Re: [FFmpeg-devel] [PATCH v3 0/6] ffmpeg: late A/V encoder init, AVFrame metadata usage

2020-10-21 Thread Jan Ekström
On Fri, Oct 16, 2020, 16:16 Jan Ekström wrote: > This patch set started with a very simple wish to not have to set color > related values manually each time when utilizing ffmpeg.c. > > As of the third iteration, the following changes were done since the > second: > 1. A simple mistake was correc

Re: [FFmpeg-devel] [PATCH 2/4] avformat/wtvdec: Check dir_length

2020-10-21 Thread Michael Niedermayer
On Tue, Oct 20, 2020 at 07:54:23AM +1100, Peter Ross wrote: > On Mon, Oct 19, 2020 at 05:59:53PM +0200, Michael Niedermayer wrote: > > Fixes: Infinite loop > > Fixes: > > 26445/clusterfuzz-testcase-minimized-ffmpeg_dem_WTV_fuzzer-5125558331244544 > > > > Found-by: continuous fuzzing process > >

Re: [FFmpeg-devel] [PATCH 05/11] avformat/ffmetadec: finalize AVBPrint on errors

2020-10-21 Thread Michael Niedermayer
On Wed, Oct 21, 2020 at 06:07:42AM +0200, Andreas Rheinhardt wrote: > Michael Niedermayer: > > Fixes: memleak > > Fixes: > > 26450/clusterfuzz-testcase-minimized-ffmpeg_dem_FFMETADATA_fuzzer-6249850443923456 > > > > Found-by: continuous fuzzing process > > https://github.com/google/oss-fuzz/tree

Re: [FFmpeg-devel] [PATCH 4/5] avcodec/on2avc: Avoid indirection when calling float dsp function

2020-10-21 Thread Anton Khirnov
Quoting Andreas Rheinhardt (2020-10-21 07:37:03) > Anton Khirnov: > > Quoting Andreas Rheinhardt (2020-10-19 13:07:05) > >> Signed-off-by: Andreas Rheinhardt > >> --- > >> libavcodec/on2avc.c | 35 ++- > >> 1 file changed, 18 insertions(+), 17 deletions(-) > >> > >

Re: [FFmpeg-devel] [PATCH 03/11] avformat/mxfdec: Check type before setting sub_descriptors_refs

2020-10-21 Thread Andreas Rheinhardt
Michael Niedermayer: > Fixes: memleak > Fixes: > 26352/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5201158714687488 > > Found-by: continuous fuzzing process > https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg > Signed-off-by: Michael Niedermayer > --- > libavformat/mxfdec

Re: [FFmpeg-devel] [PATCH v1] libavformat/hls: During operation, the user exits and interrupts, causing pls->segment to be released, resulting in a null pointer crash

2020-10-21 Thread 徐慧书
Steven Liu 于2020年10月17日周六 下午4:57写道： > > > > 在 2020年10月17日，13:38，徐慧书 写道： > > > > Andreas Rheinhardt 于2020年10月16日周五 > 下午9:32写道： > > > >> javashu2...@gmail.com: > >>> From: bevis > >>> > >>> Signed-off-by: bevis > >>> --- > >>> libavformat/hls.c | 5 +++-- > >>> 1 file changed, 3 insertions(+), 2

Re: [FFmpeg-devel] [PATCH 3/7] avutil/common: Implement av_sat_add64_c() with fewer branches

2020-10-21 Thread Andreas Rheinhardt
Michael Niedermayer: > No benchmark because this is not used in any speed relevant pathes nor is it > used where __builtin_add_overflow is available. > So I do not know how to realistically benchmark it. > > Signed-off-by: Michael Niedermayer > --- > libavutil/common.h | 9 - > 1 file ch

Re: [FFmpeg-devel] [PATCH 06/11] avutil/eval: Unconditionally check argument of e_div

