ffmpeg | branch: master | Henrik Gramner | Sun Mar 24
20:04:17 2024 +0100| [7c003b63c85aadf57ca10be849bd9895a4eeeb87] | committer:
Henrik Gramner
avcodec/x86/h264_idct: Fix incorrect xmm spilling on win64
Broken in afa471d0efed1df5dca6eeeb2fcdd211ae4cad4e. It just happened
to work before due
ffmpeg | branch: master | Henrik Gramner | Sat Mar 16
15:48:28 2024 +0100| [782c4df28dc91a2b5160fe7a35ad18541e8c5029] | committer:
Henrik Gramner
x86: Avoid using 'd' as an argument name
x86inc.asm adds defines for {b,w,d,q} which clashes with
the nasm d{b,w,d,q} pseudo-instru
ffmpeg | branch: master | Henrik Gramner | Sat Mar 16
16:39:37 2024 +0100| [afa471d0efed1df5dca6eeeb2fcdd211ae4cad4e] | committer:
Henrik Gramner
x86: Update x86inc.asm
Make things up-to-date with upstream.
https://code.videolan.org/videolan/x86inc.asm
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Henrik Gramner | Sun Mar 17
11:52:57 2024 +0100| [c3d3f0e697f27a14898e74db4fcffa5b35bcaf94] | committer:
Henrik Gramner
avutil/x86util: Fix broken pre-SSE4.1 PMINSD emulation
Fixes yadif-16 which allows FATE to pass.
Broken since
ffmpeg | branch: master | Henrik Gramner | Fri Sep 29
11:36:23 2023 +| [ed8ddf0bd3b188a08ae0b4e21235f7e3e6ce501b] | committer:
Anton Khirnov
x86inc: Add REPX macro to repeat instructions/operations
When operating on large blocks of data it's common to repeatedly use
an instructi
ffmpeg | branch: master | Henrik Gramner | Mon Dec 20
23:52:44 2021 +0100| [15cfb4eee316a1d6a0764f4460409f0258fd94cb] | committer:
Henrik Gramner
checkasm: Use the correct AVTXContext in av_tx tests
Keep a reference to the correct associated context of the reference
function and use that
ffmpeg | branch: release/4.3 | Henrik Gramner | Thu
Jul 9 11:47:35 2020 -0300| [bb3490e7f9645babab4cf84fdb2b2dd4922d81a6] |
committer: James Almer
avutil/x86inc: fix warnings when assembling with Nasm 2.15
Some new warnings regarding use of empty macro parameters has
been added, so adjust
ffmpeg | branch: master | Henrik Gramner | Thu Jul 9
11:47:35 2020 -0300| [0b2b03568f22fdb361d9a44c262bfb9269335f80] | committer:
James Almer
avutil/x86inc: fix warnings when assembling with Nasm 2.15
Some new warnings regarding use of empty macro parameters has
been added, so adjust some
ffmpeg | branch: master | Henrik Gramner | Sat Aug 5
00:09:52 2017 +0200| [3a02cbe3faccad9e346d63cf40c009e664657304] | committer:
Henrik Gramner
x86inc: Enable AVX emulation for floating-point pseudo-instructions
There are 32 pseudo-instructions for each floating-point comparison
instruction
ffmpeg | branch: master | Henrik Gramner | Wed Aug 16
15:59:16 2017 +0200| [6b6edd121699a87c17bd5eca9e94cdd125088c0e] | committer:
Henrik Gramner
x86inc: Support creating global symbols from local labels
On ELF platforms such symbols needs to be flagged as functions with the
correct
ffmpeg | branch: master | Henrik Gramner | Tue Aug 15
16:11:32 2017 +0200| [9e4b3675f226238ae132cada2bb82bcb00110aa6] | committer:
Henrik Gramner
x86inc: Use .rdata instead of .rodata on Windows
The standard section for read-only data on Windows is .rdata. Nasm will
flag non-standard sections
ffmpeg | branch: master | Henrik Gramner | Thu Jan 18
23:20:33 2018 +0100| [6f62b0bd4ff2bd4333d899b697f82643d14ff560] | committer:
Henrik Gramner
x86inc: Drop cpuflags_slowctz
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=6f62b0bd4ff2bd4333d899b697f82643d14ff560
---
libavu
ffmpeg | branch: master | Henrik Gramner | Sat Jan 6
17:47:42 2018 +0100| [eb5f063e7ccc93062a70faac2402a533bb9669fd] | committer:
Henrik Gramner
x86inc: Correctly set mmreg variables
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=eb5f063e7ccc93062a70faac2402a533bb966
