perf tools: Introduce perf_evlist__start_workload_ex()

PATCH In review Namhyung Kim 2 participants 4 messages May 20, 2015 perf events

Share & tools

Namhyung Kim

author maintainer

2015-05-20 02:58 UTC #0 lore raw

PATCH 1/3 perf tools: Introduce perf_evlist__start_workload_ex()

The perf_evlist__start_workload_ex() does same as __start_work() but also invokes callback which does additional work for each command.

Signed-off-by Namhyung Kim <[email protected]>

---
 tools/perf/util/evlist.c | 11 +++++++++++
 tools/perf/util/evlist.h |  4 ++++
 2 files changed, 15 insertions(+)

tools/perf/util/evlist.c +11 -0

index dc1dc2c181ef..bfe455ec673b 100644
@@ -1505,6 +1505,17 @@ int perf_evlist__start_workload(struct perf_evlist *evlist)
 	return 0;
 }
 
+int perf_evlist__start_workload_ex(struct perf_evlist *evlist,
+				   workload_callback_t callback, void *arg)
+{
+	int ret = callback(evlist, arg);
+
+	if (ret == 0)
+		ret = perf_evlist__start_workload(evlist);
+
+	return ret;
+}
+
 int perf_evlist__parse_sample(struct perf_evlist *evlist, union perf_event *event,
 			      struct perf_sample *sample)
 {

tools/perf/util/evlist.h +4 -0

index 955bf31b7dd3..6212f036c134 100644
@@ -123,6 +123,10 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist,
 						     void *ucontext));
 int perf_evlist__start_workload(struct perf_evlist *evlist);
 
+typedef int (*workload_callback_t)(struct perf_evlist *evlist, void *arg);
+int perf_evlist__start_workload_ex(struct perf_evlist *evlist,
+				   workload_callback_t callback, void *arg);
+
 struct option;
 
 int __perf_evlist__parse_mmap_pages(unsigned int *mmap_pages, const char *str);

Signature · 1 line

2.4.0

Namhyung Kim

author maintainer

2015-05-20 02:58 UTC #1 lore raw

PATCH 2/3 perf trace: Create struct thread for command line workload

When perf creates a new child to profile, the events are enabled on exec(). And in this case, it doesn't synthesize any event for the child since they'll be generated during exec(). But there's an window between the enabling and the event generation. This leads to some early event not having a comm but pid (like ':15328').

Fix it by using perf_evlist__start_workload_ex() to create corresponding thread at the same time.

Signed-off-by Namhyung Kim <[email protected]>

---
 tools/perf/builtin-trace.c |  4 +++-
 tools/perf/util/evlist.c   | 14 ++++++++++++++
 tools/perf/util/evlist.h   |  2 ++
 3 files changed, 19 insertions(+), 1 deletion(-)

tools/perf/builtin-trace.c +3 -1

index a05490d06374..e30d0ed1e60f 100644
@@ -2282,7 +2282,9 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 		perf_evlist__enable(evlist);
 
 	if (forks)
-		perf_evlist__start_workload(evlist);
+		perf_evlist__start_workload_ex(evlist,
+					       perf_evlist__create_workload_thread,
+					       trace->host);
 
 	trace->multiple_threads = evlist->threads->map[0] == -1 ||
 				  evlist->threads->nr > 1 ||

tools/perf/util/evlist.c +14 -0

index bfe455ec673b..f84305915e47 100644
@@ -1516,6 +1516,20 @@ int perf_evlist__start_workload_ex(struct perf_evlist *evlist,
 	return ret;
 }
 
+int perf_evlist__create_workload_thread(struct perf_evlist *evlist, void *arg)
+{
+	struct machine *machine = arg;
+	int pid = evlist->workload.pid;
+	struct thread *thread;
+
+	thread = machine__findnew_thread(machine, pid, pid);
+	if (thread == NULL)
+		return -1;
+
+	thread__set_comm(thread, program_invocation_short_name, 0);
+	return 0;
+}
+
 int perf_evlist__parse_sample(struct perf_evlist *evlist, union perf_event *event,
 			      struct perf_sample *sample)
 {

tools/perf/util/evlist.h +2 -0

index 6212f036c134..a02b0f4ba4f9 100644
@@ -126,6 +126,8 @@ int perf_evlist__start_workload(struct perf_evlist *evlist);
 typedef int (*workload_callback_t)(struct perf_evlist *evlist, void *arg);
 int perf_evlist__start_workload_ex(struct perf_evlist *evlist,
 				   workload_callback_t callback, void *arg);
+/* workload_callbacks */
+int perf_evlist__create_workload_thread(struct perf_evlist *evlist, void *arg);
 
 struct option;

Signature · 1 line

2.4.0

Namhyung Kim

author maintainer

2015-05-20 02:58 UTC #2 lore raw

PATCH 3/3 perf record: Synthesize COMM event for a command line workload

In reply to Namhyung Kim

It used to be overcome since samples are only in kernel (so we always have the map) and the comm is overridden by a later COMM event. However it won't work if events are processed and displayed before the COMM event overrides like in 'perf script'. This leads to those early samples (like native_write_msr_safe) not having a comm but pid (like ':15328').

