Merge commit '8606e881b02bec2ac24943d22c8afe11d641fac8'

* commit '8606e881b02bec2ac24943d22c8afe11d641fac8':
  mpeg12: Move the vlc bits to a stand alone file

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

Merge commit 'f1fa1eed2abdc8dfb0af318a43f5d293b81141bd'

* commit 'f1fa1eed2abdc8dfb0af318a43f5d293b81141bd':
  mpegvideo: Expand macro

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

h264: update avctx width/height/pix_fmt when returning frame

Inconsistencies between the dimensions/pixel format of avctx and the
frame can confuse API users.
For example this can crash the demuxing_decoding example.

Back up the previous values and restore them, when decoding the next
frame. This is necessary, because these can be different between the
returned frame and the last decoded frame.

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

Merge commit 'e7af52a68dde9144b273a9598b60bf0f56e1323b'

* commit 'e7af52a68dde9144b273a9598b60bf0f56e1323b':
  mpegvideo: rv10: Move function declaration to a separate header

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

Merge commit 'e3d0f49abb20a551bf6d885f75c354d6d0bbeb9d'

* commit 'e3d0f49abb20a551bf6d885f75c354d6d0bbeb9d':
  mpegvideo: h263: Move all tables to a single file

Conflicts:
libavcodec/h263.h
libavcodec/h263data.h

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

Merge commit '8b5007a31b8d1ddbe3661bf45a732336450b7d25'

* commit '8b5007a31b8d1ddbe3661bf45a732336450b7d25':
  mpegvideo: Move ER functions to a separate file

Conflicts:
libavcodec/mpegvideo.c

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

libavcodec/imgconvert.c: support left band while cropping

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

avformat/mxfdec: Detect jpeg2000 through codec_ul too

Fixes Ticket2345

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

doc/filters: add one more zoompan example

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

matroskadec: verify seekhead IDs

Some files have SeekHead elements with broken IDs. They mismatch with
the ID of the destination element. These files are written by
"IDMmkvlib0.1" (as identified by the MuxingApp and WritingApp elements),
and the SeekHead IDs are actually endian-swapped.

This confuses the SeekHead logic of the demuxer. It will read some
elements twice, because the SeekHead ID is used to identify and remember
already read elements. With the file at hand, the stream list was
duplicated by reading the Tracks element twice.

Fix this by rejecting invalid EBML IDs in SeekHead entries. (This fix is
relatively specific to the broken file at hand, and doesn't protect
against some other cases of broken SeekHead, such as valid but
mismatching target element IDs.)

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

avfilter/vf_zoompan: support planar rgb pixel formats

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

avcodec/jpeg2000dec: Reduce verbosity of get_plt()

The existence of such element does not represent an error

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

avcodec/jpeg2000dec: Fallback to yuv if no matching xyz format exists

Fixes Ticket4471

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

avcodec/jpeg2000dec: Do not abort if prc is outside limits

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

ffmpeg_opt: Check for localtime() failure

Found-by: Daemon404
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Mark vectors as NAN instead of dereferencing NULL pointers on malloc failure

Found-by: Daemon404
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swscale: Assert that pixel format descriptor is not NULL

This may help static analyzers, the pixel format is checked
during initialization

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

avcodec/jpeg2000dec: Assert that pixel format descriptor is not NULL

We only look up pixel formats from a fixed list so none should ever
fail to be found

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

avformat/mpegts: recognizes and export private streams

Based on patch by Wolfgang Lorenz <wl-chmw@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

configure: Disable VSX on unspecified / generic CPUs

Fixes fate tests on PPC64be

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

mpeg12: Move the vlc bits to a stand alone file

mpegvideo: Expand macro

Having this macro in an header only facilitates the use of such header.
The code increase is minimal and files have one less dependency
on mpegvideo.h.

mpegvideo: rv10: Move function declaration to a separate header

mpegvideo: h263: Move all tables to a single file

mpegvideo: Move ER functions to a separate file

avcodec/jpeg2000dec: Do not hardcode tile part usage

Fixes part of Ticket 4605

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

avcodec/jpeg2000dec: Do not print a warning for RLCP

It seems to be working now

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

avcodec/jpeg2000dec: Remove unused variable and argument

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

avcodec/jpeg2000dec: Fix Selective arithmetic coding bypass and Multiple codeword segments

These 2 are highly related so they are in the same commit
Fixes part of Ticket4605
Fixes p0_04.j2k

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

avcodec/mqcdec: Support raw bypass and non reseting init

Some broken code that used the old mqc is removed

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

avcodec/jpeg2000dec: add some sanity checking on newpasses

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

avcodec/jpeg2000dec: Add some additional checking on lengthinc

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

avcodec/jpeg2000dec: Print warning for "Selective arithmetic coding bypass"

This is implemented but i suspect does not work correctly as no file using this
seems to decode correctly.

