Merge branch 'perf/core' into auto-latest

23 files changed, 341 insertions, 148 deletions

diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
index 020d871c7934..34e255fc3e2f 100644
--- a/tools/perf/Documentation/perf-record.txt
+++ b/tools/perf/Documentation/perf-record.txt
@@ -69,8 +69,8 @@ OPTIONS
 	Output file name.
 
 -i::
---inherit::
-	Child tasks inherit counters.
+--no-inherit::
+	Child tasks do not inherit counters.
 -F::
 --freq=::
 	Profile at this frequency.
diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index 484080dd5b6f..2cab8e8c33d0 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -31,8 +31,8 @@ OPTIONS
 	 hexadecimal event descriptor.
 
 -i::
---inherit::
-        child tasks inherit counters
+--no-inherit::
+        child tasks do not inherit counters
 -p::
 --pid=<pid>::
         stat events on existing pid
diff --git a/tools/perf/Documentation/perf-trace-python.txt b/tools/perf/Documentation/perf-trace-python.txt
index 864aac283a7b..693be804dd3d 100644
--- a/tools/perf/Documentation/perf-trace-python.txt
+++ b/tools/perf/Documentation/perf-trace-python.txt
@@ -182,7 +182,7 @@ mean either that the record step recorded event types that it wasn't
 really interested in, or the script was run against a trace file that
 doesn't correspond to the script.
 
-The script generated by -g option option simply prints a line for each
+The script generated by -g option simply prints a line for each
 event found in the trace stream i.e. it basically just dumps the event
 and its parameter values to stdout.  The print_header() function is
 simply a utility function used for that purpose.  Let's rename the
@@ -582,7 +582,7 @@ files:
   flag_str(event_name, field_name, field_value) - returns the string represention corresponding to field_value for the flag field field_name of event event_name
   symbol_str(event_name, field_name, field_value) - returns the string represention corresponding to field_value for the symbolic field field_name of event event_name
 
-The *autodict* function returns a special special kind of Python
+The *autodict* function returns a special kind of Python
 dictionary that implements Perl's 'autovivifying' hashes in Python
 i.e. with autovivifying hashes, you can assign nested hash values
 without having to go to the trouble of creating intermediate levels if
diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 9c4dc30cdc13..a9281cca4114 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -416,6 +416,7 @@ LIB_H += util/thread.h
 LIB_H += util/trace-event.h
 LIB_H += util/probe-finder.h
 LIB_H += util/probe-event.h
+LIB_H += util/pstack.h
 LIB_H += util/cpumap.h
 
 LIB_OBJS += $(OUTPUT)util/abspath.o
@@ -451,6 +452,7 @@ LIB_OBJS += $(OUTPUT)util/callchain.o
 LIB_OBJS += $(OUTPUT)util/values.o
 LIB_OBJS += $(OUTPUT)util/debug.o
 LIB_OBJS += $(OUTPUT)util/map.o
+LIB_OBJS += $(OUTPUT)util/pstack.o
 LIB_OBJS += $(OUTPUT)util/session.o
 LIB_OBJS += $(OUTPUT)util/thread.o
 LIB_OBJS += $(OUTPUT)util/trace-event-parse.o
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index fd1b786c8f35..77bcc9b130f5 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -365,7 +365,7 @@ static int __cmd_annotate(void)
 		goto out_delete;
 
 	if (dump_trace) {
-		event__print_totals();
+		perf_session__fprintf_nr_events(session, stdout);
 		goto out_delete;
 	}
 
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 3a95a0260a5b..a6e2fdc7a04e 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -23,9 +23,9 @@ static bool  force;
 static bool show_displacement;
 
 static int hists__add_entry(struct hists *self,
-			    struct addr_location *al, u64 count)
+			    struct addr_location *al, u64 period)
 {
-	if (__hists__add_entry(self, al, NULL, count) != NULL)
+	if (__hists__add_entry(self, al, NULL, period) != NULL)
 		return 0;
 	return -ENOMEM;
 }
@@ -50,11 +50,11 @@ static int diff__process_sample_event(event_t *event, struct perf_session *sessi
 	event__parse_sample(event, session->sample_type, &data);
 
 	if (hists__add_entry(&session->hists, &al, data.period)) {
-		pr_warning("problem incrementing symbol count, skipping event\n");
+		pr_warning("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
 
-	session->hists.stats.total += data.period;
+	session->hists.stats.total_period += data.period;
 	return 0;
 }
 
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 6b77b285fe10..0f467cf7aa72 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -54,7 +54,7 @@ static pid_t			target_tid			=     -1;
 static pid_t			*all_tids			=      NULL;
 static int			thread_num			=      0;
 static pid_t			child_pid			=     -1;
-static bool			inherit				=   true;
+static bool			no_inherit			=  false;
 static enum write_mode_t	write_mode			= WRITE_FORCE;
 static bool			call_graph			=  false;
 static bool			inherit_stat			=  false;
@@ -298,8 +298,8 @@ static void create_counter(int counter, int cpu)
 
 	attr->mmap		= track;
 	attr->comm		= track;
-	attr->inherit		= inherit;
-	if (target_pid == -1 && !system_wide) {
+	attr->inherit		= !no_inherit;
+	if (target_pid == -1 && target_tid == -1 && !system_wide) {
 		attr->disabled = 1;
 		attr->enable_on_exec = 1;
 	}
@@ -641,7 +641,7 @@ static int __cmd_record(int argc, const char **argv)
 		close(child_ready_pipe[0]);
 	}
 
