[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2036d7c3-8d4f-415a-9530-0758d440523f@linaro.org>
Date: Mon, 14 Oct 2024 10:17:47 +0100
From: James Clark <james.clark@...aro.org>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Adrian Hunter <adrian.hunter@...el.com>,
Alan Maguire <alan.maguire@...cle.com>, Howard Chu <howardchu95@...il.com>,
Ian Rogers <irogers@...gle.com>, Jiri Olsa <jolsa@...nel.org>,
Kan Liang <kan.liang@...ux.intel.com>, Namhyung Kim <namhyung@...nel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/1 perf-tools] perf trace: The return from 'write' isn't
a pid
On 11/10/2024 8:21 pm, Arnaldo Carvalho de Melo wrote:
> When adding a explicit beautifier for the 'write' syscall when the BPF
> based buffer collector was introduced there was a cut'n'paste error that
> carried the syscall_fmt->errpid setting from a nearby syscall (waitid)
> that returns a pid.
>
> So the write return was being suppressed by the return pretty printer,
> remove that field, reverting it back to the default return handler, that
> prints positive numbers as-is and interpret negative values as errnos.
>
> I actually introduced the problem while making Howard's original patch
> work just with the 'write' syscall, as we couldn't just look for any
> buffers, the ones that are filled in by the kernel couldn't use the same
> sys_enter BPF collector.
>
> Fixes: b257fac12f38d7f5 ("perf trace: Pretty print buffer data")
> Reported-by: James Clark <james.clark@...aro.org>
> Link: https://lore.kernel.org/lkml/bcf50648-3c7e-4513-8717-0d14492c53b9@linaro.org
> Cc: Adrian Hunter <adrian.hunter@...el.com>
> Cc: Alan Maguire <alan.maguire@...cle.com>
> Cc: Howard Chu <howardchu95@...il.com>
> Cc: Ian Rogers <irogers@...gle.com>
> Cc: Jiri Olsa <jolsa@...nel.org>
> Cc: Kan Liang <kan.liang@...ux.intel.com>
> Cc: Namhyung Kim <namhyung@...nel.org>
> Signed-off-by: Arnaldo Carvalho de Melo <acme@...hat.com>
> ---
> tools/perf/builtin-trace.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
> index f6e8475290739a10..d3f11b90d0255c7e 100644
> --- a/tools/perf/builtin-trace.c
> +++ b/tools/perf/builtin-trace.c
> @@ -1399,7 +1399,7 @@ static const struct syscall_fmt syscall_fmts[] = {
> .arg = { [2] = { .scnprintf = SCA_WAITID_OPTIONS, /* options */ }, }, },
> { .name = "waitid", .errpid = true,
> .arg = { [3] = { .scnprintf = SCA_WAITID_OPTIONS, /* options */ }, }, },
> - { .name = "write", .errpid = true,
> + { .name = "write",
> .arg = { [1] = { .scnprintf = SCA_BUF /* buf */, .from_user = true, }, }, },
> };
>
Fixes the return value:
0.000 ( 0.010 ms): echo/100910 write(fd: 1, buf: , count: 6)
= 6
I still see the missing address for buf, but I suppose that's a separate
issue that's introduced in the same commit. Also whether the address was
even useful I don't know.
Reviewed-by: James Clark <james.clark@...aro.org>
Powered by blists - more mailing lists