Jan Ekström [Fri, 24 Nov 2017 00:44:44 +0000 (02:44 +0200)]
hwcontext_d3d11va: properly reset values after release/close
Makes the uninit function re-entrable, which can be a common case
when an API user first tries to initialize its context, fails, and
then finally unrefs the AVHWDevice.
Mikulas Patocka [Thu, 23 Nov 2017 19:12:55 +0000 (20:12 +0100)]
avcodec/fft: fix INTERL macro on 3dnow
The commit b7c16a3f2c4921f613319938b8ee0e3d6fa83e8d ("x86: fft: Port to
cpuflags") breaks the opus decoder in ffmpeg when compiling for 3dnow. The
output is audible, but there's a lot of noise.
The reason for the breakage is that the commit unintentionally changed the
INTERL macro so that it is empty when compiling for 3dnow. This patch
fixes it.
Signed-off-by: Mikulas Patocka <mikulas@twibright.com> Signed-off-by: James Almer <jamrial@gmail.com>
Mark Thompson [Fri, 24 Nov 2017 12:36:09 +0000 (12:36 +0000)]
lavu/pixfmt: Remove gap in the middle of enum AVPixelFormat
This was added for compatibility with libav, by leaving a space for
formats added in libav to be merged. Since that feature has been
removed, we don't need a gap here.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Mikulas Patocka [Thu, 23 Nov 2017 19:22:56 +0000 (20:22 +0100)]
ffmpeg libopusdec: fix missing include file in libopusdec.c
This patch fixes the following error when compiling mplayer with libopus.
libavcodec/libopusdec.c: In function 'libopus_decode_init':
libavcodec/libopusdec.c:130:27: error: implicit declaration of function 'ff_exp10'; did you mean 'ff_exp2fi'? [-Werror=implicit-function-declaration]
double gain_lin = ff_exp10(gain_db / (20.0 * 256));
^~~~~~~~
ff_exp2fi
Signed-off-by: Mikulas Patocka <mikulas@twibright.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: signed integer overflow: 2147483598 + 64 cannot be represented in type 'int' Fixes: 4337/clusterfuzz-testcase-minimized-6192658616680448 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Philip Langdale [Fri, 24 Nov 2017 18:05:49 +0000 (10:05 -0800)]
avcodec/nvdec: Round up odd width/height values
nvdec will not produce odd width/height output, and while this is
basically never an issue with most codecs, due to internal alignment
requirements, you can get odd sized jpegs.
If an odd-sized jpeg is encountered, nvdec will actually round down
internally and produce output that is slightly smaller. This isn't
the end of the world, as long as you know the output size doesn't
match the original image resolution.
However, with an hwaccel, we don't know. The decoder controls
the reported output size and the hwaccel cannot change it. I was
able to trigger an error in mpv where it tries to copy the output
surface as part of rendering and triggers a cuda error because
cuda knows the output frame is smaller than expected.
To fix this, we can round up the configured width/height passed
to nvdec so that the frames are always at least as large as the
decoder's reported size, and data can be copied out safely.
In this particular jpeg case, you end up with a blank (green) line
at the bottom due to nvdec refusing to decode the last line, but
the behaviour matches cuviddec, so it's as good as you're going to
get.
Li, Zhong [Fri, 10 Nov 2017 08:03:42 +0000 (03:03 -0500)]
qsv/hevcdec: Load hw plugin by default on non-windows os
Software plugin is not available on Linux, Only works on Windows.
Similar changes have been applied to qsv hevc encoder by b05128f3c953bd66483e697d60a2e7e45ee9cfa0.
Signed-off-by: Zhong Li <zhong.li@intel.com> Signed-off-by: Maxym Dmytrychenko <maxim.d33@gmail.com>
Mark Thompson [Tue, 14 Nov 2017 19:47:30 +0000 (19:47 +0000)]
lavu,lavfi,ffmpeg: Remove experimental OpenCL API
This was added in early 2013 and abandoned several months later; as far as
I can tell, there are no external users. Future OpenCL use will be via
hwcontext, which requires neither special OpenCL-only API nor global state
in libavutil.