-	if ((!system_wide && !inherit) || profile_cpu != -1) {
+	if ((!system_wide && no_inherit) || profile_cpu != -1) {
 		open_counters(profile_cpu);
 	} else {
 		nr_cpus = read_cpu_map();
@@ -821,8 +821,8 @@ static const struct option options[] = {
 		    "event period to sample"),
 	OPT_STRING('o', "output", &output_name, "file",
 		    "output file name"),
-	OPT_BOOLEAN('i', "inherit", &inherit,
-		    "child tasks inherit counters"),
+	OPT_BOOLEAN('i', "no-inherit", &no_inherit,
+		    "child tasks do not inherit counters"),
 	OPT_INTEGER('F', "freq", &user_freq,
 		    "profile at this frequency"),
 	OPT_INTEGER('m', "mmap-pages", &mmap_pages,
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 04de3387de3f..68265120ee07 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -138,8 +138,14 @@ static int add_event_total(struct perf_session *session,
 	if (!hists)
 		return -ENOMEM;
 
-	hists->stats.total += data->period;
-	session->hists.stats.total += data->period;
+	hists->stats.total_period += data->period;
+	/*
+	 * FIXME: add_event_total should be moved from here to
+	 * perf_session__process_event so that the proper hist is passed to
+	 * the event_op methods.
+	 */
+	hists__inc_nr_events(hists, PERF_RECORD_SAMPLE);
+	session->hists.stats.total_period += data->period;
 	return 0;
 }
 
@@ -182,14 +188,14 @@ static int process_sample_event(event_t *event, struct perf_session *session)
 		return 0;
 
 	if (perf_session__add_hist_entry(session, &al, &data)) {
-		pr_debug("problem incrementing symbol count, skipping event\n");
+		pr_debug("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
 
 	attr = perf_header__find_attr(data.id, &session->header);
 
 	if (add_event_total(session, &data, attr)) {
-		pr_debug("problem adding event count\n");
+		pr_debug("problem adding event period\n");
 		return -1;
 	}
 
@@ -263,11 +269,25 @@ static struct perf_event_ops event_ops = {
 
 extern volatile int session_done;
 
-static void sig_handler(int sig __attribute__((__unused__)))
+static void sig_handler(int sig __used)
 {
 	session_done = 1;
 }
 
+static size_t hists__fprintf_nr_sample_events(struct hists *self,
+					      const char *evname, FILE *fp)
+{
+	size_t ret;
+	char unit;
+	unsigned long nr_events = self->stats.nr_events[PERF_RECORD_SAMPLE];
+
+	nr_events = convert_unit(nr_events, &unit);
+	ret = fprintf(fp, "# Events: %lu%c", nr_events, unit);
+	if (evname != NULL)
+		ret += fprintf(fp, " %s", evname);
+	return ret + fprintf(fp, "\n#\n");
+}
+
 static int __cmd_report(void)
 {
 	int ret = -EINVAL;
@@ -293,7 +313,7 @@ static int __cmd_report(void)
 		goto out_delete;
 
 	if (dump_trace) {
-		event__print_totals();
+		perf_session__fprintf_nr_events(session, stdout);
 		goto out_delete;
 	}
 
@@ -313,14 +333,12 @@ static int __cmd_report(void)
 		if (use_browser)
 			hists__browse(hists, help, input_name);
 		else {
-			if (rb_first(&session->hists.entries) ==
+			const char *evname = NULL;
+			if (rb_first(&session->hists.entries) !=
 			    rb_last(&session->hists.entries))
-				fprintf(stdout, "# Samples: %Ld\n#\n",
-					hists->stats.total);
-			else
-				fprintf(stdout, "# Samples: %Ld %s\n#\n",
-					hists->stats.total,
-					__event_name(hists->type, hists->config));
+				evname = __event_name(hists->type, hists->config);
+
+			hists__fprintf_nr_sample_events(hists, evname, stdout);
 
 			hists__fprintf(hists, NULL, false, stdout);
 			fprintf(stdout, "\n\n");
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index aef6ed0e119c..be7bc9264710 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -1641,19 +1641,10 @@ static int process_sample_event(event_t *event, struct perf_session *session)
 	return 0;
 }
 
-static int process_lost_event(event_t *event __used,
-			      struct perf_session *session __used)
-{
-	nr_lost_chunks++;
-	nr_lost_events += event->lost.lost;
-
-	return 0;
-}
-
 static struct perf_event_ops event_ops = {
 	.sample			= process_sample_event,
 	.comm			= event__process_comm,
-	.lost			= process_lost_event,
+	.lost			= event__process_lost,
 	.ordered_samples	= true,
 };
 
@@ -1664,8 +1655,12 @@ static int read_events(void)
 	if (session == NULL)
 		return -ENOMEM;
 
-	if (perf_session__has_traces(session, "record -R"))
+	if (perf_session__has_traces(session, "record -R")) {
 		err = perf_session__process_events(session, &event_ops);
+		nr_events      = session->hists.stats.nr_events[0];
+		nr_lost_events = session->hists.stats.total_lost;
+		nr_lost_chunks = session->hists.stats.nr_events[PERF_RECORD_LOST];
+	}
 
 	perf_session__delete(session);
 	return err;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index e619ac89dff5..ff8c413b7e73 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -72,7 +72,7 @@ static unsigned int		nr_cpus				=  0;
 static int			run_idx				=  0;
 