2020-10-21 Thread Michael Niedermayer
On Wed, Oct 21, 2020 at 07:20:41AM +0200, Andreas Rheinhardt wrote: > Michael Niedermayer: > > Fixes: division by zero > > Fixes: > > 26451/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVO_fuzzer-4756955832516608 > > > > Found-by: continuous fuzzing process > > https://github.com/google/oss-fuzz/t

Re: [FFmpeg-devel] [PATCH v1] libavformat/hls: During operation, the user exits and interrupts, causing pls->segment to be released, resulting in a null pointer crash

2020-10-21 Thread Steven Liu
> 2020年10月21日 下午5:16，徐慧书 写道： > > Steven Liu 于2020年10月17日周六 下午4:57写道： > >> >> >>> 在 2020年10月17日，13:38，徐慧书 写道： >>> >>> Andreas Rheinhardt 于2020年10月16日周五 >> 下午9:32写道： >>> javashu2...@gmail.com: > From: bevis > > Signed-off-by: bevis > --- > libavformat/hls.c | 5

Re: [FFmpeg-devel] [PATCH 3/7] avutil/common: Implement av_sat_add64_c() with fewer branches

2020-10-21 Thread Michael Niedermayer
On Wed, Oct 21, 2020 at 12:17:41PM +0200, Andreas Rheinhardt wrote: > Michael Niedermayer: > > No benchmark because this is not used in any speed relevant pathes nor is it > > used where __builtin_add_overflow is available. > > So I do not know how to realistically benchmark it. > > > > Signed-off

Re: [FFmpeg-devel] [PATCH 2/3] ffmpeg.c: refine picking default video stream

2020-10-21 Thread Michael Niedermayer
On Tue, Oct 20, 2020 at 01:41:50AM -0300, James Almer wrote: > On 10/19/2020 11:42 AM, Anton Khirnov wrote: > > Quoting James Almer (2020-10-14 19:26:51) > >> On 10/14/2020 2:09 PM, Michael Niedermayer wrote: > >>> On Wed, Oct 14, 2020 at 10:53:10AM +0200, Anton Khirnov wrote: > Use a floating

[FFmpeg-devel] Looking for contractors to improve DPX decoding

2020-10-21 Thread Dedy Yohann
Hello all, Sending my request here since I don't know which main dev is more suited to answer it. We (AV and digital department of the Cinémathèque française) are running into a couple of DPX decoding bugs and are looking for someone that can solve it. Our workflow is heavily relying on the con

Re: [FFmpeg-devel] [PATCH 2/3] ffmpeg.c: refine picking default video stream

2020-10-21 Thread James Almer
On 10/21/2020 8:17 AM, Michael Niedermayer wrote: > On Tue, Oct 20, 2020 at 01:41:50AM -0300, James Almer wrote: >> On 10/19/2020 11:42 AM, Anton Khirnov wrote: >>> Quoting James Almer (2020-10-14 19:26:51) On 10/14/2020 2:09 PM, Michael Niedermayer wrote: > On Wed, Oct 14, 2020 at 10:53:1

[FFmpeg-devel] [PATCH 2/3] avformat/mpegtsenc: make first_pcr sync with the first valid dts

2020-10-21 Thread lance . lmwang
From: Limin Wang now first_pts assume dts will start from zero, if it's not true(copyts is enable), too many null packet will be inserted for cbr output. Please test with below command, you'll get huge test.ts without the patch: ./ffmpeg -y -copyts -i ../fate-suite/mpegts/loewe.ts -c:v libx264

Re: [FFmpeg-devel] [PATCH 6/6] avformat/udp: cosmetics

2020-10-21 Thread lance . lmwang
On Wed, Oct 14, 2020 at 10:35:03PM +0800, lance.lmw...@gmail.com wrote: > From: Limin Wang > > Signed-off-by: Limin Wang > --- > libavformat/udp.c | 31 +++ > 1 file changed, 15 insertions(+), 16 deletions(-) > > diff --git a/libavformat/udp.c b/libavformat/udp.c >

