lavfi/vf_select: remove looping on request_frame().

lavfi/vf_fps: remove looping on request_frame().

lavfi/vf_alphamerge: remove looping on request_frame().

lavfi/avf_showwaves: remove looping on request_frame().

lavfi/avf_showspectrum: remove looping on request_frame().

lavfi/avf_showcqt: remove looping on request_frame().

lavfi/af_aresample: remove looping on request_frame().

swresample/resample: manually unroll the main loop in bessel()

About 10% faster

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

swresample/resample: merge first iteration into init in bessel()

speedup of about 1%

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

avcodec/mips: build fix for MSA 64bit

Modified datatype of function argument (pitch from int32_t to ptrdiff_t)

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mips: build fix for MSA

Modified sps and pps access from old HEVCContext(s) structure to newly introduced HEVCParamSets(ps)

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/encoders: Fix libkvazaar documentation

The -threads option is ignored with libkvazaar since it does not have
any of the AV_CODEC_CAP_{FRAME,SLICE,AUTO}_THREADS capabilities. This
commit removes the incorrect documentation as well as the no-op of
setting the number of threads in libkvazaar encoder.

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Fix setting framerate

The divisor and dividend in the equation had been swapped, making the
result the inverse of the actual framerate.

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Use av_image_copy for copying pixels

Replaces a for loop for copying pixels by a call to av_image_copy.

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Set pts and dts

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Replace asserts with proper errors

Changes function libkvazaar_encode to return proper error codes instead
of crashing when the video dimensions or pixel format change in the
middle of encoding.

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Remove unnecessary NULL checks

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

configure: Add version check for libkvazaar

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

libkvazaar: Update to work with the latest version

Function encoder_encode in Kvazaar API was changed to have new output
parameters: source picture and frame info. Frame info is used to set the
keyframe flag and source picture is ignored.

Signed-off-by: Arttu Ylä-Outinen <arttu.yla-outinen@tut.fi>

vp9: don't keep a stack pointer if we don't need it.

This saves one register in a few cases on 32bit builds with unaligned
stack (e.g. MSVC), making the code slightly easier to maintain.

(Can someone please test this on 32bit+msvc and confirm make fate-vp9
and tests/checkasm/checkasm still work after this patch?)

avcodec/libx264: silence -Waddress

This patch moves the pointer validity check outside the macro,
and silences the -Waddress observed with GCC 5.2.

Note that this changes the error message slightly, from:
"bad option..." to "Error parsing option...".

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>

doc/developer: use https instead of http

Change to https for FFmpeg websites.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/pngenc: Check that there is at least 1 frame

Fixes null pointer dereference
Fixes CID1322330

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

avcodec/pngenc: Initialize fctl_chunk to 0

The structure is copied around and that triggers warnings if it is uninitialized
Fixes CID1322360

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

avformat/flvdec: Print stream type in case a new stream is discovered after the header

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

checkasm: add alacdsp tests

Signed-off-by: James Almer <jamrial@gmail.com>

x86/alacdsp: add simd optimized functions

Signed-off-by: James Almer <jamrial@gmail.com>

avformat/httpauth: Add space after commas in HTTP/RTSP auth header

This fixes access to Grandstream cameras, which return 401 to ffmpeg
otherwise.
VLC sends Authorization: header with spaces between parameters, and it
is known to work with Grandstream devices and broad range of other HTTP
and RTSP servers, so author considers switching to such behaviour safe.
Just for record - RFC 2617 (HTTP Auth) does not specify the need in
spaces, so this is not a bug of FFmpeg.

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

avfilter/vf_stereo3d: add x86 SIMD for anaglyph outputs

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

README: replace http with https

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

afilter/af_extrastereo: remove dead initialization

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

avfilter:audio: fix shadowed AVFrame *out

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

avfilter/af_extrastereo: Fix shadowed variable

Fixes CID1325675

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

avformat/flvdec: Remove dead loop

Fixes CID1325682

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

vp9: fix msvc build by using 6 GPRs on 32bit if stack!=aligned.

avformat/omadec: Fix { typo

Fixes CID1324299

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

ffplay: use correct context for av_log

Recent commits c3e8de1c248f8c742dd9e61a0c71ee56bba22c28 and
8dc6e92c3dc67a85026f3010045c7a28b1c0adc8 used av_log incorrectly.
This fixes such usage.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

ffplay: more robust condition variable creation

SDL_CreateCond can fail:
https://wiki.libsdl.org/SDL_CreateCond.
This patch makes creation more robust in one instance.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

ffplay: log SDL error messages

This logs the SDL error messages on failure of creation of SDL_CreateMutex,
SDL_CreateCond, and SDL_CreateThread.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

avutil/attributes: add av_warn_unused_result

GCC 3.4 introduced an attribute warn_unused_result to warn when a programmer
discards the return value. Applying this judiciously across the codebase can help
in fixing a lot of problems. At a high level, functions which return error codes
should always be checked. More concretely, consider the functions ff_add_format
and the like in avfilter/formats.h. A quick examination shows that a large portion
of libavfilter fails to handle the associated errors, usually AVERROR(ENOMEM).
The above example was where I observed the utility of this, but it should be
useful in many places across the code base.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

dnxhd: add better support for CIDs 1270 to 1274

These are DNxHR profiles with the following properties:
- Variable size in a profile (property added in a previous commit),
  requiring variable-sized macroblock table;
- Variable bitdepth, up to 12 bits.
- Better validation of buffer sizes and positions

Signed-off-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

dnxhd: add CID 1270

This a 4:4:4 10 bits profile, where image size is not fixed by the
profile, and which strays a bit outside the old frame header parsing
code.