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

avcodec/jpeg2000dec: Implement progression order RLCP

This works with one of 2 RLCP samples i have, the other appears to also
use selective arithemtic coding bypass which may or may not be the reason why it does
not work

Sample: p0_16.j2k

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

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

This patch adds MSA (MIPS-SIMD-Arch) optimizations for AVC idct functions in new file h264idct_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

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 AVC intra prediction functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for AVC intra prediction functions in new file h264pred_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

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

avutil/softfloat: Move av_sf2double() to header

It was not intended to be a static private function

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

avutil/softfloat: properly separate testcode

This matches how other self tests are implemented

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

avformat/avio: move short seek threshold to the context

This allows us to adjust it internally.

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

avformat/aiffdec: avoid double and ldexp()

There is no support for non integer sample rates, using doubles/floats currently could
only lead to rounding differences between platforms

Previous version Reviewed-by: Mark Harris <mark.hsj@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mips: Add 'const' to static arrays in HEVC MSA code

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 AVC chroma mc functions

s patch adds MSA (MIPS-SIMD-Arch) optimizations for AVC chroma mc functions in new file h264chroma_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

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

libvpxenc: add overshoot-pct option

balances the presence of undershoot-pct

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Zern <jzern@google.com>

libvpxenc: add undershoot-pct option

replaces deprecated rc_buffer_aggressivity

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Zern <jzern@google.com>

avcodec/pcm-dvd: Fix debug == FF_DEBUG_PICT_INFO

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

avformat: Fix bug in parse_rps for HEVC.

Make the logic in libavformat/hevc.c parse_rps align with libavcodec/hevc_ps.c ff_hevc_decode_short_term_rps

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

avdevice/openal-dec: Make al_format_info const

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

avformat/movenc: Make static packed_size, const static

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

avcodec/mpeg4videodec: Mark static table as const

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

avformat/mpegts: Do not ignore registration descriptors when probing is enabled

This is required for the (not yet in git) private stream detection/export,
no other testcase known

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

doc/ffmpeg: remove tips

Ancient and misleading.

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

Merge commit 'c0e6b8cab874db97b6818007bc86507c8d213083'

* commit 'c0e6b8cab874db97b6818007bc86507c8d213083':
  mpegvideo: Mark one function as static

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

Merge commit 'd1dd0d404c085f4fce7b8358b4aea677761c5d88'

* commit 'd1dd0d404c085f4fce7b8358b4aea677761c5d88':
  mpegvideo: Move block permutation function where is used

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

Merge commit 'c6aa0554b0c3e31fec4580b68ea85b66966cd381'

* commit 'c6aa0554b0c3e31fec4580b68ea85b66966cd381':
  mpegvideo: Drop release_unused_pictures() function

Conflicts:
libavcodec/mpegvideo.c

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

Merge commit '7061bf0925a7d6c5899bbe710b2b8cdae9005274'

* commit '7061bf0925a7d6c5899bbe710b2b8cdae9005274':
  mpegvideo: Move Picture-related functions to a separate file

Conflicts:
libavcodec/mpegvideo.c
libavcodec/mpegvideo.h

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

Merge commit '529c05698e88b057f0bea61e0d85f2b42925b5ea'

* commit '529c05698e88b057f0bea61e0d85f2b42925b5ea':
  movenc: fixes a questionable valgrind uninitialized value warning

See: 3c5c6870a7ffa4997928caddb0802c7a1e5fcfeb
Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavf/http: Correctly terminate session with HTTP POST client.

Send a footer to correctly close client sockets.
This fixes network errors in client applications.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>
Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mips: remove ;;

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

libavformat/flvdec.c: don't build index_entries for input stream if AVIOContext is not seekable

Signed-off-by: Igor Derzhavin <igor.derzhavin@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

mpegvideo: Mark one function as static

mpegvideo: Move block permutation function where is used

Mark it as static.

mpegvideo: Drop release_unused_pictures() function

It is only used in one place, and it is small enough that it might be
called directly.

mpegvideo: Move Picture-related functions to a separate file

movenc: fixes a questionable valgrind uninitialized value warning

display_matrix_size is only initialized when av_stream_get_side_data()
returns a side data pointer. The code is safe since the only effect this
has is setting the display_matrix pointer to NULL which it was already
anyway.

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for HEVC intra prediction functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for HEVC intra predition functions in new file hevcpred_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

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 HEVC loop filter and sao functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for HEVC loop filter and sao functions in new file hevc_lpf_sao_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

In this patch, in comparision with previous patch, duplicated c functions are removed.

