[00:06] rickspencer3, which is the best day of the week for me to send my Intel upstream bugs report to Intel? Monday? === KiBi_ is now known as KiBi [00:35] What's the kernel option to force VESA from within grub? [00:36] Sarvatt's been working to make xforcevesa do that again. But depending on your card, there are other ways. [00:37] Hmm. [00:37] nomodeset maybe? [00:37] THis is an nvidia card that needs a vesa driver for at least another two days. [00:37] The beta freeze being lifted will likely change this, but... [00:37] (bleeding edge hardware) [00:38] Yeah. Nomodeset will disable nouveau, because the X component doesn't work without kms. [00:38] You'll fall back on either nv or vesa, depending on which one works. [00:38] Nope, same black screen issue. [00:39] What card is it that doesn't work with nouveau, by the way? [00:39] startx from single user mode works with vesa declared in xorg.conf, but no mouse or keyboard. [00:39] nvs3100m [00:39] You can get to single user mode, though? What sort of resolution is your terminal there? [00:40] Garbage, unfortunately. [00:40] But I have to throw a special option to get there. [00:40] text or single doesn't work. [00:40] init=/bin/bash is all that does. :-/ [00:41] Probably nomodeset would work as well. You could also try passing nouveau.noaccel=1, which would disable everything but modesetting, which is more likely to work. [00:41] blacklist=driver will use vesa [00:41] or it should :D [00:41] like blacklist=nouveau [00:43] RAOF, that's a brand new quadro chip he's got [00:43] only supported by the 195 blob [00:43] bjsnider: Only supported *with acceleration* by the 195 blob; modesetting changes a lot less than accel. [00:43] yes but i figured that's why nouveau doesn't support it [00:44] RAOF: Yeah, nomodeset fails [00:44] I don't suppose it's possible for you to grab a dmesg when it's failing? :) [00:44] So pass it blacklist=noveau? [00:45] I don't *think* so. [00:45] IE: got a second machine you can ssh into it with? :) [00:45] Actually yes. [00:45] Oh, well then. Unless it's actually locking the box, you can grab all the logs we could possibly want :P [00:46] And just for added bonus debug: you're using the 2.6.32-16 kernel, right? Previous kernel versions had all sorts of vga16fb-related nouveau problems. [00:46] Let me verify. [00:47] Can't start the ssh server because it's unable to connect to Upstart [00:47] Got it working. [00:47] (It's the X stuff that kills me, I know my CLI fairly well) [00:48] Ah heck, because it's ONLY got a shell, all I can hit is PTY allocation request failed on channel 0 [00:48] during ssh. [00:48] *ponder* [00:48] Let me reboot, see if I can ssh into it and grab the x log that dies. [00:48] That'd be what we're after :) [00:49] Holy crap I'm in. [00:49] We'd also want dmesg, because modesetting is now a kernel affair. [00:49] Full Have the whole thing, installing pastebinit [00:50] http://pastebin.com/tEvhp2Dn [00:50] http://pastebin.com/esBhQR7L [00:51] Yeah, looks like 2.6.32-14-generic [00:51] The following packages have been kept back: [00:51] Forgive my lack of debian knowledge, why would it do that? :-) [00:52] Because you need to pass dist-upgrade (to apt-get) or full-upgrade (to aptitude) for it to install new packages during the upgrade? [00:52] dis-tupgrade [00:52] Yeah. [00:52] Thanks. :-) [00:52] Sorry, googling as fast as I can. [00:52] It looks like you might also want to remove your xorg.conf [00:52] RAOF: Completely? [00:53] Yup. [00:53] Along with xorg.conf.failsave? [00:53] That can happily stay. [00:53] k [00:53] Update complete, rebooting. [00:53] Hey, at least it reboots when told to. [00:54] Holy crap. [00:54] It boots now. [00:54] Sound works. [00:54] Video works. [00:54] Mouse and keyboard work. [00:54] Wireless doesn't, but that's expected. [00:54] (Intel is releasing new firmware shortly that rectifies this) [00:55] RAOF: Okay, I'll bite. Why did I want to get rid of my xorg.conf? [00:55] Because it looked like it was confusing X, and for anything but the binary drivers it's unnecessary; X autodetects the correct drivers. [00:56] They have a restricted driver update for nvidia. [00:56] Is that going to hose me, or