Merge remote-tracking branch 'tip/perf/core' into kvm-updates/3.3

* tip/perf/core: (66 commits)
  perf, x86: Expose perf capability to other modules
  perf, x86: Implement arch event mask as quirk
  x86, perf: Disable non available architectural events
  jump_label: Provide jump_label_key initializers
  jump_label, x86: Fix section mismatch
  perf, core: Rate limit perf_sched_events jump_label patching
  perf: Fix enable_on_exec for sibling events
  perf: Remove superfluous arguments
  perf, x86: Prefer fixed-purpose counters when scheduling
  perf, x86: Fix event scheduler for constraints with overlapping counters
  perf, x86: Implement event scheduler helper functions
  perf: Avoid a useless pmu_disable() in the perf-tick
  x86/tools: Add decoded instruction dump mode
  x86: Update instruction decoder to support new AVX formats
  x86/tools: Fix insn_sanity message outputs
  x86/tools: Fix instruction decoder message output
  x86: Fix instruction decoder to handle grouped AVX instructions
  x86/tools: Fix Makefile to build all test tools
  perf test: Soft errors shouldn't stop the "Validate PERF_RECORD_" test
  perf test: Validate PERF_RECORD_ events and perf_sample fields
  ...
Signed-off-by: Avi Kivity <avi@redhat.com>

* commit 'b3d9468a': (66 commits)
  perf, x86: Expose perf capability to other modules
  perf, x86: Implement arch event mask as quirk
  x86, perf: Disable non available architectural events
  jump_label: Provide jump_label_key initializers
  jump_label, x86: Fix section mismatch
  perf, core: Rate limit perf_sched_events jump_label patching
  perf: Fix enable_on_exec for sibling events
  perf: Remove superfluous arguments
  perf, x86: Prefer fixed-purpose counters when scheduling
  perf, x86: Fix event scheduler for constraints with overlapping counters
  perf, x86: Implement event scheduler helper functions
  perf: Avoid a useless pmu_disable() in the perf-tick
  x86/tools: Add decoded instruction dump mode
  x86: Update instruction decoder to support new AVX formats
  x86/tools: Fix insn_sanity message outputs
  x86/tools: Fix instruction decoder message output
  x86: Fix instruction decoder to handle grouped AVX instructions
  x86/tools: Fix Makefile to build all test tools
  perf test: Soft errors shouldn't stop the "Validate PERF_RECORD_" test
  perf test: Validate PERF_RECORD_ events and perf_sample fields
  ...

Documentation/kernel-parameters.txt

@@ -1882,6 +1882,9 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			arch_perfmon: [X86] Force use of architectural
 				perfmon on Intel CPUs instead of the
 				CPU specific event set.
+			timer: [X86] Force use of architectural NMI
+				timer mode (see also oprofile.timer
+				for generic hr timer mode)
 
 	oops=panic	Always panic on oopses. Default is to just kill the
 			process, but there is a small probability of

Documentation/trace/events.txt

@@ -191,8 +191,6 @@ And for string fields they are:
 
 Currently, only exact string matches are supported.
 
-Currently, the maximum number of predicates in a filter is 16.
-
 5.2 Setting filters
 -------------------
 

arch/Kconfig

@@ -30,6 +30,10 @@ config OPROFILE_EVENT_MULTIPLEX
 config HAVE_OPROFILE
 	bool
 
+config OPROFILE_NMI_TIMER
+	def_bool y
+	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
+
 config KPROBES
 	bool "Kprobes"
 	depends on MODULES

arch/x86/include/asm/insn.h

@@ -137,6 +137,13 @@ static inline int insn_is_avx(struct insn *insn)
 	return (insn->vex_prefix.value != 0);
 }
 
+/* Ensure this instruction is decoded completely */
+static inline int insn_complete(struct insn *insn)
+{
+	return insn->opcode.got && insn->modrm.got && insn->sib.got &&
+		insn->displacement.got && insn->immediate.got;
+}
+
 static inline insn_byte_t insn_vex_m_bits(struct insn *insn)
 {
 	if (insn->vex_prefix.nbytes == 2)	/* 2 bytes VEX */

arch/x86/include/asm/perf_event.h

@@ -57,6 +57,7 @@
 		(1 << (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX))
 
 #define ARCH_PERFMON_BRANCH_MISSES_RETIRED		6
+#define ARCH_PERFMON_EVENTS_COUNT			7
 
 /*
  * Intel "Architectural Performance Monitoring" CPUID
@@ -72,6 +73,19 @@ union cpuid10_eax {
 	unsigned int full;
 };
 
+union cpuid10_ebx {
+	struct {
+		unsigned int no_unhalted_core_cycles:1;
+		unsigned int no_instructions_retired:1;
+		unsigned int no_unhalted_reference_cycles:1;
+		unsigned int no_llc_reference:1;
+		unsigned int no_llc_misses:1;
+		unsigned int no_branch_instruction_retired:1;
+		unsigned int no_branch_misses_retired:1;
+	} split;
+	unsigned int full;
+};
+
 union cpuid10_edx {
 	struct {
 		unsigned int num_counters_fixed:5;
@@ -81,6 +95,15 @@ union cpuid10_edx {
 	unsigned int full;
 };
 
+struct x86_pmu_capability {
+	int		version;
+	int		num_counters_gp;
+	int		num_counters_fixed;
+	int		bit_width_gp;
+	int		bit_width_fixed;
+	unsigned int	events_mask;
+	int		events_mask_len;
+};
 
 /*
  * Fixed-purpose performance events:
@@ -202,6 +225,7 @@ struct perf_guest_switch_msr {
 };
 
 extern struct perf_guest_switch_msr *perf_guest_get_msrs(int *nr);
+extern void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap);
 #else
 static inline perf_guest_switch_msr *perf_guest_get_msrs(int *nr)
 {
@@ -209,6 +233,11 @@ static inline perf_guest_switch_msr *perf_guest_get_msrs(int *nr)
 	return NULL;
 }
 
+static inline void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap)
+{
+	memset(cap, 0, sizeof(*cap));
+}
+
 static inline void perf_events_lapic_init(void)	{ }
 #endif
 

arch/x86/kernel/cpu/perf_event.c

@@ -484,18 +484,195 @@ static inline int is_x86_event(struct perf_event *event)
 	return event->pmu == &pmu;
 }
 
+/*
+ * Event scheduler state:
+ *
+ * Assign events iterating over all events and counters, beginning
+ * with events with least weights first. Keep the current iterator
+ * state in struct sched_state.
+ */
+struct sched_state {
+	int	weight;
+	int	event;		/* event index */
+	int	counter;	/* counter index */
+	int	unassigned;	/* number of events to be assigned left */
+	unsigned long used[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
+};
+
+/* Total max is X86_PMC_IDX_MAX, but we are O(n!) limited */
+#define	SCHED_STATES_MAX	2
+
+struct perf_sched {
+	int			max_weight;
+	int			max_events;
+	struct event_constraint	**constraints;
+	struct sched_state	state;
+	int			saved_states;
+	struct sched_state	saved[SCHED_STATES_MAX];
+};
+
+/*
+ * Initialize interator that runs through all events and counters.
+ */
+static void perf_sched_init(struct perf_sched *sched, struct event_constraint **c,
+			    int num, int wmin, int wmax)
+{
+	int idx;
+
+	memset(sched, 0, sizeof(*sched));
+	sched->max_events	= num;
+	sched->max_weight	= wmax;
+	sched->constraints	= c;
+
+	for (idx = 0; idx < num; idx++) {
+		if (c[idx]->weight == wmin)
+			break;
+	}
+
+	sched->state.event	= idx;		/* start with min weight */
+	sched->state.weight	= wmin;
+	sched->state.unassigned	= num;
+}
+
+static void perf_sched_save_state(struct perf_sched *sched)
+{
+	if (WARN_ON_ONCE(sched->saved_states >= SCHED_STATES_MAX))
+		return;
+
+	sched->saved[sched->saved_states] = sched->state;
+	sched->saved_states++;
+}
+
+static bool perf_sched_restore_state(struct perf_sched *sched)
+{
+	if (!sched->saved_states)
+		return false;
+
+	sched->saved_states--;
+	sched->state = sched->saved[sched->saved_states];
+
+	/* continue with next counter: */
+	clear_bit(sched->state.counter++, sched->state.used);
+
+	return true;
+}
+
+/*
+ * Select a counter for the current event to schedule. Return true on
+ * success.
+ */
+static bool __perf_sched_find_counter(struct perf_sched *sched)
+{
+	struct event_constraint *c;
+	int idx;
+
+	if (!sched->state.unassigned)
+		return false;
+
+	if (sched->state.event >= sched->max_events)
+		return false;
+
+	c = sched->constraints[sched->state.event];
+
+	/* Prefer fixed purpose counters */
+	if (x86_pmu.num_counters_fixed) {
+		idx = X86_PMC_IDX_FIXED;
+		for_each_set_bit_cont(idx, c->idxmsk, X86_PMC_IDX_MAX) {
+			if (!__test_and_set_bit(idx, sched->state.used))
+				goto done;
+		}
+	}
+	/* Grab the first unused counter starting with idx */
+	idx = sched->state.counter;
+	for_each_set_bit_cont(idx, c->idxmsk, X86_PMC_IDX_FIXED) {
+		if (!__test_and_set_bit(idx, sched->state.used))
+			goto done;
+	}
+
+	return false;
+
+done:
+	sched->state.counter = idx;
+
+	if (c->overlap)
+		perf_sched_save_state(sched);
+
+	return true;
+}
+
+static bool perf_sched_find_counter(struct perf_sched *sched)
+{
+	while (!__perf_sched_find_counter(sched)) {
+		if (!perf_sched_restore_state(sched))
+			return false;
+	}
+
+	return true;
+}
+
+/*
+ * Go through all unassigned events and find the next one to schedule.
+ * Take events with the least weight first. Return true on success.
+ */
+static bool perf_sched_next_event(struct perf_sched *sched)
+{
+	struct event_constraint *c;
+
+	if (!sched->state.unassigned || !--sched->state.unassigned)
+		return false;
+
+	do {
+		/* next event */
+		sched->state.event++;
+		if (sched->state.event >= sched->max_events) {
+			/* next weight */
+			sched->state.event = 0;
+			sched->state.weight++;
+			if (sched->state.weight > sched->max_weight)
+				return false;
+		}
+		c = sched->constraints[sched->state.event];
+	} while (c->weight != sched->state.weight);
+
+	sched->state.counter = 0;	/* start with first counter */
+
+	return true;
+}
+
+/*
+ * Assign a counter for each event.
+ */
+static int perf_assign_events(struct event_constraint **constraints, int n,
+			      int wmin, int wmax, int *assign)
+{
+	struct perf_sched sched;
+
+	perf_sched_init(&sched, constraints, n, wmin, wmax);
+
+	do {
+		if (!perf_sched_find_counter(&sched))
+			break;	/* failed */
+		if (assign)
+			assign[sched.state.event] = sched.state.counter;
+	} while (perf_sched_next_event(&sched));
+
+	return sched.state.unassigned;
+}
+
 int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
 {
 	struct event_constraint *c, *constraints[X86_PMC_IDX_MAX];
 	unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
-	int i, j, w, wmax, num = 0;
+	int i, wmin, wmax, num = 0;
 	struct hw_perf_event *hwc;
 
 	bitmap_zero(used_mask, X86_PMC_IDX_MAX);
 
-	for (i = 0; i < n; i++) {
+	for (i = 0, wmin = X86_PMC_IDX_MAX, wmax = 0; i < n; i++) {
 		c = x86_pmu.get_event_constraints(cpuc, cpuc->event_list[i]);
 		constraints[i] = c;
+		wmin = min(wmin, c->weight);
+		wmax = max(wmax, c->weight);
 	}
 
 	/*
@@ -521,59 +698,11 @@ int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
 		if (assign)
 			assign[i] = hwc->idx;
 	}
-	if (i == n)
-		goto done;
-
-	/*
-	 * begin slow path
-	 */
-
-	bitmap_zero(used_mask, X86_PMC_IDX_MAX);
-
-	/*
-	 * weight = number of possible counters
-	 *
-	 * 1    = most constrained, only works on one counter
-	 * wmax = least constrained, works on any counter
-	 *
-	 * assign events to counters starting with most
-	 * constrained 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_counters_fixed)
-		wmax++;
-
-	for (w = 1, num = n; num && w <= wmax; w++) {
-		/* for each event */
-		for (i = 0; num && i < n; i++) {
-			c = constraints[i];
-			hwc = &cpuc->event_list[i]->hw;
-
-			if (c->weight != w)
-				continue;
-
-			for_each_set_bit(j, c->idxmsk, X86_PMC_IDX_MAX) {
-				if (!test_bit(j, used_mask))
-					break;
-			}
-
-			if (j == X86_PMC_IDX_MAX)
-				break;
 
-			__set_bit(j, used_mask);
+	/* slow path */
+	if (i != n)
+		num = perf_assign_events(constraints, n, wmin, wmax, assign);
 
-			if (assign)
-				assign[i] = j;
-			num--;
-		}
-	}
-done:
 	/*
 	 * scheduling failed or is just a simulation,
 	 * free resources if necessary
@@ -1119,6 +1248,7 @@ static void __init pmu_check_apic(void)
 
 static int __init init_hw_perf_events(void)
 {
+	struct x86_pmu_quirk *quirk;
 	struct event_constraint *c;
 	int err;
 
@@ -1147,8 +1277,8 @@ static int __init init_hw_perf_events(void)
 
 	pr_cont("%s PMU driver.\n", x86_pmu.name);
 
