[GIT PULL 00/17] perf/core improvements and fixes

[GIT PULL 00/17] perf/core improvements and fixes

[Arnaldo Carvalho de Melo]

The following changes since commit bab57e994d6311298b4e3915d2c75296cd81638c:

  Merge branch 'core' of git://git.kernel.org/pub/scm/linux/kernel/git/rric/oprofile into perf/core (2012-09-05 08:29:56 +0200)

are available in the git repository at:


  git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux tags/perf-core-for-mingo

for you to fetch changes up to 7a4ec938857cf534270b23545495300fbac7f5de:

  perf tools: Allow user to indicate path to objdump in command line (2012-09-05 19:41:55 -0300)

----------------------------------------------------------------
perf/core improvements and fixes:

1) Rename libtraceevent 'private' struct member to 'priv' so that it works
   in C++, from Steven Rostedt

2) Remove lots of exit()/die() calls from tools so that the main perf exit
   routine can take place, from David Ahern

3) Fix x86 build on x86-64, from David Ahern.

4) Remove some headers that prevented perf from building on Android,
   from David Ahern

5) {int,str,rb}list fixes from Suzuki K Poulose

6) perf.data header fixes from Namhyung Kim

7) Replace needless mempcpy with memcpy, to allow build on Android, from Irina Tirdea

8) Allow user to indicate objdump path, needed in cross environments, from
   Maciek Borzecki

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

----------------------------------------------------------------
David Ahern (9):
      perf session: flush_sample_queue needs to handle errors from handlers
      perf tool: handle errors in synthesized event functions
      perf lock: Remove use of die and handle errors
      perf stat: Remove use of die/exit and handle errors
      perf help: Remove use of die and handle errors
      perf script: Remove use of die/exit
      perf record: Remove use of die/exit
      perf tools: Fix x86 builds with ARCH specified on the command line
      perf tools: remove unneeded include of network header files

Irina Tirdea (1):
      perf tools: Replace mempcpy with memcpy

Maciek Borzecki (1):
      perf tools: Allow user to indicate path to objdump in command line

Namhyung Kim (3):
      perf header: Use evlist->nr_entries on write_event_desc()
      perf header: Set tracepoint event name only if not set
      perf header: Swap pmu mapping numbers if needed

Steven Rostedt (1):
      tools lib traceevent: Modify header to work in C++ programs

Suzuki K. Poulose (2):
      perf tools: Fix intlist node removal
      perf tools: Remove the node from rblist in strlist__remove

 tools/lib/traceevent/event-parse.h                 |    4 +-
 tools/perf/Documentation/perf-annotate.txt         |    3 +
 tools/perf/Documentation/perf-report.txt           |    3 +
 tools/perf/Makefile                                |    4 +-
 tools/perf/builtin-annotate.c                      |    2 +
 tools/perf/builtin-help.c                          |   48 ++++--
 tools/perf/builtin-lock.c                          |  181 ++++++++++++++------
 tools/perf/builtin-record.c                        |  158 +++++++++++------
 tools/perf/builtin-report.c                        |    2 +
 tools/perf/builtin-script.c                        |   60 +++++--
 tools/perf/builtin-stat.c                          |    7 +-
 tools/perf/util/annotate.c                         |    4 +-
 tools/perf/util/annotate.h                         |    1 +
 tools/perf/util/event.c                            |   35 +++-
 tools/perf/util/header.c                           |   13 +-
 tools/perf/util/intlist.c                          |    4 +-
 .../perf/util/scripting-engines/trace-event-perl.c |    8 +-
 tools/perf/util/session.c                          |   24 ++-
 tools/perf/util/strlist.c                          |    2 +-
 tools/perf/util/target.c                           |    4 +-
 tools/perf/util/util.h                             |    5 -
 21 files changed, 389 insertions(+), 183 deletions(-)

[PATCH 01/17] tools lib traceevent: Modify header to work in C++ programs

[Arnaldo Carvalho de Melo]

From: Steven Rostedt <rostedt@goodmis.org>

Replace keyword "private" to "priv" in event-parse.h to allow it to be
used in C++ programs.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1345735321.5069.62.camel@gandalf.local.home
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/lib/traceevent/event-parse.h |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/lib/traceevent/event-parse.h b/tools/lib/traceevent/event-parse.h
index 527df03..863a0bb 100644
--- a/tools/lib/traceevent/event-parse.h
+++ b/tools/lib/traceevent/event-parse.h
@@ -49,7 +49,7 @@ struct pevent_record {
 	int			cpu;
 	int			ref_count;
 	int			locked;		/* Do not free, even if ref_count is zero */
-	void			*private;
+	void			*priv;
 #if DEBUG_RECORD
 	struct pevent_record	*prev;
 	struct pevent_record	*next;
@@ -106,7 +106,7 @@ struct plugin_option {
 	char				*plugin_alias;
 	char				*description;
 	char				*value;
-	void				*private;
+	void				*priv;
 	int				set;
 };
 
-- 
1.7.1

[PATCH 02/17] perf session: flush_sample_queue needs to handle errors from handlers

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows errors to propogate through event processing code and back to
commands.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-2-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/session.c |   24 +++++++++++++++---------
 1 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index f7bb7ae..9453758 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -692,7 +692,7 @@ static int perf_session_deliver_event(struct perf_session *session,
 				      struct perf_tool *tool,
 				      u64 file_offset);
 
-static void flush_sample_queue(struct perf_session *s,
+static int flush_sample_queue(struct perf_session *s,
 			       struct perf_tool *tool)
 {
 	struct ordered_samples *os = &s->ordered_samples;
@@ -705,7 +705,7 @@ static void flush_sample_queue(struct perf_session *s,
 	int ret;
 
 	if (!tool->ordered_samples || !limit)
-		return;
+		return 0;
 
 	list_for_each_entry_safe(iter, tmp, head, list) {
 		if (iter->timestamp > limit)
@@ -715,9 +715,12 @@ static void flush_sample_queue(struct perf_session *s,
 						s->header.needs_swap);
 		if (ret)
 			pr_err("Can't parse sample, err = %d\n", ret);
-		else
-			perf_session_deliver_event(s, iter->event, &sample, tool,
-						   iter->file_offset);
+		else {
+			ret = perf_session_deliver_event(s, iter->event, &sample, tool,
+							 iter->file_offset);
+			if (ret)
+				return ret;
+		}
 
 		os->last_flush = iter->timestamp;
 		list_del(&iter->list);
@@ -737,6 +740,8 @@ static void flush_sample_queue(struct perf_session *s,
 	}
 
 	os->nr_samples = 0;
+
+	return 0;
 }
 
 /*
@@ -782,10 +787,11 @@ static int process_finished_round(struct perf_tool *tool,
 				  union perf_event *event __used,
 				  struct perf_session *session)
 {
-	flush_sample_queue(session, tool);
-	session->ordered_samples.next_flush = session->ordered_samples.max_timestamp;
+	int ret = flush_sample_queue(session, tool);
+	if (!ret)
+		session->ordered_samples.next_flush = session->ordered_samples.max_timestamp;
 
-	return 0;
+	return ret;
 }
 
 /* The queue is ordered by time */
@@ -1443,7 +1449,7 @@ more:
 	err = 0;
 	/* do the final flush for ordered samples */
 	session->ordered_samples.next_flush = ULLONG_MAX;
-	flush_sample_queue(session, tool);
+	err = flush_sample_queue(session, tool);
 out_err:
 	perf_session__warn_about_errors(session, tool);
 	perf_session_free_sample_buffers(session);
-- 
1.7.1

[PATCH 03/17] perf tool: handle errors in synthesized event functions

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Handle error from process callback and propagate back to caller.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-3-git-send-email-dsahern@gmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/event.c |   35 ++++++++++++++++++++++++++---------
 1 files changed, 26 insertions(+), 9 deletions(-)

diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index 3a0f1a5..84ff6f1 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -120,7 +120,9 @@ static pid_t perf_event__synthesize_comm(struct perf_tool *tool,
 	if (!full) {
 		event->comm.tid = pid;
 
-		process(tool, event, &synth_sample, machine);
+		if (process(tool, event, &synth_sample, machine) != 0)
+			return -1;
+
 		goto out;
 	}
 
@@ -151,7 +153,10 @@ static pid_t perf_event__synthesize_comm(struct perf_tool *tool,
 
 		event->comm.tid = pid;
 
-		process(tool, event, &synth_sample, machine);
+		if (process(tool, event, &synth_sample, machine) != 0) {
+			tgid = -1;
+			break;
+		}
 	}
 
 	closedir(tasks);
@@ -167,6 +172,7 @@ static int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 {
 	char filename[PATH_MAX];
 	FILE *fp;
+	int rc = 0;
 
 	snprintf(filename, sizeof(filename), "/proc/%d/maps", pid);
 
@@ -231,18 +237,22 @@ static int perf_event__synthesize_mmap_events(struct perf_tool *tool,
 			event->mmap.pid = tgid;
 			event->mmap.tid = pid;
 
-			process(tool, event, &synth_sample, machine);
+			if (process(tool, event, &synth_sample, machine) != 0) {
+				rc = -1;
+				break;
+			}
 		}
 	}
 
 	fclose(fp);
-	return 0;
+	return rc;
 }
 
 int perf_event__synthesize_modules(struct perf_tool *tool,
 				   perf_event__handler_t process,
 				   struct machine *machine)
 {
+	int rc = 0;
 	struct rb_node *nd;
 	struct map_groups *kmaps = &machine->kmaps;
 	union perf_event *event = zalloc((sizeof(event->mmap) +
@@ -284,11 +294,14 @@ int perf_event__synthesize_modules(struct perf_tool *tool,
 
 		memcpy(event->mmap.filename, pos->dso->long_name,
 		       pos->dso->long_name_len + 1);
-		process(tool, event, &synth_sample, machine);
+		if (process(tool, event, &synth_sample, machine) != 0) {
+			rc = -1;
+			break;
+		}
 	}
 
 	free(event);
-	return 0;
+	return rc;
 }
 
 static int __event__synthesize_thread(union perf_event *comm_event,
@@ -392,12 +405,16 @@ int perf_event__synthesize_threads(struct perf_tool *tool,
 		if (*end) /* only interested in proper numerical dirents */
 			continue;
 
-		__event__synthesize_thread(comm_event, mmap_event, pid, 1,
-					   process, tool, machine);
+		if (__event__synthesize_thread(comm_event, mmap_event, pid, 1,
+					   process, tool, machine) != 0) {
+			err = -1;
+			goto out_closedir;
+		}
 	}
 
-	closedir(proc);
 	err = 0;
+out_closedir:
+	closedir(proc);
 out_free_mmap:
 	free(mmap_event);
 out_free_comm:
-- 
1.7.1

[PATCH 04/17] perf lock: Remove use of die and handle errors

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows perf to clean up properly on exit.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-4-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-lock.c |  181 +++++++++++++++++++++++++++++++--------------
 1 files changed, 124 insertions(+), 57 deletions(-)

diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index 585aae2..75153c8 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -161,8 +161,10 @@ static struct thread_stat *thread_stat_findnew_after_first(u32 tid)
 		return st;
 
 	st = zalloc(sizeof(struct thread_stat));
-	if (!st)
-		die("memory allocation failed\n");
+	if (!st) {
+		pr_err("memory allocation failed\n");
+		return NULL;
+	}
 
 	st->tid = tid;
 	INIT_LIST_HEAD(&st->seq_list);
@@ -181,8 +183,10 @@ static struct thread_stat *thread_stat_findnew_first(u32 tid)
 	struct thread_stat *st;
 
 	st = zalloc(sizeof(struct thread_stat));
-	if (!st)
-		die("memory allocation failed\n");
+	if (!st) {
+		pr_err("memory allocation failed\n");
+		return NULL;
+	}
 	st->tid = tid;
 	INIT_LIST_HEAD(&st->seq_list);
 
@@ -248,18 +252,20 @@ struct lock_key keys[] = {
 	{ NULL, NULL }
 };
 
-static void select_key(void)
+static int select_key(void)
 {
 	int i;
 
 	for (i = 0; keys[i].name; i++) {
 		if (!strcmp(keys[i].name, sort_key)) {
 			compare = keys[i].key;
-			return;
+			return 0;
 		}
 	}
 
-	die("Unknown compare key:%s\n", sort_key);
+	pr_err("Unknown compare key: %s\n", sort_key);
+
+	return -1;
 }
 
 static void insert_to_result(struct lock_stat *st,
@@ -324,7 +330,8 @@ static struct lock_stat *lock_stat_findnew(void *addr, const char *name)
 	return new;
 
 alloc_failed:
-	die("memory allocation failed\n");
+	pr_err("memory allocation failed\n");
+	return NULL;
 }
 
 static const char *input_name;
@@ -356,16 +363,16 @@ struct trace_release_event {
 };
 
 struct trace_lock_handler {
-	void (*acquire_event)(struct trace_acquire_event *,
+	int (*acquire_event)(struct trace_acquire_event *,
 			      const struct perf_sample *sample);
 
-	void (*acquired_event)(struct trace_acquired_event *,
+	int (*acquired_event)(struct trace_acquired_event *,
 			       const struct perf_sample *sample);
 
-	void (*contended_event)(struct trace_contended_event *,
+	int (*contended_event)(struct trace_contended_event *,
 				const struct perf_sample *sample);
 
-	void (*release_event)(struct trace_release_event *,
+	int (*release_event)(struct trace_release_event *,
 			      const struct perf_sample *sample);
 };
 
@@ -379,8 +386,10 @@ static struct lock_seq_stat *get_seq(struct thread_stat *ts, void *addr)
 	}
 
 	seq = zalloc(sizeof(struct lock_seq_stat));
-	if (!seq)
-		die("Not enough memory\n");
+	if (!seq) {
+		pr_err("memory allocation failed\n");
+		return NULL;
+	}
 	seq->state = SEQ_STATE_UNINITIALIZED;
 	seq->addr = addr;
 
@@ -403,7 +412,7 @@ enum acquire_flags {
 	READ_LOCK = 2,
 };
 
-static void
+static int
 report_lock_acquire_event(struct trace_acquire_event *acquire_event,
 			  const struct perf_sample *sample)
 {
@@ -412,11 +421,18 @@ report_lock_acquire_event(struct trace_acquire_event *acquire_event,
 	struct lock_seq_stat *seq;
 
 	ls = lock_stat_findnew(acquire_event->addr, acquire_event->name);
+	if (!ls)
+		return -1;
 	if (ls->discard)
-		return;
+		return 0;
 
 	ts = thread_stat_findnew(sample->tid);
+	if (!ts)
+		return -1;
+
 	seq = get_seq(ts, acquire_event->addr);
+	if (!seq)
+		return -1;
 
 	switch (seq->state) {
 	case SEQ_STATE_UNINITIALIZED:
@@ -461,10 +477,10 @@ broken:
 	ls->nr_acquire++;
 	seq->prev_event_time = sample->time;
 end:
-	return;
+	return 0;
 }
 
-static void
+static int
 report_lock_acquired_event(struct trace_acquired_event *acquired_event,
 			   const struct perf_sample *sample)
 {
@@ -475,16 +491,23 @@ report_lock_acquired_event(struct trace_acquired_event *acquired_event,
 	u64 contended_term;
 
 	ls = lock_stat_findnew(acquired_event->addr, acquired_event->name);
+	if (!ls)
+		return -1;
 	if (ls->discard)
-		return;
+		return 0;
 
 	ts = thread_stat_findnew(sample->tid);
+	if (!ts)
+		return -1;
+
 	seq = get_seq(ts, acquired_event->addr);
+	if (!seq)
+		return -1;
 
 	switch (seq->state) {
 	case SEQ_STATE_UNINITIALIZED:
 		/* orphan event, do nothing */
-		return;
+		return 0;
 	case SEQ_STATE_ACQUIRING:
 		break;
 	case SEQ_STATE_CONTENDED:
@@ -515,10 +538,10 @@ report_lock_acquired_event(struct trace_acquired_event *acquired_event,
 	ls->nr_acquired++;
 	seq->prev_event_time = timestamp;
 end:
-	return;
+	return 0;
 }
 
-static void
+static int
 report_lock_contended_event(struct trace_contended_event *contended_event,
 			    const struct perf_sample *sample)
 {
@@ -527,16 +550,23 @@ report_lock_contended_event(struct trace_contended_event *contended_event,
 	struct lock_seq_stat *seq;
 
 	ls = lock_stat_findnew(contended_event->addr, contended_event->name);
+	if (!ls)
+		return -1;
 	if (ls->discard)
-		return;
+		return 0;
 
 	ts = thread_stat_findnew(sample->tid);
+	if (!ts)
+		return -1;
+
 	seq = get_seq(ts, contended_event->addr);
+	if (!seq)
+		return -1;
 
 	switch (seq->state) {
 	case SEQ_STATE_UNINITIALIZED:
 		/* orphan event, do nothing */
-		return;
+		return 0;
 	case SEQ_STATE_ACQUIRING:
 		break;
 	case SEQ_STATE_RELEASED:
@@ -559,10 +589,10 @@ report_lock_contended_event(struct trace_contended_event *contended_event,
 	ls->nr_contended++;
 	seq->prev_event_time = sample->time;
 end:
-	return;
+	return 0;
 }
 
-static void
+static int
 report_lock_release_event(struct trace_release_event *release_event,
 			  const struct perf_sample *sample)
 {
@@ -571,11 +601,18 @@ report_lock_release_event(struct trace_release_event *release_event,
 	struct lock_seq_stat *seq;
 
 	ls = lock_stat_findnew(release_event->addr, release_event->name);
+	if (!ls)
+		return -1;
 	if (ls->discard)
-		return;
+		return 0;
 
 	ts = thread_stat_findnew(sample->tid);
+	if (!ts)
+		return -1;
+
 	seq = get_seq(ts, release_event->addr);
+	if (!seq)
+		return -1;
 
 	switch (seq->state) {
 	case SEQ_STATE_UNINITIALIZED:
@@ -609,7 +646,7 @@ free_seq:
 	list_del(&seq->list);
 	free(seq);
 end:
-	return;
+	return 0;
 }
 
 /* lock oriented handlers */
@@ -623,13 +660,14 @@ static struct trace_lock_handler report_lock_ops  = {
 
 static struct trace_lock_handler *trace_handler;
 
