On Mon, 1 Aug 2016, Janne Grunau wrote:
On 2016-08-01 10:09:40 +0300, Martin Storsjö wrote:
AV_WN64 is meant for unaligned data, but the existing av_alias* unions
(without a definition for the av_alias attribute - we don't have one
for MSVC) indicate to the compiler that they would have suffici
On 2016-08-01 10:09:40 +0300, Martin Storsjö wrote:
> AV_WN64 is meant for unaligned data, but the existing av_alias* unions
> (without a definition for the av_alias attribute - we don't have one
> for MSVC) indicate to the compiler that they would have sufficient
> alignment for normal access, i.e
On 2016-08-01 09:15:30 +0300, Martin Storsjö wrote:
> This avoids issues with expanding the argument multiple times,
> and makes sure that it is of the right type for the following shifts.
>
> Even if the caller of a macro could be expected not to pass parameters
> that have side effects if expand
Quoting Diego Biurrun (2016-08-01 13:48:47)
> On Sat, Jul 30, 2016 at 10:02:47PM +0200, Anton Khirnov wrote:
> > --- a/libavcodec/x86/vp9dsp_init.c
> > +++ b/libavcodec/x86/vp9dsp_init.c
> > @@ -95,7 +98,9 @@ ff_ ## avg ## _8tap_1d_ ## dir ## _ ## sz ## _ ##
> > opt(uint8_t *dst, \
> >
On 8/1/2016 8:19 AM, Diego Biurrun wrote:
> On Sat, Jul 30, 2016 at 10:02:45PM +0200, Anton Khirnov wrote:
>> ---
>> configure | 6 +-
>> libavcodec/utils.c | 4 +++-
>> 2 files changed, 8 insertions(+), 2 deletions(-)
>
> OK
>
>> --- a/configure
>> +++ b/configure
>> @@ -5032,7 +50
On Sat, Jul 30, 2016 at 10:02:50PM +0200, Anton Khirnov wrote:
> From: Clément Bœsch
>
> Signed-off-by: Anton Khirnov
> ---
> libavcodec/x86/vp9dsp_init.c | 94
> ++--
> libavcodec/x86/vp9mc.asm | 14 +++
> 2 files changed, 54 insertions(+), 54 d
On Sat, Jul 30, 2016 at 10:02:47PM +0200, Anton Khirnov wrote:
> --- a/libavcodec/x86/vp9dsp_init.c
> +++ b/libavcodec/x86/vp9dsp_init.c
> @@ -95,7 +98,9 @@ ff_ ## avg ## _8tap_1d_ ## dir ## _ ## sz ## _ ##
> opt(uint8_t *dst, \
> mc_rep_func(put, sz, hsz, v, ssse3); \
> mc_rep_f
On Sat, Jul 30, 2016 at 10:02:49PM +0200, Anton Khirnov wrote:
> From: James Almer
>
> pavgb is an sse integer instruction, so the mmxext flag is enough
>
> Signed-off-by: James Almer
> Reviewed-by: "Ronald S. Bultje"
> Signed-off-by: Anton Khirnov
> ---
> libavcodec/x86/vp9dsp_init.c | 11 +
On Mon, Aug 01, 2016 at 01:20:39PM +0200, Diego Biurrun wrote:
> On Sat, Jul 30, 2016 at 10:02:46PM +0200, Anton Khirnov wrote:
> > It only contains the MC SIMD, other SIMD will go into different files.
> > ---
> > libavcodec/x86/Makefile | 2 +-
> > libavcodec/x86/{vp9dsp.asm =>
On Sat, Jul 30, 2016 at 10:02:46PM +0200, Anton Khirnov wrote:
> It only contains the MC SIMD, other SIMD will go into different files.
