Bug 12170

Done.

Status: NEW => RESOLVED
Resolution: (none) => FIXED

The package seems to have never built. I updated it today to 0.3.4 version, and tested it against future MGA5 cauldron : it works, without having to export any VDPAU variable.

CC: (none) => lists.jjorge

As this was an ask for MGA4, I convert this bug in a backport request.

I have tested this backport in MGA4 x86_64 with Intel vaapi, this way :

vainfo (to ensure you have VAAPI hardware)
export VDPAU_DRIVER=va_gl
mplayer -vo vdpau -vc ffh264vdpau,ffmpeg12vdpau,ffvc1vdpau,... file.mkv

It draws 15 watts instead of 18 when playing 1080p on my hardware, that's 20% more battery time.


Suggested advisory:
========================
Backported libvdpau-va-gl allows VAAPI use with mplayer and flash-player-plugin.
See hints on http://davidrosca.blogspot.fr/2013/05/va-api-backend-for-vdpau.html
========================

Updated packages in {core,tainted}/updates_testing:
========================
libvdpau-va-gl-0.3.4-1.mga4.x86_64
libvdpau-va-gl-0.3.4-1.mga4.i586

Source RPMs: 
libvdpau-va-gl-0.3.4-1.mga4.srpm

Status: RESOLVED => REOPENED
Component: New RPM package request => Backports
Version: Cauldron => 4
Resolution: FIXED => (none)
Assignee: lists.jjorge => qa-bugs
Target Milestone: --- => Mageia 4

updates or backport testing ?

(In reply to Manuel Hiebel from comment #4)
> updates or backport testing ?

Sorry, copy/paste error : 

Updated packages in core/backports_testing  of course.

I could test on another MGA4_64 machine. All is Ok, including flash-player accelerated. Procedure below for flash-player-plugin test :

Firefox with YouTube shows software video in stats.

AS ROOT :
mkdir /etc/adobe
echo EnableLinuxHWVideoDecode=1 > /etc/adobe/mms.cfg

AS USER :
echo export VDPAU_DRIVER=va_gl >> ~/.bash_profile
logout/login

Firefox with YouTube shows accelerated video in stats.

Whiteboard: (none) => has_procedure mga4-64-ok

Sorry José, we are unable to accept backports as there is a bug in rpmdrake (bug 12766) which makes the release packages invisible if a backport exists. 

I'm adding sysadmin as the packages should probably be removed from backports_testing until then.

This could possibly be considered as an update rather than backport as it's essentially a bugfix but please discuss on dev ML before doing so in case it can affect other systems.

I'm assigning back to you for the moment. Please reassign when you're ready. Thanks

CC: (none) => eeeemail, sysadmin-bugs
Depends on: (none) => 12766
Assignee: qa-bugs => lists.jjorge

Well, as no release package exists, I thought this backport could be pushed : it will not hide anything!

Why not ask for an exception to push this as an update instead José. Email dev ML about it.

No, please dont start abusing updates, or we never get the the backports thing fixed...

As this wont trigger the backports bug, it can be handled as a real backport

There is a patch proposed by anaselli to fix bug 12766 awaiting review btw

Taking this one back then.

Assignee: lists.jjorge => qa-bugs

Hervé we have a policy of asking the person requesting the backport to assist with testing it, could you ensure this package does what it's expected to please.

You'll find it in the Core Backports Testing media.

It works as expected, thanks.

Status: REOPENED => RESOLVED
Resolution: (none) => FIXED

it'll need to be pushed then.

Did you test x86_64 or i586 please?

Status: RESOLVED => REOPENED
Resolution: FIXED => (none)

I tested the x86_64 package, I have no MGA4 i586 installed.

Testing Mageia 4 i586.

