ping on this and 6/6
--
Anton Khirnov
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
On Thu, 13 Sep 2012, Luca Barbato wrote:
> Both in usage of FASTDIV the denominator might be 1.
>
> Using a branch would make the function slower than using a normal
> division.
Which CPU do you have in mind where a division is faster than a branch
miss, let alone a predictable branch?
(No object
---
This iteration fixes some #includes.
libavcodec/x86/Makefile |1 +
libavcodec/x86/dsputil_mmx.c | 32 +---
libavcodec/x86/dsputil_mmx.h |1 +
libavcodec/x86/idct_init.c | 55 ++
4 files changed, 59 insertions(+), 3
---
libavcodec/x86/Makefile |1 +
libavcodec/x86/dsputil_mmx.c | 33 ++---
libavcodec/x86/dsputil_mmx.h |1 +
libavcodec/x86/idct_init.c | 54 ++
4 files changed, 59 insertions(+), 30 deletions(-)
create mode 100644 li
On Fri, Sep 14, 2012 at 12:36:24AM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > On Fri, Sep 14, 2012 at 12:14:45AM +0100, Måns Rullgård wrote:
> >> Diego Biurrun writes:
> >> >
> >> > --- a/libavcodec/x86/Makefile
> >> > +++ b/libavcodec/x86/Makefile
> >> > @@ -7,7 +7,6 @@ OBJS-$(CONFI
On Fri, Sep 14, 2012 at 12:37:40AM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > On Fri, Sep 14, 2012 at 12:17:51AM +0100, Måns Rullgård wrote:
> >> Diego Biurrun writes:
> >> > ---
> >> > The init function might better be moved to a separate file, but I don't
> >> > see a suitable one.
Diego Biurrun writes:
> ---
> libavcodec/x86/dsputil_mmx.c | 27 ---
> libavcodec/x86/idct_mmx_xvid.c | 25 +
> libavcodec/x86/idct_xvid.h |8
> 3 files changed, 33 insertions(+), 27 deletions(-)
OK
--
Måns Rullgård
m...@m
Diego Biurrun writes:
> On Fri, Sep 14, 2012 at 12:17:51AM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> > ---
>> > The init function might better be moved to a separate file, but I don't
>> > see a suitable one...
>> >
>> > libavcodec/x86/dsputil_mmx.c | 62
>> > ++
Diego Biurrun writes:
> On Fri, Sep 14, 2012 at 12:14:45AM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> >
>> > --- a/libavcodec/x86/Makefile
>> > +++ b/libavcodec/x86/Makefile
>> > @@ -7,7 +7,6 @@ OBJS-$(CONFIG_VP3DSP) += x86/vp3dsp_init.o
>> >
>> > MMX-OBJS
---
libavcodec/x86/dsputil_mmx.c | 27 ---
libavcodec/x86/idct_mmx_xvid.c | 25 +
libavcodec/x86/idct_xvid.h |8
3 files changed, 33 insertions(+), 27 deletions(-)
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsp
On Fri, Sep 14, 2012 at 12:17:51AM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > ---
> > The init function might better be moved to a separate file, but I don't
> > see a suitable one...
> >
> > libavcodec/x86/dsputil_mmx.c | 62
> > +++---
> > 1 f
On Fri, Sep 14, 2012 at 12:14:45AM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> >
> > --- a/libavcodec/x86/Makefile
> > +++ b/libavcodec/x86/Makefile
> > @@ -7,7 +7,6 @@ OBJS-$(CONFIG_VP3DSP) += x86/vp3dsp_init.o
> >
> > MMX-OBJS += x86/ds
Diego Biurrun writes:
> ---
> The init function might better be moved to a separate file, but I don't
> see a suitable one...
>
> libavcodec/x86/dsputil_mmx.c | 62
> +++---
> 1 files changed, 34 insertions(+), 28 deletions(-)
What is the purpose of this?