-static void perf_evsel__process_lock_acquire(struct perf_evsel *evsel,
+static int perf_evsel__process_lock_acquire(struct perf_evsel *evsel,
 					     struct perf_sample *sample)
 {
 	struct trace_acquire_event acquire_event;
 	struct event_format *event = evsel->tp_format;
 	void *data = sample->raw_data;
 	u64 tmp;		/* this is required for casting... */
+	int rc = 0;
 
 	tmp = raw_field_value(event, "lockdep_addr", data);
 	memcpy(&acquire_event.addr, &tmp, sizeof(void *));
@@ -637,70 +675,84 @@ static void perf_evsel__process_lock_acquire(struct perf_evsel *evsel,
 	acquire_event.flag = (int)raw_field_value(event, "flag", data);
 
 	if (trace_handler->acquire_event)
-		trace_handler->acquire_event(&acquire_event, sample);
+		rc = trace_handler->acquire_event(&acquire_event, sample);
+
+	return rc;
 }
 
-static void perf_evsel__process_lock_acquired(struct perf_evsel *evsel,
+static int perf_evsel__process_lock_acquired(struct perf_evsel *evsel,
 					      struct perf_sample *sample)
 {
 	struct trace_acquired_event acquired_event;
 	struct event_format *event = evsel->tp_format;
 	void *data = sample->raw_data;
 	u64 tmp;		/* this is required for casting... */
+	int rc = 0;
 
 	tmp = raw_field_value(event, "lockdep_addr", data);
 	memcpy(&acquired_event.addr, &tmp, sizeof(void *));
 	acquired_event.name = (char *)raw_field_ptr(event, "name", data);
 
-	if (trace_handler->acquire_event)
-		trace_handler->acquired_event(&acquired_event, sample);
+	if (trace_handler->acquired_event)
+		rc = trace_handler->acquired_event(&acquired_event, sample);
+
+	return rc;
 }
 
-static void perf_evsel__process_lock_contended(struct perf_evsel *evsel,
+static int perf_evsel__process_lock_contended(struct perf_evsel *evsel,
 					       struct perf_sample *sample)
 {
 	struct trace_contended_event contended_event;
 	struct event_format *event = evsel->tp_format;
 	void *data = sample->raw_data;
 	u64 tmp;		/* this is required for casting... */
+	int rc = 0;
 
 	tmp = raw_field_value(event, "lockdep_addr", data);
 	memcpy(&contended_event.addr, &tmp, sizeof(void *));
 	contended_event.name = (char *)raw_field_ptr(event, "name", data);
 
-	if (trace_handler->acquire_event)
-		trace_handler->contended_event(&contended_event, sample);
+	if (trace_handler->contended_event)
+		rc = trace_handler->contended_event(&contended_event, sample);
+
+	return rc;
 }
 
-static void perf_evsel__process_lock_release(struct perf_evsel *evsel,
+static int perf_evsel__process_lock_release(struct perf_evsel *evsel,
 					     struct perf_sample *sample)
 {
 	struct trace_release_event release_event;
 	struct event_format *event = evsel->tp_format;
 	void *data = sample->raw_data;
 	u64 tmp;		/* this is required for casting... */
+	int rc = 0;
 
 	tmp = raw_field_value(event, "lockdep_addr", data);
 	memcpy(&release_event.addr, &tmp, sizeof(void *));
 	release_event.name = (char *)raw_field_ptr(event, "name", data);
 
-	if (trace_handler->acquire_event)
-		trace_handler->release_event(&release_event, sample);
+	if (trace_handler->release_event)
+		rc = trace_handler->release_event(&release_event, sample);
+
+	return rc;
 }
 
-static void perf_evsel__process_lock_event(struct perf_evsel *evsel,
+static int perf_evsel__process_lock_event(struct perf_evsel *evsel,
 					   struct perf_sample *sample)
 {
 	struct event_format *event = evsel->tp_format;
+	int rc = 0;
 
 	if (!strcmp(event->name, "lock_acquire"))
-		perf_evsel__process_lock_acquire(evsel, sample);
+		rc = perf_evsel__process_lock_acquire(evsel, sample);
 	if (!strcmp(event->name, "lock_acquired"))
-		perf_evsel__process_lock_acquired(evsel, sample);
+		rc = perf_evsel__process_lock_acquired(evsel, sample);
 	if (!strcmp(event->name, "lock_contended"))
-		perf_evsel__process_lock_contended(evsel, sample);
+		rc = perf_evsel__process_lock_contended(evsel, sample);
 	if (!strcmp(event->name, "lock_release"))
-		perf_evsel__process_lock_release(evsel, sample);
+		rc = perf_evsel__process_lock_release(evsel, sample);
+
+	return rc;
 }
 
 static void print_bad_events(int bad, int total)
@@ -802,14 +854,20 @@ static void dump_map(void)
 	}
 }
 
-static void dump_info(void)
+static int dump_info(void)
 {
+	int rc = 0;
+
 	if (info_threads)
 		dump_threads();
 	else if (info_map)
 		dump_map();
-	else
-		die("Unknown type of information\n");
+	else {
+		rc = -1;
+		pr_err("Unknown type of information\n");
+	}
+
+	return rc;
 }
 
 static int process_sample_event(struct perf_tool *tool __used,
@@ -826,8 +884,7 @@ static int process_sample_event(struct perf_tool *tool __used,
 		return -1;
 	}
 
-	perf_evsel__process_lock_event(evsel, sample);
-	return 0;
+	return perf_evsel__process_lock_event(evsel, sample);
 }
 
 static struct perf_tool eops = {
@@ -839,8 +896,10 @@ static struct perf_tool eops = {
 static int read_events(void)
 {
 	session = perf_session__new(input_name, O_RDONLY, 0, false, &eops);
-	if (!session)
-		die("Initializing perf session failed\n");
+	if (!session) {
+		pr_err("Initializing perf session failed\n");
+		return -1;
+	}
 
 	return perf_session__process_events(session, &eops);
 }
@@ -857,13 +916,18 @@ static void sort_result(void)
 	}
 }
 
-static void __cmd_report(void)
+static int __cmd_report(void)
 {
 	setup_pager();
-	select_key();
-	read_events();
+
+	if ((select_key() != 0) ||
+	    (read_events() != 0))
+		return -1;
+
 	sort_result();
 	print_result();
+
+	return 0;
 }
 
 static const char * const report_usage[] = {
@@ -959,6 +1023,7 @@ static int __cmd_record(int argc, const char **argv)
 int cmd_lock(int argc, const char **argv, const char *prefix __used)
 {
 	unsigned int i;
+	int rc = 0;
 
 	symbol__init();
 	for (i = 0; i < LOCKHASH_SIZE; i++)
@@ -993,11 +1058,13 @@ int cmd_lock(int argc, const char **argv, const char *prefix __used)
 		/* recycling report_lock_ops */
 		trace_handler = &report_lock_ops;
 		setup_pager();
-		read_events();
-		dump_info();
+		if (read_events() != 0)
+			rc = -1;
+		else
+			rc = dump_info();
 	} else {
 		usage_with_options(lock_usage, lock_options);
 	}
 
-	return 0;
+	return rc;
 }
-- 
1.7.1

[PATCH 05/17] perf stat: Remove use of die/exit and handle errors

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows perf to clean up properly on program termination.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-5-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c |    7 ++++---
 1 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index d53d8ab..02f49eb 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -428,7 +428,7 @@ static int run_perf_stat(int argc __used, const char **argv)
 
 	if (forks && (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0)) {
 		perror("failed to create pipes");
-		exit(1);
+		return -1;
 	}
 
 	if (forks) {
@@ -510,7 +510,8 @@ static int run_perf_stat(int argc __used, const char **argv)
 			}
 			if (child_pid != -1)
 				kill(child_pid, SIGTERM);
-			die("Not all events could be opened.\n");
+
+			pr_err("Not all events could be opened.\n");
 			return -1;
 		}
 		counter->supported = true;
@@ -1189,7 +1190,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 		output = fopen(output_name, mode);
 		if (!output) {
 			perror("failed to create output file");
-			exit(-1);
+			return -1;
 		}
 		clock_gettime(CLOCK_REALTIME, &tm);
 		fprintf(output, "# started on %s\n", ctime(&tm.tv_sec));
-- 
1.7.1