[FFmpeg-devel] [PATCH 1/3] avformat/mpegtsenc: use total_size instead of avio_tell()

2020-10-21 Thread lance . lmwang
From: Limin Wang avio_tell() failed to get corrected data size in some condition. Please test with below command: $ ./ffmpeg -f lavfi -i testsrc=duration=100:size=vga -c:v libx264 -x264opts \ nal-hrd=cbr:force-cfr=1 -b:v 500k -minrate 500k -maxrate 500k -bufsize 500k \ -f hls -hls_time 10

Re: [FFmpeg-devel] [PATCH 03/11] avformat/mxfdec: Check type before setting sub_descriptors_refs

2020-10-21 Thread Tomas Härdin
tis 2020-10-20 klockan 22:56 +0200 skrev Michael Niedermayer: > Fixes: memleak > Fixes: > 26352/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5201158714687488 > > Found-by: continuous fuzzing process > https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg > Signed-off-by: Michael

[FFmpeg-devel] [PATCH 3/3] fftools/ffmpeg: fix the wrong bitrate and speed stats

2020-10-21 Thread lance . lmwang
From: Limin Wang The bitrate and speed stats are wrong If the copyts is enabled. Please test with below command: wget http://samples.mplayerhq.hu/MPEG2/foxksaz.ts ./ffmpeg -y -copyts -i ./foxksaz.ts -c:v libx264 -x264opts \ nal-hrd=cbr:force-cfr=1 -b:v 3500k -minrate 3500k -maxrate 3500k -buf

Re: [FFmpeg-devel] [PATCH 03/11] avformat/mxfdec: Check type before setting sub_descriptors_refs

2020-10-21 Thread Andreas Rheinhardt
Tomas Härdin: > tis 2020-10-20 klockan 22:56 +0200 skrev Michael Niedermayer: >> Fixes: memleak >> Fixes: >> 26352/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5201158714687488 >> >> Found-by: continuous fuzzing process >> https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg >>

Re: [FFmpeg-devel] [PATCH 03/11] avformat/mxfdec: Check type before setting sub_descriptors_refs

2020-10-21 Thread Tomas Härdin
ons 2020-10-21 klockan 16:48 +0200 skrev Andreas Rheinhardt: > Tomas Härdin: > > tis 2020-10-20 klockan 22:56 +0200 skrev Michael Niedermayer: > > > Fixes: memleak > > > Fixes: > > > 26352/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5201158714687488 > > > > > > Found-by: continuous fuzzi

Re: [FFmpeg-devel] [PATCH] avformat/rtsp: don't overwrite session control uri

2020-10-21 Thread Andriy Gelman
On Sat, 10. Oct 19:22, Andriy Gelman wrote: > From: Andriy Gelman > > Fixes #1941 > > Currently the session control uri gets overwritten by the media's uri > when mpegts is signalled in the media description. This happens because > s->nb_streams doesn't count mpegts which is instead part of the

Re: [FFmpeg-devel] [PATCH] Preventing Buffer Overflow for RTSP Links Increasing the buffer size of control uri, used when storing the input argument RTSP link. Following the Web URI standards, lengths

2020-10-21 Thread git
Hi, I've just checked the state of this patch on Patchwork. There are three builds listed, two successful and one warning. I've checked the logs for the warning, none seems to be about the submitted patch. I've also thoroughly tested the code prior to submitting. This is the first patch I'm su

Re: [FFmpeg-devel] Looking for contractors to improve DPX decoding

2020-10-21 Thread Carl Eugen Hoyos
Am Mi., 21. Okt. 2020 um 14:01 Uhr schrieb Dedy Yohann : > The ticket I opened a couple of month ago is one of those case : > https://trac.ffmpeg.org/ticket/8390 Two patches exist for this issue. Please provide samples for the other issues you see so we know what you actually need. Carl Eugen _

Re: [FFmpeg-devel] [PATCH 2/3] avformat/mpegtsenc: make first_pcr sync with the first valid dts

2020-10-21 Thread Devin Heitmueller
On Wed, Oct 21, 2020 at 9:48 AM wrote: > > From: Limin Wang > > now first_pts assume dts will start from zero, if it's not true(copyts is > enable), > too many null packet will be inserted for cbr output. > > Please test with below command, you'll get huge test.ts without the patch: > ./ffmpeg -

Re: [FFmpeg-devel] [PATCH 1/3] avformat/mpegtsenc: use total_size instead of avio_tell()

2020-10-21 Thread Marton Balint
On Wed, 21 Oct 2020, lance.lmw...@gmail.com wrote: From: Limin Wang avio_tell() failed to get corrected data size in some condition. Please explain. E.g. avio_tell() fails to get the amount of data written so far when the underlying IO context is flushed to segments. Please test with

[FFmpeg-devel] [PATCH 2/9] avformat/rsd: Check size and start before computing duration

2020-10-21 Thread Michael Niedermayer
Fixes: signed integer overflow: 100794754 * 28 cannot be represented in type 'int' Fixes: 26474/clusterfuzz-testcase-minimized-ffmpeg_dem_RSD_fuzzer-5181797606096896 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Nieder

[FFmpeg-devel] [PATCH 3/9] avformat/wavdec: Refuse to read chunks bigger than the filesize in w64_read_header()

2020-10-21 Thread Michael Niedermayer
Fixes: OOM Fixes: 26414/clusterfuzz-testcase-minimized-ffmpeg_dem_FWSE_fuzzer-5070632544632832 Fixes: 26475/clusterfuzz-testcase-minimized-ffmpeg_dem_W64_fuzzer-5770207722995712 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Mi

[FFmpeg-devel] [PATCH 5/9] avformat/vividas use avpriv_set_pts_info()

2020-10-21 Thread Michael Niedermayer
Fixes: assertion failure Fixes: 26482/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVIDAS_fuzzer-4905102324006912 Fixes: 26491/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVIDAS_fuzzer-6002953141616640 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffm

[FFmpeg-devel] [PATCH 4/9] avformat/xwma: Check for EOF in dpds_table read code

2020-10-21 Thread Michael Niedermayer
Fixes: Timeout (>30 -> 140ms) Fixes: 26478/clusterfuzz-testcase-minimized-ffmpeg_dem_XWMA_fuzzer-5918147066200064 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer --- libavformat/xwma.c | 4 1 file chang

[FFmpeg-devel] [PATCH 9/9] avformat/vividas: better check of current_sb_entry

2020-10-21 Thread Michael Niedermayer
This is the simplest fix for the problem, it is possible to instead check this when the variables are set and propagate errors and then fail earlier Fixes: out of array access Fixes: 26490/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVIDAS_fuzzer-5723367078100992 Found-by: continuous fuzzing proc

[FFmpeg-devel] [PATCH 6/9] avformat/vividas: Check for zero v_size

2020-10-21 Thread Michael Niedermayer
Fixes: SEGV on unknown address 0x Fixes: 26482/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVIDAS_fuzzer-4905102324006912 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer --- libavformat/vividas.c

[FFmpeg-devel] [PATCH 7/9] avformat: Do not search for child metadata in every packet

2020-10-21 Thread Michael Niedermayer
Fixes: Timeout Fixes: 26485/clusterfuzz-testcase-minimized-ffmpeg_dem_IFF_fuzzer-5126561373880320 Not tested due to lack of a testcase for the metadata transfer code Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Nieder

