avfilter: add sierpinski video source

avfilter/vf_premultiply: forward status back from output to all input links

avfilter/vf_fieldmatch: forward status back from output to all input links

Fixes #8199.

avfilter/vf_fieldmatch: fix more leaks

avfilter/vf_v360: add sinusoidal format

lavc/tiff: correct the default value of YCbCrSubsampling to 2x2

lavf/4xm: fix memory leak in error handing path

need to free the header in error path.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Jun Zhao <barryjzhao@tencent.com>

libavcodec/libx265: add a flag to output ROI warnings only once.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>

libavcodec/libx264: add a flag to output ROI warnings only once.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>

avcodec/utils: Optimize ff_color_frame() using memcpy()

4650975 -> 4493240 dezicycles

This optimizes lines 2 and later. Line 1 still uses av_memcpy_backptr()
This change originally fixed ossfuzz 10790 but this is now fixed by other
optimizations already

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

avcodec/aacdec: Check if we run out of input in read_stream_mux_config()

Fixes: Infinite loop
Fixes: 16920/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AAC_LATM_fuzzer-5653421289373696

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_slice: clear frame only on gaps when it is not otherwise initilaized

Fixes: Timeout (53sec -> 31sec)
Fixes: 16908/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-5711207859748864

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

avcodec/utils: Use av_memcpy_backptr() in ff_color_frame()

Fixes: Timeout (191sec -> 53sec)
Fixes: 16908/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-5711207859748864
Fixes: 10709/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-5630617975259136

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

avcodec/smacker: Fix integer overflow in signed int multiply in SMK_BLK_FILL

Fixes: signed integer overflow: 238 * 16843009 cannot be represented in type 'int'
Fixes: 16958/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SMACKER_fuzzer-5193905355620352

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

avcodec/alac: Fix invalid shifts in 20/24 bps

Fixes: left shift of negative value -256
Fixes: 16892/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ALAC_fuzzer-4880802642395136

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

avcodec/alac: fix undefined behavior with INT_MIN in lpc_prediction()

Fixes: signed integer overflow: -2147483648 * -1 cannot be represented in type 'int'
Fixes: 16786/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ALAC_fuzzer-5632818851348480

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

avformat/utils: Do not assume duration is non negative in compute_pkt_fields()

Several subtitle demuxers set negative durations

Fixes: signed integer overflow: 9223372036854775807 - -1 cannot be represented in type 'long'
Fixes: 16925/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5766519790764032

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

avcodec/ffwavesynth: Fix integer overflow in timestamps

Fixes: signed integer overflow: 9223371075321077760 * 2 cannot be represented in type 'long'
Fixes: 16447/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_FFWAVESYNTH_fuzzer-5698937431785472

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

avformat/vividas: Test size and packet numbers a bit more

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

avformat/vividas: Check n_sb_blocks against input space

Fixes: OOM
Fixes: 16726/clusterfuzz-testcase-minimized-ffmpeg_DEMUXER_fuzzer-5719320750981120

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

avcodec/aacdec_fixed: Add FF_CODEC_CAP_INIT_CLEANUP

Fixes: memleaks
Fixes: 16788/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AAC_FIXED_fuzzer-5649873898045440

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

tools/target_dec_fuzzer: consider potential padding/edge in pixel threshold

Fixes: Timeout (73sec ->30ms)
Fixes: 16921/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_DIRAC_fuzzer-5689384594046976

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

libavcodec/mpeg12dec: Check input for minimal frame size

Fixes: Timeout (35sec -> 6sec)
Fixes: 16901/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MPEGVIDEO_fuzzer-5729024726269952
Fixes: 16901/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MPEGVIDEO_fuzzer-5642388592132096

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

tools/target_dec_fuzzer: Adjust threshold for MSS2

The decoder is slow

Fixes: Timeout (94sec -> 7sec)
Fixes: 16417/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MSS2_fuzzer-5711668050395136

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

avcodec/cbs_h264: Automatically free SEI payload on error

If adding an SEI message to an access unit fails, said SEI message was
not touched, so that the caller had to free any data associated with it
that might need to be freed. But given that ff_cbs_h264_add_sei_message
can simply call cbs_h264_free_sei_payload, one can easily free
the content of the SEI payload.

This fixes a memleak when inserting a user data unregistered string for
h264_metadata fails.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>

avcodec/dxv: Check op_offset in both directions

Fixes: signed integer overflow: 61 + 2147483647 cannot be represented in type 'int'
Fixes: 15311/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_DXV_fuzzer-5742552826773504

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

avcodec/truespeech: Eliminate some left shifts

This avoids some invalid shifts

Reviewed-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/adpcm: Check number of channels for MTAF

Fixes: out of array access
Fixes: 17608/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ADPCM_MTAF_fuzzer-5074936267276288

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/nvenc: fix typo in new Windows driver version

Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>

avfilter/vf_v360: do not use mod where it is not needed

avcodec/exr: cosmetics

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/hashenc: fix incorrect use of av_mallocz_array()

Fixes CID 1453867, CID 1453866, CID 1453865.