should I install it? [00:57] I don't know. I think that if hardware drivers offers it then that means the nvidia-glx drivers claim to support your card. [00:57] How hard is it to roll this back if it blows up again? [00:57] bryceh, Monday is good, yes [00:57] Not terribly hard; the worst case is that you run jockey-text in ssh to remove the nvidia drivers. [00:59] Restarting, let's see if it works. [00:59] Worked. [01:00] And COmpiz works too. [01:01] RAOF: I owe you a beer. [01:01] * RAOF will be at UDS :) [01:01] UDS? [01:01] Eh, it's all other people's work there. [01:02] Incidentally, before you installed the nvidia drivers, did you get the ubuntu splash screen? Did kms work? [01:02] Okay, I should probably clarify my position a bit. I'm a RedHat admin, never used Ubuntu before. Making the switch on the desktop away from MacOS. [01:02] RAOF: No to both. After, yes. [01:02] UDS == Ubuntu Developer's Summit [01:02] Ok. You'll have got keybuk's awesome VGA text-mode splash screen then. [01:03] !uds [01:03] The Ubuntu Developer Summit is being held from May 10th - 14th in Brussels, Belgium - See https://wiki.ubuntu.com/UDS for more information. [01:03] I live in Los Angeles. LONG way to go for a beer. :-) [01:03] RAOF: Tell me where to go, I'll do it. [01:04] The splash screen, not the summit. [01:04] It's near Brussels, but it's probably not interesting for you unless you've got a burning desire to help shape the next release of Ubuntu :) [01:04] Hmm, plymouthd crashed apparently. [01:05] And there's already a bug report open. [01:05] Slick [01:05] Apport is 32 separate shades of awesome. [01:06] Yeah, this is insanely slick. [01:06] Y'know, once you get a GUI working and all. [01:06] Also, the launchpad-ubuntu-package-bugs opensearch plugin + Do's opensearch plugin is love. [01:07] Plugin for what exactly? [01:07] Firefox, or Ubuntu? [01:07] The opensearch plugin is for firefox (or anything else which understands opensearch, I guess). [01:09] This is the fastest system I've ever used. [01:10] I think I'm in love. ;-) [01:10] But the purple has got to go... [01:10] * RAOF really likes the purple. [01:10] Well, aubergine :) [01:12] Yeah, but I"m used to a BLACK terminal window. [01:12] THis is a little too much of a pimping good time for my tastes. [01:17] There we go, black terminal. :-) [01:19] KB1JWQ, here's a really unfair question. why use ubuntu when you can use fedora? i mean if you're a redhat admin [01:20] bjsnider: That's a very fair question. [01:20] bjsnider: It comes down to maintainability. Fedora has this obnoxious habit of breaking things mid-cycle. [01:20] I've been meaning to learn debianisms anyway, and the hardware support seems to hit Ubuntu a little faster. [01:21] The forced upgrade death-march of an 18 month lifecycle with no deviation means that I'd be perpetually chasing something I can't maintain. [01:21] Not saying I'll never use it on this box, but as of today, Ubuntu seems to be the best option. [01:22] bjsnider: The community's another huge plus. People like you and RAOF just helped me get something set up that I'd have spent days working on. [01:27] but aren't you far more knowledgeable on fedora? you could even use centos i suppose, although that rig's hardware might be too new for its kernel [01:30] bjsnider: By that theory I'd have stuck with FreeBSD. :-) [01:30] And most of what I do day in and day out is application specific. Postfix is postfix, for instance, no matter what it runs on. [01:36] that isn't what the apache guys told me [01:36] but they're ornery [01:37] KB1JWQ, well, we're glad to have ya :-) [01:38] not that i'm arguing that ubuntu is the superior choice for a home system. that's very obvious [01:50] apw: i'm actually going to give this a shot to see if thats all thats needed in my case - http://sarvatt.com/downloads/patches/0001-drm-i915-Disable-FBC-on-915GM-and-945GM.patch [01:57] bjsnider: Precisely. It's a desktop box, and coming from Mac OS, I'd prefer something that's widely supported. [01:57] Ubuntu serves as a reference platform for a lot of things. [01:57] As to the Apache folks, I had dinner last night with one of their infrastructure guys. He raises "whining" to an