On Sat, Jun 21, 2025 at 11:15:19PM +0200, Michael Niedermayer wrote:
> Fixes: signed integer overflow: 2147483647 + 1913651185 cannot be represented in type 'int'
> Fixes: 418335931/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_RV60_fuzzer-6568264620900352
> 
> Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
> ---
>  libavcodec/rv60dec.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/libavcodec/rv60dec.c b/libavcodec/rv60dec.c
> index 2bbcb1d6209..dbafa7eed02 100644
> --- a/libavcodec/rv60dec.c
> +++ b/libavcodec/rv60dec.c
> @@ -409,6 +409,8 @@ static int read_slice_sizes(RV60Context *s, GetBitContext *gb)
>          if (last_size <= 0)
>              return AVERROR_INVALIDDATA;
>          s->slice[i].size = last_size;
> +        if (s->slice[i].size > INT32_MAX - sum)
> +            return AVERROR_INVALIDDATA;
>          sum += s->slice[i].size;

As it turns out, sum is not used for any purpose.
This patch (https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2025-June/345621.html)
drops sum, obviating the need for your patch.

However later on in rv60dec.c, line ~2354, there is a similar summation
of slice sizes. This might need a INT_MAX guard?

      for (int i = 0; i < s->cu_height; i++) {
          if (header_size + ofs >= avpkt->size)
              return AVERROR_INVALIDDATA;
          s->slice[i].data = avpkt->data + header_size + ofs;
          s->slice[i].data_size = FFMIN(s->slice[i].size, avpkt->size - header_size - ofs);
          ofs += s->slice[i].size;
      }


-- Peter
(A907 E02F A6E5 0CD2 34CD 20D2 6760 79C5 AC40 DD6B)