[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Yexl4R5qsyi9sEQf@kernel.org>
Date: Sat, 22 Jan 2022 17:15:29 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: German Gomez <german.gomez@....com>
Cc: linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org,
Chase Conklin <chase.conklin@....com>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
Martin KaFai Lau <kafai@...com>,
Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>, Ian Rogers <irogers@...gle.com>,
Arnaldo Carvalho de Melo <acme@...hat.com>,
Stephane Eranian <eranian@...gle.com>, netdev@...r.kernel.org,
bpf@...r.kernel.org
Subject: Re: [PATCH v2] perf record/arm-spe: Override attr->sample_period for
non-libpfm4 events
Em Tue, Jan 18, 2022 at 02:40:54PM +0000, German Gomez escreveu:
> A previous patch preventing "attr->sample_period" values from being
> overridden in pfm events changed a related behaviour in arm-spe.
>
> Before said patch:
> perf record -c 10000 -e arm_spe_0// -- sleep 1
>
> Would yield an SPE event with period=10000. After the patch, the period
> in "-c 10000" was being ignored because the arm-spe code initializes
> sample_period to a non-zero value.
>
> This patch restores the previous behaviour for non-libpfm4 events.
>
> Reported-by: Chase Conklin <chase.conklin@....com>
> Fixes: ae5dcc8abe31 (“perf record: Prevent override of attr->sample_period for libpfm4 events”)
> Signed-off-by: German Gomez <german.gomez@....com>
> ---
> As suggested by Arnaldo, this v2 doesn't include a test in order to not
> block this fix for longer than necessary. So the test can be sent as a
> separate submission later.
Got it, will apply.
> Changes since v1.
> - Update commit message (James Clark)
> ---
> tools/perf/util/evsel.c | 25 +++++++++++++++++--------
> 1 file changed, 17 insertions(+), 8 deletions(-)
>
> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> index a59fb2ecb84e..86ab038f020f 100644
> --- a/tools/perf/util/evsel.c
> +++ b/tools/perf/util/evsel.c
> @@ -1065,6 +1065,17 @@ void __weak arch_evsel__fixup_new_cycles(struct perf_event_attr *attr __maybe_un
> {
> }
>
> +static void evsel__set_default_freq_period(struct record_opts *opts,
> + struct perf_event_attr *attr)
> +{
> + if (opts->freq) {
> + attr->freq = 1;
> + attr->sample_freq = opts->freq;
> + } else {
> + attr->sample_period = opts->default_interval;
> + }
> +}
> +
> /*
> * The enable_on_exec/disabled value strategy:
> *
> @@ -1131,14 +1142,12 @@ void evsel__config(struct evsel *evsel, struct record_opts *opts,
> * We default some events to have a default interval. But keep
> * it a weak assumption overridable by the user.
> */
> - if (!attr->sample_period) {
> - if (opts->freq) {
> - attr->freq = 1;
> - attr->sample_freq = opts->freq;
> - } else {
> - attr->sample_period = opts->default_interval;
> - }
> - }
> + if ((evsel->is_libpfm_event && !attr->sample_period) ||
> + (!evsel->is_libpfm_event && (!attr->sample_period ||
> + opts->user_freq != UINT_MAX ||
> + opts->user_interval != ULLONG_MAX)))
> + evsel__set_default_freq_period(opts, attr);
> +
> /*
> * If attr->freq was set (here or earlier), ask for period
> * to be sampled.
> --
> 2.25.1
--
- Arnaldo
Powered by blists - more mailing lists