I've found out that `in_fmt` is equal to -1 at the place of error, so we just need to insert check at the beginning of `swr_init` function to check fmts positivity.

On 2023-08-02 13:51, Andreas Rheinhardt wrote:

kobrineli:

From: Eli Kobrin <kobrin...@ispras.ru>

We've been fuzzing torchvision with [sydr-fuzz](https://github.com/ispras/oss-sydr-fuzz)
and found out of bounds error in ffmpeg project at audioconvert.c:51.
To prevent error we need to insert corresponding check.

Signed-off-by: Eli Kobrin <kobrin...@ispras.ru>
---
libswresample/audioconvert.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/libswresample/audioconvert.c b/libswresample/audioconvert.c
index 1d75ba1495..701f4808a0 100644
--- a/libswresample/audioconvert.c
+++ b/libswresample/audioconvert.c
@@ -148,7 +148,12 @@ AudioConvert *swri_audio_convert_alloc(enum AVSampleFormat out_fmt,
int flags)
{
AudioConvert *ctx;
- conv_func_type *f = fmt_pair_to_conv_functions[av_get_packed_sample_fmt(out_fmt) + AV_SAMPLE_FMT_NB*av_get_packed_sample_fmt(in_fmt)];
+
+ size_t idx = av_get_packed_sample_fmt(out_fmt) + AV_SAMPLE_FMT_NB * av_get_packed_sample_fmt(in_fmt);
+    if (idx >= AV_SAMPLE_FMT_NB * AV_SAMPLE_FMT_NB)
+        return NULL;
+
+    conv_func_type *f = fmt_pair_to_conv_functions[idx];

if (!f)
return NULL;

Something seems to be using an invalid sample format (either out_fmt or
in_fmt). You should investigate where this comes from.
(Given that this is a public function, we should probably validate user
input; and maybe stop using AV_SAMPLE_FMT_NB altogether.)

- 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-requ...@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

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

Reply via email to