Fixes ticket #4581 (DNxHR is not stricly supported, but that sequence is).

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

dnxhd: add decoder support for DNxHR

Signed-off-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

isom: add support for DNxHR codec family

This is actually similar to DNxHD.

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

dnxhddec: Introduce DNXHD_VARIABLE

Currently not used, but will be used to indicate that a CIDEntry field
is not set, because it is variable, and that checks should be adapted.

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

dnxhdenc: do not select 4:4:4 profiles

The encoder can only deal with 4:2:2.

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

dnxhd: profile flags

Move the 'interlaced' flag to this element (arbitrarily set to 16bits).
This should allow better detection/selection of profiles.

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

dnxhddec: cleanup frame header parsing

Rely more on the actual syntax from the specs (also seen in the
encoder code).

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

avcodec/mmaldec: Use av_assert0() instead of assert()

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

Merge commit '39f01e346cab464ef6c0d4ec58cc13b7123e60d8'

* commit '39f01e346cab464ef6c0d4ec58cc13b7123e60d8':
  mmaldec: be more tolerant against MMAL not returning decoded output

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '65db4899fa8790049bec3af16ecdb75dd81051fd'

* commit '65db4899fa8790049bec3af16ecdb75dd81051fd':
  mmaldec: refactor to have more context per MMAL input buffer

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'eae58428bd199f7d4670bf471f56ed204430878e'

* commit 'eae58428bd199f7d4670bf471f56ed204430878e':
  avcodec: Do not lock during open for codecs marked as having threadsafe init

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'abe9adfb31566c415fd830a8d4977c79512d4385'

* commit 'abe9adfb31566c415fd830a8d4977c79512d4385':
  rangecoder: Use AV_RB16 instead of bytestream_get_be16

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'b8b13acd70a1c1b6c044f967d78cb82f1a53887b'

* commit 'b8b13acd70a1c1b6c044f967d78cb82f1a53887b':
  hevc: Simplify logical check

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '10aa6f9db44e96977e7155fdc3e62619e0e00a15'

* commit '10aa6f9db44e96977e7155fdc3e62619e0e00a15':
  h264: Use the correct argument context in debug log

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '4628443ca3534060888dd0015b229337eac13fd2'

* commit '4628443ca3534060888dd0015b229337eac13fd2':
  h263: Drop uninitialized variable use from log message

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '1da2a20763ae9ca579d5fd20763065871ddf6311'

* commit '1da2a20763ae9ca579d5fd20763065871ddf6311':
  dvbsubdec: Fix function return type

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'cab63a8b594cdc365bb2581a12b3ac8e6dd480b2'

* commit 'cab63a8b594cdc365bb2581a12b3ac8e6dd480b2':
  dv: Mark internal frame reference as const

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '4c160fa23996c05efcd952ccfac2359311d8a1bc'

* commit '4c160fa23996c05efcd952ccfac2359311d8a1bc':
  dcadec: Always initialize return variable

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'bb198c4997d5036f3bf91de51e44f807115677d0'

* commit 'bb198c4997d5036f3bf91de51e44f807115677d0':
  d3d11va: make av_d3d11va_alloc_context() available at all times

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'f3202871598f59b570b31b01cfeb64b8fedbd700'

* commit 'f3202871598f59b570b31b01cfeb64b8fedbd700':
  d3d11va: check for malloc failure

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '2c3dbff1d89bbf8300c121c06524c014cb6e0915'

* commit '2c3dbff1d89bbf8300c121c06524c014cb6e0915':
  avutil/avstring: Inline some tiny functions

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'f56a08559334b7eb6b3fedbc0cc741887f6067ae'

* commit 'f56a08559334b7eb6b3fedbc0cc741887f6067ae':
  matroskaenc: Don't write a track language tag

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '8bb376cf6b4ab8645daedb8becaa7163656436a4'

* commit '8bb376cf6b4ab8645daedb8becaa7163656436a4':
  checkasm: Fix the function name sorting algorithm

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

ffserver: avoid leaking pathname at exit

Fixes Coverity CID 1325681

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: fix up rate units in server messages

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: wrap around bandwidth exceeded msg text

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: wrap around too-busy message text

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: unify exit path in _write_index()

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: fix line wrapping on function decls

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffplay: more robust mutex creation

SDL_CreateMutex can fail:
https://wiki.libsdl.org/SDL_CreateMutex.
This patch makes creation more robust in one instance.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

concatdec: fix metadata memleak on error

Fixes Coverity CID 1323077.

Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Marton Balint <cus@passwd.hu>

libavformat/tls_securetransport: fix argument evalulation order UB

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

avcodec/tta: Un-break build without threads

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

avfilter: add displace video filter

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

mmaldec: be more tolerant against MMAL not returning decoded output

In some situations, MMAL won't return a decoded frame for certain input
frames. This can happen if a frame fails to decode, or if a packet does
not actually contain a complete frame. In these situations, we would
deadlock (or actually timeout) waiting for an expected output frame,
which is not ideal. On the other hand, there are situations where we
definitely have to block to avoid deadlocks. (This mess is a
consequence of trying to map MMAL's asynchronous and flexible
dataflow to libavcodec, which is more static and rigid.)

Solve this by doing a blocking wait only if the amount of buffered data
is too big. The whole purpose of the blocking wait is to avoid excessive
buffering of input data, so we can skip it if it appears to be low. The
consequence is that libavcodec can gracefully return no frame to the
API user.

We want to track the number of full packets to make our heuristic work.
But MMAL buffers are fixed-size, requiring splitting large packets. This
is why the previous commit is needed. We use the ..._FRAME_END flag to
remember packet boundaries, but MMAL does not preserve these buffer
flags when returning buffers to the user.

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

mmaldec: refactor to have more context per MMAL input buffer

The next commit needs 1 bit of additional information per MMAL buffer
sent to the MMAL input port. This information will be needed when the
buffer is recycled (i.e. returned by the input port's callback).
Normally, we could use MMAL_BUFFER_HEADER_FLAG_USER0, but that is
unexpectedly not preserved.

Do this by storing a pointer to FFBufferEntry in the MMAL buffer's
user data, instead of an AVBufferRef. This also changes the lifetime
of FFBufferEntry.

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

avcodec/alacdec: split off decorrelate_stereo and append_extra_bits as alacdsp

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

avcodec: use HAVE_THREADS header guards to silence -Wunused-function

When compiled with --disable-pthreads, e.g
http://fate.ffmpeg.org/report.cgi?time=20150917015044&slot=alpha-debian-qemu-gcc-4.7,
a bunch of -Wunused-functions are reported due to missing header guards
around threading related functions.
This patch should silence such warnings.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>

all: add _DEFAULT_SOURCE locally wherever needed

Glibc 2.20 onwards generates a deprecation warning for usage of _BSD_SOURCE and _SVID_SOURCE.
The solution from man feature_test_macros is to define both _DEFAULT_SOURCE and the old macros.
This solution is on the lines of the one in commit af1818276ef271af98e2e2bbabb4dc875b4fa7d8.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

ffplay: more robust thread creation

SDL_CreateThread can fail:
https://wiki.libsdl.org/SDL_CreateThread.
This patch makes thread creation more robust in one instance.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

Changelog: add note on ffplay dynamic volume control

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

checkasm: Fix compilation with --disable-avcodec

avformat/flacdec: support fast-seek

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

avcodec: Do not lock during open for codecs marked as having threadsafe init

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avfilter/ebur128: use AV_OPT_TYPE_BOOL for video option

avfilter/ebur128: add dualmono measurement option

rangecoder: Use AV_RB16 instead of bytestream_get_be16

Silence an incompatible-pointer-types-discards-qualifiers warning from clang.
rangecoder.c:58:34: warning: passing
      'uint8_t **' (aka 'unsigned char **') to parameter of type
      'const uint8_t **' (aka 'const unsigned char **') discards qualifiers in
      nested pointer types [-Wincompatible-pointer-types-discards-qualifiers]

hevc: Simplify logical check

The intended meaning is "if this block is the first block in a slice then
its left boundary is a slice boundary". Silence a logical-not-parentheses
warning from gcc.

h264: Use the correct argument context in debug log

h263: Drop uninitialized variable use from log message

dvbsubdec: Fix function return type

dv: Mark internal frame reference as const

Silence a warning due to frame assignment in dvenc. All uses of the
reference in dvdec are read only, except the ones in the main decoding
function, so use the frame pointer directly there.

dcadec: Always initialize return variable

Silence an uninitialized warning from clang.

lavf: add chromaprint muxer

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

libavformat/hlsenc: Use of uninitialized memory unlinking old files

Fixes ticket#4900

Signed-off-by: DHE <git@dehacked.net>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

libx264: copy A53 closed captions from source

Assumes 'GA94' format (ATSC standard)

Signed-off-by: DHE <git@dehacked.net>
Tested-by: Anshul <anshul.ffmpeg@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

blockdsp: reindent after parameter removal

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

avcodec/apedec: fix bug introduced in commit d3e5fbb1406995e07fccbff3ca8c1e24f57a1f7b

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

lavf/mlpdec: Simplify mlp/thd probe function.

Move the demuxers into their own file.