perf fixes/updates

index 125fc1a..7626fa7 100644
@@ -880,6 +880,7 @@
 #define PV_970		0x0039
 #define PV_POWER5	0x003A
 #define PV_POWER5p	0x003B
+#define PV_POWER7	0x003F
 #define PV_970FX	0x003C
 #define PV_630		0x0040
 #define PV_630p	0x0041

index ab6f6be..97e0ae4 100644
@@ -1269,6 +1269,28 @@ unsigned long perf_instruction_pointer(struct pt_regs *regs)
 	return ip;
 }
 
+static bool pmc_overflow(unsigned long val)
+{
+	if ((int)val < 0)
+		return true;
+
+	/*
+	 * Events on POWER7 can roll back if a speculative event doesn't
+	 * eventually complete. Unfortunately in some rare cases they will
+	 * raise a performance monitor exception. We need to catch this to
+	 * ensure we reset the PMC. In all cases the PMC will be 256 or less
+	 * cycles from overflow.
+	 *
+	 * We only do this if the first pass fails to find any overflowing
+	 * PMCs because a user might set a period of less than 256 and we
+	 * don't want to mistakenly reset them.
+	 */
+	if (__is_processor(PV_POWER7) && ((0x80000000 - val) <= 256))
+		return true;
+
+	return false;
+}
+
 /*
  * Performance monitor interrupt stuff
  */

@@ -1316,7 +1338,7 @@ static void perf_event_interrupt(struct pt_regs *regs)
 			if (is_limited_pmc(i + 1))
 				continue;
 			val = read_pmc(i + 1);
-			if ((int)val < 0)
+			if (pmc_overflow(val))
 				write_pmc(i + 1, 0);
 		}
 	}

1,256 lines not on this page · raw message

113 lines not on this page · raw message

185 lines not on this page · raw message

index 2660418..e8dbe17 100644
@@ -178,8 +178,6 @@ struct cpu_hw_events {
  */
 #define INTEL_UEVENT_CONSTRAINT(c, n)	\
 	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)
-#define PEBS_EVENT_CONSTRAINT(c, n)	\
-	INTEL_UEVENT_CONSTRAINT(c, n)
 
 #define EVENT_CONSTRAINT_END		\
 	EVENT_CONSTRAINT(0, 0, 0)

134 lines not on this page · raw message

index 59f5544..b8ef8dd 100644
@@ -258,8 +258,10 @@ op_add_sample(struct oprofile_cpu_buffer *cpu_buf,
  */
 static int
 log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc,
-	   unsigned long backtrace, int is_kernel, unsigned long event)
+	   unsigned long backtrace, int is_kernel, unsigned long event,
+	   struct task_struct *task)
 {
+	struct task_struct *tsk = task ? task : current;
 	cpu_buf->sample_received++;
 
 	if (pc == ESCAPE_CODE) {

@@ -267,7 +269,7 @@ log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc,
 		return 0;
 	}
 
-	if (op_add_code(cpu_buf, backtrace, is_kernel, current))
+	if (op_add_code(cpu_buf, backtrace, is_kernel, tsk))
 		goto fail;
 
 	if (op_add_sample(cpu_buf, pc, event))

@@ -292,7 +294,8 @@ static inline void oprofile_end_trace(struct oprofile_cpu_buffer *cpu_buf)
 
 static inline void
 __oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
-			  unsigned long event, int is_kernel)
+			  unsigned long event, int is_kernel,
+			  struct task_struct *task)
 {
 	struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(op_cpu_buffer);
 	unsigned long backtrace = oprofile_backtrace_depth;

@@ -301,7 +304,7 @@ __oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
 	 * if log_sample() fail we can't backtrace since we lost the
 	 * source of this event
 	 */
-	if (!log_sample(cpu_buf, pc, backtrace, is_kernel, event))
+	if (!log_sample(cpu_buf, pc, backtrace, is_kernel, event, task))
 		/* failed */
 		return;
 

