ffmpeg.git
2 years agoMerge commit 'e723dce6f8ba1e8260433b6ecfe5a3262f4c7a99'
James Almer [Thu, 16 Mar 2017 00:35:36 +0000 (21:35 -0300)]
Merge commit 'e723dce6f8ba1e8260433b6ecfe5a3262f4c7a99'

* commit 'e723dce6f8ba1e8260433b6ecfe5a3262f4c7a99':
  dvbsubdec: Use NULL instead of 0 as pointer value

This commit is a noop. The affected code isn't in our tree.

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit 'fc94a1acc27ab7296edce3fa81ef36691af5c134'
James Almer [Thu, 16 Mar 2017 00:33:59 +0000 (21:33 -0300)]
Merge commit 'fc94a1acc27ab7296edce3fa81ef36691af5c134'

* commit 'fc94a1acc27ab7296edce3fa81ef36691af5c134':
  Revert "libavutil: Use an intermediate variable in AV_COPY*U"

This commit is a noop.

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit '9806b9ab5c7fb2ac5efd8ffa8713fea0c5fd218d'
James Almer [Thu, 16 Mar 2017 00:33:13 +0000 (21:33 -0300)]
Merge commit '9806b9ab5c7fb2ac5efd8ffa8713fea0c5fd218d'

* commit '9806b9ab5c7fb2ac5efd8ffa8713fea0c5fd218d':
  Revert "Don't use expressions with side effects in macro parameters"

This commit is a noop.

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit 'f79d847400d218cfd0b95f10358fe6e65ec3c9c4'
James Almer [Thu, 16 Mar 2017 00:20:44 +0000 (21:20 -0300)]
Merge commit 'f79d847400d218cfd0b95f10358fe6e65ec3c9c4'

* commit 'f79d847400d218cfd0b95f10358fe6e65ec3c9c4':
  intreadwrite: Use the __unaligned keyword on MSVC for ARM and x86_64

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit '230b1c070baa3b6d4bd590426a365b843d60ff50'
James Almer [Thu, 16 Mar 2017 00:17:37 +0000 (21:17 -0300)]
Merge commit '230b1c070baa3b6d4bd590426a365b843d60ff50'

* commit '230b1c070baa3b6d4bd590426a365b843d60ff50':
  intreadwrite: Add intermediate variables in the byteswise AV_W*() macros

Mostly a noop. Merged for cosmetic purposes.
See d83ff76ca0e1ec905ca2d5b27bd0973991eadbd9

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit '014773b66bdff4de24f384066d1a85d2a5bb6774'
James Almer [Thu, 16 Mar 2017 00:14:51 +0000 (21:14 -0300)]
Merge commit '014773b66bdff4de24f384066d1a85d2a5bb6774'

* commit '014773b66bdff4de24f384066d1a85d2a5bb6774':
  libavutil: Use an intermediate variable in AV_COPY*U

This commit is a noop. It would be reverted in a future merge either
way.

Merged-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit '25bacd0a0c32ae682e6f411b1ac9020aeaabca72'
James Almer [Thu, 16 Mar 2017 00:13:54 +0000 (21:13 -0300)]
Merge commit '25bacd0a0c32ae682e6f411b1ac9020aeaabca72'

* commit '25bacd0a0c32ae682e6f411b1ac9020aeaabca72':
  Don't use expressions with side effects in macro parameters

This commit is a noop. It would be reverted in a future merge either
way.

Merged-by: James Almer <jamrial@gmail.com>
2 years agoavcodec/tiff: Check palette shift
Michael Niedermayer [Tue, 14 Mar 2017 00:34:14 +0000 (01:34 +0100)]
avcodec/tiff: Check palette shift

Fixes multiple  runtime error: shift exponent 792 is too large for 32-bit type 'unsigned int'

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/wavpack: Fix multiple integer overflows
Michael Niedermayer [Tue, 14 Mar 2017 00:17:53 +0000 (01:17 +0100)]
avcodec/wavpack: Fix multiple integer overflows

Fixes: 839/clusterfuzz-testcase-4871084446842880

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/simple_idct_template: Fix several integer overflows
Michael Niedermayer [Mon, 13 Mar 2017 01:25:42 +0000 (02:25 +0100)]
avcodec/simple_idct_template: Fix several integer overflows

Benchmarks with START_TIMER indicate that the code is faster with unsigned, (that is
with the patch), there was quite some fluctuation in the numbers so this may be just
random

Fixes: 811/clusterfuzz-testcase-6465493076541440

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoconfigure: libnpp is always nonfree, even with LGPL
Ricardo Constantino [Wed, 15 Mar 2017 22:47:58 +0000 (22:47 +0000)]
configure: libnpp is always nonfree, even with LGPL

libnpp was erroneously grouped up with libfdk-aac and openssl to check
if --enable-nonfree wasn't passed only with --enable-gpl in
9f28db47accb31bfec40a56dd2dc19ffd366a6be. The latter two are compatible
with LGPL, libnpp is not.