All internal users are also deleted - this is just the unsharp filter
(replaced by unsharp_opencl, which is more flexible) and the deshake filter
(no replacement).
Mark Thompson [Tue, 14 Nov 2017 19:47:27 +0000 (19:47 +0000)]
lavfi: Add OpenCL unsharp mask filter
Intended to replace existing opencl mode of the unsharp filter.
Supports many more pixel formats and works without immediate upload
and download of frame data. The options are compatible with the
existing filter.
Mark Thompson [Sun, 10 Sep 2017 12:37:20 +0000 (13:37 +0100)]
hwcontext_opencl: DRM to OpenCL mapping for ARM
Using cl_arm_import_memory. Unfortunately, despite this not being a
standard extension, the function clImportMemoryARM() is not accessible
via clGetExtensionFunctionAddressForPlatform(). This means that it has
to be linked directly to the ARM OpenCL binary, so making a portable
binary is not possible as it is with all other mapping extensions.
Jim DeLaHunt [Wed, 15 Nov 2017 08:43:30 +0000 (00:43 -0800)]
doc: Add FAQs about running in background (rev 2)
Add two FAQs about running FFmpeg in the background.
The first explains the use of the -nostdin option in
a straightforward way. Text revised based on review.
The second FAQ starts from a confusing error message,
and leads to the solution, use of the -nostdin option.
The purpose of the second FAQ is to attract web searches
from people having the problem, and offer them a solution.
Add an anchor to the Main Options section of the ffmpeg
documentation, so that the FAQs can link directly there.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Dale Curtis [Tue, 21 Nov 2017 23:40:22 +0000 (15:40 -0800)]
avformat/mov: Increment stsd_count while processing stsd data; avoids leaks.
In the event of ff_mov_read_stsd_entries() failure, sc->stsd_count
is not updated, even if the function allocates extradata memory.
Instead update the sc->stsd_count as entries are parsed so that
mov_read_close() can do the right thing.
Signed-off-by: Dale Curtis <dalecurtis@chromium.org> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/jpeg2000: Dynamically allocate codeblock data
Fixes: OOM Fixes: 3541/clusterfuzz-testcase-minimized-6469958596820992
Adds support for decoding codeblock data larger than 8kb
Reduces decoder memory consumption
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Dale Curtis [Sat, 18 Nov 2017 00:05:30 +0000 (16:05 -0800)]
Fix undefined shift on assumed 8-bit input.
decode_user_data() attempts to create an integer |build|
value with 8 bits of spacing for 3 components. However
each component is an int32_t, so shifting each component
is undefined for values outside of the 8 bit range.
This patch simply clamps input to 8-bits per component
and prints out a warning that the values were clamped.
Signed-off-by: Dale Curtis <dalecurtis@chromium.org> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/hevcdsp_template: Fix invalid shift in put_hevc_epel_bi_w_v()
Fixes: runtime error: left shift of negative value -255 Fixes: 4037/clusterfuzz-testcase-minimized-5290998163832832 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Aman Gupta [Wed, 15 Nov 2017 20:50:31 +0000 (12:50 -0800)]
libavcodec/videotoolbox: fix decoding of h264 streams with minor SPS changes
Previously the codec kept an entire copy of the SPS, and restarted the VT decoder
session whenever it changed. This fixed decoding errors in [1], as
described in 9519983c. On further inspection, that sample features an SPS change
from High/4.0 to High/3.2 while moving from one scene to another.
Yesterday I received [2], which contains minor SPS changes where the
profile and level do not change. These occur frequently and are not associated with
scene changes. After 9519983c, the VT decoder session is recreated unnecessarily when
these are encountered causing visual glitches.
This commit simplifies the state kept in the VTContext to include just the first three
bytes of the SPS, containing the profile and level details. This is populated initially
when the VT decoder session is created, and used to detect changes and force a restart.
This means minor SPS changes are fed directly into the existing decoder, whereas
profile/level changes force the decoder session to be recreated with the new parameters.
After this commit, both samples [1] and [2] playback as expected.