| AnAnt | Hello, is Intel X45000HD graphics chip supported by Ubuntu (wether jaunty or karmic) ? | 06:16 |
|---|---|---|
| diverse_izzue | hey guys. i'm testing the radeon KMS code. other than resuming from standby being broken, the biggest issue for me is performance. DRI performance is 2-4-fold reduced for me, depending on the situation, compiz feels sluggish at times. what does upstream say about that - do they think they will be able to speed that up somewhat in time for release? | 11:14 |
| diverse_izzue | also, scrolling in GTK apps for example is much slower than with the non-KMS code. | 11:14 |
| jcristau | they'll probably have things working in f12 | 11:15 |
| diverse_izzue | that's sort of the same timeframe as karmic, right? | 11:16 |
| tjaalton | a bit later | 11:17 |
| tjaalton | and they'll likely use a newer kernel | 11:17 |
| tjaalton | tseliot: I haven't, no hw for that | 11:18 |
| tseliot | tjaalton: ok, no problem. I was wondering if we could use nouveau to replace nvidia-glx-71 (which should be removed from karmic) | 11:20 |
| tjaalton | tseliot: ok, why should it be removed? | 11:21 |
| tseliot | tjaalton: because Nvidia didn't (and won't) add the support for xserver 1.6.x to that driver | 11:22 |
| tjaalton | ah, so it still hasn't.. ok | 11:22 |
| tjaalton | supporting nouveau like that would imo mean shipping the drm driver with the kernel | 11:23 |
| tseliot | ah so we don't do that | 11:23 |
| tjaalton | which should be doable, just a bit of work to make sure there are no collisions | 11:24 |
| tjaalton | nope | 11:24 |
| RAOF | Although we'd need to pull in a newer drm.ko & ttm.ko. | 11:24 |
| tseliot | can't that be done with dkms instead? | 11:24 |
| tjaalton | there's nouveau-kernel-source for that | 11:24 |
| RAOF | That's what's currently done. | 11:24 |
| tseliot | and what's the problem with using the current approach then? | 11:24 |
| tjaalton | it should be moved to main | 11:25 |
| * tseliot is just trying to understand | 11:25 | |
| tjaalton | I guess | 11:25 |
| tseliot | ok, do we still default to "nv" for nvidia? | 11:25 |
| tjaalton | yes | 11:25 |
| tseliot | and does nouveau have 3D acceleration? (the version in Karmic) | 11:26 |
| tjaalton | no | 11:26 |
| tjaalton | needs some crackhead branch for mesa | 11:26 |
| tseliot | ok, then we should keep using "nv". Update manager will switch to "nv" with nvidia-common | 11:27 |
| tjaalton | yes | 11:27 |
| tseliot | good | 11:27 |
| RAOF | Upstream would be extremely unhappy with us if we shipped nouveau's 3D component. | 11:29 |
| tjaalton | yes, it's totally unsupported | 11:30 |
| RAOF | But we might want to use nouveau over nv, given that it's got 2d acceleration (and kms). | 11:30 |
| RAOF | Problem is, I've got no hardware that's only supported by -71, so I don't know nouveau's status on that hardware. | 11:31 |
| RAOF | Nouveau will be a bit more fragile than nv, too, since it recently regained a hard-dependency on the kernel module. | 11:32 |
| tjaalton | I should probably try nouveau again | 11:34 |
| RAOF | The edgers ppa version supports kms on everything I've got access to (and works nicely). | 11:36 |
| === james_w` is now known as james_w | ||
| Ng | --help | 14:32 |
| EagleScreen | can i disable UXA in karmic? | 17:17 |
| hyperair | no, i don't think you can | 17:20 |
| hyperair | why would you want to anyway? | 17:20 |
| EagleScreen | because UXA kill/hangs kdm on logout | 17:21 |
| jcristau | you can configure kdm to not do that | 17:21 |
| EagleScreen | what? configure kdm to not crash X server? | 17:22 |
| jcristau | yes | 17:22 |
| EagleScreen | what are you talking about? | 17:22 |
| jcristau | specifically, to terminate the server and using a new one, instead of reusing the same | 17:22 |
| EagleScreen | i think it is done by default now.. | 17:23 |
| jcristau | therefore avoiding the intel bug where it crashes when you init a second session | 17:23 |
| EagleScreen | any tip to do that configuration? | 17:23 |
| jcristau | there was something on the fdo bug | 17:24 |
| EagleScreen | fdo bug? | 17:25 |
| EagleScreen | which bug? | 17:25 |
| EagleScreen | i only know Bug #371500 | 17:26 |
| ubottu | Launchpad bug 371500 in xserver-xorg-video-intel "[i965gm] X server crash at closing session if kdm is in use. [UXA bug]" [Medium,Confirmed] https://launchpad.net/bugs/371500 | 17:26 |
| jcristau | 20516 iirc | 17:26 |
| jcristau | shouldn't be too hard to find.. | 17:27 |
| EagleScreen | sorry by my ingnorance, but, what does "20516 iirc" mean? | 17:29 |
| EagleScreen | i cannot find that option in kdmrc | 17:55 |
| micahg | is anyone here using the xorg crack on Jaunty? | 19:28 |
| Duke` | xorg crack? | 19:29 |
| micahg | https://edge.launchpad.net/~xorg-edgers/+archive/ppa | 19:29 |
| micahg | would the xorg crack drivers fix this problem: http://pastebin.ubuntu.com/250969/ | 19:33 |
| micahg | I'm running the ubuntu-x-swat intel 2.7 driver | 19:33 |
| micahg | ping bryce ^^^^ | 19:34 |
| tjaalton | maybe, if you get a new kernel too | 19:35 |
| micahg | yeah, well that ppa comes with | 19:35 |
| micahg | 2.6.30 | 19:35 |
| bryce | micahg, dunno not an error I've seen before | 19:35 |
| micahg | ah, ok | 19:35 |
| micahg | well, has anyone actually used the xorg-edgers stuff for a while? | 19:36 |
| micahg | xorg-video-intel 2.7 got my dual screens working, but it started crashing | 19:36 |
| tjaalton | karmic is more sensible :) | 19:36 |
| micahg | I'm wondering if 2.8 wiil fix it | 19:37 |
| micahg | I can't go to karmic as this is my prod laptop | 19:37 |
| tjaalton | well the errors are from the kernel | 19:37 |
| micahg | ah | 19:37 |
| micahg | I thought they were from the graphics driver module though | 19:37 |
| tjaalton | just try getting the kernel and nothing else | 19:37 |
| jcristau | tjaalton: do you guys have 865 issues in karmic? bgoglin says X hangs on startup.. | 19:37 |
| micahg | tjaalton: kernel from where? | 19:38 |
| micahg | xorg-edgers? | 19:38 |
| tjaalton | micahg: yes | 19:38 |
| micahg | ok | 19:38 |
| tjaalton | jcristau: I haven't triaged -intel for a while, but I remember someone on the forums saying that 8xx have issues with kms and disabling it worked around the problem | 19:39 |
| jcristau | k | 19:39 |
| tjaalton | s/for/in/ | 19:40 |
| micahg | tjaalton: it seems to have worked | 20:14 |
| micahg | xscreensaver no longer crashes my box | 20:15 |
| micahg | the kernel worked | 20:15 |
| bryce | jcristau, yes I sent a slew of 8xx bugs upstream over the past few weeks | 21:36 |
| bryce | jcristau, I speculate some quirks didn't get copied into the kernel, but most of the issues I sent were existing issues from pre-KMS | 21:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!