Diego Biurrun writes:
> ---
> libavcodec/x86/Makefile |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
> index be4fd54..1b0574b 100644
> --- a/libavcodec/x86/Makefile
> +++ b/libavcodec/x86/Makefile
> @@ -7,7 +7,6 @@
---
The init function might better be moved to a separate file, but I don't
see a suitable one...
libavcodec/x86/dsputil_mmx.c | 62 +++---
1 files changed, 34 insertions(+), 28 deletions(-)
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_
---
libavcodec/x86/dsputil_mmx.h|2 ++
libavcodec/x86/dsputilenc_mmx.c | 13 +
libavcodec/x86/fdct.c | 16
3 files changed, 19 insertions(+), 12 deletions(-)
diff --git a/libavcodec/x86/dsputil_mmx.h b/libavcodec/x86/dsputil_mmx.h
index 00c7fb1..
---
libavcodec/x86/Makefile |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index be4fd54..1b0574b 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -7,7 +7,6 @@ OBJS-$(CONFIG_VP3DSP) +
On 2012-09-13 23:58:38 +0200, Luca Barbato wrote:
> In both usages of FASTDIV the denominator might be 1.
>
> Using a branch would make the function slower than using a normal
> division.
>
> Both denominator and numerator can be multiplied by 2 safely and
> using shifts is faster than using a br
In both usages of FASTDIV the denominator might be 1.
Using a branch would make the function slower than using a normal
division.
Both denominator and numerator can be multiplied by 2 safely and
using shifts is faster than using a branch.
---
libavcodec/vorbisdec.c | 5 ++---
1 file changed, 2 i
Diego Biurrun writes:
> On Thu, Sep 13, 2012 at 12:12:19PM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> > ---
>> > This could alternatively be squashed together with 2/3.
>> >
>> > libavcodec/dsputil.c|4 +++-
>> > libavcodec/dsputil.h|1 +
>> > libavc
Diego Biurrun writes:
> On Thu, Sep 13, 2012 at 12:07:05PM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> > On Wed, Sep 12, 2012 at 12:02:45PM +0100, Måns Rullgård wrote:
>> >> Diego Biurrun writes:
>> >> > On Tue, Sep 11, 2012 at 05:17:11PM +0300, Martin Storsjö wrote:
>> >> >> On Tue
On Thu, Sep 13, 2012 at 12:12:19PM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > ---
> > This could alternatively be squashed together with 2/3.
> >
> > libavcodec/dsputil.c|4 +++-
> > libavcodec/dsputil.h|1 +
> > libavcodec/x86/dsputil_mmx.h|2
On Thu, Sep 13, 2012 at 12:07:05PM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > On Wed, Sep 12, 2012 at 12:02:45PM +0100, Måns Rullgård wrote:
> >> Diego Biurrun writes:
> >> > On Tue, Sep 11, 2012 at 05:17:11PM +0300, Martin Storsjö wrote:
> >> >> On Tue, 11 Sep 2012, Diego Biurrun wr
On Thu, Sep 13, 2012 at 02:25:39PM +0200, Luca Barbato wrote:
> On 9/13/12 1:01 PM, Diego Biurrun wrote:
> >+return AVERROR_UNKNOWN;
> AVERROR(ENOMEM) maybe?
Sounds sensible - Nathan?
Diego
___
libav-devel mailing list
libav-devel@libav.org
http
On Thu, Sep 13, 2012 at 08:18:56PM +0100, Mans Rullgard wrote:
> Some compilers are extra strict about register usage in main(),
> disallowing ebp in inline asm there while allowing it elsewhere.
> This change makes the test better reflect actual usage.
>
> Signed-off-by: Mans Rullgard
> ---
> c
Signed-off-by: Mans Rullgard
---
libavcodec/ac3dec.c | 32 ++--
libavcodec/ac3dec.h | 1 -
2 files changed, 6 insertions(+), 27 deletions(-)
diff --git a/libavcodec/ac3dec.c b/libavcodec/ac3dec.c
index 37426c6..f1bda7f 100644
--- a/libavcodec/ac3dec.c
+++ b/libavcode
---
libavcodec/ac3dec.c | 14 +++--
libavcodec/ac3dec.h | 4 +++
libavcodec/ac3dsp.c | 2 +-
libavcodec/ac3dsp.h | 2 +-
libavcodec/x86/ac3dsp_init.c | 70 ++--
5 files changed, 59 insertions(+), 33 deletions(-)
di
Signed-off-by: Mans Rullgard
---
libavcodec/ac3dec.c | 27 ---
1 file changed, 20 insertions(+), 7 deletions(-)
diff --git a/libavcodec/ac3dec.c b/libavcodec/ac3dec.c
index 8ddf354..3f6c92d 100644
--- a/libavcodec/ac3dec.c
+++ b/libavcodec/ac3dec.c
@@ -188,7 +188,6 @@ sta
Some compilers are extra strict about register usage in main(),
disallowing ebp in inline asm there while allowing it elsewhere.
