linux-kernel - Re: [PATCH 1/9] perf evlist: Fix cpu/thread map leak

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: <20200907112925.GJ1199773@krava>
Date:   Mon, 7 Sep 2020 13:29:25 +0200
From:   Jiri Olsa <jolsa@...hat.com>
To:     Namhyung Kim <namhyung@...nel.org>
Cc:     Arnaldo Carvalho de Melo <acme@...nel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <a.p.zijlstra@...llo.nl>,
        Mark Rutland <mark.rutland@....com>,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Stephane Eranian <eranian@...gle.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Andi Kleen <andi@...stfloor.org>,
        Ian Rogers <irogers@...gle.com>
Subject: Re: [PATCH 1/9] perf evlist: Fix cpu/thread map leak

On Mon, Sep 07, 2020 at 12:44:54PM +0900, Namhyung Kim wrote:
> Asan reported leak of cpu and thread maps as they have one more
> refcount than released.  I found that after setting evlist maps it
> should release it's refcount.
> 
> It seems to be broken from the beginning so I chose the original
> commit as the culprit.  But not sure how it's applied to stable trees
> since there are many changes in the code after that.
> 
> Fixes: 7e2ed097538c5 ("perf evlist: Store pointer to the cpu and thread maps")
> Fixes: 4112eb1899c0e ("perf evlist: Default to syswide target when no thread/cpu maps set")
> Signed-off-by: Namhyung Kim <namhyung@...nel.org>
> ---
>  tools/perf/util/evlist.c | 11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> index e3fa3bf7498a..c0768c61eb43 100644
> --- a/tools/perf/util/evlist.c
> +++ b/tools/perf/util/evlist.c
> @@ -946,6 +946,10 @@ int perf_evlist__create_maps(struct evlist *evlist, struct target *target)
>  
>  	perf_evlist__set_maps(&evlist->core, cpus, threads);
>  
> +	/* as evlist now has references, put count here */
> +	perf_cpu_map__put(cpus);
> +	perf_thread_map__put(threads);

nice catch, I checked perf_evlist__create_syswide_maps is doing this
correctly, but I think we might have the same issue in script's
set_maps function

thanks,
jirka

> +
>  	return 0;
>  
>  out_delete_threads:
> @@ -1273,11 +1277,12 @@ static int perf_evlist__create_syswide_maps(struct evlist *evlist)
>  		goto out_put;
>  
>  	perf_evlist__set_maps(&evlist->core, cpus, threads);
> -out:
> -	return err;
> +
> +	perf_thread_map__put(threads);
>  out_put:
>  	perf_cpu_map__put(cpus);
> -	goto out;
> +out:
> +	return err;
>  }
>  
>  int evlist__open(struct evlist *evlist)
> -- 
> 2.28.0.526.ge36021eeef-goog
>