Merge commit '1410eeb6ea6bc5784e40032430afcdf54a79aedb'

* commit '1410eeb6ea6bc5784e40032430afcdf54a79aedb':
  imc: Use correct position for flcoeffs2 calculation

See: 75fd5ce4c1c0b2d96d71c74b650cefaaef519d27
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mpeg4videodec: Check P cbpy

Fixes undefined behavior
Fixes: signal_sigsegv_c3097a_991_xtrem_e2_m64q15_a32sxx.3gp
Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

imc: use correct position for flcoeffs2 calculation

flcoeffs2[pos] should be the log2 of flcoeffs1[pos].
flcoeffs1[0] can be 0 here, thus flcoeffs2[pos] gets set to -inf,
causing problems further down.

This seems to have been copied from imc_decode_level_coefficients in
commit 4eb4bb3 without updating the position.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

avfilter/vf_ssim: Mark constant tables as const

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

avcodec/j2kenc: remove unused variable

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

imc: Use correct position for flcoeffs2 calculation

flcoeffs2[pos] should be the log2 of flcoeffs1[pos].
flcoeffs1[0] can be 0 here, thus flcoeffs2[pos] gets set to -inf,
causing problems further down.

This seems to have been copied from imc_decode_level_coefficients in
commit 4eb4bb3 without updating the position.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

hevc: check slice address length

It is used as get_bits argument and reading 0 bits isn't supported.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

Merge commit '872fab4a3df48e7e6484333ee2228c684e319634'

* commit '872fab4a3df48e7e6484333ee2228c684e319634':
  asfdec: Fix reading from the pipe

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

avcodec/utils: Document 32 min for h264 width

Suggested-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/g2meet: Check R/G/B values in epic_decode_pixel_pred()

Fixes: asan_double-free_d34593_861_smp3.wmv
Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

snow: remove an obsolete av_assert2

It asserts that the frame linesize is larger than 37, but it can be
smaller and decoding such frames works.

Before commit cc884a35 src_stride > 7*MB_SIZE was necessary, because the
blocks were interleaved in the tmp buffer and the last block was added
with an offset of 6*MB_SIZE.
It was changed for src_stride <= 7*MB_SIZE to write the blocks
sequentially, hence the larger tmp_step.
After that the assert was only necessary to make sure that the buffer
remained large enough.
Since commit bd2b6b33 s->scratchbuf is used as tmp buffer.
As part of commit 86e107a7 the minimal scratchbuf size was increased to
256*7*MB_SIZE, which is enough for any src_stride <= 7*MB_SIZE.

Also add a comment explaining the tmp_step calculation.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

avcodec/utils: use a minimum 32pixel width in  avcodec_align_dimensions2() for H.264

Fixes Assertion failure
Found-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/motion_est_template: Fix undefined behavior in small_diamond_search()

Fixes: asan_heap-oob_394322e_138_cov_4265020547_CVPCMNL1_SVA_C.264
Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

asfdec: Fix reading from the pipe

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avfilter/vf_removegrain: clip to uint16 instead to uint8

This is how original filter behaves.

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

avcodec/libopenh264enc: Do not truncate frame rate

Suggested-by: Gregory J Wolfe <gregory.wolfe@kodakalaris.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '67c884eb07c7e9f2cb72bb8d447d945e5ac8cac7'

* commit '67c884eb07c7e9f2cb72bb8d447d945e5ac8cac7':
  libvpx: Add the library header

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '461b45efd04859b2672238bc8a6ecab9e9a14948'

* commit '461b45efd04859b2672238bc8a6ecab9e9a14948':
  lavc: Add nvenc.h to the skipheader

Conflicts:
libavcodec/Makefile

No change as there is no nvenc.h in FFmpeg

Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/qsvenc.c: improving handling for return codes of MFXVideoENCODE_EncodeFrameAsync

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mpegvideo: Clear pointers in ff_mpv_common_init()

This ensures that no stale pointers leak through on any path

Fixes: signal_sigsegv_c3097a_991_xtrem_e2_m64q15_a32sxx.3gp
Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libvpx: Add the library header

Unbreak make checkheaders

lavc: Add nvenc.h to the skipheader

Unbreak make checkheaders

fate: add removegrain tests

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

lavf/rtpenc_jpeg: Error out for non-standard Huffman tables.