[FFmpeg-devel] [PATCH 8/9] avformat/iff: More completely check body_size

2020-10-21 Thread Michael Niedermayer
Fixes: infinite loop Fixes: 26485/clusterfuzz-testcase-minimized-ffmpeg_dem_IFF_fuzzer-5126561373880320 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer --- libavformat/iff.c | 4 +++- 1 file changed, 3 inser

[FFmpeg-devel] [PATCH 1/9] avformat/segafilm: Check that there is a stream

2020-10-21 Thread Michael Niedermayer
Fixes: assertion failure Fixes: 26472/clusterfuzz-testcase-minimized-ffmpeg_dem_SEGAFILM_fuzzer-5759751591559168 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer --- libavformat/segafilm.c | 3 +++ 1 file cha

Re: [FFmpeg-devel] [PATCH 2/3] avformat/mpegtsenc: make first_pcr sync with the first valid dts

2020-10-21 Thread lance . lmwang
On Wed, Oct 21, 2020 at 03:41:00PM -0400, Devin Heitmueller wrote: > On Wed, Oct 21, 2020 at 9:48 AM wrote: > > > > From: Limin Wang > > > > now first_pts assume dts will start from zero, if it's not true(copyts is > > enable), > > too many null packet will be inserted for cbr output. > > > > Pl

Re: [FFmpeg-devel] [PATCH 1/3] avformat/mpegtsenc: use total_size instead of avio_tell()

2020-10-21 Thread lance . lmwang
On Wed, Oct 21, 2020 at 11:39:10PM +0200, Marton Balint wrote: > > > On Wed, 21 Oct 2020, lance.lmw...@gmail.com wrote: > > > From: Limin Wang > > > > avio_tell() failed to get corrected data size in some condition. > > Please explain. E.g. avio_tell() fails to get the amount of data written

Re: [FFmpeg-devel] [PATCH 1/3] avformat/mpegtsenc: use total_size instead of avio_tell()

2020-10-21 Thread lance . lmwang
On Wed, Oct 21, 2020 at 11:39:10PM +0200, Marton Balint wrote: > > > On Wed, 21 Oct 2020, lance.lmw...@gmail.com wrote: > > > From: Limin Wang > > > > avio_tell() failed to get corrected data size in some condition. > > Please explain. E.g. avio_tell() fails to get the amount of data written

[FFmpeg-devel] [PATCH] ffmpeg_qsv: respect hwaccel_output_format setting

2020-10-21 Thread Haihao Xiang
Without this change, the output pixel format from qsv decoder is always 'qsv' regardless of the hwaccel_output_format setting, user will get errors when FFmpeg inserts an auto filter to convert 'qsv' format to the specified format Fox example: ffmpeg -loglevel verbose -hwaccel qsv -c:v h264_qsv -

[FFmpeg-devel] [RFC] support half-precision float in *_cuda filters

2020-10-21 Thread leozhang
Half precision float is supported for those devices sm capacity >=5.3. Theoretically, half float can achieve 2x performance speed up. So is there any plan to support half float in FFmpeg? Or any problem so that half float can't be met up? BR leozhang __

Re: [FFmpeg-devel] [PATCH 1/9] avformat/segafilm: Check that there is a stream

2020-10-21 Thread Andreas Rheinhardt
Michael Niedermayer: > Fixes: assertion failure > Fixes: > 26472/clusterfuzz-testcase-minimized-ffmpeg_dem_SEGAFILM_fuzzer-5759751591559168 > > Found-by: continuous fuzzing process > https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg > Signed-off-by: Michael Niedermayer > --- > lib

Re: [FFmpeg-devel] [PATCH 6/9] avformat/vividas: Check for zero v_size

2020-10-21 Thread Andreas Rheinhardt
Michael Niedermayer: > Fixes: SEGV on unknown address 0x > Fixes: > 26482/clusterfuzz-testcase-minimized-ffmpeg_dem_VIVIDAS_fuzzer-4905102324006912 > > Found-by: continuous fuzzing process > https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg > Signed-off-by: Michael Niede