixup_ip(&regs))
+		regs.flags |= PERF_EFLAGS_EXACT;
+	else
+		regs.flags &= ~PERF_EFLAGS_EXACT;
+
+	if (perf_event_overflow(event, 1, &data, &regs))
+		x86_pmu_stop(event);
+}
+
+static void intel_pmu_drain_pebs_core(struct pt_regs *iregs)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	struct debug_store *ds = cpuc->ds;
+	struct perf_event *event = cpuc->events[0]; /* PMC0 only */
+	struct pebs_record_core *at, *top;
+	int n;
+
+	if (!ds || !x86_pmu.pebs)
+		return;
+
+	at  = (struct pebs_record_core *)(unsigned long)ds->pebs_buffer_base;
+	top = (struct pebs_record_core *)(unsigned long)ds->pebs_index;
+
+	/*
+	 * Whatever else happens, drain the thing
+	 */
+	ds->pebs_index = ds->pebs_buffer_base;
+
+	if (!test_bit(0, cpuc->active_mask))
+		return;
+
+	WARN_ON_ONCE(!event);
+
+	if (!event->attr.precise_ip)
+		return;
+
+	n = top - at;
+	if (n <= 0)
+		return;
+
+	/*
+	 * Should not happen, we program the threshold at 1 and do not
+	 * set a reset value.
+	 */
+	WARN_ON_ONCE(n > 1);
+	at += n - 1;
+
+	__intel_pmu_pebs_event(event, iregs, at);
+}
+
+static void intel_pmu_drain_pebs_nhm(struct pt_regs *iregs)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	struct debug_store *ds = cpuc->ds;
+	struct pebs_record_nhm *at, *top;
+	struct perf_event *event = NULL;
+	u64 status = 0;
+	int bit, n;
+
+	if (!ds || !x86_pmu.pebs)
+		return;
+
+	at  = (struct pebs_record_nhm *)(unsigned long)ds->pebs_buffer_base;
+	top = (struct pebs_record_nhm *)(unsigned long)ds->pebs_index;
+
+	ds->pebs_index = ds->pebs_buffer_base;
+
+	n = top - at;
+	if (n <= 0)
+		return;
+
+	/*
+	 * Should not happen, we program the threshold at 1 and do not
+	 * set a reset value.
+	 */
+	WARN_ON_ONCE(n > MAX_PEBS_EVENTS);
+
+	for ( ; at < top; at++) {
+		for_each_set_bit(bit, (unsigned long *)&at->status, MAX_PEBS_EVENTS) {
+			event = cpuc->events[bit];
+			if (!test_bit(bit, cpuc->active_mask))
+				continue;
+
+			WARN_ON_ONCE(!event);
+
+			if (!event->attr.precise_ip)
+				continue;
+
+			if (__test_and_set_bit(bit, (unsigned long *)&status))
+				continue;
+
+			break;
+		}
+
+		if (!event || bit >= MAX_PEBS_EVENTS)
+			continue;
+
+		__intel_pmu_pebs_event(event, iregs, at);
+	}
+}
+
+/*
+ * BTS, PEBS probe and setup
+ */
+
+static void intel_ds_init(void)
+{
+	/*
+	 * No support for 32bit formats
+	 */
+	if (!boot_cpu_has(X86_FEATURE_DTES64))
+		return;
+
+	x86_pmu.bts  = boot_cpu_has(X86_FEATURE_BTS);
+	x86_pmu.pebs = boot_cpu_has(X86_FEATURE_PEBS);
+	if (x86_pmu.pebs) {
+		char pebs_type = x86_pmu.intel_cap.pebs_trap ?  '+' : '-';
+		int format = x86_pmu.intel_cap.pebs_format;
+
+		switch (format) {
+		case 0:
+			printk(KERN_CONT "PEBS fmt0%c, ", pebs_type);
+			x86_pmu.pebs_record_size = sizeof(struct pebs_record_core);
+			x86_pmu.drain_pebs = intel_pmu_drain_pebs_core;
+			x86_pmu.pebs_constraints = intel_core_pebs_events;
+			break;
+
+		case 1:
+			printk(KERN_CONT "PEBS fmt1%c, ", pebs_type);
+			x86_pmu.pebs_record_size = sizeof(struct pebs_record_nhm);
+			x86_pmu.drain_pebs = intel_pmu_drain_pebs_nhm;
+			x86_pmu.pebs_constraints = intel_nehalem_pebs_events;
+			break;
+
+		default:
+			printk(KERN_CONT "no PEBS fmt%d%c, ", format, pebs_type);
+			x86_pmu.pebs = 0;
+			break;
+		}
+	}
+}
+
+#else /* CONFIG_CPU_SUP_INTEL */
+
+static int reserve_ds_buffers(void)
+{
+	return 0;
+}
+
+static void release_ds_buffers(void)
+{
+}
+
+#endif /* CONFIG_CPU_SUP_INTEL */
diff --git a/arch/x86/kernel/cpu/perf_event_intel_lbr.c b/arch/x86/kernel/cpu/perf_event_intel_lbr.c
new file mode 100644
index 00000000000..d202c1bece1
--- /dev/null
+++ b/arch/x86/kernel/cpu/perf_event_intel_lbr.c
@@ -0,0 +1,218 @@
+#ifdef CONFIG_CPU_SUP_INTEL
+
+enum {
+	LBR_FORMAT_32		= 0x00,
+	LBR_FORMAT_LIP		= 0x01,
+	LBR_FORMAT_EIP		= 0x02,
+	LBR_FORMAT_EIP_FLAGS	= 0x03,
+};
+
+/*
+ * We only support LBR implementations that have FREEZE_LBRS_ON_PMI
+ * otherwise it becomes near impossible to get a reliable stack.
+ */
+
+static void __intel_pmu_lbr_enable(void)
+{
+	u64 debugctl;
+
+	rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+	debugctl |= (DEBUGCTLMSR_LBR | DEBUGCTLMSR_FREEZE_LBRS_ON_PMI);
+	wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+}
+
+static void __intel_pmu_lbr_disable(void)
+{
+	u64 debugctl;
+
+	rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+	debugctl &= ~(DEBUGCTLMSR_LBR | DEBUGCTLMSR_FREEZE_LBRS_ON_PMI);
+	wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
+}
+
+static void intel_pmu_lbr_reset_32(void)
+{
+	int i;
+
+	for (i = 0; i < x86_pmu.lbr_nr; i++)
+		wrmsrl(x86_pmu.lbr_from + i, 0);
+}
+
+static void intel_pmu_lbr_reset_64(void)
+{
+	int i;
+
+	for (i = 0; i < x86_pmu.lbr_nr; i++) {
+		wrmsrl(x86_pmu.lbr_from + i, 0);
+		wrmsrl(x86_pmu.lbr_to   + i, 0);
+	}
+}
+
+static void intel_pmu_lbr_reset(void)
+{
+	if (!x86_pmu.lbr_nr)
+		return;
+
+	if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_32)
+		intel_pmu_lbr_reset_32();
+	else
+		intel_pmu_lbr_reset_64();
+}
+
+static void intel_pmu_lbr_enable(struct perf_event *event)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (!x86_pmu.lbr_nr)
+		return;
+
+	WARN_ON_ONCE(cpuc->enabled);
+
+	/*
+	 * Reset the LBR stack if we changed task context to
+	 * avoid data leaks.
+	 */
+
+	if (event->ctx->task && cpuc->lbr_context != event->ctx) {
+		intel_pmu_lbr_reset();
+		cpuc->lbr_context = event->ctx;
+	}
+
+	cpuc->lbr_users++;
+}
+
+static void intel_pmu_lbr_disable(struct perf_event *event)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (!x86_pmu.lbr_nr)
+		return;
+
+	cpuc->lbr_users--;
+	WARN_ON_ONCE(cpuc->lbr_users < 0);
+
+	if (cpuc->enabled && !cpuc->lbr_users)
+		__intel_pmu_lbr_disable();
+}
+
+static void intel_pmu_lbr_enable_all(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (cpuc->lbr_users)
+		__intel_pmu_lbr_enable();
+}
+
+static void intel_pmu_lbr_disable_all(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (cpuc->lbr_users)
+		__intel_pmu_lbr_disable();
+}
+
+static inline u64 intel_pmu_lbr_tos(void)
+{
+	u64 tos;
+
+	rdmsrl(x86_pmu.lbr_tos, tos);
+
+	return tos;
+}
+
+static void intel_pmu_lbr_read_32(struct cpu_hw_events *cpuc)
+{
+	unsigned long mask = x86_pmu.lbr_nr - 1;
+	u64 tos = intel_pmu_lbr_tos();
+	int i;
+
+	for (i = 0; i < x86_pmu.lbr_nr; i++) {
+		unsigned long lbr_idx = (tos - i) & mask;
+		union {
+			struct {
+				u32 from;
+				u32 to;
+			};
+			u64     lbr;
+		} msr_lastbranch;
+
+		rdmsrl(x86_pmu.lbr_from + lbr_idx, msr_lastbranch.lbr);
+
+		cpuc->lbr_entries[i].from  = msr_lastbranch.from;
+		cpuc->lbr_entries[i].to    = msr_lastbranch.to;
+		cpuc->lbr_entries[i].flags = 0;
+	}
+	cpuc->lbr_stack.nr = i;
+}
+
+#define LBR_FROM_FLAG_MISPRED  (1ULL << 63)
+
+/*
+ * Due to lack of segmentation in Linux the effective address (offset)
+ * is the same as the linear address, allowing us to merge the LIP and EIP
+ * LBR formats.
+ */
+static void intel_pmu_lbr_read_64(struct cpu_hw_events *cpuc)
+{
+	unsigned long mask = x86_pmu.lbr_nr - 1;
+	int lbr_format = x86_pmu.intel_cap.lbr_format;
+	u64 tos = intel_pmu_lbr_tos();
+	int i;
+
+	for (i = 0; i < x86_pmu.lbr_nr; i++) {
+		unsigned long lbr_idx = (tos - i) & mask;
+		u64 from, to, flags = 0;
+
+		rdmsrl(x86_pmu.lbr_from + lbr_idx, from);
+		rdmsrl(x86_pmu.lbr_to   + lbr_idx, to);
+
+		if (lbr_format == LBR_FORMAT_EIP_FLAGS) {
+			flags = !!(from & LBR_FROM_FLAG_MISPRED);
+			from = (u64)((((s64)from) << 1) >> 1);
+		}
+
+		cpuc->lbr_entries[i].from  = from;
+		cpuc->lbr_entries[i].to    = to;
+		cpuc->lbr_entries[i].flags = flags;
+	}
+	cpuc->lbr_stack.nr = i;
+}
+
+static void intel_pmu_lbr_read(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+
+	if (!cpuc->lbr_users)
+		return;
+
+	if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_32)
+		intel_pmu_lbr_read_32(cpuc);
+	else
+		intel_pmu_lbr_read_64(cpuc);
+}
+
+static void intel_pmu_lbr_init_core(void)
+{
+	x86_pmu.lbr_nr     = 4;
+	x86_pmu.lbr_tos    = 0x01c9;
+	x86_pmu.lbr_from   = 0x40;
+	x86_pmu.lbr_to     = 0x60;
+}
+
+static void intel_pmu_lbr_init_nhm(void)
+{
+	x86_pmu.lbr_nr     = 16;
+	x86_pmu.lbr_tos    = 0x01c9;
+	x86_pmu.lbr_from   = 0x680;
+	x86_pmu.lbr_to     = 0x6c0;
+}
+
+static void intel_pmu_lbr_init_atom(void)
+{
+	x86_pmu.lbr_nr	   = 8;
+	x86_pmu.lbr_tos    = 0x01c9;
+	x86_pmu.lbr_from   = 0x40;
+	x86_pmu.lbr_to     = 0x60;
+}
+
+#endif /* CONFIG_CPU_SUP_INTEL */
diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c
new file mode 100644
index 00000000000..424fc8de68e
--- /dev/null
+++ b/arch/x86/kernel/cpu/perf_event_p4.c
@@ -0,0 +1,857 @@
+/*
+ * Netburst Perfomance Events (P4, old Xeon)
+ *
+ *  Copyright (C) 2010 Parallels, Inc., Cyrill Gorcunov <gorcunov@openvz.org>
+ *  Copyright (C) 2010 Intel Corporation, Lin Ming <ming.m.lin@intel.com>
+ *
+ *  For licencing details see kernel-base/COPYING
+ */
+
+#ifdef CONFIG_CPU_SUP_INTEL
+
+#include <asm/perf_event_p4.h>
+
+#define P4_CNTR_LIMIT 3
+/*
+ * array indices: 0,1 - HT threads, used with HT enabled cpu
+ */
+struct p4_event_bind {
+	unsigned int opcode;			/* Event code and ESCR selector */
+	unsigned int escr_msr[2];		/* ESCR MSR for this event */
+	char cntr[2][P4_CNTR_LIMIT];		/* counter index (offset), -1 on abscence */
+};
+
+struct p4_cache_event_bind {
+	unsigned int metric_pebs;
+	unsigned int metric_vert;
+};
+
+#define P4_GEN_CACHE_EVENT_BIND(name)		\
+	[P4_CACHE__##name] = {			\
+		.metric_pebs = P4_PEBS__##name,	\
+		.metric_vert = P4_VERT__##name,	\
+	}
+
+static struct p4_cache_event_bind p4_cache_event_bind_map[] = {
+	P4_GEN_CACHE_EVENT_BIND(1stl_cache_load_miss_retired),
+	P4_GEN_CACHE_EVENT_BIND(2ndl_cache_load_miss_retired),
+	P4_GEN_CACHE_EVENT_BIND(dtlb_load_miss_retired),
+	P4_GEN_CACHE_EVENT_BIND(dtlb_store_miss_retired),
+};
+
+/*
+ * Note that we don't use CCCR1 here, there is an
+ * exception for P4_BSQ_ALLOCATION but we just have
+ * no workaround
+ *
+ * consider this binding as resources which particular
+ * event may borrow, it doesn't contain EventMask,
+ * Tags and friends -- they are left to a caller
+ */
+static struct p4_event_bind p4_event_bind_map[] = {
+	[P4_EVENT_TC_DELIVER_MODE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_TC_DELIVER_MODE),
+		.escr_msr	= { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_BPU_FETCH_REQUEST] = {
+		.opcode		= P4_OPCODE(P4_EVENT_BPU_FETCH_REQUEST),
+		.escr_msr	= { MSR_P4_BPU_ESCR0, MSR_P4_BPU_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_ITLB_REFERENCE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_ITLB_REFERENCE),
+		.escr_msr	= { MSR_P4_ITLB_ESCR0, MSR_P4_ITLB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_MEMORY_CANCEL] = {
+		.opcode		= P4_OPCODE(P4_EVENT_MEMORY_CANCEL),
+		.escr_msr	= { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_MEMORY_COMPLETE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_MEMORY_COMPLETE),
+		.escr_msr	= { MSR_P4_SAAT_ESCR0 , MSR_P4_SAAT_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_LOAD_PORT_REPLAY] = {
+		.opcode		= P4_OPCODE(P4_EVENT_LOAD_PORT_REPLAY),
+		.escr_msr	= { MSR_P4_SAAT_ESCR0, MSR_P4_SAAT_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_STORE_PORT_REPLAY] = {
+		.opcode		= P4_OPCODE(P4_EVENT_STORE_PORT_REPLAY),
+		.escr_msr	= { MSR_P4_SAAT_ESCR0 ,  MSR_P4_SAAT_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_MOB_LOAD_REPLAY] = {
+		.opcode		= P4_OPCODE(P4_EVENT_MOB_LOAD_REPLAY),
+		.escr_msr	= { MSR_P4_MOB_ESCR0, MSR_P4_MOB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_PAGE_WALK_TYPE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_PAGE_WALK_TYPE),
+		.escr_msr	= { MSR_P4_PMH_ESCR0, MSR_P4_PMH_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_BSQ_CACHE_REFERENCE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_BSQ_CACHE_REFERENCE),
+		.escr_msr	= { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_IOQ_ALLOCATION] = {
+		.opcode		= P4_OPCODE(P4_EVENT_IOQ_ALLOCATION),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_IOQ_ACTIVE_ENTRIES] = {	/* shared ESCR */
+		.opcode		= P4_OPCODE(P4_EVENT_IOQ_ACTIVE_ENTRIES),
+		.escr_msr	= { MSR_P4_FSB_ESCR1,  MSR_P4_FSB_ESCR1 },
+		.cntr		= { {2, -1, -1}, {3, -1, -1} },
+	},
+	[P4_EVENT_FSB_DATA_ACTIVITY] = {
+		.opcode		= P4_OPCODE(P4_EVENT_FSB_DATA_ACTIVITY),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_BSQ_ALLOCATION] = {		/* shared ESCR, broken CCCR1 */
+		.opcode		= P4_OPCODE(P4_EVENT_BSQ_ALLOCATION),
+		.escr_msr	= { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR0 },
+		.cntr		= { {0, -1, -1}, {1, -1, -1} },
+	},
+	[P4_EVENT_BSQ_ACTIVE_ENTRIES] = {	/* shared ESCR */
+		.opcode		= P4_OPCODE(P4_EVENT_BSQ_ACTIVE_ENTRIES),
+		.escr_msr	= { MSR_P4_BSU_ESCR1 , MSR_P4_BSU_ESCR1 },
+		.cntr		= { {2, -1, -1}, {3, -1, -1} },
+	},
+	[P4_EVENT_SSE_INPUT_ASSIST] = {
+		.opcode		= P4_OPCODE(P4_EVENT_SSE_INPUT_ASSIST),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_PACKED_SP_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_PACKED_SP_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_PACKED_DP_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_PACKED_DP_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_SCALAR_SP_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_SCALAR_SP_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_SCALAR_DP_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_SCALAR_DP_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_64BIT_MMX_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_64BIT_MMX_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_128BIT_MMX_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_128BIT_MMX_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_X87_FP_UOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_X87_FP_UOP),
+		.escr_msr	= { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_TC_MISC] = {
+		.opcode		= P4_OPCODE(P4_EVENT_TC_MISC),
+		.escr_msr	= { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_GLOBAL_POWER_EVENTS] = {
+		.opcode		= P4_OPCODE(P4_EVENT_GLOBAL_POWER_EVENTS),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_TC_MS_XFER] = {
+		.opcode		= P4_OPCODE(P4_EVENT_TC_MS_XFER),
+		.escr_msr	= { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_UOP_QUEUE_WRITES] = {
+		.opcode		= P4_OPCODE(P4_EVENT_UOP_QUEUE_WRITES),
+		.escr_msr	= { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE),
+		.escr_msr	= { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR0 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_RETIRED_BRANCH_TYPE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_RETIRED_BRANCH_TYPE),
+		.escr_msr	= { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR1 },
+		.cntr		= { {4, 5, -1}, {6, 7, -1} },
+	},
+	[P4_EVENT_RESOURCE_STALL] = {
+		.opcode		= P4_OPCODE(P4_EVENT_RESOURCE_STALL),
+		.escr_msr	= { MSR_P4_ALF_ESCR0, MSR_P4_ALF_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_WC_BUFFER] = {
+		.opcode		= P4_OPCODE(P4_EVENT_WC_BUFFER),
+		.escr_msr	= { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 },
+		.cntr		= { {8, 9, -1}, {10, 11, -1} },
+	},
+	[P4_EVENT_B2B_CYCLES] = {
+		.opcode		= P4_OPCODE(P4_EVENT_B2B_CYCLES),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_BNR] = {
+		.opcode		= P4_OPCODE(P4_EVENT_BNR),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_SNOOP] = {
+		.opcode		= P4_OPCODE(P4_EVENT_SNOOP),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_RESPONSE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_RESPONSE),
+		.escr_msr	= { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 },
+		.cntr		= { {0, -1, -1}, {2, -1, -1} },
+	},
+	[P4_EVENT_FRONT_END_EVENT] = {
+		.opcode		= P4_OPCODE(P4_EVENT_FRONT_END_EVENT),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_EXECUTION_EVENT] = {
+		.opcode		= P4_OPCODE(P4_EVENT_EXECUTION_EVENT),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_REPLAY_EVENT] = {
+		.opcode		= P4_OPCODE(P4_EVENT_REPLAY_EVENT),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_INSTR_RETIRED] = {
+		.opcode		= P4_OPCODE(P4_EVENT_INSTR_RETIRED),
+		.escr_msr	= { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_UOPS_RETIRED] = {
+		.opcode		= P4_OPCODE(P4_EVENT_UOPS_RETIRED),
+		.escr_msr	= { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_UOP_TYPE] = {
+		.opcode		= P4_OPCODE(P4_EVENT_UOP_TYPE),
+		.escr_msr	= { MSR_P4_RAT_ESCR0, MSR_P4_RAT_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_BRANCH_RETIRED] = {
+		.opcode		= P4_OPCODE(P4_EVENT_BRANCH_RETIRED),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_MISPRED_BRANCH_RETIRED] = {
+		.opcode		= P4_OPCODE(P4_EVENT_MISPRED_BRANCH_RETIRED),
+		.escr_msr	= { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_X87_ASSIST] = {
+		.opcode		= P4_OPCODE(P4_EVENT_X87_ASSIST),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_MACHINE_CLEAR] = {
+		.opcode		= P4_OPCODE(P4_EVENT_MACHINE_CLEAR),
+		.escr_msr	= { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+	[P4_EVENT_INSTR_COMPLETED] = {
+		.opcode		= P4_OPCODE(P4_EVENT_INSTR_COMPLETED),
+		.escr_msr	= { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 },
+		.cntr		= { {12, 13, 16}, {14, 15, 17} },
+	},
+};
+
+#define P4_GEN_CACHE_EVENT(event, bit, cache_event)			  \
