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: (123 commits)
  perf: Remove the nmi parameter from the oprofile_perf backend
  x86, perf: Make copy_from_user_nmi() a library function
  perf: Remove perf_event_attr::type check
  x86, perf: P4 PMU - Fix typos in comments and style cleanup
  perf tools: Make test use the preset debugfs path
  perf tools: Add automated tests for events parsing
  perf tools: De-opt the parse_events function
  perf script: Fix display of IP address for non-callchain path
  perf tools: Fix endian conversion reading event attr from file header
  perf tools: Add missing 'node' alias to the hw_cache[] array
  perf probe: Support adding probes on offline kernel modules
  perf probe: Add probed module in front of function
  perf probe: Introduce debuginfo to encapsulate dwarf information
  perf-probe: Move dwarf library routines to dwarf-aux.{c, h}
  perf probe: Remove redundant dwarf functions
  perf probe: Move strtailcmp to string.c
  perf probe: Rename DIE_FIND_CB_FOUND to DIE_FIND_CB_END
  tracing/kprobe: Update symbol reference when loading module
  tracing/kprobes: Support module init function probing
  kprobes: Return -ENOENT if probe point doesn't exist
  ...

19 files changed, 633 insertions, 300 deletions

diff --git a/arch/x86/include/asm/irqflags.h b/arch/x86/include/asm/irqflags.h
index 5745ce8bf10..bba3cf88e62 100644
--- a/arch/x86/include/asm/irqflags.h
+++ b/arch/x86/include/asm/irqflags.h
@@ -60,23 +60,24 @@ static inline void native_halt(void)
 #include <asm/paravirt.h>
 #else
 #ifndef __ASSEMBLY__
+#include <linux/types.h>
 
-static inline unsigned long arch_local_save_flags(void)
+static inline notrace unsigned long arch_local_save_flags(void)
 {
 	return native_save_fl();
 }
 
-static inline void arch_local_irq_restore(unsigned long flags)
+static inline notrace void arch_local_irq_restore(unsigned long flags)
 {
 	native_restore_fl(flags);
 }
 
-static inline void arch_local_irq_disable(void)
+static inline notrace void arch_local_irq_disable(void)
 {
 	native_irq_disable();
 }
 
-static inline void arch_local_irq_enable(void)
+static inline notrace void arch_local_irq_enable(void)
 {
 	native_irq_enable();
 }
@@ -102,7 +103,7 @@ static inline void halt(void)
 /*
  * For spinlocks, etc:
  */
-static inline unsigned long arch_local_irq_save(void)
+static inline notrace unsigned long arch_local_irq_save(void)
 {
 	unsigned long flags = arch_local_save_flags();
 	arch_local_irq_disable();
diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h
index d9d4dae305f..094fb30817a 100644
--- a/arch/x86/include/asm/perf_event.h
+++ b/arch/x86/include/asm/perf_event.h
@@ -152,6 +152,11 @@ extern unsigned long perf_misc_flags(struct pt_regs *regs);
 	(regs)->bp = caller_frame_pointer();			\
 	(regs)->cs = __KERNEL_CS;				\
 	regs->flags = 0;					\
+	asm volatile(						\
+		_ASM_MOV "%%"_ASM_SP ", %0\n"			\
+		: "=m" ((regs)->sp)				\
+		:: "memory"					\
+	);							\
 }
 
 #else
diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h
index 56fd9e3abbd..4f7e67e2345 100644
--- a/arch/x86/include/asm/perf_event_p4.h
+++ b/arch/x86/include/asm/perf_event_p4.h
@@ -102,6 +102,14 @@
 #define P4_CONFIG_HT			(1ULL << P4_CONFIG_HT_SHIFT)
 
 /*
+ * If an event has alias it should be marked
+ * with a special bit. (Don't forget to check
+ * P4_PEBS_CONFIG_MASK and related bits on
+ * modification.)
+ */
+#define P4_CONFIG_ALIASABLE		(1 << 9)
+
+/*
  * The bits we allow to pass for RAW events
  */
 #define P4_CONFIG_MASK_ESCR		\
@@ -123,6 +131,31 @@
 	(p4_config_pack_escr(P4_CONFIG_MASK_ESCR))	| \
 	(p4_config_pack_cccr(P4_CONFIG_MASK_CCCR))
 
+/*
+ * In case of event aliasing we need to preserve some
+ * caller bits, otherwise the mapping won't be complete.
+ */
+#define P4_CONFIG_EVENT_ALIAS_MASK			  \
+	(p4_config_pack_escr(P4_CONFIG_MASK_ESCR)	| \
+	 p4_config_pack_cccr(P4_CCCR_EDGE		| \
+			     P4_CCCR_THRESHOLD_MASK	| \
+			     P4_CCCR_COMPLEMENT		| \
+			     P4_CCCR_COMPARE))
+
+#define  P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS		  \
+	((P4_CONFIG_HT)					| \
+	 p4_config_pack_escr(P4_ESCR_T0_OS		| \
+			     P4_ESCR_T0_USR		| \
+			     P4_ESCR_T1_OS		| \
+			     P4_ESCR_T1_USR)		| \
+	 p4_config_pack_cccr(P4_CCCR_OVF		| \
+			     P4_CCCR_CASCADE		| \
+			     P4_CCCR_FORCE_OVF		| \
+			     P4_CCCR_THREAD_ANY		| \
+			     P4_CCCR_OVF_PMI_T0		| \
+			     P4_CCCR_OVF_PMI_T1		| \
+			     P4_CONFIG_ALIASABLE))
+
 static inline bool p4_is_event_cascaded(u64 config)
 {
 	u32 cccr = p4_config_unpack_cccr(config);
diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h
index 99ddd148a76..36361bf6fdd 100644
--- a/arch/x86/include/asm/uaccess.h
+++ b/arch/x86/include/asm/uaccess.h
@@ -555,6 +555,9 @@ struct __large_struct { unsigned long buf[100]; };
 
 #endif /* CONFIG_X86_WP_WORKS_OK */
 
+extern unsigned long
+copy_from_user_nmi(void *to, const void __user *from, unsigned long n);
+
 /*
  * movsl can be slow when source and dest are not both 8-byte aligned
  */
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 3a0338b4b17..4ee3abf20ed 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -22,7 +22,6 @@
 #include <linux/sched.h>
 #include <linux/uaccess.h>
 #include <linux/slab.h>
-#include <linux/highmem.h>
 #include <linux/cpu.h>
 #include <linux/bitops.h>
 
