=== yofel_ is now known as yofel
[10:57] <tjaalton> RAOF: have you tried llvmpipe yet? unity3d doesn't seem too useful with it, there's actually a bug report already, bug 926859
[10:57] <ubot4> Launchpad bug 926859 in xorg (Ubuntu) "Desktop almost useless - due to corruption and flikering (affects: 1) (heat: 6)" [Undecided,Incomplete] https://launchpad.net/bugs/926859
[10:57] <tjaalton> but I'm not sure if it's the driver or unity failing
[10:57] <RAOF> tjaalton: A bit of both - compiz does weird, weird things.
[10:57] <tjaalton> in any case, it might be wise to blacklist it?
[10:58] <RAOF> We already do, right?
[10:58] <RAOF> Or not?
[10:58] <RAOF> If it's not blacklisted, it should be.
[10:58] <tjaalton> I didn't need to do anything to use it
[10:58] <tjaalton> yeah
[10:58] <RAOF> Oh, the renderer string changed, so the existing software rasteriser blacklist doesn't kick in.
[10:58] <RAOF> Yeah.  One bug to file against unity!
[10:59] <tjaalton> not that it has those already :)
[10:59] <tjaalton> hmm that wasn't right
[11:00] <RAOF> One *more* bug to file against unity :P
[11:01] <tjaalton> :)
[11:01] <tjaalton> unity2d was quite snappy on the sis-crap though
[11:01] <tjaalton> using vesa, of course
[11:01] <RAOF> Heh.
[11:01] <RAOF> Yeah, Qt's renderer is pretty hot stuff.
=== shadeslayer is now known as shadeslayer_
=== shadeslayer_ is now known as shadeslayer
[20:22] <stgraber> any reason why I'm getting compiz running in my Precise VMs now? :) (obviously it's all broken and full of artifacts)
[20:25] <stgraber> (it's rather annoying when you try to debug the live environment and you don't see what you're doing or just get apport windows popping up every 10s ;))
[20:29] <Sarvatt> stgraber: software rendering uses llvmpipe instead of the old swrast, theres probably a hardcoded check for the old GL renderer string somewhere, hmm
[20:30] <Sarvatt> looks like unity-support-test in nux
[20:31] <stgraber> Sarvatt: I "think" we may want to change that check ;) It's certainly interesting to have an option of running compiz on the llvmpipe stuff but it doesn't seem quite ready at this point :)
[20:32] <Prf_Jakob> stgraber: which host software?
[20:32] <stgraber> (I wouldn't mind compiz running in my VM if it wasn't using 50% of CPU to render garbage)
[20:32] <Sarvatt> yep, was hopeful it'd work but sounds like its a nogo
[20:32] <stgraber> Prf_Jakob: kvm
[20:32] <Sarvatt> Prf_Jakob: i'm sure he's using kvm or virtualbox
[20:32] <Prf_Jakob> ok, nm then.
[20:32] <Sarvatt> stgraber: btw vmware 3D passthrough works great out of the box with unity-3d now :P
[20:33] <stgraber> Sarvatt: apparently we'll be getting a new compiz soon, but I don't know if it's going to work any better under llvmpipe
[20:33] <Prf_Jakob> Sarvatt: In case he wasn't I couldn't let the oppertunity to brag about that ;-)
[20:33] <Prf_Jakob> go*
[20:33] <Sarvatt> yep definitely needs to be fixed in unity-support-test in nux
[20:33] <Sarvatt>   // Check for software rendering.
[20:33] <Sarvatt>   if (results->renderer != NULL &&
[20:33] <Sarvatt>       (strncmp (results->renderer, "Software Rasterizer", 19) == 0 ||
[20:33] <Sarvatt>        strncmp (results->renderer, "Mesa X11", 8) == 0 ||
[20:33] <Sarvatt>        strstr (results->renderer, "on softpipe") != NULL)) {
[20:33] <Sarvatt>     results->flags |= FLAG_SOFTWARE_RENDERING;
[20:40] <cnd> urgh, X synaptics has a new motion regression algo, and it's completely broken
[20:40] <cnd> the question is whether we should fix it, or just revert the algo
[20:41] <Sarvatt> stgraber: https://bugs.launchpad.net/ubuntu/+source/nux/+bug/926859
[20:41] <ubot4> Launchpad bug 926859 in nux (Ubuntu) "llvmpipe software rendering needs blacklisting in unity-support-test (affects: 1) (heat: 6)" [High,Triaged]
[20:42] <stgraber> Sarvatt: thanks
[22:53] <bryceh> hey tjaalton, for bug #745112 is it correct that it got fixed in 2.6.39 (and 2.6.38 via natty-updates) but then has regressed starting with 3.0?
[22:53] <ubot4> Launchpad bug 745112 in xserver-xorg-video-intel (Ubuntu Precise) (and 6 other projects) "[arrandale] desktop is messed up with external monitors (x86_64) (affects: 133) (dups: 13) (heat: 581)" [Medium,Invalid] https://launchpad.net/bugs/745112
[22:53] <bryceh> tjaalton, Sarvatt , RAOF - any of you have arrandale laptops with docking stations that can reproduce ^^ this bug?
[22:53] <RAOF> bryceh: Sorry, no arrandale here.
[22:53] <Sarvatt> no docking stations and no arrandale
[22:54]  * RAOF runs a SandyBridge and GM45 shop.
[22:56] <broder> what generation of cpu would go with arrandale?
[22:57] <RAOF> That's the original Core iX chips, IIRC.
[22:58] <bryceh> http://en.wikipedia.org/wiki/Arrandale
[22:58] <bryceh> nehalem/westmere
[22:59] <Sarvatt> 3 digit numbers after the i3/i5/i7
[23:00] <broder> ok. i have a 1st-gen i7 machine here. i can probably test that for you guys, but will take a bit because somebody's using it at the moment
[23:03] <bryceh> broder, does it have a displayport and docking station?
[23:03] <bryceh> broder, the testing is probably going to be of the kernel bisection variety
[23:04] <broder> it has a DP and a docking station, which has DP, DVI, and VGA
[23:06] <bryceh> broder, ok cool; when you get a chance see if you can reproduce the bug, which appears to occur just when you dock the laptop (possibly when the docking station has 2 external monitors configured).  It should black-screen fairly reliably within 1-2 tries
[23:06] <broder> precise or oneiric?
[23:07] <bryceh> broder, apparently the bug affects both
[23:07] <bryceh> broder, however I'm most interested in precise
[23:07] <broder> ok. i'll go ahead and grab a new daily then
[23:08] <bryceh> if it can be repro'd on precise, next thing to test after that would be http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-fixes/