The flag was and is intended to be a sane default and thus does not enable
time consuming checks. This also matches how the flag is used in decoders and
demuxers.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Don't use word-size multiplications if size == 2, and if we're using
SIMD instructions (size >= 8), complete leftover 4byte sets using movd,
not mov. Both of these changes lead to minor speedups.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
signed integer overflow is undefined in C
Fixes the following gcc warnings:
In file included from libavcodec/hevcpred.c:27:0:
libavcodec/hevcpred_template.c: In function ‘intra_pred_8’:
libavcodec/hevcpred_template.c:302:9: warning: assuming signed overflow does not occur when assuming that (X - c) > X is always false [-Wstrict-overflow]
In file included from libavcodec/hevcpred.c:31:0:
libavcodec/hevcpred_template.c: In function ‘intra_pred_9’:
libavcodec/hevcpred_template.c:302:9: warning: assuming signed overflow does not occur when assuming that (X - c) > X is always false [-Wstrict-overflow]
In file included from libavcodec/hevcpred.c:35:0:
libavcodec/hevcpred_template.c: In function ‘intra_pred_10’:
libavcodec/hevcpred_template.c:302:9: warning: assuming signed overflow does not occur when assuming that (X - c) > X is always false [-Wstrict-overflow]
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'afa93d198aaf2cc661c4df6d4095cd030265d30a':
hevc_parser: Set pict_type, key_frame and output_picture_number.
hevc: Search start code in decode_nal_units().
Merged-by: Michael Niedermayer <michaelni@gmx.at>
this might fix Ticket2999 as well as some fate clients
untested as the original patch submitter no longer has the environment to test
this should be reverted if it does not fix the issues
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Do not consider it an error if we have no frames and should discard one.
This condition can easily happen when decoding is started from an I frame
Fixes Ticket2811
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Should fix crashes or corrupt output on pre-SSE2 CPUs when they were
using SSE2-code (e.g. AMD Athlon XP 2400+ or Intel Pentium III) in
hfix or hvar single-edge (left/right) extension functions.
Tested-by: Ingo Brückl <ib@wupperonline.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Fixes out of array read
This issue can be reproduced with:
Lucky.asf
teuflisch.asf
from http://samples.ffmpeg.org/A-codecs/VoxWare/
as well as other samples
fcb16m has 320 elements but the code before the patch can address a
table of 384 elements in its place.
I dont know if theres some elements missing, or its maybe the wrong
table, or if the parameters affecting the table size are wrong, or
if theres something else wrong.
if someone reading this has easy access to some reference. (binary
with the tables in it, spec, whatever, please contact us)
This solution fixes the inconsistency by padding the table with zeros,
this variant was suggested by Vitor Sessak. And ensures that nothing
currently working breaks.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
That makes avformat & avcodec thread safe without the need to explicitly
register a lock manager.
Reviewed-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
