On Tue, 2010-03-30 at 12:33 +0800, Kite Lau wrote:
Hi list:
libva is not working and you might find some information below. I wonder
whether it is correct to use i915 module for intel i3-530 arrandale GPU?
Will libva work with it?
bash-4.1$ vainfo
libva: libva version 0.31.0
vainfo:
On Mon, 2010-03-29 at 18:58 +0800, Gwenole Beauchesne wrote:
Hi,
On Mon, 29 Mar 2010, Xiang, Haihao wrote:
The reference frame array in mplayer-vaapi-20090828 is filled with
h-ref_list however this array in mplayer-vaapi-20100224 is
filled with h-short_ref/h-long_ref except GMA500
On Tue, 2010-04-06 at 22:29 +0800, Kite Lau wrote:
于 2010年04月06日 17:27, Xiang, Haihao 写道:
On Sun, 2010-04-04 at 10:33 +0800, Kite Lau wrote:
于 2010年04月04日 10:22, Kite Lau 写道:
Now mplayer-vaapi plays h264 files, but the screen broken. See attached
screenshot.
screenshot is here
On Mon, 2010-07-12 at 00:00 +0800, Gwenole Beauchesne wrote:
Hi,
Here is a series of patches for Intel Ironlake driver. Those are the
patches from my tree and considered complete. Others will be
proposed later. e.g. vaPutImage() implementation.
Thanks a lot.
It seems some patches
On Sun, 2010-07-11 at 23:44 +0800, Gwenole Beauchesne wrote:
Hi,
This patch adds RGBA subpictures. Should we support more than 1
subpicture per surface, this will also fix Gnash (Flash Player) support.
Applied.
Thanks
Haihao
Regards,
Gwenole.
On Sun, 2010-07-11 at 23:48 +0800, Gwenole Beauchesne wrote:
Hi,
This patch adds vaGetImage() support. This now supports sub-regions of
a VA surface. vaGetImage() is used is used by VLC and possibly another
player.
Applied
Thanks
Haihao.
Regards,
Gwenole.
On Wed, 2010-09-08 at 05:31 +0800, Andrés Revilla wrote:
Hi ..
Currently i am compiling and using these :
libva: libva version 0.31.1-sds1
libva: va_getDriverName() returns 0
libva: Trying to open /usr/lib/va/drivers/i965_drv_
video.so
libva: va_openDriver() returns 0
vainfo: VA API
://bugs.freedesktop.org/show_bug.cgi?id=29977#c7
Signed-off-by: Kristian Høgsberg k...@bitplanet.net
---
Could someone experienced find out the problem and report to mesa
developers please?
On 6 November 2010 15:52, Xiang, Haihao haihao.xi...@intel.com wrote
Hi,
It seems mplayer_vaapi always uses the first profile in the supported
list and passes it to vaapi driver via vaCreateConfig. Sometimes the
vaapi driver needs to know which profile the decoded bitstream is in. Is
there a way to pass the right profile to vaapi driver?
Thanks
Haihao
On Sun, 2010-11-21 at 18:59 +0800, Oleksiy Krivoshey wrote:
Do you know when your patch is going to be incorporated? Or maybe is
there a git branch already?
Please try with the latest master branch.
Thanks
Haihao
On 8 November 2010 07:38, Xiang, Haihao haihao.xi...@intel.com wrote:
I have
Hi, Austin
VC1 has a profile flag in the sequence layer, however
VAPictureParameterBufferVC1 hasn't a member for this flag. The i965
video driver needs this flag, does the parameter 'profile' passed in
vaCreateConfig match the profile specified in the sequence layer?
Thanks
Haihao
Hi,
On Tue, 2010-12-21 at 11:20 +0800, Yuan, Shengquan wrote:
On Fri, Dec 17, 2010 at 12:12 AM, Damien Lespiau
damien.lesp...@gmail.com wrote:
Hi,
I've a few patches around for build related little details mainly.
1/2/3 were committed. Haihao can look at 4/5
Applied.
