avcodec/ffv1dec: Fix data races emanating from copying whole context
ffmpeg | branch: master | Andreas Rheinhardt <andreas.rheinhardt@outlook.com> | Wed Apr 21 19:25:22 2021 +0200| [0cfc7418bbf55f7b316862598c4e59434c65b3a8] | committer: Andreas Rheinhardt avcodec/ffv1dec: Fix data races emanating from copying whole context When using frame threading, the FFV1 decoder's update_thread_context() function copies the whole context and afterwards restores some allocated fields with backups made earlier. Among these fields are the ThreadFrames and the source context's ThreadFrames can change concurrently without any synchronization, leading to data races which are undefined behaviour even if they don't lead to problems in practice (as the destination's own ThreadFrames are restored directly thereafter). Fix this by only copying the actually needed fields. Reviewed-by: Michael Niedermayer <michael@niedermayer.cc> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=0cfc7418bbf55f7b31...
libavcodec/ffv1dec.c | 19 ++++++------------- 1 file changed, 6 insertions(+), 13 deletions(-) diff --git a/libavcodec/ffv1dec.c b/libavcodec/ffv1dec.c index cdf72a25dd..95beae09ad 100644 --- a/libavcodec/ffv1dec.c +++ b/libavcodec/ffv1dec.c @@ -1072,18 +1072,12 @@ static int update_thread_context(AVCodecContext *dst, const AVCodecContext *src) if (dst == src) return 0; - { - ThreadFrame picture = fdst->picture, last_picture = fdst->last_picture; - uint8_t (*initial_states[MAX_QUANT_TABLES])[32]; - struct FFV1Context *slice_context[MAX_SLICES]; - memcpy(initial_states, fdst->initial_states, sizeof(fdst->initial_states)); - memcpy(slice_context, fdst->slice_context , sizeof(fdst->slice_context)); - - memcpy(fdst, fsrc, sizeof(*fdst)); - memcpy(fdst->initial_states, initial_states, sizeof(fdst->initial_states)); - memcpy(fdst->slice_context, slice_context , sizeof(fdst->slice_context)); - fdst->picture = picture; - fdst->last_picture = last_picture; + copy_fields(fdst, fsrc, fsrc); + fdst->use32bit = fsrc->use32bit; + memcpy(fdst->state_transition, fsrc->state_transition, + sizeof(fdst->state_transition)); + memcpy(fdst->quant_table, fsrc->quant_table, sizeof(fsrc->quant_table)); + for (i = 0; i<fdst->num_h_slices * fdst->num_v_slices; i++) { FFV1Context *fssrc = fsrc->slice_context[i]; FFV1Context *fsdst = fdst->slice_context[i]; @@ -1091,7 +1085,6 @@ static int update_thread_context(AVCodecContext *dst, const AVCodecContext *src) } av_assert0(!fdst->plane[0].state); av_assert0(!fdst->sample_buffer); - } av_assert1(fdst->max_slice_count == fsrc->max_slice_count);
participants (1)
-
Andreas Rheinhardt