This change makes the test better reflect actual usage.
Signed-off-by: Mans Rullgard
---
configure | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --gi
On 09/13/2012 08:30 PM, Diego Elio Pettenò wrote:
> On 13/09/2012 10:45, Luca Barbato wrote:
>> Both in usage of FASTDIV the denominator might be 1.
>
> "In both usages of FASTDIV" I guess?
>
Yes.
___
libav-devel mailing list
libav-devel@libav.org
http
On 13/09/2012 10:45, Luca Barbato wrote:
> Both in usage of FASTDIV the denominator might be 1.
"In both usages of FASTDIV" I guess?
--
Diego Elio Pettenò — Flameeyes
flamee...@flameeyes.eu — http://blog.flameeyes.eu/
___
libav-devel mailing list
libav
On 09/13/2012 07:51 PM, Janne Grunau wrote:
> On 2012-09-13 19:45:21 +0200, Luca Barbato wrote:
>> Both in usage of FASTDIV the denominator might be 1.
>>
>> Using a branch would make the function slower than using a normal
>> division.
>>
>> Both denominator and numerator can be multiplied by 2 sa
On 2012-09-13 19:45:21 +0200, Luca Barbato wrote:
> Both in usage of FASTDIV the denominator might be 1.
>
> Using a branch would make the function slower than using a normal
> division.
>
> Both denominator and numerator can be multiplied by 2 safely and
> using shifts is faster than using a bra
Both in usage of FASTDIV the denominator might be 1.
Using a branch would make the function slower than using a normal
division.
Both denominator and numerator can be multiplied by 2 safely and
using shifts is faster than using a branch.
---
libavcodec/vorbisdec.c |7 ---
1 files changed
On 2012-09-10 16:13:26 +0100, Måns Rullgård wrote:
> Signed-off-by: Mans Rullgard
> ---
> libavresample/arm/audio_convert_init.c | 5 +
> libavresample/arm/audio_convert_neon.S | 233
> +
> 2 files changed, 238 insertions(+)
lgtm
Janne
Hi,
On Sep 13, 2012 4:16 AM, "Martin Storsjö" wrote:
>
> On Thu, 13 Sep 2012, Martin Storsjö wrote:
>
>> This fixes building of this tool on mingw32ce.
>> ---
>> tools/trasher.c |2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/tools/trasher.c b/tools/trasher.c
>> index 11605b8..356
On Thu, 13 Sep 2012, Martin Storsjö wrote:
On Thu, 13 Sep 2012, Samuel Pitoiset wrote:
+av_log(s1, AV_LOG_WARNING,
+ "Only 8-bit precision is supported.\n");
+
+/* a quantization table is 64 bytes long */
+qtable_len = AV_RB16(&buf[
On 9/13/12 1:01 PM, Diego Biurrun wrote:
+return AVERROR_UNKNOWN;
AVERROR(ENOMEM) maybe?
lu
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
On Thu, 13 Sep 2012, Samuel Pitoiset wrote:
---
Changelog | 1 +
libavformat/Makefile | 1 +
libavformat/rtpenc.c | 4 ++
libavformat/rtpenc.h | 1 +
libavformat/rtpenc_jpeg.c | 120 ++
libavformat/sdp.c | 5
Martin Storsjö writes:
> ---
> configure|2 ++
> libavformat/os_support.h |6 +++---
> 2 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/configure b/configure
> index 5afdab9..2e28124 100755
> --- a/configure
> +++ b/configure
> @@ -1132,6 +1132,7 @@ HAVE_
---
configure|2 ++
libavformat/os_support.h |6 +++---
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/configure b/configure
index 5afdab9..2e28124 100755
--- a/configure
+++ b/configure
@@ -1132,6 +1132,7 @@ HAVE_LIST="
dev_ic_bt8xx_h
dev_video_bk
On Thu, 13 Sep 2012, Martin Storsjö wrote:
This fixes building of this tool on mingw32ce.