+	p4_config_pack_escr(P4_ESCR_EVENT(event)			| \
+			    P4_ESCR_EMASK_BIT(event, bit))		| \
+	p4_config_pack_cccr(cache_event					| \
+			    P4_CCCR_ESEL(P4_OPCODE_ESEL(P4_OPCODE(event))))
+
+static __initconst const u64 p4_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) ] = 0x0,
+		[ C(RESULT_MISS)   ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS,
+						P4_CACHE__1stl_cache_load_miss_retired),
+	},
+ },
+ [ C(LL  ) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS,
+						P4_CACHE__2ndl_cache_load_miss_retired),
+	},
+},
+ [ C(DTLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS,
+						P4_CACHE__dtlb_load_miss_retired),
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x0,
+		[ C(RESULT_MISS)   ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS,
+						P4_CACHE__dtlb_store_miss_retired),
+	},
+ },
+ [ C(ITLB) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_ITLB_REFERENCE, HIT,
+						P4_CACHE__itlb_reference_hit),
+		[ C(RESULT_MISS)   ] = P4_GEN_CACHE_EVENT(P4_EVENT_ITLB_REFERENCE, MISS,
+						P4_CACHE__itlb_reference_miss),
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+};
+
+static u64 p4_general_events[PERF_COUNT_HW_MAX] = {
+  /* non-halted CPU clocks */
+  [PERF_COUNT_HW_CPU_CYCLES] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)),
+
+  /*
+   * retired instructions
+   * in a sake of simplicity we don't use the FSB tagging
+   */
+  [PERF_COUNT_HW_INSTRUCTIONS] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_INSTR_RETIRED)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, NBOGUSNTAG)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, BOGUSNTAG)),
+
+  /* cache hits */
+  [PERF_COUNT_HW_CACHE_REFERENCES] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_BSQ_CACHE_REFERENCE)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITS)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITE)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITM)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITS)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITE)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITM)),
+
+  /* cache misses */
+  [PERF_COUNT_HW_CACHE_MISSES] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_BSQ_CACHE_REFERENCE)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_MISS)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_MISS)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, WR_2ndL_MISS)),
+
+  /* branch instructions retired */
+  [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_RETIRED_BRANCH_TYPE)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CONDITIONAL)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CALL)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, RETURN)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, INDIRECT)),
+
+  /* mispredicted branches retired */
+  [PERF_COUNT_HW_BRANCH_MISSES]	=
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_MISPRED_BRANCH_RETIRED)	|
+		P4_ESCR_EMASK_BIT(P4_EVENT_MISPRED_BRANCH_RETIRED, NBOGUS)),
+
+  /* bus ready clocks (cpu is driving #DRDY_DRV\#DRDY_OWN):  */
+  [PERF_COUNT_HW_BUS_CYCLES] =
+	p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_FSB_DATA_ACTIVITY)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_DRV)		|
+		P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OWN))	|
+	p4_config_pack_cccr(P4_CCCR_EDGE | P4_CCCR_COMPARE),
+};
+
+static struct p4_event_bind *p4_config_get_bind(u64 config)
+{
+	unsigned int evnt = p4_config_unpack_event(config);
+	struct p4_event_bind *bind = NULL;
+
+	if (evnt < ARRAY_SIZE(p4_event_bind_map))
+		bind = &p4_event_bind_map[evnt];
+
+	return bind;
+}
+
+static u64 p4_pmu_event_map(int hw_event)
+{
+	struct p4_event_bind *bind;
+	unsigned int esel;
+	u64 config;
+
+	config = p4_general_events[hw_event];
+	bind = p4_config_get_bind(config);
+	esel = P4_OPCODE_ESEL(bind->opcode);
+	config |= p4_config_pack_cccr(P4_CCCR_ESEL(esel));
+
+	return config;
+}
+
+static int p4_hw_config(struct perf_event *event)
+{
+	int cpu = get_cpu();
+	int rc = 0;
+	unsigned int evnt;
+	u32 escr, cccr;
+
+	/*
+	 * the reason we use cpu that early is that: if we get scheduled
+	 * first time on the same cpu -- we will not need swap thread
+	 * specific flags in config (and will save some cpu cycles)
+	 */
+
+	cccr = p4_default_cccr_conf(cpu);
+	escr = p4_default_escr_conf(cpu, event->attr.exclude_kernel,
+					 event->attr.exclude_user);
+	event->hw.config = p4_config_pack_escr(escr) |
+			   p4_config_pack_cccr(cccr);
+
+	if (p4_ht_active() && p4_ht_thread(cpu))
+		event->hw.config = p4_set_ht_bit(event->hw.config);
+
+	if (event->attr.type == PERF_TYPE_RAW) {
+
+		/* user data may have out-of-bound event index */
+		evnt = p4_config_unpack_event(event->attr.config);
+		if (evnt >= ARRAY_SIZE(p4_event_bind_map)) {
+			rc = -EINVAL;
+			goto out;
+		}
+
+		/*
+		 * We don't control raw events so it's up to the caller
+		 * to pass sane values (and we don't count the thread number
+		 * on HT machine but allow HT-compatible specifics to be
+		 * passed on)
+		 *
+		 * XXX: HT wide things should check perf_paranoid_cpu() &&
+		 *      CAP_SYS_ADMIN
+		 */
+		event->hw.config |= event->attr.config &
+			(p4_config_pack_escr(P4_ESCR_MASK_HT) |
+			 p4_config_pack_cccr(P4_CCCR_MASK_HT));
+	}
+
+	rc = x86_setup_perfctr(event);
+out:
+	put_cpu();
+	return rc;
+}
+
+static inline void p4_pmu_clear_cccr_ovf(struct hw_perf_event *hwc)
+{
+	unsigned long dummy;
+
+	rdmsrl(hwc->config_base + hwc->idx, dummy);
+	if (dummy & P4_CCCR_OVF) {
+		(void)checking_wrmsrl(hwc->config_base + hwc->idx,
+			((u64)dummy) & ~P4_CCCR_OVF);
+	}
+}
+
+static inline void p4_pmu_disable_event(struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+
+	/*
+	 * If event gets disabled while counter is in overflowed
+	 * state we need to clear P4_CCCR_OVF, otherwise interrupt get
+	 * asserted again and again
+	 */
+	(void)checking_wrmsrl(hwc->config_base + hwc->idx,
+		(u64)(p4_config_unpack_cccr(hwc->config)) &
+			~P4_CCCR_ENABLE & ~P4_CCCR_OVF & ~P4_CCCR_RESERVED);
+}
+
+static void p4_pmu_disable_all(void)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	int idx;
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		struct perf_event *event = cpuc->events[idx];
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+		p4_pmu_disable_event(event);
+	}
+}
+
+static void p4_pmu_enable_event(struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	int thread = p4_ht_config_thread(hwc->config);
+	u64 escr_conf = p4_config_unpack_escr(p4_clear_ht_bit(hwc->config));
+	unsigned int idx = p4_config_unpack_event(hwc->config);
+	unsigned int idx_cache = p4_config_unpack_cache_event(hwc->config);
+	struct p4_event_bind *bind;
+	struct p4_cache_event_bind *bind_cache;
+	u64 escr_addr, cccr;
+
+	bind = &p4_event_bind_map[idx];
+	escr_addr = (u64)bind->escr_msr[thread];
+
+	/*
+	 * - we dont support cascaded counters yet
+	 * - and counter 1 is broken (erratum)
+	 */
+	WARN_ON_ONCE(p4_is_event_cascaded(hwc->config));
+	WARN_ON_ONCE(hwc->idx == 1);
+
+	/* we need a real Event value */
+	escr_conf &= ~P4_ESCR_EVENT_MASK;
+	escr_conf |= P4_ESCR_EVENT(P4_OPCODE_EVNT(bind->opcode));
+
+	cccr = p4_config_unpack_cccr(hwc->config);
+
+	/*
+	 * it could be Cache event so that we need to
+	 * set metrics into additional MSRs
+	 */
+	BUILD_BUG_ON(P4_CACHE__MAX > P4_CCCR_CACHE_OPS_MASK);
+	if (idx_cache > P4_CACHE__NONE &&
+		idx_cache < ARRAY_SIZE(p4_cache_event_bind_map)) {
+		bind_cache = &p4_cache_event_bind_map[idx_cache];
+		(void)checking_wrmsrl(MSR_IA32_PEBS_ENABLE, (u64)bind_cache->metric_pebs);
+		(void)checking_wrmsrl(MSR_P4_PEBS_MATRIX_VERT, (u64)bind_cache->metric_vert);
+	}
+
+	(void)checking_wrmsrl(escr_addr, escr_conf);
+	(void)checking_wrmsrl(hwc->config_base + hwc->idx,
+				(cccr & ~P4_CCCR_RESERVED) | P4_CCCR_ENABLE);
+}
+
+static void p4_pmu_enable_all(int added)
+{
+	struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+	int idx;
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+		struct perf_event *event = cpuc->events[idx];
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+		p4_pmu_enable_event(event);
+	}
+}
+
+static int p4_pmu_handle_irq(struct pt_regs *regs)
+{
+	struct perf_sample_data data;
+	struct cpu_hw_events *cpuc;
+	struct perf_event *event;
+	struct hw_perf_event *hwc;
+	int idx, handled = 0;
+	u64 val;
+
+	data.addr = 0;
+	data.raw = NULL;
+
+	cpuc = &__get_cpu_var(cpu_hw_events);
+
+	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
+
+		if (!test_bit(idx, cpuc->active_mask))
+			continue;
+
+		event = cpuc->events[idx];
+		hwc = &event->hw;
+
+		WARN_ON_ONCE(hwc->idx != idx);
+
+		/*
+		 * FIXME: Redundant call, actually not needed
+		 * but just to check if we're screwed
+		 */
+		p4_pmu_clear_cccr_ovf(hwc);
+
+		val = x86_perf_event_update(event);
+		if (val & (1ULL << (x86_pmu.cntval_bits - 1)))
+			continue;
+
+		/*
+		 * event overflow
+		 */
+		handled		= 1;
+		data.period	= event->hw.last_period;
+
+		if (!x86_perf_event_set_period(event))
+			continue;
+		if (perf_event_overflow(event, 1, &data, regs))
+			p4_pmu_disable_event(event);
+	}
+
+	if (handled) {
+		/* p4 quirk: unmask it again */
+		apic_write(APIC_LVTPC, apic_read(APIC_LVTPC) & ~APIC_LVT_MASKED);
+		inc_irq_stat(apic_perf_irqs);
+	}
+
+	return handled;
+}
+
+/*
+ * swap thread specific fields according to a thread
+ * we are going to run on
+ */
+static void p4_pmu_swap_config_ts(struct hw_perf_event *hwc, int cpu)
+{
+	u32 escr, cccr;
+
+	/*
+	 * we either lucky and continue on same cpu or no HT support
+	 */
+	if (!p4_should_swap_ts(hwc->config, cpu))
+		return;
+
+	/*
+	 * the event is migrated from an another logical
+	 * cpu, so we need to swap thread specific flags
+	 */
+
+	escr = p4_config_unpack_escr(hwc->config);
+	cccr = p4_config_unpack_cccr(hwc->config);
+
+	if (p4_ht_thread(cpu)) {
+		cccr &= ~P4_CCCR_OVF_PMI_T0;
+		cccr |= P4_CCCR_OVF_PMI_T1;
+		if (escr & P4_ESCR_T0_OS) {
+			escr &= ~P4_ESCR_T0_OS;
+			escr |= P4_ESCR_T1_OS;
+		}
+		if (escr & P4_ESCR_T0_USR) {
+			escr &= ~P4_ESCR_T0_USR;
+			escr |= P4_ESCR_T1_USR;
+		}
+		hwc->config  = p4_config_pack_escr(escr);
+		hwc->config |= p4_config_pack_cccr(cccr);
+		hwc->config |= P4_CONFIG_HT;
+	} else {
+		cccr &= ~P4_CCCR_OVF_PMI_T1;
+		cccr |= P4_CCCR_OVF_PMI_T0;
+		if (escr & P4_ESCR_T1_OS) {
+			escr &= ~P4_ESCR_T1_OS;
+			escr |= P4_ESCR_T0_OS;
+		}
+		if (escr & P4_ESCR_T1_USR) {
+			escr &= ~P4_ESCR_T1_USR;
+			escr |= P4_ESCR_T0_USR;
+		}
+		hwc->config  = p4_config_pack_escr(escr);
+		hwc->config |= p4_config_pack_cccr(cccr);
+		hwc->config &= ~P4_CONFIG_HT;
+	}
+}
+
+/*
+ * ESCR address hashing is tricky, ESCRs are not sequential
+ * in memory but all starts from MSR_P4_BSU_ESCR0 (0x03e0) and
+ * the metric between any ESCRs is laid in range [0xa0,0xe1]
+ *
+ * so we make ~70% filled hashtable
+ */
+
+#define P4_ESCR_MSR_BASE		0x000003a0
+#define P4_ESCR_MSR_MAX			0x000003e1
+#define P4_ESCR_MSR_TABLE_SIZE		(P4_ESCR_MSR_MAX - P4_ESCR_MSR_BASE + 1)
+#define P4_ESCR_MSR_IDX(msr)		(msr - P4_ESCR_MSR_BASE)
+#define P4_ESCR_MSR_TABLE_ENTRY(msr)	[P4_ESCR_MSR_IDX(msr)] = msr
+
+static const unsigned int p4_escr_table[P4_ESCR_MSR_TABLE_SIZE] = {
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_ALF_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_ALF_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_BPU_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_BPU_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_BSU_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_BSU_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR2),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR3),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR4),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_CRU_ESCR5),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_DAC_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_DAC_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FIRM_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FIRM_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FLAME_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FLAME_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FSB_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_FSB_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IQ_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IQ_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IS_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IS_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_ITLB_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_ITLB_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IX_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_IX_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_MOB_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_MOB_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_MS_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_MS_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_PMH_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_PMH_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_RAT_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_RAT_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_SAAT_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_SAAT_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_SSU_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_SSU_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_TBPU_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_TBPU_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_TC_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_TC_ESCR1),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_U2L_ESCR0),
+	P4_ESCR_MSR_TABLE_ENTRY(MSR_P4_U2L_ESCR1),
+};
+
+static int p4_get_escr_idx(unsigned int addr)
+{
+	unsigned int idx = P4_ESCR_MSR_IDX(addr);
+
+	if (unlikely(idx >= P4_ESCR_MSR_TABLE_SIZE ||
+			!p4_escr_table[idx])) {
+		WARN_ONCE(1, "P4 PMU: Wrong address passed: %x\n", addr);
+		return -1;
+	}
+
+	return idx;
+}
+
+static int p4_next_cntr(int thread, unsigned long *used_mask,
+			struct p4_event_bind *bind)
+{
+	int i, j;
+
+	for (i = 0; i < P4_CNTR_LIMIT; i++) {
+		j = bind->cntr[thread][i];
+		if (j != -1 && !test_bit(j, used_mask))
+			return j;
+	}
+
+	return -1;
+}
+
+static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
+{
+	unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
+	unsigned long escr_mask[BITS_TO_LONGS(P4_ESCR_MSR_TABLE_SIZE)];
+	int cpu = raw_smp_processor_id();
+	struct hw_perf_event *hwc;
+	struct p4_event_bind *bind;
+	unsigned int i, thread, num;
+	int cntr_idx, escr_idx;
+
+	bitmap_zero(used_mask, X86_PMC_IDX_MAX);
+	bitmap_zero(escr_mask, P4_ESCR_MSR_TABLE_SIZE);
+
+	for (i = 0, num = n; i < n; i++, num--) {
+
+		hwc = &cpuc->event_list[i]->hw;
+		thread = p4_ht_thread(cpu);
+		bind = p4_config_get_bind(hwc->config);
+		escr_idx = p4_get_escr_idx(bind->escr_msr[thread]);
+		if (unlikely(escr_idx == -1))
+			goto done;
+
+		if (hwc->idx != -1 && !p4_should_swap_ts(hwc->config, cpu)) {
+			cntr_idx = hwc->idx;
+			if (assign)
+				assign[i] = hwc->idx;
+			goto reserve;
+		}
+
+		cntr_idx = p4_next_cntr(thread, used_mask, bind);
+		if (cntr_idx == -1 || test_bit(escr_idx, escr_mask))
+			goto done;
+
+		p4_pmu_swap_config_ts(hwc, cpu);
+		if (assign)
+			assign[i] = cntr_idx;
+reserve:
+		set_bit(cntr_idx, used_mask);
+		set_bit(escr_idx, escr_mask);
+	}
+
+done:
+	return num ? -ENOSPC : 0;
+}
+
+static __initconst const struct x86_pmu p4_pmu = {
+	.name			= "Netburst P4/Xeon",
+	.handle_irq		= p4_pmu_handle_irq,
+	.disable_all		= p4_pmu_disable_all,
+	.enable_all		= p4_pmu_enable_all,
+	.enable			= p4_pmu_enable_event,
+	.disable		= p4_pmu_disable_event,
+	.eventsel		= MSR_P4_BPU_CCCR0,
+	.perfctr		= MSR_P4_BPU_PERFCTR0,
+	.event_map		= p4_pmu_event_map,
+	.max_events		= ARRAY_SIZE(p4_general_events),
+	.get_event_constraints	= x86_get_event_constraints,
+	/*
+	 * IF HT disabled we may need to use all
+	 * ARCH_P4_MAX_CCCR counters simulaneously
+	 * though leave it restricted at moment assuming
+	 * HT is on
+	 */
+	.num_counters		= ARCH_P4_MAX_CCCR,
+	.apic			= 1,
+	.cntval_bits		= 40,
+	.cntval_mask		= (1ULL << 40) - 1,
+	.max_period		= (1ULL << 39) - 1,
+	.hw_config		= p4_hw_config,
+	.schedule_events	= p4_pmu_schedule_events,
+};
+
+static __init int p4_pmu_init(void)
+{
+	unsigned int low, high;
+
+	/* If we get stripped -- indexig fails */
+	BUILD_BUG_ON(ARCH_P4_MAX_CCCR > X86_PMC_MAX_GENERIC);
+
+	rdmsr(MSR_IA32_MISC_ENABLE, low, high);
+	if (!(low & (1 << 7))) {
+		pr_cont("unsupported Netburst CPU model %d ",
+			boot_cpu_data.x86_model);
+		return -ENODEV;
+	}
+
+	memcpy(hw_cache_event_ids, p4_hw_cache_event_ids,
+		sizeof(hw_cache_event_ids));
+
+	pr_cont("Netburst events, ");
+
+	x86_pmu = p4_pmu;
+
+	return 0;
+}
+
+#endif /* CONFIG_CPU_SUP_INTEL */
diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c
index a330485d14d..34ba07be2cd 100644
--- a/arch/x86/kernel/cpu/perf_event_p6.c
+++ b/arch/x86/kernel/cpu/perf_event_p6.c
@@ -27,24 +27,6 @@ static u64 p6_pmu_event_map(int hw_event)
  */
 #define P6_NOP_EVENT			0x0000002EULL
 