> ---
> libavcodec/x86/Makefile | 2 +-
> libavcodec/x86/{vp9dsp.asm => vp9mc.asm} | 0
> 2 files changed, 1 insertion(+), 1 deletion(-)
> renam
On Sat, Jul 30, 2016 at 10:02:45PM +0200, Anton Khirnov wrote:
> ---
> configure | 6 +-
> libavcodec/utils.c | 4 +++-
> 2 files changed, 8 insertions(+), 2 deletions(-)
OK
> --- a/configure
> +++ b/configure
> @@ -5032,7 +5036,7 @@ enabled_all dxva2 CoTaskMemFree &&
> enable
On Sat, Jul 30, 2016 at 10:02:44PM +0200, Anton Khirnov wrote:
> From: James Almer
>
> Signed-off-by: James Almer
> Signed-off-by: Anton Khirnov
> ---
> configure| 5 +++--
> libavutil/internal.h | 6 ++
> 2 files changed, 9 insertions(+), 2 deletions(-)
OK
Diego
On Sat, Jul 30, 2016 at 10:02:43PM +0200, Anton Khirnov wrote:
> From: Christophe Gisquet
>
> Signed-off-by: Anton Khirnov
> ---
> libavcodec/x86/constants.c | 2 ++
> libavcodec/x86/constants.h | 3 +++
> libavcodec/x86/h264_intrapred_10bit.asm | 2 +-
> libavcodec/x8
On Sun, Jul 31, 2016 at 11:25:04PM +0100, Mark Thompson wrote:
> Most of the functionality here has moved into lavc.
> ---
> This doesn't actually work until codecs have been converted too. (Maybe this
> should go at the end of the series because of that?)
Probably yes.
Diego
__
On Sun, Jul 31, 2016 at 11:30:02PM +0100, Mark Thompson wrote:
> Also constify lots of pointers.
Maybe the constification can be a separate patch? Or does it depend on
the conversion?
Diego
___
libav-devel mailing list
libav-devel@libav.org
https://lis
On Sun, Jul 31, 2016 at 11:22:46PM +0100, Mark Thompson wrote:
>
> --- /dev/null
> +++ b/libavcodec/vaapi_decode.h
> @@ -0,0 +1,95 @@
> +
> +#ifndef AVCODEC_VAAPI_DECODE_H
> +#define AVCODEC_VAAPI_DECODE_H
> +
> +#include
> +
> +#include "libavutil/hwcontext.h"
> +#include "libavutil/hwcontext_va
Hi Anton,
Nice to see your work on race/UB removal.
I have been working on removing races from libavcodec/pthread_frame.c
as part of a wider effort to remove TSan-detected races from Firefox,
which uses ffmpeg-3.0.2. This is tracked starting at
https://bugzilla.mozilla.org/show_bug.cgi?id=127425
Hi Anton,
Nice to see your work on race/UB removal.
I have been working on removing races from libavcodec/pthread_frame.c
as part of a wider effort to remove TSan-detected races from Firefox,
which uses ffmpeg-3.0.2. This is tracked starting at
https://bugzilla.mozilla.org/show_bug.cgi?id=127425
On 31/07/16 23:22, Mark Thompson wrote:
> +void *ff_vaapi_decode_alloc_slice_buffer(AVCodecContext *avctx,
> + VAAPIDecodePicture *pic,
> + size_t params_size,
> + const void *sli
On 01/08/16 06:44, Anton Khirnov wrote:
> Fixes hwaccel decoding of files with multiple slices.
>
> Found-By: Mark Thompson
> ---
> libavcodec/h264dec.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
> index 0e318a5..2c5
On 01/08/16 10:12, Martin Storsjö wrote:
> I guess that would work as well, although it feels more like coincidence
> that they happen to be similar. Or maybe not (since this MSVC attribute
> came from itanium first I think). Since they're quite different
> compilers I'd rather have separate blocks
On Mon, 1 Aug 2016, Luca Barbato wrote:
On 01/08/16 09:09, Martin Storsjö wrote:
AV_WN64 is meant for unaligned data, but the existing av_alias* unions
(without a definition for the av_alias attribute - we don't have one
for MSVC) indicate to the compiler that they would have sufficient
alignme
On 01/08/16 09:09, Martin Storsjö wrote:
> AV_WN64 is meant for unaligned data, but the existing av_alias* unions
> (without a definition for the av_alias attribute - we don't have one
> for MSVC) indicate to the compiler that they would have sufficient
> alignment for normal access, i.e. the compi
On 01/08/16 07:44, Anton Khirnov wrote:
> Fixes hwaccel decoding of files with multiple slices.
>
> Found-By: Mark Thompson
> ---
> libavcodec/h264dec.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
> index 0e318a5..2c5
AV_WN64 is meant for unaligned data, but the existing av_alias* unions
(without a definition for the av_alias attribute - we don't have one
for MSVC) indicate to the compiler that they would have sufficient
alignment for normal access, i.e. the compiler is free to assume
8 byte alignment.
On ARM,
25 matches
Mail list logo