@@ -45,38 +44,27 @@ do {								\
 #endif
 
 /*
- * best effort, GUP based copy_from_user() that assumes IRQ or NMI context
+ *          |   NHM/WSM    |      SNB     |
+ * register -------------------------------
+ *          |  HT  | no HT |  HT  | no HT |
+ *-----------------------------------------
+ * offcore  | core | core  | cpu  | core  |
+ * lbr_sel  | core | core  | cpu  | core  |
+ * ld_lat   | cpu  | core  | cpu  | core  |
+ *-----------------------------------------
+ *
+ * Given that there is a small number of shared regs,
+ * we can pre-allocate their slot in the per-cpu
+ * per-core reg tables.
  */
-static unsigned long
-copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
-{
-	unsigned long offset, addr = (unsigned long)from;
-	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);
-		memcpy(to, map+offset, size);
-		kunmap_atomic(map);
-		put_page(page);
+enum extra_reg_type {
+	EXTRA_REG_NONE  = -1,	/* not used */
 
-		len  += size;
-		to   += size;
-		addr += size;
+	EXTRA_REG_RSP_0 = 0,	/* offcore_response_0 */
+	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */
 
-	} while (len < n);
-
-	return len;
-}
+	EXTRA_REG_MAX		/* number of entries needed */
+};
 
 struct event_constraint {
 	union {
@@ -132,11 +120,10 @@ struct cpu_hw_events {
 	struct perf_branch_entry	lbr_entries[MAX_LBR_ENTRIES];
 
 	/*
-	 * Intel percore register state.
-	 * Coordinate shared resources between HT threads.
+	 * manage shared (per-core, per-cpu) registers
+	 * used on Intel NHM/WSM/SNB
 	 */
-	int				percore_used; /* Used by this CPU? */
-	struct intel_percore		*per_core;
+	struct intel_shared_regs	*shared_regs;
 
 	/*
 	 * AMD specific bits
@@ -187,26 +174,45 @@ struct cpu_hw_events {
 	for ((e) = (c); (e)->weight; (e)++)
 
 /*
+ * Per register state.
+ */
+struct er_account {
+	raw_spinlock_t		lock;	/* per-core: protect structure */
+	u64			config;	/* extra MSR config */
+	u64			reg;	/* extra MSR number */
+	atomic_t		ref;	/* reference count */
+};
+
+/*
  * Extra registers for specific events.
+ *
  * Some events need large masks and require external MSRs.
- * Define a mapping to these extra registers.
+ * Those extra MSRs end up being shared for all events on
+ * a PMU and sometimes between PMU of sibling HT threads.
+ * In either case, the kernel needs to handle conflicting
+ * accesses to those extra, shared, regs. The data structure
+ * to manage those registers is stored in cpu_hw_event.
  */
 struct extra_reg {
 	unsigned int		event;
 	unsigned int		msr;
 	u64			config_mask;
 	u64			valid_mask;
+	int			idx;  /* per_xxx->regs[] reg index */
 };
 
-#define EVENT_EXTRA_REG(e, ms, m, vm) {	\
+#define EVENT_EXTRA_REG(e, ms, m, vm, i) {	\
 	.event = (e),		\
 	.msr = (ms),		\
 	.config_mask = (m),	\
 	.valid_mask = (vm),	\
+	.idx = EXTRA_REG_##i	\
 	}
-#define INTEL_EVENT_EXTRA_REG(event, msr, vm)	\
-	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm)
-#define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0)
+
+#define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\
+	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx)
+
+#define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0)
 
 union perf_capabilities {
 	struct {
@@ -252,7 +258,6 @@ struct x86_pmu {
 	void		(*put_event_constraints)(struct cpu_hw_events *cpuc,
 						 struct perf_event *event);
 	struct event_constraint *event_constraints;
-	struct event_constraint *percore_constraints;
 	void		(*quirks)(void);
 	int		perfctr_second_write;
 
@@ -286,8 +291,12 @@ struct x86_pmu {
 	 * Extra registers for events
 	 */
 	struct extra_reg *extra_regs;
+	unsigned int er_flags;
 };
 
+#define ERF_NO_HT_SHARING	1
+#define ERF_HAS_RSP_1		2
+
 static struct x86_pmu x86_pmu __read_mostly;
 
 static DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = {
@@ -393,10 +402,10 @@ static inline unsigned int x86_pmu_event_addr(int index)
  */
 static int x86_pmu_extra_regs(u64 config, struct perf_event *event)
 {
+	struct hw_perf_event_extra *reg;
 	struct extra_reg *er;
 
-	event->hw.extra_reg = 0;
-	event->hw.extra_config = 0;
+	reg = &event->hw.extra_reg;
 
 	if (!x86_pmu.extra_regs)
 		return 0;
@@ -406,8 +415,10 @@ static int x86_pmu_extra_regs(u64 config, struct perf_event *event)
 			continue;
 		if (event->attr.config1 & ~er->valid_mask)
 			return -EINVAL;
-		event->hw.extra_reg = er->msr;
-		event->hw.extra_config = event->attr.config1;
+
+		reg->idx = er->idx;
+		reg->config = event->attr.config1;
+		reg->reg = er->msr;
 		break;
 	}
 	return 0;
@@ -706,6 +717,9 @@ static int __x86_pmu_event_init(struct perf_event *event)
 	event->hw.last_cpu = -1;
 	event->hw.last_tag = ~0ULL;
 
+	/* mark unused */
+	event->hw.extra_reg.idx = EXTRA_REG_NONE;
+
 	return x86_pmu.hw_config(event);
 }
 
@@ -747,8 +761,8 @@ static void x86_pmu_disable(struct pmu *pmu)
 static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc,
 					  u64 enable_mask)
 {
-	if (hwc->extra_reg)
-		wrmsrl(hwc->extra_reg, hwc->extra_config);
+	if (hwc->extra_reg.reg)
+		wrmsrl(hwc->extra_reg.reg, hwc->extra_reg.config);
 	wrmsrl(hwc->config_base, hwc->config | enable_mask);
 }
 
@@ -1332,7 +1346,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
 		if (!x86_perf_event_set_period(event))
 			continue;
 
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
@@ -1637,6 +1651,40 @@ static int x86_pmu_commit_txn(struct pmu *pmu)
 	perf_pmu_enable(pmu);
 	return 0;
 }
+/*
+ * a fake_cpuc is used to validate event groups. Due to
+ * the extra reg logic, we need to also allocate a fake
+ * per_core and per_cpu structure. Otherwise, group events
+ * using extra reg may conflict without the kernel being
+ * able to catch this when the last event gets added to
+ * the group.
+ */
+static void free_fake_cpuc(struct cpu_hw_events *cpuc)
+{
+	kfree(cpuc->shared_regs);
+	kfree(cpuc);
+}
+
+static struct cpu_hw_events *allocate_fake_cpuc(void)
+{
+	struct cpu_hw_events *cpuc;
+	int cpu = raw_smp_processor_id();
+
+	cpuc = kzalloc(sizeof(*cpuc), GFP_KERNEL);
+	if (!cpuc)
+		return ERR_PTR(-ENOMEM);
+
+	/* only needed, if we have extra_regs */
+	if (x86_pmu.extra_regs) {
+		cpuc->shared_regs = allocate_shared_regs(cpu);
+		if (!cpuc->shared_regs)
+			goto error;
+	}
+	return cpuc;
+error:
+	free_fake_cpuc(cpuc);
+	return ERR_PTR(-ENOMEM);
+}
 