-static u64 p6_pmu_raw_event(u64 hw_event)
-{
-#define P6_EVNTSEL_EVENT_MASK		0x000000FFULL
-#define P6_EVNTSEL_UNIT_MASK		0x0000FF00ULL
-#define P6_EVNTSEL_EDGE_MASK		0x00040000ULL
-#define P6_EVNTSEL_INV_MASK		0x00800000ULL
-#define P6_EVNTSEL_REG_MASK		0xFF000000ULL
-
-#define P6_EVNTSEL_MASK			\
-	(P6_EVNTSEL_EVENT_MASK |	\
-	 P6_EVNTSEL_UNIT_MASK  |	\
-	 P6_EVNTSEL_EDGE_MASK  |	\
-	 P6_EVNTSEL_INV_MASK   |	\
-	 P6_EVNTSEL_REG_MASK)
-
-	return hw_event & P6_EVNTSEL_MASK;
-}
-
 static struct event_constraint p6_event_constraints[] =
 {
 	INTEL_EVENT_CONSTRAINT(0xc1, 0x1),	/* FLOPS */
@@ -66,7 +48,7 @@ static void p6_pmu_disable_all(void)
 	wrmsrl(MSR_P6_EVNTSEL0, val);
 }
 
-static void p6_pmu_enable_all(void)
+static void p6_pmu_enable_all(int added)
 {
 	unsigned long val;
 
@@ -102,22 +84,23 @@ static void p6_pmu_enable_event(struct perf_event *event)
 	(void)checking_wrmsrl(hwc->config_base + hwc->idx, val);
 }
 
