[FFmpeg-devel,4/5] avcodec/atrac3: Check for huge block aligns

Submitted by Michael Niedermayer on Nov. 2, 2019, 4:06 p.m.

Details

Message ID 20191102160601.14711-4-michael@niedermayer.cc
State New
Headers show

Commit Message

Michael Niedermayer Nov. 2, 2019, 4:06 p.m.
The largest documented frame size = block align is 1024 bytes
(https://wiki.multimedia.cx/index.php/ATRAC3)

Without a limit this can allocate arbitrary memory and trigger OOM
Fixes: OOM
Fixes: 18337/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ATRAC3_fuzzer-5763861478637568
Fixes: 18556/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ATRAC3AL_fuzzer-5646183334936576

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
---
 libavcodec/atrac3.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Michael Niedermayer Nov. 20, 2019, 2:10 p.m.
On Sat, Nov 02, 2019 at 05:06:00PM +0100, Michael Niedermayer wrote:
> The largest documented frame size = block align is 1024 bytes
> (https://wiki.multimedia.cx/index.php/ATRAC3)
> 
> Without a limit this can allocate arbitrary memory and trigger OOM
> Fixes: OOM
> Fixes: 18337/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ATRAC3_fuzzer-5763861478637568
> Fixes: 18556/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ATRAC3AL_fuzzer-5646183334936576
> 
> Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
> ---
>  libavcodec/atrac3.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

will apply

[...]

Patch hide | download patch | download mbox

diff --git a/libavcodec/atrac3.c b/libavcodec/atrac3.c
index dc19a3863e..067aa23f1f 100644
--- a/libavcodec/atrac3.c
+++ b/libavcodec/atrac3.c
@@ -964,7 +964,7 @@  static av_cold int atrac3_decode_init(AVCodecContext *avctx)
         return AVERROR_INVALIDDATA;
     }
 
-    if (avctx->block_align >= UINT_MAX / 2 || avctx->block_align <= 0)
+    if (avctx->block_align > 1024 || avctx->block_align <= 0)
         return AVERROR(EINVAL);
 
     q->decoded_bytes_buffer = av_mallocz(FFALIGN(avctx->block_align, 4) +