Signed-off-by: James Almer <jamrial@gmail.com>
2 years agoMerge commit '7ebdffc353f3f0827864e8e3461fdc00cc243b14'
Clément Bœsch [Wed, 15 Mar 2017 22:26:10 +0000 (23:26 +0100)]
Merge commit '7ebdffc353f3f0827864e8e3461fdc00cc243b14'

* commit '7ebdffc353f3f0827864e8e3461fdc00cc243b14':
  dxv: Check to make sure we don't overrun buffers on corrupt inputs

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'e328178da90f44690e0076f4dbfd16da9175f441'
Clément Bœsch [Wed, 15 Mar 2017 22:22:50 +0000 (23:22 +0100)]
Merge commit 'e328178da90f44690e0076f4dbfd16da9175f441'

* commit 'e328178da90f44690e0076f4dbfd16da9175f441':
  qsvdec: only access hwaccel_context is the pixel format is QSV

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'ed1cd81076434b76f37576d4d806973476a8e96c'
Clément Bœsch [Wed, 15 Mar 2017 22:16:23 +0000 (23:16 +0100)]
Merge commit 'ed1cd81076434b76f37576d4d806973476a8e96c'

* commit 'ed1cd81076434b76f37576d4d806973476a8e96c':
  flac demuxer: improve probing

Suggested commit very closely matches our code, except with regards to
AVPROBE_SCORE_EXTENSION. The code layout is mostly merged but preserves
our behaviour.

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '5ebef79abecc3ffcc4ab0d46e203d13b068107c9'
Clément Bœsch [Wed, 15 Mar 2017 22:00:50 +0000 (23:00 +0100)]
Merge commit '5ebef79abecc3ffcc4ab0d46e203d13b068107c9'

* commit '5ebef79abecc3ffcc4ab0d46e203d13b068107c9':
  Fix instances of broken indentation found by gcc 6

Noop, see 21d3f0c026089c44a2

Merged-by: Clément Bœsch <u@pkh.me>
2 years agodoc: remove remaining legacy x11grab references
Clément Bœsch [Wed, 15 Mar 2017 21:23:00 +0000 (22:23 +0100)]
doc: remove remaining legacy x11grab references

2 years agoMerge commit '5ed4644d6de7f6112431dc2d9a5cfe9a0a75a688'
Clément Bœsch [Wed, 15 Mar 2017 21:17:29 +0000 (22:17 +0100)]
Merge commit '5ed4644d6de7f6112431dc2d9a5cfe9a0a75a688'

* commit '5ed4644d6de7f6112431dc2d9a5cfe9a0a75a688':
  x11grab: Rename internal component to "xcbgrab"

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '4fef648d10bf3bcfd4b8fa5755c1128966a2427c'
Clément Bœsch [Wed, 15 Mar 2017 21:07:47 +0000 (22:07 +0100)]
Merge commit '4fef648d10bf3bcfd4b8fa5755c1128966a2427c'

* commit '4fef648d10bf3bcfd4b8fa5755c1128966a2427c':
  Remove the legacy X11 screen grabber

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '8495d84f0101464b15517860db33e8605586d87e'
Clément Bœsch [Wed, 15 Mar 2017 20:58:13 +0000 (21:58 +0100)]
Merge commit '8495d84f0101464b15517860db33e8605586d87e'

* commit '8495d84f0101464b15517860db33e8605586d87e':
  ac3dec: Add some inline hints

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '3db51bf671defd47f2ec5ab67b11fb7730fb5e5a'
Clément Bœsch [Wed, 15 Mar 2017 20:57:57 +0000 (21:57 +0100)]
Merge commit '3db51bf671defd47f2ec5ab67b11fb7730fb5e5a'

* commit '3db51bf671defd47f2ec5ab67b11fb7730fb5e5a':
  ac3dec: Simplify skipping

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'f0ccc65bc9ab9ddf1366066395564c71bcc825ee'
Clément Bœsch [Wed, 15 Mar 2017 20:56:47 +0000 (21:56 +0100)]
Merge commit 'f0ccc65bc9ab9ddf1366066395564c71bcc825ee'

* commit 'f0ccc65bc9ab9ddf1366066395564c71bcc825ee':
  ac3dec: Split coupling-specific code from decode_audio_block()

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'a5ebe5d1217942238c641c83b24ef1106e53934a'
Clément Bœsch [Wed, 15 Mar 2017 20:43:59 +0000 (21:43 +0100)]
Merge commit 'a5ebe5d1217942238c641c83b24ef1106e53934a'

* commit 'a5ebe5d1217942238c641c83b24ef1106e53934a':
  ac3dec: Split spx-specific code from decode_audio_block()

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '2ac00d2d1d51047c6ce69d5fbe1a08392d142658'
Clément Bœsch [Wed, 15 Mar 2017 20:29:56 +0000 (21:29 +0100)]
Merge commit '2ac00d2d1d51047c6ce69d5fbe1a08392d142658'

* commit '2ac00d2d1d51047c6ce69d5fbe1a08392d142658':
  mov: Validate the ID number

This commit is a noop as the modified check is not present in FFmpeg.
See d30870cc730337b7017cc194d696f68ee48f7f1d.

