lavc/dpx: Reset n_datum for every new line when decoding 12bit.

Fixes odd unpacked 12bit decoding.
Reported-by: Piotr Bandurski

ffmpeg: assert that audio packet duration in process_input_packet() is non negative

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

fftools/ffmpeg: Fallback to duration if sample rate is unavailable

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>

avformat/mov: Only set pkt->duration to non negative values

Reviewed-by: Sasi Inguva <isasi@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/formats: get fflags values up-to-date

qt-faststart: add fate test for stco overflow

verify that the stco atom is upgraded to co64 when the addition of moov
size to the offsets results in an overflow

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

qt-faststart - stco offset bug fix

when the last offsets in the stco atom are close to 4GB, the addition of
the moov atom size can overflow, causing corruption near the end of the
mp4 file.
this patch upgrades all stco atoms to co64 when such an edge case is
detected. in order to accomplish this, the implementation was changed to
walk the atom tree, instead of searching for the strings 'stco'/'co64'.
this was required since when an stco atom is changed to co64, its size
changes, and the sizes of all containing atoms (moov, trak, etc.) have
to be updated as well.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

qt-faststart: add validation on ftyp atom size

avoid trying to allocate an unreasonably sized buffer on corrupt files

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mpeg4videodec: Clear bits_per_raw_sample if it has originated from a previous instance

Fixes: assertion failure
Fixes: ffmpeg_crash_5.avi
Found-by: Thuan Pham <thuanpv@comp.nus.edu.sg>, Marcel Böhme, Andrew Santosa and Alexandru Razvan Caciulescu with AFLSmart
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/devices: order alphabetically and add missing v4l2 outdev

Signed-off-by: Lou Logan <lou@lrcd.com>

avformat/movenc: fix recognization of cover image streams

For chapter images, the mov demux produces streams with disposition set
to attached_pic+timed_thumbnails. This patch fixes to properly recognize
streams that should be encoded as cover image (ones with only and only
attached_pic disposition set).

Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/movenc: properly handle cover image codecs

Find codec tag for attached images using appropriate list of
supported image formats.

This fixes writing the cover image to m4v/m4a and other container
formats that do not allow these codecs as a track.

Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/mov: Add check for per-sample IV size.

Found by Chrome's ClusterFuzz: http://crbug.com/849062.

Signed-off-by: Jacob Trimble <modmaker@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

configure: The eac3_core bitstream filter needs the ac3 parser.

Fixes linking with "--disable-everything --enable-bsf=eac3_core".

avformat: deprecate unused MP4A_LATM flag

A generic lavf flag for AAC LATM packetization for the RTP muxer was
added in ef409645f0 and then made inert 20 days later in 0832122880
when a private muxer option was added and the generic flag no longer
read.

avdevice/decklink_dec: capture timecode to metadata when requested

If the user provides a valid timecode_format look for timecode of that
format in the capture and if found store it on the video avstream's
metadata.

Slightly modified by Marton Balint to capture per-frame timecode as well.

Signed-off-by: Marton Balint <cus@passwd.hu>

avdevice/decklink_common: Move DECKLINK_* string functions into header

This allows other decklink source access to these cross-platform
convenience functions.

Signed-off-by: Marton Balint <cus@passwd.hu>

avdevice/decklink_dec: use std::atomic for decklink_input_callback refcounting

Also remove the callback from the context, and add proper error handling.

Signed-off-by: Marton Balint <cus@passwd.hu>

avdevice/decklink_dec: use a custom memory allocator

The default memory allocator is limited in the max number of frames available,
and therefore caused frame drops if the frames were not freed fast enough.

Signed-off-by: Marton Balint <cus@passwd.hu>

lavc/qsv: suppress code scan complain

Suppress the complain "variables 'type' is used but maybe uninitialized".

pixdesc: Only check against valid entries when iterating over lists of enums

Some of these enums have gaps in between their values, since they correspond
to the values in various specs, instead of being an incrementing list.

Fixes segfaults when, for example, using the valid API call:

   av_color_primaries_from_name("jecdec-p22");

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>

avcodec/h264_slice: Fix overflow in recovery_frame computation

Fixes: signed integer overflow: 15 + 2147483646 cannot be represented in type 'int'
Fixes: 8381/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-6225533137321984
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_ps: Move MAX_LOG2_MAX_FRAME_NUM to header so it can be used in h264_sei

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_mc_template: Only prefetch motion if the list is used.

