[00:00] I'm having trouble with coming out of resume using nvidia on Lucid [00:00] Anyone else able to replicateit? [00:00] I have a blinking cursor in upper-left corner, and after several attempts at ctrl-alt-f1-f6, I can eventually hit alt-f7 and get to the gdm login menu. [00:03] just curious, what does cat /etc/default/console-setup | grep XKBOPTIONS return for you? [00:03] # cat /etc/default/console-setup | grep XKBOPTIONS [00:03] XKBOPTIONS="lv3:ralt_switch" [00:03] it returns 'useless use of cat' ;) [00:03] console is a mess at boot, unless I pass vga=0... this is a big difference from karmic and jaunty [00:04] good point :D i was curious because alt+fkey is actually what I need to do to switch VT's outside of X and I had an option breaking control+alt+fkey [00:04] I had a 1920x1200 console at boot with those, but with Karmic, unless I pass vga=0, I get all kinds of colored pixel garble, in two identical vertical columns [00:06] afaik resume on nvidia is broken for everyone [00:07] we were messing with it last week at the developer sprint but couldn't see any nvidia user get it to resume successfully [00:07] It resumes fine, just takes a lot of vt switching to get it to go back into X [00:07] it's always worked great, until Lucid [00:08] I can also resume with the proprietary, but I get graphical artifacts if compositing is enabled [00:08] *proprietary driver [00:08] Coming out of resume on Karmic and Jaunty to the gdm menu was about 3-4 seconds after opening the laptop. This takes about 30 seconds and a lot of ctrl-alt-fx'ing [00:09] setuid, johanbr, then I think count yourself lucky ;-) [00:09] if you are interested in participating in debugging and patching the issue, contact tseliot [00:10] Didn't he die in 1965? [00:11] i'd like to figure out why alt+fkey is whats needed for vt switching now instead of control+alt+fkey :) when I had grp:ctrl_alt_toggle in my xkboptions I wasn't able to use control+alt+f at all, without it it at least ignores the control [00:14] bryceh: that seemed like it all, just added the backing out libdrm/plymouth/udev patches for lucid+1 part [00:14] ok thanks [00:19] https://wiki.ubuntu.com/X/Drivers will this page be updated for karmic/lucid? [00:19] raevol, go ahead if you'd like to update it [00:20] if I have some time later I'll add to it [00:20] :[ i don't actually know anything about the drivers aside from Karmic not having 3d support [00:21] I hadn't really given much priority to updating that page since you are actually the first person I have evidence of having even looked at it ;-) [00:22] hahah ;D [00:22] maybe it should be deleted [00:22] planet ubuntu just had a post about testing the proprietyary drivers, and linked the ubtuntu X wiki page for info on the open source ones [00:22] and i clicked through to that because i am hoping against hope we get 3d support for r600 in lucid [00:23] I've been looking around for some 'live wallpaper' ideas for the Ubuntu desktop... back in the day, we used to run things like xplanet, glmatrix, etc. in the root window, but that's not the same thing. [00:23] Anyone know of a Nexus One style live wallpaper app for Ubuntu? [00:26] setuid, not really the right channel for that question, try #ubuntu-desktop maybe [00:26] hahah, first I tried #ubuntu+1 because the problems I'm hitting are in Lucid only, they told me to go here, because they're X related... so I ask an X question, and that belongs in #ubuntu-desktop :) [00:27] gotta run, bye [05:43] RAOF, Sarvatt, https://wiki.ubuntu.com/X/Testing/NouveauEvaluation [05:44] I think you meant “grep _dr*v*.so /var/log/Xorg.0.log...” [05:45] (If you're still editing that page) [05:45] modified [05:45] is that actually needed? I've seen only _drv.so [05:46] Oh, I meant you currently have grep _dri.so rather than grep _drv.so [05:47] Also, users should be able to get away with just installing xserver-xorg-video-nouveau (& possibly the appropriate linux-backports-modules-nouveau); novueau-firmware