Commit Graph

769 Commits

Author SHA1 Message Date
Ramiro Polla
d31dbec374 Rename CHECKED_ALLOC(Z) to FF_ALLOC(Z)_OR_GOTO and add context and label
parameters.

Originally committed as revision 19776 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-09-06 00:08:19 +00:00
Lars Täuber
35f6c15405 Make 16bit YUV formats compatible with NE avcodec_get_pix_fmt().
Patch by Lars Täuber <lars <dot> taeuber <at> gmx <dot> net>.

Originally committed as revision 19775 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-09-05 23:33:25 +00:00
Ramiro Polla
5ce6934e0a Add a newline in error message in CHECKED_ALLOC(Z).
Originally committed as revision 19758 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-09-04 16:59:28 +00:00
Diego Biurrun
ee1550116b whitespace cosmetics, prettyprinting
Originally committed as revision 19751 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-09-01 16:00:45 +00:00
Thilo Borgmann
c6dbbc4576 Add function to compute ceil(log2(x)).
Patch by Thilo Borgmann <thilo.borgmann _at_ googlemail.com>

Originally committed as revision 19747 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-30 15:50:03 +00:00
Ramiro Polla
e48a0966ca Add CHECKED_ALLOC macro.
It works the same as CHECKED_ALLOCZ except that it does not zero the allocated
memory.

Originally committed as revision 19742 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-29 22:38:48 +00:00
Måns Rullgård
d691da9508 AVR32: optimisations for intreadwrite.h
Originally committed as revision 19665 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-17 00:15:53 +00:00
Måns Rullgård
e482c2e2fb AVR32: inline asm for bswap.h
Originally committed as revision 19664 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-17 00:15:51 +00:00
Måns Rullgård
9dc6bb7b9f MIPS: 100L: fix AV_[RW]N64 on MIPS64
Originally committed as revision 19660 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 18:43:36 +00:00
Måns Rullgård
4a051891f6 cosmetics: reformat intreadwrite.h
Originally committed as revision 19659 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 18:06:25 +00:00
Måns Rullgård
63826ceb32 intreadwrite: ensure arch-specific versions are always used if defined
The per-arch headers can define any combination of B/L/N variants.
This ensures that whatever is defined in an arch header gets used
for all equivalents not defined there.  E.g. on a little-endian
machine, AV_RN and AV_RL should give the same code.

Originally committed as revision 19658 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 15:51:55 +00:00
Måns Rullgård
57c36bdcea intreadwrite: allow arch-specific 24-bit access macros
Originally committed as revision 19657 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 15:51:52 +00:00
Måns Rullgård
e6956a6e48 ARM: first value loaded in AV_RN64 needs to be early-clobber
Originally committed as revision 19656 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 15:51:50 +00:00
Ramiro Polla
f51fb686de Introduce av_clip_uint16().
Originally committed as revision 19636 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-13 18:47:13 +00:00
Diego Biurrun
3d7b15e450 Remove disabled code cruft.
Originally committed as revision 19616 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-10 13:42:16 +00:00
Måns Rullgård
bd41c06f18 Define INFINITIY and NAN in mathematics.h if missing
Originally committed as revision 19611 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-10 01:27:27 +00:00
Diego Biurrun
75af07495a Revert addition of '#undef av_always_inline' to config.h in the small case.
Instead, #include config.h at the top of common.h so that av_always_inline
does not get doubly defined.

Originally committed as revision 19553 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-30 21:58:43 +00:00
Diego Biurrun
8b4e7c28fb Mark conditionally used variable T2 as av_unused; avoids the warning:
libavutil/sha.c:183: warning: unused variable `T2'

Originally committed as revision 19551 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-30 21:21:30 +00:00
Måns Rullgård
b6ec1f47c7 PPC: fix asm operand constraints in AV_RL64
The target register of first word loaded must be marked early-clobber.

Originally committed as revision 19536 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-29 19:06:55 +00:00
Måns Rullgård
63613fe615 Replace WORDS_BIGENDIAN with HAVE_BIGENDIAN
Originally committed as revision 19508 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-26 12:20:04 +00:00
Måns Rullgård
e7ea5e3d8d Check for __attribute__((packed)) support
Not all compilers claiming to be gcc support this attribute.

Originally committed as revision 19453 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-17 08:21:43 +00:00
Kostya Shishkov
87d718aefc cosmetics: reindent after last commit
Originally committed as revision 19450 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-17 04:30:49 +00:00
Kostya Shishkov
aa5a99ae0c Add SHA-2 hashing
Originally committed as revision 19449 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-17 04:30:06 +00:00
Stefano Sabatini
68485d2728 Remove useless definition of struct AVCLASS.
Originally committed as revision 19402 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-12 11:30:42 +00:00
Alex Converse
291fd18a6e Add a log2 replacement for systems that haven't quite caught up with the C99
standard library.

Originally committed as revision 19395 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-10 19:20:06 +00:00
Kostya Shishkov
451ae25724 Prepare SHA code to handle SHA-2 as well. For now rename files and functions
and leave av_sha1_* functions for compatibility until next major bump.

Originally committed as revision 19394 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-10 15:50:49 +00:00
Kostya Shishkov
0d73abb86a Rename function to sha1_transform so it won't be confused with SHA-2 transform
when it's added.

