Merge commit '76ce9bd8e26dcb3652240a1072840ff4011d7cdc'

* commit '76ce9bd8e26dcb3652240a1072840ff4011d7cdc':
  libavutil: Add ARM av_clip_intp2_arm

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

Merge commit 'bf07d813f6c88b5a76980f321cf7272d799c4216'

* commit 'bf07d813f6c88b5a76980f321cf7272d799c4216':
  libavutil: Add av_clip_intp2

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

Makefile: improve DBG option for asm

This improves the cleanup, dependency generation and DBG command issuing.

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

Merge commit '3fe26bb60858951710a42065c8183cce100d632a'

* commit '3fe26bb60858951710a42065c8183cce100d632a':
  hlsenc: write playlist into a temp file and replace the original atomically

Conflicts:
libavformat/hlsenc.c

See: da97e4c64a289ac83c3d53554ff36cdd6586537c
Merged-by: Michael Niedermayer <michaelni@gmx.at>

avformat/hlsenc: Use ff_rename() only for local files

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

hlsenc: write playlist into a temp file and replace the original atomically

Reviewed-by: Lukasz Marek <lukasz.m.luki2@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'ff48ae94f90f6862dbe5b1e7397a1bda3150e200'

* commit 'ff48ae94f90f6862dbe5b1e7397a1bda3150e200':
  hlsenc: remove the AVIOContext for the playlist from the muxer context

Conflicts:
libavformat/hlsenc.c

See: 28bf05e7b50e0920ff9ea4deded1fd1d7b817616
Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavutil: Add ARM av_clip_intp2_arm

add ARM code for implementing av_clip_intp2 using the ssat instruction

on Cortex-A8, av_clip_intp2_arm() is faster than av_clip_intp2_c() and
the generic av_clip(), about -19%

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

libavutil: Add av_clip_intp2

there already is a function, av_clip_uintp2() that clips a signed integer
to an unsigned power-of-two range, i.e. 0,2^p-1

this patch adds a function av_clip_intp2() that clips a signed integer
to a signed power-of-two range, i.e. -(2^p),(2^p-1)

the new function can be used as a special case for av_clip(), e.g.
av_clip(x, -8192, 8191) can be rewritten as av_clip_intp2(x, 13)

there are ARM instructions, usat and ssat resp., which map nicely to these
functions (see next patch)

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

swresample: add av_cold to init functions

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

lavf/ffmenc: do not fail on missing codec

ffm encoder fails when codec is not found.
It may happen when stream is being copied.
This commit allows to store such stream and provides
backward compatibility with version prior 2.5 release.

fixes #4266

Signed-off-by: Lukasz Marek <lukasz.m.luki2@gmail.com>

hlsenc: write playlist into a temp file and replace the original atomically

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

hlsenc: remove the AVIOContext for the playlist from the muxer context

Its only used in one function, having it in the context serves no purpose.

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

Merge commit 'b8d18a94376c44dac9523dc3317259a30fc92f8b'

* commit 'b8d18a94376c44dac9523dc3317259a30fc92f8b':
  libavcodec: Don't use av_cold annotation in twinvq header file

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

Merge commit 'df0891fc8f32db3ced797cd6ebff9492fda01b99'

* commit 'df0891fc8f32db3ced797cd6ebff9492fda01b99':
  libavresample: Annotate AARCH64 init function with av_cold

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

Merge commit '46ad9ac9641d1fe8292ec1f46bbd2e4456621ca7'

* commit '46ad9ac9641d1fe8292ec1f46bbd2e4456621ca7':
  rtpdec_h264: Move a leftover comment into h264_handle_packet_stap_a

Conflicts:
libavformat/rtpdec_h264.c

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

Merge commit 'a335ed767161c6da2815371177cfd5e40f78e5b7'

* commit 'a335ed767161c6da2815371177cfd5e40f78e5b7':
  rtpdec_h264: Remove an unnecessary check

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

Merge commit '176903ce833ce7469f411640e9748a0d549b5285'

* commit '176903ce833ce7469f411640e9748a0d549b5285':
  rtpdec_h264: Return immediately on errors in h264_handle_packet_stap_a

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

Merge commit 'bb8c6ac840afa69dd37860fdf85da9f4cf1e0ae4'

* commit 'bb8c6ac840afa69dd37860fdf85da9f4cf1e0ae4':
  rtpdec_h264: Make a parameter pointer const

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

Merge commit '7650caf013f45ebebf128855735a0c6350836ea4'

* commit '7650caf013f45ebebf128855735a0c6350836ea4':
  rtpdec_h264: Use av_realloc instead of av_malloc+mempcy

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

Merge commit '8bdbf49c6f4d9473183a3c45ec70d611eb6183cd'

* commit '8bdbf49c6f4d9473183a3c45ec70d611eb6183cd':
  rtpdec_h264: Include the right header for AV_RB16

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

Merge commit 'ba2e07909b848e583245856a59d0fe1dd08f917e'

