The x86 optimized dct_quantize only calculates the last nonzero
coefficient correctly if the zigzag scan order is used. For the
alternate scan order, this value is incorrect.
To work around this, the dct_unquantize functions process the entire
block if the alternate scan order is used.
But a second workaround (bb198e198a) was added that recalculates the
last nonzero coefficient after dct_quantize is called if the alternate
scan order is used.
This commit removes the first workaround, which became redundant.
It's safe to assume that LOCAL_ALIGNED_16 does indeed align. Otherwise
we would have many more problems...
This assert was added in f8188626 all the way back in 2003.
Fixes:
https://fate.ffmpeg.org/report.cgi?slot=x86_64-archlinux-gcc-tsan&time=20240823175808
Reproduction steps:
./configure --enable-memory-poisoning --toolchain=gcc-tsan --disable-stripping && make fate-vvc
Root cause:
We hold the current frame's lock while updating progress for other frames,
which also requires acquiring other frame locks. This could potentially lead to a deadlock.
However, I don't think this will happen in practice because progress updates are one-way, with no cyclic dependencies.
But we need this patch to make FATE happy.
Implement load/save loop filter and segment parameters defined in
section 7.20 and 7.21 in spec for show_existing_frame frames.
Fixes ticket #11151.
Signed-off-by: Fei Wang <fei.w.wang@intel.com>
If the demuxer does not provide per-stream indexes, the generic seek search can
attempt to read the whole media file from the beginning when seeking. For large
MXF files this can cause huge lockups for a seek after the last timestamp,
which will eventually fail. So let's disable the generic seek for mxf, the
demuxer's own seek code should handle seeking just fine.
Signed-off-by: Marton Balint <cus@passwd.hu>
Recently, I have been experiencing an increasing number of user that use ffmpeg
to retrive RTSP stream from personal mediaproxies (e.g. MediaMtx) with
authorization based on JWT. The current length of PATH does not permit to
insert the token in the URL failing the authorization with no possibilities to
get the video.
VLC has just modified the RSTP max URL length, and it permits to use token
inside the URL.
For these reasons, I propose this patch to extend the PATH buffer from 1024 to
2048 in order to use tokens and the authorization process based on JWT.
Signed-off-by: Marton Balint <cus@passwd.hu>
It differs from query_func() in accepting arrays of input/output format
configurations to be filled as callback parameters. This allows to mark
the filter context as const, ensuring it is not modified by this
function, as it is not supposed to have any side effects beyond
returning the supported formats.
This ensures that if a codec isn't on codec_whitelist, trying to open it
will not trigger ff_codec_close(), which could invalidate useful
information still present in the context.
Signed-off-by: Dale Curtis <dalecurtis@chromium.org>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
The ret value here is not yet intialized so the return would return
uninitialized data. What was probably meant to be checked here was the
return value of ffio_read_size, which can return an error.
Introduced in 38bcb3ba7b
Fixes: CID1618758
Signed-off-by: James Almer <jamrial@gmail.com>
Text was stolen from the linux kernel
This is thus identical to the kernel just a different more compact format.
I am very happy also to switch the file entirely to the format of the linux kernel maintainer list
if people prefer
This allows specifying a git repository for an area
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Text was stolen from the linux kernel
This is thus identical to the kernel just a different more compact format.
I am very happy also to switch the file entirely to the format of the linux kernel maintainer list
if people prefer
This allows specifying more details than what fits on a single line
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Text was stolen from the linux kernel
This is thus identical to the kernel just a different more compact format.
I am very happy also to switch the file entirely to the format of the linux kernel maintainer list
if people prefer
This allows specifying a bug tracker. It would allow a maintainer to
use gitlab, github, or something else to keep track of bugs
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Text was stolen from the linux kernel
This is thus identical to the kernel just a different more compact format.
I am very happy also to switch the file entirely to the format of the linux kernel maintainer list
if people prefer
This allows specifying a webpage for an area
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Text was stolen from the linux kernel
This is thus identical to the kernel just a different more compact format.
I am very happy also to switch the file entirely to the format of the linux kernel maintainer list
if people prefer
This allows specifying a more specific mailing list, if a specific area has
such a list.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
The output of put_bits_left() here can be as big as 27, which is a bit
count not supported by get_bits().
Fixes fate-iamf-stereo-demux when using --assert-level=2
Signed-off-by: James Almer <jamrial@gmail.com>
Use ff_mp4_read_descr() to read both the tags and the vlc value
that comes after it, which was not being taken into account.
Ref: https://github.com/AOMediaCodec/libiamf/issues/119
Signed-off-by: James Almer <jamrial@gmail.com>
Implement the missing requirements from H.266 (V3) p. 106 on the
position and size of subpictures whose dimensions are provided
explicitly.
Signed-off-by: Frank Plowman <post@frankplowman.com>
The previous logic relied on the subpicture boundaries coinciding with
the tile boundaries. Per 6.3.1 of H.266 (V3), vertical subpicture
boundaries are always tile boundaries however the same cannot be said
for horizontal subpicture boundaries. Furthermore, it is possible to
construct an illegal bitstream where vertical subpicture boundaries are
not coincident with tile boundaries. In these cases, the condition of
the while loop would never be satisfied resulting in an OOB read on
col_bd/row_bd.
Patch fixes this issue by replacing != with <, thereby not requiring
subpicture boundaries and tile boundaries to be coincident.
Signed-off-by: Frank Plowman <post@frankplowman.com>
fix
==135000== Conditional jump or move depends on uninitialised value(s)
==135000== at 0x169FF95: vvc_deblock_bs (filter.c:699)
and
==135000== Conditional jump or move depends on uninitialised value(s)
==135000== at 0x16A2E72: ff_vvc_alf_filter (filter.c:1217)
Reported-by: James Almer <jamrial@gmail.com>