ffmpeg | branch: master | Henrik Gramner | Sat Mar 25
10:16:09 2017 +0100| [f7197f68dc61753be67ff68b27bbd3b15fe43da6] | committer:
James Darnley
x86inc: AVX-512 support
AVX-512 consists of a plethora of different extensions, but in order to keep
things a bit more manageable we group together
ffmpeg | branch: master | Henrik Gramner | Sun Sep 17
22:52:13 2017 -0300| [18821e3ba1baa8e0fe037e11c77459ebc73f7e37] | committer:
James Almer
x86/exrdsp: optimize ff_reorder_pixels_avx2()
Tested with "checkasm --test=exrdsp -bench"
Before:
reorder_pixels_c: 5187.8
reorder_pixels_s
ffmpeg | branch: master | Henrik Gramner | Sat May 6
12:26:56 2017 +0200| [aad1b6786e73ebbf908c98099a23e4b061678c4b] | committer:
James Darnley
x86inc: Add some additional cpuflag relations
Simplifies writing assembly code that depends on available instructions.
LZCNT implies SSE2
BMI1
ffmpeg | branch: master | Henrik Gramner | Sat Apr 22
20:30:35 2017 +0200| [cd4ca8245963c6c1845a87b658822b0be0c99048] | committer:
James Darnley
x86inc: Prefer r14/r15 over r12/r13 on x86-64
Due to a peculiarity in the ModR/M addressing encoding, the r12 and r13
registers sometimes requires
ffmpeg | branch: master | Henrik Gramner | Thu Apr 20
19:16:51 2017 +0200| [88dcdfad0964a954b6f185936c023dfd1c5d5e99] | committer:
James Darnley
x86inc: Make REP_RET identical to RET in SSSE3+ functions
There's no point in emitting a rep prefix before ret on modern CPUs.
&
ffmpeg | branch: master | Henrik Gramner | Wed Mar 29
16:43:57 2017 +0200| [406e0ddc0b9b5cc3867ee65924ab211be3770ed6] | committer:
James Darnley
x86inc: Fix call with memory operands
We overload the `call` instruction with a macro, but it would misbehave when
the macro argument wasn't a
ffmpeg | branch: master | Henrik Gramner | Sun Dec 25
20:24:59 2016 +0100| [3cba1ad76d362c994fa98fb686e04e20826fb579] | committer:
Anton Khirnov
x86inc: Avoid using eax/rax for storing the stack pointer
When allocating stack space with an alignment requirement that is larger
than the current
ffmpeg | branch: master | Henrik Gramner | Tue Dec 20
21:33:12 2016 +0100| [cd09e3b34903db7b7e37fdb4d3e10549bf8b2e82] | committer:
Henrik Gramner
x86inc: Avoid using eax/rax for storing the stack pointer
When allocating stack space with an alignment requirement that is larger
than the current
ffmpeg | branch: master | Henrik Gramner | Wed Apr 20
19:35:31 2016 +0200| [ba3eb745cc236765c916b74b39c1874da5c3aefc] | committer:
Anton Khirnov
x86inc: Fix AVX emulation of scalar float instructions
Those instructions are not commutative since they only change the first
element in the vector
ffmpeg | branch: master | Henrik Gramner | Fri Mar 4
17:51:41 2016 +0100| [8dd3ee9dddcaf1d538252d77dd2af3ea139f4f96] | committer:
Henrik Gramner
x86inc: Fix AVX emulation of scalar float instructions
Those instructions are not commutative since they only change the first
element in the
ffmpeg | branch: master | Henrik Gramner | Thu Jan 28
17:11:21 2016 +0100| [389b79842c67b1f5730215a752a5f89cb1b8d9a3] | committer:
Luca Barbato
msvc: Fix libx264 linking
Signed-off-by: Luca Barbato
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Fri Feb 5
21:01:49 2016 +0100| [aa751573fef5a313fd485a74746ab664bd537581] | committer:
Henrik Gramner
avcodec/h264: Fix segfault in 4:2:2 chroma deblock with 32-bit msvc
Using rNm and x86inc's stack allocation with a negative value at the same
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:44 2016 +0100| [7adcd4e841f351aad6c902a08e13d51f7137ad6e] | committer:
Anton Khirnov
x86inc: Make cpuflag() and notcpuflag() return 0 or 1
Makes it possible to use them in arithmetic expressions.
