mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Adrian Hunter <adrian.hunter@intel.com>
To: "James O. D. Hunt" <james.o.hunt@intel.com>,
	<linux-perf-users@vger.kernel.org>,
	Ingo Molnar <mingo@redhat.com>,
	"Arnaldo Carvalho de Melo" <acme@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	"Mark Rutland" <mark.rutland@arm.com>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Jiri Olsa <jolsa@kernel.org>, "Ian Rogers" <irogers@google.com>,
	"Liang, Kan" <kan.liang@linux.intel.com>
Cc: <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] perf: Fix JSON output crash with intel_pt// samples
Date: Thu, 17 Apr 2025 09:25:11 +0300	[thread overview]
Message-ID: <cf18217c-a650-405b-86fb-7c60472a4b64@intel.com> (raw)
In-Reply-To: <20250407162211.172951-1-james.o.hunt@intel.com>

Subject usually contains the relevant perf command.
e.g.
	perf data: Fix JSON output crash with intel_pt// samples
or
	perf data convert: Fix JSON output crash with intel_pt// samples

On 7/04/25 19:22, James O. D. Hunt wrote:
> Current behaviour:
> 
> ```bash

Might want to go easy on the backticks.

> $ perf record -e intel_pt// true
> $ perf data convert --to-json /tmp/perf.json
> Segmentation fault (core dumped)
> ```
> 
> With fix applied:
> 
> ```bash
> $ perf record -e intel_pt// true
> $ perf data convert --to-json /tmp/perf.json
> $ jq < /tmp/perf.json &>/dev/null && echo ok
> ok
> ```
> 
> The crash actually occurs in `intel_pt_process_auxtrace_info()` where
> `session->itrace_synth_opts->set` is unconditionally dereferenced. This
> platform-specific code could be changed to fix the issue, but this patch
> fixes the problem generally.
> 
> Fixes: f6986c95af84ff2a76847910b4322f542b793bbf ("perf session: Add instruction tracing options")

Full commit hash is not necessary and no one uses it for Fixes tags.
Should be 12+ characters but not much more.

Note it is the wrong commit.  The code was added by
commit d0713d4ca3e94 ("perf data: Add JSON export")

> Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
> ---
>  tools/perf/util/data-convert-json.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
> index d9f805bf6fb0..b0aaa23659a0 100644
> --- a/tools/perf/util/data-convert-json.c
> +++ b/tools/perf/util/data-convert-json.c
> @@ -326,6 +326,10 @@ int bt_convert__perf2json(const char *input_name, const char *output_name,
>  		.force = opts->force,
>  	};
>  
> +	struct itrace_synth_opts itrace_synth_opts = {
> +		.set = 0,
> +	};

The problem is that this sub-command sets up the auxtrace callbacks:

	c2c.tool.auxtrace_info  = perf_event__process_auxtrace_info;
	c2c.tool.auxtrace       = perf_event__process_auxtrace;
	c2c.tool.auxtrace_error = perf_event__process_auxtrace_error;

but not itrace options.

Other sub-commands either support the --itrace option:
	perf script
	perf report
	perf inject
	perf annotate
Or are limited to memory events only:
	perf c2c report
	perf mem report

In this case it looks like it should either support the --itrace
option or remove the auxtrace callbacks.

> +
>  	perf_tool__init(&c.tool, /*ordered_events=*/true);
>  	c.tool.sample         = process_sample_event;
>  	c.tool.mmap           = perf_event__process_mmap;
> @@ -377,6 +381,8 @@ int bt_convert__perf2json(const char *input_name, const char *output_name,
>  		goto err_fclose;
>  	}
>  
> +	session->itrace_synth_opts = &itrace_synth_opts;
> +
>  	if (symbol__init(&session->header.env) < 0) {
>  		fprintf(stderr, "Symbol init error!\n");
>  		goto err_session_delete;


      reply	other threads:[~2025-04-17  6:25 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-07 16:22 James O. D. Hunt
2025-04-17  6:25 ` Adrian Hunter [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=cf18217c-a650-405b-86fb-7c60472a4b64@intel.com \
    --to=adrian.hunter@intel.com \
    --cc=acme@kernel.org \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=irogers@google.com \
    --cc=james.o.hunt@intel.com \
    --cc=jolsa@kernel.org \
    --cc=kan.liang@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®