-	if (x86_pmu.quirks)
-		x86_pmu.quirks();
+	for (quirk = x86_pmu.quirks; quirk; quirk = quirk->next)
+		quirk->func();
 
 	if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) {
 		WARN(1, KERN_ERR "hw perf events %d > max(%d), clipping!",
@@ -1171,7 +1301,7 @@ static int __init init_hw_perf_events(void)
 
 	unconstrained = (struct event_constraint)
 		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
-				   0, x86_pmu.num_counters);
+				   0, x86_pmu.num_counters, 0);
 
 	if (x86_pmu.event_constraints) {
 		for_each_event_constraint(c, x86_pmu.event_constraints) {
@@ -1566,3 +1696,15 @@ unsigned long perf_misc_flags(struct pt_regs *regs)
 
 	return misc;
 }
+
+void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap)
+{
+	cap->version		= x86_pmu.version;
+	cap->num_counters_gp	= x86_pmu.num_counters;
+	cap->num_counters_fixed	= x86_pmu.num_counters_fixed;
+	cap->bit_width_gp	= x86_pmu.cntval_bits;
+	cap->bit_width_fixed	= x86_pmu.cntval_bits;
+	cap->events_mask	= (unsigned int)x86_pmu.events_maskl;
+	cap->events_mask_len	= x86_pmu.events_mask_len;
+}
+EXPORT_SYMBOL_GPL(perf_get_x86_pmu_capability);

arch/x86/kernel/cpu/perf_event.h

@@ -45,6 +45,7 @@ struct event_constraint {
 	u64	code;
 	u64	cmask;
 	int	weight;
+	int	overlap;
 };
 
 struct amd_nb {
@@ -151,15 +152,40 @@ struct cpu_hw_events {
 	void				*kfree_on_online;
 };
 
-#define __EVENT_CONSTRAINT(c, n, m, w) {\
+#define __EVENT_CONSTRAINT(c, n, m, w, o) {\
 	{ .idxmsk64 = (n) },		\
 	.code = (c),			\
 	.cmask = (m),			\
 	.weight = (w),			\
+	.overlap = (o),			\
 }
 
 #define EVENT_CONSTRAINT(c, n, m)	\
-	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n))
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0)
+
+/*
+ * The overlap flag marks event constraints with overlapping counter
+ * masks. This is the case if the counter mask of such an event is not
+ * a subset of any other counter mask of a constraint with an equal or
+ * higher weight, e.g.:
+ *
+ *  c_overlaps = EVENT_CONSTRAINT_OVERLAP(0, 0x09, 0);
+ *  c_another1 = EVENT_CONSTRAINT(0, 0x07, 0);
+ *  c_another2 = EVENT_CONSTRAINT(0, 0x38, 0);
+ *
+ * The event scheduler may not select the correct counter in the first
+ * cycle because it needs to know which subsequent events will be
+ * scheduled. It may fail to schedule the events then. So we set the
+ * overlap flag for such constraints to give the scheduler a hint which
+ * events to select for counter rescheduling.
+ *
+ * Care must be taken as the rescheduling algorithm is O(n!) which
+ * will increase scheduling cycles for an over-commited system
+ * dramatically.  The number of such EVENT_CONSTRAINT_OVERLAP() macros
+ * and its counter masks must be kept at a minimum.
+ */
+#define EVENT_CONSTRAINT_OVERLAP(c, n, m)	\
+	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1)
 
 /*
  * Constraint on the Event code.
@@ -235,6 +261,11 @@ union perf_capabilities {
 	u64	capabilities;
 };
 
+struct x86_pmu_quirk {
+	struct x86_pmu_quirk *next;
+	void (*func)(void);
+};
+
 /*
  * struct x86_pmu - generic x86 pmu
  */
@@ -259,6 +290,11 @@ struct x86_pmu {
 	int		num_counters_fixed;
 	int		cntval_bits;
 	u64		cntval_mask;
+	union {
+			unsigned long events_maskl;
+			unsigned long events_mask[BITS_TO_LONGS(ARCH_PERFMON_EVENTS_COUNT)];
+	};
+	int		events_mask_len;
 	int		apic;
 	u64		max_period;
 	struct event_constraint *
@@ -268,7 +304,7 @@ struct x86_pmu {
 	void		(*put_event_constraints)(struct cpu_hw_events *cpuc,
 						 struct perf_event *event);
 	struct event_constraint *event_constraints;
-	void		(*quirks)(void);
+	struct x86_pmu_quirk *quirks;
 	int		perfctr_second_write;
 
 	int		(*cpu_prepare)(int cpu);
@@ -309,6 +345,15 @@ struct x86_pmu {
 	struct perf_guest_switch_msr *(*guest_get_msrs)(int *nr);
 };
 
+#define x86_add_quirk(func_)						\
+do {									\
+	static struct x86_pmu_quirk __quirk __initdata = {		\
+		.func = func_,						\
+	};								\
+	__quirk.next = x86_pmu.quirks;					\
+	x86_pmu.quirks = &__quirk;					\
+} while (0)
+
 #define ERF_NO_HT_SHARING	1
 #define ERF_HAS_RSP_1		2
 

arch/x86/kernel/cpu/perf_event_amd.c

@@ -492,7 +492,7 @@ static __initconst const struct x86_pmu amd_pmu = {
 static struct event_constraint amd_f15_PMC0  = EVENT_CONSTRAINT(0, 0x01, 0);
 static struct event_constraint amd_f15_PMC20 = EVENT_CONSTRAINT(0, 0x07, 0);
 static struct event_constraint amd_f15_PMC3  = EVENT_CONSTRAINT(0, 0x08, 0);
-static struct event_constraint amd_f15_PMC30 = EVENT_CONSTRAINT(0, 0x09, 0);
+static struct event_constraint amd_f15_PMC30 = EVENT_CONSTRAINT_OVERLAP(0, 0x09, 0);
 static struct event_constraint amd_f15_PMC50 = EVENT_CONSTRAINT(0, 0x3F, 0);
 static struct event_constraint amd_f15_PMC53 = EVENT_CONSTRAINT(0, 0x38, 0);
 

arch/x86/kernel/cpu/perf_event_intel.c

@@ -1519,7 +1519,7 @@ static __initconst const struct x86_pmu intel_pmu = {
 	.guest_get_msrs		= intel_guest_get_msrs,
 };
 
-static void intel_clovertown_quirks(void)
+static __init void intel_clovertown_quirk(void)
 {
 	/*
 	 * PEBS is unreliable due to:
@@ -1545,19 +1545,60 @@ static void intel_clovertown_quirks(void)
 	x86_pmu.pebs_constraints = NULL;
 }
 
-static void intel_sandybridge_quirks(void)
+static __init void intel_sandybridge_quirk(void)
 {
 	printk(KERN_WARNING "PEBS disabled due to CPU errata.\n");
 	x86_pmu.pebs = 0;
 	x86_pmu.pebs_constraints = NULL;
 }
 
+static const struct { int id; char *name; } intel_arch_events_map[] __initconst = {
+	{ PERF_COUNT_HW_CPU_CYCLES, "cpu cycles" },
+	{ PERF_COUNT_HW_INSTRUCTIONS, "instructions" },
+	{ PERF_COUNT_HW_BUS_CYCLES, "bus cycles" },
+	{ PERF_COUNT_HW_CACHE_REFERENCES, "cache references" },
+	{ PERF_COUNT_HW_CACHE_MISSES, "cache misses" },
+	{ PERF_COUNT_HW_BRANCH_INSTRUCTIONS, "branch instructions" },
+	{ PERF_COUNT_HW_BRANCH_MISSES, "branch misses" },
+};
+
+static __init void intel_arch_events_quirk(void)
+{
+	int bit;
+
+	/* disable event that reported as not presend by cpuid */
+	for_each_set_bit(bit, x86_pmu.events_mask, ARRAY_SIZE(intel_arch_events_map)) {
+		intel_perfmon_event_map[intel_arch_events_map[bit].id] = 0;
+		printk(KERN_WARNING "CPUID marked event: \'%s\' unavailable\n",
+				intel_arch_events_map[bit].name);
+	}
+}
+
+static __init void intel_nehalem_quirk(void)
+{
+	union cpuid10_ebx ebx;
+
+	ebx.full = x86_pmu.events_maskl;
+	if (ebx.split.no_branch_misses_retired) {
+		/*
+		 * Erratum AAJ80 detected, we work it around by using
+		 * the BR_MISP_EXEC.ANY event. This will over-count
+		 * branch-misses, but it's still much better than the
+		 * architectural event which is often completely bogus:
+		 */
+		intel_perfmon_event_map[PERF_COUNT_HW_BRANCH_MISSES] = 0x7f89;
+		ebx.split.no_branch_misses_retired = 0;
+		x86_pmu.events_maskl = ebx.full;
+		printk(KERN_INFO "CPU erratum AAJ80 worked around\n");
+	}
+}
+
 __init int intel_pmu_init(void)
 {
 	union cpuid10_edx edx;
 	union cpuid10_eax eax;
+	union cpuid10_ebx ebx;
 	unsigned int unused;
-	unsigned int ebx;
 	int version;
 
 	if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
@@ -1574,8 +1615,8 @@ __init int intel_pmu_init(void)
 	 * Check whether the Architectural PerfMon supports
 	 * Branch Misses Retired hw_event or not.
 	 */
-	cpuid(10, &eax.full, &ebx, &unused, &edx.full);
-	if (eax.split.mask_length <= ARCH_PERFMON_BRANCH_MISSES_RETIRED)
+	cpuid(10, &eax.full, &ebx.full, &unused, &edx.full);
+	if (eax.split.mask_length < ARCH_PERFMON_EVENTS_COUNT)
 		return -ENODEV;
 
 	version = eax.split.version_id;
@@ -1589,6 +1630,9 @@ __init int intel_pmu_init(void)
 	x86_pmu.cntval_bits		= eax.split.bit_width;
 	x86_pmu.cntval_mask		= (1ULL << eax.split.bit_width) - 1;
 
+	x86_pmu.events_maskl		= ebx.full;
+	x86_pmu.events_mask_len		= eax.split.mask_length;
+
 	/*
 	 * Quirk: v2 perfmon does not report fixed-purpose events, so
 	 * assume at least 3 events:
@@ -1608,6 +1652,8 @@ __init int intel_pmu_init(void)
 
 	intel_ds_init();
 
+	x86_add_quirk(intel_arch_events_quirk); /* Install first, so it runs last */
+
 	/*
 	 * Install the hw-cache-events table:
 	 */
@@ -1617,7 +1663,7 @@ __init int intel_pmu_init(void)
 		break;
 
 	case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */
-		x86_pmu.quirks = intel_clovertown_quirks;
+		x86_add_quirk(intel_clovertown_quirk);
 	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" */
@@ -1651,17 +1697,8 @@ __init int intel_pmu_init(void)
 		/* UOPS_EXECUTED.CORE_ACTIVE_CYCLES,c=1,i=1 */
 		intel_perfmon_event_map[PERF_COUNT_HW_STALLED_CYCLES_BACKEND] = 0x1803fb1;
 
-		if (ebx & 0x40) {
-			/*
-			 * Erratum AAJ80 detected, we work it around by using
-			 * the BR_MISP_EXEC.ANY event. This will over-count
-			 * branch-misses, but it's still much better than the
-			 * architectural event which is often completely bogus:
-			 */
-			intel_perfmon_event_map[PERF_COUNT_HW_BRANCH_MISSES] = 0x7f89;
+		x86_add_quirk(intel_nehalem_quirk);
 
-			pr_cont("erratum AAJ80 worked around, ");
-		}
 		pr_cont("Nehalem events, ");
 		break;
 
@@ -1701,7 +1738,7 @@ __init int intel_pmu_init(void)
 		break;
 
 	case 42: /* SandyBridge */
-		x86_pmu.quirks = intel_sandybridge_quirks;
+		x86_add_quirk(intel_sandybridge_quirk);
 	case 45: /* SandyBridge, "Romely-EP" */
 		memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
 		       sizeof(hw_cache_event_ids));
@@ -1738,5 +1775,6 @@ __init int intel_pmu_init(void)
 			break;
 		}
 	}
+
 	return 0;
 }

arch/x86/kernel/jump_label.c

@@ -50,7 +50,7 @@ void arch_jump_label_transform(struct jump_entry *entry,
 	put_online_cpus();
 }
 
-void arch_jump_label_transform_static(struct jump_entry *entry,
+__init_or_module void arch_jump_label_transform_static(struct jump_entry *entry,
 				      enum jump_label_type type)
 {
 	__jump_label_transform(entry, type, text_poke_early);

arch/x86/lib/inat.c

@@ -82,9 +82,16 @@ insn_attr_t inat_get_avx_attribute(insn_byte_t opcode, insn_byte_t vex_m,
 	const insn_attr_t *table;
 	if (vex_m > X86_VEX_M_MAX || vex_p > INAT_LSTPFX_MAX)
 		return 0;
-	table = inat_avx_tables[vex_m][vex_p];
+	/* At first, this checks the master table */
+	table = inat_avx_tables[vex_m][0];
 	if (!table)
 		return 0;
+	if (!inat_is_group(table[opcode]) && vex_p) {
+		/* If this is not a group, get attribute directly */
+		table = inat_avx_tables[vex_m][vex_p];
+		if (!table)
+			return 0;
+	}
 	return table[opcode];
 }
 

arch/x86/lib/insn.c

@@ -202,7 +202,7 @@ void insn_get_opcode(struct insn *insn)
 		m = insn_vex_m_bits(insn);
 		p = insn_vex_p_bits(insn);
 		insn->attr = inat_get_avx_attribute(op, m, p);
-		if (!inat_accept_vex(insn->attr))
+		if (!inat_accept_vex(insn->attr) && !inat_is_group(insn->attr))
 			insn->attr = 0;	/* This instruction is bad */
 		goto end;	/* VEX has only 1 byte for opcode */
 	}
@@ -249,6 +249,8 @@ void insn_get_modrm(struct insn *insn)
 			pfx = insn_last_prefix(insn);
 			insn->attr = inat_get_group_attribute(mod, pfx,
 							      insn->attr);
+			if (insn_is_avx(insn) && !inat_accept_vex(insn->attr))
+				insn->attr = 0;	/* This is bad */
 		}
 	}
 

arch/x86/oprofile/Makefile

@@ -4,9 +4,8 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \
 		oprof.o cpu_buffer.o buffer_sync.o \
 		event_buffer.o oprofile_files.o \
 		oprofilefs.o oprofile_stats.o  \
-		timer_int.o )
+		timer_int.o nmi_timer_int.o )
 
 oprofile-y				:= $(DRIVER_OBJS) init.o backtrace.o
 oprofile-$(CONFIG_X86_LOCAL_APIC) 	+= nmi_int.o op_model_amd.o \
 					   op_model_ppro.o op_model_p4.o
-oprofile-$(CONFIG_X86_IO_APIC)		+= nmi_timer_int.o

arch/x86/oprofile/init.c

@@ -16,37 +16,23 @@
  * with the NMI mode driver.
  */
 
+#ifdef CONFIG_X86_LOCAL_APIC
 extern int op_nmi_init(struct oprofile_operations *ops);
-extern int op_nmi_timer_init(struct oprofile_operations *ops);
 extern void op_nmi_exit(void);
-extern void x86_backtrace(struct pt_regs * const regs, unsigned int depth);
+#else
+static int op_nmi_init(struct oprofile_operations *ops) { return -ENODEV; }
+static void op_nmi_exit(void) { }
+#endif
 