[PATCH 06/17] perf help: Remove use of die and handle errors

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows perf to clean up properly on exit.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-6-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-help.c |   48 +++++++++++++++++++++++++++++++-------------
 1 files changed, 34 insertions(+), 14 deletions(-)

diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c
index 6d5a8a7..f9daae5 100644
--- a/tools/perf/builtin-help.c
+++ b/tools/perf/builtin-help.c
@@ -24,13 +24,14 @@ static struct man_viewer_info_list {
 } *man_viewer_info_list;
 
 enum help_format {
+	HELP_FORMAT_NONE,
 	HELP_FORMAT_MAN,
 	HELP_FORMAT_INFO,
 	HELP_FORMAT_WEB,
 };
 
 static bool show_all = false;
-static enum help_format help_format = HELP_FORMAT_MAN;
+static enum help_format help_format = HELP_FORMAT_NONE;
 static struct option builtin_help_options[] = {
 	OPT_BOOLEAN('a', "all", &show_all, "print all available commands"),
 	OPT_SET_UINT('m', "man", &help_format, "show man page", HELP_FORMAT_MAN),
@@ -54,7 +55,9 @@ static enum help_format parse_help_format(const char *format)
 		return HELP_FORMAT_INFO;
 	if (!strcmp(format, "web") || !strcmp(format, "html"))
 		return HELP_FORMAT_WEB;
-	die("unrecognized help format '%s'", format);
+
+	pr_err("unrecognized help format '%s'", format);
+	return HELP_FORMAT_NONE;
 }
 
 static const char *get_man_viewer_info(const char *name)
@@ -259,6 +262,8 @@ static int perf_help_config(const char *var, const char *value, void *cb)
 		if (!value)
 			return config_error_nonbool(var);
 		help_format = parse_help_format(value);
+		if (help_format == HELP_FORMAT_NONE)
+			return -1;
 		return 0;
 	}
 	if (!strcmp(var, "man.viewer")) {
@@ -352,7 +357,7 @@ static void exec_viewer(const char *name, const char *page)
 		warning("'%s': unknown man viewer.", name);
 }
 
-static void show_man_page(const char *perf_cmd)
+static int show_man_page(const char *perf_cmd)
 {
 	struct man_viewer_list *viewer;
 	const char *page = cmd_to_page(perf_cmd);
@@ -365,28 +370,35 @@ static void show_man_page(const char *perf_cmd)
 	if (fallback)
 		exec_viewer(fallback, page);
 	exec_viewer("man", page);
-	die("no man viewer handled the request");
+
+	pr_err("no man viewer handled the request");
+	return -1;
 }
 
-static void show_info_page(const char *perf_cmd)
+static int show_info_page(const char *perf_cmd)
 {
 	const char *page = cmd_to_page(perf_cmd);
 	setenv("INFOPATH", system_path(PERF_INFO_PATH), 1);
 	execlp("info", "info", "perfman", page, NULL);
+	return -1;
 }
 
-static void get_html_page_path(struct strbuf *page_path, const char *page)
+static int get_html_page_path(struct strbuf *page_path, const char *page)
 {
 	struct stat st;
 	const char *html_path = system_path(PERF_HTML_PATH);
 
 	/* Check that we have a perf documentation directory. */
 	if (stat(mkpath("%s/perf.html", html_path), &st)
-	    || !S_ISREG(st.st_mode))
-		die("'%s': not a documentation directory.", html_path);
+	    || !S_ISREG(st.st_mode)) {
+		pr_err("'%s': not a documentation directory.", html_path);
+		return -1;
+	}
 
 	strbuf_init(page_path, 0);
 	strbuf_addf(page_path, "%s/%s.html", html_path, page);
+
+	return 0;
 }
 
 /*
@@ -401,19 +413,23 @@ static void open_html(const char *path)
 }
 #endif
 
-static void show_html_page(const char *perf_cmd)
+static int show_html_page(const char *perf_cmd)
 {
 	const char *page = cmd_to_page(perf_cmd);
 	struct strbuf page_path; /* it leaks but we exec bellow */
 
-	get_html_page_path(&page_path, page);
+	if (get_html_page_path(&page_path, page) != 0)
+		return -1;
 
 	open_html(page_path.buf);
+
+	return 0;
 }
 
 int cmd_help(int argc, const char **argv, const char *prefix __used)
 {
 	const char *alias;
+	int rc = 0;
 
 	load_command_list("perf-", &main_cmds, &other_cmds);
 
@@ -444,16 +460,20 @@ int cmd_help(int argc, const char **argv, const char *prefix __used)
 
 	switch (help_format) {
 	case HELP_FORMAT_MAN:
-		show_man_page(argv[0]);
+		rc = show_man_page(argv[0]);
 		break;
 	case HELP_FORMAT_INFO:
-		show_info_page(argv[0]);
+		rc = show_info_page(argv[0]);
 		break;
 	case HELP_FORMAT_WEB:
-		show_html_page(argv[0]);
+		rc = show_html_page(argv[0]);
+		break;
+	case HELP_FORMAT_NONE:
+		/* fall-through */
 	default:
+		rc = -1;
 		break;
 	}
 
-	return 0;
+	return rc;
 }
-- 
1.7.1

[PATCH 07/17] perf script: Remove use of die/exit

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows perf to clean up properly on exit. Only exits left are exec
failures which are appropriate and usage callbacks that list available
options.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-7-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-script.c |   60 +++++++++++++++++++++++++++++-------------
 1 files changed, 41 insertions(+), 19 deletions(-)

diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 2d6e3b2..c350cfe 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -1153,18 +1153,23 @@ static const struct option options[] = {
 	OPT_END()
 };
 
-static bool have_cmd(int argc, const char **argv)
+static int have_cmd(int argc, const char **argv)
 {
 	char **__argv = malloc(sizeof(const char *) * argc);
 
-	if (!__argv)
-		die("malloc");
+	if (!__argv) {
+		pr_err("malloc failed\n");
+		return -1;
+	}
+
 	memcpy(__argv, argv, sizeof(const char *) * argc);
 	argc = parse_options(argc, (const char **)__argv, record_options,
 			     NULL, PARSE_OPT_STOP_AT_NON_OPTION);
 	free(__argv);
 
-	return argc != 0;
+	system_wide = (argc == 0);
+
+	return 0;
 }
 
 int cmd_script(int argc, const char **argv, const char *prefix __used)
@@ -1231,13 +1236,13 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
 
 		if (pipe(live_pipe) < 0) {
 			perror("failed to create pipe");
-			exit(-1);
+			return -1;
 		}
 
 		pid = fork();
 		if (pid < 0) {
 			perror("failed to fork");
-			exit(-1);
+			return -1;
 		}
 
 		if (!pid) {
@@ -1249,13 +1254,18 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
 			if (is_top_script(argv[0])) {
 				system_wide = true;
 			} else if (!system_wide) {
-				system_wide = !have_cmd(argc - rep_args,
-							&argv[rep_args]);
+				if (have_cmd(argc - rep_args, &argv[rep_args]) != 0) {
+					err = -1;
+					goto out;
+				}
 			}
 
 			__argv = malloc((argc + 6) * sizeof(const char *));
-			if (!__argv)
-				die("malloc");
+			if (!__argv) {
+				pr_err("malloc failed\n");
+				err = -ENOMEM;
+				goto out;
+			}
 
 			__argv[j++] = "/bin/sh";
 			__argv[j++] = rec_script_path;
@@ -1277,8 +1287,12 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
 		close(live_pipe[1]);
 
 		__argv = malloc((argc + 4) * sizeof(const char *));
-		if (!__argv)
-			die("malloc");
+		if (!__argv) {
+			pr_err("malloc failed\n");
+			err = -ENOMEM;
+			goto out;
+		}
+
 		j = 0;
 		__argv[j++] = "/bin/sh";
 		__argv[j++] = rep_script_path;
@@ -1303,12 +1317,20 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
 
 		if (!rec_script_path)
 			system_wide = false;
-		else if (!system_wide)
-			system_wide = !have_cmd(argc - 1, &argv[1]);
+		else if (!system_wide) {
+			if (have_cmd(argc - 1, &argv[1]) != 0) {
+				err = -1;
+				goto out;
+			}
+		}
 
 		__argv = malloc((argc + 2) * sizeof(const char *));
-		if (!__argv)
-			die("malloc");
+		if (!__argv) {
+			pr_err("malloc failed\n");
+			err = -ENOMEM;
+			goto out;
+		}
+
 		__argv[j++] = "/bin/sh";
 		__argv[j++] = script_path;
 		if (system_wide)
@@ -1357,18 +1379,18 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
 		input = open(session->filename, O_RDONLY);	/* input_name */
 		if (input < 0) {
 			perror("failed to open file");
-			exit(-1);
+			return -1;
 		}
 
 		err = fstat(input, &perf_stat);
 		if (err < 0) {
 			perror("failed to stat file");
-			exit(-1);
+			return -1;
 		}
 
 		if (!perf_stat.st_size) {
 			fprintf(stderr, "zero-sized file, nothing to do!\n");
-			exit(0);
+			return 0;
 		}
 
 		scripting_ops = script_spec__lookup(generate_script_lang);
-- 
1.7.1

[PATCH 08/17] perf record: Remove use of die/exit

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Allows perf to clean up properly on exit. If perf-record is exiting due
to failure, the on_exit should not run as the session has been deleted.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346005487-62961-8-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |  158 +++++++++++++++++++++++++++++-------------
 1 files changed, 109 insertions(+), 49 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 479ff2a..7b8b891 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -71,19 +71,23 @@ static void advance_output(struct perf_record *rec, size_t size)
 	rec->bytes_written += size;
 }
 
-static void write_output(struct perf_record *rec, void *buf, size_t size)
+static int write_output(struct perf_record *rec, void *buf, size_t size)
 {
 	while (size) {
 		int ret = write(rec->output, buf, size);
 