will get pulled in by l-b-m-n. [05:47] ahh [05:47] Hm. And it's missing a apt-get upgrade; the newer X won't get pulled in automatically. [05:48] `grep _dr*v*.so /var/log/Xorg.0.log` actually did work on the cmdline :-D [05:48] :) [05:49] RAOF, go ahead and edit those switching directions into correctness [05:49] this is actually based off my experience pulling from xorg-edgers last week, I just cribbed it to the nouveau ppa, I've not actually run through and tested it yet (on the todo list for tomorrow) [05:50] Ok. Is there a particular reason you didn't use “add-apt-repository”, or can I change that, too? [05:51] go ahead [05:51] I just wasn't aware of that tool [05:51] Oh, it's really neet. [05:52] my lamitude is irredeemable [05:56] There we go. [06:04] thanks [06:05] btw, did you get good feedback from your call from testing the other week? [06:10] Is this a reasonable way to advise people to shut off KMS if needed? [06:11] # Intel: [06:11] echo i915.modeset=0 > /etc/modprob.d/i915-kms.conf [06:11] # ATI Radeon: [06:11] echo radeon.modeset=0 > /etc/modprob.d/radeon-kms.conf [06:11] # Nouveau: [06:11] echo lbm_nouveau.modeset=0 > /etc/modprob.d/lbm_nouveau-kms.conf [06:12] nouveau only supports kms [06:12] tjaalton, ah, right [07:08] bryceh: I got a much smaller response to the call for testing than I expected. No one reported show-stoppers, though. [07:09] well that's good [07:10] then we should do another call with this new ppa and the testing page. I've had a few people promise to do nouveau testing when it's ready, so maybe I'll call in those chits [07:11] RAOF: I've got gf8600/9600/some quadro, need testing on those? [07:12] quadro nvs 295 or so [07:14] tjaalton, https://wiki.ubuntu.com/X/Testing/NouveauEvaluation [07:14] if you ran through that to sanity check the directions it would be quite helpful [07:14] sure [07:23] bryceh, xorg as package (from where you can reroot them) is a good choice for people to report bugs on the prop drivers project? [07:25] tjaalton: Everything's worth testing. Frankly, I think everything that's not an IGP is going to be solid; it's the IGPs I'd worry about. [07:25] ara, that's fine [07:26] RAOF: ok [15:06] a nvaa IGP seems to be working now surprisingly [15:08] tried my 8200 out, didn't try 3D out but 2D was working. it's in my HTPC and the wife wanted the tv back so I couldn't try it long :D === yofel_ is now known as yofel [16:09] apw: regarding https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/404064 -- a fix has been cced to stable http://lists.freedesktop.org/archives/intel-gfx/2010-February/005803.html [16:09] Ubuntu bug 404064 in linux "KMS error message while intializing modesetting (during boot and resume) - render error detected, EIR: 0x00000010 [i915]" [Medium,In progress] [16:09] Sarvatt, hi [16:12] heyo! sorry to ping you, i was just scanning through bugs fixed by recent patches going into drm-intel-next and that one was assigned to you so thought you might want to know [16:15] ahh i see that now, missed the second link on first view [16:16] Sarvatt, i've added a link in the bug [16:18] the patch is actually attached in the upstream bug report thats already linked too [16:18] sweet [16:33] morning [16:33] any way to search the logs attached to bug reports on launchpad through the web interface? doesn't look like a search checks those [16:33] i'm looking for errors with [drm:i915_gem_object_pin] *ERROR* Failure to install fence: -28 fixed by a recent libdrm commit [16:42] morning bryceh [16:42] no, the web interface doesn't provide a way to search them [16:43] btw, how do we get the hwdb id for the nouveauevaluation wiki page? [16:43] there have been launchpadlib tools written which do that though, if you'd like to try your hand at crafting them [16:43] Sarvatt, I'm a bit fuzzy on how that works. I think the user gets it when they sign up [16:43] and if we have the id, we can look up the info [16:44] but there isn't a way for us, knowing the