Signed-off-by: Anton Khirnov
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:48 2016 +0100| [5ca8e195e51c43dcc50ca1560ac8f2b33e51] | committer:
Anton Khirnov
x86inc: Use more consistent indentation
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Thu Jan 21
00:05:15 2016 +0100| [91ed050f426ba9747ea42b14e762ed412a28da76] | committer:
Anton Khirnov
x86inc: Preserve arguments when allocating stack space
When allocating stack space with a larger alignment than the known stack
alignment a
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:45 2016 +0100| [f60f06d9894b06d5371729c84f8fd7219909e2bf] | committer:
Anton Khirnov
x86inc: Be more verbose in assertion failures
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:46 2016 +0100| [715eb7ca24c9fda8aed88a3276f5d11567c355bc] | committer:
Anton Khirnov
x86inc: Improve FMA instruction handling
* Correctly handle FMA instructions with memory operands.
* Print a warning if FMA instructions are used
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:49 2016 +0100| [fd6ecac38eb382d1ed41c2ceafa052a3eb6593fc] | committer:
Anton Khirnov
x86inc: Simplify AUTO_REP_RET
cpuflags is never undefined any more, it's set to 0 instead.
Also fix an incorrect comment.
Signed-off-by:
ffmpeg | branch: master | Henrik Gramner | Mon Jan 18
00:21:50 2016 +0100| [002c47798da0c43a053822c8041144798d49ed84] | committer:
Anton Khirnov
x86inc: Avoid creating unnecessary local labels
The REP_RET workaround is only needed on old AMD cpus, and the labels clutter
up the symbol table
ffmpeg | branch: master | Henrik Gramner | Wed Jan 27
22:41:43 2016 +0100| [58313f2dda0a01add2d3d3d33f0b9eadfee1ff05] | committer:
Henrik Gramner
msvc: Fix libx264 linking
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=58313f2dda0a01add2d3d3d33f0b9eadfee1ff05
---
config
ffmpeg | branch: master | Henrik Gramner | Mon Oct 12
21:55:11 2015 +0200| [2d60b18cf0827c599123b1418efc2d0f0a346d01] | committer:
Henrik Gramner
x86inc: Use more consistent indentation
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=2d60b18cf0827c599123b1418efc2d0f0a346
ffmpeg | branch: master | Henrik Gramner | Sun Jan 17
00:25:47 2016 +0100| [b1496008ee173a62c14b49627123ccd4cc486a53] | committer:
Henrik Gramner
x86inc: Improve FMA instruction handling
* Correctly handle FMA instructions with memory operands.
