[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAP-5=fXTiHY6-Tb5kLGxQq7MGmW_qQpM9ZYV3Oaz=hyfw0sxYg@mail.gmail.com>
Date: Wed, 6 Nov 2024 09:28:58 -0800
From: Ian Rogers <irogers@...gle.com>
To: Namhyung Kim <namhyung@...nel.org>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>, Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>,
Adrian Hunter <adrian.hunter@...el.com>, Kan Liang <kan.liang@...ux.intel.com>,
Thomas Richter <tmricht@...ux.ibm.com>, James Clark <james.clark@...aro.org>,
linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/2] perf test: Skip annotate test for sanitizer builds
On Wed, Oct 23, 2024 at 4:31 PM Namhyung Kim <namhyung@...nel.org> wrote:
>
> On Tue, Oct 22, 2024 at 10:48:38AM -0700, Ian Rogers wrote:
> > Symbols vary and the test breaks.
> >
> > Closes: https://lore.kernel.org/lkml/CAP-5=fU04PAN4T=7KuHA4h+po=oTy+6Nbee-Gvx9hCsEf2Lh0w@mail.gmail.com/
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > ---
> > tools/perf/tests/shell/annotate.sh | 6 ++++++
> > 1 file changed, 6 insertions(+)
> >
> > diff --git a/tools/perf/tests/shell/annotate.sh b/tools/perf/tests/shell/annotate.sh
> > index 1590a37363de..199f547e656d 100755
> > --- a/tools/perf/tests/shell/annotate.sh
> > +++ b/tools/perf/tests/shell/annotate.sh
> > @@ -4,6 +4,12 @@
> >
> > set -e
> >
> > +if perf check feature -q sanitizer
> > +then
> > + echo "Skip test with sanitizers due to differing assembly code"
>
> I don't think it's because of different assembly code.
> It should be the return value from ASAN leak detector.
>
> Maybe we can enable it using "ASAN_OPTIONS=detect_leaks=0"?
> Probably with a comment that mentions we don't call
> perf_session__delete() in perf annotate for a performance reason.
So doing things like not deleting things for performance reasons is
okay but should be guarded by an "#ifndef NDEBUG" or worse case
"#ifndef LEAK_SANITIZER", not always be on. The comment about a
performance optimization belongs there. We should be trying to avoid
disabling sanitizers in tests as otherwise we're missing the benefits
of sanitizers. I'll see if doing this can to avoid the test skipping.
It'd be nice if other people tested with sanitizers. I think we still
want the features for if we get other issues in the future.
Thanks,
Ian
Powered by blists - more mailing lists