Originally committed as revision 19389 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 07:29:03 +00:00
Kostya Shishkov
2c6361e009 Use pointer to hash transform function to make adding SHA-2 support easier.
Originally committed as revision 19388 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 07:23:43 +00:00
Kostya Shishkov
3a7c65077d cosmetics: vertical align AVSHA1 members
Originally committed as revision 19387 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 07:15:30 +00:00
Kostya Shishkov
01cc628845 Extend AVSHA1 so it can be used in future SHA-2 code as well
Originally committed as revision 19386 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 07:14:07 +00:00
Kostya Shishkov
4364fc9a0d Document SHA-1 functions and structures
Originally committed as revision 19385 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 07:07:01 +00:00
Ramiro Polla
16c2e21488 Move DECLARE_ALIGNED and DECLARE_ASM_CONST back from lavu/internal.h to
lavu/mem.h.

Originally committed as revision 19384 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 02:20:29 +00:00
Ramiro Polla
4b2be143cd Do not throw compiler error if asm code needs alignment, but we do not know
how to do it for this compiler. This removes a dependency of config.h on the
DECLARE_* macros.

Originally committed as revision 19383 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-09 02:18:29 +00:00
Måns Rullgård
300ad6c16e ARM: use gcc inline asm in bswap.h only when available
Originally committed as revision 19311 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-07-01 00:55:15 +00:00
Carl Eugen Hoyos
a3a55b7848 Icc does not support attribute flatten, do not try to use it.
Originally committed as revision 19286 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-28 18:28:27 +00:00
Carl Eugen Hoyos
06be9d9d8e Icc 11.1 does not support attributes force_align_arg_pointer, alloc_size and cold.
Originally committed as revision 19282 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-27 08:34:04 +00:00
Jeff Downs
d468ff0fdf Fix cast of byte buffer to uint32 that was disregarding alignment
requirements.
Now calculates crc byte at a time until aligned, then continues with uint32
optimized calculation.
This fixes crashes during mlp decoding on sparc (at least, maybe others).

Originally committed as revision 19160 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-12 02:13:12 +00:00
Diego Biurrun
0ef37cd52e cosmetics: Reformat file to K&R style upon Kostya's request.
Originally committed as revision 19154 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-11 10:16:51 +00:00
Diego Biurrun
9d58d4717a cosmetics: Add missing ending backslash to installed header list.
Originally committed as revision 19135 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-09 07:18:58 +00:00
Stefano Sabatini
bf7e799c9e Remove '\p', '\c' and '\e' doxygen markup from doxy, as it should
improve plain text doxy readability.

See the thread: "[RFC] Should we use doxygen markup?".

Originally committed as revision 19122 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-06-06 09:35:15 +00:00
Måns Rullgård
530456bf31 MIPS: inline asm for intreadwrite.h
Originally committed as revision 18980 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-28 23:19:35 +00:00
Ramiro Polla
7d9beec775 Add LABEL_MANGLE() to export label symbols from inside asm blocks.
Originally committed as revision 18934 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-25 01:50:37 +00:00
David Conrad
ef516f7377 Move ALIGN macro to libavutil/common.h and use it in various places
Originally committed as revision 18898 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-22 21:32:13 +00:00
Stefano Sabatini
7a79e51cd6 Add @code/@endcode commands to markup C code in av_tree_insert() doxy.
Originally committed as revision 18776 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-09 23:02:20 +00:00
Stefano Sabatini
9c5a4ea28c Add missing '*' to av_tree_insert() doxy.
Originally committed as revision 18775 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-09 23:00:46 +00:00
Michael Niedermayer
0ee51c556e Fix 10l typo (usage of PIX_FMT_NE()).
Should fix FATE, i forgot to commit this.

Originally committed as revision 18656 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-22 02:29:11 +00:00
Michael Niedermayer
6c2a860875 Add 420,422 and 444 planar 16bit per component pix formats.
Originally committed as revision 18636 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-21 01:02:48 +00:00
Måns Rullgård
cc078b5d8f PPC: check for x-form asm constraint support
Originally committed as revision 18606 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-18 13:57:30 +00:00
Måns Rullgård
9f5ff83f2a PPC asm for AV_RL*()
PPC is normally big endian but has special little endian load/store
instructions.  Using these avoids a separate byteswap.  This makes the
vorbis decoder about 5% faster.  Not much else uses little-endian
read/write extensively.

GCC generates horrible PPC code for the default AV_[RW]B64 (which uses
a packed struct), so we override it with a plain pointer cast.

Originally committed as revision 18602 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-18 00:00:31 +00:00
Måns Rullgård
3c55ce039d ARM asm for AV_RN*()
ARMv6 and later support unaligned loads and stores for single
word/halfword but not double/multiple.  GCC is ignorant of this and
will always use bytewise accesses for unaligned data.  Casting to an
int32_t pointer is dangerous since a load/store double or multiple
instruction might be used (this happens with some code in FFmpeg).
Implementing the AV_[RW]* macros with inline asm using only supported
instructions gives fast and safe unaligned accesses.  ARM RVCT does
the right thing with generic code.

This gives an overall speedup of up to 10%.

Originally committed as revision 18601 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-18 00:00:28 +00:00