 /*
  * validate that we can schedule this event
@@ -1647,9 +1695,9 @@ static int validate_event(struct perf_event *event)
 	struct event_constraint *c;
 	int ret = 0;
 
-	fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO);
-	if (!fake_cpuc)
-		return -ENOMEM;
+	fake_cpuc = allocate_fake_cpuc();
+	if (IS_ERR(fake_cpuc))
+		return PTR_ERR(fake_cpuc);
 
 	c = x86_pmu.get_event_constraints(fake_cpuc, event);
 
@@ -1659,7 +1707,7 @@ static int validate_event(struct perf_event *event)
 	if (x86_pmu.put_event_constraints)
 		x86_pmu.put_event_constraints(fake_cpuc, event);
 
-	kfree(fake_cpuc);
+	free_fake_cpuc(fake_cpuc);
 
 	return ret;
 }
@@ -1679,36 +1727,32 @@ static int validate_group(struct perf_event *event)
 {
 	struct perf_event *leader = event->group_leader;
 	struct cpu_hw_events *fake_cpuc;
-	int ret, n;
-
-	ret = -ENOMEM;
-	fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO);
-	if (!fake_cpuc)
-		goto out;
+	int ret = -ENOSPC, n;
 
+	fake_cpuc = allocate_fake_cpuc();
+	if (IS_ERR(fake_cpuc))
+		return PTR_ERR(fake_cpuc);
 	/*
 	 * the event is not yet connected with its
 	 * siblings therefore we must first collect
 	 * existing siblings, then add the new event
 	 * before we can simulate the scheduling
 	 */
-	ret = -ENOSPC;
 	n = collect_events(fake_cpuc, leader, true);
 	if (n < 0)
-		goto out_free;
+		goto out;
 
 	fake_cpuc->n_events = n;
 	n = collect_events(fake_cpuc, event, false);
 	if (n < 0)
-		goto out_free;
+		goto out;
 
 	fake_cpuc->n_events = n;
 
 	ret = x86_pmu.schedule_events(fake_cpuc, n, NULL);
 
-out_free:
-	kfree(fake_cpuc);
 out:
+	free_fake_cpuc(fake_cpuc);
 	return ret;
 }
 
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index fe29c1d2219..941caa2e449 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -89,6 +89,20 @@ static __initconst const u64 amd_hw_cache_event_ids
 		[ C(RESULT_MISS)   ] = -1,
 	},
  },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0xb8e9, /* CPU Request to Memory, l+r */
+		[ C(RESULT_MISS)   ] = 0x98e9, /* CPU Request to Memory, r   */
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
 };
 
 /*
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 41178c826c4..45fbb8f7f54 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1,25 +1,15 @@
 #ifdef CONFIG_CPU_SUP_INTEL
 
-#define MAX_EXTRA_REGS 2
-
-/*
- * Per register state.
- */
-struct er_account {
-	int			ref;		/* reference count */
-	unsigned int		extra_reg;	/* extra MSR number */
-	u64			extra_config;	/* extra MSR config */
-};
-
 /*
- * Per core state
- * This used to coordinate shared registers for HT threads.
+ * Per core/cpu state
+ *
+ * Used to coordinate shared registers between HT threads or
+ * among events on a single PMU.
  */
-struct intel_percore {
-	raw_spinlock_t		lock;		/* protect structure */
-	struct er_account	regs[MAX_EXTRA_REGS];
-	int			refcnt;		/* number of threads */
-	unsigned		core_id;
+struct intel_shared_regs {
+	struct er_account       regs[EXTRA_REG_MAX];
+	int                     refcnt;		/* per-core: #HT threads */
+	unsigned                core_id;	/* per-core: core id */
 };
 
 /*
@@ -88,16 +78,10 @@ static struct event_constraint intel_nehalem_event_constraints[] __read_mostly =
 
 static struct extra_reg intel_nehalem_extra_regs[] __read_mostly =
 {
-	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff),
+	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
 	EVENT_EXTRA_END
 };
 
-static struct event_constraint intel_nehalem_percore_constraints[] __read_mostly =
-{
-	INTEL_EVENT_CONSTRAINT(0xb7, 0),
-	EVENT_CONSTRAINT_END
-};
-
 static struct event_constraint intel_westmere_event_constraints[] __read_mostly =
 {
 	FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */
@@ -116,8 +100,6 @@ static struct event_constraint intel_snb_event_constraints[] __read_mostly =
 	FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */
 	/* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */
 	INTEL_EVENT_CONSTRAINT(0x48, 0x4), /* L1D_PEND_MISS.PENDING */
-	INTEL_EVENT_CONSTRAINT(0xb7, 0x1), /* OFF_CORE_RESPONSE_0 */
-	INTEL_EVENT_CONSTRAINT(0xbb, 0x8), /* OFF_CORE_RESPONSE_1 */
 	INTEL_UEVENT_CONSTRAINT(0x01c0, 0x2), /* INST_RETIRED.PREC_DIST */
 	INTEL_EVENT_CONSTRAINT(0xcd, 0x8), /* MEM_TRANS_RETIRED.LOAD_LATENCY */
 	EVENT_CONSTRAINT_END
@@ -125,15 +107,13 @@ static struct event_constraint intel_snb_event_constraints[] __read_mostly =
 
 static struct extra_reg intel_westmere_extra_regs[] __read_mostly =
 {
-	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff),
-	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff),
+	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0xffff, RSP_0),
+	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0xffff, RSP_1),
 	EVENT_EXTRA_END
 };
 
-static struct event_constraint intel_westmere_percore_constraints[] __read_mostly =
+static struct event_constraint intel_v1_event_constraints[] __read_mostly =
 {
-	INTEL_EVENT_CONSTRAINT(0xb7, 0),
-	INTEL_EVENT_CONSTRAINT(0xbb, 0),
 	EVENT_CONSTRAINT_END
 };
 
@@ -145,6 +125,12 @@ static struct event_constraint intel_gen_event_constraints[] __read_mostly =
 	EVENT_CONSTRAINT_END
 };
 
+static struct extra_reg intel_snb_extra_regs[] __read_mostly = {
+	INTEL_EVENT_EXTRA_REG(0xb7, MSR_OFFCORE_RSP_0, 0x3fffffffffull, RSP_0),
+	INTEL_EVENT_EXTRA_REG(0xbb, MSR_OFFCORE_RSP_1, 0x3fffffffffull, RSP_1),
+	EVENT_EXTRA_END
+};
+
 static u64 intel_pmu_event_map(int hw_event)
 {
 	return intel_perfmon_event_map[hw_event];
@@ -245,6 +231,21 @@ static __initconst const u64 snb_hw_cache_event_ids
 		[ C(RESULT_MISS)   ] = -1,
 	},
  },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
+
 };
 
 static __initconst const u64 westmere_hw_cache_event_ids
@@ -346,6 +347,20 @@ static __initconst const u64 westmere_hw_cache_event_ids
 		[ C(RESULT_MISS)   ] = -1,
 	},
  },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+ },
 };
 
