[00:10] <RAOF> Hey, cool. UVD support for radeon.
[01:29] <bjsnider> i don't think any player supports that
[02:21] <RAOF> bjsnider: UVD isn't the client-facing API, it's the hardware decode engine.
[02:21] <RAOF> ie: It's now possible for mesa's vdpau support to work on radeon, and use the actual hardware decode engines.
[02:30] <Dandel> some major patches just landed for radeon hardware w/ mesa. ( I wonder how long it'll be before test packages are made )
[02:34] <RAOF> xorg-edgers gets updated roughly nightly, so...
[02:36] <Dandel> RAOF: Yea, well this is a set of patches that modifies the 3.8 kernel
[02:37] <RAOF> So, you'll need the kernel nightlies from the pseudo-ppa then :)
[02:37] <Dandel> it involves 10 patches to the kernel and adds UVD support. 
[02:37] <Dandel> Oh, how would i go about setting that up.
[02:39] <RAOF> You'd be looking for http://kernel.ubuntu.com/~kernel-ppa/mainline/
[02:41] <Sarvatt> and mesa hasnt been updated in edgers in over a month, lack of caring on my part :) plus it doesnt package mesa vdpau
[02:42] <Sarvatt> one more thing to enable that breaks the build every week
[02:42] <RAOF> Heh
[02:42] <Sarvatt> mesa doesn't have enough of those already
[02:42] <RAOF> Got sick of clover, then? ☺
[02:43] <Sarvatt> radeon llvm shader compiler this month
[02:43] <Dandel> I sort of noticed that.
[02:43] <Sarvatt> needs a new llvm checkout every few weeks :)
[02:43] <bjsnider> is mesa vdpau actually getting commits and whatnot?
[02:43] <Dandel> bjsnider: yes, and a MAJOR one just landed.
[02:44] <Sarvatt> not really but the build system has enough churn its annoying for fringe compile time options
[02:44] <RAOF> The build system should have seen the majority of its near-term churn by now though, right?
[02:45] <Sarvatt> i havent built it in a month, maybe? :)
[02:46] <Dandel> I dunno about the compile issues, but I think getting a test repo up would be good since the patches are aimed at all cards from amd that have some form of UVD decoder included.
[02:46]  * Sarvatt will look tomorrow and see about enabling vdpau in the ubuntu+1 branch if nothing critical work related comes along
[02:47] <bjsnider> haha
[02:47] <bjsnider> always conditiions
[02:48] <Dandel> bjsnider: well that's not too bad. Many radeon HD2000 line cards have UVD, and it's guaranteed on radeon hd4000+
[02:48] <Sarvatt> s/many/all/
[02:48] <Sarvatt> or is it limited to newer UVD?
[02:48] <Dandel> No, UVD/UVD2/UVD3
[02:49] <Dandel> RV610/RS780C - good example of unsupported
[02:49] <bjsnider> with the libav-mt, i'm not sure any of this is necessary anymore
[02:49] <Dandel> bjsnider: you'd be surprised
[02:50] <Dandel> ever try running 1080p on the amd e series?
[02:50] <bjsnider> you must be joking
[02:50] <Dandel> No.
[02:50] <Dandel> i'm dead serious.
[02:50] <Dandel> that's what is used by many htpc users
[02:50] <bjsnider> only intel
[02:51] <bjsnider> hm, strange choice
[02:51] <Sarvatt> good to finally have a real use for mesa vdpau at least, there have been bugs about enabling it in debian/ubuntu for years :)
[02:52] <Dandel> AMD C-50 ( found in netbooks ) is a good example... it's a dual core at 1ghz. ( the C-30 is a single core at 1.2ghz )
[02:53] <Sarvatt> yeah fat lot of good libav-mt will do there
[02:53] <bjsnider> why not use intel/nvidia hardware?
[02:53] <bjsnider> too expensive?
[02:53] <Dandel> Too expensive.
[02:53] <bjsnider> is there an echo in here?
[02:53] <Dandel> not an echo. lol... it is that it is too pricy
[02:53] <Sarvatt> C-50 is more people buying cheap $199 laptops thinking they are actually useful..
[02:54] <bjsnider> well, i say pay the extra ducats
[02:54] <Dandel> well for real world, I know that the AMD E-350 is what a lot of htpc's have and that uses a dual core at 1.6ghz
[02:55] <bjsnider> but nvidia hardware would have worked since 2008
[02:55] <bjsnider> with vdpau i mean
[02:56] <Sarvatt> err, "The R6xx and RS780/RS880 chipset generations are currently not supported, but might be added in the future."?