-		if (ret < 0)
-			die("failed to write");
+		if (ret < 0) {
+			pr_err("failed to write\n");
+			return -1;
+		}
 
 		size -= ret;
 		buf += ret;
 
 		rec->bytes_written += ret;
 	}
+
+	return 0;
 }
 
 static int process_synthesized_event(struct perf_tool *tool,
@@ -92,11 +96,13 @@ static int process_synthesized_event(struct perf_tool *tool,
 				     struct machine *machine __used)
 {
 	struct perf_record *rec = container_of(tool, struct perf_record, tool);
-	write_output(rec, event, event->header.size);
+	if (write_output(rec, event, event->header.size) < 0)
+		return -1;
+
 	return 0;
 }
 
-static void perf_record__mmap_read(struct perf_record *rec,
+static int perf_record__mmap_read(struct perf_record *rec,
 				   struct perf_mmap *md)
 {
 	unsigned int head = perf_mmap__read_head(md);
@@ -104,9 +110,10 @@ static void perf_record__mmap_read(struct perf_record *rec,
 	unsigned char *data = md->base + rec->page_size;
 	unsigned long size;
 	void *buf;
+	int rc = 0;
 
 	if (old == head)
-		return;
+		return 0;
 
 	rec->samples++;
 
@@ -117,17 +124,26 @@ static void perf_record__mmap_read(struct perf_record *rec,
 		size = md->mask + 1 - (old & md->mask);
 		old += size;
 
-		write_output(rec, buf, size);
+		if (write_output(rec, buf, size) < 0) {
+			rc = -1;
+			goto out;
+		}
 	}
 
 	buf = &data[old & md->mask];
 	size = head - old;
 	old += size;
 
-	write_output(rec, buf, size);
+	if (write_output(rec, buf, size) < 0) {
+		rc = -1;
+		goto out;
+	}
 
 	md->prev = old;
 	perf_mmap__write_tail(md, old);
+
+out:
+	return rc;
 }
 
 static volatile int done = 0;
@@ -183,12 +199,13 @@ static bool perf_evlist__equal(struct perf_evlist *evlist,
 	return true;
 }
 
-static void perf_record__open(struct perf_record *rec)
+static int perf_record__open(struct perf_record *rec)
 {
 	struct perf_evsel *pos;
 	struct perf_evlist *evlist = rec->evlist;
 	struct perf_session *session = rec->session;
 	struct perf_record_opts *opts = &rec->opts;
+	int rc = 0;
 
 	perf_evlist__config_attrs(evlist, opts);
 
@@ -222,10 +239,13 @@ try_again:
 
 			if (err == EPERM || err == EACCES) {
 				ui__error_paranoid();
-				exit(EXIT_FAILURE);
+				rc = -err;
+				goto out;
 			} else if (err ==  ENODEV && opts->target.cpu_list) {
-				die("No such device - did you specify"
-					" an out-of-range profile CPU?\n");
+				pr_err("No such device - did you specify"
+				       " an out-of-range profile CPU?\n");
+				rc = -err;
+				goto out;
 			} else if (err == EINVAL) {
 				if (!opts->exclude_guest_missing &&
 				    (attr->exclude_guest || attr->exclude_host)) {
@@ -272,7 +292,8 @@ try_again:
 			if (err == ENOENT) {
 				ui__error("The %s event is not supported.\n",
 					  perf_evsel__name(pos));
-				exit(EXIT_FAILURE);
+				rc = -err;
+				goto out;
 			}
 
 			printf("\n");
@@ -280,34 +301,46 @@ try_again:
 			      err, strerror(err));
 
 #if defined(__i386__) || defined(__x86_64__)
-			if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP)
-				die("No hardware sampling interrupt available."
-				    " No APIC? If so then you can boot the kernel"
-				    " with the \"lapic\" boot parameter to"
-				    " force-enable it.\n");
+			if (attr->type == PERF_TYPE_HARDWARE &&
+			    err == EOPNOTSUPP) {
+				pr_err("No hardware sampling interrupt available."
+				       " No APIC? If so then you can boot the kernel"
+				       " with the \"lapic\" boot parameter to"
+				       " force-enable it.\n");
+				rc = -err;
+				goto out;
+			}
 #endif
 
-			die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
+			pr_err("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
+			rc = -err;
+			goto out;
 		}
 	}
 
 	if (perf_evlist__set_filters(evlist)) {
 		error("failed to set filter with %d (%s)\n", errno,
 			strerror(errno));
-		exit(-1);
+		rc = -1;
+		goto out;
 	}
 
 	if (perf_evlist__mmap(evlist, opts->mmap_pages, false) < 0) {
-		if (errno == EPERM)
-			die("Permission error mapping pages.\n"
-			    "Consider increasing "
-			    "/proc/sys/kernel/perf_event_mlock_kb,\n"
-			    "or try again with a smaller value of -m/--mmap_pages.\n"
-			    "(current value: %d)\n", opts->mmap_pages);
-		else if (!is_power_of_2(opts->mmap_pages))
-			die("--mmap_pages/-m value must be a power of two.");
-
-		die("failed to mmap with %d (%s)\n", errno, strerror(errno));
+		if (errno == EPERM) {
+			pr_err("Permission error mapping pages.\n"
+			       "Consider increasing "
+			       "/proc/sys/kernel/perf_event_mlock_kb,\n"
+			       "or try again with a smaller value of -m/--mmap_pages.\n"
+			       "(current value: %d)\n", opts->mmap_pages);
+			rc = -errno;
+		} else if (!is_power_of_2(opts->mmap_pages)) {
+			pr_err("--mmap_pages/-m value must be a power of two.");
+			rc = -EINVAL;
+		} else {
+			pr_err("failed to mmap with %d (%s)\n", errno, strerror(errno));
+			rc = -errno;
+		}
+		goto out;
 	}
 
 	if (rec->file_new)
@@ -315,11 +348,14 @@ try_again:
 	else {
 		if (!perf_evlist__equal(session->evlist, evlist)) {
 			fprintf(stderr, "incompatible append\n");
-			exit(-1);
+			rc = -1;
+			goto out;
 		}
  	}
 
 	perf_session__set_id_hdr_size(session);
+out:
+	return rc;
 }
 
 static int process_buildids(struct perf_record *rec)
@@ -335,10 +371,13 @@ static int process_buildids(struct perf_record *rec)
 					      size, &build_id__mark_dso_hit_ops);
 }
 
-static void perf_record__exit(int status __used, void *arg)
+static void perf_record__exit(int status, void *arg)
 {
 	struct perf_record *rec = arg;
 
+	if (status != 0)
+		return;
+
 	if (!rec->opts.pipe_output) {
 		rec->session->header.data_size += rec->bytes_written;
 
@@ -393,17 +432,26 @@ static struct perf_event_header finished_round_event = {
 	.type = PERF_RECORD_FINISHED_ROUND,
 };
 
-static void perf_record__mmap_read_all(struct perf_record *rec)
+static int perf_record__mmap_read_all(struct perf_record *rec)
 {
 	int i;
+	int rc = 0;
 
 	for (i = 0; i < rec->evlist->nr_mmaps; i++) {
-		if (rec->evlist->mmap[i].base)
-			perf_record__mmap_read(rec, &rec->evlist->mmap[i]);
+		if (rec->evlist->mmap[i].base) {
+			if (perf_record__mmap_read(rec, &rec->evlist->mmap[i]) != 0) {
+				rc = -1;
+				goto out;
+			}
+		}
 	}
 
 	if (perf_header__has_feat(&rec->session->header, HEADER_TRACING_DATA))
-		write_output(rec, &finished_round_event, sizeof(finished_round_event));
+		rc = write_output(rec, &finished_round_event,
+				  sizeof(finished_round_event));
+
+out:
+	return rc;
 }
 
 static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
@@ -463,7 +511,7 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 		output = open(output_name, flags, S_IRUSR | S_IWUSR);
 	if (output < 0) {
 		perror("failed to create output file");
-		exit(-1);
+		return -1;
 	}
 
 	rec->output = output;
@@ -503,7 +551,10 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 		}
 	}
 
-	perf_record__open(rec);
+	if (perf_record__open(rec) != 0) {
+		err = -1;
+		goto out_delete_session;
+	}
 
 	/*
 	 * perf_session__delete(session) will be called at perf_record__exit()
@@ -513,19 +564,20 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 	if (opts->pipe_output) {
 		err = perf_header__write_pipe(output);
 		if (err < 0)
-			return err;
+			goto out_delete_session;
 	} else if (rec->file_new) {
 		err = perf_session__write_header(session, evsel_list,
 						 output, false);
 		if (err < 0)
-			return err;
+			goto out_delete_session;
 	}
 
 	if (!rec->no_buildid
 	    && !perf_header__has_feat(&session->header, HEADER_BUILD_ID)) {
 		pr_err("Couldn't generate buildids. "
 		       "Use --no-buildid to profile anyway.\n");
-		return -1;
+		err = -1;
+		goto out_delete_session;
 	}
 
 	rec->post_processing_offset = lseek(output, 0, SEEK_CUR);
@@ -533,7 +585,8 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 	machine = perf_session__find_host_machine(session);
 	if (!machine) {
 		pr_err("Couldn't find native kernel information.\n");
-		return -1;
+		err = -1;
+		goto out_delete_session;
 	}
 
 	if (opts->pipe_output) {
@@ -541,14 +594,14 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 						   process_synthesized_event);
 		if (err < 0) {
 			pr_err("Couldn't synthesize attrs.\n");
-			return err;
+			goto out_delete_session;
 		}
 
 		err = perf_event__synthesize_event_types(tool, process_synthesized_event,
 							 machine);
 		if (err < 0) {
 			pr_err("Couldn't synthesize event_types.\n");
-			return err;
+			goto out_delete_session;
 		}
 
 		if (have_tracepoints(&evsel_list->entries)) {
@@ -564,7 +617,7 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 								  process_synthesized_event);
 			if (err <= 0) {
 				pr_err("Couldn't record tracing data.\n");
-				return err;
+				goto out_delete_session;
 			}
 			advance_output(rec, err);
 		}
@@ -592,20 +645,24 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 					       perf_event__synthesize_guest_os);
 
 	if (!opts->target.system_wide)
-		perf_event__synthesize_thread_map(tool, evsel_list->threads,
+		err = perf_event__synthesize_thread_map(tool, evsel_list->threads,
 						  process_synthesized_event,
 						  machine);
 	else
-		perf_event__synthesize_threads(tool, process_synthesized_event,
+		err = perf_event__synthesize_threads(tool, process_synthesized_event,
 					       machine);
 