@@ -313,10 +316,17 @@ __oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
 	oprofile_end_trace(cpu_buf);
 }
 
+void oprofile_add_ext_hw_sample(unsigned long pc, struct pt_regs * const regs,
+				unsigned long event, int is_kernel,
+				struct task_struct *task)
+{
+	__oprofile_add_ext_sample(pc, regs, event, is_kernel, task);
+}
+
 void oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
 			     unsigned long event, int is_kernel)
 {
-	__oprofile_add_ext_sample(pc, regs, event, is_kernel);
+	__oprofile_add_ext_sample(pc, regs, event, is_kernel, NULL);
 }
 
 void oprofile_add_sample(struct pt_regs * const regs, unsigned long event)

@@ -332,7 +342,7 @@ void oprofile_add_sample(struct pt_regs * const regs, unsigned long event)
 		pc = ESCAPE_CODE; /* as this causes an early return. */
 	}
 
-	__oprofile_add_ext_sample(pc, regs, event, is_kernel);
+	__oprofile_add_ext_sample(pc, regs, event, is_kernel, NULL);
 }
 
 /*

@@ -403,7 +413,7 @@ int oprofile_write_commit(struct op_entry *entry)
 void oprofile_add_pc(unsigned long pc, int is_kernel, unsigned long event)
 {
 	struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(op_cpu_buffer);
-	log_sample(cpu_buf, pc, 0, is_kernel, event);
+	log_sample(cpu_buf, pc, 0, is_kernel, event, NULL);
 }
 
 void oprofile_add_trace(unsigned long pc)

index 0107251..3ef4462 100644
@@ -97,7 +97,7 @@ static struct notifier_block __refdata oprofile_cpu_notifier = {
 	.notifier_call = oprofile_cpu_notify,
 };
 
-int __init oprofile_timer_init(struct oprofile_operations *ops)
+int oprofile_timer_init(struct oprofile_operations *ops)
 {
 	int rc;
 

@@ -113,7 +113,7 @@ int __init oprofile_timer_init(struct oprofile_operations *ops)
 	return 0;
 }
 
-void __exit oprofile_timer_exit(void)
+void oprofile_timer_exit(void)
 {
 	unregister_hotcpu_notifier(&oprofile_cpu_notifier);
 }

index 1ca6411..7f5cfd3 100644
@@ -106,6 +106,13 @@ void oprofile_add_sample(struct pt_regs * const regs, unsigned long event);
 void oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
 				unsigned long event, int is_kernel);
 
+/**
+ * Add an hardware sample.
+ */
+void oprofile_add_ext_hw_sample(unsigned long pc, struct pt_regs * const regs,
+	unsigned long event, int is_kernel,
+	struct task_struct *task);
+
 /* Use this instead when the PC value is not from the regs. Doesn't
  * backtrace. */
 void oprofile_add_pc(unsigned long pc, int is_kernel, unsigned long event);

index 614615b..f495c01 100644
@@ -878,8 +878,8 @@ enum perf_event_context_type {
  * Used as a container for task events and CPU events as well:
  */
 struct perf_event_context {
-	enum perf_event_context_type	type;
 	struct pmu			*pmu;
+	enum perf_event_context_type	type;
 	/*
 	 * Protect the states of the events in the list,
 	 * nr_active, and the list:

index ed253aa..3472bb1 100644
@@ -5122,7 +5122,7 @@ static int perf_exclude_event(struct perf_event *event,
 			      struct pt_regs *regs)
 {
 	if (event->hw.state & PERF_HES_STOPPED)
-		return 0;
+		return 1;
 
 	if (regs) {
 		if (event->attr.exclude_user && user_mode(regs))

@@ -5478,6 +5478,8 @@ static int perf_tp_event_match(struct perf_event *event,
 				struct perf_sample_data *data,
 				struct pt_regs *regs)
 {
+	if (event->hw.state & PERF_HES_STOPPED)
+		return 0;
 	/*
 	 * All tracepoints are from kernel-space.
 	 */