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/nvenc: add driver version info for SDK 9.1

Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>

avcodec/bsf: check that AVBSFInternal was allocated before dereferencing it

This can happen when av_bsf_free() is called on av_bsf_alloc() failure.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

avfilter/vf_v360: fix regression introduced in b342678bc47b4e5

avformat/dashdec: fix segfault when parsing segmentlist

index into segmentlists_tab was specified as 4 instead of 3 causing invalid access

further fix to: 8135

Reviewed-by: Steven Liu <lq@onvideo.cn>
Signed-off-by: vectronic <hello.vectronic@gmail.com>

avformat/dashdec: fix pointer being freed was not allocated

prevent attempt to call xmlFree if val was not allocated
fixes: 8135

Reviewed-by: Steven Liu <lq@onvideo.cn>
Signed-off-by: vectronic <hello.vectronic@gmail.com>

avfilter/vf_v360: adjust h for mercator input/output

avfilter/vf_v360: fix xyz_to_mercator()

avfilter/vf_v360: fix mercator_to_xyz()

avfilter/v360: reduce size of some struct members

avformat/y4m: do not try to seek if pts is less than 0

Fixes #8193.

avcodec/sunrast: Fix indention

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

avcodec/sunrast: Fix return type for "unsupported (compression) type"

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/opt: Fix type specifier

This bug has been introduced in 9e0a071e.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/vf_v360: add fixed pixel padding options

avfilter/vf_v360: remove intermediate variables

tools/patcheck: remove gcc-2.95 compat check

this conflicts with the current contributing guidance:
http://ffmpeg.org/developer.html#Contributing
FFmpeg is programmed in the ISO C90 language with a few additional
features from ISO C99, namely:
...
for loops with variable definition (‘for (int i = 0; i < 8; i++)’);

Signed-off-by: James Zern <jzern@google.com>

avfilter/vf_v360: fix comparison with float value

avfilter/vf_v360: simplify xyz_to_dfisheye() calculations

avfilter/vf_v360: simplify dfisheye_to_xyz() calculations

avfilter/avf_showfreqs: fix check for failed allocation

avformat/latmenc: fix context used for logging

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

avfilter/vf_v360: simplify ball_to_xyz() calculations

avfilter/vf_v360: simplify xyz_to_ball() calculations

avfilter/vf_v360: add hammer projection

avfilter/vf_v360: use lrintf() in kernel

avfilter/vf_v360: fix small artifacts between corners in EAC format

avformat/hashenc: add streamhash muxer

Implemented as a variant of the hash muxer, reusing most functions,
and making use of the previously introduced array of hashes.

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/hashenc: use an array of hashes

Only the first element of the array is used currently, the other
elements are in preparation for a new muxer calculating multiple
hashes.

Also move alloc/init code from the write_header() functions to
dedicated init() functions, and the cleanup code from the
write_trailer() functions to dedicated deinit() functions.

hash_free() and framehash_free() turn out to be identical here,
but will differ in the subsequent commit, so they are not consolidated.

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/hashenc: rearrange options definition

Only the frame* muxers support the format_version option.
Use macros to ease the proliferation of identical options to
coming muxers as well.

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/encryption_info: Don't pass NULL to memcpy

The pointer arguments to memcpy (and several other functions of the
C standard library) are not allowed to be NULL, not even when the number
of bytes to copy is zero. An AVEncryptionInitInfo's data pointer is
explicitly allowed to be NULL and yet av_encryption_init_info_add_side_data
unconditionally used it as a source pointer to copy from. This commit changes
this so that copying is only done if the number of bytes to copy is > 0.

Fixes ticket #8141 as well as a part of ticket #8150.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

tools/target_dec_fuzzer: fuzz channels and sample_rate too

This should increase coverage over more audio decoders.

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

avcodec/utils: Check channels fully earlier

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

avformat/rtpenc_mpegts: copy metadata to mpegts sub-muxer

Fixes #7293.

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/aiffdec: parse replaygain metadata

Signed-off-by: Moritz Barsnick <barsnick@gmx.net>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/opt: Print out numeric values of option constants

It's often not obvious how option constants relate to numerical values.
Defaults are sometimes printed as numbers and from/to are always printed as numbers.
Printing the numeric values of options constants avoids this confusion.
It also allows to see which constants are equivalent.