[02:56] <Sarvatt> whee
[02:57] <Dandel> Radeon 3000/3100
[02:57] <Dandel> igp
[02:57] <Sarvatt> 2-3k and igp excluded yeah, i see what you mean now
[02:58] <Dandel> Anyways I know of two applications that are going to benefit from this a lot... it'll be flash and xbmc
[02:59] <Sarvatt> so we need glamor packaged, ddx building with glamor support, new firmware and mesa 9.2, fun stuff for 13.10
[02:59] <bjsnider> surely not flash
[02:59] <Sarvatt> flash will use it yeah
[03:00] <Sarvatt> i was kind of worried about packaging mesa vdpau because things might try to actually use it and fail, like nouveau needs extra firmware too
[03:00] <Sarvatt> not sure if it gracefully falls back or flat out fails
[03:00] <bjsnider> adobe has stabilized and enabled gpu acceleration on linux?
[03:01] <Sarvatt> vdpau only for years actually, yeah
[03:01] <Sarvatt> dont remember the smurf bug? :)
[03:01] <bjsnider> i think you are wrong
[03:01] <bjsnider> it's there, but not stable and not enabled by default
[03:02] <bjsnider> and if you enable it, you will get an x lockup
[03:02] <Sarvatt> https://bugs.launchpad.net/ubuntu/+source/flashplugin-nonfree/+bug/968647
[03:02] <Dandel> vdpau is enabled by default in flash ( at least with chrome )
[03:02] <Sarvatt> well https://bugs.launchpad.net/ubuntu/+source/adobe-flashplugin/+bug/967091
[03:03] <bjsnider> Dandel, is that with the internal flash?
[03:03] <Dandel> it actually is a flash plugin. ( just happens to be that firefox fails to implement the api to use it )
[03:03] <Sarvatt> vdpau was broken out of the box on nvidia until libvdpau got updated to work around it
[03:04] <Sarvatt> err flash was broken, used vdpau by default
[03:04] <bjsnider> i didn't have the red/blue reverse bug, i had the one that took down x
[03:04] <bjsnider> at the same time as adobe announced they would no longer be developing linux flash
[03:05] <bjsnider> Dandel, have you got an nvidia card to test that on?
[03:05] <Dandel> google took that up in chrome, so you get version 11.7.700.141
[03:05] <Dandel> bjsnider: Yea, but i doubt it'll be any good... 7300gs
[03:06] <Sarvatt> bjsnider: pretty sure you said you were on 11.04 or 11.10 at the time? it was during the precise dev cycle
[03:06] <bjsnider> well, google may have fixed it by now, who knows
[03:06] <Dandel> bjsnider: adobe still maintains flash on mozilla for linux ( security updates only )
[03:08] <Sarvatt> crappy but i just found out flash 11.2 in the distro doesnt work with amazon prime videos tonight, no more prime videos on linux (they block chrome flash because no drm)
[03:08] <bjsnider> i know
[03:09] <Sarvatt> anyway have a good night everyone :)
[03:09] <Dandel> actually, depending on the ffmpeg setup, lightspark can get accelerated decoding.
[03:13] <Dandel> ah and gnash too 0o' so it appears a lot of places are ready.
[03:27] <bjsnider> ffm-what?
[03:29] <Dandel> bjsnider: i say ffmpeg but it's libav or whatever else is in use.
[05:12] <mlankhorst> morning
[05:18] <Dandel> good morning. 
[10:59] <mlankhorst> wow.. people still use xcompmgr?
[11:25] <ogra_> with openbox ;)
[11:39] <seb128> ogra_, is openbox hype or something?
[11:39] <seb128> ogra_, I learnt yesterday that some french people were working on an openbox ubuntu remix :p
[11:40] <seb128> ogra_, http://beta.linuxvillage.net/index.php/topic,248.0.html
[11:40] <ogra_> haha
[11:40] <seb128> OBUbuntu remox
[11:40] <seb128> remix
[11:41] <ogra_> its not hype, but people that find lubuntu to bloated (?!?) seem to like such a setup
[11:41] <ogra_> i wonder what happened to fluxbuntu 
[11:41] <seb128> right, I chatted a bit with one of the people working on it and asked about lubuntu
[11:41] <ogra_> which would be largely the same 
[11:41] <seb128> it seems lubuntu needs 512M ram to run nowadays
[11:42] <seb128> they claim they can do better
[11:42] <seb128> I'm waiting for the remix which claims that 640k is enough :p
[11:42] <ogra_> hnn, my lubuntu test installs on the ac100 never use more than ~120M
[11:42] <ogra_> HAHA !
