Quoting Mark Thompson (2016-09-04 14:43:21)
> Also adds some extra fields to the main context structure that may
> be needed by a hwaccel decoder.
> ---
> libavcodec/vp8.c | 190
> +++
> libavcodec/vp8.h | 25
> 2 files changed, 162
Quoting Mark Thompson (2016-09-04 14:46:37)
> ---
> Tested on Skylake.
>
> Passes the conformance tests, except for 6 and 14 (the ones with odd frame
> dimensions).
Looks ok other than the stuff already mentioned.
--
Anton Khirnov
___
libav-devel
Quoting Mark Thompson (2016-09-04 14:48:07)
> Enables VP8 decoding - the decoder places the the bitstream version
> in the profile field, which we want to ignore.
> ---
> An alternative would be defining FF_PROFILE_VP8_VERSION[0-3] in avcodec.h and
> then using those here, but I think that is
Also adds some extra fields to the main context structure that may
be needed by a hwaccel decoder.
---
v2: review comments incorporated.
On 06/09/16 09:57, Anton Khirnov wrote:
> Quoting Mark Thompson (2016-09-04 14:43:21)
>> @@ -313,13 +336,19 @@ static void get_quants(VP8Context *s)
>>
On 06/09/16 21:54, Mark Thompson wrote:
> ---
> configure | 2 +
> libavcodec/Makefile| 1 +
> libavcodec/allcodecs.c | 1 +
> libavcodec/vaapi_vp8.c | 231
> +
> libavcodec/vp8.c | 3 +
> 5 files changed, 238
Enables VP8 decoding - the decoder places the the bitstream version
in the profile field, which we want to ignore.
---
libavcodec/vaapi_decode.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/libavcodec/vaapi_decode.c b/libavcodec/vaapi_decode.c
index 51b6d47..ab8445a
On Mon, Sep 05, 2016 at 01:02:42PM +0200, Anton Khirnov wrote:
> This will make the x86 asm simpler.
>
> ARM conversion by Martin Storsjö and Janne Grunau
>
> ---
> libavcodec/ac3enc_float.c | 2 +-
> libavcodec/arm/audiodsp_init_neon.c | 3
On Mon, Sep 05, 2016 at 01:02:43PM +0200, Anton Khirnov wrote:
> ---
> libavcodec/x86/Makefile | 1 -
> libavcodec/x86/audiodsp.asm | 42 +++
> libavcodec/x86/audiodsp_mmx.c | 58
> ---
> 3 files changed, 42
Quoting Mark Thompson (2016-09-05 13:38:19)
> ---
> On 05/09/16 12:10, Diego Biurrun wrote:
> > Module: libav
> > Branch: master
> > Commit: 2610c9528f86286e4c6e174411a26ff5b4815cde
> >
> > Author:Diego Biurrun
> > Committer: Diego Biurrun
> > Date:
On 06/09/16 10:40, Anton Khirnov wrote:
> I think the idea is to only auto-enable things that do not carry runtime
> library dependencies.
>
* Donning his Gentoo hat
lu> Automagic dependencies are murder!
Jokes aside, yes we should not auto-enable external dependencies that
aren't expected
On Mon, Sep 05, 2016 at 01:02:41PM +0200, Anton Khirnov wrote:
> The x86 version processes 16 floats per iteration, so len must be a
> multiple of 16.
> ---
> libavcodec/audiodsp.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
OK
Diego
___
On Tue, Sep 06, 2016 at 12:35:02AM +0200, Luca Barbato wrote:
> On 05/09/16 20:31, Diego Biurrun wrote:
> > --- a/configure
> > +++ b/configure
> > @@ -4456,7 +4456,8 @@ check_func ${malloc_prefix}memalign&&
> > enable memalign
> >
> > check_cpp_condition unistd.h
On Mon, Sep 05, 2016 at 01:02:36PM +0200, Anton Khirnov wrote:
> --- /dev/null
> +++ b/tests/checkasm/blockdsp.c
> @@ -0,0 +1,68 @@
> +/*
> + * Copyright (c) 2015 Henrik Gramner
Henrik?
> +#include
> +
> +#include "checkasm.h"
> +
> +#include "libavcodec/blockdsp.h"
> +
> +#include
On Mon, Sep 05, 2016 at 01:02:35PM +0200, Anton Khirnov wrote:
> --- a/libavcodec/blockdsp.h
> +++ b/libavcodec/blockdsp.h
> @@ -40,13 +40,13 @@ typedef struct BlockDSPContext {
>
> #if FF_API_XVMC
> -void ff_blockdsp_init_x86(BlockDSPContext *c, unsigned high_bit_depth,
> +void
On Tue, Sep 6, 2016 at 11:39 AM, Anton Khirnov wrote:
>> Use 3-arg maxps instead of mova.