 static int			run_count			=  1;
-static bool			inherit				=  true;
+static bool			no_inherit			= false;
 static bool			scale				=  true;
 static pid_t			target_pid			= -1;
 static pid_t			target_tid			= -1;
@@ -167,8 +167,8 @@ static int create_perf_stat_counter(int counter)
 				++ncreated;
 		}
 	} else {
-		attr->inherit	     = inherit;
-		if (target_pid == -1) {
+		attr->inherit = !no_inherit;
+		if (target_pid == -1 && target_tid == -1) {
 			attr->disabled = 1;
 			attr->enable_on_exec = 1;
 		}
@@ -518,8 +518,8 @@ static const struct option options[] = {
 	OPT_CALLBACK('e', "event", NULL, "event",
 		     "event selector. use 'perf list' to list available events",
 		     parse_events),
-	OPT_BOOLEAN('i', "inherit", &inherit,
-		    "child tasks inherit counters"),
+	OPT_BOOLEAN('i', "no-inherit", &no_inherit,
+		    "child tasks do not inherit counters"),
 	OPT_INTEGER('p', "pid", &target_pid,
 		    "stat events on existing process id"),
 	OPT_INTEGER('t', "tid", &target_tid,
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 95fcb0517a9d..dddf3f01b5ab 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -109,7 +109,7 @@ static int process_sample_event(event_t *event, struct perf_session *session)
 					     data.time, thread->comm);
 	}
 
-	session->hists.stats.total += data.period;
+	session->hists.stats.total_period += data.period;
 	return 0;
 }
 
diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index cce006ec8f05..50771b5813ee 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -7,6 +7,23 @@
 #include "strlist.h"
 #include "thread.h"
 
+const char *event__name[] = {
+	[0]			 = "TOTAL",
+	[PERF_RECORD_MMAP]	 = "MMAP",
+	[PERF_RECORD_LOST]	 = "LOST",
+	[PERF_RECORD_COMM]	 = "COMM",
+	[PERF_RECORD_EXIT]	 = "EXIT",
+	[PERF_RECORD_THROTTLE]	 = "THROTTLE",
+	[PERF_RECORD_UNTHROTTLE] = "UNTHROTTLE",
+	[PERF_RECORD_FORK]	 = "FORK",
+	[PERF_RECORD_READ]	 = "READ",
+	[PERF_RECORD_SAMPLE]	 = "SAMPLE",
+	[PERF_RECORD_HEADER_ATTR]	 = "ATTR",
+	[PERF_RECORD_HEADER_EVENT_TYPE]	 = "EVENT_TYPE",
+	[PERF_RECORD_HEADER_TRACING_DATA]	 = "TRACING_DATA",
+	[PERF_RECORD_HEADER_BUILD_ID]	 = "BUILD_ID",
+};
+
 static pid_t event__synthesize_comm(pid_t pid, int full,
 				    event__handler_t process,
 				    struct perf_session *session)
@@ -368,7 +385,7 @@ int event__process_comm(event_t *self, struct perf_session *session)
 int event__process_lost(event_t *self, struct perf_session *session)
 {
 	dump_printf(": id:%Ld: lost:%Ld\n", self->lost.id, self->lost.lost);
-	session->hists.stats.lost += self->lost.lost;
+	session->hists.stats.total_lost += self->lost.lost;
 	return 0;
 }
 
diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
index 48c2cc9dae4f..8577085db067 100644
--- a/tools/perf/util/event.h
+++ b/tools/perf/util/event.h
@@ -160,4 +160,6 @@ int event__preprocess_sample(const event_t *self, struct perf_session *session,
 			     struct addr_location *al, symbol_filter_t filter);
 int event__parse_sample(event_t *event, u64 type, struct sample_data *data);
 
+extern const char *event__name[];
+
 #endif /* __PERF_RECORD_H */
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 451d2e45d843..f75c5f62401c 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -1,3 +1,4 @@
+#include "util.h"
 #include "hist.h"
 #include "session.h"
 #include "sort.h"
@@ -8,21 +9,21 @@ struct callchain_param	callchain_param = {
 	.min_percent = 0.5
 };
 
-static void hist_entry__add_cpumode_count(struct hist_entry *self,
-					  unsigned int cpumode, u64 count)
+static void hist_entry__add_cpumode_period(struct hist_entry *self,
+					   unsigned int cpumode, u64 period)
 {
 	switch (cpumode) {
 	case PERF_RECORD_MISC_KERNEL:
-		self->count_sys += count;
+		self->period_sys += period;
 		break;
 	case PERF_RECORD_MISC_USER:
-		self->count_us += count;
+		self->period_us += period;
 		break;
 	case PERF_RECORD_MISC_GUEST_KERNEL:
-		self->count_guest_sys += count;
+		self->period_guest_sys += period;
 		break;
 	case PERF_RECORD_MISC_GUEST_USER:
-		self->count_guest_us += count;
+		self->period_guest_us += period;
 		break;
 	default:
 		break;
@@ -30,7 +31,7 @@ static void hist_entry__add_cpumode_count(struct hist_entry *self,
 }
 
 /*
- * histogram, sorted on item, collects counts
+ * histogram, sorted on item, collects periods
  */
 
 static struct hist_entry *hist_entry__new(struct hist_entry *template)
@@ -40,6 +41,7 @@ static struct hist_entry *hist_entry__new(struct hist_entry *template)
 
 	if (self != NULL) {
 		*self = *template;
+		self->nr_events = 1;
 		if (symbol_conf.use_callchain)
 			callchain_init(self->callchain);
 	}
@@ -56,7 +58,7 @@ static void hists__inc_nr_entries(struct hists *self, struct hist_entry *entry)
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *sym_parent, u64 count)
+				      struct symbol *sym_parent, u64 period)
 {
 	struct rb_node **p = &self->entries.rb_node;
 	struct rb_node *parent = NULL;
@@ -69,7 +71,7 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 		},
 		.ip	= al->addr,
 		.level	= al->level,
-		.count	= count,
+		.period	= period,
 		.parent = sym_parent,
 	};
 	int cmp;
@@ -81,7 +83,8 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 		cmp = hist_entry__cmp(&entry, he);
 
 		if (!cmp) {
-			he->count += count;
+			he->period += period;
+			++he->nr_events;
 			goto out;
 		}
 