-static int nmi_timer;
+extern void x86_backtrace(struct pt_regs * const regs, unsigned int depth);
 
 int __init oprofile_arch_init(struct oprofile_operations *ops)
 {
-	int ret;
-
-	ret = -ENODEV;
-
-#ifdef CONFIG_X86_LOCAL_APIC
-	ret = op_nmi_init(ops);
-#endif
-	nmi_timer = (ret != 0);
-#ifdef CONFIG_X86_IO_APIC
-	if (nmi_timer)
-		ret = op_nmi_timer_init(ops);
-#endif
 	ops->backtrace = x86_backtrace;
-
-	return ret;
+	return op_nmi_init(ops);
 }
 
-
 void oprofile_arch_exit(void)
 {
-#ifdef CONFIG_X86_LOCAL_APIC
-	if (!nmi_timer)
-		op_nmi_exit();
-#endif
+	op_nmi_exit();
 }

arch/x86/oprofile/nmi_int.c

@@ -595,24 +595,36 @@ static int __init p4_init(char **cpu_type)
 	return 0;
 }
 
-static int force_arch_perfmon;
-static int force_cpu_type(const char *str, struct kernel_param *kp)
+enum __force_cpu_type {
+	reserved = 0,		/* do not force */
+	timer,
+	arch_perfmon,
+};
+
+static int force_cpu_type;
+
+static int set_cpu_type(const char *str, struct kernel_param *kp)
 {
-	if (!strcmp(str, "arch_perfmon")) {
-		force_arch_perfmon = 1;
+	if (!strcmp(str, "timer")) {
+		force_cpu_type = timer;
+		printk(KERN_INFO "oprofile: forcing NMI timer mode\n");
+	} else if (!strcmp(str, "arch_perfmon")) {
+		force_cpu_type = arch_perfmon;
 		printk(KERN_INFO "oprofile: forcing architectural perfmon\n");
+	} else {
+		force_cpu_type = 0;
 	}
 
 	return 0;
 }
-module_param_call(cpu_type, force_cpu_type, NULL, NULL, 0);
+module_param_call(cpu_type, set_cpu_type, NULL, NULL, 0);
 
 static int __init ppro_init(char **cpu_type)
 {
 	__u8 cpu_model = boot_cpu_data.x86_model;
 	struct op_x86_model_spec *spec = &op_ppro_spec;	/* default */
 
-	if (force_arch_perfmon && cpu_has_arch_perfmon)
+	if (force_cpu_type == arch_perfmon && cpu_has_arch_perfmon)
 		return 0;
 
 	/*
@@ -679,6 +691,9 @@ int __init op_nmi_init(struct oprofile_operations *ops)
 	if (!cpu_has_apic)
 		return -ENODEV;
 
+	if (force_cpu_type == timer)
+		return -ENODEV;
+
 	switch (vendor) {
 	case X86_VENDOR_AMD:
 		/* Needs to be at least an Athlon (or hammer in 32bit mode) */

arch/x86/oprofile/nmi_timer_int.c

-/**
- * @file nmi_timer_int.c
- *
- * @remark Copyright 2003 OProfile authors
- * @remark Read the file COPYING
- *
- * @author Zwane Mwaikambo <zwane@linuxpower.ca>
- */
-
-#include <linux/init.h>
-#include <linux/smp.h>
-#include <linux/errno.h>
-#include <linux/oprofile.h>
-#include <linux/rcupdate.h>
-#include <linux/kdebug.h>
-
-#include <asm/nmi.h>
-#include <asm/apic.h>
-#include <asm/ptrace.h>
-
-static int profile_timer_exceptions_notify(unsigned int val, struct pt_regs *regs)
-{
-	oprofile_add_sample(regs, 0);
-	return NMI_HANDLED;
-}
-
-static int timer_start(void)
-{
-	if (register_nmi_handler(NMI_LOCAL, profile_timer_exceptions_notify,
-					0, "oprofile-timer"))
-		return 1;
-	return 0;
-}
-
-
-static void timer_stop(void)
-{
-	unregister_nmi_handler(NMI_LOCAL, "oprofile-timer");
-	synchronize_sched();  /* Allow already-started NMIs to complete. */
-}
-
-
-int __init op_nmi_timer_init(struct oprofile_operations *ops)
-{
-	ops->start = timer_start;
-	ops->stop = timer_stop;
-	ops->cpu_type = "timer";
-	printk(KERN_INFO "oprofile: using NMI timer interrupt.\n");
-	return 0;
-}

arch/x86/tools/Makefile

@@ -18,14 +18,21 @@ chkobjdump = $(srctree)/arch/x86/tools/chkobjdump.awk
 quiet_cmd_posttest = TEST    $@
       cmd_posttest = ($(OBJDUMP) -v | $(AWK) -f $(chkobjdump)) || $(OBJDUMP) -d -j .text $(objtree)/vmlinux | $(AWK) -f $(distill_awk) | $(obj)/test_get_len $(posttest_64bit) $(posttest_verbose)
 
-posttest: $(obj)/test_get_len vmlinux
+quiet_cmd_sanitytest = TEST    $@
+      cmd_sanitytest = $(obj)/insn_sanity $(posttest_64bit) -m 1000000
+
+posttest: $(obj)/test_get_len vmlinux $(obj)/insn_sanity
 	$(call cmd,posttest)
+	$(call cmd,sanitytest)
 
-hostprogs-y	:= test_get_len
+hostprogs-y	+= test_get_len insn_sanity
 
 # -I needed for generated C source and C source which in the kernel tree.
 HOSTCFLAGS_test_get_len.o := -Wall -I$(objtree)/arch/x86/lib/ -I$(srctree)/arch/x86/include/ -I$(srctree)/arch/x86/lib/ -I$(srctree)/include/
 
+HOSTCFLAGS_insn_sanity.o := -Wall -I$(objtree)/arch/x86/lib/ -I$(srctree)/arch/x86/include/ -I$(srctree)/arch/x86/lib/ -I$(srctree)/include/
+
 # Dependencies are also needed.
 $(obj)/test_get_len.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/inat.c $(srctree)/arch/x86/include/asm/inat_types.h $(srctree)/arch/x86/include/asm/inat.h $(srctree)/arch/x86/include/asm/insn.h $(objtree)/arch/x86/lib/inat-tables.c
 
+$(obj)/insn_sanity.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/inat.c $(srctree)/arch/x86/include/asm/inat_types.h $(srctree)/arch/x86/include/asm/inat.h $(srctree)/arch/x86/include/asm/insn.h $(objtree)/arch/x86/lib/inat-tables.c

arch/x86/tools/gen-insn-attr-x86.awk

@@ -47,7 +47,7 @@ BEGIN {
 	sep_expr = "^\\|$"
 	group_expr = "^Grp[0-9A-Za-z]+"
 
-	imm_expr = "^[IJAO][a-z]"
+	imm_expr = "^[IJAOL][a-z]"
 	imm_flag["Ib"] = "INAT_MAKE_IMM(INAT_IMM_BYTE)"
 	imm_flag["Jb"] = "INAT_MAKE_IMM(INAT_IMM_BYTE)"
 	imm_flag["Iw"] = "INAT_MAKE_IMM(INAT_IMM_WORD)"
@@ -59,6 +59,7 @@ BEGIN {
 	imm_flag["Iv"] = "INAT_MAKE_IMM(INAT_IMM_VWORD)"
 	imm_flag["Ob"] = "INAT_MOFFSET"
 	imm_flag["Ov"] = "INAT_MOFFSET"
+	imm_flag["Lx"] = "INAT_MAKE_IMM(INAT_IMM_BYTE)"
 
 	modrm_expr = "^([CDEGMNPQRSUVW/][a-z]+|NTA|T[012])"
 	force64_expr = "\\([df]64\\)"
@@ -70,8 +71,12 @@ BEGIN {
 	lprefix3_expr = "\\(F2\\)"
 	max_lprefix = 4
 
-	vexok_expr = "\\(VEX\\)"
-	vexonly_expr = "\\(oVEX\\)"
+	# All opcodes starting with lower-case 'v' or with (v1) superscript
+	# accepts VEX prefix
+	vexok_opcode_expr = "^v.*"
+	vexok_expr = "\\(v1\\)"
+	# All opcodes with (v) superscript supports *only* VEX prefix
+	vexonly_expr = "\\(v\\)"
 
 	prefix_expr = "\\(Prefix\\)"
 	prefix_num["Operand-Size"] = "INAT_PFX_OPNDSZ"
@@ -85,8 +90,8 @@ BEGIN {
 	prefix_num["SEG=GS"] = "INAT_PFX_GS"
 	prefix_num["SEG=SS"] = "INAT_PFX_SS"
 	prefix_num["Address-Size"] = "INAT_PFX_ADDRSZ"
-	prefix_num["2bytes-VEX"] = "INAT_PFX_VEX2"
-	prefix_num["3bytes-VEX"] = "INAT_PFX_VEX3"
+	prefix_num["VEX+1byte"] = "INAT_PFX_VEX2"
+	prefix_num["VEX+2byte"] = "INAT_PFX_VEX3"
 
 	clear_vars()
 }
@@ -310,12 +315,10 @@ function convert_operands(count,opnd,       i,j,imm,mod)
 		if (match(opcode, fpu_expr))
 			flags = add_flags(flags, "INAT_MODRM")
 
-		# check VEX only code
+		# check VEX codes
 		if (match(ext, vexonly_expr))
 			flags = add_flags(flags, "INAT_VEXOK | INAT_VEXONLY")
-
-		# check VEX only code
-		if (match(ext, vexok_expr))
+		else if (match(ext, vexok_expr) || match(opcode, vexok_opcode_expr))
 			flags = add_flags(flags, "INAT_VEXOK")
 