Before this patch:

  -segment_list_type <int>        E........ set the segment list type (from -1 to 4) (default -1)
     flat                         E........ flat format
     csv                          E........ csv format
     ext                          E........ extended format
     ffconcat                     E........ ffconcat format
     m3u8                         E........ M3U8 format
     hls                          E........ Apple HTTP Live Streaming compatible

Afterwards:

  -segment_list_type <int>        E........ set the segment list type (from -1 to 4) (default -1)
     flat            0            E........ flat format
     csv             1            E........ csv format
     ext             3            E........ extended format
     ffconcat        4            E........ ffconcat format
     m3u8            2            E........ M3U8 format
     hls             2            E........ Apple HTTP Live Streaming compatible

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

avutil/opt: Fix printing integer option defaults

Integer values should not be printed using format specifier '%g' which leads to inexact display in case of higher values.

Before this patch:
  -trans_color       <int>        .D.V..... color value [...] (default 1.67772e+07)

Afterwards:
  -trans_color       <int>        .D.V..... color value [...] (default 16777215)

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

avcodec/ac3enc: Fix invalid shift

Fixes the FATE-tests unknown_layout-ac3, ac3-fixed-encode, ac3-encode
and eac3-encode. It furthermore fixes the ac3-encoder bugs mentioned in
tickets #7994, #8144 and #8159.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/g723_1dec: Fix invalid shift

Fixes the FATE-tests g723_1-dec-1, g723_1-dec-2 and g723_1-dec-4.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/tdsc: Fix undefined shifts

Fixes the tdsc FATE-test.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/vf_v360: remove uneeded array element

avfilter/vf_v360: add support for mirroring to ball input format

avfilter/vf_v360: adjust kernel after latest change

avfilter/vf_v360: fix various artifacts with cubemaps

avfilter/vf_v360: improve description about s in comments

libavfilter/dnn: support multiple outputs for native mode

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

libavfilter/dnn/dnn_backend_native: find the input operand according to input name

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

FATE/dnn: add unit test for layer maximum

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

libavfilter/dnn: add layer maximum for native mode.

The reason to add this layer is that it is used by srcnn in vf_sr.
This layer is currently ignored in native mode. After this patch,
we can add multiple outputs support for native mode.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

Delete unused branch in libaomenc

this branch is leftover from libvpxenc+vp8

Signed-off-by: James Zern <jzern@google.com>

avcodec/libvpxenc: add ROI-based encoding support for VP8/VP9

example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M tmp.webm

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: James Zern <jzern@google.com>

avformat/movenc: don't hardcode the colr box size

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

lavc/cook: Remove an incorrect comment.

Multichannel cook was implemented in 67da3182.

avformat/mpegts: add support for EPG extraction from mpegts

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

avcodec: add EPG codec ID

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

avformat/mpegts: add all well known PIDs and TIDs to mpegts.h

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

avfilter/dnn: fix inclusion guard in dnn/dnn_backend_native_layer_depth2space.h

Fixes fate-source failure.

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

avfilter/vf_v360: add ball projection format

FATE/dnn: add unit test for dnn depth_to_space layer

'make fate-dnn-layer-depth2space' to run the test

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

libavfilter/dnn: separate depth_to_space layer from dnn_backend_native.c to a new file

the logic is that one layer in one separated source file to make
the source files simple for maintaining.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

FATE/dnn: add unit test for dnn conv2d layer

'make fate-dnn-layer-conv2d' to run the test

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

libavfilter/dnn: separate conv2d layer from dnn_backend_native.c to a new file

the logic is that one layer in one separated source file to make
the source files simple for maintaining.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>

avcodec/v210enc: move v210_enc_8/10 function to template file

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/v210enc: make 8bit and 10bit process consistent

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/v210enc: add depth parameter for WRITE_PIXELS and CLIP

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/common: Fix undefined shift

av_mod_uintp2_c uses a bitwise AND with (1 << p) - 1 to clear the high
bits of an unsigned int. But this is undefined if p == 31, because 1 is
an int and 2^31 is not representable in an int. So make 1 unsigned.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/avpacket: Improve allocating packets

av_mallocz + av_init_packet leads to the same result as av_mallocz +
av_packet_unref, but faster.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mjpegdec: Unify switch statements

This has been forgotten in d5a3a20d.

Found via PVS-Studio (see ticket #8156).

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/utils: Preserve integrity of linked list

1. Instead of relying on ff_packet_list_get to get the oldest element in
an AVPacketList, ff_read_packet used its own ad-hoc code. Said code
forgot to set the end of the list to NULL if the last element of the
list has been removed, thereby leaving the list in an inconsistent state.
2. Furthermore, if the list was not empty, the oldest element of the
list would always be copied into another packet structure before it was
known whether the oldest entry of the list would be removed. This makes
the ownership confusing and potentially copies unnecessarily.

Both of these issues have been fixed. ff_packet_list_get is used now.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>