I followed the testcase described in comment 3, and it works as expected, VDPAU backend is now provided:

> [VS] Software VDPAU backend library initialized
> libva info: VA-API version 0.34.0
> libva info: va_getDriverName() returns 0
> libva info: Trying to open /usr/lib/dri/i965_drv_video.so
> libva info: Found init function __vaDriverInit_0_34
> libva info: va_openDriver() returns 0
> ==========================================================================
> Forced video codec: ffh264vdpau
> Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
> libavcodec version 55.18.102 (external)
> Selected video codec: [ffh264vdpau] vfm: ffmpeg (FFmpeg H.264 (VDPAU))
> ==========================================================================
> ==========================================================================
> Trying to force audio codec driver family libmad...
> Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
> AUDIO: 48000 Hz, 2 ch, floatle, 0.0 kbit/0.00% (ratio: 0->384000)
> Selected audio codec: [ffaac] afm: ffmpeg (FFmpeg AAC (MPEG-2/MPEG-4 Audio))
> ==========================================================================
> AO: [pulse] 48000Hz 2ch floatle (4 bytes per sample)
> Starting playback...
> [VD_FFMPEG] Trying pixfmt=0.
> [VD_FFMPEG] XVMC-accelerated MPEG-2.
> Movie-Aspect is 2.00:1 - prescaling to correct movie aspect.
> VO: [vdpau] 1280x640 => 1280x640 H.264 VDPAU acceleration  [zoom]


Following hints on http://davidrosca.blogspot.fr/2013/05/va-api-backend-for-vdpau.html I also configured the VDPAU hw acceleration for flash-player-plugin.

To do so, I followed the procedure described in comment 6 (actually not, because I had overlooked this comment and figured out the procedure myself, but I did the same things ;). Afterwards Youtube shows that I have both "hardware video rendering, hardware video decoding" when right-clicking on the video and choosing "Advanced statistics".

CC: (none) => remi

And I confirm that there is a clear improvement (at least for Flash videos in fullscreen, it was unusable without this backport, now it's fine).

@José: Wouldn't it be better if the libvdpau-va-gl package would set the required configuration for flash-player-plugin and "export VDPAU_DRIVER=va_gl" by itself? As it is now the backport candidate is only usable by advanced users, but I guess that end-users installing this package would expect it to work out of the box?

In MGA5, it works  without the hack, so the problem is solved...

Ah, it might be that my test was a bit too short on Cauldron (I installed the package, and since it wasn't working out of the box I did the same hack as described above). Probably I was just missing a reboot.

Or maybe my hardware (Intel HD Graphics 4000) doesn't qualify as i965? (IIUC the following line is what should make it work in mga5)

> #This automatically enable vdpau backend on intel i965
> ln -s libvdpau_va_gl.so.1 $RPM_BUILD_ROOT%{_libdir}/vdpau/libvdpau_i965.so.1

Would it also work out of the box for Flash player?

Yes but you have to configure vdpau to get it working with flash player.
mkdir /etc/adobe
echo -e "EnableLinuxHWVideoDecode = 1\nOverrideGPUValidation = 1" | tee /etc/adobe/mms.cfg

http://www.reddit.com/r/Ubuntu/comments/1wpavp/how_to_enable_vdpau_acceleration_for_accelerated/

CC: (none) => ozkyster

This isn't forgotten btw, it is on hold awaiting a new bugzilla keyword and some info from sysadmins.

It's removed from backports testing so it needs to be rebuild from src rpm too.

(In reply to Otto Leipälä from comment #21)
> Yes but you have to configure vdpau to get it working with flash player.

Well, reverse it : flash player does not default using vdpau if it is avalaible, so one has to force it in flash player, this is not a vdpau configuration.

(In reply to claire robinson from comment #22)
> This isn't forgotten btw, it is on hold awaiting a new bugzilla keyword and
> some info from sysadmins.

So is this still awaiting a bugzilla keyword, or sysadmins info?

Yes, we're unable to push backports currently. There are 3 waiting.

Adding feedback marker until we can progress

Whiteboard: has_procedure mga4-32-ok mga4-64-ok => has_procedure feedback mga4-32-ok mga4-64-ok

Mageia 4 EOL is coming:
http://blog.mageia.org/en/2015/09/10/mageia-4-about-to-reach-its-end-of-life/

Status: REOPENED => RESOLVED
Resolution: (none) => OLD