 /*
@@ -398,7 +413,21 @@ static __initconst const u64 nehalem_hw_cache_extra_regs
 		[ C(RESULT_ACCESS) ] = NHM_DMND_PREFETCH|NHM_L3_ACCESS,
 		[ C(RESULT_MISS)   ] = NHM_DMND_PREFETCH|NHM_L3_MISS,
 	},
- }
+ },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = NHM_DMND_READ|NHM_ALL_DRAM,
+		[ C(RESULT_MISS)   ] = NHM_DMND_READ|NHM_REMOTE_DRAM,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = NHM_DMND_WRITE|NHM_ALL_DRAM,
+		[ C(RESULT_MISS)   ] = NHM_DMND_WRITE|NHM_REMOTE_DRAM,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = NHM_DMND_PREFETCH|NHM_ALL_DRAM,
+		[ C(RESULT_MISS)   ] = NHM_DMND_PREFETCH|NHM_REMOTE_DRAM,
+	},
+ },
 };
 
 static __initconst const u64 nehalem_hw_cache_event_ids
@@ -500,6 +529,20 @@ static __initconst const u64 nehalem_hw_cache_event_ids
 		[ C(RESULT_MISS)   ] = -1,
 	},
  },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = 0x01b7,
+		[ C(RESULT_MISS)   ] = 0x01b7,
+	},
+ },
 };
 
 static __initconst const u64 core2_hw_cache_event_ids
@@ -1003,7 +1046,7 @@ again:
 
 		data.period = event->hw.last_period;
 
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
@@ -1037,65 +1080,121 @@ intel_bts_constraints(struct perf_event *event)
 	return NULL;
 }
 
+static bool intel_try_alt_er(struct perf_event *event, int orig_idx)
+{
+	if (!(x86_pmu.er_flags & ERF_HAS_RSP_1))
+		return false;
+
+	if (event->hw.extra_reg.idx == EXTRA_REG_RSP_0) {
+		event->hw.config &= ~INTEL_ARCH_EVENT_MASK;
+		event->hw.config |= 0x01bb;
+		event->hw.extra_reg.idx = EXTRA_REG_RSP_1;
+		event->hw.extra_reg.reg = MSR_OFFCORE_RSP_1;
+	} else if (event->hw.extra_reg.idx == EXTRA_REG_RSP_1) {
+		event->hw.config &= ~INTEL_ARCH_EVENT_MASK;
+		event->hw.config |= 0x01b7;
+		event->hw.extra_reg.idx = EXTRA_REG_RSP_0;
+		event->hw.extra_reg.reg = MSR_OFFCORE_RSP_0;
+	}
+
+	if (event->hw.extra_reg.idx == orig_idx)
+		return false;
+
+	return true;
+}
+
+/*
+ * manage allocation of shared extra msr for certain events
+ *
+ * sharing can be:
+ * per-cpu: to be shared between the various events on a single PMU
+ * per-core: per-cpu + shared by HT threads
+ */
 static struct event_constraint *
-intel_percore_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
+__intel_shared_reg_get_constraints(struct cpu_hw_events *cpuc,
+				   struct perf_event *event)
 {
-	struct hw_perf_event *hwc = &event->hw;
-	unsigned int e = hwc->config & ARCH_PERFMON_EVENTSEL_EVENT;
-	struct event_constraint *c;
-	struct intel_percore *pc;
+	struct event_constraint *c = &emptyconstraint;
+	struct hw_perf_event_extra *reg = &event->hw.extra_reg;
 	struct er_account *era;
-	int i;
-	int free_slot;
-	int found;
+	unsigned long flags;
+	int orig_idx = reg->idx;
 
-	if (!x86_pmu.percore_constraints || hwc->extra_alloc)
-		return NULL;
+	/* already allocated shared msr */
+	if (reg->alloc)
+		return &unconstrained;
 
-	for (c = x86_pmu.percore_constraints; c->cmask; c++) {
-		if (e != c->code)
-			continue;
+again:
+	era = &cpuc->shared_regs->regs[reg->idx];
+	/*
+	 * we use spin_lock_irqsave() to avoid lockdep issues when
+	 * passing a fake cpuc
+	 */
+	raw_spin_lock_irqsave(&era->lock, flags);
+
+	if (!atomic_read(&era->ref) || era->config == reg->config) {
+
+		/* lock in msr value */
+		era->config = reg->config;
+		era->reg = reg->reg;
+
+		/* one more user */
+		atomic_inc(&era->ref);
+
+		/* no need to reallocate during incremental event scheduling */
+		reg->alloc = 1;
 
 		/*
-		 * Allocate resource per core.
+		 * All events using extra_reg are unconstrained.
+		 * Avoids calling x86_get_event_constraints()
+		 *
+		 * Must revisit if extra_reg controlling events
+		 * ever have constraints. Worst case we go through
+		 * the regular event constraint table.
 		 */
-		pc = cpuc->per_core;
-		if (!pc)
-			break;
-		c = &emptyconstraint;
-		raw_spin_lock(&pc->lock);
-		free_slot = -1;
-		found = 0;
-		for (i = 0; i < MAX_EXTRA_REGS; i++) {
-			era = &pc->regs[i];
-			if (era->ref > 0 && hwc->extra_reg == era->extra_reg) {
-				/* Allow sharing same config */
-				if (hwc->extra_config == era->extra_config) {
-					era->ref++;
-					cpuc->percore_used = 1;
-					hwc->extra_alloc = 1;
-					c = NULL;
-				}
-				/* else conflict */
-				found = 1;
-				break;
-			} else if (era->ref == 0 && free_slot == -1)
-				free_slot = i;
-		}
-		if (!found && free_slot != -1) {
-			era = &pc->regs[free_slot];
-			era->ref = 1;
-			era->extra_reg = hwc->extra_reg;
-			era->extra_config = hwc->extra_config;
-			cpuc->percore_used = 1;
-			hwc->extra_alloc = 1;
-			c = NULL;
-		}
-		raw_spin_unlock(&pc->lock);
-		return c;
+		c = &unconstrained;
+	} else if (intel_try_alt_er(event, orig_idx)) {
+		raw_spin_unlock(&era->lock);
+		goto again;
 	}
+	raw_spin_unlock_irqrestore(&era->lock, flags);
 