 		# check prefixes

arch/x86/tools/insn_sanity.c

+/*
+ * x86 decoder sanity test - based on test_get_insn.c
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2009
+ * Copyright (C) Hitachi, Ltd., 2011
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <assert.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+
+#define unlikely(cond) (cond)
+#define ARRAY_SIZE(a)	(sizeof(a)/sizeof(a[0]))
+
+#include <asm/insn.h>
+#include <inat.c>
+#include <insn.c>
+
+/*
+ * Test of instruction analysis against tampering.
+ * Feed random binary to instruction decoder and ensure not to
+ * access out-of-instruction-buffer.
+ */
+
+#define DEFAULT_MAX_ITER	10000
+#define INSN_NOP 0x90
+
+static const char	*prog;		/* Program name */
+static int		verbose;	/* Verbosity */
+static int		x86_64;		/* x86-64 bit mode flag */
+static unsigned int	seed;		/* Random seed */
+static unsigned long	iter_start;	/* Start of iteration number */
+static unsigned long	iter_end = DEFAULT_MAX_ITER;	/* End of iteration number */
+static FILE		*input_file;	/* Input file name */
+
+static void usage(const char *err)
+{
+	if (err)
+		fprintf(stderr, "Error: %s\n\n", err);
+	fprintf(stderr, "Usage: %s [-y|-n|-v] [-s seed[,no]] [-m max] [-i input]\n", prog);
+	fprintf(stderr, "\t-y	64bit mode\n");
+	fprintf(stderr, "\t-n	32bit mode\n");
+	fprintf(stderr, "\t-v	Verbosity(-vv dumps any decoded result)\n");
+	fprintf(stderr, "\t-s	Give a random seed (and iteration number)\n");
+	fprintf(stderr, "\t-m	Give a maximum iteration number\n");
+	fprintf(stderr, "\t-i	Give an input file with decoded binary\n");
+	exit(1);
+}
+
+static void dump_field(FILE *fp, const char *name, const char *indent,
+		       struct insn_field *field)
+{
+	fprintf(fp, "%s.%s = {\n", indent, name);
+	fprintf(fp, "%s\t.value = %d, bytes[] = {%x, %x, %x, %x},\n",
+		indent, field->value, field->bytes[0], field->bytes[1],
+		field->bytes[2], field->bytes[3]);
+	fprintf(fp, "%s\t.got = %d, .nbytes = %d},\n", indent,
+		field->got, field->nbytes);
+}
+
+static void dump_insn(FILE *fp, struct insn *insn)
+{
+	fprintf(fp, "Instruction = {\n");
+	dump_field(fp, "prefixes", "\t",	&insn->prefixes);
+	dump_field(fp, "rex_prefix", "\t",	&insn->rex_prefix);
+	dump_field(fp, "vex_prefix", "\t",	&insn->vex_prefix);
+	dump_field(fp, "opcode", "\t",		&insn->opcode);
+	dump_field(fp, "modrm", "\t",		&insn->modrm);
+	dump_field(fp, "sib", "\t",		&insn->sib);
+	dump_field(fp, "displacement", "\t",	&insn->displacement);
+	dump_field(fp, "immediate1", "\t",	&insn->immediate1);
+	dump_field(fp, "immediate2", "\t",	&insn->immediate2);
+	fprintf(fp, "\t.attr = %x, .opnd_bytes = %d, .addr_bytes = %d,\n",
+		insn->attr, insn->opnd_bytes, insn->addr_bytes);
+	fprintf(fp, "\t.length = %d, .x86_64 = %d, .kaddr = %p}\n",
+		insn->length, insn->x86_64, insn->kaddr);
+}
+
+static void dump_stream(FILE *fp, const char *msg, unsigned long nr_iter,
+			unsigned char *insn_buf, struct insn *insn)
+{
+	int i;
+
+	fprintf(fp, "%s:\n", msg);
+
+	dump_insn(fp, insn);
+
+	fprintf(fp, "You can reproduce this with below command(s);\n");
+
+	/* Input a decoded instruction sequence directly */
+	fprintf(fp, " $ echo ");
+	for (i = 0; i < MAX_INSN_SIZE; i++)
+		fprintf(fp, " %02x", insn_buf[i]);
+	fprintf(fp, " | %s -i -\n", prog);
+
+	if (!input_file) {
+		fprintf(fp, "Or \n");
+		/* Give a seed and iteration number */
+		fprintf(fp, " $ %s -s 0x%x,%lu\n", prog, seed, nr_iter);
+	}
+}
+
+static void init_random_seed(void)
+{
+	int fd;
+
+	fd = open("/dev/urandom", O_RDONLY);
+	if (fd < 0)
+		goto fail;
+
+	if (read(fd, &seed, sizeof(seed)) != sizeof(seed))
+		goto fail;
+
+	close(fd);
+	return;
+fail:
+	usage("Failed to open /dev/urandom");
+}
+
+/* Read given instruction sequence from the input file */
+static int read_next_insn(unsigned char *insn_buf)
+{
+	char buf[256]  = "", *tmp;
+	int i;
+
+	tmp = fgets(buf, ARRAY_SIZE(buf), input_file);
+	if (tmp == NULL || feof(input_file))
+		return 0;
+
+	for (i = 0; i < MAX_INSN_SIZE; i++) {
+		insn_buf[i] = (unsigned char)strtoul(tmp, &tmp, 16);
+		if (*tmp != ' ')
+			break;
+	}
+
+	return i;
+}
+
+static int generate_insn(unsigned char *insn_buf)
+{
+	int i;
+
+	if (input_file)
+		return read_next_insn(insn_buf);
+
+	/* Fills buffer with random binary up to MAX_INSN_SIZE */
+	for (i = 0; i < MAX_INSN_SIZE - 1; i += 2)
+		*(unsigned short *)(&insn_buf[i]) = random() & 0xffff;
+
+	while (i < MAX_INSN_SIZE)
+		insn_buf[i++] = random() & 0xff;
+
+	return i;
+}
+
+static void parse_args(int argc, char **argv)
+{
+	int c;
+	char *tmp = NULL;
+	int set_seed = 0;
+
+	prog = argv[0];
+	while ((c = getopt(argc, argv, "ynvs:m:i:")) != -1) {
+		switch (c) {
+		case 'y':
+			x86_64 = 1;
+			break;
+		case 'n':
+			x86_64 = 0;
+			break;
+		case 'v':
+			verbose++;
+			break;
+		case 'i':
+			if (strcmp("-", optarg) == 0)
+				input_file = stdin;
+			else
+				input_file = fopen(optarg, "r");
+			if (!input_file)
+				usage("Failed to open input file");
+			break;
+		case 's':
+			seed = (unsigned int)strtoul(optarg, &tmp, 0);
+			if (*tmp == ',') {
+				optarg = tmp + 1;
+				iter_start = strtoul(optarg, &tmp, 0);
+			}
+			if (*tmp != '\0' || tmp == optarg)
+				usage("Failed to parse seed");
+			set_seed = 1;
+			break;
+		case 'm':
+			iter_end = strtoul(optarg, &tmp, 0);
+			if (*tmp != '\0' || tmp == optarg)
+				usage("Failed to parse max_iter");
+			break;
+		default:
+			usage(NULL);
+		}
+	}
+
+	/* Check errors */
+	if (iter_end < iter_start)
+		usage("Max iteration number must be bigger than iter-num");
+
+	if (set_seed && input_file)
+		usage("Don't use input file (-i) with random seed (-s)");
+
+	/* Initialize random seed */
+	if (!input_file) {
+		if (!set_seed)	/* No seed is given */
+			init_random_seed();
+		srand(seed);
+	}
+}
+
+int main(int argc, char **argv)
+{
+	struct insn insn;
+	int insns = 0;
+	int errors = 0;
+	unsigned long i;
+	unsigned char insn_buf[MAX_INSN_SIZE * 2];
+
+	parse_args(argc, argv);
+
+	/* Prepare stop bytes with NOPs */
+	memset(insn_buf + MAX_INSN_SIZE, INSN_NOP, MAX_INSN_SIZE);
+
+	for (i = 0; i < iter_end; i++) {
+		if (generate_insn(insn_buf) <= 0)
+			break;
+
+		if (i < iter_start)	/* Skip to given iteration number */
+			continue;
+
+		/* Decode an instruction */
+		insn_init(&insn, insn_buf, x86_64);
+		insn_get_length(&insn);
+
+		if (insn.next_byte <= insn.kaddr ||
+		    insn.kaddr + MAX_INSN_SIZE < insn.next_byte) {
+			/* Access out-of-range memory */
+			dump_stream(stderr, "Error: Found an access violation", i, insn_buf, &insn);
+			errors++;
+		} else if (verbose && !insn_complete(&insn))
+			dump_stream(stdout, "Info: Found an undecodable input", i, insn_buf, &insn);
+		else if (verbose >= 2)
+			dump_insn(stdout, &insn);
+		insns++;
+	}
+
+	fprintf(stdout, "%s: decoded and checked %d %s instructions with %d errors (seed:0x%x)\n", (errors) ? "Failure" : "Success", insns, (input_file) ? "given" : "random", errors, seed);
+
+	return errors ? 1 : 0;
+}

drivers/oprofile/nmi_timer_int.c

+/**
+ * @file nmi_timer_int.c
+ *
+ * @remark Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * @author Robert Richter <robert.richter@amd.com>
+ */
+
+#include <linux/init.h>
+#include <linux/smp.h>
+#include <linux/errno.h>
+#include <linux/oprofile.h>
+#include <linux/perf_event.h>
+
+#ifdef CONFIG_OPROFILE_NMI_TIMER
+
+static DEFINE_PER_CPU(struct perf_event *, nmi_timer_events);
+static int ctr_running;
+
+static struct perf_event_attr nmi_timer_attr = {
+	.type           = PERF_TYPE_HARDWARE,
+	.config         = PERF_COUNT_HW_CPU_CYCLES,
+	.size           = sizeof(struct perf_event_attr),
+	.pinned         = 1,
+	.disabled       = 1,
+};
+
+static void nmi_timer_callback(struct perf_event *event,
+			       struct perf_sample_data *data,
+			       struct pt_regs *regs)
+{
+	event->hw.interrupts = 0;       /* don't throttle interrupts */
+	oprofile_add_sample(regs, 0);
+}
+
+static int nmi_timer_start_cpu(int cpu)
+{
+	struct perf_event *event = per_cpu(nmi_timer_events, cpu);
+
+	if (!event) {
+		event = perf_event_create_kernel_counter(&nmi_timer_attr, cpu, NULL,
+							 nmi_timer_callback, NULL);
+		if (IS_ERR(event))
+			return PTR_ERR(event);
+		per_cpu(nmi_timer_events, cpu) = event;
+	}
+
+	if (event && ctr_running)
+		perf_event_enable(event);
+
+	return 0;
+}
+
+static void nmi_timer_stop_cpu(int cpu)
+{
+	struct perf_event *event = per_cpu(nmi_timer_events, cpu);
+
+	if (event && ctr_running)
+		perf_event_disable(event);
+}
+
+static int nmi_timer_cpu_notifier(struct notifier_block *b, unsigned long action,
+				  void *data)
+{
+	int cpu = (unsigned long)data;
+	switch (action) {
+	case CPU_DOWN_FAILED:
+	case CPU_ONLINE:
+		nmi_timer_start_cpu(cpu);
+		break;
+	case CPU_DOWN_PREPARE:
+		nmi_timer_stop_cpu(cpu);
+		break;
+	}
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block nmi_timer_cpu_nb = {
+	.notifier_call = nmi_timer_cpu_notifier
+};
+
+static int nmi_timer_start(void)
+{
+	int cpu;
+
+	get_online_cpus();
+	ctr_running = 1;
+	for_each_online_cpu(cpu)
+		nmi_timer_start_cpu(cpu);
+	put_online_cpus();
+
+	return 0;
+}
+
+static void nmi_timer_stop(void)
+{
+	int cpu;
+
+	get_online_cpus();
+	for_each_online_cpu(cpu)
+		nmi_timer_stop_cpu(cpu);
+	ctr_running = 0;
+	put_online_cpus();
+}
+
+static void nmi_timer_shutdown(void)
+{
+	struct perf_event *event;
+	int cpu;
+
+	get_online_cpus();
+	unregister_cpu_notifier(&nmi_timer_cpu_nb);
+	for_each_possible_cpu(cpu) {
+		event = per_cpu(nmi_timer_events, cpu);
+		if (!event)
+			continue;
+		perf_event_disable(event);
+		per_cpu(nmi_timer_events, cpu) = NULL;
+		perf_event_release_kernel(event);
+	}
+
+	put_online_cpus();
+}
+
+static int nmi_timer_setup(void)
+{
+	int cpu, err;
+	u64 period;
+
+	/* clock cycles per tick: */
+	period = (u64)cpu_khz * 1000;
+	do_div(period, HZ);
+	nmi_timer_attr.sample_period = period;
+
+	get_online_cpus();
+	err = register_cpu_notifier(&nmi_timer_cpu_nb);
+	if (err)
+		goto out;
+	/* can't attach events to offline cpus: */
+	for_each_online_cpu(cpu) {
+		err = nmi_timer_start_cpu(cpu);
+		if (err)
+			break;
+	}
+	if (err)
+		nmi_timer_shutdown();
+out:
+	put_online_cpus();
+	return err;
+}
+
+int __init op_nmi_timer_init(struct oprofile_operations *ops)
+{
+	int err = 0;
+
+	err = nmi_timer_setup();
+	if (err)
+		return err;
+	nmi_timer_shutdown();		/* only check, don't alloc */
+
+	ops->create_files	= NULL;
+	ops->setup		= nmi_timer_setup;
+	ops->shutdown		= nmi_timer_shutdown;
+	ops->start		= nmi_timer_start;
+	ops->stop		= nmi_timer_stop;
+	ops->cpu_type		= "timer";
+
+	printk(KERN_INFO "oprofile: using NMI timer interrupt.\n");
+
+	return 0;
+}
+
+#endif

drivers/oprofile/oprof.c

@@ -246,37 +246,31 @@ static int __init oprofile_init(void)
 	int err;
 
 	/* always init architecture to setup backtrace support */
+	timer_mode = 0;
 	err = oprofile_arch_init(&oprofile_ops);
+	if (!err) {
+		if (!timer && !oprofilefs_register())
+			return 0;
+		oprofile_arch_exit();
+	}
 
-	timer_mode = err || timer;	/* fall back to timer mode on errors */
-	if (timer_mode) {
-		if (!err)
-			oprofile_arch_exit();
+	/* setup timer mode: */
+	timer_mode = 1;
+	/* no nmi timer mode if oprofile.timer is set */
+	if (timer || op_nmi_timer_init(&oprofile_ops)) {
 		err = oprofile_timer_init(&oprofile_ops);
 		if (err)
 			return err;
 	}
 
-	err = oprofilefs_register();
-	if (!err)
-		return 0;
-
-	/* failed */
-	if (timer_mode)
-		oprofile_timer_exit();
-	else
-		oprofile_arch_exit();
-
-	return err;
+	return oprofilefs_register();
 }
 
 
 static void __exit oprofile_exit(void)
 {
 	oprofilefs_unregister();
-	if (timer_mode)
-		oprofile_timer_exit();
-	else
+	if (!timer_mode)
 		oprofile_arch_exit();
 }
 

drivers/oprofile/oprof.h

@@ -36,6 +36,15 @@ struct dentry;
 void oprofile_create_files(struct super_block *sb, struct dentry *root);
 int oprofile_timer_init(struct oprofile_operations *ops);
 void oprofile_timer_exit(void);
+#ifdef CONFIG_OPROFILE_NMI_TIMER
+int op_nmi_timer_init(struct oprofile_operations *ops);
+#else
+static inline int op_nmi_timer_init(struct oprofile_operations *ops)
+{
+	return -ENODEV;
+}
+#endif
+
 
 int oprofile_set_ulong(unsigned long *addr, unsigned long val);
 int oprofile_set_timeout(unsigned long time);

drivers/oprofile/timer_int.c

@@ -97,24 +97,24 @@ static struct notifier_block __refdata oprofile_cpu_notifier = {
 	.notifier_call = oprofile_cpu_notify,
 };
 
-int oprofile_timer_init(struct oprofile_operations *ops)
+static int oprofile_hrtimer_setup(void)
 {
-	int rc;
-
-	rc = register_hotcpu_notifier(&oprofile_cpu_notifier);
-	if (rc)
-		return rc;
-	ops->create_files = NULL;
-	ops->setup = NULL;
-	ops->shutdown = NULL;
-	ops->start = oprofile_hrtimer_start;
-	ops->stop = oprofile_hrtimer_stop;
-	ops->cpu_type = "timer";
-	printk(KERN_INFO "oprofile: using timer interrupt.\n");
-	return 0;
+	return register_hotcpu_notifier(&oprofile_cpu_notifier);
 }
 
-void oprofile_timer_exit(void)
+static void oprofile_hrtimer_shutdown(void)
 {
 	unregister_hotcpu_notifier(&oprofile_cpu_notifier);
 }
+
+int oprofile_timer_init(struct oprofile_operations *ops)
+{
+	ops->create_files	= NULL;
+	ops->setup		= oprofile_hrtimer_setup;
+	ops->shutdown		= oprofile_hrtimer_shutdown;
+	ops->start		= oprofile_hrtimer_start;
+	ops->stop		= oprofile_hrtimer_stop;
+	ops->cpu_type		= "timer";
+	printk(KERN_INFO "oprofile: using timer interrupt.\n");
+	return 0;
+}

include/linux/bitops.h

@@ -22,8 +22,14 @@ extern unsigned long __sw_hweight64(__u64 w);
 #include <asm/bitops.h>
 
