configure: add encoding to --enable-libopus doc.

avutil/dict: delay addition of const from e12a73246d8ce7d0fc4036522688934e26de4bb1 until next major ABI bump

This unbreaks API

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

avutil/fifo: delay addition of const from 78d3453c4a2efef9bc079e8f3458653beafcf990 until next major ABI bump

This unbreaks API, for example audacity has more build errors due to this

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

avutil/version: add FF_CONST_AVUTIL53

This will become "const" on the next major API version

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

Merge commit '7e90133f6420b1c53652f972b9561600822881ee'

* commit '7e90133f6420b1c53652f972b9561600822881ee':
  build: do not append $(FFLIBS-) to $(FFLIBS) if $(NAME) is not defined

Conflicts:
common.mak

See: efa95968317411179b0016af54745906029c2295
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'a738540366c9b114949b7914c0d08e2c28982cfb'

* commit 'a738540366c9b114949b7914c0d08e2c28982cfb':
  lavf: properly document the distinction between flags and ctx_flags

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

build: do not append $(FFLIBS-) to $(FFLIBS) if $(NAME) is not defined

Avoids including disabled library Makefiles.

avfilter/lut3d: move the scale calc out of the inner loop.

lavf: properly document the distinction between flags and ctx_flags

avfilter/vf_removelogo: use av_malloc_array()

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

avfilter/vf_hqdn3: use av_malloc_array()

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

avcodec/libx264: Implement reference frame count limiting based on level

This makes libavcodec/libx264.c behave more similar to the x264 command line
util

Fixes Ticket3307

Implementation based on x264

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

swscale/utils: Add check that ensures that the hardcoded struct offsets are valid

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

build: move ebur128 swr dep to configure.

configure: swresample depends on avutil

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

common.mak: fix (not)building disabled libs

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

Merge commit '449511740f06a4675b0066730fa45cdb764ffafc'

* commit '449511740f06a4675b0066730fa45cdb764ffafc':
  build: handle library dependencies in configure

Conflicts:
common.mak
configure
libavdevice/Makefile
libavfilter/Makefile

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

dpx: use intfloat.h instead of deprecated intfloat_readwrite.h.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Merge commit '9aa4592076d4dbb29d1198b0e258f9f85c0c00b5'

* commit '9aa4592076d4dbb29d1198b0e258f9f85c0c00b5':
  aarch64: assembler in clang-3.4 ignores the division by two

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

build: handle library dependencies in configure

Instead of setting FFLIBS in each library Makefile configure
exports FFLIBS-$library in config.mak.

aarch64: assembler in clang-3.4 ignores the division by two

Values are positive powers of two, so just replace it with right shift.

Use rac_get_prob branchy version when used within a conditional branch.

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

Support broadcast destination for udp protocol

Use the required socket option SO_BROADCAST to be able to stream to a broadcast
address.

Prior to the patch, trying to stream to a broadcast address was resulting to the
following error:

av_interleaved_write_frame(): Permission denied

The patch has been tested with:

ffmpeg -f v4l2 -framerate 30 -input_format yuyv422 -video_size 640x480 -i /dev/video0 \
 -c:v libx264 -profile:v high -preset ultrafast -tune zerolatency -b:v 500k -pix_fmt yuv420p \
 -f mpegts udp://192.168.1.255:5004?broadcast=1

I have added an option to let the user explicitly request broadcast in order to avoid
ffmpeg to broadcast unintentionally.

Signed-off-by: Olivier Langlois <olivier@trillion01.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

hevc: fixing code indenting in hevcpred(cherry picked from commit 28a5916845928de2951013b5398465573b181925)

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

hevc/intrapred: fix indent(cherry picked from commit ab167f3158cf37bd679bda28566170e2c3691686)

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

hevc: a much faster implementation of intra prediction (up to 1s on
haswell 2.6GHz on basket ball drive all intra qp 27)(cherry picked from commit 5de9739176f0eb4c205e80a91628a0196c9924b2)

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

avformat/asfenc: honor the maximum of 63 payloads

Fixes Ticket3634

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

avcodec: add option to make is_intra_more_likely() from error concealment return "no"

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

avformat/mp3dec: check the number of frames per buffer size

Fixes misdetection of issue3327-libc-2.17.so

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

avformat/utils: if the id3 tag is too large for data after it to become available, use the full AVPROBE_SCORE_EXTENSION if extension matches

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

avformat/utils: limit probe score in the id3 case only if there is a chance that data after the id3 tag will become available

Fixes ticket3327 (sample.mp3)

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

Use Matroska document version 2 for WebM files if possible.