-	return NULL;
+	return c;
+}
+
+static void
+__intel_shared_reg_put_constraints(struct cpu_hw_events *cpuc,
+				   struct hw_perf_event_extra *reg)
+{
+	struct er_account *era;
+
+	/*
+	 * only put constraint if extra reg was actually
+	 * allocated. Also takes care of event which do
+	 * not use an extra shared reg
+	 */
+	if (!reg->alloc)
+		return;
+
+	era = &cpuc->shared_regs->regs[reg->idx];
+
+	/* one fewer user */
+	atomic_dec(&era->ref);
+
+	/* allocate again next time */
+	reg->alloc = 0;
+}
+
+static struct event_constraint *
+intel_shared_regs_constraints(struct cpu_hw_events *cpuc,
+			      struct perf_event *event)
+{
+	struct event_constraint *c = NULL;
+
+	if (event->hw.extra_reg.idx != EXTRA_REG_NONE)
+		c = __intel_shared_reg_get_constraints(cpuc, event);
+
+	return c;
 }
 
 static struct event_constraint *
@@ -1111,49 +1210,28 @@ intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event
 	if (c)
 		return c;
 
-	c = intel_percore_constraints(cpuc, event);
+	c = intel_shared_regs_constraints(cpuc, event);
 	if (c)
 		return c;
 
 	return x86_get_event_constraints(cpuc, event);
 }
 
-static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
+static void
+intel_put_shared_regs_event_constraints(struct cpu_hw_events *cpuc,
 					struct perf_event *event)
 {
-	struct extra_reg *er;
-	struct intel_percore *pc;
-	struct er_account *era;
-	struct hw_perf_event *hwc = &event->hw;
-	int i, allref;
+	struct hw_perf_event_extra *reg;
 
-	if (!cpuc->percore_used)
-		return;
-
-	for (er = x86_pmu.extra_regs; er->msr; er++) {
-		if (er->event != (hwc->config & er->config_mask))
-			continue;
+	reg = &event->hw.extra_reg;
+	if (reg->idx != EXTRA_REG_NONE)
+		__intel_shared_reg_put_constraints(cpuc, reg);
+}
 
-		pc = cpuc->per_core;
-		raw_spin_lock(&pc->lock);
-		for (i = 0; i < MAX_EXTRA_REGS; i++) {
-			era = &pc->regs[i];
-			if (era->ref > 0 &&
-			    era->extra_config == hwc->extra_config &&
-			    era->extra_reg == er->msr) {
-				era->ref--;
-				hwc->extra_alloc = 0;
-				break;
-			}
-		}
-		allref = 0;
-		for (i = 0; i < MAX_EXTRA_REGS; i++)
-			allref += pc->regs[i].ref;
-		if (allref == 0)
-			cpuc->percore_used = 0;
-		raw_spin_unlock(&pc->lock);
-		break;
-	}
+static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
+					struct perf_event *event)
+{
+	intel_put_shared_regs_event_constraints(cpuc, event);
 }
 
 static int intel_pmu_hw_config(struct perf_event *event)
@@ -1231,20 +1309,36 @@ static __initconst const struct x86_pmu core_pmu = {
 	.event_constraints	= intel_core_event_constraints,
 };
 
+static struct intel_shared_regs *allocate_shared_regs(int cpu)
+{
+	struct intel_shared_regs *regs;
+	int i;
+
+	regs = kzalloc_node(sizeof(struct intel_shared_regs),
+			    GFP_KERNEL, cpu_to_node(cpu));
+	if (regs) {
+		/*
+		 * initialize the locks to keep lockdep happy
+		 */
+		for (i = 0; i < EXTRA_REG_MAX; i++)
+			raw_spin_lock_init(&regs->regs[i].lock);
+
+		regs->core_id = -1;
+	}
+	return regs;
+}
+
 static int intel_pmu_cpu_prepare(int cpu)
 {
 	struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
 
-	if (!cpu_has_ht_siblings())
+	if (!x86_pmu.extra_regs)
 		return NOTIFY_OK;
 
-	cpuc->per_core = kzalloc_node(sizeof(struct intel_percore),
-				      GFP_KERNEL, cpu_to_node(cpu));
-	if (!cpuc->per_core)
+	cpuc->shared_regs = allocate_shared_regs(cpu);
+	if (!cpuc->shared_regs)
 		return NOTIFY_BAD;
 
-	raw_spin_lock_init(&cpuc->per_core->lock);
-	cpuc->per_core->core_id = -1;
 	return NOTIFY_OK;
 }
 
@@ -1260,32 +1354,34 @@ static void intel_pmu_cpu_starting(int cpu)
 	 */
 	intel_pmu_lbr_reset();
 
-	if (!cpu_has_ht_siblings())
+	if (!cpuc->shared_regs || (x86_pmu.er_flags & ERF_NO_HT_SHARING))
 		return;
 
 	for_each_cpu(i, topology_thread_cpumask(cpu)) {
-		struct intel_percore *pc = per_cpu(cpu_hw_events, i).per_core;
+		struct intel_shared_regs *pc;
 
+		pc = per_cpu(cpu_hw_events, i).shared_regs;
 		if (pc && pc->core_id == core_id) {
-			kfree(cpuc->per_core);
-			cpuc->per_core = pc;
+			kfree(cpuc->shared_regs);
+			cpuc->shared_regs = pc;
 			break;
 		}
 	}
 
-	cpuc->per_core->core_id = core_id;
-	cpuc->per_core->refcnt++;
+	cpuc->shared_regs->core_id = core_id;
+	cpuc->shared_regs->refcnt++;
 }
 
 static void intel_pmu_cpu_dying(int cpu)
 {
 	struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
-	struct intel_percore *pc = cpuc->per_core;
+	struct intel_shared_regs *pc;
 
+	pc = cpuc->shared_regs;
 	if (pc) {
 		if (pc->core_id == -1 || --pc->refcnt == 0)
 			kfree(pc);
-		cpuc->per_core = NULL;
+		cpuc->shared_regs = NULL;
 	}
 
 	fini_debug_store_on_cpu(cpu);
@@ -1436,7 +1532,6 @@ static __init int intel_pmu_init(void)
 
 		x86_pmu.event_constraints = intel_nehalem_event_constraints;
 		x86_pmu.pebs_constraints = intel_nehalem_pebs_event_constraints;
-		x86_pmu.percore_constraints = intel_nehalem_percore_constraints;
 		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		x86_pmu.extra_regs = intel_nehalem_extra_regs;
 
@@ -1481,10 +1576,10 @@ static __init int intel_pmu_init(void)
 		intel_pmu_lbr_init_nhm();
 
 		x86_pmu.event_constraints = intel_westmere_event_constraints;
-		x86_pmu.percore_constraints = intel_westmere_percore_constraints;
 		x86_pmu.enable_all = intel_pmu_nhm_enable_all;
 		x86_pmu.pebs_constraints = intel_westmere_pebs_event_constraints;
 		x86_pmu.extra_regs = intel_westmere_extra_regs;
+		x86_pmu.er_flags |= ERF_HAS_RSP_1;
 
 		/* UOPS_ISSUED.STALLED_CYCLES */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = 0x180010e;
@@ -1502,6 +1597,10 @@ static __init int intel_pmu_init(void)
 
 		x86_pmu.event_constraints = intel_snb_event_constraints;
 		x86_pmu.pebs_constraints = intel_snb_pebs_events;
+		x86_pmu.extra_regs = intel_snb_extra_regs;
+		/* all extra regs are per-cpu when HT is on */
+		x86_pmu.er_flags |= ERF_HAS_RSP_1;
+		x86_pmu.er_flags |= ERF_NO_HT_SHARING;
 