@@ -6720,17 +6722,20 @@ __perf_event_exit_task(struct perf_event *child_event,
 			 struct perf_event_context *child_ctx,
 			 struct task_struct *child)
 {
-	struct perf_event *parent_event;
+	if (child_event->parent) {
+		raw_spin_lock_irq(&child_ctx->lock);
+		perf_group_detach(child_event);
+		raw_spin_unlock_irq(&child_ctx->lock);
+	}
 
 	perf_remove_from_context(child_event);
 
-	parent_event = child_event->parent;
 	/*
-	 * It can happen that parent exits first, and has events
+	 * It can happen that the parent exits first, and has events
 	 * that are still around due to the child reference. These
-	 * events need to be zapped - but otherwise linger.
+	 * events need to be zapped.
 	 */
-	if (parent_event) {
+	if (child_event->parent) {
 		sync_child_event(child_event, child);
 		free_event(child_event);
 	}

index 3249b4f..8008ddc 100644
@@ -391,8 +391,8 @@ static int process_ops(struct filter_pred *preds,
 		       struct filter_pred *op, void *rec)
 {
 	struct filter_pred *pred;
+	int match = 0;
 	int type;
-	int match;
 	int i;
 
 	/*

54 lines not on this page · raw message

index 2e93f99..7a2a79d 100644
@@ -202,9 +202,20 @@ static struct thread_stat *thread_stat_findnew_first(u32 tid)
 SINGLE_KEY(nr_acquired)
 SINGLE_KEY(nr_contended)
 SINGLE_KEY(wait_time_total)
-SINGLE_KEY(wait_time_min)
 SINGLE_KEY(wait_time_max)
 
+static int lock_stat_key_wait_time_min(struct lock_stat *one,
+					struct lock_stat *two)
+{
+	u64 s1 = one->wait_time_min;
+	u64 s2 = two->wait_time_min;
+	if (s1 == ULLONG_MAX)
+		s1 = 0;
+	if (s2 == ULLONG_MAX)
+		s2 = 0;
+	return s1 > s2;
+}
+
 struct lock_key {
 	/*
 	 * name: the value for specify by user

355 lines not on this page · raw message

index 21c0252..e2109f9 100644
@@ -333,6 +333,12 @@ static int run_perf_stat(int argc __used, const char **argv)
 		}
 	}
 
+	if (perf_evlist__set_filters(evsel_list)) {
+		error("failed to set filter with %d (%s)\n", errno,
+			strerror(errno));
+		return -1;
+	}
+
 	/*
 	 * Enable counters and exec the command:
 	 */

@@ -634,6 +640,8 @@ static const struct option options[] = {
 	OPT_CALLBACK('e', "event", &evsel_list, "event",
 		     "event selector. use 'perf list' to list available events",
 		     parse_events),
+	OPT_CALLBACK(0, "filter", &evsel_list, "filter",
+		     "event filter", parse_filter),
 	OPT_BOOLEAN('i', "no-inherit", &no_inherit,
 		    "child tasks do not inherit counters"),
 	OPT_INTEGER('p', "pid", &target_pid,

index 80c9e06..70f1075 100644
@@ -152,7 +152,7 @@ static int parse_source(struct sym_entry *syme)
 	/*
 	 * We can't annotate with just /proc/kallsyms
 	 */
-	if (map->dso->origin == DSO__ORIG_KERNEL) {
+	if (map->dso->symtab_type == SYMTAB__KALLSYMS) {
 		pr_err("Can't annotate %s: No vmlinux file was found in the "
 		       "path\n", sym->name);
 		sleep(1);

@@ -515,24 +515,25 @@ static void handle_keypress(struct perf_session *session, int c)
 			break;
 		case 'E':
 			if (top.evlist->nr_entries > 1) {
+				int counter;
 				fprintf(stderr, "\nAvailable events:");
 
 				list_for_each_entry(top.sym_evsel, &top.evlist->entries, node)
 					fprintf(stderr, "\n\t%d %s", top.sym_evsel->idx, event_name(top.sym_evsel));
 