-static __initconst struct x86_pmu p6_pmu = {
+static __initconst const struct x86_pmu p6_pmu = {
 	.name			= "p6",
 	.handle_irq		= x86_pmu_handle_irq,
 	.disable_all		= p6_pmu_disable_all,
 	.enable_all		= p6_pmu_enable_all,
 	.enable			= p6_pmu_enable_event,
 	.disable		= p6_pmu_disable_event,
+	.hw_config		= x86_pmu_hw_config,
+	.schedule_events	= x86_schedule_events,
 	.eventsel		= MSR_P6_EVNTSEL0,
 	.perfctr		= MSR_P6_PERFCTR0,
 	.event_map		= p6_pmu_event_map,
-	.raw_event		= p6_pmu_raw_event,
 	.max_events		= ARRAY_SIZE(p6_perfmon_event_map),
 	.apic			= 1,
 	.max_period		= (1ULL << 31) - 1,
 	.version		= 0,
-	.num_events		= 2,
+	.num_counters		= 2,
 	/*
 	 * Events have 40 bits implemented. However they are designed such
 	 * that bits [32-39] are sign extensions of bit 31. As such the
@@ -125,8 +108,8 @@ static __initconst struct x86_pmu p6_pmu = {
 	 *
 	 * See IA-32 Intel Architecture Software developer manual Vol 3B
 	 */
-	.event_bits		= 32,
-	.event_mask		= (1ULL << 32) - 1,
+	.cntval_bits		= 32,
+	.cntval_mask		= (1ULL << 32) - 1,
 	.get_event_constraints	= x86_get_event_constraints,
 	.event_constraints	= p6_event_constraints,
 };
diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c
deleted file mode 100644
index 1c47390dd0e..00000000000
--- a/arch/x86/kernel/ds.c
+++ /dev/null
@@ -1,1437 +0,0 @@
-/*
- * Debug Store support
- *
- * This provides a low-level interface to the hardware's Debug Store
- * feature that is used for branch trace store (BTS) and
- * precise-event based sampling (PEBS).
- *
- * It manages:
- * - DS and BTS hardware configuration
- * - buffer overflow handling (to be done)
- * - buffer access
- *
- * It does not do:
- * - security checking (is the caller allowed to trace the task)
- * - buffer allocation (memory accounting)
- *
- *
- * Copyright (C) 2007-2009 Intel Corporation.
- * Markus Metzger <markus.t.metzger@intel.com>, 2007-2009
- */
-
-#include <linux/kernel.h>
-#include <linux/string.h>
-#include <linux/errno.h>
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/mm.h>
-#include <linux/trace_clock.h>
-
-#include <asm/ds.h>
-
-#include "ds_selftest.h"
-
-/*
- * The configuration for a particular DS hardware implementation:
- */
-struct ds_configuration {
-	/* The name of the configuration: */
-	const char		*name;
-
-	/* The size of pointer-typed fields in DS, BTS, and PEBS: */
-	unsigned char		sizeof_ptr_field;
-
-	/* The size of a BTS/PEBS record in bytes: */
-	unsigned char		sizeof_rec[2];
-
-	/* The number of pebs counter reset values in the DS structure. */
-	unsigned char		nr_counter_reset;
-
-	/* Control bit-masks indexed by enum ds_feature: */
-	unsigned long		ctl[dsf_ctl_max];
-};
-static struct ds_configuration ds_cfg __read_mostly;
-
-
-/* Maximal size of a DS configuration: */
-#define MAX_SIZEOF_DS		0x80
-
-/* Maximal size of a BTS record: */
-#define MAX_SIZEOF_BTS		(3 * 8)
-
-/* BTS and PEBS buffer alignment: */
-#define DS_ALIGNMENT		(1 << 3)
-
-/* Number of buffer pointers in DS: */
-#define NUM_DS_PTR_FIELDS	8
-
-/* Size of a pebs reset value in DS: */
-#define PEBS_RESET_FIELD_SIZE	8
-
-/* Mask of control bits in the DS MSR register: */
-#define BTS_CONTROL				  \
-	( ds_cfg.ctl[dsf_bts]			| \
-	  ds_cfg.ctl[dsf_bts_kernel]		| \
-	  ds_cfg.ctl[dsf_bts_user]		| \
-	  ds_cfg.ctl[dsf_bts_overflow] )
-
-/*
- * A BTS or PEBS tracer.
- *
- * This holds the configuration of the tracer and serves as a handle
- * to identify tracers.
- */
-struct ds_tracer {
-	/* The DS context (partially) owned by this tracer. */
-	struct ds_context	*context;
-	/* The buffer provided on ds_request() and its size in bytes. */
-	void			*buffer;
-	size_t			size;
-};
-
-struct bts_tracer {
-	/* The common DS part: */
-	struct ds_tracer	ds;
-
-	/* The trace including the DS configuration: */
-	struct bts_trace	trace;
-
-	/* Buffer overflow notification function: */
-	bts_ovfl_callback_t	ovfl;
-
-	/* Active flags affecting trace collection. */
-	unsigned int		flags;
-};
-
-struct pebs_tracer {
-	/* The common DS part: */
-	struct ds_tracer	ds;
-
-	/* The trace including the DS configuration: */
-	struct pebs_trace	trace;
-
-	/* Buffer overflow notification function: */
-	pebs_ovfl_callback_t	ovfl;
-};
-
-/*
- * Debug Store (DS) save area configuration (see Intel64 and IA32
- * Architectures Software Developer's Manual, section 18.5)
- *
- * The DS configuration consists of the following fields; different
- * architetures vary in the size of those fields.
- *
- * - double-word aligned base linear address of the BTS buffer
- * - write pointer into the BTS buffer
- * - end linear address of the BTS buffer (one byte beyond the end of
- *   the buffer)
- * - interrupt pointer into BTS buffer
- *   (interrupt occurs when write pointer passes interrupt pointer)
- * - double-word aligned base linear address of the PEBS buffer
- * - write pointer into the PEBS buffer
- * - end linear address of the PEBS buffer (one byte beyond the end of
- *   the buffer)
- * - interrupt pointer into PEBS buffer
- *   (interrupt occurs when write pointer passes interrupt pointer)
- * - value to which counter is reset following counter overflow
- *
- * Later architectures use 64bit pointers throughout, whereas earlier
- * architectures use 32bit pointers in 32bit mode.
- *
- *
- * We compute the base address for the first 8 fields based on:
- * - the field size stored in the DS configuration
- * - the relative field position
- * - an offset giving the start of the respective region
- *
- * This offset is further used to index various arrays holding
- * information for BTS and PEBS at the respective index.
- *
- * On later 32bit processors, we only access the lower 32bit of the
- * 64bit pointer fields. The upper halves will be zeroed out.
- */
-
-enum ds_field {
-	ds_buffer_base = 0,
-	ds_index,
-	ds_absolute_maximum,
-	ds_interrupt_threshold,
-};
-
-enum ds_qualifier {
-	ds_bts = 0,
-	ds_pebs
-};
-
-static inline unsigned long
-ds_get(const unsigned char *base, enum ds_qualifier qual, enum ds_field field)
-{
-	base += (ds_cfg.sizeof_ptr_field * (field + (4 * qual)));
-	return *(unsigned long *)base;
-}
-
-static inline void
-ds_set(unsigned char *base, enum ds_qualifier qual, enum ds_field field,
-       unsigned long value)
-{
-	base += (ds_cfg.sizeof_ptr_field * (field + (4 * qual)));
-	(*(unsigned long *)base) = value;
-}
-
-
-/*
- * Locking is done only for allocating BTS or PEBS resources.
- */
-static DEFINE_SPINLOCK(ds_lock);
-
-/*
- * We either support (system-wide) per-cpu or per-thread allocation.
- * We distinguish the two based on the task_struct pointer, where a
- * NULL pointer indicates per-cpu allocation for the current cpu.
- *
- * Allocations are use-counted. As soon as resources are allocated,
- * further allocations must be of the same type (per-cpu or
- * per-thread). We model this by counting allocations (i.e. the number
- * of tracers of a certain type) for one type negatively:
- *   =0  no tracers
- *   >0  number of per-thread tracers
- *   <0  number of per-cpu tracers
- *
- * Tracers essentially gives the number of ds contexts for a certain
- * type of allocation.
- */
-static atomic_t tracers = ATOMIC_INIT(0);
-
-static inline int get_tracer(struct task_struct *task)
-{
-	int error;
-
-	spin_lock_irq(&ds_lock);
-
-	if (task) {
-		error = -EPERM;
-		if (atomic_read(&tracers) < 0)
-			goto out;
-		atomic_inc(&tracers);
-	} else {
-		error = -EPERM;
-		if (atomic_read(&tracers) > 0)
-			goto out;
-		atomic_dec(&tracers);
-	}
-
-	error = 0;
-out:
-	spin_unlock_irq(&ds_lock);
-	return error;
-}
-
-static inline void put_tracer(struct task_struct *task)
-{
-	if (task)
-		atomic_dec(&tracers);
-	else
-		atomic_inc(&tracers);
-}
-
-/*
- * The DS context is either attached to a thread or to a cpu:
- * - in the former case, the thread_struct contains a pointer to the
- *   attached context.
- * - in the latter case, we use a static array of per-cpu context
- *   pointers.
- *
- * Contexts are use-counted. They are allocated on first access and
- * deallocated when the last user puts the context.
- */
-struct ds_context {
-	/* The DS configuration; goes into MSR_IA32_DS_AREA: */
-	unsigned char		ds[MAX_SIZEOF_DS];
-
-	/* The owner of the BTS and PEBS configuration, respectively: */
-	struct bts_tracer	*bts_master;
-	struct pebs_tracer	*pebs_master;
-
-	/* Use count: */
-	unsigned long		count;
-
-	/* Pointer to the context pointer field: */
-	struct ds_context	**this;
-
-	/* The traced task; NULL for cpu tracing: */
-	struct task_struct	*task;
-
-	/* The traced cpu; only valid if task is NULL: */
-	int			cpu;
-};
-
-static DEFINE_PER_CPU(struct ds_context *, cpu_ds_context);
-
-
-static struct ds_context *ds_get_context(struct task_struct *task, int cpu)
-{
-	struct ds_context **p_context =
-		(task ? &task->thread.ds_ctx : &per_cpu(cpu_ds_context, cpu));
-	struct ds_context *context = NULL;
-	struct ds_context *new_context = NULL;
-
-	/* Chances are small that we already have a context. */
-	new_context = kzalloc(sizeof(*new_context), GFP_KERNEL);
-	if (!new_context)
-		return NULL;
-
-	spin_lock_irq(&ds_lock);
-
-	context = *p_context;
-	if (likely(!context)) {
-		context = new_context;
-
-		context->this = p_context;
-		context->task = task;
-		context->cpu = cpu;
-		context->count = 0;
-
-		*p_context = context;
-	}
-
-	context->count++;
-
-	spin_unlock_irq(&ds_lock);
-
-	if (context != new_context)
-		kfree(new_context);
-
-	return context;
-}
-
-static void ds_put_context(struct ds_context *context)
-{
-	struct task_struct *task;
-	unsigned long irq;
-
-	if (!context)
-		return;
-
-	spin_lock_irqsave(&ds_lock, irq);
-
-	if (--context->count) {
-		spin_unlock_irqrestore(&ds_lock, irq);
-		return;
-	}
-
-	*(context->this) = NULL;
-
-	task = context->task;
-
-	if (task)
-		clear_tsk_thread_flag(task, TIF_DS_AREA_MSR);
-
-	/*
-	 * We leave the (now dangling) pointer to the DS configuration in
-	 * the DS_AREA msr. This is as good or as bad as replacing it with
-	 * NULL - the hardware would crash if we enabled tracing.
-	 *
-	 * This saves us some problems with having to write an msr on a
-	 * different cpu while preventing others from doing the same for the
-	 * next context for that same cpu.
-	 */
-
-	spin_unlock_irqrestore(&ds_lock, irq);
-
-	/* The context might still be in use for context switching. */
-	if (task && (task != current))
-		wait_task_context_switch(task);
-
-	kfree(context);
-}
-
-static void ds_install_ds_area(struct ds_context *context)
-{
-	unsigned long ds;
-
-	ds = (unsigned long)context->ds;
-
-	/*
-	 * There is a race between the bts master and the pebs master.
-	 *
-	 * The thread/cpu access is synchronized via get/put_cpu() for
-	 * task tracing and via wrmsr_on_cpu for cpu tracing.
-	 *
-	 * If bts and pebs are collected for the same task or same cpu,
-	 * the same confiuration is written twice.
-	 */
-	if (context->task) {
-		get_cpu();
-		if (context->task == current)
-			wrmsrl(MSR_IA32_DS_AREA, ds);
-		set_tsk_thread_flag(context->task, TIF_DS_AREA_MSR);
-		put_cpu();
-	} else
-		wrmsr_on_cpu(context->cpu, MSR_IA32_DS_AREA,
-			     (u32)((u64)ds), (u32)((u64)ds >> 32));
-}
-
-/*
- * Call the tracer's callback on a buffer overflow.
- *
- * context: the ds context
- * qual: the buffer type
- */
-static void ds_overflow(struct ds_context *context, enum ds_qualifier qual)
-{
-	switch (qual) {
-	case ds_bts:
-		if (context->bts_master &&
-		    context->bts_master->ovfl)
-			context->bts_master->ovfl(context->bts_master);
-		break;
-	case ds_pebs:
-		if (context->pebs_master &&
-		    context->pebs_master->ovfl)
-			context->pebs_master->ovfl(context->pebs_master);
-		break;
-	}
-}
-
-
-/*
- * Write raw data into the BTS or PEBS buffer.
- *
- * The remainder of any partially written record is zeroed out.
- *
- * context: the DS context
- * qual:    the buffer type
- * record:  the data to write
- * size:    the size of the data
- */
-static int ds_write(struct ds_context *context, enum ds_qualifier qual,
-		    const void *record, size_t size)
-{
-	int bytes_written = 0;
-
-	if (!record)
-		return -EINVAL;
-
-	while (size) {
-		unsigned long base, index, end, write_end, int_th;
-		unsigned long write_size, adj_write_size;
-
-		/*
-		 * Write as much as possible without producing an
-		 * overflow interrupt.
-		 *
-		 * Interrupt_threshold must either be
-		 * - bigger than absolute_maximum or
-		 * - point to a record between buffer_base and absolute_maximum
-		 *
-		 * Index points to a valid record.
-		 */
-		base   = ds_get(context->ds, qual, ds_buffer_base);
-		index  = ds_get(context->ds, qual, ds_index);
-		end    = ds_get(context->ds, qual, ds_absolute_maximum);
-		int_th = ds_get(context->ds, qual, ds_interrupt_threshold);
-
-		write_end = min(end, int_th);
-
-		/*
-		 * If we are already beyond the interrupt threshold,
-		 * we fill the entire buffer.
-		 */
-		if (write_end <= index)
-			write_end = end;
-
-		if (write_end <= index)
-			break;
-
-		write_size = min((unsigned long) size, write_end - index);
-		memcpy((void *)index, record, write_size);
-
-		record = (const char *)record + write_size;
-		size -= write_size;
-		bytes_written += write_size;
-
-		adj_write_size = write_size / ds_cfg.sizeof_rec[qual];
-		adj_write_size *= ds_cfg.sizeof_rec[qual];
-
-		/* Zero out trailing bytes. */
-		memset((char *)index + write_size, 0,
-		       adj_write_size - write_size);
-		index += adj_write_size;
-
-		if (index >= end)
-			index = base;
-		ds_set(context->ds, qual, ds_index, index);
-
-		if (index >= int_th)
-			ds_overflow(context, qual);
-	}
-
-	return bytes_written;
-}
-
-
-/*
- * Branch Trace Store (BTS) uses the following format. Different
- * architectures vary in the size of those fields.
- * - source linear address
- * - destination linear address
- * - flags
- *
- * Later architectures use 64bit pointers throughout, whereas earlier
- * architectures use 32bit pointers in 32bit mode.
- *
- * We compute the base address for the fields based on:
- * - the field size stored in the DS configuration
- * - the relative field position
- *
- * In order to store additional information in the BTS buffer, we use
- * a special source address to indicate that the record requires
- * special interpretation.
- *
- * Netburst indicated via a bit in the flags field whether the branch
- * was predicted; this is ignored.
- *
- * We use two levels of abstraction:
- * - the raw data level defined here
- * - an arch-independent level defined in ds.h
- */
-
-enum bts_field {
-	bts_from,
-	bts_to,
-	bts_flags,
-
-	bts_qual		= bts_from,
-	bts_clock		= bts_to,
-	bts_pid			= bts_flags,
-
-	bts_qual_mask		= (bts_qual_max - 1),
-	bts_escape		= ((unsigned long)-1 & ~bts_qual_mask)
-};
-
-static inline unsigned long bts_get(const char *base, unsigned long field)
-{
-	base += (ds_cfg.sizeof_ptr_field * field);
-	return *(unsigned long *)base;
-}
-
-static inline void bts_set(char *base, unsigned long field, unsigned long val)
-{
-	base += (ds_cfg.sizeof_ptr_field * field);
-	(*(unsigned long *)base) = val;
-}
-
-
-/*
- * The raw BTS data is architecture dependent.
- *
- * For higher-level users, we give an arch-independent view.
- * - ds.h defines struct bts_struct
- * - bts_read translates one raw bts record into a bts_struct
- * - bts_write translates one bts_struct into the raw format and
- *   writes it into the top of the parameter tracer's buffer.
- *
- * return: bytes read/written on success; -Eerrno, otherwise
- */
-static int
-bts_read(struct bts_tracer *tracer, const void *at, struct bts_struct *out)
-{
-	if (!tracer)
-		return -EINVAL;
