Re: [libav-devel] [PATCH 09/12] avcodec/huffyuvdsp: Change w to intptr in add_hfyu_median_pred() and add_hfyu_left_pred()

2016-10-21 Thread Luca Barbato
On 16/10/2016 20:10, Janne Grunau wrote: > From: Michael Niedermayer > > This avoids potential issues with the high 32bits being random in x86-64 asm > > Signed-off-by: Michael Niedermayer > Signed-off-by: Janne Grunau > --- >

Re: [libav-devel] [PATCH] mov: Update colr values

2016-10-21 Thread Luca Barbato
On 21/10/2016 22:20, Vittorio Giovara wrote: > For 'nclx', the latest edition of the standard switched from JPEG XR > to 23001-8, which matches the current order of our entries. Bound > checking is preserved as a sanity check. > > For 'nclc', qtff edition 2016-09-13 introduced a few new entries.

[libav-devel] [PATCH] mov: Update colr values

2016-10-21 Thread Vittorio Giovara
For 'nclx', the latest edition of the standard switched from JPEG XR to 23001-8, which matches the current order of our entries. Bound checking is preserved as a sanity check. For 'nclc', qtff edition 2016-09-13 introduced a few new entries. Signed-off-by: Vittorio Giovara

Re: [libav-devel] [PATCH 09/12] avcodec/huffyuvdsp: Change w to intptr in add_hfyu_median_pred() and add_hfyu_left_pred()

2016-10-21 Thread Diego Biurrun
On Sun, Oct 16, 2016 at 08:10:59PM +0200, Janne Grunau wrote: > From: Michael Niedermayer > > This avoids potential issues with the high 32bits being random in x86-64 asm > > Signed-off-by: Michael Niedermayer > Signed-off-by: Janne Grunau

Re: [libav-devel] [PATCH 11/12] x86util: add and use RSHIFT/LSHIFT macros

2016-10-21 Thread Diego Biurrun
On Sun, Oct 16, 2016 at 08:11:01PM +0200, Janne Grunau wrote: > From: Christophe Gisquet > > Those macros take a byte number as shift argument, as this argument > differs between MMX and SSE2 instructions. > > Signed-off-by: Michael Niedermayer >

Re: [libav-devel] [PATCH 10/12] x86: huffyuvdsp: fewer functions for x86_64

2016-10-21 Thread Diego Biurrun
On Sun, Oct 16, 2016 at 08:11:00PM +0200, Janne Grunau wrote: > --- a/libavcodec/x86/huffyuvdsp_init.c > +++ b/libavcodec/x86/huffyuvdsp_init.c > @@ -76,17 +76,18 @@ static void add_hfyu_median_pred_cmov(uint8_t *dst, const > uint8_t *top, > > +#if ARCH_X86_32 > if

[libav-devel] [PATCH] fate: Add --ignore-tests configure option for omitting specific FATE tests

2016-10-21 Thread Diego Biurrun
This can be useful to filter out noise in known-broken scenarios like miscompilation by legacy compilers and similar. --- Applied suggestions by Janne and added config file support. configure | 10 ++ doc/fate.texi | 1 + tests/Makefile | 4 +++- tests/fate.sh | 1 + 4 files

Re: [libav-devel] [PATCH] bktr: Use memset(0) instead of zero-initialization for struct sigaction

2016-10-21 Thread Diego Biurrun
On Fri, Oct 21, 2016 at 03:41:39PM +0300, Martin Storsjö wrote: > On Fri, 21 Oct 2016, Diego Biurrun wrote: > > >sigaction is not defined in standards as a struct starting with another > >struct. > > ... except that some BSD in practice does define in such a way, causing a > warning, which this

Re: [libav-devel] [PATCH] fate: Add --skip-tests configure option for omitting specific FATE tests

2016-10-21 Thread Janne Grunau
On 2016-10-18 13:55:53 +0200, Diego Biurrun wrote: > This can be useful to filter out noise in known-broken scenarios like > miscompilation by legacy compilers and similar. > --- > > Now adds a "fate-" prefix to the test name on its own. > > configure | 4 > tests/Makefile | 2 ++ > 2