* commit 'ba2e07909b848e583245856a59d0fe1dd08f917e':
  rtpdec_vp8: Set the keyframe flag

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

avformat/omadec: Use 64bit for ret to avoid overflow

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

avformat/vqf: Use 64bit for ret to avoid overflow

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

avformat/wtvdec: Use 64bit for ret to avoid overflow

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

avformat/mvdec: Use 64bit for ret to avoid overflow

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

avformat/apngdec: Use 64bit for ret to avoid overflow

Missed this hunk in first commit

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

avformat/gxf: Use 64bit for res to avoid overflow

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

avformat/idcin: Use 64bit for ret to avoid overflow

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

avformat/apngdec: Use 64bit ret to avoid overflow

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

avformat/asfdec: Use 64bit ret to avoid overflow

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

wtvdec: fix integer overflow resulting in errors with large files

This fixes a regression in 9fbc613f0df1628e7e78bca791fa8833846f8210

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

libavcodec: Don't use av_cold annotation in twinvq header file

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Martin Storsjö <martin@martin.st>

libavresample: Annotate AARCH64 init function with av_cold

Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Martin Storsjö <martin@martin.st>

rtpdec_h264: Move a leftover comment into h264_handle_packet_stap_a

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

rtpdec_h264: Remove an unnecessary check

If src_len is too small for nal_size, we already print a warning
above, and the next step is to check the while loop condition
anyway, so this one serves no purpose.

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

rtpdec_h264: Return immediately on errors in h264_handle_packet_stap_a

Previously, errors were only logged but the code kept on trying,
and never actually returning the error as a return value.

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

rtpdec_h264: Make a parameter pointer const

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

rtpdec_h264: Use av_realloc instead of av_malloc+mempcy

This is similar to what was done for rtpdec_hevc in ced7238cd01.

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

rtpdec_h264: Include the right header for AV_RB16

Including libavcodec/get_bits.h is superfluous for AV_RB16 - nothing
in this file uses any actual bitstream reader.

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

rtpdec_vp8: Set the keyframe flag

This allows the output to be used with stream copy, which discards
packet from the start until the first keyframe.

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

MAINTAINERS: Add Oleksij as maintainer for DSS*

Also add a 2nd maintainer at the authors request (due to lack of time)

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

hlsenc: remove the AVIOContext for the playlist from the muxer context

Its only used in one function, having it in the context serves no purpose.

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

avformat/movenc: Move avid DNxHD padding to the correct spot

Outputting DNxHD into .mov containers 'corrupts' following atoms until end of stsd

ffmpeg and qtdump could not decode pasp/colr atoms in the files made by ffmpeg,
when outputting DNxHD due to the incorrect padding placement. Now we add the
padding in the correct place

Tidy up FATE changes due to padding changes.

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

avcodec/dca: move remaining tables to dcadata

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

avcodec/snowdec: Fix avmv_index increment

Fixes reading uninitialized memory

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

avfilter/vf_eq: Add process_command to eq.

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

avcodec/dss_sp: Avoid a slow division and modulo in the pitch_lag calculation

Also check the combined_pitch for validity (which is needed to keep the behavior
identical in the error case)

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

avcodec/dss_sp: de-obfuscate index inversion

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

avcodec/dss_sp: Protect DSS_SP_FORMULA() with a set of outside brackets

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

avformat/dss: Use av_freep() to avoid leaving stale pointers in memory

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

avformat/dss: Check sscanf() return value in dss_read_metadata_date()

Fixes writing uninitialized values into metadata without error in case
parsing fails

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

Merge commit 'b727b0eeb73cd78533a7fd387ceccd505e48fa6c'

* commit 'b727b0eeb73cd78533a7fd387ceccd505e48fa6c':
  fate: Add tests for DSS

Conflicts:
tests/fate/audio.mak

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

Merge commit '062cd5a975ff7bd6fb91f9b4d1d9d102a7545499'

* commit '062cd5a975ff7bd6fb91f9b4d1d9d102a7545499':
  lavf: Add DSS demuxer

Conflicts:
libavformat/Makefile
libavformat/version.h

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

Merge commit 'c56b9b1eb278c5ef89d3f0832a56dfe4732cb68b'

* commit 'c56b9b1eb278c5ef89d3f0832a56dfe4732cb68b':
  lavc: Add DSS SP decoder

Conflicts:
libavcodec/avcodec.h
libavcodec/version.h

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

Merge commit '0fbb271318899a0fb1fbcbb3db8292e909b91e23'

* commit '0fbb271318899a0fb1fbcbb3db8292e909b91e23':
  qsv_api: fix building with libmfx disabled

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

Merge commit '11c22dfd495bbd3f38df639772aa4bb9a1be882d'

* commit '11c22dfd495bbd3f38df639772aa4bb9a1be882d':
  configure: add missing h264_qsv hwaccel deps

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

Merge commit '4e08c821106fc1d6d358864abf0d8488b12f38c6'