-
-	if (at < tracer->trace.ds.begin)
-		return -EINVAL;
-
-	if (tracer->trace.ds.end < (at + tracer->trace.ds.size))
-		return -EINVAL;
-
-	memset(out, 0, sizeof(*out));
-	if ((bts_get(at, bts_qual) & ~bts_qual_mask) == bts_escape) {
-		out->qualifier = (bts_get(at, bts_qual) & bts_qual_mask);
-		out->variant.event.clock = bts_get(at, bts_clock);
-		out->variant.event.pid = bts_get(at, bts_pid);
-	} else {
-		out->qualifier = bts_branch;
-		out->variant.lbr.from = bts_get(at, bts_from);
-		out->variant.lbr.to   = bts_get(at, bts_to);
-
-		if (!out->variant.lbr.from && !out->variant.lbr.to)
-			out->qualifier = bts_invalid;
-	}
-
-	return ds_cfg.sizeof_rec[ds_bts];
-}
-
-static int bts_write(struct bts_tracer *tracer, const struct bts_struct *in)
-{
-	unsigned char raw[MAX_SIZEOF_BTS];
-
-	if (!tracer)
-		return -EINVAL;
-
-	if (MAX_SIZEOF_BTS < ds_cfg.sizeof_rec[ds_bts])
-		return -EOVERFLOW;
-
-	switch (in->qualifier) {
-	case bts_invalid:
-		bts_set(raw, bts_from, 0);
-		bts_set(raw, bts_to, 0);
-		bts_set(raw, bts_flags, 0);
-		break;
-	case bts_branch:
-		bts_set(raw, bts_from, in->variant.lbr.from);
-		bts_set(raw, bts_to,   in->variant.lbr.to);
-		bts_set(raw, bts_flags, 0);
-		break;
-	case bts_task_arrives:
-	case bts_task_departs:
-		bts_set(raw, bts_qual, (bts_escape | in->qualifier));
-		bts_set(raw, bts_clock, in->variant.event.clock);
-		bts_set(raw, bts_pid, in->variant.event.pid);
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	return ds_write(tracer->ds.context, ds_bts, raw,
-			ds_cfg.sizeof_rec[ds_bts]);
-}
-
-
-static void ds_write_config(struct ds_context *context,
-			    struct ds_trace *cfg, enum ds_qualifier qual)
-{
-	unsigned char *ds = context->ds;
-
-	ds_set(ds, qual, ds_buffer_base, (unsigned long)cfg->begin);
-	ds_set(ds, qual, ds_index, (unsigned long)cfg->top);
-	ds_set(ds, qual, ds_absolute_maximum, (unsigned long)cfg->end);
-	ds_set(ds, qual, ds_interrupt_threshold, (unsigned long)cfg->ith);
-}
-
-static void ds_read_config(struct ds_context *context,
-			   struct ds_trace *cfg, enum ds_qualifier qual)
-{
-	unsigned char *ds = context->ds;
-
-	cfg->begin = (void *)ds_get(ds, qual, ds_buffer_base);
-	cfg->top = (void *)ds_get(ds, qual, ds_index);
-	cfg->end = (void *)ds_get(ds, qual, ds_absolute_maximum);
-	cfg->ith = (void *)ds_get(ds, qual, ds_interrupt_threshold);
-}
-
-static void ds_init_ds_trace(struct ds_trace *trace, enum ds_qualifier qual,
-			     void *base, size_t size, size_t ith,
-			     unsigned int flags) {
-	unsigned long buffer, adj;
-
-	/*
-	 * Adjust the buffer address and size to meet alignment
-	 * constraints:
-	 * - buffer is double-word aligned
-	 * - size is multiple of record size
-	 *
-	 * We checked the size at the very beginning; we have enough
-	 * space to do the adjustment.
-	 */
-	buffer = (unsigned long)base;
-
-	adj = ALIGN(buffer, DS_ALIGNMENT) - buffer;
-	buffer += adj;
-	size   -= adj;
-
-	trace->n = size / ds_cfg.sizeof_rec[qual];
-	trace->size = ds_cfg.sizeof_rec[qual];
-
-	size = (trace->n * trace->size);
-
-	trace->begin = (void *)buffer;
-	trace->top = trace->begin;
-	trace->end = (void *)(buffer + size);
-	/*
-	 * The value for 'no threshold' is -1, which will set the
-	 * threshold outside of the buffer, just like we want it.
-	 */
-	ith *= ds_cfg.sizeof_rec[qual];
-	trace->ith = (void *)(buffer + size - ith);
-
-	trace->flags = flags;
-}
-
-
-static int ds_request(struct ds_tracer *tracer, struct ds_trace *trace,
-		      enum ds_qualifier qual, struct task_struct *task,
-		      int cpu, void *base, size_t size, size_t th)
-{
-	struct ds_context *context;
-	int error;
-	size_t req_size;
-
-	error = -EOPNOTSUPP;
-	if (!ds_cfg.sizeof_rec[qual])
-		goto out;
-
-	error = -EINVAL;
-	if (!base)
-		goto out;
-
-	req_size = ds_cfg.sizeof_rec[qual];
-	/* We might need space for alignment adjustments. */
-	if (!IS_ALIGNED((unsigned long)base, DS_ALIGNMENT))
-		req_size += DS_ALIGNMENT;
-
-	error = -EINVAL;
-	if (size < req_size)
-		goto out;
-
-	if (th != (size_t)-1) {
-		th *= ds_cfg.sizeof_rec[qual];
-
-		error = -EINVAL;
-		if (size <= th)
-			goto out;
-	}
-
-	tracer->buffer = base;
-	tracer->size = size;
-
-	error = -ENOMEM;
-	context = ds_get_context(task, cpu);
-	if (!context)
-		goto out;
-	tracer->context = context;
-
-	/*
-	 * Defer any tracer-specific initialization work for the context until
-	 * context ownership has been clarified.
-	 */
-
-	error = 0;
- out:
-	return error;
-}
-
-static struct bts_tracer *ds_request_bts(struct task_struct *task, int cpu,
-					 void *base, size_t size,
-					 bts_ovfl_callback_t ovfl, size_t th,
-					 unsigned int flags)
-{
-	struct bts_tracer *tracer;
-	int error;
-
-	/* Buffer overflow notification is not yet implemented. */
-	error = -EOPNOTSUPP;
-	if (ovfl)
-		goto out;
-
-	error = get_tracer(task);
-	if (error < 0)
-		goto out;
-
-	error = -ENOMEM;
-	tracer = kzalloc(sizeof(*tracer), GFP_KERNEL);
-	if (!tracer)
-		goto out_put_tracer;
-	tracer->ovfl = ovfl;
-
-	/* Do some more error checking and acquire a tracing context. */
-	error = ds_request(&tracer->ds, &tracer->trace.ds,
-			   ds_bts, task, cpu, base, size, th);
-	if (error < 0)
-		goto out_tracer;
-
-	/* Claim the bts part of the tracing context we acquired above. */
-	spin_lock_irq(&ds_lock);
-
-	error = -EPERM;
-	if (tracer->ds.context->bts_master)
-		goto out_unlock;
-	tracer->ds.context->bts_master = tracer;
-
-	spin_unlock_irq(&ds_lock);
-
-	/*
-	 * Now that we own the bts part of the context, let's complete the
-	 * initialization for that part.
-	 */
-	ds_init_ds_trace(&tracer->trace.ds, ds_bts, base, size, th, flags);
-	ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts);
-	ds_install_ds_area(tracer->ds.context);
-
-	tracer->trace.read  = bts_read;
-	tracer->trace.write = bts_write;
-
-	/* Start tracing. */
-	ds_resume_bts(tracer);
-
-	return tracer;
-
- out_unlock:
-	spin_unlock_irq(&ds_lock);
-	ds_put_context(tracer->ds.context);
- out_tracer:
-	kfree(tracer);
- out_put_tracer:
-	put_tracer(task);
- out:
-	return ERR_PTR(error);
-}
-
-struct bts_tracer *ds_request_bts_task(struct task_struct *task,
-				       void *base, size_t size,
-				       bts_ovfl_callback_t ovfl,
-				       size_t th, unsigned int flags)
-{
-	return ds_request_bts(task, 0, base, size, ovfl, th, flags);
-}
-
-struct bts_tracer *ds_request_bts_cpu(int cpu, void *base, size_t size,
-				      bts_ovfl_callback_t ovfl,
-				      size_t th, unsigned int flags)
-{
-	return ds_request_bts(NULL, cpu, base, size, ovfl, th, flags);
-}
-
-static struct pebs_tracer *ds_request_pebs(struct task_struct *task, int cpu,
-					   void *base, size_t size,
-					   pebs_ovfl_callback_t ovfl, size_t th,
-					   unsigned int flags)
-{
-	struct pebs_tracer *tracer;
-	int error;
-
-	/* Buffer overflow notification is not yet implemented. */
-	error = -EOPNOTSUPP;
-	if (ovfl)
-		goto out;
-
-	error = get_tracer(task);
-	if (error < 0)
-		goto out;
-
-	error = -ENOMEM;
-	tracer = kzalloc(sizeof(*tracer), GFP_KERNEL);
-	if (!tracer)
-		goto out_put_tracer;
-	tracer->ovfl = ovfl;
-
-	/* Do some more error checking and acquire a tracing context. */
-	error = ds_request(&tracer->ds, &tracer->trace.ds,
-			   ds_pebs, task, cpu, base, size, th);
-	if (error < 0)
-		goto out_tracer;
-
-	/* Claim the pebs part of the tracing context we acquired above. */
-	spin_lock_irq(&ds_lock);
-
-	error = -EPERM;
-	if (tracer->ds.context->pebs_master)
-		goto out_unlock;
-	tracer->ds.context->pebs_master = tracer;
-
-	spin_unlock_irq(&ds_lock);
-
-	/*
-	 * Now that we own the pebs part of the context, let's complete the
-	 * initialization for that part.
-	 */
-	ds_init_ds_trace(&tracer->trace.ds, ds_pebs, base, size, th, flags);
-	ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_pebs);
-	ds_install_ds_area(tracer->ds.context);
-
-	/* Start tracing. */
-	ds_resume_pebs(tracer);
-
-	return tracer;
-
- out_unlock:
-	spin_unlock_irq(&ds_lock);
-	ds_put_context(tracer->ds.context);
- out_tracer:
-	kfree(tracer);
- out_put_tracer:
-	put_tracer(task);
- out:
-	return ERR_PTR(error);
-}
-
-struct pebs_tracer *ds_request_pebs_task(struct task_struct *task,
-					 void *base, size_t size,
-					 pebs_ovfl_callback_t ovfl,
-					 size_t th, unsigned int flags)
-{
-	return ds_request_pebs(task, 0, base, size, ovfl, th, flags);
-}
-
-struct pebs_tracer *ds_request_pebs_cpu(int cpu, void *base, size_t size,
-					pebs_ovfl_callback_t ovfl,
-					size_t th, unsigned int flags)
-{
-	return ds_request_pebs(NULL, cpu, base, size, ovfl, th, flags);
-}
-
-static void ds_free_bts(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-
-	task = tracer->ds.context->task;
-
-	WARN_ON_ONCE(tracer->ds.context->bts_master != tracer);
-	tracer->ds.context->bts_master = NULL;
-
-	/* Make sure tracing stopped and the tracer is not in use. */
-	if (task && (task != current))
-		wait_task_context_switch(task);
-
-	ds_put_context(tracer->ds.context);
-	put_tracer(task);
-
-	kfree(tracer);
-}
-
-void ds_release_bts(struct bts_tracer *tracer)
-{
-	might_sleep();
-
-	if (!tracer)
-		return;
-
-	ds_suspend_bts(tracer);
-	ds_free_bts(tracer);
-}
-
-int ds_release_bts_noirq(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long irq;
-	int error;
-
-	if (!tracer)
-		return 0;
-
-	task = tracer->ds.context->task;
-
-	local_irq_save(irq);
-
-	error = -EPERM;
-	if (!task &&
-	    (tracer->ds.context->cpu != smp_processor_id()))
-		goto out;
-
-	error = -EPERM;
-	if (task && (task != current))
-		goto out;
-
-	ds_suspend_bts_noirq(tracer);
-	ds_free_bts(tracer);
-
-	error = 0;
- out:
-	local_irq_restore(irq);
-	return error;
-}
-
-static void update_task_debugctlmsr(struct task_struct *task,
-				    unsigned long debugctlmsr)
-{
-	task->thread.debugctlmsr = debugctlmsr;
-
-	get_cpu();
-	if (task == current)
-		update_debugctlmsr(debugctlmsr);
-	put_cpu();
-}
-
-void ds_suspend_bts(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long debugctlmsr;
-	int cpu;
-
-	if (!tracer)
-		return;
-
-	tracer->flags = 0;
-
-	task = tracer->ds.context->task;
-	cpu  = tracer->ds.context->cpu;
-
-	WARN_ON(!task && irqs_disabled());
-
-	debugctlmsr = (task ?
-		       task->thread.debugctlmsr :
-		       get_debugctlmsr_on_cpu(cpu));
-	debugctlmsr &= ~BTS_CONTROL;
-
-	if (task)
-		update_task_debugctlmsr(task, debugctlmsr);
-	else
-		update_debugctlmsr_on_cpu(cpu, debugctlmsr);
-}
-
-int ds_suspend_bts_noirq(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long debugctlmsr, irq;
-	int cpu, error = 0;
-
-	if (!tracer)
-		return 0;
-
-	tracer->flags = 0;
-
-	task = tracer->ds.context->task;
-	cpu  = tracer->ds.context->cpu;
-
-	local_irq_save(irq);
-
-	error = -EPERM;
-	if (!task && (cpu != smp_processor_id()))
-		goto out;
-
-	debugctlmsr = (task ?
-		       task->thread.debugctlmsr :
-		       get_debugctlmsr());
-	debugctlmsr &= ~BTS_CONTROL;
-
-	if (task)
-		update_task_debugctlmsr(task, debugctlmsr);
-	else
-		update_debugctlmsr(debugctlmsr);
-
-	error = 0;
- out:
-	local_irq_restore(irq);
-	return error;
-}
-
-static unsigned long ds_bts_control(struct bts_tracer *tracer)
-{
-	unsigned long control;
-
-	control = ds_cfg.ctl[dsf_bts];
-	if (!(tracer->trace.ds.flags & BTS_KERNEL))
-		control |= ds_cfg.ctl[dsf_bts_kernel];
-	if (!(tracer->trace.ds.flags & BTS_USER))
-		control |= ds_cfg.ctl[dsf_bts_user];
-
-	return control;
-}
-
-void ds_resume_bts(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long debugctlmsr;
-	int cpu;
-
-	if (!tracer)
-		return;
-
-	tracer->flags = tracer->trace.ds.flags;
-
-	task = tracer->ds.context->task;
-	cpu  = tracer->ds.context->cpu;
-
-	WARN_ON(!task && irqs_disabled());
-
-	debugctlmsr = (task ?
-		       task->thread.debugctlmsr :
-		       get_debugctlmsr_on_cpu(cpu));
-	debugctlmsr |= ds_bts_control(tracer);
-
-	if (task)
-		update_task_debugctlmsr(task, debugctlmsr);
-	else
-		update_debugctlmsr_on_cpu(cpu, debugctlmsr);
-}
-
-int ds_resume_bts_noirq(struct bts_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long debugctlmsr, irq;
-	int cpu, error = 0;
-
-	if (!tracer)
-		return 0;
-
-	tracer->flags = tracer->trace.ds.flags;
-
-	task = tracer->ds.context->task;
-	cpu  = tracer->ds.context->cpu;
-
-	local_irq_save(irq);
-
-	error = -EPERM;
-	if (!task && (cpu != smp_processor_id()))
-		goto out;
-
-	debugctlmsr = (task ?
-		       task->thread.debugctlmsr :
-		       get_debugctlmsr());
-	debugctlmsr |= ds_bts_control(tracer);
-
-	if (task)
-		update_task_debugctlmsr(task, debugctlmsr);
-	else
-		update_debugctlmsr(debugctlmsr);
-
-	error = 0;
- out:
-	local_irq_restore(irq);
-	return error;
-}
-
-static void ds_free_pebs(struct pebs_tracer *tracer)
-{
-	struct task_struct *task;
-
-	task = tracer->ds.context->task;
-
-	WARN_ON_ONCE(tracer->ds.context->pebs_master != tracer);
-	tracer->ds.context->pebs_master = NULL;
-
-	ds_put_context(tracer->ds.context);
-	put_tracer(task);
-
-	kfree(tracer);
-}
-
-void ds_release_pebs(struct pebs_tracer *tracer)
-{
-	might_sleep();
-
-	if (!tracer)
-		return;
-
-	ds_suspend_pebs(tracer);
-	ds_free_pebs(tracer);
-}
-
-int ds_release_pebs_noirq(struct pebs_tracer *tracer)
-{
-	struct task_struct *task;
-	unsigned long irq;
-	int error;
-
-	if (!tracer)
-		return 0;
-
-	task = tracer->ds.context->task;
-
-	local_irq_save(irq);
-
-	error = -EPERM;
-	if (!task &&
-	    (tracer->ds.context->cpu != smp_processor_id()))
-		goto out;
-
-	error = -EPERM;
-	if (task && (task != current))
-		goto out;
-
-	ds_suspend_pebs_noirq(tracer);
-	ds_free_pebs(tracer);
-
-	error = 0;
- out:
-	local_irq_restore(irq);
-	return error;
-}
-
-void ds_suspend_pebs(struct pebs_tracer *tracer)
-{
-
-}
-
-int ds_suspend_pebs_noirq(struct pebs_tracer *tracer)
-{
-	return 0;
-}
-
-void ds_resume_pebs(struct pebs_tracer *tracer)
-{
-
-}
-
-int ds_resume_pebs_noirq(struct pebs_tracer *tracer)
-{
-	return 0;
-}
-
-const struct bts_trace *ds_read_bts(struct bts_tracer *tracer)
-{
-	if (!tracer)
-		return NULL;
-
-	ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_bts);
-	return &tracer->trace;
-}
-
-const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer)
-{
-	if (!tracer)
-		return NULL;
-
-	ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_pebs);
-
-	tracer->trace.counters = ds_cfg.nr_counter_reset;
-	memcpy(tracer->trace.counter_reset,
-	       tracer->ds.context->ds +
-	       (NUM_DS_PTR_FIELDS * ds_cfg.sizeof_ptr_field),
-	       ds_cfg.nr_counter_reset * PEBS_RESET_FIELD_SIZE);
-
-	return &tracer->trace;
-}
-
-int ds_reset_bts(struct bts_tracer *tracer)
-{
-	if (!tracer)
-		return -EINVAL;
-
-	tracer->trace.ds.top = tracer->trace.ds.begin;
-
-	ds_set(tracer->ds.context->ds, ds_bts, ds_index,
-	       (unsigned long)tracer->trace.ds.top);
-
-	return 0;
-}
-
-int ds_reset_pebs(struct pebs_tracer *tracer)
-{
-	if (!tracer)
-		return -EINVAL;
-
-	tracer->trace.ds.top = tracer->trace.ds.begin;
-
-	ds_set(tracer->ds.context->ds, ds_pebs, ds_index,
-	       (unsigned long)tracer->trace.ds.top);
-
-	return 0;
-}
-
-int ds_set_pebs_reset(struct pebs_tracer *tracer,
-		      unsigned int counter, u64 value)
-{
-	if (!tracer)
-		return -EINVAL;
-
-	if (ds_cfg.nr_counter_reset < counter)
-		return -EINVAL;
-
-	*(u64 *)(tracer->ds.context->ds +
-		 (NUM_DS_PTR_FIELDS * ds_cfg.sizeof_ptr_field) +
-		 (counter * PEBS_RESET_FIELD_SIZE)) = value;
-
-	return 0;
-}
-
-static const struct ds_configuration ds_cfg_netburst = {
-	.name = "Netburst",
-	.ctl[dsf_bts]		= (1 << 2) | (1 << 3),
-	.ctl[dsf_bts_kernel]	= (1 << 5),
-	.ctl[dsf_bts_user]	= (1 << 6),
-	.nr_counter_reset	= 1,
-};
-static const struct ds_configuration ds_cfg_pentium_m = {