Re: [libav-devel] [PATCH 2/2] mov: Update colr values

2016-10-21 Thread Luca Barbato
On 19/10/2016 00:21, Vittorio Giovara wrote: > For 'nclx', the latest edition of the standard switched from JPEG XR > to 23001-8, which matches the current order of our entries. > For 'nclc', qtff edition 2016-09-13 introduced a few new entries. > > Signed-off-by: Vittorio Giovara

Re: [libav-devel] [PATCH 1/2] rtmpdh: Don't use the OpenSSL DH struct

2016-10-21 Thread Luca Barbato
On 21/10/2016 13:02, Martin Storsjö wrote: > Instead use our own struct, which we already use when using > gcrypt and gnutls. > > In OpenSSL 1.1, the DH struct has been made opaque. > --- > This produces the following warnings when building with gnutls: > warning: unused label 'fail'

Re: [libav-devel] [PATCH 2/2] openssl: Support version 1.1.0.

2016-10-21 Thread Luca Barbato
On 21/10/2016 13:02, Martin Storsjö wrote: > From: Matt Oliver > > Further simplifications by Martin Storsjö, to minimize the > diff. > --- > configure | 3 ++- > libavformat/tls_openssl.c | 37 +++-- > 2 files changed, 37

Re: [libav-devel] [PATCH] Replace leftover uses -aframes|-dframes|-vframes with -frames:a|d|v

2016-10-21 Thread Luca Barbato
On 21/10/2016 13:49, Diego Biurrun wrote: > --- > doc/avconv.texi | 13 - > tests/fate/h264.mak | 2 +- > tests/fate/video.mak | 4 ++-- > 3 files changed, 11 insertions(+), 8 deletions(-) > Looks good. Thank you! lu ___

Re: [libav-devel] [PATCH] bktr: Use memset(0) instead of zero-initialization for struct sigaction

2016-10-21 Thread Martin Storsjö
On Fri, 21 Oct 2016, Diego Biurrun wrote: sigaction is not defined in standards as a struct starting with another struct. ... except that some BSD in practice does define in such a way, causing a warning, which this commit fixes. This partially reverts

Re: [libav-devel] [PATCH 12/12] x86/yadif-10: remove duplicate ABS macro

2016-10-21 Thread Diego Biurrun
On Sun, Oct 16, 2016 at 08:11:02PM +0200, Janne Grunau wrote: > From: James Almer > > And use the x86util ones instead, which are optimized for mmxext/sse2. > About ~1% increase in performance on pre SSSE3 processors. > > Signed-off-by: James Almer >

[libav-devel] [PATCH] bktr: Use memset(0) instead of zero-initialization for struct sigaction

2016-10-21 Thread Diego Biurrun
sigaction is not defined in standards as a struct starting with another struct. This partially reverts a92be9b856bd11b081041c43c25d442028fe9a63. --- Alternative version as suggested by Rémi. libavdevice/bktr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git

[libav-devel] [PATCH] Replace leftover uses -aframes|-dframes|-vframes with -frames:a|d|v

2016-10-21 Thread Diego Biurrun
--- doc/avconv.texi | 13 - tests/fate/h264.mak | 2 +- tests/fate/video.mak | 4 ++-- 3 files changed, 11 insertions(+), 8 deletions(-) diff --git a/doc/avconv.texi b/doc/avconv.texi index 9f2f295..002dba3 100644 --- a/doc/avconv.texi +++ b/doc/avconv.texi @@ -343,7 +343,8

Re: [libav-devel] [PATCH 1/2] rtmpdh: Don't use the OpenSSL DH struct

2016-10-21 Thread Diego Biurrun
On Fri, Oct 21, 2016 at 02:02:43PM +0300, Martin Storsjö wrote: > Instead use our own struct, which we already use when using > gcrypt and gnutls. > > In OpenSSL 1.1, the DH struct has been made opaque. > --- > This produces the following warnings when building with gnutls: > warning: unused