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'a115eb9e750543f1d8bf951414d291069bf396c2'
Clément Bœsch [Wed, 15 Mar 2017 20:23:30 +0000 (21:23 +0100)]
Merge commit 'a115eb9e750543f1d8bf951414d291069bf396c2'

* commit 'a115eb9e750543f1d8bf951414d291069bf396c2':
  mimic: do not release the newly obsolete reference at the end of decoding

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'ae90119c6701fa09ff747cca35238e36b2d2ab2f'
Clément Bœsch [Wed, 15 Mar 2017 20:09:06 +0000 (21:09 +0100)]
Merge commit 'ae90119c6701fa09ff747cca35238e36b2d2ab2f'

* commit 'ae90119c6701fa09ff747cca35238e36b2d2ab2f':
  configure: Simplify license incompatibility check

An extra GPLV3 list is added for libsmbclient as having it in both GPL
and VERSION3 lists would cause a duplicate in the final config list.

Also, for consistency, libnpp is treated the same as the other nonfree
component (libfdk_aac and openssl).

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoavcodec/videotoolboxenc: add rc_max_bitrate control into videotoolbox
Steven Liu [Tue, 14 Mar 2017 14:53:43 +0000 (22:53 +0800)]
avcodec/videotoolboxenc: add rc_max_bitrate control into videotoolbox

add kVTCompressionPropertyKey_DataRateLimits support by rc_max_bitrate

Reviewed-by: Rick Kern <kernrj@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoMerge commit 'e46a6fb7732a7caef97a916a4f765ec0f779d195'
Clément Bœsch [Wed, 15 Mar 2017 13:25:46 +0000 (14:25 +0100)]
Merge commit 'e46a6fb7732a7caef97a916a4f765ec0f779d195'

* commit 'e46a6fb7732a7caef97a916a4f765ec0f779d195':
  avconv: Check that muxing_queue exists before reading from it

Mostly noop. This was fixed in FFmpeg in 7f7c494a3.

The merge makes the cosmetics match but does not include the weird
av_log().

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit '100fb0ddfda958da70f98feac81f924c02483789'
Clément Bœsch [Wed, 15 Mar 2017 11:49:03 +0000 (12:49 +0100)]
Merge commit '100fb0ddfda958da70f98feac81f924c02483789'

* commit '100fb0ddfda958da70f98feac81f924c02483789':
  configure: Allow detecting and using LLVM lld-link as linker for windows

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit '79fb0692992c74214c6cf8e81350fc93eeffc5ec'
Clément Bœsch [Wed, 15 Mar 2017 11:44:35 +0000 (12:44 +0100)]
Merge commit '79fb0692992c74214c6cf8e81350fc93eeffc5ec'

* commit '79fb0692992c74214c6cf8e81350fc93eeffc5ec':
  configure: Move defines for controlling MSVCRT headers to the CRT detection section

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit '56af0bc10f49654b5b5f3efe82c69a13bf15fc8b'
Clément Bœsch [Wed, 15 Mar 2017 11:39:34 +0000 (12:39 +0100)]
Merge commit '56af0bc10f49654b5b5f3efe82c69a13bf15fc8b'

* commit '56af0bc10f49654b5b5f3efe82c69a13bf15fc8b':
  configure: Check for strtoll and redirect to _strtoi64 in the msvcrt block

Also includes _strtoui64 in the check.

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit '48b80f8393d418ad35d73f5a36f5011de1928f3c'
Clément Bœsch [Wed, 15 Mar 2017 11:34:51 +0000 (12:34 +0100)]
Merge commit '48b80f8393d418ad35d73f5a36f5011de1928f3c'

* commit '48b80f8393d418ad35d73f5a36f5011de1928f3c':
  hpeldsp: Explain why put_no_rnd_pixels_tab is larger than necessary

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit '40ad05bab206c932a32171d45581080c914b06ec'
Clément Bœsch [Wed, 15 Mar 2017 11:32:15 +0000 (12:32 +0100)]
Merge commit '40ad05bab206c932a32171d45581080c914b06ec'

* commit '40ad05bab206c932a32171d45581080c914b06ec':
  checkasm: Cast unsigned to signed

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agoMerge commit 'b183abfb5b6366b177cf44f244c66156257a6fd6'
Clément Bœsch [Wed, 15 Mar 2017 11:23:18 +0000 (12:23 +0100)]
Merge commit 'b183abfb5b6366b177cf44f244c66156257a6fd6'

* commit 'b183abfb5b6366b177cf44f244c66156257a6fd6':
  vpx: Support color range

Decoder chunk not merged as the framework automatically copies avctx
color range to the frame color range. And we already set the avctx field
since cbcc88c039.

Merged-by: Clément Bœsch <cboesch@gopro.com>
2 years agolavc/h264dec: use OFFSET macro
Matthieu Bouron [Mon, 6 Feb 2017 16:14:57 +0000 (17:14 +0100)]
lavc/h264dec: use OFFSET macro

2 years agoavformat/hlsenc: fix duration wrong when no pkt duration
Steven Liu [Tue, 14 Mar 2017 23:37:11 +0000 (07:37 +0800)]
avformat/hlsenc: fix duration wrong when no pkt duration

when cannot get pkt duration, hlsenc segments duration will
be set to 0, this patch can fix it.

Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoffmpeg, ffprobe: don't "merge" side data into packet data by default
wm4 [Thu, 9 Mar 2017 14:43:30 +0000 (15:43 +0100)]
ffmpeg, ffprobe: don't "merge" side data into packet data by default

Preparation for potentially disabling merged side data by default in the
libs. Do this in particular because it affects fate tests.

The changed tests either reflect added packet side data, or the changed
packet size due to merged side data removal reducing the packet size.