Thanks
Haihao
Signed-off-by: Xiang, Haihao haihao.xi...@intel.com
---
test/vainfo/vainfo.c | 24
1 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/test/vainfo/vainfo.c b/test/vainfo/vainfo.c
index 96b7142..20e80a1 100644
--- a/test/vainfo/vainfo.c
+++ b/test/vainfo
Hi,
The following BSD check should be moved to intel_batchbuffer_init(). Others are
ok for me.
+if (intel-has_bsd) {
+intel-batch_bcs = calloc(1, sizeof(*(intel-batch_bcs)));
+assert(intel-batch_bcs);
+intel-batch_bcs-intel = intel;
+intel-batch_bcs-flag =
-
From: gbeauche...@splitted-desktop.com
[mailto:gbeauche...@splitted-desktop.com]
Sent: Wednesday, November 24, 2010 3:35 PM
To: Xiang, Haihao
Cc: Yuan, Shengquan; Yuan, Shengquan; libva@lists.freedesktop.org
Subject: Re: [Libva] profile used in mplayer vaapi
Hi,
On Tue, 2010-11-23 at 14:25
On Thu, 2011-02-10 at 08:53 +0800, Paul Bender wrote:
Commit 95da3e7
http://cgit.freedesktop.org/libva/commit/?id=95da3e742bee58383f575bdf1ebcb424fc1ad1a8,
which makes vtable a pointer breaks the compilation of i965_drv_video
because it was not updated to treat vtable as a pointer.
I have
Hi,
We want to support CABAC/CAVLC for encoding in our libva driver,
however I can't find a flag for it in va.h. Do you know which flag can
be used to specify coding mode for encoding in VA API?
Thanks
Haihao
___
Libva mailing list
On Thu, 2011-05-05 at 09:18 +0800, hank peng wrote:
Hi, guys:
My platform's GUI is based on framebuffer, now I want to make use of
Sandy Bridge i5 processor's GPU to implement a video player and there
is no X11 running on my platform.
I noticed that on libva's web site it says the current
,
Alexander
P.S.
BTW, debug prints are switched in '/etc/libva.conf' file or in environment by
4 defined (details are in libva's readme).
I sent my proposals last week. My e-mail was banned because I was outside
'libva' group. I have recent it again.
-Original Message-
From: Xiang
Hi,
The objects used in i965 driver are allocated via special ioctl, so
valgrind can not track the pointer without patching libdrm.
Shuang has a patch for libdrm, maybe he can help you.
Thanks
Haihao
Hi,
it looks like libva is leaking memory, I played around with my xine vaapi
You need this kernel patch
drm/i915: fix user irq miss in BSD ring on g4x which has been added to
the 2.6.39-stable tree.
Thanks
Haihao
Hi,
when using the g45-h264 branch and the vaapi patched mplayer I get
below 1 FPS and the kernel is printing lots of errors. The used
Hardware is an
On Sat, 2011-06-11 at 14:54 +0800, gimli wrote:
This patch fixes the following memory leaks in libva :
1.) The dri drawables are allocated but not freed
2.) The drm buffer manager is allocated but not freed
3.) The device name is allocated but not freed
Signed-off-by: Edgar Hucek gimli at
Only a small piece of code is X11 related. scaling and color space
conversion all are independent of window system. What you need to do is
to define your drawable (parameter 'draw' of vaPutSurface), then render
the surface on this drawable.
2011/6/14 Bian, Jonathan jonathan.b...@intel.com:
On Wed, 2011-06-15 at 19:54 +0800, gimli wrote:
Are there any plans to support deinterlacing for mpeg2 content ?
You can play interlaced videos with wave effect. on SNB, we support
advanced deinterlacing for MPEG2.