---
tools/trasher.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/trasher.c b/tools/trasher.c
index 11605b8..35625e9 100644
--- a/tools/trasher.c
+++ b/tools/trasher.c
@@ -21,6 +21,8 @@
#inc
This is required for for reading high bit depth raw yuv data.
---
libavformat/rawdec.c | 1 +
libavformat/rawdec.h | 1 +
libavformat/rawvideodec.c | 1 +
3 files changed, 3 insertions(+)
diff --git a/libavformat/rawdec.c b/libavformat/rawdec.c
index a64c406..541a1d9 100644
--- a/libavf
Martin Storsjö writes:
> On Thu, 13 Sep 2012, Måns Rullgård wrote:
>
>> Martin Storsjö writes:
>>
>>> Windows CE doesn't have neither mkdir nor _mkdir officially (only
>>> CreateDirectoryW), but mingw32ce has compat wrappers with these names
>>> (declared in io.h since direct.h is unavailable).
Diego Biurrun writes:
> ---
> This could alternatively be squashed together with 2/3.
>
> libavcodec/dsputil.c|4 +++-
> libavcodec/dsputil.h|1 +
> libavcodec/x86/dsputil_mmx.h|2 --
> libavcodec/x86/dsputilenc_mmx.c |2 --
> 4 files changed, 4 insert
Diego Biurrun writes:
> ---
> This could alternatively be squashed together with 3/3.
>
> libavcodec/dsputil.c |3 +++
> libavcodec/dsputil.h |3 +++
> libavcodec/x86/dsputil_mmx.c |3 ---
> libavcodec/x86/dsputil_mmx.h |1 -
> 4 files changed, 6 insertions(+), 4
On Thu, 13 Sep 2012, Måns Rullgård wrote:
Martin Storsjö writes:
Windows CE doesn't have neither mkdir nor _mkdir officially (only
CreateDirectoryW), but mingw32ce has compat wrappers with these names
(declared in io.h since direct.h is unavailable).
---
libavformat/os_support.h |4
Diego Biurrun writes:
> ---
> libavcodec/dct-test.c|4
> libavcodec/x86/dsputil_mmx.h |4
> 2 files changed, 0 insertions(+), 8 deletions(-)
>
> diff --git a/libavcodec/dct-test.c b/libavcodec/dct-test.c
> index 72d2d80..848ba8a 100644
> --- a/libavcodec/dct-test.c
> ++
Diego Biurrun writes:
> On Wed, Sep 12, 2012 at 12:02:45PM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> > On Tue, Sep 11, 2012 at 05:17:11PM +0300, Martin Storsjö wrote:
>> >> On Tue, 11 Sep 2012, Diego Biurrun wrote:
>> >> >On Mon, Sep 03, 2012 at 11:54:20AM +0200, Diego Biurrun wrot
Martin Storsjö writes:
> Windows CE doesn't have neither mkdir nor _mkdir officially (only
> CreateDirectoryW), but mingw32ce has compat wrappers with these names
> (declared in io.h since direct.h is unavailable).
> ---
> libavformat/os_support.h |4
> 1 file changed, 4 insertions(+)
>
This fixes building of this tool on mingw32ce.
---
tools/trasher.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/trasher.c b/tools/trasher.c
index 11605b8..35625e9 100644
--- a/tools/trasher.c
+++ b/tools/trasher.c
@@ -21,6 +21,8 @@
#include
#include
#include
+#include
+#in
---
This could alternatively be squashed together with 3/3.
libavcodec/dsputil.c |3 +++
libavcodec/dsputil.h |3 +++
libavcodec/x86/dsputil_mmx.c |3 ---
libavcodec/x86/dsputil_mmx.h |1 -
4 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/libavcodec/d
---
libavcodec/dct-test.c|4
libavcodec/x86/dsputil_mmx.h |4
2 files changed, 0 insertions(+), 8 deletions(-)
diff --git a/libavcodec/dct-test.c b/libavcodec/dct-test.c
index 72d2d80..848ba8a 100644
--- a/libavcodec/dct-test.c
+++ b/libavcodec/dct-test.c
@@ -48,16 +48,1
---
This could alternatively be squashed together with 2/3.
libavcodec/dsputil.c|4 +++-
libavcodec/dsputil.h|1 +
libavcodec/x86/dsputil_mmx.h|2 --
libavcodec/x86/dsputilenc_mmx.c |2 --
4 files changed, 4 insertions(+), 5 deletions(-)
diff --git a/l
From: Nicolas George
---
Rebased on top of HEAD; some cosmetics.