 #define for_each_set_bit(bit, addr, size) \
-	for ((bit) = find_first_bit((addr), (size)); \
-	     (bit) < (size); \
+	for ((bit) = find_first_bit((addr), (size));		\
+	     (bit) < (size);					\
+	     (bit) = find_next_bit((addr), (size), (bit) + 1))
+
+/* same as for_each_set_bit() but use bit as value to start with */
+#define for_each_set_bit_cont(bit, addr, size) \
+	for ((bit) = find_next_bit((addr), (size), (bit));	\
+	     (bit) < (size);					\
 	     (bit) = find_next_bit((addr), (size), (bit) + 1))
 
 static __inline__ int get_bitmask_order(unsigned int count)

include/linux/jump_label.h

@@ -3,6 +3,7 @@
 
 #include <linux/types.h>
 #include <linux/compiler.h>
+#include <linux/workqueue.h>
 
 #if defined(CC_HAVE_ASM_GOTO) && defined(CONFIG_JUMP_LABEL)
 
@@ -14,6 +15,12 @@ struct jump_label_key {
 #endif
 };
 
+struct jump_label_key_deferred {
+	struct jump_label_key key;
+	unsigned long timeout;
+	struct delayed_work work;
+};
+
 # include <asm/jump_label.h>
 # define HAVE_JUMP_LABEL
 #endif	/* CC_HAVE_ASM_GOTO && CONFIG_JUMP_LABEL */
@@ -51,8 +58,11 @@ extern void arch_jump_label_transform_static(struct jump_entry *entry,
 extern int jump_label_text_reserved(void *start, void *end);
 extern void jump_label_inc(struct jump_label_key *key);
 extern void jump_label_dec(struct jump_label_key *key);
+extern void jump_label_dec_deferred(struct jump_label_key_deferred *key);
 extern bool jump_label_enabled(struct jump_label_key *key);
 extern void jump_label_apply_nops(struct module *mod);
+extern void jump_label_rate_limit(struct jump_label_key_deferred *key,
+		unsigned long rl);
 
 #else  /* !HAVE_JUMP_LABEL */
 
@@ -68,6 +78,10 @@ static __always_inline void jump_label_init(void)
 {
 }
 
+struct jump_label_key_deferred {
+	struct jump_label_key  key;
+};
+
 static __always_inline bool static_branch(struct jump_label_key *key)
 {
 	if (unlikely(atomic_read(&key->enabled)))
@@ -85,6 +99,11 @@ static inline void jump_label_dec(struct jump_label_key *key)
 	atomic_dec(&key->enabled);
 }
 
+static inline void jump_label_dec_deferred(struct jump_label_key_deferred *key)
+{
+	jump_label_dec(&key->key);
+}
+
 static inline int jump_label_text_reserved(void *start, void *end)
 {
 	return 0;
@@ -102,6 +121,14 @@ static inline int jump_label_apply_nops(struct module *mod)
 {
 	return 0;
 }
+
+static inline void jump_label_rate_limit(struct jump_label_key_deferred *key,
+		unsigned long rl)
+{
+}
 #endif	/* HAVE_JUMP_LABEL */
 
+#define jump_label_key_enabled	((struct jump_label_key){ .enabled = ATOMIC_INIT(1), })
+#define jump_label_key_disabled	((struct jump_label_key){ .enabled = ATOMIC_INIT(0), })
+
 #endif	/* _LINUX_JUMP_LABEL_H */

include/linux/perf_event.h

@@ -890,6 +890,7 @@ struct perf_event_context {
 	int				nr_active;
 	int				is_active;
 	int				nr_stat;
+	int				nr_freq;
 	int				rotate_disable;
 	atomic_t			refcount;
 	struct task_struct		*task;
@@ -1063,12 +1064,12 @@ perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr)
 	}
 }
 
-extern struct jump_label_key perf_sched_events;
+extern struct jump_label_key_deferred perf_sched_events;
 
 static inline void perf_event_task_sched_in(struct task_struct *prev,
 					    struct task_struct *task)
 {
-	if (static_branch(&perf_sched_events))
+	if (static_branch(&perf_sched_events.key))
 		__perf_event_task_sched_in(prev, task);
 }
 
@@ -1077,7 +1078,7 @@ static inline void perf_event_task_sched_out(struct task_struct *prev,
 {
 	perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, NULL, 0);
 
-	if (static_branch(&perf_sched_events))
+	if (static_branch(&perf_sched_events.key))
 		__perf_event_task_sched_out(prev, next);
 }
 

kernel/events/Makefile

@@ -2,5 +2,5 @@ ifdef CONFIG_FUNCTION_TRACER
 CFLAGS_REMOVE_core.o = -pg
 endif
 
-obj-y := core.o ring_buffer.o
+obj-y := core.o ring_buffer.o callchain.o
 obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o

kernel/events/callchain.c

+/*
+ * Performance events callchain code, extracted from core.c:
+ *
+ *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
+ *  Copyright (C) 2008-2011 Red Hat, Inc., Ingo Molnar
+ *  Copyright (C) 2008-2011 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ *  Copyright    2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
+ *
+ * For licensing details see kernel-base/COPYING
+ */
+
+#include <linux/perf_event.h>
+#include <linux/slab.h>
+#include "internal.h"
+
+struct callchain_cpus_entries {
+	struct rcu_head			rcu_head;
+	struct perf_callchain_entry	*cpu_entries[0];
+};
+
+static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]);
+static atomic_t nr_callchain_events;
+static DEFINE_MUTEX(callchain_mutex);
+static struct callchain_cpus_entries *callchain_cpus_entries;
+
+
+__weak void perf_callchain_kernel(struct perf_callchain_entry *entry,
+				  struct pt_regs *regs)
+{
+}
+
+__weak void perf_callchain_user(struct perf_callchain_entry *entry,
+				struct pt_regs *regs)
+{
+}
+
+static void release_callchain_buffers_rcu(struct rcu_head *head)
+{
+	struct callchain_cpus_entries *entries;
+	int cpu;
+
+	entries = container_of(head, struct callchain_cpus_entries, rcu_head);
+
+	for_each_possible_cpu(cpu)
+		kfree(entries->cpu_entries[cpu]);
+
+	kfree(entries);
+}
+
+static void release_callchain_buffers(void)
+{
+	struct callchain_cpus_entries *entries;
+
+	entries = callchain_cpus_entries;
+	rcu_assign_pointer(callchain_cpus_entries, NULL);
+	call_rcu(&entries->rcu_head, release_callchain_buffers_rcu);
+}
+
+static int alloc_callchain_buffers(void)
+{
+	int cpu;
+	int size;
+	struct callchain_cpus_entries *entries;
+
+	/*
+	 * We can't use the percpu allocation API for data that can be
+	 * accessed from NMI. Use a temporary manual per cpu allocation
+	 * until that gets sorted out.
+	 */
+	size = offsetof(struct callchain_cpus_entries, cpu_entries[nr_cpu_ids]);
+
+	entries = kzalloc(size, GFP_KERNEL);
+	if (!entries)
+		return -ENOMEM;
+
+	size = sizeof(struct perf_callchain_entry) * PERF_NR_CONTEXTS;
+
+	for_each_possible_cpu(cpu) {
+		entries->cpu_entries[cpu] = kmalloc_node(size, GFP_KERNEL,
+							 cpu_to_node(cpu));
+		if (!entries->cpu_entries[cpu])
+			goto fail;
+	}
+
+	rcu_assign_pointer(callchain_cpus_entries, entries);
+
+	return 0;
+
+fail:
+	for_each_possible_cpu(cpu)
+		kfree(entries->cpu_entries[cpu]);
+	kfree(entries);
+
+	return -ENOMEM;
+}
+
+int get_callchain_buffers(void)
+{
+	int err = 0;
+	int count;
+
+	mutex_lock(&callchain_mutex);
+
+	count = atomic_inc_return(&nr_callchain_events);
+	if (WARN_ON_ONCE(count < 1)) {
+		err = -EINVAL;
+		goto exit;
+	}
+
+	if (count > 1) {
+		/* If the allocation failed, give up */
+		if (!callchain_cpus_entries)
+			err = -ENOMEM;
+		goto exit;
+	}
+
+	err = alloc_callchain_buffers();
+	if (err)
+		release_callchain_buffers();
+exit:
+	mutex_unlock(&callchain_mutex);
+
+	return err;
+}
+
+void put_callchain_buffers(void)
+{
+	if (atomic_dec_and_mutex_lock(&nr_callchain_events, &callchain_mutex)) {
+		release_callchain_buffers();
+		mutex_unlock(&callchain_mutex);
+	}
+}
+
+static struct perf_callchain_entry *get_callchain_entry(int *rctx)
+{
+	int cpu;
+	struct callchain_cpus_entries *entries;
+
+	*rctx = get_recursion_context(__get_cpu_var(callchain_recursion));
+	if (*rctx == -1)
+		return NULL;
+
+	entries = rcu_dereference(callchain_cpus_entries);
+	if (!entries)
+		return NULL;
+
+	cpu = smp_processor_id();
+
+	return &entries->cpu_entries[cpu][*rctx];
+}
+
+static void
+put_callchain_entry(int rctx)
+{
+	put_recursion_context(__get_cpu_var(callchain_recursion), rctx);
+}
+
+struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
+{
+	int rctx;
+	struct perf_callchain_entry *entry;
+
+
+	entry = get_callchain_entry(&rctx);
+	if (rctx == -1)
+		return NULL;
+
+	if (!entry)
+		goto exit_put;
+
+	entry->nr = 0;
+
+	if (!user_mode(regs)) {
+		perf_callchain_store(entry, PERF_CONTEXT_KERNEL);
+		perf_callchain_kernel(entry, regs);
+		if (current->mm)
+			regs = task_pt_regs(current);
+		else
+			regs = NULL;
+	}
+
+	if (regs) {
+		perf_callchain_store(entry, PERF_CONTEXT_USER);
+		perf_callchain_user(entry, regs);
+	}
+
+exit_put:
+	put_callchain_entry(rctx);
+
+	return entry;
+}

kernel/events/internal.h

 #ifndef _KERNEL_EVENTS_INTERNAL_H
 #define _KERNEL_EVENTS_INTERNAL_H
 
+#include <linux/hardirq.h>
+
+/* Buffer handling */
+
 #define RING_BUFFER_WRITABLE		0x01
 
 struct ring_buffer {
@@ -67,7 +71,7 @@ static inline int page_order(struct ring_buffer *rb)
 }
 #endif
 
-static unsigned long perf_data_size(struct ring_buffer *rb)
+static inline unsigned long perf_data_size(struct ring_buffer *rb)
 {
 	return rb->nr_pages << (PAGE_SHIFT + page_order(rb));
 }
@@ -96,4 +100,37 @@ __output_copy(struct perf_output_handle *handle,
 	} while (len);
 }
 
+/* Callchain handling */
+extern struct perf_callchain_entry *perf_callchain(struct pt_regs *regs);
+extern int get_callchain_buffers(void);
+extern void put_callchain_buffers(void);
+
+static inline int get_recursion_context(int *recursion)
+{
+	int rctx;
+
+	if (in_nmi())
+		rctx = 3;
+	else if (in_irq())
+		rctx = 2;
+	else if (in_softirq())
+		rctx = 1;
+	else
+		rctx = 0;
+
+	if (recursion[rctx])
+		return -1;
+
+	recursion[rctx]++;
+	barrier();
+
+	return rctx;
+}
+
+static inline void put_recursion_context(int *recursion, int rctx)
+{
+	barrier();
+	recursion[rctx]--;
+}
+
 #endif /* _KERNEL_EVENTS_INTERNAL_H */

kernel/jump_label.c

@@ -73,16 +73,47 @@ void jump_label_inc(struct jump_label_key *key)
 }
 EXPORT_SYMBOL_GPL(jump_label_inc);
 
-void jump_label_dec(struct jump_label_key *key)
+static void __jump_label_dec(struct jump_label_key *key,
+		unsigned long rate_limit, struct delayed_work *work)
 {
 	if (!atomic_dec_and_mutex_lock(&key->enabled, &jump_label_mutex))
 		return;
 
-	jump_label_update(key, JUMP_LABEL_DISABLE);
+	if (rate_limit) {
+		atomic_inc(&key->enabled);
+		schedule_delayed_work(work, rate_limit);
+	} else
+		jump_label_update(key, JUMP_LABEL_DISABLE);
+
 	jump_label_unlock();
 }
 EXPORT_SYMBOL_GPL(jump_label_dec);
 
+static void jump_label_update_timeout(struct work_struct *work)
+{
+	struct jump_label_key_deferred *key =
+		container_of(work, struct jump_label_key_deferred, work.work);
+	__jump_label_dec(&key->key, 0, NULL);
+}
+
+void jump_label_dec(struct jump_label_key *key)
+{
+	__jump_label_dec(key, 0, NULL);
+}
+
+void jump_label_dec_deferred(struct jump_label_key_deferred *key)
+{
+	__jump_label_dec(&key->key, key->timeout, &key->work);
+}
+
+
+void jump_label_rate_limit(struct jump_label_key_deferred *key,
+		unsigned long rl)
+{
+	key->timeout = rl;
+	INIT_DELAYED_WORK(&key->work, jump_label_update_timeout);
+}
+
 static int addr_conflict(struct jump_entry *entry, void *start, void *end)
 {
 	if (entry->code <= (unsigned long)end &&
@@ -113,7 +144,7 @@ static int __jump_label_text_reserved(struct jump_entry *iter_start,
  * running code can override this to make the non-live update case
  * cheaper.
  */
-void __weak arch_jump_label_transform_static(struct jump_entry *entry,
+void __weak __init_or_module arch_jump_label_transform_static(struct jump_entry *entry,
 					    enum jump_label_type type)
 {
 	arch_jump_label_transform(entry, type);	
@@ -219,8 +250,13 @@ void jump_label_apply_nops(struct module *mod)
 	if (iter_start == iter_stop)
 		return;
 
-	for (iter = iter_start; iter < iter_stop; iter++)
-		arch_jump_label_transform_static(iter, JUMP_LABEL_DISABLE);
+	for (iter = iter_start; iter < iter_stop; iter++) {
+		struct jump_label_key *iterk;
+
+		iterk = (struct jump_label_key *)(unsigned long)iter->key;
+		arch_jump_label_transform_static(iter, jump_label_enabled(iterk) ?
+				JUMP_LABEL_ENABLE : JUMP_LABEL_DISABLE);
+	}
 }
 
 static int jump_label_add_module(struct module *mod)
@@ -260,8 +296,7 @@ static int jump_label_add_module(struct module *mod)
 		key->next = jlm;
 
 		if (jump_label_enabled(key))
-			__jump_label_update(key, iter, iter_stop,
-					    JUMP_LABEL_ENABLE);
+			__jump_label_update(key, iter, iter_stop, JUMP_LABEL_ENABLE);
 	}
 
 	return 0;

kernel/lockdep.c

@@ -499,36 +499,32 @@ void get_usage_chars(struct lock_class *class, char usage[LOCK_USAGE_CHARS])
 	usage[i] = '\0';
 }
 
