Re: [FFmpeg-devel] [PATCH 1/3] avcodec/cbs_jpeg: Assert that cbs_jpeg_assemble_fragment() stays within the array

From: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
To: ffmpeg-devel@ffmpeg.org
Subject: Re: [FFmpeg-devel] [PATCH 1/3] avcodec/cbs_jpeg: Assert that cbs_jpeg_assemble_fragment() stays within the array
Date: Wed, 8 May 2024 11:46:45 +0200
Message-ID: <GV1P250MB073713D9C61F31CA1F27B3088FE52@GV1P250MB0737.EURP250.PROD.OUTLOOK.COM> (raw)
In-Reply-To: <20240508023923.28209-1-michael@niedermayer.cc>

Michael Niedermayer:
> Inspired by: CID1473561 Untrusted pointer write
> 
> Sponsored-by: Sovereign Tech Fund
> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
> ---
>  libavcodec/cbs_jpeg.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/libavcodec/cbs_jpeg.c b/libavcodec/cbs_jpeg.c
> index b1b58dcd65e..628841c5f37 100644
> --- a/libavcodec/cbs_jpeg.c
> +++ b/libavcodec/cbs_jpeg.c
> @@ -401,6 +401,7 @@ static int cbs_jpeg_assemble_fragment(CodedBitstreamContext *ctx,
>                  } else {
>                      data[dp++] = unit->data[sp];
>                  }
> +                av_assert0(dp <= size - 2);
>              }
>          }
>      }

You want to add an av_assert0 to a hot loop (or rather: to what would be
a hot loop in case this code were executed) just because Coverity thinks
that reading data with a different endianness taints this data? (That
the supposedly tainted variable has actually already been checked via an
assert makes this even more crazy.)

- Andreas

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".