[11:50] <ogra_> wow ... they want to run in less than 512M but ship all the LibreOffice bits by default ?
[11:51] <ogra_> and a ton of gnome bits too that will fire up g-s-d and friends in the background once you start them
[11:51] <ogra_> phew, but at least it will have xsnow ...
[11:51]  * ogra_ is relived
[11:57] <seb128> lol
[11:57] <seb128> ogra_, well, they said that having nothing to run is what let them run libreoffice on the liveCD on a 512M box :p
[11:58] <seb128> which seems like a selling point for them compared to xubuntu/lubuntu ;-)
[11:58] <ogra_> well, to me it looks like lubuntu with lxpanel replaced with tint2 ... and shipping LibO instead of goffice
[11:58] <seb128> yeah, I would just contribute to lubuntu
[11:59] <ogra_> ++
[11:59] <seb128> but it's their pet project, if they like it, good for them ;-)
[11:59] <ogra_> but up to them 
[14:19] <dholbach> hiya
[14:19] <dholbach> so... I just had apport come up
[14:19] <dholbach> and it directed me to https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1099394
[14:19] <dholbach> which has been closed since 2 months
[14:19] <mlankhorst> oh fake hang?
[14:19] <dholbach> I'm not quite sure why I answer all of apport's questions when it directs me to an already closed bug :)
[14:20] <dholbach> so either it's a different bug, or it's not fixed yet
[14:21] <seb128> I get regular gpu hangs on intel/raring as well
[14:21] <tjaalton> apport can dupe the bugs if the two numbers match, but there could be something else going on making it a new bug
[14:21] <seb128> got one yesterday after docking back my laptop and doing a dnd of xchat between both screens
[14:21] <dholbach> https://errors.ubuntu.com/ counts 286 times  "xserver-xorg-video-intel [sandybridge-m-gt2] GPU lockup" bugs
[14:22] <tjaalton> I get a compiz hang when returning to my session, fixed by mesa 9.1 but it's getting late to push it to raring
[14:22] <dholbach> it directs to https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1159536 which is a duplicate and also fix released
[14:22] <seb128> tjaalton, can we backport the patch?
[14:23] <tjaalton> seb128: which patch?
[14:23] <tjaalton> the one on the bug above is in -intel, already in raring
[14:23] <tjaalton> dholbach: Sarvatt duped it :)
[14:23] <dholbach> I get the feeling that these lockups happen quite often, but that the data can't be forwarded to LP correctly
[14:24] <dholbach> tjaalton, if it's a dup that's fine with me - it's apparently just not fixed :)
[14:24] <seb128> tjaalton, the one for your hand when returning to your session fixed by mesa 9.1
[14:24] <seb128> hand->hang
[14:24] <tjaalton> ah, not sure yet
[14:25] <tjaalton> doesn't happen every time, but maybe 1/3
[14:25] <tjaalton> killing compiz fixes it
[14:25] <tjaalton> with all the issues it brings
[14:25] <tjaalton> like global menu not working anymore :)
[14:31] <tjaalton> the reason why 9.1 is not in raring yet is that blur was broken/slow on newer intels
[14:31] <tjaalton> it's fixed now..
[14:36] <seb128> I'm familiar with that compiz lock
[14:36] <seb128> I get it several time a week when coming back from guest session :p
[14:36] <tjaalton> yeah
[14:36] <tjaalton> the same
[14:36] <seb128> tjaalton, one way to avoid the menus, etc not working is to sig11 compiz and go back to your session
[14:36] <seb128> if gnome-session respawns it for you, you win ;-)
[14:37] <seb128> the env is the same and menus, etc keep working
[14:38] <tjaalton> ah, cool
[14:43] <bjsnider> good thing i don't run compiz
[16:14] <sforshee> I'm trying to figure out what interface into the kernel is used when userspace wants to power off a display. Can anyone here tell me?
[17:38] <agrestringere> does anyone know the status of the recent CVE and Nvidia drivers 304.88 and 310.44?
[17:38] <agrestringere> http://nvidia.custhelp.com/app/answers/detail/a_id/3290
[17:42] <tjaalton> being worked on
[17:43] <agrestringere> okay, so I don't need to file a bug?
[17:46] <agrestringere> Thank you for the information, have a good one...