Fixes: index 59 out of bounds for type 'H264Ref [48]'
Fixes: 8232/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-5703295145345024
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/xwddec: Use ff_set_dimensions()

Fixes: OOM
Fixes: 8178/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_XWD_fuzzer-4844793342459904
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/wavpack: Fix overflow in adding tail

Fixes: signed integer overflow: 2146907204 + 26846088 cannot be represented in type 'int'
Fixes: 8105/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_WAVPACK_fuzzer-6233036682166272
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/libzvbi-teletextdec: add support for selecting subtitle pages only

Signed-off-by: Marton Balint <cus@passwd.hu>

lsws/rgb2rgb_template: Do not compile unneeded shuffle functions on big-endian.

Fixes the following warnings:
In file included from libswscale/rgb2rgb.c:128:0:
libswscale/rgb2rgb_template.c:346:13: warning: 'shuffle_bytes_3210_c' defined but not used
libswscale/rgb2rgb_template.c:346:13: warning: 'shuffle_bytes_3012_c' defined but not used
libswscale/rgb2rgb_template.c:346:13: warning: 'shuffle_bytes_1230_c' defined but not used

lavc/aacsbr: fix make checkheaders warning

move the the function aacsbr_tableinit definition from header file
to .c file to fix make checkheaders warning.

Signed-off-by: Jun Zhao <mypopydev@gmail.com>

lavc/dolby_e: fix make checkheaders warning

move the the function init_tables() definitions from header file
to .c file to fix make checkheaders warning.

Signed-off-by: Jun Zhao <mypopydev@gmail.com>

ffmpeg: Fail if the user requested impossible subtitle encoding.

Fixes ticket #7239.

avformat/mov: Fix reading saio/saiz for clear content.

This validates that the common encryption saio/saiz atoms only appear
when the data is actually encrypted.  This also ignores those atoms
in clear content.

Found by Chrome's ClusterFuzz: http://crbug.com/850389

Signed-off-by: Jacob Trimble <modmaker@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/filters: change subtitles example to hint at correct color syntax

libass changed their color syntax in 2015 to TBGR

avformat/mxfdec: add support for recognizing timed text streams

Signed-off-by: Marton Balint <cus@passwd.hu>

avcodec/avcodec.h: add AV_CODEC_ID_TTML

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: remove check for NULL MXFTrack in mxf_set_pts

It cannot happen for video streams.

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: only disallow seek on metadata streams

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: recognize SMPTE 436 VBI data

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: use MXFCodecUL struct to store essence description for data_essence_container_uls

Signed-off-by: Marton Balint <cus@passwd.hu>

ffmpeg: mark sseof as an input-only option

libavfilter/dnn_interface.c: fix av_freep dnn_module.

Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

libavfilter/vf_srcnn.c: adds DNN module backend AVOption, changes AV_LOG_INFO message to AV_LOG_VERBOSE.

Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

MAINTAINERS: add myself to qsv section

Signed-off-by: Zhong Li <zhong.li@intel.com>

avcodec/shorten: Fix multiple integer overflows

Fixes: signed integer overflow: 3 * 1006632960 cannot be represented in type 'int'
Fixes: 8278/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SHORTEN_fuzzer-5692857166856192
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/shorten: Fix undefined shift in fix_bitshift()

Fixes: left shift of negative value -9
Fixes: 8571/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SHORTEN_fuzzer-5715966875926528
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/shorten: Fix a negative left shift in shorten_decode_frame()

Fixes: left shift of negative value -9057
Fixes: 8527/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SHORTEN_fuzzer-5666853924896768
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/shorten: Sanity check nmeans

Fixes: OOM
Fixes: 8195/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SHORTEN_fuzzer-5179785826271232
The reference software appears to use longs for 32bits and it uses int for nmeans
hinting that the intended maximum size was not 32bit.

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

configure: fix check for opencl

add pkg-config support for opencl check.