>
> Isn't that AVX-only?
It is, x86inc will simply convert it to mova+minps when assembling it
as non-AVX code but it reduces the line count. It's certainly not
worth to go into
On Mon, Sep 05, 2016 at 01:02:37PM +0200, Anton Khirnov wrote:
> --- a/libavcodec/x86/blockdsp.c
> +++ /dev/null
> @@ -1,118 +0,0 @@
> -av_cold void ff_blockdsp_init_x86(BlockDSPContext *c)
> -{
> -if (INLINE_MMX(cpu_flags)) {
> -c->clear_block = clear_block_mmx;
> -
On Mon, Sep 05, 2016 at 01:02:40PM +0200, Anton Khirnov wrote:
> --- /dev/null
> +++ b/tests/checkasm/audiodsp.c
> @@ -0,0 +1,139 @@
> +
> +#include
> +#include
> +#include
> +#include
nit: order
> +#include "libavcodec/audiodsp.h"
> +
> +#include "libavutil/common.h"
> +#include
On Tue, Sep 6, 2016 at 11:44 AM, Anton Khirnov wrote:
> Also change shl to add, since it can be faster on some CPUs.
>
> CC: libav-sta...@libav.org
> ---
> libavcodec/x86/audiodsp.asm | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
Ok.
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/hq_hqadsp.c | 3 ++-
libavcodec/hq_hqadsp.h | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/libavcodec/hq_hqadsp.c b/libavcodec/hq_hqadsp.c
index 93fc067..3fb2a76 100644
---
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/loco.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libavcodec/loco.c b/libavcodec/loco.c
index f25ef61..c031ac6 100644
--- a/libavcodec/loco.c
+++ b/libavcodec/loco.c
@@ -114,7 +114,7 @@
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/on2avc.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/libavcodec/on2avc.c b/libavcodec/on2avc.c
index 2a528c6..dd65dec 100644
--- a/libavcodec/on2avc.c
+++ b/libavcodec/on2avc.c
@@ -317,8
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/qdrw.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/qdrw.c b/libavcodec/qdrw.c
index b7493e4..3605a5a 100644
--- a/libavcodec/qdrw.c
+++ b/libavcodec/qdrw.c
@@ -67,7 +67,7 @@ static int
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/fic.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libavcodec/fic.c b/libavcodec/fic.c
index b1286eb..fc4ffac 100644
--- a/libavcodec/fic.c
+++ b/libavcodec/fic.c
@@ -79,7 +79,7 @@ static
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/dv.h | 4 ++--
libavcodec/dv_profile.h | 4 ++--
libavcodec/dvdec.c | 4 ++--
libavcodec/dvenc.c | 9 +
4 files changed, 11 insertions(+), 10 deletions(-)
diff --git a/libavcodec/dv.h
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/hapenc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/hapenc.c b/libavcodec/hapenc.c
index 9ebad4a..c05e1c8 100644
--- a/libavcodec/hapenc.c
+++ b/libavcodec/hapenc.c
@@ -61,7 +61,7 @@
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/error_resilience.c | 21 -
libavcodec/error_resilience.h | 4 ++--
2 files changed, 14 insertions(+), 11 deletions(-)
diff --git a/libavcodec/error_resilience.c b/libavcodec/error_resilience.c
index
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/rectangle.h | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/libavcodec/rectangle.h b/libavcodec/rectangle.h
index 616a637..2fc264d 100644
--- a/libavcodec/rectangle.h
+++ b/libavcodec/rectangle.h
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/dxv.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/dxv.c b/libavcodec/dxv.c
index 99327df..fce7f62 100644
--- a/libavcodec/dxv.c
+++ b/libavcodec/dxv.c
@@ -36,7 +36,7 @@ typedef struct
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/svq3.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/libavcodec/svq3.c b/libavcodec/svq3.c
index aa85e7c..7eeb94c 100644
--- a/libavcodec/svq3.c
+++ b/libavcodec/svq3.c
@@ -118,8 +118,9 @@
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/utvideo.h| 2 +-
libavcodec/utvideodec.c | 12 ++--
libavcodec/utvideoenc.c | 13 +++--
3 files changed, 14 insertions(+), 13 deletions(-)
diff --git a/libavcodec/utvideo.h b/libavcodec/utvideo.h
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/iirfilter.c | 7 +--
libavcodec/iirfilter.h | 8 +---
2 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/libavcodec/iirfilter.c b/libavcodec/iirfilter.c
index 442c837..4116d5c 100644
---
This avoids SIMD-optimized functions having to sign-extend their
line size argument manually to be able to do pointer arithmetic.