Google's plugin for the Internet Explorer refuses to play
files with another document version.

Fixes ticket #3583.

Fix compilation with --enable-lto --enable-shared with icc.

hevcdsp: correctly indicate that hevc_put_hevc_bi_epel_h uses 9 GPRs

Fixes FATE on Windows.

Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avfilter/buffer: Use av_malloc_array()

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

avcodec/buffersrc: use av_mallocz_array()

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

configure: Ensure log2 is disable based on msvcrt version.

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

configure: Allow log2 with msvc12 onwards.

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

avcodec/vda_h264: fix bistream typo

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

Merge commit '1839fafa951308237c4672ffd9a94f927b26eac7'

* commit '1839fafa951308237c4672ffd9a94f927b26eac7':
  avconv: Support VDA hwaccel

Conflicts:
Makefile

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

avcodec/vda_h264: fix null pointer dereference

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

Merge commit '67afcefb35932b420998f6f3fda46c7c85848a3f'

* commit '67afcefb35932b420998f6f3fda46c7c85848a3f':
  lavc: Add new VDA hwaccel

Conflicts:
configure
libavcodec/vda.h
libavcodec/vda_h264.c

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

Merge commit '31a46750c7eafe6bce6a3e4f09853cd0a585292b'

* commit '31a46750c7eafe6bce6a3e4f09853cd0a585292b':
  vda: use hwaccel private data for internal bitstream buffer

Conflicts:
libavcodec/vda.h
libavcodec/vda_h264.c

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

Merge commit '66e6c8a7b4e3ac926302529606bd3e323c28b20f'

* commit '66e6c8a7b4e3ac926302529606bd3e323c28b20f':
  lavu: Add a pixel format for new VDA hwaccel

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

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

Merge commit 'dd2d3b766b20196d0b65a82e3d897ccecbf7adb8'

* commit 'dd2d3b766b20196d0b65a82e3d897ccecbf7adb8':
  lavc: Add hwaccel private data and init/uninit callbacks

Conflicts:
libavcodec/avcodec.h
libavcodec/internal.h
libavcodec/pthread_frame.c

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

Merge commit 'ebc29519d1634bfeb386c20a5d8a52837aae2436'

* commit 'ebc29519d1634bfeb386c20a5d8a52837aae2436':
  hwaccel: Support specific frame allocators

Conflicts:
libavcodec/utils.c

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

Merge commit 'a871ef0cc936612bdb0e643196be72efcf0c8986'

* commit 'a871ef0cc936612bdb0e643196be72efcf0c8986':
  hwaccel: Rename priv_data_size to frame_priv_data_size

Conflicts:
libavcodec/avcodec.h

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

Merge commit '08bb2567586599d62a282c4cd31567a0807f3d92'

* commit '08bb2567586599d62a282c4cd31567a0807f3d92':
  lavc: document which parts of AVHWAccel are private.

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

Merge commit '5c1d7246cd65dc4db1b6dc36e29ce39fc1068f3f'

* commit '5c1d7246cd65dc4db1b6dc36e29ce39fc1068f3f':
  lavc: set AVCodecContext.hwaccel in ff_get_format()

Conflicts:
libavcodec/mpeg12dec.c

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

Merge commit '632ad2248e2e5d8cd4b51e6c87c943a38c3da425'

* commit '632ad2248e2e5d8cd4b51e6c87c943a38c3da425':
  lavc: Add an internal wrapper around get_format()

Conflicts:
libavcodec/h264_slice.c
libavcodec/mpeg12dec.c

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

avcodec/dct: use av_malloc_array()

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

avcodec/cinepakenc: use av_malloc_array()

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

avconv: Support VDA hwaccel

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

lavc: Add new VDA hwaccel

It leverages the new hwaccel 1.2 features:

- get_buffer2 is never called
- the internal context is automatically initialized/deinitialized

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

vda: use hwaccel private data for internal bitstream buffer

lavu: Add a pixel format for new VDA hwaccel

The current hwaccel is broken and cannot be fixed in a compatible
way. It will be deprecated and replaced with a new one.

lavc: Add hwaccel private data and init/uninit callbacks

hwaccel: Support specific frame allocators

It would reduce the boilerplate code users have to write.

hwaccel: Rename priv_data_size to frame_priv_data_size

This describes more accurately what this field is for.

lavc: document which parts of AVHWAccel are private.

lavc: set AVCodecContext.hwaccel in ff_get_format()

This way each decoder does not have to do the same thing manually.

lavc: Add an internal wrapper around get_format()

It will be useful in the following commits.

configure: Fix floating point conformance ith icl.