2 years agolavf/avio: Be more explicit in logging white/black list matches
Alexander Strasser [Sat, 11 Mar 2017 14:43:39 +0000 (15:43 +0100)]
lavf/avio: Be more explicit in logging white/black list matches

The current form of the messages indicating matches in the white
or black lists seems to be a bit too much relying on context.

Make the messages more explicit.

Signed-off-by: Alexander Strasser <eclipse7@gmx.net>
2 years agolavf/avio: Remove unnecessary escaping of ' in string literals
Alexander Strasser [Sat, 11 Mar 2017 11:02:32 +0000 (12:02 +0100)]
lavf/avio: Remove unnecessary escaping of ' in string literals

Signed-off-by: Alexander Strasser <eclipse7@gmx.net>
2 years agoavcodec/mpegaudiodec_template: Make l3_unscale() work with e=0
Michael Niedermayer [Mon, 13 Mar 2017 19:45:09 +0000 (20:45 +0100)]
avcodec/mpegaudiodec_template: Make l3_unscale() work with e=0

Fixes undefined behavior
Fixes: 830/clusterfuzz-testcase-6253175327686656

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/tiff: Check for multiple geo key directories
Michael Niedermayer [Mon, 13 Mar 2017 19:45:08 +0000 (20:45 +0100)]
avcodec/tiff: Check for multiple geo key directories

Fixes memleak
Fixes: 826/clusterfuzz-testcase-5316921379520512

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/pictordec: Fix runtime error: left shift of 805306368 by 2 places cannot...
Michael Niedermayer [Mon, 13 Mar 2017 19:45:07 +0000 (20:45 +0100)]
avcodec/pictordec: Fix runtime error: left shift of 805306368 by 2 places cannot be represented in type 'int'

Fixes: 823/clusterfuzz-testcase-6727060074528768

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agodoc/filters: Add colourspace values for colormatrix filter
Katherine Nagels [Sun, 12 Mar 2017 22:57:11 +0000 (11:57 +1300)]
doc/filters: Add colourspace values for colormatrix filter

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Kieran Kunhya <kierank@obe.tv>
2 years agoavformat/flvenc: flx flvflags no_metadata bug
Steven Liu [Tue, 14 Mar 2017 10:11:20 +0000 (18:11 +0800)]
avformat/flvenc: flx flvflags no_metadata bug

When use flvflags no_metadata , the FLV header will be cover by write tailer
This commit fix the bug

Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoavformat/hlsenc: fix ticket 6231
Steven Liu [Tue, 14 Mar 2017 02:16:00 +0000 (10:16 +0800)]
avformat/hlsenc: fix ticket 6231

check if the hls_flags is byterange_mode and check if should close fd

Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoavcodec/wavpack: Fix runtime error: shift exponent 32 is too large for 32-bit type...
Michael Niedermayer [Mon, 13 Mar 2017 01:51:15 +0000 (02:51 +0100)]
avcodec/wavpack: Fix runtime error: shift exponent 32 is too large for 32-bit type 'int'

Fixes: 822/clusterfuzz-testcase-4873433189974016

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/targa: Skip hflip on blank images
Michael Niedermayer [Sun, 12 Mar 2017 23:18:04 +0000 (00:18 +0100)]
avcodec/targa: Skip hflip on blank images

Fixes: timeout with 810/clusterfuzz-testcase-5249282825256960

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/wavpack: Fix runtime error: shift exponent 137 is too large for 32-bit type...
Michael Niedermayer [Sun, 12 Mar 2017 22:45:54 +0000 (23:45 +0100)]
avcodec/wavpack: Fix runtime error: shift exponent 137 is too large for 32-bit type 'int'

Fixes: 808/clusterfuzz-testcase-4715513349406720

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/h264_direct: Fix runtime error: signed integer overflow: 2147483647 - -14133...
Michael Niedermayer [Fri, 10 Mar 2017 14:24:52 +0000 (15:24 +0100)]
avcodec/h264_direct: Fix runtime error: signed integer overflow: 2147483647 - -14133 cannot be represented in type 'int'

Fixes: 755/clusterfuzz-testcase-5369072516595712

See: [FFmpeg-devel] [PATCH 1/2] avcodec/h264_direct: Fix runtime error: signed integer overflow: 2147483647 - -14133 cannot be represented in type 'int'

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/xpmdec: there are XPM files with dos line endings
Paul B Mahol [Sun, 12 Mar 2017 22:13:26 +0000 (23:13 +0100)]
avcodec/xpmdec: there are XPM files with dos line endings

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavfilter/vf_lut3d: actually skip lines when encountering DOMAIN_ string
Paul B Mahol [Sun, 12 Mar 2017 17:59:05 +0000 (18:59 +0100)]
avfilter/vf_lut3d: actually skip lines when encountering DOMAIN_ string

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavformat/hlsenc: second_levels flags process function extract
Steven Liu [Mon, 13 Mar 2017 03:58:34 +0000 (11:58 +0800)]
avformat/hlsenc: second_levels flags process function extract

the SECOND_LEVEL* flags process and name is too long
extract all of them output to funtions, make code clear

Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoavcodec/vp6: clear dimensions on failed resolution change in vp6_parse_header()
Michael Niedermayer [Sun, 12 Mar 2017 02:04:06 +0000 (03:04 +0100)]
avcodec/vp6: clear dimensions on failed resolution change in vp6_parse_header()