Edgar (gimli) Hucek
___
Libva
On Thu, 2011-06-16 at 21:03 +0800, Gwenole Beauchesne wrote:
---
va/x11/va_x11.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/va/x11/va_x11.c b/va/x11/va_x11.c
index 0d62534..947b5b1 100644
--- a/va/x11/va_x11.c
+++ b/va/x11/va_x11.c
@@ -63,7 +63,7 @@ static
Hi,
I pulled the libva into sandybridge with the xorg.conf as follows.
start of file-
Section Device
IdentifierIntel Graphics Driver
Driverintel
EndSection
Section Monitor
IdentifierGeneric Monitor
Option
On Sat, 2011-06-18 at 00:08 +0800, keller-tec wrote:
Hi.
I have installed libva from g45-h264 branch, but vainfo still prints this:
libva: libva version 0.32.0
libva: va_getDriverName() returns 0
libva: Trying to open /usr/local/lib/dri/i965_drv_video.so
libva: va_openDriver() returns 0
.
tor 2011-06-09 klockan 10:41 +0800 skrev Xiang, Haihao:
You need this kernel patch
drm/i915: fix user irq miss in BSD ring on g4x which has been added to
the 2.6.39-stable tree.
Thanks
Haihao
Hi,
when using the g45-h264 branch and the vaapi patched mplayer I get
below
On Fri, 2011-07-01 at 11:44 +0800, hank peng wrote:
Hi:
I am using libva to decode H264 video on sandy bridge platform, I
found that interleaved_uv is set to 1 while decoding H264 and output
format is NV12. Since my display format is YUYV422 packed format, I
use libswscale provided by
call vaGetImage directly
This will make the conformance test fail because we need to hook
i965_PutSurface to dump the decoded yuv file. Is there any other method?
Thanks
Hai Lan
-Original Message-
From: libva-bounces+hai.lan=intel@lists.freedesktop.org
On Sat, 2011-08-13 at 23:28 +0800, hank peng wrote:
Hi, all:
I noticed that the dri_bo of a object surface is allocated using
drm_intel_bo_alloc_tiled function which returned USWC buffer, can I
use drm_intel_bo_alloc_for_render instead and then dri_bo_map to get
real decoded data back?
No
On Mon, 2011-08-15 at 15:57 +0800, hank peng wrote:
2011/8/15 Xiang, Haihao haihao.xi...@intel.com:
On Sat, 2011-08-13 at 23:28 +0800, hank peng wrote:
Hi, all:
I noticed that the dri_bo of a object surface is allocated using
drm_intel_bo_alloc_tiled function which returned USWC buffer
@lists.freedesktop.org
Cc: Xiang, Haihao
Subject: New intel-driver repository
Hi,
I have uploaded the Intel driver to new VA drivers home:
git://git.freedesktop.org/git/vaapi/intel-driver.git
The original tree still lays in the libva repository but I am waiting
for fd.o admins to fix the gitweb
Hi, Gwenole
There is an implementation of vaPutImage on vaapi-ext branch, the
implementation directly uses some internal post processing functions, it
supports pixel format conversion between NV12, YV12 and I420. Currently
it doesn't work on CTG due to lack of GPU kernels on CTG. Could you try
On Fri, 2011-09-09 at 12:57 +0800, Gwenole Beauchesne wrote:
Hi,
2011/9/6 Xiang, Haihao haihao.xi...@intel.com:
There is an implementation of vaPutImage on vaapi-ext branch, the
implementation directly uses some internal post processing functions, it
supports pixel format conversion
@lists.freedesktop.org] On
Behalf Of Sebastian Reichel
Sent: Monday, October 03, 2011 12:42 AM
To: libva@lists.freedesktop.org
Subject: Re: [Libva] G45 h264 decoding
On Sat, Oct 01, 2011 at 05:15:30PM +0800, Xiang, Haihao wrote:
There are some fixes for G45 h264 decoding on g45-h264 branch, please
Hi,
Works fine in my Vaio TT, although only with videos up to 720. With
1080 videos, cannot cope with the rate. The CPU is fine (around 20%)
so looks like it's an issue with the driver, as they play fine in
Windows.