* Print a warning if FMA instructions are used
ffmpeg | branch: master | Henrik Gramner | Fri Oct 16
21:28:49 2015 +0200| [d3662777e070b297ead7c23883d1a7e77d37193f] | committer:
Henrik Gramner
x86inc: Avoid creating unnecessary local labels
The REP_RET workaround is only needed on old AMD cpus, and the labels clutter
up the symbol table
ffmpeg | branch: master | Henrik Gramner | Mon Oct 12
20:15:18 2015 +0200| [dfe771dc5a09286a2cfeb68bdb4c328ecfe1c551] | committer:
Henrik Gramner
x86inc: Preserve arguments when allocating stack space
When allocating stack space with a larger alignment than the known stack
alignment a
ffmpeg | branch: master | Henrik Gramner | Thu Oct 15
17:42:49 2015 +0200| [87b587d4fe91064325884bbd4130ef079e6c3abe] | committer:
Henrik Gramner
x86inc: Simplify AUTO_REP_RET
cpuflags is never undefined any more, it's set to 0 instead.
Also fix an incorrect comment.
&
ffmpeg | branch: master | Henrik Gramner | Sun Oct 11
22:31:53 2015 +0200| [6cbd0fdf2859ce1c65a6d79de473922dce95d11a] | committer:
Henrik Gramner
x86inc: Be more verbose in assertion failures
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=6cbd0fdf2859ce1c65a6d79de473922dce95d
ffmpeg | branch: master | Henrik Gramner | Sat Sep 26
22:27:36 2015 +0200| [2c3dbff1d89bbf8300c121c06524c014cb6e0915] | committer:
Anton Khirnov
avutil/avstring: Inline some tiny functions
They're short enough that inlining them actually reduces code size due to
all the overhead assoc
ffmpeg | branch: master | Henrik Gramner | Mon Sep 28
16:50:31 2015 +0200| [8bb376cf6b4ab8645daedb8becaa7163656436a4] | committer:
Anton Khirnov
checkasm: Fix the function name sorting algorithm
The previous implementation was behaving incorrectly in some corner cases.
Signed-off-by: Anton
ffmpeg | branch: master | Henrik Gramner | Thu Oct 1
19:25:33 2015 +0200| [ec85153f252ce1cc1ce0612237b9df26cb37f293] | committer:
Henrik Gramner
checkasm: Fix compilation with --disable-avcodec
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Sat Sep 26
15:43:49 2015 +0200| [99982524f93a5fc5f8eadd3e8f9b4e3af446cdaa] | committer:
Henrik Gramner
checkasm: Remove use of deprecated av_set_cpu_flags_mask()
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Wed Sep 30
23:17:41 2015 +0200| [17710550c483cc7779ed9ce6129bd85c5fa83a8e] | committer:
Henrik Gramner
x86inc: Make cpuflag() and notcpuflag() return 0 or 1
Makes it possible to use them in arithmetic expressions.
> http://git.videolan.
ffmpeg | branch: master | Henrik Gramner | Fri Sep 25
21:35:35 2015 +0200| [5405584b7b54ca889c341743de1d58792449830d] | committer:
Anton Khirnov
checkasm: Use a self-balancing tree
Tested functions are internally kept in a binary search tree for efficient
lookups. The downside of the current
ffmpeg | branch: master | Henrik Gramner | Wed Sep 23
17:22:38 2015 +0200| [cc28552120ab237d183dc3a26f99fce3030f] | committer:
Anton Khirnov
checkasm/x86: Correctly handle variadic functions
The System V ABI on x86-64 specifies that the al register contains an upper
bound of the number of
ffmpeg | branch: master | Henrik Gramner | Mon Sep 28
18:50:13 2015 +0200| [69e456d7fbc5fff88acf747d135bf15c8e511c59] | committer:
Henrik Gramner
checkasm/vp9dsp: Fix iszero() to read the correct data
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Sat Sep 26
20:15:35 2015 +0200| [19b28d047daa2bc9eaac0cca391bb5932fcb7059] | committer:
Henrik Gramner
checkasm: Fix the function name sorting algorithm
The previous implementation was behaving incorrectly in some corner cases.