+	if (err != 0)
+		goto out_delete_session;
+
 	if (rec->realtime_prio) {
 		struct sched_param param;
 
 		param.sched_priority = rec->realtime_prio;
 		if (sched_setscheduler(0, SCHED_FIFO, &param)) {
 			pr_err("Could not set realtime priority.\n");
-			exit(-1);
+			err = -1;
+			goto out_delete_session;
 		}
 	}
 
@@ -620,7 +677,10 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv)
 	for (;;) {
 		int hits = rec->samples;
 
-		perf_record__mmap_read_all(rec);
+		if (perf_record__mmap_read_all(rec) < 0) {
+			err = -1;
+			goto out_delete_session;
+		}
 
 		if (hits == rec->samples) {
 			if (done)
-- 
1.7.1

[PATCH 09/17] perf tools: Fix x86 builds with ARCH specified on the command line

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

e.g., compiling i386 on x86_64 using:
$ make -C tools/perf ARCH=i386

fails with:

    CC /tmp/pbuild/util/evsel.o
In file included from util/evsel.c:21:0:
util/perf_regs.h:5:23: fatal error: perf_regs.h: No such file or directory
compilation terminated.

Adding V=1 you see that the include argument for the arch is
'-Iarch/i386/include' is wrong. It is supposed to be -Iarch/x86/include
per the redefinition of ARCH in the Makefile.

According to the make manual,
http://www.gnu.org/software/make/manual/make.html#Override-Directive:
  "If a variable has been set with a command argument (see Overriding
   Variables), then ordinary assignments in the makefile are ignored. If
   you want to set the variable in the makefile even though it was set
   with a command argument, you can use an override directive ..."

Make it so.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346094354-74356-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 722ddee..939cf6d 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -64,12 +64,12 @@ AR = $(CROSS_COMPILE)ar
 
 # Additional ARCH settings for x86
 ifeq ($(ARCH),i386)
-	ARCH := x86
+	override ARCH := x86
 	NO_PERF_REGS := 0
 	LIBUNWIND_LIBS = -lunwind -lunwind-x86
 endif
 ifeq ($(ARCH),x86_64)
-	ARCH := x86
+	override ARCH := x86
 	IS_X86_64 := 0
 	ifeq (, $(findstring m32,$(EXTRA_CFLAGS)))
 		IS_X86_64 := $(shell echo __x86_64__ | ${CC} -E -xc - | tail -n 1)
-- 
1.7.1

[PATCH 10/17] perf tools: Fix intlist node removal

[Arnaldo Carvalho de Melo]

From: Suzuki K. Poulose <suzuki@in.ibm.com>

Similar to the one in :
	https://lkml.org/lkml/2012/8/29/27

Make sure we remove the node from the rblist before we delete the node.
The rblist__remove_node() will invoke rblist->node_delete, which  will
take care of deleting the node with the suitable function provided by
the user.

Signed-off-by: Suzuki K Poulose <suzuki@in.ibm.com>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Suzuki K Poulose <suzuki@in.ibm.com>
Link: http://lkml.kernel.org/r/20120831065840.5167.90318.stgit@suzukikp.in.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/intlist.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/intlist.c b/tools/perf/util/intlist.c
index fd530dc..77c504f 100644
--- a/tools/perf/util/intlist.c
+++ b/tools/perf/util/intlist.c
@@ -52,9 +52,9 @@ int intlist__add(struct intlist *ilist, int i)
 	return rblist__add_node(&ilist->rblist, (void *)((long)i));
 }
 
-void intlist__remove(struct intlist *ilist __used, struct int_node *node)
+void intlist__remove(struct intlist *ilist, struct int_node *node)
 {
-	int_node__delete(node);
+	rblist__remove_node(&ilist->rblist, &node->rb_node);
 }
 
 struct int_node *intlist__find(struct intlist *ilist, int i)
-- 
1.7.1

[PATCH 11/17] perf tools: Remove the node from rblist in strlist__remove

[Arnaldo Carvalho de Melo]

From: Suzuki K. Poulose <suzuki@in.ibm.com>

The following commit:

author	David Ahern <dsahern@gmail.com>
	Tue, 31 Jul 2012 04:31:33 +0000 (22:31 -0600)
committer	Arnaldo Carvalho de Melo <acme@redhat.com>
	Fri, 3 Aug 2012 13:39:51 +0000 (10:39 -0300)
commit	ee8dd3ca43f151d9fbe1edeef68fb8a77eb9f047

causes a double free during a probe deletion as the node is never
removed from the list via strlist__remove(), even though it gets
'deleted' (read free()'d). This causes a double free when we do
strlist__delete() as the node is already deleted but present in the
rblist.

[suzukikp@suzukikp perf]$ sudo ./perf probe -a do_fork
Added new event:
  probe:do_fork        (on do_fork)

You can now use it in all perf tools, such as:

	perf record -e probe:do_fork -aR sleep 1

[suzukikp@suzukikp perf]$ sudo ./perf probe -d do_fork
Removed event: probe:do_fork
*** glibc detected *** ./perf: double free or corruption (fasttop): 0x000000000133d600 ***
======= Backtrace: =========
/lib64/libc.so.6[0x38eec7dda6]
./perf(rblist__delete+0x5c)[0x47d3dc]
./perf(del_perf_probe_events+0xb6)[0x47b826]
./perf(cmd_probe+0x471)[0x42c8d1]
./perf[0x4150b3]
./perf(main+0x501)[0x4148e1]
/lib64/libc.so.6(__libc_start_main+0xed)[0x38eec2169d]
./perf[0x414a61]

Make sure we remove the node from the rblist before we delete the node.
The rblist__remove_node() will invoke rblist->node_delete, which  will
take care of deleting the node with the suitable function provided by
the user.

Reported-by: Ananth N. Mavinakayanahalli <ananth@in.ibm.com>
Signed-off-by: Suzuki K. Poulose <suzuki@in.ibm.com>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20120829055840.7802.1459.stgit@suzukikp.in.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/strlist.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/tools/perf/util/strlist.c b/tools/perf/util/strlist.c
index 95856ff..155d8b7 100644
--- a/tools/perf/util/strlist.c
+++ b/tools/perf/util/strlist.c
@@ -93,7 +93,7 @@ out:
 
 void strlist__remove(struct strlist *slist, struct str_node *snode)
 {
-	str_node__delete(snode, slist->dupstr);
+	rblist__remove_node(&slist->rblist, &snode->rb_node);
 }
 
 struct str_node *strlist__find(struct strlist *slist, const char *entry)
-- 
1.7.1

[PATCH 12/17] perf tools: remove unneeded include of network header files

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

perf does not have networking related functionality, and the inclusion
of these headers is one of the causes of compile failures for Android:

  https://lkml.org/lkml/2012/8/23/316
  https://lkml.org/lkml/2012/8/28/293

So, remove them.

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1346255732-93246-1-git-send-email-dsahern@gmail.com
[ committer note: fix trace-event-perl.c compile failure by reordering includes ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 .../perf/util/scripting-engines/trace-event-perl.c |    8 ++++----
 tools/perf/util/util.h                             |    5 -----
 2 files changed, 4 insertions(+), 9 deletions(-)

diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
index d280010..94e6736 100644
--- a/tools/perf/util/scripting-engines/trace-event-perl.c
+++ b/tools/perf/util/scripting-engines/trace-event-perl.c
@@ -25,16 +25,16 @@
 #include <ctype.h>
 #include <errno.h>
 
-#include "../../perf.h"
 #include "../util.h"
+#include <EXTERN.h>
+#include <perl.h>
+
+#include "../../perf.h"
 #include "../thread.h"
 #include "../event.h"
 #include "../trace-event.h"
 #include "../evsel.h"
 
-#include <EXTERN.h>
-#include <perl.h>
-
 void boot_Perf__Trace__Context(pTHX_ CV *cv);
 void boot_DynaLoader(pTHX_ CV *cv);
 typedef PerlInterpreter * INTERP;
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index 00a93a9..67a3713 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -69,11 +69,6 @@
 #include <sys/poll.h>
 #include <sys/socket.h>
 #include <sys/ioctl.h>
-#include <sys/select.h>
-#include <netinet/in.h>
-#include <netinet/tcp.h>
-#include <arpa/inet.h>
-#include <netdb.h>
 #include <inttypes.h>
 #include "../../../include/linux/magic.h"
 #include "types.h"
-- 
1.7.1

[PATCH 13/17] perf header: Use evlist->nr_entries on write_event_desc()

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

Number of events (evsels) in a evlist is kept on nr_entries field
so that we don't need to recalculate it.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1346821373-31621-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/header.c |    5 ++---
 1 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index 9696e64..a124b93 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -610,11 +610,10 @@ static int write_event_desc(int fd, struct perf_header *h __used,
 			    struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
-	u32 nre = 0, nri, sz;
+	u32 nre, nri, sz;
 	int ret;
 
-	list_for_each_entry(evsel, &evlist->entries, node)
-		nre++;
+	nre = evlist->nr_entries;
 
 	/*
 	 * write number of events
-- 
1.7.1

[PATCH 14/17] perf header: Set tracepoint event name only if not set

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

The event name can be set already by processing a event_desc data.

