From: Namhyung Kim <namhyung@kernel.org>
To: Ian Rogers <irogers@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
Jiri Olsa <jolsa@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>,
James Clark <james.clark@linaro.org>,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v1] perf test: Stat std output match 2 metrics for stalled-cycles-backend
Date: Mon, 6 Apr 2026 22:24:27 -0700 [thread overview]
Message-ID: <adSVC2UqC6IO3Z3Z@z2> (raw)
In-Reply-To: <CAP-5=fU50qG6BB+eG=EkE55Hn+rFPBPeHe9P=0hYdACQ=yviUg@mail.gmail.com>
On Mon, Apr 06, 2026 at 11:25:49AM -0700, Ian Rogers wrote:
> On Mon, Apr 6, 2026 at 11:16 AM Namhyung Kim <namhyung@kernel.org> wrote:
> >
> > On Thu, Apr 02, 2026 at 03:04:06PM -0700, Ian Rogers wrote:
> > > Seen on ARM, the stalled-cycles-backend event can be given as the
> > > event for both backend_cycles_idle and stalled_cycles_per_instruction
> > > metrics. Make the metric match sufficiently generic to allow this.
> > >
> > > Signed-off-by: Ian Rogers <irogers@google.com>
> > > ---
> > > tools/perf/tests/shell/stat+std_output.sh | 7 +++++--
> > > 1 file changed, 5 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/tools/perf/tests/shell/stat+std_output.sh b/tools/perf/tests/shell/stat+std_output.sh
> > > index 9c4b92ecf448..5a6bda436dd4 100755
> > > --- a/tools/perf/tests/shell/stat+std_output.sh
> > > +++ b/tools/perf/tests/shell/stat+std_output.sh
> > > @@ -11,9 +11,12 @@ set -e
> > >
> > > stat_output=$(mktemp /tmp/__perf_test.stat_output.std.XXXXX)
> > >
> > > +# For event_name[i] expect to see event_metric[i]. 'stalled-cycles-backend' may
> > > +# match both 'backend_cycles_idle' and 'stalled_cycles_per_instruction', hence
> > > +# matching "_cycles_".
> > > event_name=(cpu-clock task-clock context-switches cpu-migrations page-faults stalled-cycles-frontend stalled-cycles-backend cycles instructions branches branch-misses)
> > > -event_metric=("CPUs_utilized" "CPUs_utilized" "cs/sec" "migrations/sec" "faults/sec" "frontend_cycles_idle" "backend_cycles_idle" "GHz" "insn_per_cycle" "/sec" "branch_miss_rate")
> > > -skip_metric=("tma_" "TopdownL1")
> > > +event_metric=("CPUs_utilized" "CPUs_utilized" "cs/sec" "migrations/sec" "faults/sec" "frontend_cycles_idle" "_cycles_" "GHz" "insn_per_cycle" "/sec" "branch_miss_rate")
> >
> > Is this used for partial string match?
> >
> > Then I guess "frontend_cycles_idle" can be covered by "_cycles_" too?
> > Also as we have "/sec", can "cs/sec" and "migrations/sec" be removed?
>
> When we see the event at index 'i' in the event_name array, we check
> the corresponding metric at index 'i' in the event_metric array. So
> "cs/sec" matches the event "context-switches" while "/sec" matches
> "branches". We could weaken the "context-switches" event to match just
> "/sec" as a metric, but it seems better to leave it as "cs/sec".
I see. So it's index-based and the metric names are different on x86
and ARM.
>
> > > +skip_metric=("tma_" "TopdownL1" "percent of slots")
> >
> > Why is "percent of slots" added as well?
>
> Seen on ARM, the metric names lack an easily catchable prefix like
> tma_. Its part of the fixing this test for ARM, but could be called
> out more in the commit message.
Right, it'd be great if you could explain the problem clearly with an
example. Feel free to consider me knows nothing..
Thanks,
Namhyung
prev parent reply other threads:[~2026-04-07 5:24 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-02 22:04 [PATCH v1] perf test: Stat std output match 2 metrics for stalled-cycles-backend Ian Rogers
2026-04-06 18:16 ` Namhyung Kim
2026-04-06 18:25 ` Ian Rogers
2026-04-07 5:24 ` Namhyung Kim [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=adSVC2UqC6IO3Z3Z@z2 \
--to=namhyung@kernel.org \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=irogers@google.com \
--cc=james.clark@linaro.org \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox