Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
MIME-Version: 1.0
References: <20231207125716.1947860-1-tmricht@linux.ibm.com>
In-Reply-To: <20231207125716.1947860-1-tmricht@linux.ibm.com>
From:   Namhyung Kim <namhyung@kernel.org>
Date:   Thu, 7 Dec 2023 15:26:43 -0800
Message-ID: <CAM9d7cgJUgVGpQJ=KgH+3RzAJ8Ce50Acr=gSZh_ZLuiTa3-sEg@mail.gmail.com>
Subject: Re: [PATCH] perf test: Fix fails of perf stat --bpf-counters
 --for-each-cgroup on s390
To:     Thomas Richter <tmricht@linux.ibm.com>
Cc:     linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
        acme@kernel.org, svens@linux.ibm.com, gor@linux.ibm.com,
        sumanthk@linux.ibm.com, hca@linux.ibm.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

Hello,

On Thu, Dec 7, 2023 at 4:57 AM Thomas Richter <tmricht@linux.ibm.com> wrote:
>
> On s390 this test fails very often, as can be observed in the output
> below. This is caused by the second test function
> check_cpu_list_counted(). The perf stat is triggered for 2 CPUs
> 0 and 1.  On s390, which usually has a lot more CPUs, most often
> this ends up in no counter increments on these 2 CPUs 0 and 1.
>
> Fix this and trigger explicit workload on CPU 0 and 1 for
> systemd. This is a better approach than calculating a long
> list of CPUs (which is basicly the same as option -a), or
> wait a longer period of time.
>
> Output before:
>  # for i in $(seq 10)
>  > do ./perf test 100
>  > done
>  100: perf stat --bpf-counters --for-each-cgroup test : FAILED!
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test : FAILED!
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test : FAILED!
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test : FAILED!
>  100: perf stat --bpf-counters --for-each-cgroup test : Ok
>  #
>
> Output after:
>  # for i in $(seq 10);
>  do ./perf test 100;
>  done
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  100: perf stat --bpf-counters --for-each-cgroup test  : Ok
>  #
>
> Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
> ---
>  tools/perf/tests/shell/stat_bpf_counters_cgrp.sh | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/tools/perf/tests/shell/stat_bpf_counters_cgrp.sh b/tools/perf/tests/shell/stat_bpf_counters_cgrp.sh
> index e75d0780dc78..f67602321403 100755
> --- a/tools/perf/tests/shell/stat_bpf_counters_cgrp.sh
> +++ b/tools/perf/tests/shell/stat_bpf_counters_cgrp.sh
> @@ -60,6 +60,7 @@ check_system_wide_counted()
>
>  check_cpu_list_counted()
>  {
> +       taskset -c 0,1 systemctl daemon-reexec &

Thanks for the patch.  But I think it should support
machines without systemd (or maybe with old versions).

Also probably you want to reset the behavior after
the test.  I think we can just run some built-in test
workload like `perf test -w thloop`.

Thanks,
Namhyung


>         check_cpu_list_counted_output=$(perf stat -C 0,1 --bpf-counters --for-each-cgroup ${test_cgroups} -e cpu-clock -x, taskset -c 1 sleep 1  2>&1)
>         if echo ${check_cpu_list_counted_output} | grep -q -F "<not "; then
>                 echo "Some CPU events are not counted"
> --
> 2.43.0
>