Related to ticket #3823.

lavf/rtpenc_jpeg: Do not check the table number when checking precision.

ffmpeg: Use av_log to print benchmark output.

Cosmetics: Reindent after last commit.

lavf/matroskaenc: Do not needlessly allocate memory for cuepoints.

Fixes ticket #4690.

libavcodec/qsvenc.c: Fix for too agressive height alignment during frame encoding which may be reason of superflous frame copying.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/qsvenc_hevc: Attempt to fix error: too few arguments to function ff_hevc_extract_rbsp

Found-by: Ivan Uskov
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec: Implementation of AAC_fixed_decoder (LC-module) [4/4]

Build system modified

There are several warnings occurring during build after this patch is
applied. The cause of most of these warnings is in that some definitions
needed here are logical part of sbr module and are added in later patches.
When this patches are applied these warnings stop occurring.

The only warning that is added here and is not fixed with later patches
is warning that warns that type mismatch for table ff_aac_eld_window_480.

The reason for this warning is in that ER AAC ELD 480 is not integrated in
to the fixed point implementation at this moment and there is no fixed point
version of this table.

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec: Implementation of AAC_fixed_decoder (LC-module) [3/4]

Add fixed point implementation

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec: Implementation of AAC_fixed_decoder (LC-module) [2/4]

Add fixed point implementation of functions for generating tables

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec: Implementation of AAC_fixed_decoder (LC-module) [1/4]

Move existing code to the new template files

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/hevc_ps: Remove gotos from ff_hevc_parse_sps()

They are currently unnecessary

Found-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '84b223cc6d6ed4cc8bd295457a90f7c94a9dd784'

* commit '84b223cc6d6ed4cc8bd295457a90f7c94a9dd784':
  configure: Make the new qsv encoder depend on libmfx

Conflicts:
configure

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'd1a6cb195f610978ba5d2351e60f938f7f261d59'

* commit 'd1a6cb195f610978ba5d2351e60f938f7f261d59':
  x86: Serialize rdtsc in read_time()

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '40af330adf7fde8073271cf2b41ff9adc4c2bba9'

* commit '40af330adf7fde8073271cf2b41ff9adc4c2bba9':
  avconv: vda: Unlock the pixel buffer once it is accessed

See: c06fdacc3dc706e70d953917fea845532d3703ca
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '8fcd121b823caeadbe7597c9ae9229f6f164f949'

* commit '8fcd121b823caeadbe7597c9ae9229f6f164f949':
  doc: Use the succinct syntax for the channelmap example

Conflicts:
doc/filters.texi

Merged-by: Michael Niedermayer <michaelni@gmx.at>

configure: Fix build without libmfx

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '66acb76bb0492b263215ca9b4d927a7be39ace02'

* commit '66acb76bb0492b263215ca9b4d927a7be39ace02':
  lavc: add Intel libmfx-based HEVC encoder

Conflicts:
Changelog
configure
libavcodec/Makefile
libavcodec/allcodecs.c
libavcodec/qsv.c
libavcodec/qsvenc.c
libavcodec/version.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

configure: Fix build without libmfx

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '3a85397e8bb477eb34678d9edc52893f57003226'

* commit '3a85397e8bb477eb34678d9edc52893f57003226':
  lavc: add Intel libmfx-based MPEG2 encoder

Conflicts:
Changelog
configure
libavcodec/allcodecs.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '69ab9f53f901eac6a649e22d28cf093357870627'

* commit '69ab9f53f901eac6a649e22d28cf093357870627':
  hevc: split bitstream unescaping to a separate file

Conflicts:
libavcodec/Makefile
libavcodec/hevc.c

See: afa93d198aaf2cc661c4df6d4095cd030265d30a
Merged-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/hevc_ps: Do not return success on failures in ff_hevc_parse_sps()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'fd124d8357b1becfde3ac8d5e3320127cf97a5b7'

* commit 'fd124d8357b1becfde3ac8d5e3320127cf97a5b7':
  hevc_ps: split the code for parsing the SPS and exporting it into the context

Conflicts:
libavcodec/hevc.c
libavcodec/hevc_ps.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

avcodev/libdcadec: implement request_channel_layout

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

Merge commit '0e7c0ec344f542e68e3cc9680e8d41dffeffdb4e'

* commit '0e7c0ec344f542e68e3cc9680e8d41dffeffdb4e':
  lavf/hevc: pad the RBSP buffer as required by the bistream reader

Merged-by: Michael Niedermayer <michaelni@gmx.at>

oggparsedirac: check return value of init_get_bits

If init_get_bits fails the GetBitContext is invalid and must not be
used. Check the return value in dirac_header and propogate the error.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/qsvenc.c: fix incorrect loop condition.