[libav-devel] [PATCH 1/2] rtmpdh: Don't use the OpenSSL DH struct

2016-10-21 Thread Martin Storsjö
Instead use our own struct, which we already use when using gcrypt and gnutls. In OpenSSL 1.1, the DH struct has been made opaque. --- This produces the following warnings when building with gnutls: warning: unused label 'fail' [-Wunused-label] Any suggestions on avoiding that? ---

[libav-devel] [PATCH 2/2] openssl: Support version 1.1.0.

2016-10-21 Thread Martin Storsjö
From: Matt Oliver Further simplifications by Martin Storsjö, to minimize the diff. --- configure | 3 ++- libavformat/tls_openssl.c | 37 +++-- 2 files changed, 37 insertions(+), 3 deletions(-) diff --git a/configure

Re: [libav-devel] [PATCH 1/2] hevc: x86: Add add_residual optimizations

2016-10-21 Thread Diego Biurrun
On Fri, Oct 21, 2016 at 09:41:28AM +0200, Alexandra Hájková wrote: > --- /dev/null > +++ b/libavcodec/x86/hevc_add_res.asm > @@ -0,0 +1,371 @@ > +%macro ADD_RES_MMX_4_8 0 > +mova m0, [r1] > +mova m2, [r1+8] > +pxor m1, m1 > +pxor

Re: [libav-devel] [PATCH 01/12] yadif: x86 assembly for 16-bit samples

2016-10-21 Thread Anton Khirnov
Quoting Janne Grunau (2016-10-16 20:10:51) > From: James Darnley > > This is a fairly dumb copy of the assembly for 8-bit samples but it > works and produces identical output to the C version. The options have > been tested on an Athlon64 and a Core2Quad. > > Athlon64:

Re: [libav-devel] [PATCH 03/12] fate: yadif: add >8 bit tests

2016-10-21 Thread Anton Khirnov
Quoting Janne Grunau (2016-10-16 20:10:53) > From: Christophe Gisquet > > Signed-off-by: Janne Grunau > --- > tests/fate/filter-video.mak | 6 ++ > tests/ref/fate/filter-yadif10 | 31 +++ >

[libav-devel] [PATCH 1/2] hevc: x86: Add add_residual optimizations

2016-10-21 Thread Alexandra Hájková
From: Pierre Edouard Lepere Initially written by Pierre Edouard Lepere , extended by James Almer . Signed-off-by: Alexandra Hájková Signed-off-by: Diego Biurrun

[libav-devel] [PATCH 2/2] checkasm: Add a test for HEVC add_residual

2016-10-21 Thread Alexandra Hájková
--- Applied all Henrik's and Diego's review comments: fixed stride, and size sizes, consistent naming tests/checkasm/Makefile | 2 +- tests/checkasm/checkasm.c | 1 + tests/checkasm/checkasm.h | 1 + tests/checkasm/hevc_add_res.c | 85 +++

Re: [libav-devel] [PATCH 2/3] vaapi_decode: Clear parameter buffers to fix picture reuse

2016-10-21 Thread Anton Khirnov
Quoting Mark Thompson (2016-10-20 00:59:56) > When decoding interlaced pictures, the structure is reused to render > to the same surface twice. The parameter buffers were not being > cleared, which caused the i965 driver to error out. > --- > This fixes the error Luca was seeing (unrelated to the

Re: [libav-devel] [PATCH 21/24] examples/decode_video: remove a stray unrelated comment

2016-10-21 Thread Anton Khirnov
Quoting Diego Biurrun (2016-10-20 18:27:29) > On Thu, Oct 20, 2016 at 04:07:39PM +0200, Anton Khirnov wrote: > > --- a/doc/examples/decode_video.c > > +++ b/doc/examples/decode_video.c > > @@ -99,8 +99,6 @@ int main(int argc, char **argv) > > exit(1); > > } > > > > -/* the