Received-SPF: pass (google.com: domain of linux-kernel+bounces-214962-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Precedence: bulk
MIME-Version: 1.0
References: <20240614073517.94974-1-sieberf@amazon.com>
In-Reply-To: <20240614073517.94974-1-sieberf@amazon.com>
From: Namhyung Kim <namhyung@kernel.org>
Date: Fri, 14 Jun 2024 06:47:42 -0700
Message-ID: <CAM9d7ciYQ9C0_zgwWYt505+Mv56EC7byqxBXkPgGydomrzTHcw@mail.gmail.com>
Subject: Re: [PATCH] perf: sched map skips redundant lines with cpu filters
To: Fernand Sieber <sieberf@amazon.com>
Cc: linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org, 
	Peter Zijlstra <peterz@infradead.org>, Ingo Molnar <mingo@redhat.com>, 
	Arnaldo Carvalho de Melo <acme@kernel.org>, Alexander Shishkin <alexander.shishkin@linux.intel.com>, 
	Madadi Vineeth Reddy <vineethr@linux.ibm.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hello,

On Fri, Jun 14, 2024 at 12:49=E2=80=AFAM Fernand Sieber <sieberf@amazon.com=
> wrote:
>
> perf sched map supports cpu filter.
> However, even with cpu filters active, any context switch currently
> corresponds to a separate line.
> As result, context switches on irrelevant cpus result to redundant lines,
> which makes the output particlularly difficult to read on wide
> architectures.
>
> Fix it by skipping printing for irrelevant CPUs.
>
> Example snippet of output before fix:
>
>   *B0       1.461147 secs
>    B0
>    B0
>    B0
>   *G0       1.517139 secs
>
> After fix:
>
>   *B0       1.461147 secs
>   *G0       1.517139 secs
>
> Signed-off-by: Fernand Sieber <sieberf@amazon.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>

Acked-by: Namhyung Kim <namhyung@kernel.org>

Thanks,
Namhyung

> ---
>  tools/perf/builtin-sched.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
> index 7422c930abaf..aa59f763ca46 100644
> --- a/tools/perf/builtin-sched.c
> +++ b/tools/perf/builtin-sched.c
> @@ -1594,8 +1594,6 @@ static int map_switch_event(struct perf_sched *sche=
d, struct evsel *evsel,
>
>         sched->curr_thread[this_cpu.cpu] =3D thread__get(sched_in);
>
> -       printf("  ");
> -
>         new_shortname =3D 0;
>         if (!tr->shortname[0]) {
>                 if (!strcmp(thread__comm_str(sched_in), "swapper")) {
> @@ -1622,6 +1620,11 @@ static int map_switch_event(struct perf_sched *sch=
ed, struct evsel *evsel,
>                 new_shortname =3D 1;
>         }
>
> +       if (sched->map.cpus && !perf_cpu_map__has(sched->map.cpus, this_c=
pu))
> +               goto out;
> +
> +       printf("  ");
> +
>         for (i =3D 0; i < cpus_nr; i++) {
>                 struct perf_cpu cpu =3D {
>                         .cpu =3D sched->map.comp ? sched->map.comp_cpus[i=
].cpu : i,
> @@ -1656,9 +1659,6 @@ static int map_switch_event(struct perf_sched *sche=
d, struct evsel *evsel,
>                         color_fprintf(stdout, color, "   ");
>         }
>
> -       if (sched->map.cpus && !perf_cpu_map__has(sched->map.cpus, this_c=
pu))
> -               goto out;
> -
>         timestamp__scnprintf_usec(timestamp, stimestamp, sizeof(stimestam=
p));
>         color_fprintf(stdout, color, "  %12s secs ", stimestamp);
>         if (new_shortname || tr->comm_changed || (verbose > 0 && thread__=
tid(sched_in))) {
> @@ -1675,9 +1675,9 @@ static int map_switch_event(struct perf_sched *sche=
d, struct evsel *evsel,
>         if (sched->map.comp && new_cpu)
>                 color_fprintf(stdout, color, " (CPU %d)", this_cpu);
>
> -out:
>         color_fprintf(stdout, color, "\n");
>
> +out:
>         thread__put(sched_in);
>
>         return 0;
> --
> 2.40.1
>
>