-				prompt_integer(&top.sym_counter, "Enter details event counter");
+				prompt_integer(&counter, "Enter details event counter");
 
-				if (top.sym_counter >= top.evlist->nr_entries) {
+				if (counter >= top.evlist->nr_entries) {
 					top.sym_evsel = list_entry(top.evlist->entries.next, struct perf_evsel, node);
-					top.sym_counter = 0;
 					fprintf(stderr, "Sorry, no such event, using %s.\n", event_name(top.sym_evsel));
 					sleep(1);
 					break;
 				}
 				list_for_each_entry(top.sym_evsel, &top.evlist->entries, node)
-					if (top.sym_evsel->idx == top.sym_counter)
+					if (top.sym_evsel->idx == counter)
 						break;
-			} else top.sym_counter = 0;
+			} else
+				top.sym_evsel = list_entry(top.evlist->entries.next, struct perf_evsel, node);
 			break;
 		case 'f':
 			prompt_integer(&top.count_filter, "Enter display event count filter");

@@ -675,7 +676,7 @@ static int symbol_filter(struct map *map, struct symbol *sym)
 
 	for (i = 0; skip_symbols[i]; i++) {
 		if (!strcmp(skip_symbols[i], name)) {
-			syme->skip = 1;
+			sym->ignore = true;
 			break;
 		}
 	}

@@ -768,7 +769,7 @@ static void perf_event__process_sample(const union perf_event *event,
 			struct symbol *sym = sym_entry__symbol(top.sym_filter_entry);
 
 			pr_err("Can't annotate %s", sym->name);
-			if (top.sym_filter_entry->map->dso->origin == DSO__ORIG_KERNEL) {
+			if (top.sym_filter_entry->map->dso->symtab_type == SYMTAB__KALLSYMS) {
 				pr_err(": No vmlinux file was found in the path:\n");
 				machine__fprintf_vmlinux_path(machine, stderr);
 			} else

@@ -778,10 +779,9 @@ static void perf_event__process_sample(const union perf_event *event,
 	}
 
 	syme = symbol__priv(al.sym);
-	if (!syme->skip) {
+	if (!al.sym->ignore) {
 		struct perf_evsel *evsel;
 
-		syme->origin = origin;
 		evsel = perf_evlist__id2evsel(top.evlist, sample->id);
 		assert(evsel != NULL);
 		syme->count[evsel->idx]++;

index c7798c7..4702e24 100644
@@ -19,6 +19,7 @@ extern int cmd_bench(int argc, const char **argv, const char *prefix);
 extern int cmd_buildid_cache(int argc, const char **argv, const char *prefix);
 extern int cmd_buildid_list(int argc, const char **argv, const char *prefix);
 extern int cmd_diff(int argc, const char **argv, const char *prefix);
+extern int cmd_evlist(int argc, const char **argv, const char *prefix);
 extern int cmd_help(int argc, const char **argv, const char *prefix);
 extern int cmd_sched(int argc, const char **argv, const char *prefix);
 extern int cmd_list(int argc, const char **argv, const char *prefix);

index 595d0f4..ec635b7 100644
@@ -313,6 +313,7 @@ static void handle_internal_command(int argc, const char **argv)
 		{ "buildid-cache", cmd_buildid_cache, 0 },
 		{ "buildid-list", cmd_buildid_list, 0 },
 		{ "diff",	cmd_diff,	0 },
+		{ "evlist",	cmd_evlist,	0 },
 		{ "help",	cmd_help,	0 },
 		{ "list",	cmd_list,	0 },
 		{ "record",	cmd_record,	0 },

index 0d0830c..e01af2b 100644
@@ -294,7 +294,7 @@ fallback:
 		free_filename = false;
 	}
 
-	if (dso->origin == DSO__ORIG_KERNEL) {
+	if (dso->symtab_type == SYMTAB__KALLSYMS) {
 		char bf[BUILD_ID_SIZE * 2 + 16] = " with build id ";
 		char *build_id_msg = NULL;
 

index 54a7e26..952b4ae 100644
@@ -263,6 +263,28 @@ static char *event_cache_name(u8 cache_type, u8 cache_op, u8 cache_result)
 	return name;
 }
 
