linux-kernel - Re: [PATCH 2/2] perf inject: Fix segfault due to perf_data_

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: <YbdHBhw7lk/6vKd5@kernel.org>
Date:   Mon, 13 Dec 2021 10:13:42 -0300
From:   Arnaldo Carvalho de Melo <acme@...nel.org>
To:     Adrian Hunter <adrian.hunter@...el.com>
Cc:     Jiri Olsa <jolsa@...hat.com>, linux-kernel@...r.kernel.org,
        Riccardo Mancini <rickyman7@...il.com>,
        Namhyung Kim <namhyung@...nel.org>
Subject: Re: [PATCH 2/2] perf inject: Fix segfault due to perf_data__fd()
 without open

Em Mon, Dec 13, 2021 at 10:48:29AM +0200, Adrian Hunter escreveu:
> The fixed commit attempts to get the output file descriptor even if the
> file was never opened e.g.
> 
>  $ perf record uname
>  Linux
>  [ perf record: Woken up 1 times to write data ]
>  [ perf record: Captured and wrote 0.002 MB perf.data (7 samples) ]
>  $ perf inject -i perf.data --vm-time-correlation=dry-run
>  Segmentation fault (core dumped)
>  $ gdb --quiet perf
>  Reading symbols from perf...
>  (gdb) r inject -i perf.data --vm-time-correlation=dry-run
>  Starting program: /home/ahunter/bin/perf inject -i perf.data --vm-time-correlation=dry-run
>  [Thread debugging using libthread_db enabled]
>  Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
> 
>  Program received signal SIGSEGV, Segmentation fault.
>  __GI___fileno (fp=0x0) at fileno.c:35
>  35      fileno.c: No such file or directory.
>  (gdb) bt
>  #0  __GI___fileno (fp=0x0) at fileno.c:35
>  #1  0x00005621e48dd987 in perf_data__fd (data=0x7fff4c68bd08) at util/data.h:72
>  #2  perf_data__fd (data=0x7fff4c68bd08) at util/data.h:69
>  #3  cmd_inject (argc=<optimized out>, argv=0x7fff4c69c1f0) at builtin-inject.c:1017
>  #4  0x00005621e4936783 in run_builtin (p=0x5621e4ee6878 <commands+600>, argc=4, argv=0x7fff4c69c1f0) at perf.c:313
>  #5  0x00005621e4897d5c in handle_internal_command (argv=<optimized out>, argc=<optimized out>) at perf.c:365
>  #6  run_argv (argcp=<optimized out>, argv=<optimized out>) at perf.c:409
>  #7  main (argc=4, argv=0x7fff4c69c1f0) at perf.c:539
>  (gdb)

Thanks, applied.

- Arnaldo

 
> Fixes: 0ae03893623d ("perf tools: Pass a fd to perf_file_header__read_pipe()")
> Cc: stable@...r.kernel.org
> Signed-off-by: Adrian Hunter <adrian.hunter@...el.com>
> ---
>  tools/perf/builtin-inject.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
> index af70f1c72052..409b721666cb 100644
> --- a/tools/perf/builtin-inject.c
> +++ b/tools/perf/builtin-inject.c
> @@ -755,12 +755,16 @@ static int parse_vm_time_correlation(const struct option *opt, const char *str,
>  	return inject->itrace_synth_opts.vm_tm_corr_args ? 0 : -ENOMEM;
>  }
>  
> +static int output_fd(struct perf_inject *inject)
> +{
> +	return inject->in_place_update ? -1 : perf_data__fd(&inject->output);
> +}
> +
>  static int __cmd_inject(struct perf_inject *inject)
>  {
>  	int ret = -EINVAL;
>  	struct perf_session *session = inject->session;
> -	struct perf_data *data_out = &inject->output;
> -	int fd = inject->in_place_update ? -1 : perf_data__fd(data_out);
> +	int fd = output_fd(inject);
>  	u64 output_data_offset;
>  
>  	signal(SIGINT, sig_handler);
> @@ -1015,7 +1019,7 @@ int cmd_inject(int argc, const char **argv)
>  	}
>  
>  	inject.session = __perf_session__new(&data, repipe,
> -					     perf_data__fd(&inject.output),
> +					     output_fd(&inject),
>  					     &inject.tool);
>  	if (IS_ERR(inject.session)) {
>  		ret = PTR_ERR(inject.session);
> -- 
> 2.25.1

-- 

- Arnaldo