소스 검색

perf script: Allow creating per-event dump files

Introduce a new option to dump trace output to files named by the
monitored events and update perf-script documentation accordingly.

Shown below is output of perf script command with the newly introduced
option.

         $ perf record -e cycles -e cs -ag -- sleep 1
         $ perf script --per-event-dump
         $ ls
         perf.data.cycles.dump perf.data.cs.dump

Without per-event-dump support, drawing flamegraphs for different events
would require post processing to separate events. You can monitor only
one event at a time if you want to get flamegraphs for different events.
Using this option, you can get the trace output files named by the
monitored events, and could draw flamegraphs according to the event's
name.

Based-on-a-patch-by: yuzhoujian <yuzhoujian@didichuxing.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1508921599-10832-3-git-send-email-yuzhoujian@didichuxing.com
Link: http://lkml.kernel.org/n/tip-8ngzsjdhgiovkupl3r5yy570@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Arnaldo Carvalho de Melo 7 년 전
부모
커밋
a14390fde6
2개의 변경된 파일62개의 추가작업 그리고 1개의 파일을 삭제
  1. 4 0
      tools/perf/Documentation/perf-script.txt
  2. 58 1
      tools/perf/builtin-script.c

+ 4 - 0
tools/perf/Documentation/perf-script.txt

@@ -325,6 +325,10 @@ include::itrace.txt[]
 	Set the maximum number of program blocks to print with brstackasm for
 	Set the maximum number of program blocks to print with brstackasm for
 	each sample.
 	each sample.
 
 
+--per-event-dump::
+	Create per event files with a "perf.data.EVENT.dump" name instead of
+        printing to stdout, useful, for instance, for generating flamegraphs.
+
 --inline::
 --inline::
 	If a callgraph address belongs to an inlined function, the inline stack
 	If a callgraph address belongs to an inlined function, the inline stack
 	will be printed. Each entry has function name and file/line. Enabled by
 	will be printed. Each entry has function name and file/line. Enabled by

+ 58 - 1
tools/perf/builtin-script.c

@@ -1392,6 +1392,7 @@ struct perf_script {
 	bool			show_switch_events;
 	bool			show_switch_events;
 	bool			show_namespace_events;
 	bool			show_namespace_events;
 	bool			allocated;
 	bool			allocated;
+	bool			per_event_dump;
 	struct cpu_map		*cpus;
 	struct cpu_map		*cpus;
 	struct thread_map	*threads;
 	struct thread_map	*threads;
 	int			name_width;
 	int			name_width;
@@ -1438,7 +1439,7 @@ static void process_event(struct perf_script *script,
 	struct thread *thread = al->thread;
 	struct thread *thread = al->thread;
 	struct perf_event_attr *attr = &evsel->attr;
 	struct perf_event_attr *attr = &evsel->attr;
 	unsigned int type = output_type(attr->type);
 	unsigned int type = output_type(attr->type);
-	FILE *fp = stdout;
+	FILE *fp = evsel->priv;
 
 
 	if (output[type].fields == 0)
 	if (output[type].fields == 0)
 		return;
 		return;
@@ -1887,6 +1888,52 @@ static void sig_handler(int sig __maybe_unused)
 	session_done = 1;
 	session_done = 1;
 }
 }
 
 
+static void perf_script__fclose_per_event_dump(struct perf_script *script)
+{
+	struct perf_evlist *evlist = script->session->evlist;
+	struct perf_evsel *evsel;
+
+	evlist__for_each_entry(evlist, evsel) {
+		if (!evsel->priv)
+			break;
+		fclose(evsel->priv);
+		evsel->priv = NULL;
+	}
+}
+
+static int perf_script__fopen_per_event_dump(struct perf_script *script)
+{
+	struct perf_evsel *evsel;
+
+	evlist__for_each_entry(script->session->evlist, evsel) {
+		char filename[PATH_MAX];
+		snprintf(filename, sizeof(filename), "%s.%s.dump",
+			 script->session->file->path, perf_evsel__name(evsel));
+		evsel->priv = fopen(filename, "w");
+		if (evsel->priv == NULL)
+			goto out_err_fclose;
+	}
+
+	return 0;
+
+out_err_fclose:
+	perf_script__fclose_per_event_dump(script);
+	return -1;
+}
+
+static int perf_script__setup_per_event_dump(struct perf_script *script)
+{
+	struct perf_evsel *evsel;
+
+	if (script->per_event_dump)
+		return perf_script__fopen_per_event_dump(script);
+
+	evlist__for_each_entry(script->session->evlist, evsel)
+		evsel->priv = stdout;
+
+	return 0;
+}
+
 static int __cmd_script(struct perf_script *script)
 static int __cmd_script(struct perf_script *script)
 {
 {
 	int ret;
 	int ret;
@@ -1908,8 +1955,16 @@ static int __cmd_script(struct perf_script *script)
 	if (script->show_namespace_events)
 	if (script->show_namespace_events)
 		script->tool.namespaces = process_namespaces_event;
 		script->tool.namespaces = process_namespaces_event;
 
 
+	if (perf_script__setup_per_event_dump(script)) {
+		pr_err("Couldn't create the per event dump files\n");
+		return -1;
+	}
+
 	ret = perf_session__process_events(script->session);
 	ret = perf_session__process_events(script->session);
 
 
+	if (script->per_event_dump)
+		perf_script__fclose_per_event_dump(script);
+
 	if (debug_mode)
 	if (debug_mode)
 		pr_err("Misordered timestamps: %" PRIu64 "\n", nr_unordered);
 		pr_err("Misordered timestamps: %" PRIu64 "\n", nr_unordered);
 
 
@@ -2827,6 +2882,8 @@ int cmd_script(int argc, const char **argv)
 		    "Show context switch events (if recorded)"),
 		    "Show context switch events (if recorded)"),
 	OPT_BOOLEAN('\0', "show-namespace-events", &script.show_namespace_events,
 	OPT_BOOLEAN('\0', "show-namespace-events", &script.show_namespace_events,
 		    "Show namespace events (if recorded)"),
 		    "Show namespace events (if recorded)"),
+	OPT_BOOLEAN('\0', "per-event-dump", &script.per_event_dump,
+		    "Dump trace output to files named by the monitored events"),
 	OPT_BOOLEAN('f', "force", &symbol_conf.force, "don't complain, do it"),
 	OPT_BOOLEAN('f', "force", &symbol_conf.force, "don't complain, do it"),
 	OPT_INTEGER(0, "max-blocks", &max_blocks,
 	OPT_INTEGER(0, "max-blocks", &max_blocks,
 		    "Maximum number of code blocks to dump with brstackinsn"),
 		    "Maximum number of code blocks to dump with brstackinsn"),