@@ -98,7 +101,7 @@ struct hist_entry *__hists__add_entry(struct hists *self,
 	rb_insert_color(&he->rb_node, &self->entries);
 	hists__inc_nr_entries(self, he);
 out:
-	hist_entry__add_cpumode_count(he, al->cpumode, count);
+	hist_entry__add_cpumode_period(he, al->cpumode, period);
 	return he;
 }
 
@@ -159,7 +162,7 @@ static bool collapse__insert_entry(struct rb_root *root, struct hist_entry *he)
 		cmp = hist_entry__collapse(iter, he);
 
 		if (!cmp) {
-			iter->count += he->count;
+			iter->period += he->period;
 			hist_entry__free(he);
 			return false;
 		}
@@ -202,7 +205,7 @@ void hists__collapse_resort(struct hists *self)
 }
 
 /*
- * reverse the map, sort on count.
+ * reverse the map, sort on period.
  */
 
 static void __hists__insert_output_entry(struct rb_root *entries,
@@ -221,7 +224,7 @@ static void __hists__insert_output_entry(struct rb_root *entries,
 		parent = *p;
 		iter = rb_entry(parent, struct hist_entry, rb_node);
 
-		if (he->count > iter->count)
+		if (he->period > iter->period)
 			p = &(*p)->rb_left;
 		else
 			p = &(*p)->rb_right;
@@ -238,7 +241,7 @@ void hists__output_resort(struct hists *self)
 	struct hist_entry *n;
 	u64 min_callchain_hits;
 
-	min_callchain_hits = self->stats.total * (callchain_param.min_percent / 100);
+	min_callchain_hits = self->stats.total_period * (callchain_param.min_percent / 100);
 
 	tmp = RB_ROOT;
 	next = rb_first(&self->entries);
@@ -287,7 +290,7 @@ static size_t ipchain__fprintf_graph_line(FILE *fp, int depth, int depth_mask,
 }
 
 static size_t ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain,
-				     int depth, int depth_mask, int count,
+				     int depth, int depth_mask, int period,
 				     u64 total_samples, int hits,
 				     int left_margin)
 {
@@ -300,7 +303,7 @@ static size_t ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain,
 			ret += fprintf(fp, "|");
 		else
 			ret += fprintf(fp, " ");
-		if (!count && i == depth - 1) {
+		if (!period && i == depth - 1) {
 			double percent;
 
 			percent = hits * 100.0 / total_samples;
@@ -515,7 +518,7 @@ int hist_entry__snprintf(struct hist_entry *self, char *s, size_t size,
 			 long displacement, bool color, u64 session_total)
 {
 	struct sort_entry *se;
-	u64 count, total, count_sys, count_us, count_guest_sys, count_guest_us;
+	u64 period, total, period_sys, period_us, period_guest_sys, period_guest_us;
 	const char *sep = symbol_conf.field_sep;
 	int ret;
 
@@ -523,57 +526,57 @@ int hist_entry__snprintf(struct hist_entry *self, char *s, size_t size,
 		return 0;
 
 	if (pair_hists) {
-		count = self->pair ? self->pair->count : 0;
-		total = pair_hists->stats.total;
-		count_sys = self->pair ? self->pair->count_sys : 0;
-		count_us = self->pair ? self->pair->count_us : 0;
-		count_guest_sys = self->pair ? self->pair->count_guest_sys : 0;
-		count_guest_us = self->pair ? self->pair->count_guest_us : 0;
+		period = self->pair ? self->pair->period : 0;
+		total = pair_hists->stats.total_period;
+		period_sys = self->pair ? self->pair->period_sys : 0;
+		period_us = self->pair ? self->pair->period_us : 0;
+		period_guest_sys = self->pair ? self->pair->period_guest_sys : 0;
+		period_guest_us = self->pair ? self->pair->period_guest_us : 0;
 	} else {
-		count = self->count;
+		period = self->period;
 		total = session_total;
-		count_sys = self->count_sys;
-		count_us = self->count_us;
-		count_guest_sys = self->count_guest_sys;
-		count_guest_us = self->count_guest_us;
+		period_sys = self->period_sys;
+		period_us = self->period_us;
+		period_guest_sys = self->period_guest_sys;
+		period_guest_us = self->period_guest_us;
 	}
 
 	if (total) {
 		if (color)
 			ret = percent_color_snprintf(s, size,
 						     sep ? "%.2f" : "   %6.2f%%",
-						     (count * 100.0) / total);
+						     (period * 100.0) / total);
 		else
 			ret = snprintf(s, size, sep ? "%.2f" : "   %6.2f%%",
-				       (count * 100.0) / total);
+				       (period * 100.0) / total);
 		if (symbol_conf.show_cpu_utilization) {
 			ret += percent_color_snprintf(s + ret, size - ret,
 					sep ? "%.2f" : "   %6.2f%%",
-					(count_sys * 100.0) / total);
+					(period_sys * 100.0) / total);
 			ret += percent_color_snprintf(s + ret, size - ret,
 					sep ? "%.2f" : "   %6.2f%%",
-					(count_us * 100.0) / total);
+					(period_us * 100.0) / total);
 			if (perf_guest) {
 				ret += percent_color_snprintf(s + ret,
 						size - ret,
 						sep ? "%.2f" : "   %6.2f%%",
-						(count_guest_sys * 100.0) /
+						(period_guest_sys * 100.0) /
 								total);
 				ret += percent_color_snprintf(s + ret,
 						size - ret,
 						sep ? "%.2f" : "   %6.2f%%",
-						(count_guest_us * 100.0) /
+						(period_guest_us * 100.0) /
 								total);
 			}
 		}
 	} else
-		ret = snprintf(s, size, sep ? "%lld" : "%12lld ", count);
+		ret = snprintf(s, size, sep ? "%lld" : "%12lld ", period);
 
