| Message ID | 20180501204407.769-1-michael@niedermayer.cc |
|---|---|
| State | Accepted |
| Commit | 16d8b13b3b26c19d7f8856e039fe6662d96b4ff3 |
| Headers | show |
On Tue, 1 May 2018 22:44:07 +0200 Michael Niedermayer <michael@niedermayer.cc> wrote: > Regression since: af1761f7 > Fixes: Division by 0 > Fixes: ffmpeg_crash_1 > > Found-by: Thuan Pham, Marcel Böhme, Andrew Santosa and Alexandru Razvan Caciulescu with AFLSmart > Signed-off-by: Michael Niedermayer <michael@niedermayer.cc> > --- > fftools/ffmpeg.c | 8 ++++++-- > 1 file changed, 6 insertions(+), 2 deletions(-) > > diff --git a/fftools/ffmpeg.c b/fftools/ffmpeg.c > index 5dc198f933..15fa54f24e 100644 > --- a/fftools/ffmpeg.c > +++ b/fftools/ffmpeg.c > @@ -2706,8 +2706,12 @@ static int process_input_packet(InputStream *ist, const AVPacket *pkt, int no_eo > ist->dts = ist->next_dts; > switch (ist->dec_ctx->codec_type) { > case AVMEDIA_TYPE_AUDIO: > - ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / > - ist->dec_ctx->sample_rate; > + if (ist->dec_ctx->sample_rate) { > + ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / > + ist->dec_ctx->sample_rate; > + } else { > + ist->next_dts += av_rescale_q(pkt->duration, ist->st->time_base, AV_TIME_BASE_Q); > + } > break; > case AVMEDIA_TYPE_VIDEO: > if (ist->framerate.num) { Wouldn't it be better to error out here, instead of using yet another unreliable value and "hoping for the best"? Also I'm not sure, but I think unknown durations are sometimes set to -1 instead of 0.
On Wed, May 02, 2018 at 02:57:11AM +0200, wm4 wrote: > On Tue, 1 May 2018 22:44:07 +0200 > Michael Niedermayer <michael@niedermayer.cc> wrote: > > > Regression since: af1761f7 > > Fixes: Division by 0 > > Fixes: ffmpeg_crash_1 > > > > Found-by: Thuan Pham, Marcel Böhme, Andrew Santosa and Alexandru Razvan Caciulescu with AFLSmart > > Signed-off-by: Michael Niedermayer <michael@niedermayer.cc> > > --- > > fftools/ffmpeg.c | 8 ++++++-- > > 1 file changed, 6 insertions(+), 2 deletions(-) > > > > diff --git a/fftools/ffmpeg.c b/fftools/ffmpeg.c > > index 5dc198f933..15fa54f24e 100644 > > --- a/fftools/ffmpeg.c > > +++ b/fftools/ffmpeg.c > > @@ -2706,8 +2706,12 @@ static int process_input_packet(InputStream *ist, const AVPacket *pkt, int no_eo > > ist->dts = ist->next_dts; > > switch (ist->dec_ctx->codec_type) { > > case AVMEDIA_TYPE_AUDIO: > > - ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / > > - ist->dec_ctx->sample_rate; > > + if (ist->dec_ctx->sample_rate) { > > + ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / > > + ist->dec_ctx->sample_rate; > > + } else { > > + ist->next_dts += av_rescale_q(pkt->duration, ist->st->time_base, AV_TIME_BASE_Q); > > + } > > break; > > case AVMEDIA_TYPE_VIDEO: > > if (ist->framerate.num) { > > Wouldn't it be better to error out here, instead of using yet another > unreliable value and "hoping for the best"? > > Also I'm not sure, but I think unknown durations are sometimes set to > -1 instead of 0. The same update this patch would add is also done in the AVMEDIA_TYPE_VIDEO below the changed AVMEDIA_TYPE_AUDIO. I can add a special case and error out but then audio and video differ more about -1, ill add a assert, the existing code prior to the patch would already misbehave (in some cases) if its -1. So i will have to fix this even without the patch if it happens thx [...]
diff --git a/fftools/ffmpeg.c b/fftools/ffmpeg.c index 5dc198f933..15fa54f24e 100644 --- a/fftools/ffmpeg.c +++ b/fftools/ffmpeg.c @@ -2706,8 +2706,12 @@ static int process_input_packet(InputStream *ist, const AVPacket *pkt, int no_eo ist->dts = ist->next_dts; switch (ist->dec_ctx->codec_type) { case AVMEDIA_TYPE_AUDIO: - ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / - ist->dec_ctx->sample_rate; + if (ist->dec_ctx->sample_rate) { + ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) / + ist->dec_ctx->sample_rate; + } else { + ist->next_dts += av_rescale_q(pkt->duration, ist->st->time_base, AV_TIME_BASE_Q); + } break; case AVMEDIA_TYPE_VIDEO: if (ist->framerate.num) {
Regression since: af1761f7 Fixes: Division by 0 Fixes: ffmpeg_crash_1 Found-by: Thuan Pham, Marcel Böhme, Andrew Santosa and Alexandru Razvan Caciulescu with AFLSmart Signed-off-by: Michael Niedermayer <michael@niedermayer.cc> --- fftools/ffmpeg.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)