So check it before setting to prevent possible leak.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1346821373-31621-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/header.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index a124b93..05c9310 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -2314,7 +2314,7 @@ static int perf_evlist__set_tracepoint_names(struct perf_evlist *evlist,
 	struct perf_evsel *pos;
 
 	list_for_each_entry(pos, &evlist->entries, node) {
-		if (pos->attr.type == PERF_TYPE_TRACEPOINT &&
+		if (pos->attr.type == PERF_TYPE_TRACEPOINT && !pos->name &&
 		    perf_evsel__set_tracepoint_name(pos, pevent))
 			return -1;
 	}
-- 
1.7.1

[PATCH 15/17] perf header: Swap pmu mapping numbers if needed

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

Like others, the numbers can be saved in a different endian format than
a host machine.  Swap them if needed.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robert Richter <robert.richter@amd.com>
Link: http://lkml.kernel.org/r/1346821373-31621-4-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/header.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index 05c9310..43425b7 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -1440,6 +1440,9 @@ static void print_pmu_mappings(struct perf_header *ph, int fd, FILE *fp)
 	if (ret != sizeof(pmu_num))
 		goto error;
 
+	if (ph->needs_swap)
+		pmu_num = bswap_32(pmu_num);
+
 	if (!pmu_num) {
 		fprintf(fp, "# pmu mappings: not available\n");
 		return;
@@ -1448,6 +1451,9 @@ static void print_pmu_mappings(struct perf_header *ph, int fd, FILE *fp)
 	while (pmu_num) {
 		if (read(fd, &type, sizeof(type)) != sizeof(type))
 			break;
+		if (ph->needs_swap)
+			type = bswap_32(type);
+
 		name = do_read_string(fd, ph);
 		if (!name)
 			break;
-- 
1.7.1

[PATCH 16/17] perf tools: Replace mempcpy with memcpy

[Arnaldo Carvalho de Melo]

From: Irina Tirdea <irina.tirdea@gmail.com>

mempcpy is not supported by bionic in Android and will lead to
compilation errors.

Replacing mempcpy with memcpy so it will work in Android.

Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/CANg8OW+Y3ZMG-GdhYu2_yKOYH_XEMgw73PdCX_23UTnfYhmttA@mail.gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/target.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/target.c b/tools/perf/util/target.c
index 051eaa6..065528b 100644
--- a/tools/perf/util/target.c
+++ b/tools/perf/util/target.c
@@ -117,8 +117,8 @@ int perf_target__strerror(struct perf_target *target, int errnum,
 
 		if (err != buf) {
 			size_t len = strlen(err);
-			char *c = mempcpy(buf, err, min(buflen - 1, len));
-			*c = '\0';
+			memcpy(buf, err, min(buflen - 1, len));
+			*(buf + min(buflen - 1, len)) = '\0';
 		}
 
 		return 0;
-- 
1.7.1

[PATCH 17/17] perf tools: Allow user to indicate path to objdump in command line

[Arnaldo Carvalho de Melo]

From: Maciek Borzecki <maciek.borzecki@gmail.com>

When analyzing perf data from hosts of other architecture than one of
the local host it's useful to call objdump that is part of a toolchain
for that architecture. Instead of calling regular objdump, call one that
user specified in command line.

Signed-off-by: Maciek Borzecki <maciek.borzecki@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/1346754750.16299.3.camel@localhost.localdomain
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-annotate.txt |    3 +++
 tools/perf/Documentation/perf-report.txt   |    3 +++
 tools/perf/builtin-annotate.c              |    2 ++
 tools/perf/builtin-report.c                |    2 ++
 tools/perf/util/annotate.c                 |    4 +++-
 tools/perf/util/annotate.h                 |    1 +
 6 files changed, 14 insertions(+), 1 deletions(-)

diff --git a/tools/perf/Documentation/perf-annotate.txt b/tools/perf/Documentation/perf-annotate.txt
index c89f9e1..c8ffd9f 100644
--- a/tools/perf/Documentation/perf-annotate.txt
+++ b/tools/perf/Documentation/perf-annotate.txt
@@ -85,6 +85,9 @@ OPTIONS
 -M::
 --disassembler-style=:: Set disassembler style for objdump.
 
+--objdump=<path>::
+        Path to objdump binary.
+
 SEE ALSO
 --------
 linkperf:perf-record[1], linkperf:perf-report[1]
diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index 495210a..f4d91be 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -168,6 +168,9 @@ OPTIONS
 	branch stacks and it will automatically switch to the branch view mode,
 	unless --no-branch-stack is used.
 
+--objdump=<path>::
+        Path to objdump binary.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1], linkperf:perf-annotate[1]
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 67522cf..2f3f002 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -282,6 +282,8 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used)
 		    "Display raw encoding of assembly instructions (default)"),
 	OPT_STRING('M', "disassembler-style", &disassembler_style, "disassembler style",
 		   "Specify disassembler style (e.g. -M intel for intel syntax)"),
+	OPT_STRING(0, "objdump", &objdump_path, "path",
+		   "objdump binary to use for disassembly and annotations"),
 	OPT_END()
 	};
 
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index d618253..1f8d11b 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -638,6 +638,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __used)
 		    "Show a column with the sum of periods"),
 	OPT_CALLBACK_NOOPT('b', "branch-stack", &sort__branch_mode, "",
 		    "use branch records for histogram filling", parse_branch_mode),
+	OPT_STRING(0, "objdump", &objdump_path, "path",
+		   "objdump binary to use for disassembly and annotations"),
 	OPT_END()
 	};
 
diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c
index 3a282c0..51ef69c 100644
--- a/tools/perf/util/annotate.c
+++ b/tools/perf/util/annotate.c
@@ -17,6 +17,7 @@
 #include <pthread.h>
 
 const char 	*disassembler_style;
+const char	*objdump_path;
 
 static struct ins *ins__find(const char *name);
 static int disasm_line__parse(char *line, char **namep, char **rawp);
@@ -820,9 +821,10 @@ fallback:
 		 dso, dso->long_name, sym, sym->name);
 
 	snprintf(command, sizeof(command),
-		 "objdump %s%s --start-address=0x%016" PRIx64
+		 "%s %s%s --start-address=0x%016" PRIx64
 		 " --stop-address=0x%016" PRIx64
 		 " -d %s %s -C %s|grep -v %s|expand",
+		 objdump_path ? objdump_path : "objdump",
 		 disassembler_style ? "-M " : "",
 		 disassembler_style ? disassembler_style : "",
 		 map__rip_2objdump(map, sym->start),
diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h
index 78a5692..a6d6bc5 100644
--- a/tools/perf/util/annotate.h
+++ b/tools/perf/util/annotate.h
@@ -152,5 +152,6 @@ int symbol__tui_annotate(struct symbol *sym, struct map *map, int evidx,
 #endif
 
 extern const char	*disassembler_style;
+extern const char	*objdump_path;
 
 #endif	/* __PERF_ANNOTATE_H */
-- 
1.7.1

Re: [PATCH 11/17] perf tools: Remove the node from rblist in strlist__remove

[David Ahern]

This one needs to go through urgent.