Also adjust parameter names to be "stride" everywhere.
---
libavcodec/arm/pixblockdsp_init_arm.c | 5 +++--
libavcodec/pixblockdsp.c | 2 +-
ptrdiff_t is the correct type for array strides and similar.
---
libavcodec/fraps.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/fraps.c b/libavcodec/fraps.c
index eb61c70..af7f4ad 100644
--- a/libavcodec/fraps.c
+++ b/libavcodec/fraps.c
@@ -91,7 +91,7 @@ static
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/iirfilter.c | 7 +--
> libavcodec/iirfilter.h | 8 +---
> 2 files changed, 10 insertions(+), 5 deletions(-)
>
Seems fine.
Quoting Diego Biurrun (2016-09-06 15:54:08)
> On Mon, Sep 05, 2016 at 01:02:40PM +0200, Anton Khirnov wrote:
> > --- /dev/null
> > +++ b/tests/checkasm/audiodsp.c
> > @@ -0,0 +1,139 @@
> > +
> > +#include
> > +#include
> > +#include
> > +#include
>
> nit: order
>
> > +#include
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/rectangle.h | 5 -
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
Ok.
___
libav-devel mailing list
libav-devel@libav.org
On 06/09/16 23:19, Mark Thompson wrote:
> On 06/09/16 21:54, Mark Thompson wrote:
>> ---
>> configure | 2 +
>> libavcodec/Makefile| 1 +
>> libavcodec/allcodecs.c | 1 +
>> libavcodec/vaapi_vp8.c | 231
>> +
>>
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/svq3.c | 9 +
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/libavcodec/svq3.c b/libavcodec/svq3.c
> index aa85e7c..7eeb94c 100644
> ---
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/utvideo.h| 2 +-
> libavcodec/utvideodec.c | 12 ++--
> libavcodec/utvideoenc.c | 13 +++--
> 3 files changed, 14 insertions(+), 13 deletions(-)
>
Seems
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/dv.h | 4 ++--
> libavcodec/dv_profile.h | 4 ++--
> libavcodec/dvdec.c | 4 ++--
> libavcodec/dvenc.c | 9 +
> 4 files changed, 11 insertions(+), 10
---
configure | 2 +
libavcodec/Makefile| 1 +
libavcodec/allcodecs.c | 1 +
libavcodec/vaapi_vp8.c | 231 +
libavcodec/vp8.c | 3 +
5 files changed, 238 insertions(+)
create mode 100644 libavcodec/vaapi_vp8.c
diff
Quoting Henrik Gramner (2016-09-05 15:15:14)
> On Mon, Sep 5, 2016 at 1:02 PM, Anton Khirnov wrote:
> > +cglobal vector_clipf, 3, 3, 6, dst, src, len, min, max
> > +%if ARCH_X86_32
> > +VBROADCASTSS m0, minm
> > +VBROADCASTSS m1, maxm
> > +%else
> > +VBROADCASTSS
Also change shl to add, since it can be faster on some CPUs.
CC: libav-sta...@libav.org
---
libavcodec/x86/audiodsp.asm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/x86/audiodsp.asm b/libavcodec/x86/audiodsp.asm
index dc38ada..d7e63eb 100644
---
---
Now the benchmark is always run on the maximum buffer size, so it actually
produces consistent results
---
tests/checkasm/Makefile | 1 +
tests/checkasm/audiodsp.c | 139 ++
tests/checkasm/checkasm.c | 3 +
tests/checkasm/checkasm.h | 1 +
4
---
Some of the review comments were conflicting, so I tried to find a workable
compromise.
---
libavcodec/x86/Makefile| 1 -
libavcodec/x86/audiodsp.asm| 43 +++
libavcodec/x86/audiodsp_init.c | 2 +-
libavcodec/x86/audiodsp_mmx.c | 58
On 06/09/16 18:11, Diego Biurrun wrote:
> ptrdiff_t is the correct type for array strides and similar.
> ---
> libavcodec/error_resilience.c | 21 -
> libavcodec/error_resilience.h | 4 ++--
> 2 files changed, 14 insertions(+), 11 deletions(-)
>
Seems OK.
lu
46 matches
Mail list logo