person, to find their hw [16:44] when we sign up for what? [16:46] for the hwdb [16:48] hmm i dont see hwdb-client or hwdb-gui in lucid [17:13] Sarvatt: looks like ickle just posted a fix for the libdrm crash you mentioned yesterday [17:14] it didn't fix 945 for me, already tested it :( [17:15] damn was hoping it was fence reg overallocation [17:15] crashed after 7 minutes [17:16] http://pastebin.com/f4a62401d [17:20] commit message said it was just changing i915 and prior but i have 3 fences reserved on 945 too [17:46] actually, it was a bit different, theres no uxa_prepare_access() error before the hang like there normally was - http://sarvatt.com/downloads/intel_crashes/02092010/ [17:46] in Xorg.0.log [17:48] http://sarvatt.com/downloads/intel_crashes/Xorg.0.log [17:49] thats what the hangs looked like before that libdrm commit [18:35] Duke`: are you around? [18:36] i just set up a 14 patch series for libdrm to revert intel back to 11-25 incase you want to try it with me to see if it still works [18:36] http://sarvatt.com/downloads/intel_crashes/libdrm-reverts/ [18:41] just installed that and am building intel git against it to see if its still working right [18:42] uploaded the debs incase you wanted them too [19:05] bryceh: got enough ppa's? :D [19:05] trying to find your cairo 1.9 snapshot to see if cairo-trace is in it [19:09] woohoo it is built with your package, just not installed. that saves me alot of time :D [19:25] Sarvatt, I get some color corruption - see the end of this dmesg: http://pastebin.com/f3ab3aa06 [19:25] Sarvatt, to be more precise it's font color corruption [19:26] http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=AIII,+invalid/inactive+channel+id+128 [19:26] funny the first hits are your exact gpu :D [19:27] heh :) [19:30] were you watching a video when it happened? [19:31] or opening a folder with videos getting thumbnailed in nautilus? [19:41] kklimonda: Sarvatt: that one (i guess), there's a patch, just hasn't hit upstream yet [19:42] from #nouveau [19:42] http://cgit.freedesktop.org/nouveau/linux-2.6/commit/?id=250cc31f35f00b5dd4b76e6c0bb9425b4fe23304 [19:45] will make a new lbm-nouveau with nouveau git instead of linus git in edgers sometime [19:49] we want testing with linus git though, if anything i'll try cherry-picking that into linus' tree and build a new lbm-nouveau but i can't at the moment because cloning kernel git will take a year on a tethered phone connection :) [19:49] hehe [19:50] if anything you could grab the source, then grab http://cgit.freedesktop.org/nouveau/linux-2.6/patch/?id=250cc31f35f00b5dd4b76e6c0bb9425b4fe23304 then change the patch paths at the to a/updates/nouveau/whatever.c b/updates/nouveau/whatever.c and apply it by hand with patch -p1 and rebuild [19:50] the linux-backports-modules-nouveau-whatever source that is [19:51] Sarvatt, will do if it gets more distracting than it is now [19:54] http://sarvatt.com/downloads/patches/nouveau_pgraph.patch [19:54] that should apply with patch -p1 < nouveau_pgraph.patch from the linux-backports-modules-whatever/ directory [19:55] thanks for that though, something we should probably pick up if it doesnt go upstream in time :D [20:07] superm1: did you upload a new DKMS in Karmic with my patch to fix bug #474917 ? [20:07] Launchpad bug 474917 in nvidia-graphics-drivers-180 "nvidia drivers 185.xx compile into kernel 2.6.28 instead of 2.6.31 on update from jaunty to karmic" [Critical,In progress] https://launchpad.net/bugs/474917 [20:08] if you don't plan to upload an update I'll mark the task for karmic as a won't fix [20:09] tseliot, didn't upload to karmic, probably wont at this point [20:09] should be fixed in lucid however [20:09] superm1: yes, I know. I just wanted to double check with you [20:16] superm1: also, did you have the chance to work on fglrx? [20:18] tseliot, yeah i think most of the conversion necessary should be done, but you should double check my work [20:18] it's all on phorogit [20:18] superm1: ok, let me check [20:18] since we dont have a (working) driver for lucid right now, it was a bit hard to verify [20:45] superm1: I made these changes (on the basis of what we do with nvidia): http://pastebin.ubuntu.com/372728/ [20:46] tseliot, that first set of changes should be in dkms's common.postinst [20:46] right, but can I use them if I source from that file? [20:47] without executing the rest of the script, that is [20:47] tseliot, Oh. hm. [20:47] what's update-initramfs's behavior without the -k flag? [20:49] superm1: it updates the kernel grub defaults to [20:50] hm well i understand better why you've got that in place now... [20:52] i guess that's the only way to do it for now then [20:52] I wish there was a cleaner way [20:56] * tseliot nods [21:01] superm1: I think I'll push my changes [21:01] tseliot, okay. do you know what piece is incompatible with the current fglrx in the X stack? [21:02] perhaps you can revert that single piece to the karmic version to be able to test all this [21:02] superm1: they didn't tell me what's missing but it mustn't be trivial if it's taking so long [21:03] tseliot, well a lot of that is probably their development process too. they normally develop 2-3 releases out [21:03] superm1: yes, of course it must be a matter of priorities too [21:08] I am actually a bit surprised they did not already put out a driver that works with xserver 1.7 [21:09] tseliot: please make it provide the abi again and conflict with the current server. no reason to let people install it just to see the server crash [21:10] I would love it if they'd fix their development process to deliver released drivers in time for ubuntu. Receiving those pre-release drops ends up just making everyone paranoid about canonical [21:11] tjaalton: that's a good idea [21:11] I dunno, maybe as a patch to the packages carried in ubuntu [21:12] tseliot: just providing the old abi is enough to make it conflict [21:12] bryceh: right, but I have a *secret driver* so they're not being paranoid :-P [21:12] it's hard to gauge which releases get supported by the packaging kept in phorogit sometimes [21:12] tseliot, you're a very sinister guy [21:12] and if you forget to update one of them, then suddenly a release that should work won't [21:13] bryceh: I am. Furthermore I'm working from our secret lab in Lexington right now :-P [21:13] say hi for me :-) [21:13] superm1: do you mean a patch which changes the control file? [21:14] tseliot: btw, what's wrong with the e71?-) [21:14] bryceh: sure, I'll say hi to dr. evil :-P [21:14] tseliot, just for now maybe uploading to lucid something that conflicts with the current x server [21:14] but not making that change in the upstream packaging [21:14] btw, are you subbed to xorg-prop-drivers-testers@lists.launchpad.net ? [21:15] tjaalton: the browser is an abomination. I can't even login on launchpad or on the website which should allow me to have my wifi connection in the hotel. Furthermore I keep pressing the "p" key when I try to hang up... [21:15] tseliot: heh yeah, the browser sucks. opera might be better though [21:16] superm1: aah, ok, that would definitely be better [21:16] bryceh: yes, I did [21:16] tjaalton: I'll try webkit when I get my nexus [21:16] :-) [21:16] tseliot: boo :) [21:17] :-D [21:48] ok [22:16] Sarvatt, I'm x86_64, your packages are i386 [22:34] bryceh: Oh! I've just noticed something obvious missing from the nouveau testing page: suspend/resume. I'll add that in. [22:35] RAOF, aha thanks [22:49] can you build it yourself Duke`? [22:49] i'll send the source if you want [22:54] Duke`: grab xserver-xorg-video-intel_2.10.0+git20100209.41784e15-0ubuntu0sarvatt2~karmic* from http://sarvatt.com/downloads/intel_crashes/libdrm-reverts/ then dpkg-source -x xserver-xorg-video-intel_2.10.0+git20100209.41784e15-0ubuntu0sarvatt2~karmic.dsc [22:59] Sarvatt, I'll see what I can do tomorrow, I need to go to bed now or I'll have an hard day :) [23:00] good night