On 9/5/12 5:08 PM, Arnaldo Carvalho de Melo wrote:
> From: Suzuki K. Poulose <suzuki@in.ibm.com>
>
> The following commit:
>
> author	David Ahern <dsahern@gmail.com>
> 	Tue, 31 Jul 2012 04:31:33 +0000 (22:31 -0600)
> committer	Arnaldo Carvalho de Melo <acme@redhat.com>
> 	Fri, 3 Aug 2012 13:39:51 +0000 (10:39 -0300)
> commit	ee8dd3ca43f151d9fbe1edeef68fb8a77eb9f047
>
> causes a double free during a probe deletion as the node is never
> removed from the list via strlist__remove(), even though it gets
> 'deleted' (read free()'d). This causes a double free when we do
> strlist__delete() as the node is already deleted but present in the
> rblist.
>
> [suzukikp@suzukikp perf]$ sudo ./perf probe -a do_fork
> Added new event:
>    probe:do_fork        (on do_fork)
>
> You can now use it in all perf tools, such as:
>
> 	perf record -e probe:do_fork -aR sleep 1
>
> [suzukikp@suzukikp perf]$ sudo ./perf probe -d do_fork
> Removed event: probe:do_fork
> *** glibc detected *** ./perf: double free or corruption (fasttop): 0x000000000133d600 ***
> ======= Backtrace: =========
> /lib64/libc.so.6[0x38eec7dda6]
> ./perf(rblist__delete+0x5c)[0x47d3dc]
> ./perf(del_perf_probe_events+0xb6)[0x47b826]
> ./perf(cmd_probe+0x471)[0x42c8d1]
> ./perf[0x4150b3]
> ./perf(main+0x501)[0x4148e1]
> /lib64/libc.so.6(__libc_start_main+0xed)[0x38eec2169d]
> ./perf[0x414a61]
>
> Make sure we remove the node from the rblist before we delete the node.
> The rblist__remove_node() will invoke rblist->node_delete, which  will
> take care of deleting the node with the suitable function provided by
> the user.
>
> Reported-by: Ananth N. Mavinakayanahalli <ananth@in.ibm.com>
> Signed-off-by: Suzuki K. Poulose <suzuki@in.ibm.com>
> Acked-by: David Ahern <dsahern@gmail.com>
> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
> Cc: David Ahern <dsahern@gmail.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Ingo Molnar <mingo@kernel.org>
> Cc: Jiri Olsa <jolsa@redhat.com>
> Cc: Namhyung Kim <namhyung@kernel.org>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Link: http://lkml.kernel.org/r/20120829055840.7802.1459.stgit@suzukikp.in.ibm.com
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---
>   tools/perf/util/strlist.c |    2 +-
>   1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/tools/perf/util/strlist.c b/tools/perf/util/strlist.c
> index 95856ff..155d8b7 100644
> --- a/tools/perf/util/strlist.c
> +++ b/tools/perf/util/strlist.c
> @@ -93,7 +93,7 @@ out:
>
>   void strlist__remove(struct strlist *slist, struct str_node *snode)
>   {
> -	str_node__delete(snode, slist->dupstr);
> +	rblist__remove_node(&slist->rblist, &snode->rb_node);
>   }
>
>   struct str_node *strlist__find(struct strlist *slist, const char *entry)
>

Re: [PATCH 14/17] perf header: Set tracepoint event name only if not set

[David Ahern]

On 9/5/12 5:08 PM, Arnaldo Carvalho de Melo wrote:
> From: Namhyung Kim <namhyung.kim@lge.com>
>
> The event name can be set already by processing a event_desc data.
>
> So check it before setting to prevent possible leak.
>
> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> Cc: Ingo Molnar <mingo@kernel.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Link: http://lkml.kernel.org/r/1346821373-31621-3-git-send-email-namhyung@kernel.org
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---
>   tools/perf/util/header.c |    2 +-
>   1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
> index a124b93..05c9310 100644
> --- a/tools/perf/util/header.c
> +++ b/tools/perf/util/header.c
> @@ -2314,7 +2314,7 @@ static int perf_evlist__set_tracepoint_names(struct perf_evlist *evlist,
>   	struct perf_evsel *pos;
>
>   	list_for_each_entry(pos, &evlist->entries, node) {
> -		if (pos->attr.type == PERF_TYPE_TRACEPOINT &&
> +		if (pos->attr.type == PERF_TYPE_TRACEPOINT && !pos->name &&
>   		    perf_evsel__set_tracepoint_name(pos, pevent))
>   			return -1;
>   	}
>

I'm getting a segfault running perf-script due to this patch. Most 
likely this is exposing another problem. Back trace is:

#0  0x000000000049f4f2 in pevent_event_info (s=0x7fffffffdab0, 
event=0x0, record=0x7fffffffda70)
     at tools/lib/traceevent/event-parse.c:4183
#1  0x000000000046cd52 in event_format__print (event=0x0, cpu=<optimized 
out>, data=<optimized out>, size=<optimized out>)
     at util/trace-event-parse.c:182
#2  0x000000000042d41b in process_event (event=0x7fffef9d1e60, 
sample=0x7fffffffdc90, evsel=0x90d870, machine=0x906f80,
     al=<optimized out>) at builtin-script.c:415
#3  0x000000000042c4e1 in process_sample_event (tool=<optimized out>, 
event=0x7fffef9d1e60, sample=0x7fffffffdc90, evsel=
     0x90d870, machine=0x906f80) at builtin-script.c:515
#4  0x000000000046864d in perf_session_deliver_event (session=<optimized 
out>, event=0x7fffef9d1e60, sample=0x7fffffffdc90,
     tool=0x79d020, file_offset=<optimized out>) at util/session.c:1065
#5  0x0000000000468c5a in flush_sample_queue (s=0x906f20, tool=0x79d020) 
at util/session.c:719
#6  0x0000000000468d46 in process_finished_round (tool=<optimized out>, 
event=<optimized out>, session=0x906f20)
     at util/session.c:790
#7  0x0000000000469055 in perf_session__process_user_event 
(file_offset=541768, tool=0x79d020, event=0x7fffefa2b448,
     session=0x906f20) at util/session.c:1129
#8  perf_session__process_event (session=0x906f20, event=0x7fffefa2b448, 
tool=0x79d020, file_offset=541768)
     at util/session.c:1161
#9  0x000000000046a851 in __perf_session__process_events 
(session=0x906f20, data_offset=<optimized out>,
     data_size=<optimized out>, file_size=2564784, tool=0x79d020) at 
util/session.c:1429
#10 0x000000000042ebdc in __cmd_script (session=0x906f20) at 
builtin-script.c:548

Re: [PATCH 14/17] perf header: Set tracepoint event name only if not set

[Namhyung Kim]

On Wed, 05 Sep 2012 18:01:16 -0600, David Ahern wrote:
> On 9/5/12 5:08 PM, Arnaldo Carvalho de Melo wrote:
>> From: Namhyung Kim <namhyung.kim@lge.com>
>>
>> The event name can be set already by processing a event_desc data.
>>
>> So check it before setting to prevent possible leak.
>>
>> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
>> Cc: Ingo Molnar <mingo@kernel.org>
>> Cc: Paul Mackerras <paulus@samba.org>
>> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
>> Link: http://lkml.kernel.org/r/1346821373-31621-3-git-send-email-namhyung@kernel.org
>> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
>> ---
>>   tools/perf/util/header.c |    2 +-
>>   1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
>> index a124b93..05c9310 100644
>> --- a/tools/perf/util/header.c
>> +++ b/tools/perf/util/header.c
>> @@ -2314,7 +2314,7 @@ static int perf_evlist__set_tracepoint_names(struct perf_evlist *evlist,
>>   	struct perf_evsel *pos;
>>
>>   	list_for_each_entry(pos, &evlist->entries, node) {
>> -		if (pos->attr.type == PERF_TYPE_TRACEPOINT &&
>> +		if (pos->attr.type == PERF_TYPE_TRACEPOINT && !pos->name &&
>>   		    perf_evsel__set_tracepoint_name(pos, pevent))
>>   			return -1;
>>   	}
>>
>
> I'm getting a segfault running perf-script due to this patch. Most
> likely this is exposing another problem. Back trace is:

Hmm.  It's because evsel->tp_format being not set, not sure why I didn't
detect this during tests.  Will find out a solution.

Thanks,
Namhyung


>
> #0  0x000000000049f4f2 in pevent_event_info (s=0x7fffffffdab0,
> event=0x0, record=0x7fffffffda70)
>     at tools/lib/traceevent/event-parse.c:4183
> #1  0x000000000046cd52 in event_format__print (event=0x0,
> cpu=<optimized out>, data=<optimized out>, size=<optimized out>)
>     at util/trace-event-parse.c:182
> #2  0x000000000042d41b in process_event (event=0x7fffef9d1e60,
> sample=0x7fffffffdc90, evsel=0x90d870, machine=0x906f80,
>     al=<optimized out>) at builtin-script.c:415
> #3  0x000000000042c4e1 in process_sample_event (tool=<optimized out>,
> event=0x7fffef9d1e60, sample=0x7fffffffdc90, evsel=
>     0x90d870, machine=0x906f80) at builtin-script.c:515
> #4  0x000000000046864d in perf_session_deliver_event
> (session=<optimized out>, event=0x7fffef9d1e60, sample=0x7fffffffdc90,
>     tool=0x79d020, file_offset=<optimized out>) at util/session.c:1065
> #5  0x0000000000468c5a in flush_sample_queue (s=0x906f20,
> tool=0x79d020) at util/session.c:719
> #6  0x0000000000468d46 in process_finished_round (tool=<optimized
> out>, event=<optimized out>, session=0x906f20)
>     at util/session.c:790
> #7  0x0000000000469055 in perf_session__process_user_event
> (file_offset=541768, tool=0x79d020, event=0x7fffefa2b448,
>     session=0x906f20) at util/session.c:1129
> #8  perf_session__process_event (session=0x906f20,
> event=0x7fffefa2b448, tool=0x79d020, file_offset=541768)
>     at util/session.c:1161
> #9  0x000000000046a851 in __perf_session__process_events
> (session=0x906f20, data_offset=<optimized out>,
>     data_size=<optimized out>, file_size=2564784, tool=0x79d020) at
> util/session.c:1429
> #10 0x000000000042ebdc in __cmd_script (session=0x906f20) at
> builtin-script.c:548