-	.name = "Pentium M",
-	.ctl[dsf_bts]		= (1 << 6) | (1 << 7),
-	.nr_counter_reset	= 1,
-};
-static const struct ds_configuration ds_cfg_core2_atom = {
-	.name = "Core 2/Atom",
-	.ctl[dsf_bts]		= (1 << 6) | (1 << 7),
-	.ctl[dsf_bts_kernel]	= (1 << 9),
-	.ctl[dsf_bts_user]	= (1 << 10),
-	.nr_counter_reset	= 1,
-};
-static const struct ds_configuration ds_cfg_core_i7 = {
-	.name = "Core i7",
-	.ctl[dsf_bts]		= (1 << 6) | (1 << 7),
-	.ctl[dsf_bts_kernel]	= (1 << 9),
-	.ctl[dsf_bts_user]	= (1 << 10),
-	.nr_counter_reset	= 4,
-};
-
-static void
-ds_configure(const struct ds_configuration *cfg,
-	     struct cpuinfo_x86 *cpu)
-{
-	unsigned long nr_pebs_fields = 0;
-
-	printk(KERN_INFO "[ds] using %s configuration\n", cfg->name);
-
-#ifdef __i386__
-	nr_pebs_fields = 10;
-#else
-	nr_pebs_fields = 18;
-#endif
-
-	/*
-	 * Starting with version 2, architectural performance
-	 * monitoring supports a format specifier.
-	 */
-	if ((cpuid_eax(0xa) & 0xff) > 1) {
-		unsigned long perf_capabilities, format;
-
-		rdmsrl(MSR_IA32_PERF_CAPABILITIES, perf_capabilities);
-
-		format = (perf_capabilities >> 8) & 0xf;
-
-		switch (format) {
-		case 0:
-			nr_pebs_fields = 18;
-			break;
-		case 1:
-			nr_pebs_fields = 22;
-			break;
-		default:
-			printk(KERN_INFO
-			       "[ds] unknown PEBS format: %lu\n", format);
-			nr_pebs_fields = 0;
-			break;
-		}
-	}
-
-	memset(&ds_cfg, 0, sizeof(ds_cfg));
-	ds_cfg = *cfg;
-
-	ds_cfg.sizeof_ptr_field =
-		(cpu_has(cpu, X86_FEATURE_DTES64) ? 8 : 4);
-
-	ds_cfg.sizeof_rec[ds_bts]  = ds_cfg.sizeof_ptr_field * 3;
-	ds_cfg.sizeof_rec[ds_pebs] = ds_cfg.sizeof_ptr_field * nr_pebs_fields;
-
-	if (!cpu_has(cpu, X86_FEATURE_BTS)) {
-		ds_cfg.sizeof_rec[ds_bts] = 0;
-		printk(KERN_INFO "[ds] bts not available\n");
-	}
-	if (!cpu_has(cpu, X86_FEATURE_PEBS)) {
-		ds_cfg.sizeof_rec[ds_pebs] = 0;
-		printk(KERN_INFO "[ds] pebs not available\n");
-	}
-
-	printk(KERN_INFO "[ds] sizes: address: %u bit, ",
-	       8 * ds_cfg.sizeof_ptr_field);
-	printk("bts/pebs record: %u/%u bytes\n",
-	       ds_cfg.sizeof_rec[ds_bts], ds_cfg.sizeof_rec[ds_pebs]);
-
-	WARN_ON_ONCE(MAX_PEBS_COUNTERS < ds_cfg.nr_counter_reset);
-}
-
-void __cpuinit ds_init_intel(struct cpuinfo_x86 *c)
-{
-	/* Only configure the first cpu. Others are identical. */
-	if (ds_cfg.name)
-		return;
-
-	switch (c->x86) {
-	case 0x6:
-		switch (c->x86_model) {
-		case 0x9:
-		case 0xd: /* Pentium M */
-			ds_configure(&ds_cfg_pentium_m, c);
-			break;
-		case 0xf:
-		case 0x17: /* Core2 */
-		case 0x1c: /* Atom */
-			ds_configure(&ds_cfg_core2_atom, c);
-			break;
-		case 0x1a: /* Core i7 */
-			ds_configure(&ds_cfg_core_i7, c);
-			break;
-		default:
-			/* Sorry, don't know about them. */
-			break;
-		}
-		break;
-	case 0xf:
-		switch (c->x86_model) {
-		case 0x0:
-		case 0x1:
-		case 0x2: /* Netburst */
-			ds_configure(&ds_cfg_netburst, c);
-			break;
-		default:
-			/* Sorry, don't know about them. */
-			break;
-		}
-		break;
-	default:
-		/* Sorry, don't know about them. */
-		break;
-	}
-}
-
-static inline void ds_take_timestamp(struct ds_context *context,
-				     enum bts_qualifier qualifier,
-				     struct task_struct *task)
-{
-	struct bts_tracer *tracer = context->bts_master;
-	struct bts_struct ts;
-
-	/* Prevent compilers from reading the tracer pointer twice. */
-	barrier();
-
-	if (!tracer || !(tracer->flags & BTS_TIMESTAMPS))
-		return;
-
-	memset(&ts, 0, sizeof(ts));
-	ts.qualifier		= qualifier;
-	ts.variant.event.clock	= trace_clock_global();
-	ts.variant.event.pid	= task->pid;
-
-	bts_write(tracer, &ts);
-}
-
-/*
- * Change the DS configuration from tracing prev to tracing next.
- */
-void ds_switch_to(struct task_struct *prev, struct task_struct *next)
-{
-	struct ds_context *prev_ctx	= prev->thread.ds_ctx;
-	struct ds_context *next_ctx	= next->thread.ds_ctx;
-	unsigned long debugctlmsr	= next->thread.debugctlmsr;
-
-	/* Make sure all data is read before we start. */
-	barrier();
-
-	if (prev_ctx) {
-		update_debugctlmsr(0);
-
-		ds_take_timestamp(prev_ctx, bts_task_departs, prev);
-	}
-
-	if (next_ctx) {
-		ds_take_timestamp(next_ctx, bts_task_arrives, next);
-
-		wrmsrl(MSR_IA32_DS_AREA, (unsigned long)next_ctx->ds);
-	}
-
-	update_debugctlmsr(debugctlmsr);
-}
-
-static __init int ds_selftest(void)
-{
-	if (ds_cfg.sizeof_rec[ds_bts]) {
-		int error;
-
-		error = ds_selftest_bts();
-		if (error) {
-			WARN(1, "[ds] selftest failed. disabling bts.\n");
-			ds_cfg.sizeof_rec[ds_bts] = 0;
-		}
-	}
-
-	if (ds_cfg.sizeof_rec[ds_pebs]) {
-		int error;
-
-		error = ds_selftest_pebs();
-		if (error) {
-			WARN(1, "[ds] selftest failed. disabling pebs.\n");
-			ds_cfg.sizeof_rec[ds_pebs] = 0;
-		}
-	}
-
-	return 0;
-}
-device_initcall(ds_selftest);
diff --git a/arch/x86/kernel/ds_selftest.c b/arch/x86/kernel/ds_selftest.c
deleted file mode 100644
index 6bc7c199ab9..00000000000
--- a/arch/x86/kernel/ds_selftest.c
+++ /dev/null
@@ -1,408 +0,0 @@
-/*
- * Debug Store support - selftest
- *
- *
- * Copyright (C) 2009 Intel Corporation.
- * Markus Metzger <markus.t.metzger@intel.com>, 2009
- */
-
-#include "ds_selftest.h"
-
-#include <linux/kernel.h>
-#include <linux/string.h>
-#include <linux/smp.h>
-#include <linux/cpu.h>
-
-#include <asm/ds.h>
-
-
-#define BUFFER_SIZE		521	/* Intentionally chose an odd size. */
-#define SMALL_BUFFER_SIZE	24	/* A single bts entry. */
-
-struct ds_selftest_bts_conf {
-	struct bts_tracer *tracer;
-	int error;
-	int (*suspend)(struct bts_tracer *);
-	int (*resume)(struct bts_tracer *);
-};
-
-static int ds_selftest_bts_consistency(const struct bts_trace *trace)
-{
-	int error = 0;
-
-	if (!trace) {
-		printk(KERN_CONT "failed to access trace...");
-		/* Bail out. Other tests are pointless. */
-		return -1;
-	}
-
-	if (!trace->read) {
-		printk(KERN_CONT "bts read not available...");
-		error = -1;
-	}
-
-	/* Do some sanity checks on the trace configuration. */
-	if (!trace->ds.n) {
-		printk(KERN_CONT "empty bts buffer...");
-		error = -1;
-	}
-	if (!trace->ds.size) {
-		printk(KERN_CONT "bad bts trace setup...");
-		error = -1;
-	}
-	if (trace->ds.end !=
-	    (char *)trace->ds.begin + (trace->ds.n * trace->ds.size)) {
-		printk(KERN_CONT "bad bts buffer setup...");
-		error = -1;
-	}
-	/*
-	 * We allow top in [begin; end], since its not clear when the
-	 * overflow adjustment happens: after the increment or before the
-	 * write.
-	 */
-	if ((trace->ds.top < trace->ds.begin) ||
-	    (trace->ds.end < trace->ds.top)) {
-		printk(KERN_CONT "bts top out of bounds...");
-		error = -1;
-	}
-
-	return error;
-}
-
-static int ds_selftest_bts_read(struct bts_tracer *tracer,
-				const struct bts_trace *trace,
-				const void *from, const void *to)
-{
-	const unsigned char *at;
-
-	/*
-	 * Check a few things which do not belong to this test.
-	 * They should be covered by other tests.
-	 */
-	if (!trace)
-		return -1;
-
-	if (!trace->read)
-		return -1;
-
-	if (to < from)
-		return -1;
-
-	if (from < trace->ds.begin)
-		return -1;
-
-	if (trace->ds.end < to)
-		return -1;
-
-	if (!trace->ds.size)
-		return -1;
-
-	/* Now to the test itself. */
-	for (at = from; (void *)at < to; at += trace->ds.size) {
-		struct bts_struct bts;
-		unsigned long index;
-		int error;
-
-		if (((void *)at - trace->ds.begin) % trace->ds.size) {
-			printk(KERN_CONT
-			       "read from non-integer index...");
-			return -1;
-		}
-		index = ((void *)at - trace->ds.begin) / trace->ds.size;
-
-		memset(&bts, 0, sizeof(bts));
-		error = trace->read(tracer, at, &bts);
-		if (error < 0) {
-			printk(KERN_CONT
-			       "error reading bts trace at [%lu] (0x%p)...",
-			       index, at);
-			return error;
-		}
-
-		switch (bts.qualifier) {
-		case BTS_BRANCH:
-			break;
-		default:
-			printk(KERN_CONT
-			       "unexpected bts entry %llu at [%lu] (0x%p)...",
-			       bts.qualifier, index, at);
-			return -1;
-		}
-	}
-
-	return 0;
-}
-
-static void ds_selftest_bts_cpu(void *arg)
-{
-	struct ds_selftest_bts_conf *conf = arg;
-	const struct bts_trace *trace;
-	void *top;
-
-	if (IS_ERR(conf->tracer)) {
-		conf->error = PTR_ERR(conf->tracer);
-		conf->tracer = NULL;
-
-		printk(KERN_CONT
-		       "initialization failed (err: %d)...", conf->error);
-		return;
-	}
-
-	/* We should meanwhile have enough trace. */
-	conf->error = conf->suspend(conf->tracer);
-	if (conf->error < 0)
-		return;
-
-	/* Let's see if we can access the trace. */
-	trace = ds_read_bts(conf->tracer);
-
-	conf->error = ds_selftest_bts_consistency(trace);
-	if (conf->error < 0)
-		return;
-
-	/* If everything went well, we should have a few trace entries. */
-	if (trace->ds.top == trace->ds.begin) {
-		/*
-		 * It is possible but highly unlikely that we got a
-		 * buffer overflow and end up at exactly the same
-		 * position we started from.
-		 * Let's issue a warning, but continue.
-		 */
-		printk(KERN_CONT "no trace/overflow...");
-	}
-
-	/* Let's try to read the trace we collected. */
-	conf->error =
-		ds_selftest_bts_read(conf->tracer, trace,
-				     trace->ds.begin, trace->ds.top);
-	if (conf->error < 0)
-		return;
-
-	/*
-	 * Let's read the trace again.
-	 * Since we suspended tracing, we should get the same result.
-	 */
-	top = trace->ds.top;
-
-	trace = ds_read_bts(conf->tracer);
-	conf->error = ds_selftest_bts_consistency(trace);
-	if (conf->error < 0)
-		return;
-
-	if (top != trace->ds.top) {
-		printk(KERN_CONT "suspend not working...");
-		conf->error = -1;
-		return;
-	}
-
-	/* Let's collect some more trace - see if resume is working. */
-	conf->error = conf->resume(conf->tracer);
-	if (conf->error < 0)
-		return;
-
-	conf->error = conf->suspend(conf->tracer);
-	if (conf->error < 0)
-		return;
-
-	trace = ds_read_bts(conf->tracer);
-
-	conf->error = ds_selftest_bts_consistency(trace);
-	if (conf->error < 0)
-		return;
-
-	if (trace->ds.top == top) {
-		/*
-		 * It is possible but highly unlikely that we got a
-		 * buffer overflow and end up at exactly the same
-		 * position we started from.
-		 * Let's issue a warning and check the full trace.
-		 */
-		printk(KERN_CONT
-		       "no resume progress/overflow...");
-
-		conf->error =
-			ds_selftest_bts_read(conf->tracer, trace,
-					     trace->ds.begin, trace->ds.end);
-	} else if (trace->ds.top < top) {
-		/*
-		 * We had a buffer overflow - the entire buffer should
-		 * contain trace records.
-		 */
-		conf->error =
-			ds_selftest_bts_read(conf->tracer, trace,
-					     trace->ds.begin, trace->ds.end);
-	} else {
-		/*
-		 * It is quite likely that the buffer did not overflow.
-		 * Let's just check the delta trace.
-		 */
-		conf->error =
-			ds_selftest_bts_read(conf->tracer, trace, top,
-					     trace->ds.top);
-	}
-	if (conf->error < 0)
-		return;
-
-	conf->error = 0;
-}
-
-static int ds_suspend_bts_wrap(struct bts_tracer *tracer)
-{
-	ds_suspend_bts(tracer);
-	return 0;
-}
-
-static int ds_resume_bts_wrap(struct bts_tracer *tracer)
-{
-	ds_resume_bts(tracer);
-	return 0;
-}
-
-static void ds_release_bts_noirq_wrap(void *tracer)
-{
-	(void)ds_release_bts_noirq(tracer);
-}
-
-static int ds_selftest_bts_bad_release_noirq(int cpu,
-					     struct bts_tracer *tracer)
-{
-	int error = -EPERM;
-
-	/* Try to release the tracer on the wrong cpu. */
-	get_cpu();
-	if (cpu != smp_processor_id()) {
-		error = ds_release_bts_noirq(tracer);
-		if (error != -EPERM)
-			printk(KERN_CONT "release on wrong cpu...");
-	}
-	put_cpu();
-
-	return error ? 0 : -1;
-}
-
-static int ds_selftest_bts_bad_request_cpu(int cpu, void *buffer)
-{
-	struct bts_tracer *tracer;
-	int error;
-
-	/* Try to request cpu tracing while task tracing is active. */
-	tracer = ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE, NULL,
-				    (size_t)-1, BTS_KERNEL);
-	error = PTR_ERR(tracer);
-	if (!IS_ERR(tracer)) {
-		ds_release_bts(tracer);
-		error = 0;
-	}
-
-	if (error != -EPERM)
-		printk(KERN_CONT "cpu/task tracing overlap...");
-
-	return error ? 0 : -1;
-}
-
-static int ds_selftest_bts_bad_request_task(void *buffer)
-{
-	struct bts_tracer *tracer;
-	int error;
-
-	/* Try to request cpu tracing while task tracing is active. */
-	tracer = ds_request_bts_task(current, buffer, BUFFER_SIZE, NULL,
-				    (size_t)-1, BTS_KERNEL);
-	error = PTR_ERR(tracer);
-	if (!IS_ERR(tracer)) {
-		error = 0;
-		ds_release_bts(tracer);
-	}
-
-	if (error != -EPERM)
-		printk(KERN_CONT "task/cpu tracing overlap...");
-
-	return error ? 0 : -1;
-}
-
-int ds_selftest_bts(void)
-{
-	struct ds_selftest_bts_conf conf;
-	unsigned char buffer[BUFFER_SIZE], *small_buffer;
-	unsigned long irq;
-	int cpu;
-
-	printk(KERN_INFO "[ds] bts selftest...");
-	conf.error = 0;
-
-	small_buffer = (unsigned char *)ALIGN((unsigned long)buffer, 8) + 8;
-
-	get_online_cpus();
-	for_each_online_cpu(cpu) {
-		conf.suspend = ds_suspend_bts_wrap;
-		conf.resume = ds_resume_bts_wrap;
-		conf.tracer =
-			ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE,
-					   NULL, (size_t)-1, BTS_KERNEL);
-		ds_selftest_bts_cpu(&conf);
-		if (conf.error >= 0)
-			conf.error = ds_selftest_bts_bad_request_task(buffer);
-		ds_release_bts(conf.tracer);
-		if (conf.error < 0)
-			goto out;
-
-		conf.suspend = ds_suspend_bts_noirq;
-		conf.resume = ds_resume_bts_noirq;
-		conf.tracer =
-			ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE,
-					   NULL, (size_t)-1, BTS_KERNEL);
-		smp_call_function_single(cpu, ds_selftest_bts_cpu, &conf, 1);
-		if (conf.error >= 0) {
-			conf.error =
-				ds_selftest_bts_bad_release_noirq(cpu,
-								  conf.tracer);
-			/* We must not release the tracer twice. */
-			if (conf.error < 0)
-				conf.tracer = NULL;
-		}
-		if (conf.error >= 0)
-			conf.error = ds_selftest_bts_bad_request_task(buffer);
-		smp_call_function_single(cpu, ds_release_bts_noirq_wrap,
-					 conf.tracer, 1);
-		if (conf.error < 0)
-			goto out;
-	}
-
-	conf.suspend = ds_suspend_bts_wrap;
-	conf.resume = ds_resume_bts_wrap;
-	conf.tracer =
-		ds_request_bts_task(current, buffer, BUFFER_SIZE,
-				    NULL, (size_t)-1, BTS_KERNEL);
-	ds_selftest_bts_cpu(&conf);
-	if (conf.error >= 0)
-		conf.error = ds_selftest_bts_bad_request_cpu(0, buffer);
-	ds_release_bts(conf.tracer);
-	if (conf.error < 0)
-		goto out;
-
-	conf.suspend = ds_suspend_bts_noirq;
-	conf.resume = ds_resume_bts_noirq;
-	conf.tracer =
-		ds_request_bts_task(current, small_buffer, SMALL_BUFFER_SIZE,
-				   NULL, (size_t)-1, BTS_KERNEL);
-	local_irq_save(irq);
-	ds_selftest_bts_cpu(&conf);
-	if (conf.error >= 0)
-		conf.error = ds_selftest_bts_bad_request_cpu(0, buffer);
-	ds_release_bts_noirq(conf.tracer);
-	local_irq_restore(irq);
-	if (conf.error < 0)
-		goto out;
-
-	conf.error = 0;
- out:
-	put_online_cpus();
-	printk(KERN_CONT "%s.\n", (conf.error ? "failed" : "passed"));
-
-	return conf.error;
-}
-
-int ds_selftest_pebs(void)
-{
-	return 0;
-}
diff --git a/arch/x86/kernel/ds_selftest.h b/arch/x86/kernel/ds_selftest.h
deleted file mode 100644
index 2ba8745c666..00000000000
--- a/arch/x86/kernel/ds_selftest.h
+++ /dev/null
@@ -1,15 +0,0 @@
-/*
- * Debug Store support - selftest
- *
- *
- * Copyright (C) 2009 Intel Corporation.
- * Markus Metzger <markus.t.metzger@intel.com>, 2009
- */
-
-#ifdef CONFIG_X86_DS_SELFTEST
-extern int ds_selftest_bts(void);
-extern int ds_selftest_pebs(void);
-#else
-static inline int ds_selftest_bts(void) { return 0; }
-static inline int ds_selftest_pebs(void) { return 0; }
-#endif
diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
index 6d817554780..c89a386930b 100644
--- a/arch/x86/kernel/dumpstack.c
+++ b/arch/x86/kernel/dumpstack.c
@@ -224,11 +224,6 @@ unsigned __kprobes long oops_begin(void)
 	int cpu;
 	unsigned long flags;
 