For example, the encoder may return MFX_WRN_INCOMPATIBLE_VIDEO_PARAM warning
i.e. ret==5 old loop implementation will repeat several times until output buffer
overflow. New implementation explicitly uses loop only for device busy case.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/movenc: fix mime-types in movenc.c

https://developer.apple.com/library/ios/documentation/AppleApplications/Reference/SafariWebContent/CreatingContentforSafarioniPhone/CreatingContentforSafarioniPhone.html

says video/mp4

suggested by BBB on irc.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '1761ab838c75223a6b97d8c0720d09275374c53d'

* commit '1761ab838c75223a6b97d8c0720d09275374c53d':
  lavc: Deprecate avctx.rc_strategy

Conflicts:
libavcodec/mpegvideo.h
libavcodec/version.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '02b7c630875c0bc63cee5ec597aa33baf9bf4e20'

* commit '02b7c630875c0bc63cee5ec597aa33baf9bf4e20':
  h261: Signal freeze picture release for intra frames

Conflicts:
tests/ref/vsynth/vsynth1-h261
tests/ref/vsynth/vsynth2-h261

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'dc1de0b958836545339611e9c050a1d4fdded263'

* commit 'dc1de0b958836545339611e9c050a1d4fdded263':
  h261: Set 'still image mode off' in picture header

Conflicts:
tests/ref/vsynth/vsynth1-h261
tests/ref/vsynth/vsynth2-h261

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'c8b8271379b200c5c6fa89ca995f90f97f55c2c5'

* commit 'c8b8271379b200c5c6fa89ca995f90f97f55c2c5':
  xcbgrab: Explicitly include xcb/shape.h

See: 54170a33c2c97b0f50347f57e8f0f2ea681dca1d
Merged-by: Michael Niedermayer <michaelni@gmx.at>

configure: Make the new qsv encoder depend on libmfx

Found-by: kropping

Merge commit '161a301d44274645c2272855dac3e4664f935603'

* commit '161a301d44274645c2272855dac3e4664f935603':
  mpjpeg: Write the Content-length

Conflicts:
libavformat/mpjpeg.c

See: 0d2f4eedc8a46892471c51cbc7a78cd9a489771a
Merged-by: Michael Niedermayer <michaelni@gmx.at>

x86: Serialize rdtsc in read_time()

Improves the accuracy of measurements, especially in short sections.

To quote the Intel 64 and IA-32 Architectures Software Developer's Manual:
"The RDTSC instruction is not a serializing instruction. It does not necessarily
wait until all previous instructions have been executed before reading the counter.
Similarly, subsequent instructions may begin execution before the read operation
is performed. If software requires RDTSC to be executed only after all previous
instructions have completed locally, it can either use RDTSCP (if the processor
supports that instruction) or execute the sequence LFENCE;RDTSC."

SSE2 is a requirement for lfence so only use it on SSE2-capable systems.
Prefer lfence;rdtsc over rdtscp since rdtscp is supported on fewer systems.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avconv: vda: Unlock the pixel buffer once it is accessed

Avoid possible issues with memmapped hardware buffers in
case VDA is not doing a conversion on behalf of the user
and make the code more proper as working example.

CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

doc: Use the succinct syntax for the channelmap example

Mixing succinct and long syntax does not work.

Merge commit 'd09b4cce21cdad5ef2855698395ffd6e37445212'

* commit 'd09b4cce21cdad5ef2855698395ffd6e37445212':
  mpjpeg: Simplify using avio_printf

Conflicts:
libavformat/mpjpeg.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavc: add Intel libmfx-based HEVC encoder

lavc: add Intel libmfx-based MPEG2 encoder

hevc: split bitstream unescaping to a separate file

It will be useful in the QSV HEVC encoder.

hevc_ps: split the code for parsing the SPS and exporting it into the context

This will be useful in the later commits, where we want to parse an SPS
without having a whole decoding context.