+const char *event_type(int type)
+{
+	switch (type) {
+	case PERF_TYPE_HARDWARE:
+		return "hardware";
+
+	case PERF_TYPE_SOFTWARE:
+		return "software";
+
+	case PERF_TYPE_TRACEPOINT:
+		return "tracepoint";
+
+	case PERF_TYPE_HW_CACHE:
+		return "hardware-cache";
+
+	default:
+		break;
+	}
+
+	return "unknown";
+}
+
 const char *event_name(struct perf_evsel *evsel)
 {
 	u64 config = evsel->attr.config;

index 212f88e..746d3fc 100644
@@ -20,6 +20,7 @@ struct tracepoint_path {
 extern struct tracepoint_path *tracepoint_id_to_path(u64 config);
 extern bool have_tracepoints(struct list_head *evlist);
 
+const char *event_type(int type);
 const char *event_name(struct perf_evsel *event);
 extern const char *__event_name(int type, u64 config);
 

index 9368081..6214272 100644
@@ -245,9 +245,10 @@ static inline struct event *find_cache_event(int type)
 	return event;
 }
 
-static void perl_process_event(int cpu, void *data,
-			       int size __unused,
-			       unsigned long long nsecs, char *comm)
+static void perl_process_event(union perf_event *pevent __unused,
+			       struct perf_sample *sample,
+			       struct perf_session *session __unused,
+			       struct thread *thread)
 {
 	struct format_field *field;
 	static char handler[256];

@@ -256,6 +257,10 @@ static void perl_process_event(int cpu, void *data,
 	struct event *event;
 	int type;
 	int pid;
+	int cpu = sample->cpu;
+	void *data = sample->raw_data;
+	unsigned long long nsecs = sample->time;
+	char *comm = thread->comm;
 
 	dSP;
 

index 2040b85..1b85d60 100644
@@ -204,9 +204,10 @@ static inline struct event *find_cache_event(int type)
 	return event;
 }
 
-static void python_process_event(int cpu, void *data,
-				 int size __unused,
-				 unsigned long long nsecs, char *comm)
+static void python_process_event(union perf_event *pevent __unused,
+				 struct perf_sample *sample,
+				 struct perf_session *session __unused,
+				 struct thread *thread)
 {
 	PyObject *handler, *retval, *context, *t, *obj, *dict = NULL;
 	static char handler_name[256];

@@ -217,6 +218,10 @@ static void python_process_event(int cpu, void *data,
 	unsigned n = 0;
 	int type;
 	int pid;
+	int cpu = sample->cpu;
+	void *data = sample->raw_data;
+	unsigned long long nsecs = sample->time;
+	char *comm = thread->comm;
 
 	t = PyTuple_New(MAX_FIELDS);
 	if (!t)

64 lines not on this page · raw message

index b5b148b..0b3c9af 100644
@@ -159,4 +159,8 @@ static inline int perf_session__parse_sample(struct perf_session *session,
 					session->sample_id_all, sample);
 }
 
+void perf_session__print_symbols(union perf_event *event,
+				 struct perf_sample *sample,
+				 struct perf_session *session);
+
 #endif /* __PERF_SESSION_H */

144 lines not on this page · raw message

index 4d7ed09..713b0b4 100644
@@ -48,12 +48,17 @@ char *strxfrchar(char *s, char from, char to);
 
 #define BUILD_ID_SIZE 20
 
+/** struct symbol - symtab entry
+ *
+ * @ignore - resolvable but tools ignore it (e.g. idle routines)
+ */
 struct symbol {
 	struct rb_node	rb_node;
 	u64		start;
 	u64		end;
 	u16		namelen;
 	u8		binding;
+	bool		ignore;
 	char		name[0];
 };
 