-	/* notify the hw-branch tracer so it may disable tracing and
-	   add the last trace to the trace buffer -
-	   the earlier this happens, the more useful the trace. */
-	trace_hw_branch_oops();
-
 	oops_enter();
 
 	/* racy, but better than risking deadlock. */
diff --git a/arch/x86/kernel/hw_breakpoint.c b/arch/x86/kernel/hw_breakpoint.c
index d6cc065f519..a8f1b803d2f 100644
--- a/arch/x86/kernel/hw_breakpoint.c
+++ b/arch/x86/kernel/hw_breakpoint.c
@@ -189,25 +189,16 @@ static int get_hbp_len(u8 hbp_len)
 }
 
 /*
- * Check for virtual address in user space.
- */
-int arch_check_va_in_userspace(unsigned long va, u8 hbp_len)
-{
-	unsigned int len;
-
-	len = get_hbp_len(hbp_len);
-
-	return (va <= TASK_SIZE - len);
-}
-
-/*
  * Check for virtual address in kernel space.
  */
-static int arch_check_va_in_kernelspace(unsigned long va, u8 hbp_len)
+int arch_check_bp_in_kernelspace(struct perf_event *bp)
 {
 	unsigned int len;
+	unsigned long va;
+	struct arch_hw_breakpoint *info = counter_arch_bp(bp);
 
-	len = get_hbp_len(hbp_len);
+	va = info->address;
+	len = get_hbp_len(info->len);
 
 	return (va >= TASK_SIZE) && ((va + len - 1) >= TASK_SIZE);
 }
@@ -300,8 +291,7 @@ static int arch_build_bp_info(struct perf_event *bp)
 /*
  * Validate the arch-specific HW Breakpoint register settings
  */
-int arch_validate_hwbkpt_settings(struct perf_event *bp,
-				  struct task_struct *tsk)
+int arch_validate_hwbkpt_settings(struct perf_event *bp)
 {
 	struct arch_hw_breakpoint *info = counter_arch_bp(bp);
 	unsigned int align;
@@ -314,16 +304,6 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp,
 
 	ret = -EINVAL;
 
-	if (info->type == X86_BREAKPOINT_EXECUTE)
-		/*
-		 * Ptrace-refactoring code
-		 * For now, we'll allow instruction breakpoint only for user-space
-		 * addresses
-		 */
-		if ((!arch_check_va_in_userspace(info->address, info->len)) &&
-			info->len != X86_BREAKPOINT_EXECUTE)
-			return ret;
-
 	switch (info->len) {
 	case X86_BREAKPOINT_LEN_1:
 		align = 0;
@@ -350,15 +330,6 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp,
 	if (info->address & align)
 		return -EINVAL;
 
-	/* Check that the virtual address is in the proper range */
-	if (tsk) {
-		if (!arch_check_va_in_userspace(info->address, info->len))
-			return -EFAULT;
-	} else {
-		if (!arch_check_va_in_kernelspace(info->address, info->len))
-			return -EFAULT;
-	}
-
 	return 0;
 }
 
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 1658efdfb4e..345a4b1fe14 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -422,14 +422,22 @@ static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs,
 
 static void __kprobes clear_btf(void)
 {
-	if (test_thread_flag(TIF_DEBUGCTLMSR))
-		update_debugctlmsr(0);
+	if (test_thread_flag(TIF_BLOCKSTEP)) {
+		unsigned long debugctl = get_debugctlmsr();
+
+		debugctl &= ~DEBUGCTLMSR_BTF;
+		update_debugctlmsr(debugctl);
+	}
 }
 
 static void __kprobes restore_btf(void)
 {
-	if (test_thread_flag(TIF_DEBUGCTLMSR))
-		update_debugctlmsr(current->thread.debugctlmsr);
+	if (test_thread_flag(TIF_BLOCKSTEP)) {
+		unsigned long debugctl = get_debugctlmsr();
+
+		debugctl |= DEBUGCTLMSR_BTF;
+		update_debugctlmsr(debugctl);
+	}
 }
 
 void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 0415c3ef91b..cc6877535ef 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -20,7 +20,6 @@
 #include <asm/idle.h>
 #include <asm/uaccess.h>
 #include <asm/i387.h>
-#include <asm/ds.h>
 #include <asm/debugreg.h>
 
 unsigned long idle_halt;
@@ -50,8 +49,6 @@ void free_thread_xstate(struct task_struct *tsk)
 		kmem_cache_free(task_xstate_cachep, tsk->thread.xstate);
 		tsk->thread.xstate = NULL;
 	}
-
-	WARN(tsk->thread.ds_ctx, "leaking DS context\n");
 }
 
 void free_thread_info(struct thread_info *ti)
@@ -198,11 +195,16 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
 	prev = &prev_p->thread;
 	next = &next_p->thread;
 
-	if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) ||
-	    test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR))
-		ds_switch_to(prev_p, next_p);
-	else if (next->debugctlmsr != prev->debugctlmsr)
-		update_debugctlmsr(next->debugctlmsr);
+	if (test_tsk_thread_flag(prev_p, TIF_BLOCKSTEP) ^
+	    test_tsk_thread_flag(next_p, TIF_BLOCKSTEP)) {
+		unsigned long debugctl = get_debugctlmsr();
+
+		debugctl &= ~DEBUGCTLMSR_BTF;
+		if (test_tsk_thread_flag(next_p, TIF_BLOCKSTEP))
+			debugctl |= DEBUGCTLMSR_BTF;
+
+		update_debugctlmsr(debugctl);
+	}
 
 	if (test_tsk_thread_flag(prev_p, TIF_NOTSC) ^
 	    test_tsk_thread_flag(next_p, TIF_NOTSC)) {
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index f6c62667e30..75090c589b7 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -55,7 +55,6 @@
 #include <asm/cpu.h>
 #include <asm/idle.h>
 #include <asm/syscalls.h>
-#include <asm/ds.h>
 #include <asm/debugreg.h>
 
 asmlinkage void ret_from_fork(void) __asm__("ret_from_fork");
@@ -238,13 +237,6 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 		kfree(p->thread.io_bitmap_ptr);
 		p->thread.io_bitmap_max = 0;
 	}
-
-	clear_tsk_thread_flag(p, TIF_DS_AREA_MSR);
-	p->thread.ds_ctx = NULL;
-
-	clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR);
-	p->thread.debugctlmsr = 0;
-
 	return err;
 }
 
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 17cb3295cbf..50cc84ac0a0 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -49,7 +49,6 @@
 #include <asm/ia32.h>
 #include <asm/idle.h>
 #include <asm/syscalls.h>
-#include <asm/ds.h>
 #include <asm/debugreg.h>
 
 asmlinkage extern void ret_from_fork(void);
@@ -313,13 +312,6 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 		if (err)
 			goto out;
 	}
-
-	clear_tsk_thread_flag(p, TIF_DS_AREA_MSR);
-	p->thread.ds_ctx = NULL;
-
-	clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR);
-	p->thread.debugctlmsr = 0;
-
 	err = 0;
 out:
 	if (err && p->thread.io_bitmap_ptr) {
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 2e9b55027b7..70c4872cd8a 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -2,9 +2,6 @@
 /*
  * Pentium III FXSR, SSE support
  *	Gareth Hughes <gareth@valinux.com>, May 2000
- *
- * BTS tracing
- *	Markus Metzger <markus.t.metzger@intel.com>, Dec 2007
  */
 
 #include <linux/kernel.h>
@@ -22,7 +19,6 @@
 #include <linux/audit.h>
 #include <linux/seccomp.h>
 #include <linux/signal.h>
-#include <linux/workqueue.h>
 #include <linux/perf_event.h>
 #include <linux/hw_breakpoint.h>
 
@@ -36,7 +32,6 @@
 #include <asm/desc.h>
 #include <asm/prctl.h>
 #include <asm/proto.h>
-#include <asm/ds.h>
 #include <asm/hw_breakpoint.h>
 