 		/* UOPS_ISSUED.ANY,c=1,i=1 to count stall cycles */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = 0x180010e;
@@ -1512,11 +1611,19 @@ static __init int intel_pmu_init(void)
 		break;
 
 	default:
-		/*
-		 * default constraints for v2 and up
-		 */
-		x86_pmu.event_constraints = intel_gen_event_constraints;
-		pr_cont("generic architected perfmon, ");
+		switch (x86_pmu.version) {
+		case 1:
+			x86_pmu.event_constraints = intel_v1_event_constraints;
+			pr_cont("generic architected perfmon v1, ");
+			break;
+		default:
+			/*
+			 * default constraints for v2 and up
+			 */
+			x86_pmu.event_constraints = intel_gen_event_constraints;
+			pr_cont("generic architected perfmon, ");
+			break;
+		}
 	}
 	return 0;
 }
@@ -1528,4 +1635,8 @@ static int intel_pmu_init(void)
 	return 0;
 }
 
+static struct intel_shared_regs *allocate_shared_regs(int cpu)
+{
+	return NULL;
+}
 #endif /* CONFIG_CPU_SUP_INTEL */
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index bab491b8ee2..1b1ef3addcf 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -340,7 +340,7 @@ static int intel_pmu_drain_bts_buffer(void)
 	 */
 	perf_prepare_sample(&header, &data, event, &regs);
 
-	if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1))
+	if (perf_output_begin(&handle, event, header.size * (top - at)))
 		return 1;
 
 	for (; at < top; at++) {
@@ -616,7 +616,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 	else
 		regs.flags &= ~PERF_EFLAGS_EXACT;
 
-	if (perf_event_overflow(event, 1, &data, &regs))
+	if (perf_event_overflow(event, &data, &regs))
 		x86_pmu_stop(event, 0);
 }
 
diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c
index ead584fb6a7..7809d2bcb20 100644
--- a/arch/x86/kernel/cpu/perf_event_p4.c
+++ b/arch/x86/kernel/cpu/perf_event_p4.c
@@ -554,13 +554,102 @@ static __initconst const u64 p4_hw_cache_event_ids
 		[ C(RESULT_MISS)   ] = -1,
 	},
  },
+ [ C(NODE) ] = {
+	[ C(OP_READ) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_WRITE) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+	[ C(OP_PREFETCH) ] = {
+		[ C(RESULT_ACCESS) ] = -1,
+		[ C(RESULT_MISS)   ] = -1,
+	},
+ },
 };
 
+/*
+ * Because of Netburst being quite restricted in how many
+ * identical events may run simultaneously, we introduce event aliases,
+ * ie the different events which have the same functionality but
+ * utilize non-intersected resources (ESCR/CCCR/counter registers).
+ *
+ * This allow us to relax restrictions a bit and run two or more
+ * identical events together.
+ *
+ * Never set any custom internal bits such as P4_CONFIG_HT,
+ * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are
+ * either up to date automatically or not applicable at all.
+ */
+struct p4_event_alias {
+	u64 original;
+	u64 alternative;
+} p4_event_aliases[] = {
+	{
+		/*
+		 * Non-halted cycles can be substituted with non-sleeping cycles (see
+		 * Intel SDM Vol3b for details). We need this alias to be able
+		 * to run nmi-watchdog and 'perf top' (or any other user space tool
+		 * which is interested in running PERF_COUNT_HW_CPU_CYCLES)
+		 * simultaneously.
+		 */
+	.original	=
+		p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS)		|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)),
+	.alternative	=
+		p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_EXECUTION_EVENT)		|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3)|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2)	|
+				    P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3))|
+		p4_config_pack_cccr(P4_CCCR_THRESHOLD(15) | P4_CCCR_COMPLEMENT		|
+				    P4_CCCR_COMPARE),
+	},
+};
+
+static u64 p4_get_alias_event(u64 config)
+{
+	u64 config_match;
+	int i;
+
+	/*
+	 * Only event with special mark is allowed,
+	 * we're to be sure it didn't come as malformed
+	 * RAW event.
+	 */
+	if (!(config & P4_CONFIG_ALIASABLE))
+		return 0;
+
+	config_match = config & P4_CONFIG_EVENT_ALIAS_MASK;
+
+	for (i = 0; i < ARRAY_SIZE(p4_event_aliases); i++) {
+		if (config_match == p4_event_aliases[i].original) {
+			config_match = p4_event_aliases[i].alternative;
+			break;
+		} else if (config_match == p4_event_aliases[i].alternative) {
+			config_match = p4_event_aliases[i].original;
+			break;
+		}
+	}
+
+	if (i >= ARRAY_SIZE(p4_event_aliases))
+		return 0;
+
+	return config_match | (config & P4_CONFIG_EVENT_ALIAS_IMMUTABLE_BITS);
+}
+
 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)),
+		P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING))	|
+		P4_CONFIG_ALIASABLE,
 
   /*
    * retired instructions
@@ -945,7 +1034,7 @@ static int p4_pmu_handle_irq(struct pt_regs *regs)
 
 		if (!x86_perf_event_set_period(event))
 			continue;
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
@@ -1120,6 +1209,8 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 	struct p4_event_bind *bind;
 	unsigned int i, thread, num;
 	int cntr_idx, escr_idx;
+	u64 config_alias;
+	int pass;
 
 	bitmap_zero(used_mask, X86_PMC_IDX_MAX);
 	bitmap_zero(escr_mask, P4_ESCR_MSR_TABLE_SIZE);
@@ -1128,6 +1219,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 
 		hwc = &cpuc->event_list[i]->hw;
 		thread = p4_ht_thread(cpu);
+		pass = 0;
+
+again:
+		/*
+		 * It's possible to hit a circular lock
+		 * between original and alternative events
+		 * if both are scheduled already.
+		 */
+		if (pass > 2)
+			goto done;
+
 		bind = p4_config_get_bind(hwc->config);
 		escr_idx = p4_get_escr_idx(bind->escr_msr[thread]);
 		if (unlikely(escr_idx == -1))
@@ -1141,8 +1243,17 @@ static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign
 		}
 
 		cntr_idx = p4_next_cntr(thread, used_mask, bind);
-		if (cntr_idx == -1 || test_bit(escr_idx, escr_mask))
-			goto done;
+		if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) {
+			/*
+			 * Check whether an event alias is still available.
+			 */
+			config_alias = p4_get_alias_event(hwc->config);
+			if (!config_alias)
+				goto done;
+			hwc->config = config_alias;
+			pass++;
+			goto again;
+		}
 
 		p4_pmu_swap_config_ts(hwc, cpu);
 		if (assign)
diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c
index e71c98d3c0d..19853ad8afc 100644
--- a/arch/x86/kernel/dumpstack_64.c
+++ b/arch/x86/kernel/dumpstack_64.c
@@ -105,34 +105,6 @@ in_irq_stack(unsigned long *stack, unsigned long *irq_stack,
 }
 