@@ -137,7 +142,7 @@ struct dso {
 	u8		 annotate_warned:1;
 	u8		 sname_alloc:1;
 	u8		 lname_alloc:1;
-	unsigned char	 origin;
+	unsigned char	 symtab_type;
 	u8		 sorted_by_name;
 	u8		 loaded;
 	u8		 build_id[BUILD_ID_SIZE];

@@ -188,18 +193,18 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp);
 size_t dso__fprintf_symbols_by_name(struct dso *self, enum map_type type, FILE *fp);
 size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp);
 
-enum dso_origin {
-	DSO__ORIG_KERNEL = 0,
-	DSO__ORIG_GUEST_KERNEL,
-	DSO__ORIG_JAVA_JIT,
-	DSO__ORIG_BUILD_ID_CACHE,
-	DSO__ORIG_FEDORA,
-	DSO__ORIG_UBUNTU,
-	DSO__ORIG_BUILDID,
-	DSO__ORIG_DSO,
-	DSO__ORIG_GUEST_KMODULE,
-	DSO__ORIG_KMODULE,
-	DSO__ORIG_NOT_FOUND,
+enum symtab_type {
+	SYMTAB__KALLSYMS = 0,
+	SYMTAB__GUEST_KALLSYMS,
+	SYMTAB__JAVA_JIT,
+	SYMTAB__BUILD_ID_CACHE,
+	SYMTAB__FEDORA_DEBUGINFO,
+	SYMTAB__UBUNTU_DEBUGINFO,
+	SYMTAB__BUILDID_DEBUGINFO,
+	SYMTAB__SYSTEM_PATH_DSO,
+	SYMTAB__GUEST_KMODULE,
+	SYMTAB__SYSTEM_PATH_KMODULE,
+	SYMTAB__NOT_FOUND,
 };
 
 char dso__symtab_origin(const struct dso *self);

index 75cfe4d..a11f607 100644
@@ -171,7 +171,7 @@ float perf_top__decay_samples(struct perf_top *top, struct rb_root *root)
 {
 	struct sym_entry *syme, *n;
 	float sum_ksamples = 0.0;
-	int snap = !top->display_weighted ? top->sym_counter : 0, j;
+	int snap = !top->display_weighted ? top->sym_evsel->idx : 0, j;
 
 	/* Sort the active symbols */
 	pthread_mutex_lock(&top->active_symbols_lock);

@@ -184,9 +184,9 @@ float perf_top__decay_samples(struct perf_top *top, struct rb_root *root)
 		if (syme->snap_count != 0) {
 
 			if ((top->hide_user_symbols &&
-			     syme->origin == PERF_RECORD_MISC_USER) ||
+			     syme->map->dso->kernel == DSO_TYPE_USER) ||
 			    (top->hide_kernel_symbols &&
-			     syme->origin == PERF_RECORD_MISC_KERNEL)) {
+			     syme->map->dso->kernel == DSO_TYPE_KERNEL)) {
 				perf_top__remove_active_sym(top, syme);
 				continue;
 			}

index 96d1cb7..bfbf95b 100644
@@ -16,8 +16,6 @@ struct sym_entry {
 	struct list_head	node;
 	unsigned long		snap_count;
 	double			weight;
-	int			skip;
-	u8			origin;
 	struct map		*map;
 	unsigned long		count[0];
 };

@@ -41,7 +39,7 @@ struct perf_top {
 	u64		   exact_samples;
 	u64		   guest_us_samples, guest_kernel_samples;
 	int		   print_entries, count_filter, delay_secs;
-	int		   display_weighted, freq, rb_entries, sym_counter;
+	int		   display_weighted, freq, rb_entries;
 	pid_t		   target_pid, target_tid;
 	bool		   hide_kernel_symbols, hide_user_symbols, zero;
 	const char	   *cpu_list;

76 lines not on this page · raw message

index f7af2fc..66f4b78 100644

Lines of this file are not shown. View the raw message.

index b5f12ca..b04da57 100644

Lines of this file are not shown. View the raw message.