[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aQlzWvieU5mBktx4@google.com>
Date: Mon, 3 Nov 2025 19:30:34 -0800
From: Namhyung Kim <namhyung@...nel.org>
To: Ian Rogers <irogers@...gle.com>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
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>,
James Clark <james.clark@...aro.org>, linux-kernel@...r.kernel.org,
linux-perf-users@...r.kernel.org
Subject: Re: [PATCH v1 1/2] perf pmu: Make pmu_alias_terms weak again
On Thu, Oct 30, 2025 at 08:08:02AM -0700, Ian Rogers wrote:
> On Wed, Oct 29, 2025 at 10:47 PM Namhyung Kim <namhyung@...nel.org> wrote:
> >
> > Hi Ian,
> >
> > On Wed, Oct 29, 2025 at 03:26:37PM -0700, Ian Rogers wrote:
> > > The terms for a json event should be weak so they don't override
> > > command line options.
> >
> > Can you please give an example command line and the error?
>
> Sure, it is also covered in the test.
Yep, but lazy maintainers want to see it in the commit message. :)
>
> Before:
> ```
> $ perf record -vv -c 1000 -e uops_issued.any -o /dev/null true 2>&1
> |grep "{ sample_period, sample_freq }"
> { sample_period, sample_freq } 200003
> { sample_period, sample_freq } 2000003
> { sample_period, sample_freq } 1000
> ```
>
> After:
> ```
> $ perf record -vv -c 1000 -e uops_issued.any -o /dev/null true 2>&1
> |grep "{ sample_period, sample_freq }"
> { sample_period, sample_freq } 1000
> { sample_period, sample_freq } 1000
> { sample_period, sample_freq } 1000
> ```
Thanks, it ignored command line options. I can see the default period
is in the JSON description.
$ perf list -j | grep -C 5 uops_issued.any
"Encoding": "cpu/event=0xb1,period=2000003,umask=0x10/"
},
{
"Unit": "cpu",
"Topic": "pipeline",
"EventName": "uops_issued.any",
"EventType": "Kernel PMU event",
"BriefDescription": "Uops that RAT issues to RS",
"PublicDescription": "Counts the number of uops that the Resource Allocation Table (RAT) issues to the Reservation Station (RS)",
"Encoding": "cpu/event=0xe,period=2000003,umask=1/"
},
It'd be great if you could include this kind of info in the commit log
so that others can understand the problem clearly and test the patch
easily.
Thanks,
Namhyung
> > >
> > > Fixes: 84bae3af20d0 ("perf pmu: Don't eagerly parse event terms")
> > > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > > ---
> > > tools/perf/util/pmu.c | 8 ++++++++
> > > 1 file changed, 8 insertions(+)
> > >
> > > diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> > > index d597263fab4f..f14f2a12d061 100644
> > > --- a/tools/perf/util/pmu.c
> > > +++ b/tools/perf/util/pmu.c
> > > @@ -802,6 +802,7 @@ static int pmu_aliases_parse_eager(struct perf_pmu *pmu, int sysfs_fd)
> > > static int pmu_alias_terms(struct perf_pmu_alias *alias, struct list_head *terms)
> > > {
> > > struct parse_events_terms alias_terms;
> > > + struct parse_events_term *term;
> > > int ret;
> > >
> > > parse_events_terms__init(&alias_terms);
> > > @@ -812,6 +813,13 @@ static int pmu_alias_terms(struct perf_pmu_alias *alias, struct list_head *terms
> > > parse_events_terms__exit(&alias_terms);
> > > return ret;
> > > }
> > > + list_for_each_entry(term, &alias_terms.terms, list) {
> > > + /*
> > > + * Weak terms don't override command line options,
> > > + * which we don't want for implicit terms in aliases.
> > > + */
> > > + term->weak = true;
> > > + }
> > > list_splice_init(&alias_terms.terms, terms);
> > > parse_events_terms__exit(&alias_terms);
> > > return 0;
> > > --
> > > 2.51.1.851.g4ebd6896fd-goog
> > >
Powered by blists - more mailing lists