 /*
- * We are returning from the irq stack and go to the previous one.
- * If the previous stack is also in the irq stack, then bp in the first
- * frame of the irq stack points to the previous, interrupted one.
- * Otherwise we have another level of indirection: We first save
- * the bp of the previous stack, then we switch the stack to the irq one
- * and save a new bp that links to the previous one.
- * (See save_args())
- */
-static inline unsigned long
-fixup_bp_irq_link(unsigned long bp, unsigned long *stack,
-		  unsigned long *irq_stack, unsigned long *irq_stack_end)
-{
-#ifdef CONFIG_FRAME_POINTER
-	struct stack_frame *frame = (struct stack_frame *)bp;
-	unsigned long next;
-
-	if (!in_irq_stack(stack, irq_stack, irq_stack_end)) {
-		if (!probe_kernel_address(&frame->next_frame, next))
-			return next;
-		else
-			WARN_ONCE(1, "Perf: bad frame pointer = %p in "
-				  "callchain\n", &frame->next_frame);
-	}
-#endif
-	return bp;
-}
-
-/*
  * x86-64 can have up to three kernel stacks:
  * process stack
  * interrupt stack
@@ -155,9 +127,12 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
 		task = current;
 
 	if (!stack) {
-		stack = &dummy;
-		if (task && task != current)
+		if (regs)
+			stack = (unsigned long *)regs->sp;
+		else if (task && task != current)
 			stack = (unsigned long *)task->thread.sp;
+		else
+			stack = &dummy;
 	}
 
 	if (!bp)
@@ -205,8 +180,6 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
 				 * pointer (index -1 to end) in the IRQ stack:
 				 */
 				stack = (unsigned long *) (irq_stack_end[-1]);
-				bp = fixup_bp_irq_link(bp, stack, irq_stack,
-						       irq_stack_end);
 				irq_stack_end = NULL;
 				ops->stack(data, "EOI");
 				continue;
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 8a445a0c989..d656f68371a 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -297,27 +297,26 @@ ENDPROC(native_usergs_sysret64)
 	.endm
 
 /* save partial stack frame */
-	.pushsection .kprobes.text, "ax"
-ENTRY(save_args)
-	XCPT_FRAME
+	.macro SAVE_ARGS_IRQ
 	cld
-	/*
-	 * start from rbp in pt_regs and jump over
-	 * return address.
-	 */
-	movq_cfi rdi, RDI+8-RBP
-	movq_cfi rsi, RSI+8-RBP
-	movq_cfi rdx, RDX+8-RBP
-	movq_cfi rcx, RCX+8-RBP
-	movq_cfi rax, RAX+8-RBP
-	movq_cfi  r8,  R8+8-RBP
-	movq_cfi  r9,  R9+8-RBP
-	movq_cfi r10, R10+8-RBP
-	movq_cfi r11, R11+8-RBP
-
-	leaq -RBP+8(%rsp),%rdi	/* arg1 for handler */
-	movq_cfi rbp, 8		/* push %rbp */
-	leaq 8(%rsp), %rbp		/* mov %rsp, %ebp */
+	/* start from rbp in pt_regs and jump over */
+	movq_cfi rdi, RDI-RBP
+	movq_cfi rsi, RSI-RBP
+	movq_cfi rdx, RDX-RBP
+	movq_cfi rcx, RCX-RBP
+	movq_cfi rax, RAX-RBP
+	movq_cfi  r8,  R8-RBP
+	movq_cfi  r9,  R9-RBP
+	movq_cfi r10, R10-RBP
+	movq_cfi r11, R11-RBP
+
+	/* Save rbp so that we can unwind from get_irq_regs() */
+	movq_cfi rbp, 0
+
+	/* Save previous stack value */
+	movq %rsp, %rsi
+
+	leaq -RBP(%rsp),%rdi	/* arg1 for handler */
 	testl $3, CS(%rdi)
 	je 1f
 	SWAPGS
@@ -329,19 +328,14 @@ ENTRY(save_args)
 	 */
 1:	incl PER_CPU_VAR(irq_count)
 	jne 2f
-	popq_cfi %rax			/* move return address... */
 	mov PER_CPU_VAR(irq_stack_ptr),%rsp
 	EMPTY_FRAME 0
-	pushq_cfi %rbp			/* backlink for unwinder */
-	pushq_cfi %rax			/* ... to the new stack */
-	/*
-	 * We entered an interrupt context - irqs are off:
-	 */
-2:	TRACE_IRQS_OFF
-	ret
-	CFI_ENDPROC
-END(save_args)
-	.popsection
+
+2:	/* Store previous stack value */
+	pushq %rsi
+	/* We entered an interrupt context - irqs are off: */
+	TRACE_IRQS_OFF
+	.endm
 
 ENTRY(save_rest)
 	PARTIAL_FRAME 1 REST_SKIP+8
@@ -791,7 +785,7 @@ END(interrupt)
 	/* reserve pt_regs for scratch regs and rbp */
 	subq $ORIG_RAX-RBP, %rsp
 	CFI_ADJUST_CFA_OFFSET ORIG_RAX-RBP
-	call save_args
+	SAVE_ARGS_IRQ
 	PARTIAL_FRAME 0
 	call \func
 	.endm
@@ -814,15 +808,14 @@ ret_from_intr:
 	DISABLE_INTERRUPTS(CLBR_NONE)
 	TRACE_IRQS_OFF
 	decl PER_CPU_VAR(irq_count)
-	leaveq
 
-	CFI_RESTORE		rbp
+	/* Restore saved previous stack */
+	popq %rsi
+	leaq 16(%rsi), %rsp
+
 	CFI_DEF_CFA_REGISTER	rsp
-	CFI_ADJUST_CFA_OFFSET	-8
+	CFI_ADJUST_CFA_OFFSET	-16
 
-	/* we did not save rbx, restore only from ARGOFFSET */
-	addq $8, %rsp
-	CFI_ADJUST_CFA_OFFSET	-8
 exit_intr:
 	GET_THREAD_INFO(%rcx)
 	testl $3,CS-ARGOFFSET(%rsp)
diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c
index 5f9ecff328b..00354d4919a 100644
--- a/arch/x86/kernel/kgdb.c
+++ b/arch/x86/kernel/kgdb.c
@@ -608,7 +608,7 @@ int kgdb_arch_init(void)
 	return register_die_notifier(&kgdb_notifier);
 }
 