 #include "tls.h"
@@ -693,7 +688,7 @@ static int ptrace_set_breakpoint_addr(struct task_struct *tsk, int nr,
 	struct perf_event_attr attr;
 
 	if (!t->ptrace_bps[nr]) {
-		hw_breakpoint_init(&attr);
+		ptrace_breakpoint_init(&attr);
 		/*
 		 * Put stub len and type to register (reserve) an inactive but
 		 * correct bp
@@ -789,342 +784,6 @@ static int ioperm_get(struct task_struct *target,
 				   0, IO_BITMAP_BYTES);
 }
 
-#ifdef CONFIG_X86_PTRACE_BTS
-/*
- * A branch trace store context.
- *
- * Contexts may only be installed by ptrace_bts_config() and only for
- * ptraced tasks.
- *
- * Contexts are destroyed when the tracee is detached from the tracer.
- * The actual destruction work requires interrupts enabled, so the
- * work is deferred and will be scheduled during __ptrace_unlink().
- *
- * Contexts hold an additional task_struct reference on the traced
- * task, as well as a reference on the tracer's mm.
- *
- * Ptrace already holds a task_struct for the duration of ptrace operations,
- * but since destruction is deferred, it may be executed after both
- * tracer and tracee exited.
- */
-struct bts_context {
-	/* The branch trace handle. */
-	struct bts_tracer	*tracer;
-
-	/* The buffer used to store the branch trace and its size. */
-	void			*buffer;
-	unsigned int		size;
-
-	/* The mm that paid for the above buffer. */
-	struct mm_struct	*mm;
-
-	/* The task this context belongs to. */
-	struct task_struct	*task;
-
-	/* The signal to send on a bts buffer overflow. */
-	unsigned int		bts_ovfl_signal;
-
-	/* The work struct to destroy a context. */
-	struct work_struct	work;
-};
-
-static int alloc_bts_buffer(struct bts_context *context, unsigned int size)
-{
-	void *buffer = NULL;
-	int err = -ENOMEM;
-
-	err = account_locked_memory(current->mm, current->signal->rlim, size);
-	if (err < 0)
-		return err;
-
-	buffer = kzalloc(size, GFP_KERNEL);
-	if (!buffer)
-		goto out_refund;
-
-	context->buffer = buffer;
-	context->size = size;
-	context->mm = get_task_mm(current);
-
-	return 0;
-
- out_refund:
-	refund_locked_memory(current->mm, size);
-	return err;
-}
-
-static inline void free_bts_buffer(struct bts_context *context)
-{
-	if (!context->buffer)
-		return;
-
-	kfree(context->buffer);
-	context->buffer = NULL;
-
-	refund_locked_memory(context->mm, context->size);
-	context->size = 0;
-
-	mmput(context->mm);
-	context->mm = NULL;
-}
-
-static void free_bts_context_work(struct work_struct *w)
-{
-	struct bts_context *context;
-
-	context = container_of(w, struct bts_context, work);
-
-	ds_release_bts(context->tracer);
-	put_task_struct(context->task);
-	free_bts_buffer(context);
-	kfree(context);
-}
-
-static inline void free_bts_context(struct bts_context *context)
-{
-	INIT_WORK(&context->work, free_bts_context_work);
-	schedule_work(&context->work);
-}
-
-static inline struct bts_context *alloc_bts_context(struct task_struct *task)
-{
-	struct bts_context *context = kzalloc(sizeof(*context), GFP_KERNEL);
-	if (context) {
-		context->task = task;
-		task->bts = context;
-
-		get_task_struct(task);
-	}
-
-	return context;
-}
-
-static int ptrace_bts_read_record(struct task_struct *child, size_t index,
-				  struct bts_struct __user *out)
-{
-	struct bts_context *context;
-	const struct bts_trace *trace;
-	struct bts_struct bts;
-	const unsigned char *at;
-	int error;
-
-	context = child->bts;
-	if (!context)
-		return -ESRCH;
-
-	trace = ds_read_bts(context->tracer);
-	if (!trace)
-		return -ESRCH;
-
-	at = trace->ds.top - ((index + 1) * trace->ds.size);
-	if ((void *)at < trace->ds.begin)
-		at += (trace->ds.n * trace->ds.size);
-
-	if (!trace->read)
-		return -EOPNOTSUPP;
-
-	error = trace->read(context->tracer, at, &bts);
-	if (error < 0)
-		return error;
-
-	if (copy_to_user(out, &bts, sizeof(bts)))
-		return -EFAULT;
-
-	return sizeof(bts);
-}
-
-static int ptrace_bts_drain(struct task_struct *child,
-			    long size,
-			    struct bts_struct __user *out)
-{
-	struct bts_context *context;
-	const struct bts_trace *trace;
-	const unsigned char *at;
-	int error, drained = 0;
-
-	context = child->bts;
-	if (!context)
-		return -ESRCH;
-
-	trace = ds_read_bts(context->tracer);
-	if (!trace)
-		return -ESRCH;
-
-	if (!trace->read)
-		return -EOPNOTSUPP;
-
-	if (size < (trace->ds.top - trace->ds.begin))
-		return -EIO;
-
-	for (at = trace->ds.begin; (void *)at < trace->ds.top;
-	     out++, drained++, at += trace->ds.size) {
-		struct bts_struct bts;
-
-		error = trace->read(context->tracer, at, &bts);
-		if (error < 0)
-			return error;
-
-		if (copy_to_user(out, &bts, sizeof(bts)))
-			return -EFAULT;
-	}
-
-	memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size);
-
-	error = ds_reset_bts(context->tracer);
-	if (error < 0)
-		return error;
-
-	return drained;
-}
-
-static int ptrace_bts_config(struct task_struct *child,
-			     long cfg_size,
-			     const struct ptrace_bts_config __user *ucfg)
-{
-	struct bts_context *context;
-	struct ptrace_bts_config cfg;
-	unsigned int flags = 0;
-
-	if (cfg_size < sizeof(cfg))
-		return -EIO;
-
-	if (copy_from_user(&cfg, ucfg, sizeof(cfg)))
-		return -EFAULT;
-
-	context = child->bts;
-	if (!context)
-		context = alloc_bts_context(child);
-	if (!context)
-		return -ENOMEM;
-
-	if (cfg.flags & PTRACE_BTS_O_SIGNAL) {
-		if (!cfg.signal)
-			return -EINVAL;
-
-		return -EOPNOTSUPP;
-		context->bts_ovfl_signal = cfg.signal;
-	}
-
-	ds_release_bts(context->tracer);
-	context->tracer = NULL;
-
-	if ((cfg.flags & PTRACE_BTS_O_ALLOC) && (cfg.size != context->size)) {
-		int err;
-
-		free_bts_buffer(context);
-		if (!cfg.size)
-			return 0;
-
-		err = alloc_bts_buffer(context, cfg.size);
-		if (err < 0)
-			return err;
-	}
-
-	if (cfg.flags & PTRACE_BTS_O_TRACE)
-		flags |= BTS_USER;
-
-	if (cfg.flags & PTRACE_BTS_O_SCHED)
-		flags |= BTS_TIMESTAMPS;
-
-	context->tracer =
-		ds_request_bts_task(child, context->buffer, context->size,
-				    NULL, (size_t)-1, flags);
-	if (unlikely(IS_ERR(context->tracer))) {
-		int error = PTR_ERR(context->tracer);
-
-		free_bts_buffer(context);
-		context->tracer = NULL;
-		return error;
-	}
-
-	return sizeof(cfg);
-}
-
-static int ptrace_bts_status(struct task_struct *child,
-			     long cfg_size,
-			     struct ptrace_bts_config __user *ucfg)
-{
-	struct bts_context *context;
-	const struct bts_trace *trace;
-	struct ptrace_bts_config cfg;
-
-	context = child->bts;
-	if (!context)
-		return -ESRCH;
-
-	if (cfg_size < sizeof(cfg))
-		return -EIO;
-
-	trace = ds_read_bts(context->tracer);
-	if (!trace)
-		return -ESRCH;
-
-	memset(&cfg, 0, sizeof(cfg));
-	cfg.size	= trace->ds.end - trace->ds.begin;
-	cfg.signal	= context->bts_ovfl_signal;
-	cfg.bts_size	= sizeof(struct bts_struct);
-
-	if (cfg.signal)
-		cfg.flags |= PTRACE_BTS_O_SIGNAL;
-
-	if (trace->ds.flags & BTS_USER)
-		cfg.flags |= PTRACE_BTS_O_TRACE;
-
-	if (trace->ds.flags & BTS_TIMESTAMPS)
-		cfg.flags |= PTRACE_BTS_O_SCHED;
-
-	if (copy_to_user(ucfg, &cfg, sizeof(cfg)))
-		return -EFAULT;
-
-	return sizeof(cfg);
-}
-
-static int ptrace_bts_clear(struct task_struct *child)
-{
-	struct bts_context *context;
-	const struct bts_trace *trace;
-
-	context = child->bts;
-	if (!context)
-		return -ESRCH;
-
-	trace = ds_read_bts(context->tracer);
-	if (!trace)
-		return -ESRCH;
-
-	memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size);
-
-	return ds_reset_bts(context->tracer);
-}
-
-static int ptrace_bts_size(struct task_struct *child)
-{
-	struct bts_context *context;
-	const struct bts_trace *trace;
-
-	context = child->bts;
-	if (!context)
-		return -ESRCH;
-
-	trace = ds_read_bts(context->tracer);
-	if (!trace)
-		return -ESRCH;
-
-	return (trace->ds.top - trace->ds.begin) / trace->ds.size;
-}
-
-/*
- * Called from __ptrace_unlink() after the child has been moved back
- * to its original parent.
- */
-void ptrace_bts_untrace(struct task_struct *child)
-{
-	if (unlikely(child->bts)) {
-		free_bts_context(child->bts);
-		child->bts = NULL;
-	}
-}
-#endif /* CONFIG_X86_PTRACE_BTS */
-
 /*
  * Called by kernel/ptrace.c when detaching..
  *
@@ -1252,39 +911,6 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
 		break;
 #endif
 
-	/*
-	 * These bits need more cooking - not enabled yet:
-	 */
-#ifdef CONFIG_X86_PTRACE_BTS
-	case PTRACE_BTS_CONFIG:
-		ret = ptrace_bts_config
-			(child, data, (struct ptrace_bts_config __user *)addr);
-		break;
-
-	case PTRACE_BTS_STATUS:
-		ret = ptrace_bts_status
-			(child, data, (struct ptrace_bts_config __user *)addr);
-		break;
-
-	case PTRACE_BTS_SIZE:
-		ret = ptrace_bts_size(child);
-		break;
-
-	case PTRACE_BTS_GET:
-		ret = ptrace_bts_read_record
-			(child, data, (struct bts_struct __user *) addr);
-		break;
-
-	case PTRACE_BTS_CLEAR:
-		ret = ptrace_bts_clear(child);
-		break;
-
-	case PTRACE_BTS_DRAIN:
-		ret = ptrace_bts_drain
-			(child, data, (struct bts_struct __user *) addr);
-		break;
-#endif /* CONFIG_X86_PTRACE_BTS */
-
 	default:
 		ret = ptrace_request(child, request, addr, data);
 		break;
@@ -1544,14 +1170,6 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 
 	case PTRACE_GET_THREAD_AREA:
 	case PTRACE_SET_THREAD_AREA:
-#ifdef CONFIG_X86_PTRACE_BTS
-	case PTRACE_BTS_CONFIG:
-	case PTRACE_BTS_STATUS:
-	case PTRACE_BTS_SIZE:
-	case PTRACE_BTS_GET:
-	case PTRACE_BTS_CLEAR:
-	case PTRACE_BTS_DRAIN:
-#endif /* CONFIG_X86_PTRACE_BTS */
 		return arch_ptrace(child, request, addr, data);
 
 	default:
diff --git a/arch/x86/kernel/step.c b/arch/x86/kernel/step.c
index 3149032ff10..58de45ee08b 100644
--- a/arch/x86/kernel/step.c
+++ b/arch/x86/kernel/step.c
@@ -158,22 +158,6 @@ static int enable_single_step(struct task_struct *child)
 }
 
 /*
- * Install this value in MSR_IA32_DEBUGCTLMSR whenever child is running.
- */
-static void write_debugctlmsr(struct task_struct *child, unsigned long val)
-{
-	if (child->thread.debugctlmsr == val)
-		return;
-
-	child->thread.debugctlmsr = val;
-
-	if (child != current)
-		return;
-
-	update_debugctlmsr(val);
-}
-
-/*
  * Enable single or block step.
  */
 static void enable_step(struct task_struct *child, bool block)
@@ -186,15 +170,17 @@ static void enable_step(struct task_struct *child, bool block)
 	 * that uses user-mode single stepping itself.
 	 */
 	if (enable_single_step(child) && block) {
-		set_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
-		write_debugctlmsr(child,
-				  child->thread.debugctlmsr | DEBUGCTLMSR_BTF);
-	} else {
-		write_debugctlmsr(child,
-				  child->thread.debugctlmsr & ~DEBUGCTLMSR_BTF);
-
-		if (!child->thread.debugctlmsr)
-			clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
+		unsigned long debugctl = get_debugctlmsr();
+
+		debugctl |= DEBUGCTLMSR_BTF;
+		update_debugctlmsr(debugctl);
+		set_tsk_thread_flag(child, TIF_BLOCKSTEP);
+	} else if (test_tsk_thread_flag(child, TIF_BLOCKSTEP)) {
+		unsigned long debugctl = get_debugctlmsr();
+
+		debugctl &= ~DEBUGCTLMSR_BTF;
+		update_debugctlmsr(debugctl);
+		clear_tsk_thread_flag(child, TIF_BLOCKSTEP);
 	}
 }
 
@@ -213,11 +199,13 @@ void user_disable_single_step(struct task_struct *child)
 	/*
 	 * Make sure block stepping (BTF) is disabled.
 	 */
-	write_debugctlmsr(child,
-			  child->thread.debugctlmsr & ~DEBUGCTLMSR_BTF);
+	if (test_tsk_thread_flag(child, TIF_BLOCKSTEP)) {
+		unsigned long debugctl = get_debugctlmsr();
 
-	if (!child->thread.debugctlmsr)
-		clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR);
+		debugctl &= ~DEBUGCTLMSR_BTF;
+		update_debugctlmsr(debugctl);
+		clear_tsk_thread_flag(child, TIF_BLOCKSTEP);
+	}
 
 	/* Always clear TIF_SINGLESTEP... */
 	clear_tsk_thread_flag(child, TIF_SINGLESTEP);
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 1168e445418..36f1bd9f8e7 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -543,11 +543,11 @@ dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code)
 
 	/* DR6 may or may not be cleared by the CPU */
 	set_debugreg(0, 6);
+
 	/*
 	 * The processor cleared BTF, so don't mark that we need it set.
 	 */
-	clear_tsk_thread_flag(tsk, TIF_DEBUGCTLMSR);
-	tsk->thread.debugctlmsr = 0;
+	clear_tsk_thread_flag(tsk, TIF_BLOCKSTEP);
 
 	/* Store the virtualized DR6 value */
 	tsk->thread.debugreg6 = dr6;
author	Linus Torvalds <torvalds@linux-foundation.org>	2010-05-18 08:19:03 -0700
committer	Linus Torvalds <torvalds@linux-foundation.org>	2010-05-18 08:19:03 -0700
commit	4d7b4ac22fbec1a03206c6cde353f2fd6942f828 (patch)
tree	2d96a9e9c28cf6fa628a278decc00ad55a8b043b /arch/x86/kernel
parent	3aaf51ace5975050ab43c7d4d7e439e0ae7d13d7 (diff)
parent	94f3ca95787ada3d64339a4ecb2754236ab563f6 (diff)