On Sat, Oct 1, 2011 at 11:15 AM, Xiang, Haihao haihao.xi...@intel.com
On Sat, 2011-10-08 at 15:30 +0800, Gwenole Beauchesne wrote:
Hi,
I would suggest the following changes:
1. Remove the VA_JPEG_ROTATION_* definitions. They are the same as the
current VA_ROTATION_* definitions used as display attributes. Besides,
even if they were not, we could still use
On Tue, 2012-01-31 at 22:48 +0100, atechsys...@freenet.de wrote:
Hello,
I forward it to the vaapi mailinglist to get sure.
I will try to reproduce this attitude to check if I get also deinterlaced
output with or without the wrong color space. Till now I've not seen the
xine:
On Wed, 2012-02-01 at 10:32 +0800, Xiang, Haihao wrote:
On Tue, 2012-01-31 at 22:48 +0100, atechsys...@freenet.de wrote:
Hello,
I forward it to the vaapi mailinglist to get sure.
I will try to reproduce this attitude to check if I get also deinterlaced
output with or without
Hi,
Could you try with the latest vaapi-ext branch?
BTW many operations such as vaDeriveImage(), vaPutImage() are
unnecessary for decoding in xine-lib-vaapi
Thanks
Haihao
Xiang, Haihao haihao.xiang@... writes:
On Tue, 2012-01-31 at 22:48 +0100, atechsystem at freenet.de wrote
On Mon, 2012-01-30 at 19:27 +0100, Gwenole Beauchesne wrote:
Signed-off-by: Gwenole Beauchesne gwenole.beauche...@intel.com
---
src/gen7_mfd.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/gen7_mfd.c b/src/gen7_mfd.c
index 9308999..c289427 100644
---
On Thu, 2012-02-02 at 23:12 +0100, Christoph Evers wrote:
Hi Haihao,
Could you try with the latest vaapi-ext branch?
I am already on latest vaapi-ext branch except changes done in the last
24hours as git is not reachable.
I pushed some commits yesterday, please give a try.
Thanks
-Original Message-
From: gimli [mailto:gi...@dark-green.com]
Sent: Friday, February 03, 2012 4:19 PM
To: Xiang, Haihao
Cc: libva@lists.freedesktop.org
Subject: Re: [Libva] vaapi intel-driver (vaapi-ext): assertion failed
And you look at the source code at
https://github.com
+vpos = slice_param-slice_vertical_position;
+if (vpos = mb_height || vpos == last_vpos + 2)
+return 1;
Why is the workaround applied for vpos == last_vpos + 2 ? I think it is
right that vpos = 0, 2, 4, ... for top field and vpos = 1, 3, 5, .. for
On Thu, 2012-03-29 at 09:01 +0200, Gwenole Beauchesne wrote:
Hi,
2012/3/29 Xiang, Haihao haihao.xi...@intel.com:
+vpos = slice_param-slice_vertical_position;
+if (vpos = mb_height || vpos == last_vpos + 2)
+return 1;
Why is the workaround
-Original Message-
From: libva-bounces+haihao.xiang=intel@lists.freedesktop.org
[mailto:libva-bounces+haihao.xiang=intel@lists.freedesktop.org] On
Behalf Of Haitao Huang
Sent: Thursday, May 17, 2012 3:48 PM
To: libva@lists.freedesktop.org
Subject: [Libva] [PATCH 1/2]
It is OK for me.
Thanks
Haihao
-Original Message-
From: libva-bounces+haihao.xiang=intel@lists.freedesktop.org
[mailto:libva-bounces+haihao.xiang=intel@lists.freedesktop.org] On
Behalf Of Haitao Huang
Sent: Thursday, May 17, 2012 3:48 PM
To: libva@lists.freedesktop.org
-Original Message-
From: libva-bounces+haihao.xiang=intel@lists.freedesktop.org
[mailto:libva-bounces+haihao.xiang=intel@lists.freedesktop.org] On
Behalf Of Matt Turner
Sent: Friday, May 18, 2012 7:40 PM
To: Huang, Haitao
Cc: libva@lists.freedesktop.org
Subject: Re: [Libva]
Hi, Gwenole
JPEG decoding is broken by your revised API on staging branch (libva
and intel-driver), you can run loadjpeg in test/decode to reproduce the
issue. Could you take a look at and fix this issue ?