Changelog |1 +
configure |5 +
doc/general.texi|2 +
libavcodec/Makefile |1 +
libavcodec/allcodecs.c |1 +
libavcodec/libopusdec.c | 201 ++
On Wed, Sep 12, 2012 at 12:02:45PM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > On Tue, Sep 11, 2012 at 05:17:11PM +0300, Martin Storsjö wrote:
> >> On Tue, 11 Sep 2012, Diego Biurrun wrote:
> >> >On Mon, Sep 03, 2012 at 11:54:20AM +0200, Diego Biurrun wrote:
> >> >>---
> >> >> libavcod
On 9/13/12 12:27 PM, Anton Khirnov wrote:
On Mon, 10 Sep 2012 13:26:21 +0200, Luca Barbato wrote:
Do not crash if the input format is not allocated yet.
When would that happen?
Some people might fail before allocating it. It had been a question at
the VDD.
Another option might be try to
On Mon, 10 Sep 2012 13:26:21 +0200, Luca Barbato wrote:
> Do not crash if the input format is not allocated yet.
When would that happen?
--
Anton Khirnov
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/liba
On Thu, Sep 13, 2012 at 11:40:43AM +0300, Martin Storsjö wrote:
> This fixes building without inline assembly.
> ---
> libavcodec/x86/ac3dsp_init.c |2 ++
> 1 file changed, 2 insertions(+)
OK
Diego
___
libav-devel mailing list
libav-devel@libav.org
On 2012-09-10 13:26:21 +0200, Luca Barbato wrote:
> Do not crash if the input format is not allocated yet.
> ---
> libavformat/utils.c | 17 +
> 1 file changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/libavformat/utils.c b/libavformat/utils.c
> index 3a829a9..c0da7f2 100
On 2012-09-10 13:26:20 +0200, Luca Barbato wrote:
> avio_close checks by itself for NULL condition.
> ---
> libavformat/utils.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/libavformat/utils.c b/libavformat/utils.c
> index 81a4b34..3a829a9 100644
> --- a/libavforma
On 9/10/12 1:26 PM, Luca Barbato wrote:
avio_close checks by itself for NULL condition.
---
libavformat/utils.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavformat/utils.c b/libavformat/utils.c
index 81a4b34..3a829a9 100644
--- a/libavformat/utils.c
+++ b/libavf
On 9/13/12 11:36 AM, Martin Storsjö wrote:
Windows CE doesn't have neither mkdir nor _mkdir officially (only
CreateDirectoryW), but mingw32ce has compat wrappers with these names
(declared in io.h since direct.h is unavailable).
---
libavformat/os_support.h |4
1 file changed, 4 inser
Windows CE doesn't have neither mkdir nor _mkdir officially (only
CreateDirectoryW), but mingw32ce has compat wrappers with these names
(declared in io.h since direct.h is unavailable).
---
libavformat/os_support.h |4
1 file changed, 4 insertions(+)
diff --git a/libavformat/os_support.h
On 9/13/12 10:40 AM, Martin Storsjö wrote:
This fixes building without inline assembly.
---
libavcodec/x86/ac3dsp_init.c |2 ++
1 file changed, 2 insertions(+)
Ok.
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mail
This fixes building without inline assembly.
---
libavcodec/x86/ac3dsp_init.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/libavcodec/x86/ac3dsp_init.c b/libavcodec/x86/ac3dsp_init.c
index b23a9a1..5008d65 100644
--- a/libavcodec/x86/ac3dsp_init.c
+++ b/libavcodec/x86/ac3dsp_init.c
@@
66 matches
Mail list logo