-static int __print_lock_name(struct lock_class *class)
+static void __print_lock_name(struct lock_class *class)
 {
 	char str[KSYM_NAME_LEN];
 	const char *name;
 
-	name = class->name;
-	if (!name)
-		name = __get_key_name(class->key, str);
-
-	return printk("%s", name);
-}
-
-static void print_lock_name(struct lock_class *class)
-{
-	char str[KSYM_NAME_LEN], usage[LOCK_USAGE_CHARS];
-	const char *name;
-
-	get_usage_chars(class, usage);
-
 	name = class->name;
 	if (!name) {
 		name = __get_key_name(class->key, str);
-		printk(" (%s", name);
+		printk("%s", name);
 	} else {
-		printk(" (%s", name);
+		printk("%s", name);
 		if (class->name_version > 1)
 			printk("#%d", class->name_version);
 		if (class->subclass)
 			printk("/%d", class->subclass);
 	}
+}
+
+static void print_lock_name(struct lock_class *class)
+{
+	char usage[LOCK_USAGE_CHARS];
+
+	get_usage_chars(class, usage);
+
+	printk(" (");
+	__print_lock_name(class);
 	printk("){%s}", usage);
 }
 

kernel/trace/trace.c

@@ -338,7 +338,8 @@ static DECLARE_WAIT_QUEUE_HEAD(trace_wait);
 /* trace_flags holds trace_options default values */
 unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
 	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME |
-	TRACE_ITER_GRAPH_TIME | TRACE_ITER_RECORD_CMD | TRACE_ITER_OVERWRITE;
+	TRACE_ITER_GRAPH_TIME | TRACE_ITER_RECORD_CMD | TRACE_ITER_OVERWRITE |
+	TRACE_ITER_IRQ_INFO;
 
 static int trace_stop_count;
 static DEFINE_RAW_SPINLOCK(tracing_start_lock);
@@ -426,6 +427,7 @@ static const char *trace_options[] = {
 	"record-cmd",
 	"overwrite",
 	"disable_on_free",
+	"irq-info",
 	NULL
 };
 
@@ -1843,6 +1845,33 @@ static void s_stop(struct seq_file *m, void *p)
 	trace_event_read_unlock();
 }
 
+static void
+get_total_entries(struct trace_array *tr, unsigned long *total, unsigned long *entries)
+{
+	unsigned long count;
+	int cpu;
+
+	*total = 0;
+	*entries = 0;
+
+	for_each_tracing_cpu(cpu) {
+		count = ring_buffer_entries_cpu(tr->buffer, cpu);
+		/*
+		 * If this buffer has skipped entries, then we hold all
+		 * entries for the trace and we need to ignore the
+		 * ones before the time stamp.
+		 */
+		if (tr->data[cpu]->skipped_entries) {
+			count -= tr->data[cpu]->skipped_entries;
+			/* total is the same as the entries */
+			*total += count;
+		} else
+			*total += count +
+				ring_buffer_overrun_cpu(tr->buffer, cpu);
+		*entries += count;
+	}
+}
+
 static void print_lat_help_header(struct seq_file *m)
 {
 	seq_puts(m, "#                  _------=> CPU#            \n");
@@ -1855,12 +1884,35 @@ static void print_lat_help_header(struct seq_file *m)
 	seq_puts(m, "#     \\   /      |||||  \\    |   /           \n");
 }
 
-static void print_func_help_header(struct seq_file *m)
+static void print_event_info(struct trace_array *tr, struct seq_file *m)
+{
+	unsigned long total;
+	unsigned long entries;
+
+	get_total_entries(tr, &total, &entries);
+	seq_printf(m, "# entries-in-buffer/entries-written: %lu/%lu   #P:%d\n",
+		   entries, total, num_online_cpus());
+	seq_puts(m, "#\n");
+}
+
+static void print_func_help_header(struct trace_array *tr, struct seq_file *m)
 {
-	seq_puts(m, "#           TASK-PID    CPU#    TIMESTAMP  FUNCTION\n");
+	print_event_info(tr, m);
+	seq_puts(m, "#           TASK-PID   CPU#      TIMESTAMP  FUNCTION\n");
 	seq_puts(m, "#              | |       |          |         |\n");
 }
 
+static void print_func_help_header_irq(struct trace_array *tr, struct seq_file *m)
+{
+	print_event_info(tr, m);
+	seq_puts(m, "#                              _-----=> irqs-off\n");
+	seq_puts(m, "#                             / _----=> need-resched\n");
+	seq_puts(m, "#                            | / _---=> hardirq/softirq\n");
+	seq_puts(m, "#                            || / _--=> preempt-depth\n");
+	seq_puts(m, "#                            ||| /     delay\n");
+	seq_puts(m, "#           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION\n");
+	seq_puts(m, "#              | |       |   ||||       |         |\n");
+}
 
 void
 print_trace_header(struct seq_file *m, struct trace_iterator *iter)
@@ -1869,32 +1921,14 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter)
 	struct trace_array *tr = iter->tr;
 	struct trace_array_cpu *data = tr->data[tr->cpu];
 	struct tracer *type = current_trace;
-	unsigned long entries = 0;
-	unsigned long total = 0;
-	unsigned long count;
+	unsigned long entries;
+	unsigned long total;
 	const char *name = "preemption";
-	int cpu;
 
 	if (type)
 		name = type->name;
 
-
-	for_each_tracing_cpu(cpu) {
-		count = ring_buffer_entries_cpu(tr->buffer, cpu);
-		/*
-		 * If this buffer has skipped entries, then we hold all
-		 * entries for the trace and we need to ignore the
-		 * ones before the time stamp.
-		 */
-		if (tr->data[cpu]->skipped_entries) {
-			count -= tr->data[cpu]->skipped_entries;
-			/* total is the same as the entries */
-			total += count;
-		} else
-			total += count +
-				ring_buffer_overrun_cpu(tr->buffer, cpu);
-		entries += count;
-	}
+	get_total_entries(tr, &total, &entries);
 
 	seq_printf(m, "# %s latency trace v1.1.5 on %s\n",
 		   name, UTS_RELEASE);
@@ -2140,6 +2174,21 @@ enum print_line_t print_trace_line(struct trace_iterator *iter)
 	return print_trace_fmt(iter);
 }
 
+void trace_latency_header(struct seq_file *m)
+{
+	struct trace_iterator *iter = m->private;
+
+	/* print nothing if the buffers are empty */
+	if (trace_empty(iter))
+		return;
+
+	if (iter->iter_flags & TRACE_FILE_LAT_FMT)
+		print_trace_header(m, iter);
+
+	if (!(trace_flags & TRACE_ITER_VERBOSE))
+		print_lat_help_header(m);
+}
+
 void trace_default_header(struct seq_file *m)
 {
 	struct trace_iterator *iter = m->private;
@@ -2155,8 +2204,12 @@ void trace_default_header(struct seq_file *m)
 		if (!(trace_flags & TRACE_ITER_VERBOSE))
 			print_lat_help_header(m);
 	} else {
-		if (!(trace_flags & TRACE_ITER_VERBOSE))
-			print_func_help_header(m);
+		if (!(trace_flags & TRACE_ITER_VERBOSE)) {
+			if (trace_flags & TRACE_ITER_IRQ_INFO)
+				print_func_help_header_irq(iter->tr, m);
+			else
+				print_func_help_header(iter->tr, m);
+		}
 	}
 }
 

kernel/trace/trace.h

@@ -370,6 +370,7 @@ void trace_graph_function(struct trace_array *tr,
 		    unsigned long ip,
 		    unsigned long parent_ip,
 		    unsigned long flags, int pc);
+void trace_latency_header(struct seq_file *m);
 void trace_default_header(struct seq_file *m);
 void print_trace_header(struct seq_file *m, struct trace_iterator *iter);
 int trace_empty(struct trace_iterator *iter);
@@ -654,6 +655,7 @@ enum trace_iterator_flags {
 	TRACE_ITER_RECORD_CMD		= 0x100000,
 	TRACE_ITER_OVERWRITE		= 0x200000,
 	TRACE_ITER_STOP_ON_FREE		= 0x400000,
+	TRACE_ITER_IRQ_INFO		= 0x800000,
 };
 
 /*

kernel/trace/trace_events_filter.c

@@ -27,6 +27,12 @@
 #include "trace.h"
 #include "trace_output.h"
 
+#define DEFAULT_SYS_FILTER_MESSAGE					\
+	"### global filter ###\n"					\
+	"# Use this to set filters for multiple events.\n"		\
+	"# Only events with the given fields will be affected.\n"	\
+	"# If no events are modified, an error message will be displayed here"
+
 enum filter_op_ids
 {
 	OP_OR,
@@ -646,7 +652,7 @@ void print_subsystem_event_filter(struct event_subsystem *system,
 	if (filter && filter->filter_string)
 		trace_seq_printf(s, "%s\n", filter->filter_string);
 	else
-		trace_seq_printf(s, "none\n");
+		trace_seq_printf(s, DEFAULT_SYS_FILTER_MESSAGE "\n");
 	mutex_unlock(&event_mutex);
 }
 
@@ -1838,7 +1844,10 @@ int apply_subsystem_event_filter(struct event_subsystem *system,
 	if (!filter)
 		goto out;
 
-	replace_filter_string(filter, filter_string);
+	/* System filters just show a default message */
+	kfree(filter->filter_string);
+	filter->filter_string = NULL;
+
 	/*
 	 * No event actually uses the system filter
 	 * we can free it without synchronize_sched().
@@ -1848,14 +1857,12 @@ int apply_subsystem_event_filter(struct event_subsystem *system,
 
 	parse_init(ps, filter_ops, filter_string);
 	err = filter_parse(ps);
-	if (err) {
-		append_filter_err(ps, system->filter);
-		goto out;
-	}
+	if (err)
+		goto err_filter;
 
 	err = replace_system_preds(system, ps, filter_string);
 	if (err)
-		append_filter_err(ps, system->filter);
+		goto err_filter;
 
 out:
 	filter_opstack_clear(ps);
@@ -1865,6 +1872,11 @@ int apply_subsystem_event_filter(struct event_subsystem *system,
 	mutex_unlock(&event_mutex);
 
 	return err;
+
+err_filter:
+	replace_filter_string(filter, filter_string);
+	append_filter_err(ps, system->filter);
+	goto out;
 }
 
 #ifdef CONFIG_PERF_EVENTS

kernel/trace/trace_irqsoff.c

@@ -280,9 +280,20 @@ static enum print_line_t irqsoff_print_line(struct trace_iterator *iter)
 }
 
 static void irqsoff_graph_return(struct ftrace_graph_ret *trace) { }
-static void irqsoff_print_header(struct seq_file *s) { }
 static void irqsoff_trace_open(struct trace_iterator *iter) { }
 static void irqsoff_trace_close(struct trace_iterator *iter) { }
+
+#ifdef CONFIG_FUNCTION_TRACER
+static void irqsoff_print_header(struct seq_file *s)
+{
+	trace_default_header(s);
+}
+#else
+static void irqsoff_print_header(struct seq_file *s)
+{
+	trace_latency_header(s);
+}
+#endif /* CONFIG_FUNCTION_TRACER */
 #endif /* CONFIG_FUNCTION_GRAPH_TRACER */
 
 /*

kernel/trace/trace_output.c

@@ -627,11 +627,23 @@ int trace_print_context(struct trace_iterator *iter)
 	unsigned long usec_rem = do_div(t, USEC_PER_SEC);
 	unsigned long secs = (unsigned long)t;
 	char comm[TASK_COMM_LEN];
+	int ret;
 
 	trace_find_cmdline(entry->pid, comm);
 
-	return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ",
-				comm, entry->pid, iter->cpu, secs, usec_rem);
+	ret = trace_seq_printf(s, "%16s-%-5d [%03d] ",
+			       comm, entry->pid, iter->cpu);
+	if (!ret)
+		return 0;
+
+	if (trace_flags & TRACE_ITER_IRQ_INFO) {
+		ret = trace_print_lat_fmt(s, entry);
+		if (!ret)
+			return 0;
+	}
+
+	return trace_seq_printf(s, " %5lu.%06lu: ",
+				secs, usec_rem);
 }
 
 int trace_print_lat_context(struct trace_iterator *iter)

kernel/trace/trace_sched_wakeup.c

@@ -280,9 +280,20 @@ static enum print_line_t wakeup_print_line(struct trace_iterator *iter)
 }
 
 static void wakeup_graph_return(struct ftrace_graph_ret *trace) { }
-static void wakeup_print_header(struct seq_file *s) { }
 static void wakeup_trace_open(struct trace_iterator *iter) { }
 static void wakeup_trace_close(struct trace_iterator *iter) { }
+
+#ifdef CONFIG_FUNCTION_TRACER
+static void wakeup_print_header(struct seq_file *s)
+{
+	trace_default_header(s);
+}
+#else
+static void wakeup_print_header(struct seq_file *s)
+{
+	trace_latency_header(s);
+}
+#endif /* CONFIG_FUNCTION_TRACER */
 #endif /* CONFIG_FUNCTION_GRAPH_TRACER */
 
 /*

tools/perf/Documentation/perf-annotate.txt

@@ -66,7 +66,7 @@ OPTIONS
 	used. This interfaces starts by centering on the line with more
 	samples, TAB/UNTAB cycles through the lines with more samples.
 
--c::
+-C::
 --cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can
 	be provided as a comma-separated list with no space: 0,1. Ranges of
 	CPUs are specified with -: 0-2. Default is to report samples on all

tools/perf/Documentation/perf-report.txt

@@ -39,7 +39,7 @@ OPTIONS
 -T::
 --threads::
 	Show per-thread event counters
--C::
+-c::
 --comms=::
 	Only consider symbols in these comms. CSV that understands
 	file://filename entries.
@@ -128,7 +128,7 @@ OPTIONS
 --symfs=<directory>::
         Look for files with symbols relative to this directory.
 