Fixes: 807/clusterfuzz-testcase-6470061042696192
Fixes null pointer dereference

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/rv34: Fix runtime error: signed integer overflow: 36880 * 66288 cannot be...
Michael Niedermayer [Sun, 12 Mar 2017 02:04:05 +0000 (03:04 +0100)]
avcodec/rv34: Fix runtime error: signed integer overflow: 36880 * 66288 cannot be represented in type 'int'

Fixes: 768/clusterfuzz-testcase-4807444305805312

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/mpeg12dec: Fix runtime error: left shift of negative value -1
Michael Niedermayer [Sun, 12 Mar 2017 02:04:04 +0000 (03:04 +0100)]
avcodec/mpeg12dec: Fix runtime error: left shift of negative value -1

Fixes: 764/clusterfuzz-testcase-6273034652483584

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/xpmdec: avoid "magic" numbers in function hex_char_to_number()
Paul B Mahol [Sun, 12 Mar 2017 21:58:49 +0000 (22:58 +0100)]
avcodec/xpmdec: avoid "magic" numbers in function hex_char_to_number()

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: fix colors values which are different for X11
Paul B Mahol [Sun, 12 Mar 2017 21:56:58 +0000 (22:56 +0100)]
avcodec/xpmdec: fix colors values which are different for X11

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: do not allow number of colors to be higher than allocated
Paul B Mahol [Sun, 12 Mar 2017 21:51:00 +0000 (22:51 +0100)]
avcodec/xpmdec: do not allow number of colors to be higher than allocated

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: rename yet another function
Paul B Mahol [Sun, 12 Mar 2017 21:43:49 +0000 (22:43 +0100)]
avcodec/xpmdec: rename yet another function

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: rename convert to hex_char_to_number
Paul B Mahol [Sun, 12 Mar 2017 21:42:05 +0000 (22:42 +0100)]
avcodec/xpmdec: rename convert to hex_char_to_number

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: improve comment for one function and also fix 2 identation issues
Paul B Mahol [Sun, 12 Mar 2017 21:37:53 +0000 (22:37 +0100)]
avcodec/xpmdec: improve comment for one function and also fix 2 identation issues

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoswresample/swresample: do not reset tsf on swr_alloc_set_opts
Muhammad Faiz [Sat, 11 Mar 2017 17:24:21 +0000 (00:24 +0700)]
swresample/swresample: do not reset tsf on swr_alloc_set_opts

so tsf option in aresample will have effect
previously tsf/internal_sample_format had no effect

fate is updated
s32p previously used fltp internally
dblp previously used fltp/dblp internally

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
2 years agovf_hwupload: Add missing return value check
Jun Zhao [Fri, 3 Mar 2017 01:25:53 +0000 (09:25 +0800)]
vf_hwupload: Add missing return value check

Add missing return value checks to suppress build warning and
remove noop ff_formats_unref() calling.

Note: most filters using ff_formats_ref() didn't have a suitable
error handling, it's a potential memory leak issue.

Signed-off-by: Jun Zhao <jun.zhao@intel.com>
Signed-off-by: Mark Thompson <sw@jkqxz.net>
2 years agotests/api-seek: fix memory leak on realloc() failure
James Almer [Sat, 11 Mar 2017 23:29:40 +0000 (20:29 -0300)]
tests/api-seek: fix memory leak on realloc() failure

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
2 years agotest/api-seek: clean up properly on failure
James Almer [Sun, 12 Mar 2017 13:36:44 +0000 (10:36 -0300)]
test/api-seek: clean up properly on failure

Also propagate better error values.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
2 years agotests/api-seek: make the crc array uint32_t
James Almer [Sat, 11 Mar 2017 23:24:53 +0000 (20:24 -0300)]
tests/api-seek: make the crc array uint32_t

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
2 years agoavcodec/xpmdec: skip everything before signature
Paul B Mahol [Sun, 12 Mar 2017 14:07:02 +0000 (15:07 +0100)]
avcodec/xpmdec: skip everything before signature

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/codec_desc: add mime type for X-BitMap
Paul B Mahol [Sun, 12 Mar 2017 13:50:33 +0000 (14:50 +0100)]
avcodec/codec_desc: add mime type for X-BitMap

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/codec_desc: add mime type for X-PixMap
Paul B Mahol [Sun, 12 Mar 2017 13:49:26 +0000 (14:49 +0100)]
avcodec/codec_desc: add mime type for X-PixMap

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/xpmdec: make convert function more picky about its input
Paul B Mahol [Sun, 12 Mar 2017 11:48:01 +0000 (12:48 +0100)]
avcodec/xpmdec: make convert function more picky about its input

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec: add XPM decoder and demuxer
Paras Chadha [Sat, 11 Mar 2017 21:01:23 +0000 (02:31 +0530)]
avcodec: add XPM decoder and demuxer

Signed-off-by: Paras Chadha <paraschadha18@gmail.com>
2 years agoMerge commit 'ad71d3276fef0ee7e791e62bbfe9c4e540047417'
Mark Thompson [Sun, 12 Mar 2017 16:55:32 +0000 (16:55 +0000)]
Merge commit 'ad71d3276fef0ee7e791e62bbfe9c4e540047417'

* commit 'ad71d3276fef0ee7e791e62bbfe9c4e540047417':
  lavfi: add a QSV deinterlacing filter

Minor fixup for lavfi differences.

