summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorNamhyung Kim <namhyung@kernel.org>2023-02-22 23:01:55 -0800
committerArnaldo Carvalho de Melo <acme@redhat.com>2023-02-23 10:03:00 -0300
commitce9f1c05d2edfa6cdf2c1a510495d333e11810a8 (patch)
tree95cec947c09aa9bb2fa98aa34c49869354336a4c
parent3ef9fec011d471a259935f2e2ff58ab12886d683 (diff)
downloadlwn-ce9f1c05d2edfa6cdf2c1a510495d333e11810a8.tar.gz
lwn-ce9f1c05d2edfa6cdf2c1a510495d333e11810a8.zip
perf inject: Fix --buildid-all not to eat up MMAP2
When MMAP2 has the PERF_RECORD_MISC_MMAP_BUILD_ID flag, it means the record already has the build-id info. So it marks the DSO as hit, to skip if the same DSO is not processed if it happens to miss the build-id later. But it missed to copy the MMAP2 record itself so it'd fail to symbolize samples for those regions. For example, the following generates 249 MMAP2 events. $ perf record --buildid-mmap -o- true | perf report --stat -i- | grep MMAP2 MMAP2 events: 249 (86.8%) Adding perf inject should not change the number of events like this $ perf record --buildid-mmap -o- true | perf inject -b | \ > perf report --stat -i- | grep MMAP2 MMAP2 events: 249 (86.5%) But when --buildid-all is used, it eats most of the MMAP2 events. $ perf record --buildid-mmap -o- true | perf inject -b --buildid-all | \ > perf report --stat -i- | grep MMAP2 MMAP2 events: 1 ( 2.5%) With this patch, it shows the original number now. $ perf record --buildid-mmap -o- true | perf inject -b --buildid-all | \ > perf report --stat -i- | grep MMAP2 MMAP2 events: 249 (86.5%) Committer testing: Before: $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf inject -b | perf report --stat -i- | grep MMAP2 MMAP2 events: 58 (36.2%) $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf report --stat -i- | grep MMAP2 MMAP2 events: 58 (36.2%) $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf inject -b --buildid-all | perf report --stat -i- | grep MMAP2 MMAP2 events: 2 ( 1.9%) $ After: $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf inject -b | perf report --stat -i- | grep MMAP2 MMAP2 events: 58 (29.3%) $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf report --stat -i- | grep MMAP2 MMAP2 events: 58 (34.3%) $ perf record --buildid-mmap -o- perf stat --null sleep 1 2> /dev/null | perf inject -b --buildid-all | perf report --stat -i- | grep MMAP2 MMAP2 events: 58 (38.4%) $ Fixes: f7fc0d1c915a74ff ("perf inject: Do not inject BUILD_ID record if MMAP2 has it") Signed-off-by: Namhyung Kim <namhyung@kernel.org> Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20230223070155.54251-1-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
-rw-r--r--tools/perf/builtin-inject.c1
1 files changed, 1 insertions, 0 deletions
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index f8182417b734..10bb1d494258 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -538,6 +538,7 @@ static int perf_event__repipe_buildid_mmap2(struct perf_tool *tool,
dso->hit = 1;
}
dso__put(dso);
+ perf_event__repipe(tool, event, sample, machine);
return 0;
}