The most interesting parts are initialization in ff_MPV_common_init() and
uninitialization in ff_MPV_common_end().
ff_mpeg_unref_picture and ff_thread_release_buffer have additional NULL
checks for Picture.f, because these functions can be called on
uninitialized or partially initialized Pictures.
NULL pointer checks are added to ff_thread_release_buffer() stub function.
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
This limits ABI issues in case libavcodec is linked to a libavutil with larger AVFrame
Which can happen if they are shiped in seperate binary packages and libavutil is upgraded
A cleaner alternative would be to replace them by pointers but this would likely cause
a small speedloss
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This is a bit tricky, we allocate a correctly sized AVFrame but then only
copy the compile time AVFrame size, this is to ensure that user applications
which do not use the correct av frame API dont end with out of array reads.
Note, applications using the correct API have set extended_data and the
changed code will never be executed for them.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Fixes inconsistency and out of array accesses
Fixes: 10cdd7e63e7f66e3e66273939e0863dd-asan_heap-oob_1a4ff32_7078_cov_4056274555_mov_h264_aac__mp4box_frag.mp4
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The default get_buffer2() implementation (and possibly some
user ones) does not allocate edges when this flag is set, which may
expose bugs in some decoders. Until the 10 release is out, it is safer
to remove this part.
this allows seperate installation of shared libs that should not conflict with
whatever is already installed.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This ensures that theres just one AVFrame allocation function and libs dont
produce multiple AVFrame variants after a minor lib update
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Its possible to implement this with a few lines less code but it then
would flip the order of the list and require registration of external
codecs to be done first, also it could break user applications due to
this. Thus to maintain ABI this slighty more complex solution is
used.
Reviewed-by: Stefano Sabatini
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Use only proper AVFrame API (no assigning of whole frames, since that
hardcodes sizeof(AVFrame) into lavc).
Make a copy of the side data, so the caller can use av_frame_unref/free
on non-refcounted frames, eliminating the need for
avcodec_get_frame_defaults()/avcodec_free_frame().
Not just on failure. This is the same thing that is done in the audio
path and should prevent leaks in decoders that allocate a frame, but
then end up not writing into it.
This code changes the input packet, which is read only and can in
rare circumstances lead to decoder errors. (i run into one of these in
the audio decoder, which corrupted the packet during av_find_stream_info()
so that actual decoding that single packet failed later)
Until a better fix is implemented, this commit limits the problem.
A better fix might be to make the subtitle decoders not depend on
data[size] = 0 or to copy their input when this is not the case.
Michael Niedermayer
wm4
Anton Khirnov
John Stebbins
Diego Biurrun
Tim Walker
Nicolas George
Ben Boeckel
Luca Barbato
Stefano Sabatini