Anton Khirnov [Sun, 12 Apr 2015 17:08:23 +0000 (19:08 +0200)]
h264: copy nal_length_size in update_thread_context
It does not make sense to copy is_avc without copying this as well. This
patch should not change anything for now, but will have an effect in
later commits.
Anton Khirnov [Mon, 6 Apr 2015 18:33:58 +0000 (20:33 +0200)]
h264: eliminate ff_h264_set_parameter_from_sps()
That function currently does two things -- reinitializing the DSP
contexts and setting low_delay based on the SPS values.
The former more appropriately belongs in h264_slice_header_init(), while
the latter only really makes sense in decode_slice_header().
The third call to ff_h264_set_parameter_from_sps(), done immediately
after parsing a new SPS, appears to serve no useful purpose, so it is
just dropped.
Also, drop now unneeded H264Context.cur_chroma_format_idc.
Anton Khirnov [Sun, 5 Apr 2015 18:51:00 +0000 (20:51 +0200)]
h264: move the DPB init/uninit to init_context()/free_context()
Currently, the DPB is initialized in alloc_tables() and uninitialized in
free_tables(), but those functions manage frame size-dependent
variables, so DPB management does not logically belong in there.
Since we want the init/uninit to happen exactly once per the context
lifetime, init_context()/free_context() are the proper place for this
code.
avresample: Reallocate the internal buffer to the correct size
Fixes the corner case in which the internal buffer size
is larger than input buffer provided and resizing it
before moving the left over samples would make it write
to now unallocated memory.
nutdec: fix illegal count check in decode_main_header
The existing check has two problems:
1) i + count can overflow, so that the check '< 256' returns true.
2) In the (i == 'N') case occurs a j-- so that the loop runs once more.
This can trigger the assertion 'nut->header_len[0] == 0' or cause
segmentation faults or infinite hangs.
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
John Högberg [Tue, 28 Apr 2015 08:20:33 +0000 (10:20 +0200)]
mpegts: Update the PSI/SI table only if the version change
If a PAT is finished while a PMT section filter is opened but
not yet finished, the PMT section filter is closed and all
the received data is discarded.
This is usually not an issue but some multiplexers (With very
quick PAT/PMT repetition settings) consistently emit a PMT
section start, then a PAT, and then the rest of the PMT,
causing the aforementioned behavior to result in no PMT being
finished.
In the most pathologic situation the stream information are lost
and the probe fallback miscategorizes subtitles as mp3 audio.
Avoid the issue through eliminating redundant PSI/SI table
updates by checking their version field, which is required by
the standard to be incremented on every change no matter how
minor.
Check extended sync word for 16-bit LE and BE core streams to reduce
probability of alias sync detection. Previously sync word extension was
checked only for 14-bit streams.
This follows up the similar change in avcodec/dca_parser.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
apedec: prevent out of array writes in decode_array_0000
s->decoded_buffer is allocated with a min_size of:
2 * FFALIGN(blockstodecode, 8) * sizeof(*s->decoded_buffer)
Then it is assigned to s->decoded[0] (and s->decoded_buffer + FFALIGN(blockstodecode, 8)
to s->decoded[1]) and passed as out buffer to decode_array_0000.
In this function 64 elements of the out buffer are written
unconditionally and outside the array if blockstodecode is too small.
This causes memory corruption, leading to segmentation faults or other
crashes.
Thus change decode_array_0000 to write at most blockstodecode elements
of the out buffer.
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
AVOpenCLDeviceNode and AVOpenCLPlatformNode used fixed static buffer for holding the device and platform name.
This patch modifies these structures to use pointers instead. The memory required to hold the names is
now dynamically allocated, the size for which is determined by querying appropriate OpenCL runtime APIs.
Signed-off-by: Maneesh Gupta <maneesh.gupta@amd.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
OpenCL: Avoid potential buffer overflow in cmdutils_opencl.c
The opt_opencl_bench function copied the device name using strcpy without checking if the source string was larger.
This patch fixes this by replacing the strcpy with av_strlcpy, with the string copy size capped to the destination buffer size.
Signed-off-by: Maneesh Gupta <maneesh.gupta@amd.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Remove the direct profile from UTCTiming element. Per DASH spec,
direct profile value should be the time at which the request was
made to the server and not the time at which the manifest was
written. So ffmpeg cannot write this value. This patch removes
the direct profile and write the UTCTiming element with the http
profile only if a URL is passed as a parameter. Update the fate
test to reflect this change.
Signed-off-by: Vignesh Venkatasubramanian <vigneshv@google.com> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
tests: make sure subtitles tests are run with a rawdiff
This will test properly CRLF with make fate, make fate-subtitles and any
make fate-sub-* test. Before this commit, the rawdiff was triggered only
by make fate-subtitles.
Also make sure fate-sub-* only match the tests relying on fmtstdout
command, to at least avoid failing on MingW. See
https://ffmpeg.org/pipermail/ffmpeg-devel/2015-April/172395.html
Rong Yan [Mon, 27 Apr 2015 03:29:51 +0000 (03:29 +0000)]
avcodec/ppc/fdctdsp: POWER LE support in ff_fdct_altivec() delete macros VEC_FMERGEH() VEC_FMERGEL(), they where wrong
GCC tool had a bug of PPC intrinsic interpret, which has been fixed in GCC 4.9.1. This bug lead to
errors in two of our previous patches. We found this when we update our GCC tools to 4.9.1 and by
reading the related info on GCC website. We fix our previous error in two separate commits
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Rong Yan [Mon, 27 Apr 2015 03:19:45 +0000 (03:19 +0000)]
swscale/ppc/swscale_altivec.c: POWER LE support in yuv2planeX_8() delete macro GET_VF() it was wrong
GCC tool had a bug of PPC intrinsic interpret, which has been fixed in GCC 4.9.1. This bug lead to
errors in two of our previous patches. We found this when we update our GCC tools to 4.9.1 and by
reading the related info on GCC website. We fix our previous error in two separate commits
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
HAVE_LOONGSON is replaced by HAVE_LOONGSON3. Even Loongson-2E and 2F support
Loongson SIMD instructs but have low performance for decoding. We plan to focus
on optimizing Loongson-3A1000, 3B1500 and 3A1500, and modify the configure file
to support Loongson-2 series later by adding HAVE_LOONGSON2.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
avcodec/movtextdec: Decoding of Bold-Italic_Underlined styles for 3gpp timed text subtitles
Signed-off-by: Niklesh <niklesh.lalwani@iitb.ac.in>
Previous version reviewed-by: Philip Langdale <philipl@overt.org>
Previous version reviewed-by: Carl Eugen Hoyos <cehoyos@ag.or.at> Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Ronald S. Bultje [Thu, 23 Apr 2015 00:14:19 +0000 (20:14 -0400)]
vp9: use aligned size to write segmentation map into cache.
The unaligned size is not handled in setctx_2d(), causing edges of images
to have improper segmentation prediction, which causes visual artifacts
at image edges a few frames later.
Ronald S. Bultje [Wed, 22 Apr 2015 16:24:41 +0000 (12:24 -0400)]
vp9: add fate tests for show-existing-frame feature.
Samples available at:
http://downloads.webmproject.org/test_data/libvpx/vp90-2-10-show-existing-frame.webm
http://downloads.webmproject.org/test_data/libvpx/vp90-2-10-show-existing-frame2.webm
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>