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit '8e07c22e508b349d145b9f142aa3ee8b3ce1d3a4'
Mark Thompson [Sun, 12 Mar 2017 15:21:41 +0000 (15:21 +0000)]
Merge commit '8e07c22e508b349d145b9f142aa3ee8b3ce1d3a4'

* commit '8e07c22e508b349d145b9f142aa3ee8b3ce1d3a4':
  qsvenc: print warnings from encode/init

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit '0956fd460681e8ccbdae19f135f0d3970bf95c2f'
Mark Thompson [Sun, 12 Mar 2017 15:19:52 +0000 (15:19 +0000)]
Merge commit '0956fd460681e8ccbdae19f135f0d3970bf95c2f'

* commit '0956fd460681e8ccbdae19f135f0d3970bf95c2f':
  qsvenc: do not re-execute encoding on all positive status codes

Noop, see fb240a6276fa36fe120aadd67b4ca774e354f22b.

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit '95414eb2dc63a6f934275b4ed33dedd4369f2c49'
Mark Thompson [Sun, 12 Mar 2017 15:19:05 +0000 (15:19 +0000)]
Merge commit '95414eb2dc63a6f934275b4ed33dedd4369f2c49'

* commit '95414eb2dc63a6f934275b4ed33dedd4369f2c49':
  qsv: print more complete error messages

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit 'd9ec3c60143babe1bb77c268e1d5547d15acd69b'
Mark Thompson [Sun, 12 Mar 2017 15:06:07 +0000 (15:06 +0000)]
Merge commit 'd9ec3c60143babe1bb77c268e1d5547d15acd69b'

* commit 'd9ec3c60143babe1bb77c268e1d5547d15acd69b':
  qsvenc: take only the allocated dimensions from the frames context

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit '37a9015ee84c15fec5247ba8f6577351a25fa8d2'
Mark Thompson [Sun, 12 Mar 2017 15:04:45 +0000 (15:04 +0000)]
Merge commit '37a9015ee84c15fec5247ba8f6577351a25fa8d2'

* commit '37a9015ee84c15fec5247ba8f6577351a25fa8d2':
  qsvenc: add support for p010

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agolavfi: add a QSV scaling filter
Anton Khirnov [Sat, 26 Mar 2016 11:39:58 +0000 (12:39 +0100)]
lavfi: add a QSV scaling filter

This merges libav commit ac7bfd69678f3966e38debdb27f4bde94dc0345c,
which was previously skipped.

(cherry picked from commit ac7bfd69678f3966e38debdb27f4bde94dc0345c)
Signed-off-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit '21962261c74aed4df00ae8348a5e2d1ecb67c52d'
Mark Thompson [Sun, 12 Mar 2017 14:21:37 +0000 (14:21 +0000)]
Merge commit '21962261c74aed4df00ae8348a5e2d1ecb67c52d'

* commit '21962261c74aed4df00ae8348a5e2d1ecb67c52d':
  qsv: handle the semi-packed formats in map_fourcc as well

Merged-by: Mark Thompson <sw@jkqxz.net>
2 years agoMerge commit 'f65285aba0df7d46298abe0c945dfee05cbc6028'
Clément Bœsch [Sun, 12 Mar 2017 12:21:01 +0000 (13:21 +0100)]
Merge commit 'f65285aba0df7d46298abe0c945dfee05cbc6028'

* commit 'f65285aba0df7d46298abe0c945dfee05cbc6028':
  lavc: set sw_pix_fmt for hwaccel encoding

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'd59641abfd25a1007bdf4723d952887b1e3619c6'
Clément Bœsch [Sun, 12 Mar 2017 12:20:57 +0000 (13:20 +0100)]
Merge commit 'd59641abfd25a1007bdf4723d952887b1e3619c6'

* commit 'd59641abfd25a1007bdf4723d952887b1e3619c6':
  lavc: initialize AVCodecContext.sw_pix_fmt properly

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '8b7a9729aa162e2bbd571933f1aa40767f1ff47b'
Clément Bœsch [Sun, 12 Mar 2017 12:13:55 +0000 (13:13 +0100)]
Merge commit '8b7a9729aa162e2bbd571933f1aa40767f1ff47b'

* commit '8b7a9729aa162e2bbd571933f1aa40767f1ff47b':
  avconv_qsv: use the actual pixel format provided by lavc

This commit is a noop, see 03cef34aa66

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit '6f40181cad8ac04adff7bd10e1e1ab65f22bc1f0'
Clément Bœsch [Sun, 12 Mar 2017 12:13:05 +0000 (13:13 +0100)]
Merge commit '6f40181cad8ac04adff7bd10e1e1ab65f22bc1f0'

* commit '6f40181cad8ac04adff7bd10e1e1ab65f22bc1f0':
  avconv_qsv: align the surface size to 32

This commit is a noop, see 03cef34aa66

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoMerge commit 'b0f36a0043d76436cc7ab8ff92ab99c94595d3c0'
Clément Bœsch [Sun, 12 Mar 2017 12:08:04 +0000 (13:08 +0100)]
Merge commit 'b0f36a0043d76436cc7ab8ff92ab99c94595d3c0'

* commit 'b0f36a0043d76436cc7ab8ff92ab99c94595d3c0':
  avconv: stop using setpts for input framerate forced with -r