Reviewed-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc/on2avc: silent warnings about const params

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

avformat/mpegtsenc: fix minor typo in omit_video_pes_length description

Signed-off-by: Aman Gupta <ffmpeg@tmm1.net>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libmpcodecs/vf_uspp: check avcodec_open() return code

Fixes CID1135747

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

ffserver: don't leak poll_table if poll fails

Should fix Coverity Scan issue #732265

Signed-off-by: Reynaldo H. Verdejo Pinochet <r.verdejo@sisa.samsung.com>

Add missing "const" to pointer arrays.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

avfilter/vf_removelogo: fix memleak on failure

Fixes CID751770

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

hevcdsp: include stddef.h for ptrdiff_t definition

Including stdint.h was enough for systems like Mingw, but apparently not for Linux.
This should fix make checkheaders failures on every platform

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

opencl: Make init_mtx static with volatile lock

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

hevcdsp: add missing header include

Fixes make checkheaders

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

avfilter/vf_lut3d: check size in parse_dat()

Fixes CID1212261
Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/movenc: Check ff_put_wav_header() return code

Fixes CID1212260
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Document smoothstreaming muxer

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

avcodec: better level/index printing

avcodec: replace uses of deprecated avcodec_set_dimensions()

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

libmpcodecs: Fix compilation due to missing static in suncc.

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

Merge commit '9880a0d4b131ef36694d62f78060350a81f08b80'

* commit '9880a0d4b131ef36694d62f78060350a81f08b80':
  pcm-dvd: Fix 20bit decoding

See: 5db49fc38d9132e134de92584f296559bec3b789
Merged-by: Michael Niedermayer <michaelni@gmx.at>

pcm-dvd: Fix 20bit decoding

Increment the pointer as needed.

Bug-Id: 592

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

rtpdec: make the NTP time values unsigned.

As per RFC3550, section 4, the NTP time is provided as 64-bit unsigned
integer, so follow the same logic here.

Reviewed-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

configure: check for recent dxva2api headers with fixed COBJMACROS defines

This fixes build failures on older mingw chains (before 2012).

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

avcodec/hevc: Fix undefined shifts

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

avcodec/mpegaudiodec_template: make shift unsigned to avoid undefined behavior

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

avcodec/x86/hevcdsp_init: fix build failure with --disable-mmx

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

avcodec/mpeg4videodec: print run/level/index values

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

avcodec: include GET_RL_VLC() in trace output

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

hvcodec/x86/hevcdsp: make macros more modular to support functions that are not sse4

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

avfilter/phase: make the main loop readable.

avfilter/phase: add some constness for RO data pointers.

avfilter/phase: add timeline support.

avfilter/phase: avoid a memcpy per frame.

avfilter/phase: do not pass the whole context to analyze_plane().

Also make sure the AVFilterContext is only being used for logging by
changing the type to void*

avfilter/phase: add comment on the context cached frame.

Merge commit '0bacfa8d37710b904897e7cbeb8d6f96fbf75e2e'

* commit '0bacfa8d37710b904897e7cbeb8d6f96fbf75e2e':
  rtmpproto: Check the buffer sizes when copying app/playpath strings

Conflicts:
libavformat/rtmpproto.c

See: 0c2a6dabced53b4df57aedf47f0c1b8233035fed
See: 02ac3398eb52679301028c2fd3ebad1b6261b4da
Merged-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/exr: mark gama option as experimental / unsupported

Suggested-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/exr: Add a gamma flag to exr loader to avoid banding

This is needed to avoid banding artifacts when gammaing the picture.
Currently, if done with a video filter, the process is done on uints
instead of full float.

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

avformat/rtmpproto: add FFMIN() to av_strlcpy()

Suggested-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

rtmpproto: Check the buffer sizes when copying app/playpath strings

As pointed out by Reimar Döffinger.

CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>

avcodec/mjpegdec: zero gb to silence warning about it being possibly uninitialized

The code is not speed relevant, also its more robust if the pointers are NULL instead of random.

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