> h
ffmpeg | branch: master | Henrik Gramner | Wed Sep 23
16:06:02 2015 +0200| [7ca1de5b4f1a369a997c5cf9924b0197606f558d] | committer:
Henrik Gramner
checkasm/x86: Correctly handle variadic functions
The System V ABI on x86-64 specifies that the al register contains an upper
bound of the number
ffmpeg | branch: master | Henrik Gramner | Sat Sep 26
18:04:11 2015 +0200| [ad9a543e9351d940a53450f793458664c8a74a73] | committer:
Henrik Gramner
avutil/avstring: Inline some tiny functions
They're short enough that inlining them actually reduces code size due to
all the overhead assoc
ffmpeg | branch: master | Henrik Gramner | Fri Sep 25
18:56:00 2015 +0200| [2ab65b652dc5e69fb738f1afdc55f7a2f9cbc0e0] | committer:
Henrik Gramner
checkasm: Use a self-balancing tree
Tested functions are internally kept in a binary search tree for efficient
lookups. The downside of the current
ffmpeg | branch: master | Henrik Gramner | Wed Sep 16
14:03:01 2015 +0200| [781a25e9c419dd66063597cc5d70e8919de60561] | committer:
Vittorio Giovara
checkasm: v210: Fix array overwrite
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=781a25e9c419dd66063597cc5d70e8919de60
ffmpeg | branch: master | Henrik Gramner | Tue Sep 15
23:21:28 2015 +0200| [6115966ad3473aff2309f19c4e27f279109efe3a] | committer:
Henrik Gramner
checkasm: v210: Fix array overwrite
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=6115966ad3473aff2309f19c4e27f279109ef
ffmpeg | branch: master | Henrik Gramner | Wed Sep 16
13:48:43 2015 +0200| [985e7d8cc1091d2627aa8bf2723c1fa5db0d38c0] | committer:
Henrik Gramner
checkasm: v210: s/Libav/FFmpeg/
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=985e7d8cc1091d2627aa8bf2723c1fa5db0d38c0
---
te
ffmpeg | branch: master | Henrik Gramner | Sun Sep 6
01:06:12 2015 +0200| [3cdda78deb19b39dbbf8961ae0aec44dbb19bf6d] | committer:
Luca Barbato
checkasm: add unit tests for v210enc
Signed-off-by: Luca Barbato
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Mon Aug 24
22:53:56 2015 +0200| [c457bdebe7af643b380aa0f6add3cb4335d218dc] | committer:
Anton Khirnov
checkasm: Fix floating point arguments on 64-bit Windows
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Mon Aug 24
22:52:02 2015 +0200| [33a58d7bf40f9d78ffedf1ebb742bc0fc178e65a] | committer:
Henrik Gramner
checkasm: Fix floating point arguments on 64-bit Windows
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Thu Aug 20
17:06:21 2015 +0200| [515b69f8f8e9a24cfaee95d8c1f63f265d8582fe] | committer:
Anton Khirnov
checkasm: Explicitly declare function prototypes
Now we no longer have to rely on function pointers intentionally
declared without specified
ffmpeg | branch: master | Henrik Gramner | Sun Aug 16
17:58:02 2015 +0200| [e13da244f41610ee073b2f72bcf62b60fa402bb5] | committer:
Anton Khirnov
checkasm: x86: properly save rdx/edx in checked_call()
If the return value doesn't fit in a single register rdx/edx can in some
cases be us
ffmpeg | branch: master | Henrik Gramner | Sun Aug 16
13:00:21 2015 +0200| [18b101ff595c7f18e9571d26f8840f556b24ec03] | committer:
Henrik Gramner
checkasm: Explicitly declare function prototypes
Now we no longer have to rely on function pointers intentionally
declared without specified
ffmpeg | branch: master | Henrik Gramner | Sun Aug 16
17:09:40 2015 +0200| [e6b8797b827ce3c5eb0608725db7e7e85d78864b] | committer:
Henrik Gramner
checkasm: x86: properly save rdx/edx in checked_call()
