[FFmpeg-devel] [FFmpeg-cvslog] lavc/h264_levels: Avoid integer overflow in bitrate

[Carl Eugen Hoyos]

2018-09-25 0:16 GMT+02:00, Mark Thompson <git at videolan.org>:
> ffmpeg | branch: master | Mark Thompson <sw at jkqxz.net> | Mon Sep 24 23:03:32
> 2018 +0100| [581b4125aa187f2cf848d7a27e6128573c80dc64] | committer: Mark
> Thompson
>
> lavc/h264_levels: Avoid integer overflow in bitrate
>
> Fixes CID #1439656.
>
>> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=581b4125aa187f2cf848d7a27e6128573c80dc64
> ---
>
>  libavcodec/h264_levels.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libavcodec/h264_levels.c b/libavcodec/h264_levels.c
> index 6b4e18a914..737b7dcf06 100644
> --- a/libavcodec/h264_levels.c
> +++ b/libavcodec/h264_levels.c
> @@ -105,7 +105,7 @@ const H264LevelDescriptor *ff_h264_guess_level(int
> profile_idc,
>          if (level->constraint_set3_flag && no_cs3f)
>              continue;
>
> -        if (bitrate > level->max_br * h264_get_br_factor(profile_idc))
> +        if (bitrate > (int64_t)level->max_br *
> h264_get_br_factor(profile_idc))

If the overflow is possible at all (I doubt it), wouldn't it be cleaner
to change the type of cpb_br_nal_factor to uint32_t?

Carl Eugen