Signed-off-by: Jun Zhao <mypopydev@gmail.com>

lavfi/opencl: remove redundant header.

remove redundant header

Signed-off-by: Jun Zhao <mypopydev@gmail.com>

lavu/hwcontext_opecl: fix the build warning

fix the build warning when use Portable Computing Language (pocl).

Signed-off-by: Jun Zhao <mypopydev@gmail.com>

doc/general.texi: clarify FDK-AAC licensing & usability

Mnetion it can be used to decode as well.

lavu: add calling convention for OpenCL callback.

This fix a build error on Windows:
C2440: connot convert from 'void (__cdecl *) (...)' to 'void (__stdcall *)(...)'.

Signed-off-by: Ruiling Song <ruiling.song@intel.com>

lavc/profiles: Mention AAC-LC only once.

avcodec/shorten: Check non COMM chunk len before skip in decode_aiff_header()

Fixes: signed integer overflow: 2147483647 + 1 cannot be represented in type 'int'
Fixes: 8024/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SHORTEN_fuzzer-5109204648984576
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mjpegdec: Fix integer overflow in ljpeg_decode_rgb_scan()

Fixes: signed integer overflow: 32768 + 2147450880 cannot be represented in type 'int'
Fixes: 7885/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_THP_fuzzer-5298834394578944
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/truemotion2: Fix overflow in tm2_apply_deltas()

Fixes: signed integer overflow: 1077952576 + 1077952576 cannot be represented in type 'int'
Fixes: 7712/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_TRUEMOTION2_fuzzer-5056281753681920
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

configure: fix arm inline asm checks

Commit 8c893aa3cd5 removed quotes that were required to detect
inline asm in clang:

check_insn armv5te qadd r0, r0, r0
.../test.c:1:34: error: expected string literal in 'asm'
void foo(void){ __asm__ volatile(qadd r0, r0, r0); }

The correct code is:

void foo(void){ __asm__ volatile("qadd r0, r0, r0"); }

Commit message written by Frank Liberato <liberato@chromium.org>

Signed-off-by: James Almer <jamrial@gmail.com>

fate: add Dolby E test

Signed-off-by: James Almer <jamrial@gmail.com>

Use the same name for stereo3d frame/packet side data

lavf/mov.c: Set st->start_time for video streams explicitly.

If start_time is not set, ffmpeg takes the duration from the global
movie instead of the per stream duration.
Signed-off-by: Sasi Inguva <isasi@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

lavf/mov.c: Fix timestamps to be strictly monotonic for video also.

We already do this for audio, but it should be done for video too.
If we don't, seeking back to the start of the file, for example, can
become quite broken, since the first N packets will have repeating
and nonmonotonic PTS, yet they need to be decoded even if they are
to be discarded.

Signed-off-by: Sasi Inguva <isasi@isasi.mtv.corp.google.com>
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>

Adds TensorFlow backend for dnn inference module.

Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

avcodec/magicyuvenc: mark as not more experimental

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/magicyuvenc: write max huffman length and extradata too

Signed-off-by: Paul B Mahol <onemda@gmail.com>

doc/filters: mention required compile options for some filters

Signed-off-by: Lou Logan <lou@lrcd.com>

avcodec/opus_silk: Change silk_lsf2lpc() slightly toward silk/NLSF2A.c

Fixes: runtime error: signed integer overflow: -1440457022 - 785819492 cannot be represented in type 'int'
Fixes: 7700/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_OPUS_fuzzer-6595838684954624
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/amrwbdec: Fix division by 0 in find_hb_gain()

This restructures the code slightly toward D_UTIL_dec_synthesis()

Fixes: 7420/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AMRWB_fuzzer-6577305112543232
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

fftools/ffmpeg: fix for all forced key frames when 'copyts' is enabled

Forced key frames generation functionality was assuming the first PTS
value as zero, but, when 'copyts' is enabled, the first PTS can be any
big number. This was eventually forcing all the frames as key frames.
To resolve this issue, update has been made to use first input pts as
reference pts.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/vc1: fix out-of-bounds reference pixel replication

Out-of-bounds reference pixel replication should take into account the frame
coding mode of the reference frame(s), not the frame coding mode of the
current frame.