 	if (symbol_conf.show_nr_samples) {
 		if (sep)
-			ret += snprintf(s + ret, size - ret, "%c%lld", *sep, count);
+			ret += snprintf(s + ret, size - ret, "%c%lld", *sep, period);
 		else
-			ret += snprintf(s + ret, size - ret, "%11lld", count);
+			ret += snprintf(s + ret, size - ret, "%11lld", period);
 	}
 
 	if (pair_hists) {
@@ -581,9 +584,9 @@ int hist_entry__snprintf(struct hist_entry *self, char *s, size_t size,
 		double old_percent = 0, new_percent = 0, diff;
 
 		if (total > 0)
-			old_percent = (count * 100.0) / total;
+			old_percent = (period * 100.0) / total;
 		if (session_total > 0)
-			new_percent = (self->count * 100.0) / session_total;
+			new_percent = (self->period * 100.0) / session_total;
 
 		diff = new_percent - old_percent;
 
@@ -768,10 +771,10 @@ print_entries:
 			++position;
 		}
 		ret += hist_entry__fprintf(h, pair, show_displacement,
-					   displacement, fp, self->stats.total);
+					   displacement, fp, self->stats.total_period);
 
 		if (symbol_conf.use_callchain)
-			ret += hist_entry__fprintf_callchain(h, fp, self->stats.total);
+			ret += hist_entry__fprintf_callchain(h, fp, self->stats.total_period);
 
 		if (h->ms.map == NULL && verbose > 1) {
 			__map_groups__fprintf_maps(&h->thread->mg,
@@ -794,7 +797,8 @@ void hists__filter_by_dso(struct hists *self, const struct dso *dso)
 {
 	struct rb_node *nd;
 
-	self->nr_entries = self->stats.total = 0;
+	self->nr_entries = self->stats.total_period = 0;
+	self->stats.nr_events[PERF_RECORD_SAMPLE] = 0;
 	self->max_sym_namelen = 0;
 
 	for (nd = rb_first(&self->entries); nd; nd = rb_next(nd)) {
@@ -811,7 +815,8 @@ void hists__filter_by_dso(struct hists *self, const struct dso *dso)
 		h->filtered &= ~(1 << HIST_FILTER__DSO);
 		if (!h->filtered) {
 			++self->nr_entries;
-			self->stats.total += h->count;
+			self->stats.total_period += h->period;
+			self->stats.nr_events[PERF_RECORD_SAMPLE] += h->nr_events;
 			if (h->ms.sym &&
 			    self->max_sym_namelen < h->ms.sym->namelen)
 				self->max_sym_namelen = h->ms.sym->namelen;
@@ -823,7 +828,8 @@ void hists__filter_by_thread(struct hists *self, const struct thread *thread)
 {
 	struct rb_node *nd;
 
-	self->nr_entries = self->stats.total = 0;
+	self->nr_entries = self->stats.total_period = 0;
+	self->stats.nr_events[PERF_RECORD_SAMPLE] = 0;
 	self->max_sym_namelen = 0;
 
 	for (nd = rb_first(&self->entries); nd; nd = rb_next(nd)) {
@@ -836,7 +842,8 @@ void hists__filter_by_thread(struct hists *self, const struct thread *thread)
 		h->filtered &= ~(1 << HIST_FILTER__THREAD);
 		if (!h->filtered) {
 			++self->nr_entries;
-			self->stats.total += h->count;
+			self->stats.total_period += h->period;
+			self->stats.nr_events[PERF_RECORD_SAMPLE] += h->nr_events;
 			if (h->ms.sym &&
 			    self->max_sym_namelen < h->ms.sym->namelen)
 				self->max_sym_namelen = h->ms.sym->namelen;
@@ -880,7 +887,7 @@ int hist_entry__inc_addr_samples(struct hist_entry *self, u64 ip)
 	h->sum++;
 	h->ip[offset]++;
 
-	pr_debug3("%#Lx %s: count++ [ip: %#Lx, %#Lx] => %Ld\n", self->ms.sym->start,
+	pr_debug3("%#Lx %s: period++ [ip: %#Lx, %#Lx] => %Ld\n", self->ms.sym->start,
 		  self->ms.sym->name, ip, ip - self->ms.sym->start, h->ip[offset]);
 	return 0;
 }
@@ -1027,3 +1034,24 @@ int hist_entry__annotate(struct hist_entry *self, struct list_head *head)
 	pclose(file);
 	return 0;
 }
+
+void hists__inc_nr_events(struct hists *self, u32 type)
+{
+	++self->stats.nr_events[0];
+	++self->stats.nr_events[type];
+}
+
+size_t hists__fprintf_nr_events(struct hists *self, FILE *fp)
+{
+	int i;
+	size_t ret = 0;
+
+	for (i = 0; i < PERF_RECORD_HEADER_MAX; ++i) {
+		if (!event__name[i])
+			continue;
+		ret += fprintf(fp, "%10s events: %10d\n",
+			       event__name[i], self->stats.nr_events[i]);
+	}
+
+	return ret;
+}
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index ed9c06734965..6f17dcd8412c 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -37,9 +37,23 @@ struct sym_priv {
 	struct sym_ext	*ext;
 };
 