--c::
+-C::
 --cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can
 	be provided as a comma-separated list with no space: 0,1. Ranges of
 	CPUs are specified with -: 0-2. Default is to report samples on all

tools/perf/Documentation/perf-script.txt

@@ -182,12 +182,17 @@ OPTIONS
 --hide-call-graph::
         When printing symbols do not display call chain.
 
--c::
+-C::
 --cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can
 	be provided as a comma-separated list with no space: 0,1. Ranges of
 	CPUs are specified with -: 0-2. Default is to report samples on all
 	CPUs.
 
+-c::
+--comms=::
+	Only display events for these comms. CSV that understands
+	file://filename entries.
+
 -I::
 --show-info::
 	Display extended information about the perf.data file. This adds

tools/perf/Documentation/perf-test.txt

@@ -8,13 +8,19 @@ perf-test - Runs sanity tests.
 SYNOPSIS
 --------
 [verse]
-'perf test <options>'
+'perf test [<options>] [{list <test-name-fragment>|[<test-name-fragments>|<test-numbers>]}]'
 
 DESCRIPTION
 -----------
 This command does assorted sanity tests, initially through linked routines but
 also will look for a directory with more tests in the form of scripts.
 
+To get a list of available tests use 'perf test list', specifying a test name
+fragment will show all tests that have it.
+
+To run just specific tests, inform test name fragments or the numbers obtained
+from 'perf test list'.
+
 OPTIONS
 -------
 -v::

tools/perf/Makefile

@@ -278,6 +278,7 @@ LIB_H += util/strbuf.h
 LIB_H += util/strlist.h
 LIB_H += util/strfilter.h
 LIB_H += util/svghelper.h
+LIB_H += util/tool.h
 LIB_H += util/run-command.h
 LIB_H += util/sigchain.h
 LIB_H += util/symbol.h

tools/perf/builtin-annotate.c

@@ -27,32 +27,32 @@
 #include "util/sort.h"
 #include "util/hist.h"
 #include "util/session.h"
+#include "util/tool.h"
 
 #include <linux/bitmap.h>
 
-static char		const *input_name = "perf.data";
-
-static bool		force, use_tui, use_stdio;
-
-static bool		full_paths;
-
-static bool		print_line;
-
-static const char *sym_hist_filter;
-
-static const char	*cpu_list;
-static DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS);
+struct perf_annotate {
+	struct perf_tool tool;
+	char const *input_name;
+	bool	   force, use_tui, use_stdio;
+	bool	   full_paths;
+	bool	   print_line;
+	const char *sym_hist_filter;
+	const char *cpu_list;
+	DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS);
+};
 
-static int perf_evlist__add_sample(struct perf_evlist *evlist,
-				   struct perf_sample *sample,
-				   struct perf_evsel *evsel,
-				   struct addr_location *al)
+static int perf_evsel__add_sample(struct perf_evsel *evsel,
+				  struct perf_sample *sample,
+				  struct addr_location *al,
+				  struct perf_annotate *ann)
 {
 	struct hist_entry *he;
 	int ret;
 
-	if (sym_hist_filter != NULL &&
-	    (al->sym == NULL || strcmp(sym_hist_filter, al->sym->name) != 0)) {
+	if (ann->sym_hist_filter != NULL &&
+	    (al->sym == NULL ||
+	     strcmp(ann->sym_hist_filter, al->sym->name) != 0)) {
 		/* We're only interested in a symbol named sym_hist_filter */
 		if (al->sym != NULL) {
 			rb_erase(&al->sym->rb_node,
@@ -69,8 +69,7 @@ static int perf_evlist__add_sample(struct perf_evlist *evlist,
 	ret = 0;
 	if (he->ms.sym != NULL) {
 		struct annotation *notes = symbol__annotation(he->ms.sym);
-		if (notes->src == NULL &&
-		    symbol__alloc_hist(he->ms.sym, evlist->nr_entries) < 0)
+		if (notes->src == NULL && symbol__alloc_hist(he->ms.sym) < 0)
 			return -ENOMEM;
 
 		ret = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
@@ -81,25 +80,26 @@ static int perf_evlist__add_sample(struct perf_evlist *evlist,
 	return ret;
 }
 
-static int process_sample_event(union perf_event *event,
+static int process_sample_event(struct perf_tool *tool,
+				union perf_event *event,
 				struct perf_sample *sample,
 				struct perf_evsel *evsel,
-				struct perf_session *session)
+				struct machine *machine)
 {
+	struct perf_annotate *ann = container_of(tool, struct perf_annotate, tool);
 	struct addr_location al;
 
-	if (perf_event__preprocess_sample(event, session, &al, sample,
+	if (perf_event__preprocess_sample(event, machine, &al, sample,
 					  symbol__annotate_init) < 0) {
 		pr_warning("problem processing %d event, skipping it.\n",
 			   event->header.type);
 		return -1;
 	}
 
-	if (cpu_list && !test_bit(sample->cpu, cpu_bitmap))
+	if (ann->cpu_list && !test_bit(sample->cpu, ann->cpu_bitmap))
 		return 0;
 
-	if (!al.filtered &&
-	    perf_evlist__add_sample(session->evlist, sample, evsel, &al)) {
+	if (!al.filtered && perf_evsel__add_sample(evsel, sample, &al, ann)) {
 		pr_warning("problem incrementing symbol count, "
 			   "skipping event\n");
 		return -1;
@@ -108,14 +108,15 @@ static int process_sample_event(union perf_event *event,
 	return 0;
 }
 
-static int hist_entry__tty_annotate(struct hist_entry *he, int evidx)
+static int hist_entry__tty_annotate(struct hist_entry *he, int evidx,
+				    struct perf_annotate *ann)
 {
 	return symbol__tty_annotate(he->ms.sym, he->ms.map, evidx,
-				    print_line, full_paths, 0, 0);
+				    ann->print_line, ann->full_paths, 0, 0);
 }
 
 static void hists__find_annotations(struct hists *self, int evidx,
-				    int nr_events)
+				    struct perf_annotate *ann)
 {
 	struct rb_node *nd = rb_first(&self->entries), *next;
 	int key = K_RIGHT;
@@ -138,8 +139,7 @@ static void hists__find_annotations(struct hists *self, int evidx,
 		}
 
 		if (use_browser > 0) {
-			key = hist_entry__tui_annotate(he, evidx, nr_events,
-						       NULL, NULL, 0);
+			key = hist_entry__tui_annotate(he, evidx, NULL, NULL, 0);
 			switch (key) {
 			case K_RIGHT:
 				next = rb_next(nd);
@@ -154,7 +154,7 @@ static void hists__find_annotations(struct hists *self, int evidx,
 			if (next != NULL)
 				nd = next;
 		} else {
-			hist_entry__tty_annotate(he, evidx);
+			hist_entry__tty_annotate(he, evidx, ann);
 			nd = rb_next(nd);
 			/*
 			 * Since we have a hist_entry per IP for the same
@@ -167,33 +167,26 @@ static void hists__find_annotations(struct hists *self, int evidx,
 	}
 }
 
-static struct perf_event_ops event_ops = {
-	.sample	= process_sample_event,
-	.mmap	= perf_event__process_mmap,
-	.comm	= perf_event__process_comm,
-	.fork	= perf_event__process_task,
-	.ordered_samples = true,
-	.ordering_requires_timestamps = true,
-};
-
-static int __cmd_annotate(void)
+static int __cmd_annotate(struct perf_annotate *ann)
 {
 	int ret;
 	struct perf_session *session;
 	struct perf_evsel *pos;
 	u64 total_nr_samples;
 
-	session = perf_session__new(input_name, O_RDONLY, force, false, &event_ops);
+	session = perf_session__new(ann->input_name, O_RDONLY,
+				    ann->force, false, &ann->tool);
 	if (session == NULL)
 		return -ENOMEM;
 
-	if (cpu_list) {
-		ret = perf_session__cpu_bitmap(session, cpu_list, cpu_bitmap);
+	if (ann->cpu_list) {
+		ret = perf_session__cpu_bitmap(session, ann->cpu_list,
+					       ann->cpu_bitmap);
 		if (ret)
 			goto out_delete;
 	}
 
-	ret = perf_session__process_events(session, &event_ops);
+	ret = perf_session__process_events(session, &ann->tool);
 	if (ret)
 		goto out_delete;
 
@@ -217,13 +210,12 @@ static int __cmd_annotate(void)
 			total_nr_samples += nr_samples;
 			hists__collapse_resort(hists);
 			hists__output_resort(hists);
-			hists__find_annotations(hists, pos->idx,
-						session->evlist->nr_entries);
+			hists__find_annotations(hists, pos->idx, ann);
 		}
 	}
 
 	if (total_nr_samples == 0) {
-		ui__warning("The %s file has no samples!\n", input_name);
+		ui__warning("The %s file has no samples!\n", ann->input_name);
 		goto out_delete;
 	}
 out_delete:
@@ -247,29 +239,42 @@ static const char * const annotate_usage[] = {
 	NULL
 };
 
-static const struct option options[] = {
-	OPT_STRING('i', "input", &input_name, "file",
+int cmd_annotate(int argc, const char **argv, const char *prefix __used)
+{
+	struct perf_annotate annotate = {
+		.tool = {
+			.sample	= process_sample_event,
+			.mmap	= perf_event__process_mmap,
+			.comm	= perf_event__process_comm,
+			.fork	= perf_event__process_task,
+			.ordered_samples = true,
+			.ordering_requires_timestamps = true,
+		},
+		.input_name = "perf.data",
+	};
+	const struct option options[] = {
+	OPT_STRING('i', "input", &annotate.input_name, "file",
 		    "input file name"),
 	OPT_STRING('d', "dsos", &symbol_conf.dso_list_str, "dso[,dso...]",
 		   "only consider symbols in these dsos"),
-	OPT_STRING('s', "symbol", &sym_hist_filter, "symbol",
+	OPT_STRING('s', "symbol", &annotate.sym_hist_filter, "symbol",
 		    "symbol to annotate"),
-	OPT_BOOLEAN('f', "force", &force, "don't complain, do it"),
+	OPT_BOOLEAN('f', "force", &annotate.force, "don't complain, do it"),
 	OPT_INCR('v', "verbose", &verbose,
 		    "be more verbose (show symbol address, etc)"),
 	OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
 		    "dump raw trace in ASCII"),
-	OPT_BOOLEAN(0, "tui", &use_tui, "Use the TUI interface"),
-	OPT_BOOLEAN(0, "stdio", &use_stdio, "Use the stdio interface"),
+	OPT_BOOLEAN(0, "tui", &annotate.use_tui, "Use the TUI interface"),
+	OPT_BOOLEAN(0, "stdio", &annotate.use_stdio, "Use the stdio interface"),
 	OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
 		   "file", "vmlinux pathname"),
 	OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules,
 		    "load module symbols - WARNING: use only with -k and LIVE kernel"),
-	OPT_BOOLEAN('l', "print-line", &print_line,
+	OPT_BOOLEAN('l', "print-line", &annotate.print_line,
 		    "print matching source lines (may be slow)"),
-	OPT_BOOLEAN('P', "full-paths", &full_paths,
+	OPT_BOOLEAN('P', "full-paths", &annotate.full_paths,
 		    "Don't shorten the displayed pathnames"),
-	OPT_STRING('c', "cpu", &cpu_list, "cpu", "list of cpus to profile"),
+	OPT_STRING('C', "cpu", &annotate.cpu_list, "cpu", "list of cpus to profile"),
 	OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory",
 		   "Look for files with symbols relative to this directory"),
 	OPT_BOOLEAN(0, "source", &symbol_conf.annotate_src,
@@ -279,15 +284,13 @@ static const struct option options[] = {
 	OPT_STRING('M', "disassembler-style", &disassembler_style, "disassembler style",
 		   "Specify disassembler style (e.g. -M intel for intel syntax)"),
 	OPT_END()
-};
+	};
 
-int cmd_annotate(int argc, const char **argv, const char *prefix __used)
-{
 	argc = parse_options(argc, argv, options, annotate_usage, 0);
 
-	if (use_stdio)
+	if (annotate.use_stdio)
 		use_browser = 0;
-	else if (use_tui)
+	else if (annotate.use_tui)
 		use_browser = 1;
 
 	setup_browser(true);
@@ -308,7 +311,7 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used)
 		if (argc > 1)
 			usage_with_options(annotate_usage, options);
 
-		sym_hist_filter = argv[0];
+		annotate.sym_hist_filter = argv[0];
 	}
 
 	if (field_sep && *field_sep == '.') {
@@ -316,5 +319,5 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used)
 		return -1;
 	}
 
-	return __cmd_annotate();
+	return __cmd_annotate(&annotate);
 }

tools/perf/builtin-diff.c

@@ -9,7 +9,9 @@
 #include "util/debug.h"
 #include "util/event.h"
 #include "util/hist.h"
+#include "util/evsel.h"
 #include "util/session.h"
+#include "util/tool.h"
 #include "util/sort.h"
 #include "util/symbol.h"
 #include "util/util.h"
@@ -30,14 +32,15 @@ static int hists__add_entry(struct hists *self,
 	return -ENOMEM;
 }
 
-static int diff__process_sample_event(union perf_event *event,
+static int diff__process_sample_event(struct perf_tool *tool __used,
+				      union perf_event *event,
 				      struct perf_sample *sample,
 				      struct perf_evsel *evsel __used,
-				      struct perf_session *session)
+				      struct machine *machine)
 {
 	struct addr_location al;
 
-	if (perf_event__preprocess_sample(event, session, &al, sample, NULL) < 0) {
+	if (perf_event__preprocess_sample(event, machine, &al, sample, NULL) < 0) {
 		pr_warning("problem processing %d event, skipping it.\n",
 			   event->header.type);
 		return -1;
@@ -46,16 +49,16 @@ static int diff__process_sample_event(union perf_event *event,
 	if (al.filtered || al.sym == NULL)
 		return 0;
 
-	if (hists__add_entry(&session->hists, &al, sample->period)) {
+	if (hists__add_entry(&evsel->hists, &al, sample->period)) {
 		pr_warning("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
 
-	session->hists.stats.total_period += sample->period;
+	evsel->hists.stats.total_period += sample->period;
 	return 0;
 }
 
