[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAP-5=fWu0yCSCB__prps2g=JRDScbhWuimPJHUsU0S_c0ywDhg@mail.gmail.com>
Date: Mon, 12 Jun 2023 09:32:30 -0700
From: Ian Rogers <irogers@...gle.com>
To: James Clark <james.clark@....com>
Cc: linux-perf-users@...r.kernel.org,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
Adrian Hunter <adrian.hunter@...el.com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] perf map: Fix refcount errors on Arm with -DREFCNT_CHECKING=1
On Mon, Jun 12, 2023 at 8:05 AM James Clark <james.clark@....com> wrote:
>
> When quitting after running a perf report, the refcount checker finds
> some double frees. The issue is that map__put() is called on a function
> argument so it removes the refcount wrapper that someone else was using.
>
> Fix it by only calling map__put() on a reference that is owned by this
> function.
>
> Signed-off-by: James Clark <james.clark@....com>
Acked-by: Ian Rogers <irogers@...gle.com>
> ---
> tools/perf/util/symbol-elf.c | 9 +++++----
> tools/perf/util/symbol.c | 9 +++++----
> 2 files changed, 10 insertions(+), 8 deletions(-)
>
> diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
> index 63882a4db5c7..ec0d7810bbb0 100644
> --- a/tools/perf/util/symbol-elf.c
> +++ b/tools/perf/util/symbol-elf.c
> @@ -1365,6 +1365,7 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map,
> struct dso *curr_dso = *curr_dsop;
> struct map *curr_map;
> char dso_name[PATH_MAX];
> + struct map *map_ref;
nit: can we narrow the scope of this by moving it to the scope where it is used.
>
> /* Adjust symbol to map to file offset */
> if (adjust_kernel_syms)
> @@ -1390,10 +1391,10 @@ static int dso__process_kernel_symbol(struct dso *dso, struct map *map,
> if (kmaps) {
> int err;
>
> - map__get(map);
> - maps__remove(kmaps, map);
> - err = maps__insert(kmaps, map);
> - map__put(map);
> + map_ref = map__get(map);
> + maps__remove(kmaps, map_ref);
> + err = maps__insert(kmaps, map_ref);
> + map__put(map_ref);
> if (err)
> return err;
> }
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index 6b9c55784b56..b3034fd5c0af 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -1368,6 +1368,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
> int err, fd;
> char kcore_filename[PATH_MAX];
> u64 stext;
> + struct map *map_ref;
nit: can we narrow the scope of this by moving it to the scope where it is used.
Thanks,
Ian
>
> if (!kmaps)
> return -EINVAL;
> @@ -1464,10 +1465,10 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
> map__set_map_ip(map, map__map_ip_ptr(new_map));
> map__set_unmap_ip(map, map__unmap_ip_ptr(new_map));
> /* Ensure maps are correctly ordered */
> - map__get(map);
> - maps__remove(kmaps, map);
> - err = maps__insert(kmaps, map);
> - map__put(map);
> + map_ref = map__get(map);
> + maps__remove(kmaps, map_ref);
> + err = maps__insert(kmaps, map_ref);
> + map__put(map_ref);
> map__put(new_map);
> if (err)
> goto out_err;
> --
> 2.34.1
>
Powered by blists - more mailing lists