Signed-off-by: Jerome Borsboom <jerome.borsboom@carpalis.nl>

avfilter/drawtext: present 'hms' formatted 'pts' in 24h format

HMS is formatted as HH:MM:SS.mmm, but, HH part is not limited to
24 hours. For example, the the drawn text may look like this:
243029:20:30.342. To present the timestamp in more readable and
user friendly format, this patch provides an additional option
to limit the hour part in the range 0-23.

Note: Actually the above required format can be obtained with
format options 'localtime' and 'gmtime', but,  milliseconds part
is not supported in those formats.

v4l2_m2m: Mark V4L2 M2M decoders as unsuitable for probing

avfilter/f_drawgraph: fix drawing of first point for non-first metadata key

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/vf_zoompan: do not increase VAR_IN twice, also count from 0

Fixes #7242.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/avf_showspectrum: also show sample rate and channel layout

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/avf_showspectrum: avoid overwritting text

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/avf_showspectrum: improve axes drawing

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/vp3: Fix end of bitstream check in unpack_superblocks()

Fixes: regression
Found-by: Frank Liberato <liberato@google.com>
Tested-by: Frank Liberato <liberato@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/vf_lut3d: unbreak haldclut with planar rgb formats

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavfi/Makefile: move dnn_*.o under vf_srcnn

These files depend on libavformat, and the vf_srcnn filter
currently is the only thing utilizing these dnn_* files and
already happens to have a dependency on libavformat.

This fixes compilation in cases where libavformat is not a
dependency for libavfilter.

Reported by Kam_ on IRC.

doc/ffmpeg: rewrite Stream Selection chapter

Flesh out with details and examples to show quirks and limitations.

avfilter: add adeclick and adeclip audio filters

Signed-off-by: Paul B Mahol <onemda@gmail.com>

libavformat/mov: Fix heap buffer overflow.

Found by Chrome's ClusterFuzz: https://crbug.com/847060

Signed-off-by: Jacob Trimble <modmaker@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

lavc/libx265: allow users to set closed GOP via generic lavc flag

lavc flag 'cgop' can be used to set closed GOP.

Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>

avcodec/dpx: Support for RGBA 12-bit packed decoding

Limited to widths multiple of 2 due to lack of test files for such corner case

This partially fixes ticket #5639

avcodec/dnxhdenc: do not free nonexistent slice threads data when frame threading is used

Forgotten in 28e9ba951d1a0b0aca53b242aa64f484ca75e874.
Fixes #7241.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

doc/ffmpeg: remove non-existent vdt option

avfilter: use av_clip_uintp2_c where clip is variable

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mxfdec: add some missing picture_essence_container_uls

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: use int math for sample count calculation

This also changes the rounding to nearest.

Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mxfdec: fix klv_decode_ber_length return value usage

Signed-off-by: Marton Balint <cus@passwd.hu>

tools/crypto_bench: add missing RC4 implementation from tomcrypt

Signed-off-by: James Almer <jamrial@gmail.com>

tools/crypto_bench: add missing RC4 implementation from gcrypt

Signed-off-by: James Almer <jamrial@gmail.com>

tools/crypto_bench: add support for mbedcrypto

Requires mbed TLS 2.7.0 or newer

Signed-off-by: James Almer <jamrial@gmail.com>

avcodec/h263dec: Reinitialize idct context if it has not been setup for the active profile

The profile after reading headers can be different from when the context was initialized

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/idctdsp: Clear idct/idct_add for studio profile

This does not leave them "as before" which may be a value from a previous profile

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/idctdsp: Transmit studio_profile to init instead of using AVCodecContext profile

These 2 fields are not always the same, it is simpler to always use the same field
for detecting studio profile

Fixes: null pointer dereference
Fixes: ffmpeg_crash_3.avi
Found-by: Thuan Pham <thuanpv@comp.nus.edu.sg>, Marcel Böhme, Andrew Santosa and Alexandru RazvanCaciulescu with AFLSmart
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/af_aiir: draw IR frequency response

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/af_afir: fix picking of IR channel

Signed-off-by: Paul B Mahol <onemda@gmail.com>

flvenc: Fix sequence header update timestamps

flvenc: Factorize timestamp writing

The code is trivial but the semantics in the spec are ambiguous. This
should help keep parts of the muxer interpreting them consistently.