-static struct perf_event_ops event_ops = {
+static struct perf_tool perf_diff = {
 	.sample	= diff__process_sample_event,
 	.mmap	= perf_event__process_mmap,
 	.comm	= perf_event__process_comm,
@@ -145,13 +148,13 @@ static int __cmd_diff(void)
 	int ret, i;
 	struct perf_session *session[2];
 
-	session[0] = perf_session__new(input_old, O_RDONLY, force, false, &event_ops);
-	session[1] = perf_session__new(input_new, O_RDONLY, force, false, &event_ops);
+	session[0] = perf_session__new(input_old, O_RDONLY, force, false, &perf_diff);
+	session[1] = perf_session__new(input_new, O_RDONLY, force, false, &perf_diff);
 	if (session[0] == NULL || session[1] == NULL)
 		return -ENOMEM;
 
 	for (i = 0; i < 2; ++i) {
-		ret = perf_session__process_events(session[i], &event_ops);
+		ret = perf_session__process_events(session[i], &perf_diff);
 		if (ret)
 			goto out_delete;
 	}

tools/perf/builtin-inject.c

@@ -9,6 +9,7 @@
 
 #include "perf.h"
 #include "util/session.h"
+#include "util/tool.h"
 #include "util/debug.h"
 
 #include "util/parse-options.h"
@@ -16,8 +17,9 @@
 static char		const *input_name = "-";
 static bool		inject_build_ids;
 
-static int perf_event__repipe_synth(union perf_event *event,
-				    struct perf_session *session __used)
+static int perf_event__repipe_synth(struct perf_tool *tool __used,
+				    union perf_event *event,
+				    struct machine *machine __used)
 {
 	uint32_t size;
 	void *buf = event;
@@ -36,41 +38,70 @@ static int perf_event__repipe_synth(union perf_event *event,
 	return 0;
 }
 
-static int perf_event__repipe(union perf_event *event,
+static int perf_event__repipe_op2_synth(struct perf_tool *tool,
+					union perf_event *event,
+					struct perf_session *session __used)
+{
+	return perf_event__repipe_synth(tool, event, NULL);
+}
+
+static int perf_event__repipe_event_type_synth(struct perf_tool *tool,
+					       union perf_event *event)
+{
+	return perf_event__repipe_synth(tool, event, NULL);
+}
+
+static int perf_event__repipe_tracing_data_synth(union perf_event *event,
+						 struct perf_session *session __used)
+{
+	return perf_event__repipe_synth(NULL, event, NULL);
+}
+
+static int perf_event__repipe_attr(union perf_event *event,
+				   struct perf_evlist **pevlist __used)
+{
+	return perf_event__repipe_synth(NULL, event, NULL);
+}
+
+static int perf_event__repipe(struct perf_tool *tool,
+			      union perf_event *event,
 			      struct perf_sample *sample __used,
-			      struct perf_session *session)
+			      struct machine *machine)
 {
-	return perf_event__repipe_synth(event, session);
+	return perf_event__repipe_synth(tool, event, machine);
 }
 
-static int perf_event__repipe_sample(union perf_event *event,
+static int perf_event__repipe_sample(struct perf_tool *tool,
+				     union perf_event *event,
 			      struct perf_sample *sample __used,
 			      struct perf_evsel *evsel __used,
-			      struct perf_session *session)
+			      struct machine *machine)
 {
-	return perf_event__repipe_synth(event, session);
+	return perf_event__repipe_synth(tool, event, machine);
 }
 
-static int perf_event__repipe_mmap(union perf_event *event,
+static int perf_event__repipe_mmap(struct perf_tool *tool,
+				   union perf_event *event,
 				   struct perf_sample *sample,
-				   struct perf_session *session)
+				   struct machine *machine)
 {
 	int err;
 
-	err = perf_event__process_mmap(event, sample, session);
-	perf_event__repipe(event, sample, session);
+	err = perf_event__process_mmap(tool, event, sample, machine);
+	perf_event__repipe(tool, event, sample, machine);
 
 	return err;
 }
 
-static int perf_event__repipe_task(union perf_event *event,
+static int perf_event__repipe_task(struct perf_tool *tool,
+				   union perf_event *event,
 				   struct perf_sample *sample,
-				   struct perf_session *session)
+				   struct machine *machine)
 {
 	int err;
 
-	err = perf_event__process_task(event, sample, session);
-	perf_event__repipe(event, sample, session);
+	err = perf_event__process_task(tool, event, sample, machine);
+	perf_event__repipe(tool, event, sample, machine);
 
 	return err;
 }
@@ -80,7 +111,7 @@ static int perf_event__repipe_tracing_data(union perf_event *event,
 {
 	int err;
 
-	perf_event__repipe_synth(event, session);
+	perf_event__repipe_synth(NULL, event, NULL);
 	err = perf_event__process_tracing_data(event, session);
 
 	return err;
@@ -100,10 +131,10 @@ static int dso__read_build_id(struct dso *self)
 	return -1;
 }
 
-static int dso__inject_build_id(struct dso *self, struct perf_session *session)
+static int dso__inject_build_id(struct dso *self, struct perf_tool *tool,
+				struct machine *machine)
 {
 	u16 misc = PERF_RECORD_MISC_USER;
-	struct machine *machine;
 	int err;
 
 	if (dso__read_build_id(self) < 0) {
@@ -111,17 +142,11 @@ static int dso__inject_build_id(struct dso *self, struct perf_session *session)
 		return -1;
 	}
 
-	machine = perf_session__find_host_machine(session);
-	if (machine == NULL) {
-		pr_err("Can't find machine for session\n");
-		return -1;
-	}
-
 	if (self->kernel)
 		misc = PERF_RECORD_MISC_KERNEL;
 
-	err = perf_event__synthesize_build_id(self, misc, perf_event__repipe,
-					      machine, session);
+	err = perf_event__synthesize_build_id(tool, self, misc, perf_event__repipe,
+					      machine);
 	if (err) {
 		pr_err("Can't synthesize build_id event for %s\n", self->long_name);
 		return -1;
@@ -130,10 +155,11 @@ static int dso__inject_build_id(struct dso *self, struct perf_session *session)
 	return 0;
 }
 
-static int perf_event__inject_buildid(union perf_event *event,
+static int perf_event__inject_buildid(struct perf_tool *tool,
+				      union perf_event *event,
 				      struct perf_sample *sample,
 				      struct perf_evsel *evsel __used,
-				      struct perf_session *session)
+				      struct machine *machine)
 {
 	struct addr_location al;
 	struct thread *thread;
@@ -141,21 +167,21 @@ static int perf_event__inject_buildid(union perf_event *event,
 
 	cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
 
-	thread = perf_session__findnew(session, event->ip.pid);
+	thread = machine__findnew_thread(machine, event->ip.pid);
 	if (thread == NULL) {
 		pr_err("problem processing %d event, skipping it.\n",
 		       event->header.type);
 		goto repipe;
 	}
 
-	thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION,
-			      event->ip.pid, event->ip.ip, &al);
+	thread__find_addr_map(thread, machine, cpumode, MAP__FUNCTION,
+			      event->ip.ip, &al);
 
 	if (al.map != NULL) {
 		if (!al.map->dso->hit) {
 			al.map->dso->hit = 1;
 			if (map__load(al.map, NULL) >= 0) {
-				dso__inject_build_id(al.map->dso, session);
+				dso__inject_build_id(al.map->dso, tool, machine);
 				/*
 				 * If this fails, too bad, let the other side
 				 * account this as unresolved.
@@ -168,24 +194,24 @@ static int perf_event__inject_buildid(union perf_event *event,
 	}
 
 repipe:
-	perf_event__repipe(event, sample, session);
+	perf_event__repipe(tool, event, sample, machine);
 	return 0;
 }
 
-struct perf_event_ops inject_ops = {
+struct perf_tool perf_inject = {
 	.sample		= perf_event__repipe_sample,
 	.mmap		= perf_event__repipe,
 	.comm		= perf_event__repipe,
 	.fork		= perf_event__repipe,
 	.exit		= perf_event__repipe,
 	.lost		= perf_event__repipe,
-	.read		= perf_event__repipe,
+	.read		= perf_event__repipe_sample,
 	.throttle	= perf_event__repipe,
 	.unthrottle	= perf_event__repipe,
-	.attr		= perf_event__repipe_synth,
-	.event_type 	= perf_event__repipe_synth,
-	.tracing_data 	= perf_event__repipe_synth,
-	.build_id 	= perf_event__repipe_synth,
+	.attr		= perf_event__repipe_attr,
+	.event_type	= perf_event__repipe_event_type_synth,
+	.tracing_data	= perf_event__repipe_tracing_data_synth,
+	.build_id	= perf_event__repipe_op2_synth,
 };
 
 extern volatile int session_done;
@@ -203,17 +229,17 @@ static int __cmd_inject(void)
 	signal(SIGINT, sig_handler);
 
 	if (inject_build_ids) {
-		inject_ops.sample	= perf_event__inject_buildid;
-		inject_ops.mmap		= perf_event__repipe_mmap;
-		inject_ops.fork		= perf_event__repipe_task;
-		inject_ops.tracing_data	= perf_event__repipe_tracing_data;
+		perf_inject.sample	 = perf_event__inject_buildid;
+		perf_inject.mmap	 = perf_event__repipe_mmap;
+		perf_inject.fork	 = perf_event__repipe_task;
+		perf_inject.tracing_data = perf_event__repipe_tracing_data;
 	}
 
-	session = perf_session__new(input_name, O_RDONLY, false, true, &inject_ops);
+	session = perf_session__new(input_name, O_RDONLY, false, true, &perf_inject);
 	if (session == NULL)
 		return -ENOMEM;
 
-	ret = perf_session__process_events(session, &inject_ops);
+	ret = perf_session__process_events(session, &perf_inject);
 
 	perf_session__delete(session);
 

tools/perf/builtin-kmem.c

@@ -7,6 +7,7 @@
 #include "util/thread.h"
 #include "util/header.h"
 #include "util/session.h"
+#include "util/tool.h"
 
 #include "util/parse-options.h"
 #include "util/trace-event.h"
@@ -303,12 +304,13 @@ static void process_raw_event(union perf_event *raw_event __used, void *data,
 	}
 }
 
-static int process_sample_event(union perf_event *event,
+static int process_sample_event(struct perf_tool *tool __used,
+				union perf_event *event,
 				struct perf_sample *sample,
 				struct perf_evsel *evsel __used,
-				struct perf_session *session)
+				struct machine *machine)
 {
-	struct thread *thread = perf_session__findnew(session, event->ip.pid);
+	struct thread *thread = machine__findnew_thread(machine, event->ip.pid);
 
 	if (thread == NULL) {
 		pr_debug("problem processing %d event, skipping it.\n",
@@ -324,7 +326,7 @@ static int process_sample_event(union perf_event *event,
 	return 0;
 }
 
-static struct perf_event_ops event_ops = {
+static struct perf_tool perf_kmem = {
 	.sample			= process_sample_event,
 	.comm			= perf_event__process_comm,
 	.ordered_samples	= true,
@@ -483,7 +485,7 @@ static int __cmd_kmem(void)
 {
 	int err = -EINVAL;
 	struct perf_session *session = perf_session__new(input_name, O_RDONLY,
-							 0, false, &event_ops);
+							 0, false, &perf_kmem);
 	if (session == NULL)
 		return -ENOMEM;
 
@@ -494,7 +496,7 @@ static int __cmd_kmem(void)
 		goto out_delete;
 
 	setup_pager();
-	err = perf_session__process_events(session, &event_ops);
+	err = perf_session__process_events(session, &perf_kmem);
 	if (err != 0)
 		goto out_delete;
 	sort_result();

tools/perf/builtin-lock.c

@@ -12,6 +12,7 @@
 
 #include "util/debug.h"
 #include "util/session.h"
+#include "util/tool.h"
 
 #include <sys/types.h>
 #include <sys/prctl.h>
@@ -845,12 +846,13 @@ static void dump_info(void)
 		die("Unknown type of information\n");
 }
 
-static int process_sample_event(union perf_event *event,
+static int process_sample_event(struct perf_tool *tool __used,
+				union perf_event *event,
 				struct perf_sample *sample,
 				struct perf_evsel *evsel __used,
-				struct perf_session *s)
+				struct machine *machine)
 {
-	struct thread *thread = perf_session__findnew(s, sample->tid);
+	struct thread *thread = machine__findnew_thread(machine, sample->tid);
 
 	if (thread == NULL) {
 		pr_debug("problem processing %d event, skipping it.\n",
@@ -863,7 +865,7 @@ static int process_sample_event(union perf_event *event,
 	return 0;
 }
 
-static struct perf_event_ops eops = {
+static struct perf_tool eops = {
 	.sample			= process_sample_event,
 	.comm			= perf_event__process_comm,
 	.ordered_samples	= true,

tools/perf/builtin-probe.c

@@ -46,7 +46,6 @@
 
 #define DEFAULT_VAR_FILTER "!__k???tab_* & !__crc_*"
 #define DEFAULT_FUNC_FILTER "!_*"
-#define MAX_PATH_LEN 256
 
 /* Session management structure */
 static struct {

tools/perf/util/annotate.c

@@ -25,17 +25,17 @@ int symbol__annotate_init(struct map *map __used, struct symbol *sym)
 	return 0;
 }
 
-int symbol__alloc_hist(struct symbol *sym, int nevents)
+int symbol__alloc_hist(struct symbol *sym)
 {
 	struct annotation *notes = symbol__annotation(sym);
 	size_t sizeof_sym_hist = (sizeof(struct sym_hist) +
 				  (sym->end - sym->start) * sizeof(u64));
 
-	notes->src = zalloc(sizeof(*notes->src) + nevents * sizeof_sym_hist);
+	notes->src = zalloc(sizeof(*notes->src) + symbol_conf.nr_events * sizeof_sym_hist);
 	if (notes->src == NULL)
 		return -1;
 	notes->src->sizeof_sym_hist = sizeof_sym_hist;
-	notes->src->nr_histograms   = nevents;
+	notes->src->nr_histograms   = symbol_conf.nr_events;
 	INIT_LIST_HEAD(&notes->src->source);
 	return 0;
 }