If the return value doesn't fit in a single register rdx/edx can in some
cases be us
ffmpeg | branch: master | Henrik Gramner | Tue Aug 11
20:56:24 2015 +0200| [44b1203177690305c294be6eff8d9c668954] | committer:
Anton Khirnov
x86inc: Various minor backports from x264
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Sat Aug 1
17:27:36 2015 +0200| [ab43beefab9147673e09679e04be08431684a5db] | committer:
Anton Khirnov
x86inc: Drop SECTION_TEXT macro
The .text section is already 16-byte aligned by default on all supported
platforms so `SECTION_TEXT` isn'
ffmpeg | branch: master | Henrik Gramner | Sat Aug 1
17:27:35 2015 +0200| [1c6bb813284732d9a1acacfe99522d9f66ebf73e] | committer:
Anton Khirnov
x86inc: Disable vpbroadcastq workaround in newer yasm versions
The bug was fixed in 1.3.0, so only perform the workaround in earlier versions
ffmpeg | branch: master | Henrik Gramner | Sat Aug 1
17:27:32 2015 +0200| [9f1245eb9620a70feaa00ba745c6c7a56a839556] | committer:
Anton Khirnov
x86inc: Support arbitrary stack alignments
Change ALLOC_STACK to always align the stack before allocating stack space for
consistency. Previously
ffmpeg | branch: master | Henrik Gramner | Wed Aug 5
23:27:41 2015 +0200| [8f4a06faf45c1cbcabec610f4b47824171379934] | committer:
Anton Khirnov
checkasm: Remove unnecessary include
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Sat Aug 1
17:10:22 2015 +0200| [4a53c758d2f69302e6822b23f3e7349253b9e8d3] | committer:
Anton Khirnov
x86: dcadsp: Avoid SSE2 instructions in SSE functions
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Wed Aug 5
19:05:56 2015 +0200| [5e8e121fccc247942a42b609f1e32060950c1a65] | committer:
Henrik Gramner
checkasm: Remove unnecessary include
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=5e8e121fccc247942a42b609f1e32060950c1
ffmpeg | branch: master | Henrik Gramner | Wed May 27
21:38:14 2015 +0200| [f0b7882ceb791ff32267ad97bd21a09e9cbf83d5] | committer:
Henrik Gramner
x86inc: Drop SECTION_TEXT macro
The .text section is already 16-byte aligned by default on all supported
platforms so `SECTION_TEXT` isn'
ffmpeg | branch: master | Henrik Gramner | Mon Feb 10
00:58:04 2014 +0200| [826790f59640a84813e8dd81c2bba559516f8e4d] | committer:
Henrik Gramner
x86inc: Support arbitrary stack alignments
Change ALLOC_STACK to always align the stack before allocating stack space for
consistency. Previously
ffmpeg | branch: master | Henrik Gramner | Sun Aug 2
22:40:03 2015 +0200| [127203ba5a0b9fab652a5095e52f3d247abc98e5] | committer:
Michael Niedermayer
x86inc: Various minor backports from x264
Reviewed-by: "Ronald S. Bultje"
Signed-off-by: Michael Niedermayer
> http://git
ffmpeg | branch: master | Henrik Gramner | Sun Aug 2
22:40:01 2015 +0200| [f151fbd9e58c7af6535efb342d868bf803afef34] | committer:
Michael Niedermayer
x86inc: Disable vpbroadcastq workaround in newer yasm versions
The bug was fixed in 1.3.0, so only perform the workaround in earlier versions
ffmpeg | branch: master | Henrik Gramner | Sat Aug 1
17:12:48 2015 +0200| [ebaf571aca2dd6ce3cac4210a3fccd47e7db] | committer:
Anton Khirnov
x86: dct: Disable dct32_float_sse on x86-64
There is an SSE2 implementation so the SSE version is never used. The "SSE"
version also
ffmpeg | branch: master | Henrik Gramner | Fri Jul 24
23:46:20 2015 +0200| [65c14801527068fcaf729eeffc142ffd4682a21a] | committer:
Anton Khirnov
checkasm: Modify report format
Makes it a bit more clear where each test belongs.
Suggested by Anton Khirnov.
