perf stat: Make metric event lookup more robust

After setting up metric groups through the event parser, the metricgroup
code looks them up again in the event list.

Make sure we only look up events that haven't been used by some other
metric. The data structures currently cannot handle more than one metric
per event. This avoids problems with multiple events partially
overlapping.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Link: http://lkml.kernel.org/r/20190624193711.35241-2-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This commit is contained in:
Andi Kleen 2019-06-24 12:37:08 -07:00 committed by Arnaldo Carvalho de Melo
parent 9c10548c42
commit 145c407c80

View File

@ -304,7 +304,7 @@ static struct perf_evsel *perf_stat__find_event(struct perf_evlist *evsel_list,
struct perf_evsel *c2;
evlist__for_each_entry (evsel_list, c2) {
if (!strcasecmp(c2->name, name))
if (!strcasecmp(c2->name, name) && !c2->collect_stat)
return c2;
}
return NULL;
@ -343,7 +343,8 @@ void perf_stat__collect_metric_expr(struct perf_evlist *evsel_list)
if (leader) {
/* Search in group */
for_each_group_member (oc, leader) {
if (!strcasecmp(oc->name, metric_names[i])) {
if (!strcasecmp(oc->name, metric_names[i]) &&
!oc->collect_stat) {
found = true;
break;
}