lavf/hevc: pad the RBSP buffer as required by the bistream reader

avfilter: add removegrain

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

avcodec/h264_slice: Fix container cropping

Fixes out of array read
Fixes: asan_heap-oob_394322e_138_cov_4265020547_CVPCMNL1_SVA_C.264
Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc: Deprecate avctx.rc_strategy

Only used by libxvid in ratecontrol module, so move it to a codec
private option.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

avcodec/g2meet: Clear pointers after deallocation

Fixes double free

Found-by: Samuel Groß, Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

build: add LDLIBFLAGS

Fixes Ticket4673

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/utils: Check values in apply_param_change()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/swfdec: Fix "}else" style

Found-by: durandal_170
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avfilter/vf_colormatrix: add yuv444p support

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

avformat/swfdec: Do not error out on pixel format changes

Instead print an error and continue

Fixes Ticket4702

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

videodsp: assert that linesize is larger than width

Suggested-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

wmalosslessdec: reset frame->nb_samples on packet loss

Otherwise a frame with non-zero nb_samples but without any data can be
returned.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

wmalosslessdec: avoid reading 0 bits with get_bits

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

aacenc: fix option descriptions

Since the new PNS implementation has been merged and is no longer considered
proof of concept (as it's much more complex and better than the previous), change
the comments to reflect that. We need people testing it (since all AAC profiles
require it to be on by default) and having it tagged as proof of concept might drive some away.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/libdcadec: export matrix encoding side data

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

Put a space between string literals and macros.

When compiling libavutil/internal.h as C++11, clang warns that a space
is required between a string literal and an identifier. Put spaces
in concatenations of string literals and EXTERN_PREFIX.

Signed-off-by: Chris Watkins <watk@chromium.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

configure: loongson enabled local aligned 32

Signed-off-by: ZhouXiaoyong <zhouxiaoyong@loongson.cn>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

h261: Signal freeze picture release for intra frames

Freeze picture release should be set to 1 when we're responding to a
fast update request. For simplicity we set it for all intra frames,
including those that starts a GOP.

Fixes issue where Tandberg MXP1700 does not recover from packet loss
state since it's waiting for the freeze picture relase indication.

Bug-Id: 873
CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

h261: Set 'still image mode off' in picture header

Ref H.261 recommendation section 4.2.1.3, setting the still image flag
to 1 disables still image mode. Some decoders require this in order to
decode the bitstream as normal video.

Fixes H.261 calls to Cisco E20.

Also, reserved (aka spare) bits should be set to 1 unless specified
otherwise.

Bug-Id: 872
CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for idctdsp functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for idctdsp functions in new file idctdsp_msa.c and simple_idct_msa.c

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/qsv.c: Linux-only code part has been moved to separate function in order to avoid the "ISO C90 forbids mixed declarations and code" compiler warning.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavf/mpegtsenc: Only fail aac muxing if the first frame is invalid.

Fixes ticket #3957.

configure: add loongson2 cpu support

Signed-off-by: ZhouXiaoyong <zhouxiaoyong@loongson.cn>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

use mmi instead of loongson3 as simd-optimization flag

Signed-off-by: ZhouXiaoyong <zhouxiaoyong@loongson.cn>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec: loongson remove useless macros in mipsfpu optimization

Loongson has disabled all mipsfpu optimization as fate-test faild.

Signed-off-by: ZhouXiaoyong <zhouxiaoyong@loongson.cn>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

xcbgrab: Explicitly include xcb/shape.h

Found-By: Cheristheus

avcodec/lpc: Fix lpc_apply_welch_window_c() for odd len

Also removes assert
this fixes an assertion failure on non-x86

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for me_cmp functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for me_cmp functions in new file me_cmp_msa.c

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for mpegvideoencdsp functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for mpegvideoencdsp functions in new file mpegvideoencdsp_msa.c

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/qsvenc.c: A warning message when library will work at partial hardware acceleration.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffmpeg_opt: allow the user to ignore unused stream maps

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec: Add support for Closed Caption export in h264

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/qsvenc.c: More correct selection of alignment of a frame height depending whether an encoded sequence progressive or not.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc/utils: get rid of add_metadata_from_side_data forward declaration

Signed-off-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc/utils: call add_metadata_from_side_data in ff_init_buffer_info

This should ensure that each frame get its metadata from its proper packet
regardless of frame delays caused by reordering or threading.

Signed-off-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>