* commit '4e08c821106fc1d6d358864abf0d8488b12f38c6':
  lavc: add an Intel libmfx-based H.264 decoder

Conflicts:
configure
libavcodec/Makefile
libavcodec/allcodecs.c
libavcodec/version.h

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

Merge commit '31d2039cb42668ebcf08248bc48bbad44aa05f49'

* commit '31d2039cb42668ebcf08248bc48bbad44aa05f49':
  h264_parser: export video format and dimensions

Conflicts:
libavcodec/h264_parser.c
libavcodec/version.h

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

Merge commit '5b1d9ceec715846a58fe029bc3889ed6fa62436a'

* commit '5b1d9ceec715846a58fe029bc3889ed6fa62436a':
  pixfmt: add a pixel format for QSV hwaccel

Conflicts:
doc/APIchanges
libavutil/pixfmt.h
libavutil/version.h

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

fate: Add tests for DSS

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

lavf: Add DSS demuxer

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

lavc: Add DSS SP decoder

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

qsv_api: fix building with libmfx disabled

configure: add missing h264_qsv hwaccel deps

Merge remote-tracking branch 'cigaes/master'

* cigaes/master:
  doc/faq: explain DAR/SAR preserving.
  doc: set documentencoding on toplevel texi files.

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

avformat/mov: Add simple ACLR atom reading to set the color range of the incomming track for codec's like DNxHD that utilise AVID's proprietary atom.

On input ACLR will be used to set colour range no matter which codec
it is associated with.
No change for when it will be output.

Rework mov_read_extradata function to allow detection of truncated
atom reads by callers.

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

lavc: add an Intel libmfx-based H.264 decoder

Based on the code by Luca Barbato <lu_zero@gentoo.org> and Yukinori
Yamazoe <drocon11@gmail.com>.

h264_parser: export video format and dimensions

pixfmt: add a pixel format for QSV hwaccel

doc/faq: explain DAR/SAR preserving.

doc: set documentencoding on toplevel texi files.

Fix double UTF-8 encoding on output files.
makeinfo / texi2any does not seem to provide a way of setting
it globally for toplevel files.

avfilter/palettegen: raise cache size from 64k to 512k

(or 32k to 256k in 32-bit)

This is similar to a00bab347518d4824d51ec7078f1d84ca097c53b

Merge commit '0f532fd3f9b9ca974cea0feb8ddcaf4a7f3b23c5'

* commit '0f532fd3f9b9ca974cea0feb8ddcaf4a7f3b23c5':
  rv40: Return more meaningful error codes

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

rv40: Return more meaningful error codes

Signed-off-by: Diego Biurrun <diego@biurrun.de>

avcodec/wmalosslessdec: use sizeof() instead of a literal number

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

avfilter/vf_subtitles: Add () to protect the argument of the AA() macro

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

avutil/sha: Protect macro arguments with ()

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

ffmpeg_opt: Add missing MATCH_PER_TYPE_OPT() for data codecs

Fixes: CID1270809
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'd617e77cecc5b363ef1860955b548f4ac007add6'

* commit 'd617e77cecc5b363ef1860955b548f4ac007add6':
  libx265: Add crf private option

See: 8b96e8dd28c83448c9b4af81b834deb9b525be80
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '2de887e45b664f44b51686f5979fa8ce6dfe2ec2'

* commit '2de887e45b664f44b51686f5979fa8ce6dfe2ec2':
  libx265: Reduce the scope of some variables

See: b920db67317a31f0e8d4a010faa26dbd12b0dcdb
Merged-by: Michael Niedermayer <michaelni@gmx.at>

tests: fate: adding fate-test for twofish

Signed-off-by: Supraja Meedinti <supraja0493@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/movenc: Fix use of uninitialized variable (ret)

Fixed CID1270823

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

libx265: Add crf private option

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

libx265: Reduce the scope of some variables

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

libx265: Add crf private option

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

libx265: Reduce the scope of some variables

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

avformat/ffmenc: Use av_freep() to avoid leaving stale pointers in memory

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

avformat/ffmenc: Fix memleak of buf

Fixes CID1257015

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

avfilter/showpalette: fix leak in case of error

Fixes CID1270820
Fixes CID1270818

avfilter/palettegen: fix leak in case of error

avfilter/paletteuse: fix leak in case of error

Fixes CID1270819

avfilter/af_channelmap: Move potential dereference after NULL check in get_channel_idx()

Fixes CID1270822

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

avfilter/vf_qp: Fix leak of out_qp_table_buf

Fixes CID1270821

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

avfilter/avfiltergraph: assert that the heap_bubble index is valid

This might help coverity

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

avcodec/faxcompr: Fix memleak

Fixes CID1270816

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

avcodec/opusdec: remove unused headers

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

avfilter/vf_removelogo: Add () to protect the argument of apply_mask_fudge_factor()

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

swscale/bayer_template: Add () to protect the argument of BAYER_READ()

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

compat/avisynth/windowsPorts/windows2linux: Add () to protect macro arguments

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