Merged-by: Clément Bœsch <u@pkh.me>
2 years agoavcodec/scpr: use correct linesize for prev frame
Paul B Mahol [Sun, 12 Mar 2017 11:34:55 +0000 (12:34 +0100)]
avcodec/scpr: use correct linesize for prev frame

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agoavcodec/dca_xll: Fix runtime error: signed integer overflow: 2147286116 + 6298923...
Michael Niedermayer [Wed, 8 Mar 2017 20:35:51 +0000 (21:35 +0100)]
avcodec/dca_xll: Fix runtime error: signed integer overflow: 2147286116 + 6298923 cannot be represented in type 'int'

Fixes: 732/clusterfuzz-testcase-4872990070145024

See: [FFmpeg-devel] [PATCH 2/6] avcodec/dca_xll: Fix runtime error: signed integer overflow: 2147286116 + 6298923 cannot be represented in type 'int'
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/amrwbdec: Fix runtime error: left shift of negative value -1
Michael Niedermayer [Sat, 11 Mar 2017 02:55:39 +0000 (03:55 +0100)]
avcodec/amrwbdec: Fix  runtime error: left shift of negative value -1

Fixes: 763/clusterfuzz-testcase-6007567320875008

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/dca_xll: Fix runtime error: signed integer overflow: 1762028192 + 698372290...
Michael Niedermayer [Sat, 11 Mar 2017 02:38:01 +0000 (03:38 +0100)]
avcodec/dca_xll: Fix runtime error: signed integer overflow: 1762028192 + 698372290 cannot be represented in type 'int'

Fixes: 762/clusterfuzz-testcase-5927683747741696

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavcodec/wavpack: Fix runtime error: signed integer overflow: -2147483648 + -83886075...
Michael Niedermayer [Sat, 11 Mar 2017 02:25:41 +0000 (03:25 +0100)]
avcodec/wavpack: Fix runtime error: signed integer overflow: -2147483648 + -83886075 cannot be represented in type 'int'

Fixes: 761/clusterfuzz-testcase-5442222252097536

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agoavfilter/af_firequalizer: add av_restrict on convolution func
Muhammad Faiz [Sat, 11 Mar 2017 00:54:30 +0000 (07:54 +0700)]
avfilter/af_firequalizer: add av_restrict on convolution func

slightly improved speed

Reviewed-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
2 years agolavf/dashenc: update bitrates on dash_write_trailer
Przemysław Sobala [Fri, 3 Mar 2017 08:38:28 +0000 (09:38 +0100)]
lavf/dashenc: update bitrates on dash_write_trailer

Provides a way to change bandwidth parameter inside DASH manifest after a non-CBR H.264 encoding.
Caller now is able to compute the bitrate by itself, after all packets have been written, and then set that value in AVFormatContext->streams->codecpar->bit_rate before calling av_write_trailer. As a result that value will be set in DASH manifest.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agodoc/muxers: move hls_flags temp_file to after SECOND LEVEL hls example
Steven Liu [Sat, 11 Mar 2017 13:11:38 +0000 (21:11 +0800)]
doc/muxers: move hls_flags temp_file to after SECOND LEVEL hls example

the temp_file hls_flags describe text offset is wrong, now move it after example

Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
2 years agoaarch64: vp9itxfm: Reorder iadst16 coeffs
Martin Storsjö [Sat, 31 Dec 2016 20:27:13 +0000 (22:27 +0200)]
aarch64: vp9itxfm: Reorder iadst16 coeffs

This matches the order they are in the 16 bpp version.

There they are in this order, to make sure we access them in the
same order they are declared, easing loading only half of the
coefficients at a time.

This makes the 8 bpp version match the 16 bpp version better.

This is cherrypicked from libav commit
b8f66c0838b4c645227f23a35b4d54373da4c60a.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoarm: vp9itxfm: Reorder iadst16 coeffs
Martin Storsjö [Sat, 31 Dec 2016 20:27:13 +0000 (22:27 +0200)]
arm: vp9itxfm: Reorder iadst16 coeffs

This matches the order they are in the 16 bpp version.

There they are in this order, to make sure we access them in the
same order they are declared, easing loading only half of the
coefficients at a time.

This makes the 8 bpp version match the 16 bpp version better.

This is cherrypicked from libav commit
08074c092d8c97d71c5986e5325e97ffc956119d.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoaarch64: vp9itxfm: Reorder the idct coefficients for better pairing
Martin Storsjö [Sat, 31 Dec 2016 12:18:31 +0000 (14:18 +0200)]
aarch64: vp9itxfm: Reorder the idct coefficients for better pairing

All elements are used pairwise, except for the first one.
Previously, the 16th element was unused. Move the unused element
to the second slot, to make the later element pairs not split
across registers.

This simplifies loading only parts of the coefficients,
reducing the difference to the 16 bpp version.

This is cherrypicked from libav commit
09eb88a12e008d10a3f7a6be75d18ad98b368e68.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoarm: vp9itxfm: Reorder the idct coefficients for better pairing
Martin Storsjö [Sat, 31 Dec 2016 12:05:44 +0000 (14:05 +0200)]
arm: vp9itxfm: Reorder the idct coefficients for better pairing

All elements are used pairwise, except for the first one.
Previously, the 16th element was unused. Move the unused element
to the second slot, to make the later element pairs not split
across registers.

