[FFmpeg-devel] [PATCH 1/6] avformat/isom: Uninit layout in ff_mp4_read_dec_config_descr()

[Marton Balint]

On Mon, 1 Apr 2024, James Almer wrote:

> On 4/1/2024 8:40 PM, Michael Niedermayer wrote:
>>  On Mon, Apr 01, 2024 at 06:33:22PM -0300, James Almer wrote:
>>>  On 4/1/2024 5:56 PM, Michael Niedermayer wrote:
>>>> Fixes:  memleak
>>>> Fixes:  67442/clusterfuzz-testcase-minimized-ffmpeg_dem_CAF_fuzzer-5068813261406208
>>>>
>>>>  Found-by: continuous fuzzing process
>>>>  https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
>>>>  Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
>>>>  ---
>>>>     libavformat/isom.c | 1 +
>>>>     1 file changed, 1 insertion(+)
>>>>
>>>>  diff --git a/libavformat/isom.c b/libavformat/isom.c
>>>>  index 9fbccd4437f..c5930bd4d87 100644
>>>>  --- a/libavformat/isom.c
>>>>  +++ b/libavformat/isom.c
>>>>  @@ -359,6 +359,7 @@ int ff_mp4_read_dec_config_descr(AVFormatContext
>>>>  *fc, AVStream *st, AVIOContext
>>>>                                                     st->codecpar->extradata_size,
>>>>                 1, fc);
>>>>                 if (ret < 0)
>>>>                     return ret;
>>>>  +            av_channel_layout_uninit(&st->codecpar->ch_layout);
>>>>                 st->codecpar->ch_layout.order = AV_CHANNEL_ORDER_UNSPEC;
>>>
>>>  Not strictly related to this fix, but should we really drop the layout
>>>  here
>>>  by forcing unspec? We're also not even bothering to check if cfg.channels
>>>  matches st->codecpar->ch_layout.nb_channels.
>>
>>  I was wondering the same.
>>  I was hoping someone would know the specs well enouh to just say straight
>>  "this and that isnt legal together error out if thats true"
>
> Maybe Marton knows, since he's been working on this code.

I don't really know for sure. But it would make sense to me to only drop 
the layout if the channel count is different from what we already know.

Regards,
Marton