linux-kernel - Re: [PATCH v3 4/5] perf record off-cpu: save embedded sample type

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAP-5=fW0AATSiUxm4wckT3Y3BvAUczNrzRHd13uvEtmgSw90fw@mail.gmail.com>
Date: Fri, 26 Jul 2024 17:49:49 -0700
From: Ian Rogers <irogers@...gle.com>
To: Howard Chu <howardchu95@...il.com>
Cc: namhyung@...nel.org, acme@...nel.org, adrian.hunter@...el.com, 
	jolsa@...nel.org, kan.liang@...ux.intel.com, linux-perf-users@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/5] perf record off-cpu: save embedded sample type

On Fri, Jul 26, 2024 at 3:28 AM Howard Chu <howardchu95@...il.com> wrote:
>
> We have to save the embedded sample type for it to be consumed correctly
> by perf script or perf report.
>
> This is a bad approach because it most definitely will break some
> perf.data convertor. Another approach is to add this sample_type_embed
> to perf_event_attr, but changing perf api is above my pay grade, so
> please give me your suggestions!
>
> Signed-off-by: Howard Chu <howardchu95@...il.com>
> ---
>  tools/perf/util/header.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
>
> diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
> index 55e9553861d0..ccb5493dc515 100644
> --- a/tools/perf/util/header.c
> +++ b/tools/perf/util/header.c
> @@ -80,6 +80,7 @@ const char perf_version_string[] = PERF_VERSION;
>
>  struct perf_file_attr {
>         struct perf_event_attr  attr;
> +       __u64 embed;
>         struct perf_file_section        ids;
>  };
>
> @@ -3713,6 +3714,7 @@ static int perf_session__do_write_header(struct perf_session *session,
>                 }
>                 f_attr = (struct perf_file_attr){
>                         .attr = evsel->core.attr,
> +                       .embed = evsel->sample_type_embed,
>                         .ids  = {
>                                 .offset = evsel->id_offset,
>                                 .size   = evsel->core.ids * sizeof(u64),
> @@ -4147,6 +4149,14 @@ static int read_attr(int fd, struct perf_header *ph,
>
>                 ret = readn(fd, ptr, left);
>         }
> +
> +       ret = readn(fd, &f_attr->embed, sizeof(f_attr->embed));
> +       if (ret <= 0) {
> +               pr_debug("cannot read %d bytes of embedded sample type\n",
> +                        PERF_ATTR_SIZE_VER0);

nit: sizeof will likely be larger than this, so prefer to this
constant? Maybe errno will be more useful.

Thanks,
Ian

> +               return -1;
> +       }
> +
>         /* read perf_file_section, ids are read in caller */
>         ret = readn(fd, &f_attr->ids, sizeof(f_attr->ids));
>
> @@ -4272,6 +4282,8 @@ int perf_session__read_header(struct perf_session *session, int repipe_fd)
>                 tmp = lseek(fd, 0, SEEK_CUR);
>                 evsel = evsel__new(&f_attr.attr);
>
> +               evsel->sample_type_embed = f_attr.embed;
> +
>                 if (evsel == NULL)
>                         goto out_delete_evlist;
>
> --
> 2.45.2
>