[FFmpeg-devel] [PATCH 2/2] avcodec/arbc: Use ff_reget_buffer()

James Almer jamrial at gmail.com
Fri Aug 30 17:15:21 EEST 2019


On 8/29/2019 4:33 PM, James Almer wrote:
> On 8/29/2019 4:04 PM, Michael Niedermayer wrote:
>> Fixes: Timeout (20sec -> 0.5sec)
>> Fixes: 16404/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ARBC_fuzzer-5660216355454976
>>
>> Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
>> Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
>> ---
>>  libavcodec/arbc.c | 21 +++++++--------------
>>  1 file changed, 7 insertions(+), 14 deletions(-)
>>
>> diff --git a/libavcodec/arbc.c b/libavcodec/arbc.c
>> index 06970f140b..f0de50e243 100644
>> --- a/libavcodec/arbc.c
>> +++ b/libavcodec/arbc.c
>> @@ -136,15 +136,9 @@ static int decode_frame(AVCodecContext *avctx, void *data,
>>      if (7 * nb_segments > bytestream2_get_bytes_left(&s->gb))
>>          return AVERROR_INVALIDDATA;
>>  
>> -    if ((ret = ff_get_buffer(avctx, frame, AV_GET_BUFFER_FLAG_REF)) < 0)
>> +    if ((ret = ff_reget_buffer(avctx, s->prev_frame)) < 0)
>>          return ret;
>>  
>> -    if (s->prev_frame->data[0]) {
>> -        ret = av_frame_copy(frame, s->prev_frame);
>> -        if (ret < 0)
>> -            return ret;
>> -    }
>> -
>>      for (int i = 0; i < nb_segments; i++) {
>>          int resolution_flag;
>>          int fill;
>> @@ -161,19 +155,18 @@ static int decode_frame(AVCodecContext *avctx, void *data,
>>          resolution_flag = bytestream2_get_byte(&s->gb);
>>  
>>          if (resolution_flag & 0x10)
>> -            prev_pixels -= fill_tileX(avctx, 1024, 1024, fill, frame);
>> +            prev_pixels -= fill_tileX(avctx, 1024, 1024, fill, s->prev_frame);
>>          if (resolution_flag & 0x08)
>> -            prev_pixels -= fill_tileX(avctx, 256, 256, fill, frame);
>> +            prev_pixels -= fill_tileX(avctx, 256, 256, fill, s->prev_frame);
>>          if (resolution_flag & 0x04)
>> -            prev_pixels -= fill_tileX(avctx, 64, 64, fill, frame);
>> +            prev_pixels -= fill_tileX(avctx, 64, 64, fill, s->prev_frame);
>>          if (resolution_flag & 0x02)
>> -            prev_pixels -= fill_tileX(avctx, 16, 16, fill, frame);
>> +            prev_pixels -= fill_tileX(avctx, 16, 16, fill, s->prev_frame);
>>          if (resolution_flag & 0x01)
>> -            prev_pixels -= fill_tile4(avctx, fill, frame);
>> +            prev_pixels -= fill_tile4(avctx, fill, s->prev_frame);
>>      }
>>  
>> -    av_frame_unref(s->prev_frame);
>> -    if ((ret = av_frame_ref(s->prev_frame, frame)) < 0)
>> +    if ((ret = av_frame_ref(frame, s->prev_frame)) < 0)
>>          return ret;
>>  
>>      frame->pict_type = prev_pixels <= 0 ? AV_PICTURE_TYPE_I : AV_PICTURE_TYPE_P;
> 
> You could while at it revert 8b10f09fd53, then implement d70bbdc5fa for
> this decoder as part of this patch.

Or maybe we should do what you suggested and simplify things by either
adding a new reget function, or changing the current one to take a new
argument to request a writable buffer.
The latter sounds better, IMO, and I can look into implementing it.


More information about the ffmpeg-devel mailing list