[FFmpeg-devel] [PATCH 2/3] avformat/sbgdec: Check opt_duration and start for overflow

[James Almer]

On 6/24/2021 5:57 PM, Michael Niedermayer wrote:
> Fixes: signed integer overflow: 2788626175500000000 + 7118941284000000000 cannot be represented in type 'long'
> Fixes: 35215/clusterfuzz-testcase-minimized-ffmpeg_dem_SBG_fuzzer-6123272247836672
> 
> 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/sbgdec.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/libavformat/sbgdec.c b/libavformat/sbgdec.c
> index dafdc4a1cc..0a6e927e57 100644
> --- a/libavformat/sbgdec.c
> +++ b/libavformat/sbgdec.c
> @@ -935,6 +935,9 @@ static int expand_timestamps(void *log, struct sbg_script *s)
>       }
>       if (s->start_ts == AV_NOPTS_VALUE)
>           s->start_ts = (s->opt_start_at_first && s->tseq) ? s->tseq[0].ts.t : now;
> +    if (av_sat_add64(s->start_ts, s->opt_duration) != s->start_ts + (uint64_t)s->opt_duration)

Can't this instead be an if (s->start_ts > INT64_MAX - s->opt_duration) 
check? Both s->start_ts and s->opt_duration are apparently guaranteed to 
be positive.

> +        return AVERROR_INVALIDDATA;
> +
>       s->end_ts = s->opt_duration ? s->start_ts + s->opt_duration :
>                   AV_NOPTS_VALUE; /* may be overridden later by -E option */
>       cur_ts = now;
>