Signed-off-by: Anton Khirnov
ffmpeg | branch: master | Henrik Gramner | Fri Jul 17
18:07:57 2015 +0200| [6cc4d3e9a982e926494f4b919d9733fe29774acf] | committer:
Janne Grunau
checkasm: exit with status 0 instead of 1 if there are no tests to perform
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Thu Jul 16
00:10:27 2015 +0200| [a344e5d094ebcf9a23acf3a27c56cbbbc829db42] | committer:
Luca Barbato
x86: bswapdsp: Don't treat 32-bit integers as 64-bit
The upper halves are not guaranteed to be zero in x86-64.
Also use `test` instead of `and`
ffmpeg | branch: master | Henrik Gramner | Thu Jul 16
00:10:28 2015 +0200| [d37f23263584774e1798e9ac909a398304a05091] | committer:
Luca Barbato
checkasm: Add unit tests for bswapdsp
Signed-off-by: Luca Barbato
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Mon Jul 13
23:11:25 2015 +0200| [2cb34f82b92c15b811f5c03dc7f61a4baf6e02e3] | committer:
Luca Barbato
checkasm: Add unit tests for h264qpel
Signed-off-by: Luca Barbato
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Henrik Gramner | Sat Jul 11
20:32:11 2015 +0200| [8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717] | committer:
Janne Grunau
Checkasm: assembly testing and benchmarking tool
It provides the following features:
* verify correctness by comparing output to the C version
ffmpeg | branch: master | Henrik Gramner | Wed Jul 8
22:33:53 2015 +0200| [d1a6cb195f610978ba5d2351e60f938f7f261d59] | committer:
Luca Barbato
x86: Serialize rdtsc in read_time()
Improves the accuracy of measurements, especially in short sections.
To quote the Intel 64 and IA-32
ffmpeg | branch: master | Henrik Gramner | Wed Nov 5
00:05:42 2014 +0100| [4981baf9b803f3c4866b2e97fdadb008c62dc7ad] | committer:
Martin Storsjö
avstring: Mark some character handling functions av_const
Signed-off-by: Martin Storsjö
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=com
ffmpeg | branch: master | Henrik Gramner | Wed Oct 1
19:03:29 2014 +0200| [2d91abade29e43bb45c881d45909b8ee77e904e2] | committer:
Anton Khirnov
x86: h264_intrapred: Don't treat 32-bit integers as 64-bit
The upper halves are not guaranteed to be zero in x86-64.
Signed-off-by: Anton Kh
ffmpeg | branch: master | Henrik Gramner | Fri Sep 5
00:13:38 2014 +0200| [f629705b0239c80fddc1b0b15ed4bb9042c77d23] | committer:
Diego Biurrun
x86inc: Make INIT_CPUFLAGS support an arbitrary number of cpuflags
Previously there was a limit of two cpuflags.
Signed-off-by: Diego Biurrun
ffmpeg | branch: master | Henrik Gramner | Fri Sep 5
00:13:34 2014 +0200| [176a0fca3fd64f91d585f96137388e00d8c101b6] | committer:
Diego Biurrun
x86inc: Make ym# behave the same way as xm#
This makes more sense for future implementations of templates with zmm
registers.
Signed-off-by: Diego
ffmpeg | branch: master | Henrik Gramner | Fri Sep 5
00:13:38 2014 +0200| [428aa14a4851711ea1aa3f9b2d5bac867434993f] | committer:
Michael Niedermayer
x86inc: Make INIT_CPUFLAGS support an arbitrary number of cpuflags
Previously there was a limit of two cpuflags.
Signed-off-by: Michael
ffmpeg | branch: master | Henrik Gramner | Fri Sep 5
00:13:34 2014 +0200| [720c21d11ff0742989bae3e763fc0a85a587fdb6] | committer:
Michael Niedermayer
x86inc: Make ym# behave the same way as xm#
This makes more sense for future implementations of templates with zmm
registers.
Signed-off-by
86 matches
Mail list logo