+/*
+ * The kernel collects the number of events it couldn't send in a stretch and
+ * when possible sends this number in a PERF_RECORD_LOST event. The number of
+ * such "chunks" of lost events is stored in .nr_events[PERF_EVENT_LOST] while
+ * total_lost tells exactly how many events the kernel in fact lost, i.e. it is
+ * the sum of all struct lost_event.lost fields reported.
+ *
+ * The total_period is needed because by default auto-freq is used, so
+ * multipling nr_events[PERF_EVENT_SAMPLE] by a frequency isn't possible to get
+ * the total number of low level events, it is necessary to to sum all struct
+ * sample_event.period and stash the result in total_period.
+ */
 struct events_stats {
-	u64 total;
-	u64 lost;
+	u64 total_period;
+	u64 total_lost;
+	u32 nr_events[PERF_RECORD_HEADER_MAX];
+	u32 nr_unknown_events;
 };
 
 struct hists {
@@ -55,7 +69,7 @@ struct hists {
 
 struct hist_entry *__hists__add_entry(struct hists *self,
 				      struct addr_location *al,
-				      struct symbol *parent, u64 count);
+				      struct symbol *parent, u64 period);
 extern int64_t hist_entry__cmp(struct hist_entry *, struct hist_entry *);
 extern int64_t hist_entry__collapse(struct hist_entry *, struct hist_entry *);
 int hist_entry__fprintf(struct hist_entry *self, struct hists *pair_hists,
@@ -68,6 +82,10 @@ void hist_entry__free(struct hist_entry *);
 
 void hists__output_resort(struct hists *self);
 void hists__collapse_resort(struct hists *self);
+
+void hists__inc_nr_events(struct hists *self, u32 type);
+size_t hists__fprintf_nr_events(struct hists *self, FILE *fp);
+
 size_t hists__fprintf(struct hists *self, struct hists *pair,
 		      bool show_displacement, FILE *fp);
 
@@ -78,7 +96,7 @@ void hists__filter_by_dso(struct hists *self, const struct dso *dso);
 void hists__filter_by_thread(struct hists *self, const struct thread *thread);
 
 #ifdef NO_NEWT_SUPPORT
-static inline int hists__browse(struct hists self __used,
+static inline int hists__browse(struct hists *self __used,
 				const char *helpline __used,
 				const char *input_name __used)
 {
diff --git a/tools/perf/util/newt.c b/tools/perf/util/newt.c
index ba6acd04c082..e74df1240ef6 100644
--- a/tools/perf/util/newt.c
+++ b/tools/perf/util/newt.c
@@ -9,6 +9,7 @@
 
 #include "cache.h"
 #include "hist.h"
+#include "pstack.h"
 #include "session.h"
 #include "sort.h"
 #include "symbol.h"
@@ -680,16 +681,18 @@ static int hist_browser__populate(struct hist_browser *self, struct hists *hists
 	struct ui_progress *progress;
 	struct rb_node *nd;
 	u64 curr_hist = 0;
-	char seq[] = ".";
+	char seq[] = ".", unit;
 	char str[256];
+	unsigned long nr_events = hists->stats.nr_events[PERF_RECORD_SAMPLE];
 
 	if (self->form) {
 		newtFormDestroy(self->form);
 		newtPopWindow();
 	}
 
-	snprintf(str, sizeof(str), "Samples: %Ld                            ",
-		 hists->stats.total);
+	nr_events = convert_unit(nr_events, &unit);
+	snprintf(str, sizeof(str), "Events: %lu%c                            ",
+		 nr_events, unit);
 	newtDrawRootText(0, 0, str);
 
 	newtGetScreenSize(NULL, &rows);
@@ -718,12 +721,12 @@ static int hist_browser__populate(struct hist_browser *self, struct hists *hists
 		if (h->filtered)
 			continue;
 
-		len = hist_entry__append_browser(h, self->tree, hists->stats.total);
+		len = hist_entry__append_browser(h, self->tree, hists->stats.total_period);
 		if (len > max_len)
 			max_len = len;
 		if (symbol_conf.use_callchain)
 			hist_entry__append_callchain_browser(h, self->tree,
-							     hists->stats.total, idx++);
+							     hists->stats.total_period, idx++);
 		++curr_hist;
 		if (curr_hist % 5)
 			ui_progress__update(progress, curr_hist);
@@ -748,6 +751,7 @@ static int hist_browser__populate(struct hist_browser *self, struct hists *hists
 	newtFormAddHotKey(self->form, 'A');
 	newtFormAddHotKey(self->form, 'a');
 	newtFormAddHotKey(self->form, NEWT_KEY_RIGHT);
+	newtFormAddHotKey(self->form, NEWT_KEY_LEFT);
 	newtFormAddComponents(self->form, self->tree, NULL);
 	self->selection = newt__symbol_tree_get_current(self->tree);
 
@@ -799,6 +803,7 @@ static int hist_browser__title(char *bf, size_t size, const char *input_name,
 int hists__browse(struct hists *self, const char *helpline, const char *input_name)
 {
 	struct hist_browser *browser = hist_browser__new();
+	struct pstack *fstack = pstack__new(2);
 	const struct thread *thread_filter = NULL;
 	const struct dso *dso_filter = NULL;
 	struct newtExitStruct es;
@@ -808,12 +813,16 @@ int hists__browse(struct hists *self, const char *helpline, const char *input_na
 	if (browser == NULL)
 		return -1;
 