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

takdec: ensure chan2 is a valid channel index

If chan2 is not smaller than the number of channels, it can cause
segmentation faults due to dereferencing a NULL pointer.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h264_slice: Use AVFrame diemensions for grayscale handling

The AVFrame values are closer to the AVFrame bitmap changed instead of
the AVCodecContext values, so this should be more robust

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

avcodec/jpeg2000dec: Print the correct variable in the component error message

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

avcodec/jpeg2000dec: increase tile_parts to 6

Fixes Ticket4602

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

avcodec/jpeg2000dec: Fix shift for RGB(A)48/64

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

avdevice/lavfi: do not rescale AV_NOPTS_VALUE in lavfi_read_packet()

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

MAINTAINERS: add 2.7

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

sonic: set avctx->channels in sonic_decode_init

Otherwise it can be 0 in sonic_decode_frame, causing SIGFPE crashes.

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

Changelog: Add 2.7 entry

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

doc/APIchanges: fill in missing stuff

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

doc/APIchanges: Add 2.7 separator

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

libavutil/channel_layout: Correctly return layout when channel specification ends with a trailing 'c'.

Return layout when FF_API_GET_CHANNEL_LAYOUT_COMPAT is set even if the
layout itself is not in the deprecated style.

Signed-off-by: Simon Thelen <ffmpeg-dev@c-14.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavf/webm_chunk: Remove AVFMT_ALLOW_FLUSH

The flag was set unintentionally and the code will break if a NULL
packet is passed in.

Signed-off-by: Vignesh Venkatasubramanian <vigneshv@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/jpeg2000dec: Fix gain regression with gray16

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

fate: Force jpeg2000 decoder for dcinema

so that no external decoder is used which would break the test

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

avcodec/libopenjpegdec: Mark as experimental if <= 1.3

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

avcodec/jpeg2000: remove unused variable

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

avformat/oggenc: Fix return code in case of flushing

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

Merge commit 'b14086ca38efa1a86cb0f0c6aa147b05f698877b'

* commit 'b14086ca38efa1a86cb0f0c6aa147b05f698877b':
  mkv: Correctly report the latest packet had been flushed

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

Merge commit 'b380337020e271c5431aa8ef8f8e9dfda5e919b2'

* commit 'b380337020e271c5431aa8ef8f8e9dfda5e919b2':
  mpjpegdec: don't try to alloc an AVIOContext when probe is guaranteed to fail

See: 1382add59df149193620ca0714ceac0929208c88
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '210921722bf828b3b895ebcbc34374e6c4452c6f'

* commit '210921722bf828b3b895ebcbc34374e6c4452c6f':
  imc: add required padding for GetBitContext buffer

Conflicts:
libavcodec/imc.c

See: 7444cf9a9c0b8b2bba8198af2823521c654a48f4
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '09447f2b0fafac6d9565aab82a4c5f16fc99ee5e'

* commit '09447f2b0fafac6d9565aab82a4c5f16fc99ee5e':
  ac3_parser: add required padding for GetBitContext buffer

See: fccd85b9f30525f88692f53134eba41f1f2d90db
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'fb1473080223a634b8ac2cca48a632d037a0a69d'

* commit 'fb1473080223a634b8ac2cca48a632d037a0a69d':
  aac_parser: add required padding for GetBitContext buffer

See: fccd85b9f30525f88692f53134eba41f1f2d90db
Merged-by: Michael Niedermayer <michaelni@gmx.at>

doc/muxers: document new break_non_keyframes option

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

avcodec/jpeg2000: Disable special case for JPEG2000_QSTY_SI

The code gave apparently completely wrong values

Fixes Ticket2872

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

avcodec/jpeg2000dec: Add placeholder for PLT parsing

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

avcodec/jpeg2000dec: Handle Psot = 0

Fixes Ticket2869

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

mkv: Correctly report the latest packet had been flushed

Bug-Id: 865
CC: libav-stable@libav.org

lavf/segment: add an option to allow breaking on non-keyframes

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

mpjpegdec: don't try to alloc an AVIOContext when probe is guaranteed to fail

The first check is done without the AVIOContext, so alloc it only if said check succeeds

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Janne Grunau <janne-libav@jannau.net>

imc: add required padding for GetBitContext buffer

Fixes stack buffer overflow errors detected by address sanitizer in
fate-imc.

CC: libav-stable@libav.org

ac3_parser: add required padding for GetBitContext buffer

Fixes stack buffer overflow errors detected by address sanitizer in
various fate tests.

CC: libav-stable@libav.org

aac_parser: add required padding for GetBitContext buffer

Fixes stack buffer overflow errors detected by address sanitizer in
various fate tests.

CC: libav-stable@libav.org

avcodec/x86/h264_weight: handle weight1=128

Fix ticket4596

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