This simplifies loading only parts of the coefficients,
reducing the difference to the 16 bpp version.

This is cherrypicked from libav commit
de06bdfe6c8abd8266d5c6f5c68e4df0060b61fc.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoaarch64: vp9itxfm: Avoid reloading the idct32 coefficients
Martin Storsjö [Mon, 2 Jan 2017 20:08:41 +0000 (22:08 +0200)]
aarch64: vp9itxfm: Avoid reloading the idct32 coefficients

The idct32x32 function actually pushed d8-d15 onto the stack even
though it didn't clobber them; there are plenty of registers that
can be used to allow keeping all the idct coefficients in registers
without having to reload different subsets of them at different
stages in the transform.

After this, we still can skip pushing d12-d15.

Before:
vp9_inv_dct_dct_32x32_sub32_add_neon: 8128.3
After:
vp9_inv_dct_dct_32x32_sub32_add_neon: 8053.3

This is cherrypicked from libav commit
65aa002d54433154a6924dc13e498bec98451ad0.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoarm: vp9itxfm: Avoid reloading the idct32 coefficients
Martin Storsjö [Mon, 2 Jan 2017 20:50:38 +0000 (22:50 +0200)]
arm: vp9itxfm: Avoid reloading the idct32 coefficients

The idct32x32 function actually pushed q4-q7 onto the stack even
though it didn't clobber them; there are plenty of registers that
can be used to allow keeping all the idct coefficients in registers
without having to reload different subsets of them at different
stages in the transform.

Since the idct16 core transform avoids clobbering q4-q7 (but clobbers
q2-q3 instead, to avoid needing to back up and restore q4-q7 at all
in the idct16 function), and the lanewise vmul needs a register in
the q0-q3 range, we move the stored coefficients from q2-q3 into q4-q5
while doing idct16.

While keeping these coefficients in registers, we still can skip pushing
q7.

Before:                              Cortex A7       A8       A9      A53
vp9_inv_dct_dct_32x32_sub32_add_neon:  18553.8  17182.7  14303.3  12089.7
After:
vp9_inv_dct_dct_32x32_sub32_add_neon:  18470.3  16717.7  14173.6  11860.8

This is cherrypicked from libav commit
402546a17233a8815307df9e14ff88cd70424537.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoarm: vp9lpf: Implement the mix2_44 function with one single filter pass
Martin Storsjö [Sat, 14 Jan 2017 11:22:30 +0000 (13:22 +0200)]
arm: vp9lpf: Implement the mix2_44 function with one single filter pass

For this case, with 8 inputs but only changing 4 of them, we can fit
all 16 input pixels into a q register, and still have enough temporary
registers for doing the loop filter.

The wd=8 filters would require too many temporary registers for
processing all 16 pixels at once though.

Before:                          Cortex A7      A8     A9     A53
vp9_loop_filter_mix2_v_44_16_neon:   289.7   256.2  237.5   181.2
After:
vp9_loop_filter_mix2_v_44_16_neon:   221.2   150.5  177.7   138.0

This is cherrypicked from libav commit
575e31e931e4178e9f1e24407503c9b4ec0ef9ba.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoaarch64: vp9lpf: Use dup+rev16+uzp1 instead of dup+lsr+dup+trn1
Martin Storsjö [Thu, 23 Feb 2017 21:33:58 +0000 (23:33 +0200)]
aarch64: vp9lpf: Use dup+rev16+uzp1 instead of dup+lsr+dup+trn1

This is one cycle faster in total, and three instructions fewer.

Before:
vp9_loop_filter_mix2_v_44_16_neon: 123.2
After:
vp9_loop_filter_mix2_v_44_16_neon: 122.2

This is cherrypicked from libav commit
3bf9c48320f25f3d5557485b0202f22ae60748b0.

Signed-off-by: Martin Storsjö <martin@martin.st>
2 years agoarm/aarch64: vp9lpf: Keep the comparison to E within 8 bit
Martin Storsjö [Sat, 14 Jan 2017 18:49:19 +0000 (20:49 +0200)]
arm/aarch64: vp9lpf: Keep the comparison to E within 8 bit

The theoretical maximum value of E is 193, so we can just
saturate the addition to 255.

Before:                     Cortex A7      A8      A9     A53  A53/AArch64
vp9_loop_filter_v_4_8_neon:     143.0   127.7   114.8    88.0         87.7
vp9_loop_filter_v_8_8_neon:     241.0   197.2   173.7   140.0        136.7
vp9_loop_filter_v_16_8_neon:    497.0   419.5   379.7   293.0        275.7
vp9_loop_filter_v_16_16_neon:   965.2   818.7   731.4   579.0        452.0
After:
vp9_loop_filter_v_4_8_neon:     136.0   125.7   112.6    84.0         83.0
vp9_loop_filter_v_8_8_neon:     234.0   195.5   171.5   136.0        133.7
vp9_loop_filter_v_16_8_neon:    490.0   417.5   377.7   289.0        271.0
vp9_loop_filter_v_16_16_neon:   951.2   814.7   732.3   571.0        446.7

This is cherrypicked from libav commit
c582cb8537367721bb399a5d01b652c20142b756.

Signed-off-by: Martin Storsjö <martin@martin.st>