-static void kgdb_hw_overflow_handler(struct perf_event *event, int nmi,
+static void kgdb_hw_overflow_handler(struct perf_event *event,
 		struct perf_sample_data *data, struct pt_regs *regs)
 {
 	struct task_struct *tsk = current;
@@ -638,7 +638,7 @@ void kgdb_arch_late(void)
 	for (i = 0; i < HBP_NUM; i++) {
 		if (breakinfo[i].pev)
 			continue;
-		breakinfo[i].pev = register_wide_hw_breakpoint(&attr, NULL);
+		breakinfo[i].pev = register_wide_hw_breakpoint(&attr, NULL, NULL);
 		if (IS_ERR((void * __force)breakinfo[i].pev)) {
 			printk(KERN_ERR "kgdb: Could not allocate hw"
 			       "breakpoints\nDisabling the kernel debugger\n");
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 807c2a2b80f..82528799c5d 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -528,7 +528,7 @@ static int genregs_set(struct task_struct *target,
 	return ret;
 }
 
-static void ptrace_triggered(struct perf_event *bp, int nmi,
+static void ptrace_triggered(struct perf_event *bp,
 			     struct perf_sample_data *data,
 			     struct pt_regs *regs)
 {
@@ -715,7 +715,8 @@ static int ptrace_set_breakpoint_addr(struct task_struct *tsk, int nr,
 		attr.bp_type = HW_BREAKPOINT_W;
 		attr.disabled = 1;
 
-		bp = register_user_hw_breakpoint(&attr, ptrace_triggered, tsk);
+		bp = register_user_hw_breakpoint(&attr, ptrace_triggered,
+						 NULL, tsk);
 
 		/*
 		 * CHECKME: the previous code returned -EIO if the addr wasn't
diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c
index 55d9bc03f69..fdd0c6430e5 100644
--- a/arch/x86/kernel/stacktrace.c
+++ b/arch/x86/kernel/stacktrace.c
@@ -66,7 +66,7 @@ void save_stack_trace(struct stack_trace *trace)
 }
 EXPORT_SYMBOL_GPL(save_stack_trace);
 
-void save_stack_trace_regs(struct stack_trace *trace, struct pt_regs *regs)
+void save_stack_trace_regs(struct pt_regs *regs, struct stack_trace *trace)
 {
 	dump_trace(current, regs, NULL, 0, &save_stack_ops, trace);
 	if (trace->nr_entries < trace->max_entries)
diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile
index f2479f19ddd..6ba477342b8 100644
--- a/arch/x86/lib/Makefile
+++ b/arch/x86/lib/Makefile
@@ -18,7 +18,7 @@ obj-$(CONFIG_SMP) += msr-smp.o cache-smp.o
 
 lib-y := delay.o
 lib-y += thunk_$(BITS).o
-lib-y += usercopy_$(BITS).o getuser.o putuser.o
+lib-y += usercopy_$(BITS).o usercopy.o getuser.o putuser.o
 lib-y += memcpy_$(BITS).o
 lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o
 
diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c
new file mode 100644
index 00000000000..97be9cb5448
--- /dev/null
+++ b/arch/x86/lib/usercopy.c
@@ -0,0 +1,43 @@
+/*
+ * User address space access functions.
+ *
+ *  For licencing details see kernel-base/COPYING
+ */
+
+#include <linux/highmem.h>
+#include <linux/module.h>
+
+/*
+ * best effort, GUP based copy_from_user() that is NMI-safe
+ */
+unsigned long
+copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
+{
+	unsigned long offset, addr = (unsigned long)from;
+	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);
+		memcpy(to, map+offset, size);
+		kunmap_atomic(map);
+		put_page(page);
+
+		len  += size;
+		to   += size;
+		addr += size;
+
+	} while (len < n);
+
+	return len;
+}
+EXPORT_SYMBOL_GPL(copy_from_user_nmi);
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 2dbf6bf4c7e..4d09df054e3 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -1059,7 +1059,7 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	if (unlikely(error_code & PF_RSVD))
 		pgtable_bad(regs, error_code, address);
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	/*
 	 * If we're in an interrupt, have no user context or are running
@@ -1161,11 +1161,11 @@ good_area:
 	if (flags & FAULT_FLAG_ALLOW_RETRY) {
 		if (fault & VM_FAULT_MAJOR) {
 			tsk->maj_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				      regs, address);
 		} else {
 			tsk->min_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				      regs, address);
 		}
 		if (fault & VM_FAULT_RETRY) {
diff --git a/arch/x86/mm/kmemcheck/error.c b/arch/x86/mm/kmemcheck/error.c
index 704a37ceddd..dab41876cdd 100644
--- a/arch/x86/mm/kmemcheck/error.c
+++ b/arch/x86/mm/kmemcheck/error.c
@@ -185,7 +185,7 @@ void kmemcheck_error_save(enum kmemcheck_shadow state,
 	e->trace.entries = e->trace_entries;
 	e->trace.max_entries = ARRAY_SIZE(e->trace_entries);
 	e->trace.skip = 0;
-	save_stack_trace_regs(&e->trace, regs);
+	save_stack_trace_regs(regs, &e->trace);
 
 	/* Round address down to nearest 16 bytes */
 	shadow_copy = kmemcheck_shadow_lookup(address
diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c
index a5b64ab4cd6..bff89dfe361 100644
--- a/arch/x86/oprofile/backtrace.c
+++ b/arch/x86/oprofile/backtrace.c
@@ -11,10 +11,11 @@
 #include <linux/oprofile.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
+#include <linux/compat.h>
+#include <linux/uaccess.h>
+
 #include <asm/ptrace.h>
-#include <asm/uaccess.h>
 #include <asm/stacktrace.h>
-#include <linux/compat.h>
 
 static int backtrace_stack(void *data, char *name)
 {
@@ -40,13 +41,13 @@ static struct stacktrace_ops backtrace_ops = {
 static struct stack_frame_ia32 *
 dump_user_backtrace_32(struct stack_frame_ia32 *head)
 {
+	/* Also check accessibility of one struct frame_head beyond: */
 	struct stack_frame_ia32 bufhead[2];
 	struct stack_frame_ia32 *fp;
+	unsigned long bytes;
 
-	/* Also check accessibility of one struct frame_head beyond */
-	if (!access_ok(VERIFY_READ, head, sizeof(bufhead)))
-		return NULL;
-	if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead)))
+	bytes = copy_from_user_nmi(bufhead, head, sizeof(bufhead));
+	if (bytes != sizeof(bufhead))
 		return NULL;
 
 	fp = (struct stack_frame_ia32 *) compat_ptr(bufhead[0].next_frame);
@@ -87,12 +88,12 @@ x86_backtrace_32(struct pt_regs * const regs, unsigned int depth)
 
 static struct stack_frame *dump_user_backtrace(struct stack_frame *head)
 {
+	/* Also check accessibility of one struct frame_head beyond: */
 	struct stack_frame bufhead[2];
+	unsigned long bytes;
 
-	/* Also check accessibility of one struct stack_frame beyond */
-	if (!access_ok(VERIFY_READ, head, sizeof(bufhead)))
-		return NULL;
-	if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead)))
+	bytes = copy_from_user_nmi(bufhead, head, sizeof(bufhead));
+	if (bytes != sizeof(bufhead))
 		return NULL;
 
 	oprofile_add_trace(bufhead[0].return_address);