Thanks
Haihao
___
Libva mailing list
On Tue, 2012-06-05 at 11:07 +0200, Mario Schulz wrote:
Has this come to a halt?
All development has been moved to staging branch
Thanks
Haihao
When can we expect Motion Compensated/Adaptive from vaapi-ext in staging?
I still think this would sell some HTPC boards and processors if you
On Tue, 2012-06-26 at 10:32 +0300, Zhao Halley wrote:
---
src/i965_drv_video.c |2 +-
src/i965_drv_video.h |2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
mode change 100644 = 100755 src/i965_drv_video.h
diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
index
On Tue, 2012-06-26 at 10:00 +0200, Gwenole Beauchesne wrote:
Hi,
2012/6/26 Zhao Halley halley.z...@intel.com:
} else if (fourcc == VA_FOURCC('I', 'M', 'C', '1') ||
- fourcc == VA_FOURCC('I', 'M', 'C', '3')) {
+ fourcc == VA_FOURCC('I', 'M', 'C', '3') ||
On Tue, 2012-06-26 at 09:56 +0200, Gwenole Beauchesne wrote:
Hi,
2012/6/26 Zhao Halley halley.z...@intel.com:
-if (HAS_ACCELERATED_PUTIMAGE(i965))
+if (HAS_ACCELERATED_PUTIMAGE(i965) obj_image-image.format.fourcc
!= obj_surface-fourcc)
va_status =
On Wed, 2012-06-27 at 00:49 +, Zhao, Halley wrote:
While we are at adding new formats, please define corresponding
VA_FOURCC_* if they don't exist already. They will be propagated to
libva next, with proper comments as to what they actually represent.
Ok, I will add it to va.h.
Should
On Wed, 2012-06-27 at 15:34 +0200, Gwenole Beauchesne wrote:
Hi,
I pushed the following as obvious. Tested with GStreamer only.
MPlayer has internal limitations to 2K pixels in either dimension.
I can play a 3840x2160 video by MPlayer-vaapi on my machine.
...
VO: [vaapi] 3840x2160 =
Pushed.
Thanks
Haihao
---
src/i965_post_processing.c |8 ++--
1 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/src/i965_post_processing.c b/src/i965_post_processing.c
index 4bb2a74..c728936 100755
--- a/src/i965_post_processing.c
+++ b/src/i965_post_processing.c
Pushed.
Thanks
Haihao
---
src/i965_drv_video.c |8 +---
1 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
index 02dfacc..6b44bba 100755
--- a/src/i965_drv_video.c
+++ b/src/i965_drv_video.c
@@ -3569,7 +3569,7 @@
On Tue, 2012-07-03 at 10:24 +0200, Gwenole Beauchesne wrote:
Hi,
2012/7/2 Josep Torra n770gal...@gmail.com:
For long I'm aware of certain ambiguity in libva regarding how
vaBuffers have to be handled.
I have to concur.
Which make sense because there's no way for the vaapi client
On Wed, 2012-07-04 at 10:13 +0800, Xiang, Haihao wrote:
On Tue, 2012-07-03 at 10:24 +0200, Gwenole Beauchesne wrote:
Hi,
2012/7/2 Josep Torra n770gal...@gmail.com:
For long I'm aware of certain ambiguity in libva regarding how
vaBuffers have to be handled.
I have to concur
On Tue, 2012-07-03 at 21:28 +0100, Joe Bloggsian wrote:
It is great that intel have released proper documentation for the Ivy
bridge GPU/media engines. Kudos to them; if only all silicon vendors
were so progressive!