+	fstack = pstack__new(2);
+	if (fstack == NULL)
+		goto out;
+
 	ui_helpline__push(helpline);
 
 	hist_browser__title(msg, sizeof(msg), input_name,
 			    dso_filter, thread_filter);
 	if (hist_browser__populate(browser, self, msg) < 0)
-		goto out;
+		goto out_free_stack;
 
 	while (1) {
 		const struct thread *thread;
@@ -834,6 +843,19 @@ int hists__browse(struct hists *self, const char *helpline, const char *input_na
 				else
 					continue;
 			}
+
+			if (es.u.key == NEWT_KEY_LEFT) {
+				const void *top;
+
+				if (pstack__empty(fstack))
+					continue;
+				top = pstack__pop(fstack);
+				if (top == &dso_filter)
+					goto zoom_out_dso;
+				if (top == &thread_filter)
+					goto zoom_out_thread;
+				continue;
+			}
 		}
 
 		if (browser->selection->sym != NULL &&
@@ -886,12 +908,15 @@ do_annotate:
 			hist_entry__annotate_browser(he);
 		} else if (choice == zoom_dso) {
 			if (dso_filter) {
+				pstack__remove(fstack, &dso_filter);
+zoom_out_dso:
 				ui_helpline__pop();
 				dso_filter = NULL;
 			} else {
-				ui_helpline__fpush("To zoom out press -> + \"Zoom out of %s DSO\"",
+				ui_helpline__fpush("To zoom out press <- or -> + \"Zoom out of %s DSO\"",
 						   dso->kernel ? "the Kernel" : dso->short_name);
 				dso_filter = dso;
+				pstack__push(fstack, &dso_filter);
 			}
 			hists__filter_by_dso(self, dso_filter);
 			hist_browser__title(msg, sizeof(msg), input_name,
@@ -900,13 +925,16 @@ do_annotate:
 				goto out;
 		} else if (choice == zoom_thread) {
 			if (thread_filter) {
+				pstack__remove(fstack, &thread_filter);
+zoom_out_thread:
 				ui_helpline__pop();
 				thread_filter = NULL;
 			} else {
-				ui_helpline__fpush("To zoom out press -> + \"Zoom out of %s(%d) thread\"",
+				ui_helpline__fpush("To zoom out press <- or -> + \"Zoom out of %s(%d) thread\"",
 						   thread->comm_set ? thread->comm : "",
 						   thread->pid);
 				thread_filter = thread;
+				pstack__push(fstack, &thread_filter);
 			}
 			hists__filter_by_thread(self, thread_filter);
 			hist_browser__title(msg, sizeof(msg), input_name,
@@ -916,6 +944,8 @@ do_annotate:
 		}
 	}
 	err = 0;
+out_free_stack:
+	pstack__delete(fstack);
 out:
 	hist_browser__delete(browser);
 	return err;
diff --git a/tools/perf/util/pstack.c b/tools/perf/util/pstack.c
new file mode 100644
index 000000000000..13d36faf64eb
--- /dev/null
+++ b/tools/perf/util/pstack.c
@@ -0,0 +1,75 @@
+/*
+ * Simple pointer stack
+ *
+ * (c) 2010 Arnaldo Carvalho de Melo <acme@redhat.com>
+ */
+
+#include "util.h"
+#include "pstack.h"
+#include <linux/kernel.h>
+#include <stdlib.h>
+
+struct pstack {
+	unsigned short	top;
+	unsigned short	max_nr_entries;
+	void		*entries[0];
+};
+
+struct pstack *pstack__new(unsigned short max_nr_entries)
+{
+	struct pstack *self = zalloc((sizeof(*self) +
+				     max_nr_entries * sizeof(void *)));
+	if (self != NULL)
+		self->max_nr_entries = max_nr_entries;
+	return self;
+}
+
+void pstack__delete(struct pstack *self)
+{
+	free(self);
+}
+
+bool pstack__empty(const struct pstack *self)
+{
+	return self->top == 0;
+}
+
+void pstack__remove(struct pstack *self, void *key)
+{
+	unsigned short i = self->top, last_index = self->top - 1;
+
+	while (i-- != 0) {
+		if (self->entries[i] == key) {
+			if (i < last_index)
+				memmove(self->entries + i,
+					self->entries + i + 1,
+					(last_index - i) * sizeof(void *));
+			--self->top;
+			return;
+		}
+	}
+	pr_err("%s: %p not on the pstack!\n", __func__, key);
+}
+
+void pstack__push(struct pstack *self, void *key)
+{
+	if (self->top == self->max_nr_entries) {
+		pr_err("%s: top=%d, overflow!\n", __func__, self->top);
+		return;
+	}
+	self->entries[self->top++] = key;
+}
+
+void *pstack__pop(struct pstack *self)
+{
+	void *ret;
+
+	if (self->top == 0) {
+		pr_err("%s: underflow!\n", __func__);
+		return NULL;
+	}
+
+	ret = self->entries[--self->top];
+	self->entries[self->top] = NULL;
+	return ret;
+}
diff --git a/tools/perf/util/pstack.h b/tools/perf/util/pstack.h
new file mode 100644
index 000000000000..5ad07023504b
--- /dev/null
+++ b/tools/perf/util/pstack.h
@@ -0,0 +1,12 @@
+#ifndef _PERF_PSTACK_
+#define _PERF_PSTACK_
+
+struct pstack;
+struct pstack *pstack__new(unsigned short max_nr_entries);
+void pstack__delete(struct pstack *self);
+bool pstack__empty(const struct pstack *self);
+void pstack__remove(struct pstack *self, void *key);
+void pstack__push(struct pstack *self, void *key);
+void *pstack__pop(struct pstack *self);
+
+#endif /* _PERF_PSTACK_ */
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 72a7f6ae0293..25bfca4f10f0 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -94,7 +94,6 @@ struct perf_session *perf_session__new(const char *filename, int mode, bool forc
 	self->mmap_window = 32;
 	self->cwd = NULL;
 	self->cwdlen = 0;
-	self->unknown_events = 0;
 	self->machines = RB_ROOT;
 	self->repipe = repipe;
 	INIT_LIST_HEAD(&self->ordered_samples.samples_head);
@@ -241,36 +240,6 @@ static void perf_event_ops__fill_defaults(struct perf_event_ops *handler)
 	}
 }
 