So it needs to synthesize COMM event for the child explicitly before enabling so that it can have a correct comm. But at this time, the comm will be "perf" since it's not exec-ed yet.

Signed-off-by Namhyung Kim <[email protected]>

---
 tools/perf/builtin-record.c | 41 ++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 40 insertions(+), 1 deletion(-)

tools/perf/builtin-record.c +40 -1

index 5dfe91395617..a0821ee92e27 100644
@@ -469,6 +469,43 @@ static void workload_exec_failed_signal(int signo __maybe_unused,
 	child_finished = 1;
 }
 
+static int synthesize_workload_comm_event(struct perf_evlist *evlist, void *arg)
+{
+	union perf_event *event;
+	struct record *rec = arg;
+	struct machine *machine = &rec->session->machines.host;
+	int pid = evlist->workload.pid;
+	const char *comm_str = program_invocation_short_name;
+	size_t comm_size, total_size;
+	int ret;
+
+	comm_size = PERF_ALIGN(strlen(comm_str) + 1, sizeof(u64));
+	total_size = sizeof(event->comm) + machine->id_hdr_size;
+	/*
+	 * (aligned) comm size might be smaller than expected size
+	 * (i.e.  size of event->comm.comm[]), in that case it needs
+	 * to shrink the total size.
+	 */
+	if (comm_size < sizeof(event->comm.comm))
+		total_size -= sizeof(event->comm.comm) - comm_size;
+
+	event = zalloc(total_size);
+	if (event == NULL)
+		return -ENOMEM;
+
+	event->comm.header.type = PERF_RECORD_COMM;
+	event->comm.header.size = total_size;
+
+	event->comm.pid = pid;
+	event->comm.tid = pid;
+	strncpy(event->comm.comm, comm_str, comm_size);
+
+	ret = record__write(rec, event, total_size);
+
+	free(event);
+	return ret;
+}
+
 static void snapshot_sig_handler(int sig);
 
 static int __cmd_record(struct record *rec, int argc, const char **argv)

@@ -627,7 +664,9 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	 * Let the child rip
 	 */
 	if (forks)
-		perf_evlist__start_workload(rec->evlist);
+		perf_evlist__start_workload_ex(rec->evlist,
+					       synthesize_workload_comm_event,
+					       rec);
 
 	if (opts->initial_delay) {
 		usleep(opts->initial_delay * 1000);

Signature · 1 line

2.4.0

Jiri Olsa

Acked-by on v1 3/3

2015-05-20 07:08 UTC #3 lore raw

Re: [PATCH 3/3] perf record: Synthesize COMM event for a command line workload

Quoted 26 lines · Namhyung Kim

When perf creates a new child to profile, the events are enabled on exec(). And in this case, it doesn't synthesize any event for the child since they'll be generated during exec(). But there's an window between the enabling and the event generation.
It used to be overcome since samples are only in kernel (so we always have the map) and the comm is overridden by a later COMM event. However it won't work if events are processed and displayed before the COMM event overrides like in 'perf script'. This leads to those early samples (like native_write_msr_safe) not having a comm but pid (like ':15328').
So it needs to synthesize COMM event for the child explicitly before enabling so that it can have a correct comm. But at this time, the comm will be "perf" since it's not exec-ed yet.
Signed-off-by Namhyung Kim <[email protected]>
---
 tools/perf/builtin-record.c | 41 ++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 40 insertions(+), 1 deletion(-)
tools/perf/builtin-record.c +0 -0
index 5dfe91395617..a0821ee92e27 100644

tools/perf/builtin-record.c static void workload_exec_failed_signal(int signo __maybe_unused, · line 478

Quoted 10 lines

tools/perf/builtin-record.c +7 -0

@@ -469,6 +469,43 @@ static void workload_exec_failed_signal(int signo __maybe_unused,
 	child_finished = 1;
 }
 
+static int synthesize_workload_comm_event(struct perf_evlist *evlist, void *arg)
+{
+	union perf_event *event;
+	struct record *rec = arg;
+	struct machine *machine = &rec->session->machines.host;
+	int pid = evlist->workload.pid;
+	const char *comm_str = program_invocation_short_name;

never heard of program_invocation_short_name ;-) nice..

for the patchset:

Acked-by Jiri Olsa <[email protected]>

I tested the record/script use case.. works ok

thanks,
jirka