In particular it is very interesting to see the detailed information on
On Mon, 2012-07-02 at 18:25 +0200, Gwenole Beauchesne wrote:
Signed-off-by: Gwenole Beauchesne gwenole.beauche...@intel.com
---
configure.ac |2 +
test/Makefile.am |2 +-
test/common/Makefile.am| 48 +++
On Mon, 2012-07-02 at 18:25 +0200, Gwenole Beauchesne wrote:
From: Benjamin Franzke benjaminfran...@googlemail.com
Add va/va_drmcommon.h header for DRM-based drivers to use common
base utilities and data structures. So far, only the DRM state is
shared to that to maintain binary
On Thu, 2012-07-05 at 07:24 +0200, Gwenole Beauchesne wrote:
2012/7/5 Xiang, Haihao haihao.xi...@intel.com:
+/** \brief Base DRM state. */
+struct drm_state {
+/** \brief DRM connection descriptor. */
+int fd;
+/** \brief DRM connection type. */
+int
On Fri, 2012-07-13 at 08:41 +0300, Zhao Halley wrote:
add RGB to NV12 and NV12 to RGB conversion
- either is implemented in 2 ways: 8.8 fix point or float point
however, it doesn't show performance difference in my test
- RGB to NV12 conversion can be used to load video frames
+/** \brief Connected. Authenticated with DRI1 protocol. */
+VA_DRM_AUTH_DRI1= 1,
Is there anyone still use DRI1? If not, can we deprecate it ?
From: Benjamin Franzke benjaminfran...@googlemail.com
Add va/va_drmcommon.h header for DRM-based drivers to use common
base utilities
On Wed, 2012-07-25 at 22:53 -0700, Jorge Lucangeli Obes wrote:
On Wed, Jul 25, 2012 at 10:28 PM, Xiang, Haihao haihao.xi...@intel.com
wrote:
On Wed, 2012-07-25 at 21:15 -0700, Jorge Lucangeli Obes wrote:
On Wed, Jul 25, 2012 at 8:50 PM, Xiang, Haihao haihao.xi...@intel.com
wrote
It seems neither DRI2 nor DRI is supported on your system. Libva needs
them to get the right driver name for Intel platform on X11 window
system.
Thanks
Haihao
Hi,
I am getting following error while I give vainfo command on my
terminal:
Xlib: extension XFree86-DRI missing on display
This version brings the following changes:
* API: data structures for JPEG Baseline decoding
* API: clarify the definition of slice_data_bit_offset
* Add a simple test case for JPEG decoding
git tag: libva-1.0.16
http://cgit.freedesktop.org/vaapi/libva/snapshot/libva-1.0.16.tar.gz
This release requires libva-1.0.16 and brings the following changes:
* Add JPEG decoding on Ivy Bridge
* Add support for a new Ivy Bridge chip
* Add support for vaSyncSurface() and vaQuerySurfaceStatus() (Dmitry Ermilov)
* Fix decoding of MPEG-2 videos with implicit IQ matrices
* Fix concurrent
Applied.
Thanks
Haihao
Otherwise this leads to use-after-free of the bufmgr.
Signed-off-by: Stéphane Marchesin marc...@chromium.org
---
src/i965_drv_video.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
index
Hi,
I am trying to use libva with Intel driver back-end for decode and
encode. Decoding works like charm, but when I try to encode, I get
some problems.
Which branch are you using, staging or master ? Encoder on staging has
more features.
I got the same errors using the sample code
Applied
Thanks
Haihao
Thanks to Haihao and Gwenole's feedback.
V2 update:
- both BGRX and RGBX to/from NV12 is supported, a flag is added
in statci parameter on C programing side
- split RGBX/RGBA/BGRX/BGRA format support to a separate patch
- define more alias in
-Original Message-
From: Ratin [mailto:rat...@gmail.com]
Sent: Saturday, August 11, 2012 6:22 AM
To: Xiang, Haihao
Cc: Libva@lists.freedesktop.org
Subject: Re: [Libva] VAAPI encoding with Intel Sandybridge GPU
Hi Haihao, Thanks a lot for your quick reply!
On Thu, Aug 9, 2012
The staging branch supports raw-DRM. With it, an user can use libva
without an X server.
Thanks
Haihao
I was wondering if it was possible to use libva for gpu video
compression/acceleration on intel HD3000 on a system with no monitor?
