Skip to contentSkip to conversation

perf evlist: Default to syswide target when no thread/cpu maps set

tip:perf/urgent tip-bot for Arnaldo Carvalho de Melo 1 participant 1 message Oct 15, 2014 perf events

lore
Share & tools
lore

tip-bot for Arnaldo Carvalho de Melo

author merge bot
#0 lore raw
Commit-ID:  4112eb1899c0e711b2ab1491f51215359cf94d74
Gitweb:     http://git.kernel.org/tip/4112eb1899c0e711b2ab1491f51215359cf94d74
Author:     Arnaldo Carvalho de Melo <[email protected]>
AuthorDate: Fri, 10 Oct 2014 15:55:15 -0300
Committer:  Arnaldo Carvalho de Melo <[email protected]>
CommitDate: Tue, 14 Oct 2014 17:32:53 -0300

perf evlist: Default to syswide target when no thread/cpu maps set

If all a tool wants is to do system wide event monitoring, there is no more the need to setup thread_map and cpu_map objects, just call perf_evlist__open() and it will do create one fd per CPU monitoring all threads.

Cc Adrian Hunter <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc David Ahern <[email protected]>
Cc Don Zickus <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Jean Pihet <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Mike Galbraith <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Paul Mackerras <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Stephane Eranian <[email protected]>
Link http://lkml.kernel.org/n/[email protected] open here
Signed-off-by Arnaldo Carvalho de Melo <[email protected]>
---
 tools/perf/util/evlist.c | 40 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)
tools/perf/util/evlist.c +40 -0
index 5fc7bd4..b4b54d8 100644
@@ -1175,11 +1175,51 @@ void perf_evlist__close(struct perf_evlist *evlist)
 	}
 }
 
+static int perf_evlist__create_syswide_maps(struct perf_evlist *evlist)
+{
+	int err = -ENOMEM;
+
+	/*
+	 * Try reading /sys/devices/system/cpu/online to get
+	 * an all cpus map.
+	 *
+	 * FIXME: -ENOMEM is the best we can do here, the cpu_map
+	 * code needs an overhaul to properly forward the
+	 * error, and we may not want to do that fallback to a
+	 * default cpu identity map :-\
+	 */
+	evlist->cpus = cpu_map__new(NULL);
+	if (evlist->cpus == NULL)
+		goto out;
+
+	evlist->threads = thread_map__new_dummy();
+	if (evlist->threads == NULL)
+		goto out_free_cpus;
+
+	err = 0;
+out:
+	return err;
+out_free_cpus:
+	cpu_map__delete(evlist->cpus);
+	evlist->cpus = NULL;
+	goto out;
+}
+
 int perf_evlist__open(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
 	int err;
 
+	/*
+	 * Default: one fd per CPU, all threads, aka systemwide
+	 * as sys_perf_event_open(cpu = -1, thread = -1) is EINVAL
+	 */
+	if (evlist->threads == NULL && evlist->cpus == NULL) {
+		err = perf_evlist__create_syswide_maps(evlist);
+		if (err < 0)
+			goto out_err;
+	}
+
 	perf_evlist__update_id_pos(evlist);
 
 	evlist__for_each(evlist, evsel) {