-static const char *event__name[] = {
-	[0]			 = "TOTAL",
-	[PERF_RECORD_MMAP]	 = "MMAP",
-	[PERF_RECORD_LOST]	 = "LOST",
-	[PERF_RECORD_COMM]	 = "COMM",
-	[PERF_RECORD_EXIT]	 = "EXIT",
-	[PERF_RECORD_THROTTLE]	 = "THROTTLE",
-	[PERF_RECORD_UNTHROTTLE] = "UNTHROTTLE",
-	[PERF_RECORD_FORK]	 = "FORK",
-	[PERF_RECORD_READ]	 = "READ",
-	[PERF_RECORD_SAMPLE]	 = "SAMPLE",
-	[PERF_RECORD_HEADER_ATTR]	 = "ATTR",
-	[PERF_RECORD_HEADER_EVENT_TYPE]	 = "EVENT_TYPE",
-	[PERF_RECORD_HEADER_TRACING_DATA]	 = "TRACING_DATA",
-	[PERF_RECORD_HEADER_BUILD_ID]	 = "BUILD_ID",
-};
-
-unsigned long event__total[PERF_RECORD_HEADER_MAX];
-
-void event__print_totals(void)
-{
-	int i;
-	for (i = 0; i < PERF_RECORD_HEADER_MAX; ++i) {
-		if (!event__name[i])
-			continue;
-		pr_info("%10s events: %10ld\n",
-			event__name[i], event__total[i]);
-	}
-}
-
 void mem_bswap_64(void *src, int byte_size)
 {
 	u64 *m = src;
@@ -580,8 +549,7 @@ static int perf_session__process_event(struct perf_session *self,
 		dump_printf("%#Lx [%#x]: PERF_RECORD_%s",
 			    offset + head, event->header.size,
 			    event__name[event->header.type]);
-		++event__total[0];
-		++event__total[event->header.type];
+		hists__inc_nr_events(&self->hists, event->header.type);
 	}
 
 	if (self->header.needs_swap && event__swap_ops[event->header.type])
@@ -619,7 +587,7 @@ static int perf_session__process_event(struct perf_session *self,
 	case PERF_RECORD_FINISHED_ROUND:
 		return ops->finished_round(event, self, ops);
 	default:
-		self->unknown_events++;
+		++self->hists.stats.nr_unknown_events;
 		return -1;
 	}
 }
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index ce00fa6cdeda..e7fce486ebe2 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -30,8 +30,6 @@ struct perf_session {
 	struct machine		host_machine;
 	struct rb_root		machines;
 	struct rb_root		hists_tree;
-	unsigned long		event_total[PERF_RECORD_MAX];
-	unsigned long		unknown_events;
 	/*
 	 * FIXME: should point to the first entry in hists_tree and
 	 *        be a hists instance. Right now its only 'report'
@@ -140,4 +138,10 @@ size_t perf_session__fprintf_dsos_buildid(struct perf_session *self, FILE *fp,
 {
 	return machines__fprintf_dsos_buildid(&self->machines, fp, with_hits);
 }
+
+static inline
+size_t perf_session__fprintf_nr_events(struct perf_session *self, FILE *fp)
+{
+	return hists__fprintf_nr_events(&self->hists, fp);
+}
 #endif /* __PERF_SESSION_H */
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index af301acc461c..eab2e0b3b74e 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -43,14 +43,15 @@ extern enum sort_type sort__first_dimension;
 
 struct hist_entry {
 	struct rb_node		rb_node;
-	u64			count;
-	u64			count_sys;
-	u64			count_us;
-	u64			count_guest_sys;
-	u64			count_guest_us;
+	u64			period;
+	u64			period_sys;
+	u64			period_us;
+	u64			period_guest_sys;
+	u64			period_guest_us;
 	struct map_symbol	ms;
 	struct thread		*thread;
 	u64			ip;
+	u32			nr_events;
 	char			level;
 	u8			filtered;
 	struct symbol		*parent;
diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c
index f9b890fde681..214265674ddd 100644
--- a/tools/perf/util/util.c
+++ b/tools/perf/util/util.c
@@ -92,3 +92,25 @@ out_close_from:
 out:
 	return err;
 }
+
+unsigned long convert_unit(unsigned long value, char *unit)
+{
+	*unit = ' ';
+
+	if (value > 1000) {
+		value /= 1000;
+		*unit = 'K';
+	}
+
+	if (value > 1000) {
+		value /= 1000;
+		*unit = 'M';
+	}
+
+	if (value > 1000) {
+		value /= 1000;
+		*unit = 'G';
+	}
+
+	return value;
+}
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index fbf45d1b26f7..0795bf304b19 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -423,6 +423,7 @@ char **argv_split(const char *str, int *argcp);
 void argv_free(char **argv);
 bool strglobmatch(const char *str, const char *pat);
 bool strlazymatch(const char *str, const char *pat);
+unsigned long convert_unit(unsigned long value, char *unit);
 
 #define _STR(x) #x
 #define STR(x) _STR(x)