I have an fc-16 (linux 3.1) sytem in which I built
On Wed, 2012-08-15 at 11:31 -0700, Ratin wrote:
Is it possible to dump the decoded bitstream in BGRA from the VA
Surfaces when using the Intel Sandybridge GPU?
We will add support for vaGetImage()/vaPutImage() with RGBA/BGRA
Thanks
Haihao
Ratin
On Thu, 2012-08-16 at 10:25 +0200, Gwenole Beauchesne wrote:
Hi,
2012/8/16 Xiang, Haihao haihao.xi...@intel.com:
I see a comment in /va/drm/va_drm_auth.c
/* Default: root + master privs are needed for the following call */
Master is enough.
Master requires root privileges
-Original Message-
From: rlan...@gdrs.com [mailto:rlan...@gdrs.com]
Sent: Thursday, August 16, 2012 8:45 PM
To: gb.de...@gmail.com; Xiang, Haihao
Cc: libva@lists.freedesktop.org
Subject: RE: [Libva] using libva video compression on system with no monitor
I agree with Gwenole.
I
On Thu, 2012-08-16 at 14:57 -0700, Ratin wrote:
On Thu, Aug 16, 2012 at 1:29 AM, Gwenole Beauchesne gb.de...@gmail.com
wrote:
Hi,
2012/8/15 Ratin rat...@gmail.com:
Is it possible to dump the decoded bitstream in BGRA from the VA
Surfaces when using the Intel Sandybridge GPU?
If
Applied.
Thanks
Haihao
Client should not be required to send seq for each picture.
Change-Id: I4f24ac7dd7ad2d0657a641fb82793c9bae665694
---
src/i965_drv_video.c |2 --
1 files changed, 0 insertions(+), 2 deletions(-)
diff --git a/src/i965_drv_video.c b/src/i965_drv_video.c
index
On Fri, 2012-08-17 at 06:10 +0200, Gwenole Beauchesne wrote:
2012/8/16 Xiang, Haihao haihao.xi...@intel.com:
On Thu, 2012-08-16 at 10:25 +0200, Gwenole Beauchesne wrote:
Hi,
2012/8/16 Xiang, Haihao haihao.xi...@intel.com:
I see a comment in /va/drm/va_drm_auth.c
/* Default
Hi,
A new commit to vaapi/intel-driver (staging branch) has files with a
new more restrictive licence. This looks like a mistake to me - can
someone confirm whether this was intentional or not please?
(I'm referring to
On Thu, 2012-08-23 at 14:37 +0800, Zhao Halley wrote:
---
src/i965_post_processing.c | 42 +-
1 file changed, 33 insertions(+), 9 deletions(-)
diff --git a/src/i965_post_processing.c b/src/i965_post_processing.c
index 5ea5dc3..ab42901 100755
---
Pushed.
Thanks
Haihao
---
src/i965_post_processing.c | 42 +-
1 file changed, 33 insertions(+), 9 deletions(-)
diff --git a/src/i965_post_processing.c b/src/i965_post_processing.c
index 5ea5dc3..f5cac69 100755
--- a/src/i965_post_processing.c
Pushed
Thanks
Haihao
From: Li,Xiaowei xiaowei.a...@intel.com
RGB format surface is required for color
space conversion from YUV to RGB.
signed-off-by: Li,Xiaowei xiaowei.a...@intel.com
---
va/va.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/va/va.h b/va/va.h
index
Could you add a command option to specify the number of slices ?
Thanks
Haihao
On Wed, Aug 22, 2012 at 10:39 AM, Sean V Kelley sean.v.kel...@intel.com
wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
What are you using to test/validate?
Any additional description?
This
Pushed.
Thanks
Haihao
background:
not each pixel requires update in post processing, it happends:
- on left boundary blocks since dst horizontal offset must be dword
alignment
- on right boundary blocks when offset_x+width is not 16 aligned
- on botton boundary blocks when
The root cause for this issue is to link wrong libraries when build
libva-egl. the following patch can fix this issue:
---
va/Makefile.am |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/va/Makefile.am b/va/Makefile.am
index a2a26e6..968474f 100644
--- a/va/Makefile.am
+++
Applied
Thanks
Haihao
From: Zhao Yakui yakui.z...@intel.com
Otherwise the encoded h264 video stream can't be parsed by HRD test software.
Signed-off-by: Zhao Yakui yakui.z...@intel.com
---
test/encode/avcenc.c | 155
++-
1 file
I can apply your patch to the Intel driver, but it would be better to
fix the codec parser in case other drivers don't handle this case too.
Thanks
Haihao
Hi:
When gst-vaapi is used for vc1 playback, many stream give corruption
image,
after some debug, I found it was caused by
Hi, Gwenole
Your recent commits in libva staging branch result in DSO regression
issue.
make[3]: Entering directory `/home/cc/libva/test/encode'
CCLD h264encode
/usr/bin/ld: h264encode_x11.o: undefined reference to symbol 'XOpenDisplay'
/usr/bin/ld: note: 'XOpenDisplay' is defined in DSO
On Fri, 2012-10-05 at 17:10 +0200, Gwenole Beauchesne wrote:
Hi,
A new version of libva is now available at the following location:
http://www.freedesktop.org/software/vaapi/releases/libva/
VA-API was bumped to 0.33.0, while libva still supports VA drivers
compiled for 0.32.x API.
On Fri, 2012-10-19 at 13:02 -0400, yakui.z...@intel.com wrote:
From: Zhao Yakui yakui.z...@intel.com
It uses the variable of locked_image_id to check whether one surface is locked
or not. But as the locked_image_id is not assigned correctly, it causes that
it can't lock one surface next
On Tue, 2012-10-23 at 09:52 -0700, Charles, Daniel wrote:
On Mon, Oct 22, 2012 at 6:59 PM, ykzhao yakui.z...@intel.com wrote:
On Mon, 2012-10-22 at 15:12 -0600, Ratin wrote:
On Mon, Dec 17, 2012 at 5:04 PM, ykzhao yakui.z...@intel.com wrote:
On Wed, 2012-10-17 at 12:24 -0600, Ratin wrote:
, the output file is a raw video
stream which doesn't include fps information. So you should specify -fps
option in the command line, such as
mplayer -fps 30 /workspace/ratin/videos/SOCCER_704x576_60_orig_02.mp4
On Tue, Oct 23, 2012 at 10:45 PM, Xiang, Haihao haihao.xi...@intel.com
wrote
On Wed, 2012-10-31 at 13:29 +0800, ykzhao wrote:
On Tue, 2012-10-30 at 19:40 -0600, Xiang, Haihao wrote:
Some requirements for internal buffers on *SNB* are different from
*IVB*. please take carefully and double check the docs.
From: Zhao Yakui yakui.z...@intel.com
Signed-off
Hi,
We are planning for the next release 1.0.19 for Intel driver. Please
report any critical issue in the master branch, thanks.
The release will include the following new features and fixes:
Version 1.0.19 - xx.Nov.2012
* Add support Haswell
* Add raw DRM support (Dmitry Ermilov)
* Add
Hi,
Intel-driver 1.0.19 is available at
http://www.freedesktop.org/software/vaapi/releases/libva-intel-driver/
Intel-driver 1.0.19 is backward compatible with libva 1.0.16, however
libva 1.1.0 is recommended for new features, such as raw DRM, Wayland
etc
Version 1.0.19 - 09.Nov.2012
* Add
On Mon, 2012-11-19 at 21:37 +0200, Yannis Tsopokis wrote:
What about us with intel 4500hd? The g45-h264 branch does not even compile!
Here is bug for g45-h264 branch, and it should be fixed.
https://bugs.freedesktop.org/show_bug.cgi?id=54756
Could you give a try ?
Thanks
Haihao
On
/intel/libva/test/basic'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/home/yannis/intel/libva/test'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/yannis/intel/libva'
make: *** [all] Error 2
On 21/11/2012 03:28 πμ, Xiang, Haihao wrote:
On Mon
1 - 100 of 839 matches
Mail list logo