art form. :-D [02:01] KB1JWQ, that rig cost as much as a crackbook pro, so why get that instead? [03:11] bjsnider: Because the one I have has a spilled drink into keyboard. [03:11] There's no accidental damage applicable to it, so the hell with it. [03:12] As shipped from lenovo it was $1400 or so. [03:54] xorg-edgers member list is starting to look like a canonical party list :D [09:09] RAOF, hello, is there progress investigating the mesa-nouveau-compiz problem? [09:15] ricotz: what problem? Nouveau doesn't work with compiz, at least AFAIK [09:17] tseliot, it is working with enabled 3d support, but the changes in mesa caused it to break [09:18] ricotz: ah, the experimental 3d support [09:19] tseliot, yes, it is working with mesa of 2010-03-13 [09:19] nice [09:23] Sarvatt, about? [10:00] hi the synaptic touchpad driver in lucid is pretty borked,this was a known issue in jaunty and was fixed in karmic,but seems to have reappeared in lucid [10:01] also theres no horizontal scrolling support [10:35] ricotz: I haven't invested any effort in it; nouveau's 3d sometimes works & sometimes doesn't. [10:36] This is why it's in xorg-edgers and not the main archives. [10:39] RAOF, i know -- btw when i use lbm from edgers the used module should be lbm_nouveau? [10:42] Yup [10:42] RAOF, can i help solving this problem somehow? i tried to get a trace which leads to "st_texture.h:128" with compiz and "nv40_fragtex.c:113" with mutter [10:42] in mesa source [10:46] with mesa git c9c0baabdc653f162f9ce51cb17775aed1a707f7 [10:52] ricotz: You'd want to talk to #nouveau about that. [10:53] Have a look at those files, and see if there's anything obvious? [10:55] http://cgit.freedesktop.org/mesa/mesa/tree/src/mesa/state_tracker/st_texture.h?id=c9c0baabdc653f162f9ce51cb17775aed1a707f7#n128 [10:56] RAOF, ^, g2g === apachelogger_ is now known as apacehlogger [12:40] yeah so like i was saying : xserver-xorg-input-synaptics : seems pretty borked right now with soft lockups with the synaptic touchpad [12:41] im repoting a bug against it right now,and if you would like additional details please let me know [12:46] ok the bug is : https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-input-synaptics/+bug/541868 [12:46] Ubuntu bug 541868 in xserver-xorg-input-synaptics "Synaptic touchpad has soft lockups in lucid" [Undecided,New] [12:47] hope that helps,please ask if any more info is required === apacehlogger is now known as apachelogger === shadeslayer is now known as shadeslayer_ [13:56] * hyperair wonders if anyone is getting graphical artifacts with 2.6.34-rc1 [13:57] hyperair: Custom kernel I presume? [13:57] KB1JWQ: yes. [13:57] + xorg-edgers [13:57] i965 card [14:12] Sarvatt, i've slurped up that patch you suggested (slightly modified) and my patch and produced a pair of test kernels for them ... just sent out an email to kernel and X asking for feedback on both [14:14] apw: nice, thank you [14:14] i'm trying to reproduce the hangs with different drm debug levels [14:16] drm.debug=0x2 is useless at least, its taking an hour or two to trigger and 0x4 spams the log a few times every second [14:18] ah nice theres the i386 kernel, installing it now and will try to reproduce [14:23] Sarvatt, sweet [14:52] with drm.debug=0x02 the only thing I noticed was that it called [drm:intel_crtc_cursor_set], to turn the cursor back on after typing around the same time it hung, took 4 hours to make it hang that time on -16 with powersave=1.. may be awhile until I can really say this fbc-broken kernel is fully working [14:53] been on it post resume about 30 minutes now though [14:55] hyperair: what kind of artifacts? are you on lucid yet? [14:56] the only artifacts I get aren't edgers related and I'm on 945, they happen when I'm doing stuff in gnome-terminal and it kind of corrupts/blacks out the screen until i move the mouse back up to the panel [15:41] Sarvatt: when images are loading in firefox, scrolling makes weird horizontal lines appear in the image, kind of like the edge of the status bar got rendered onto the image. [15:42] Sarvatt: and occasionally, notify-osd appears with graphic corruption [15:42] Sarvatt: mostly the volume up/down one [15:42] Sarvatt: hovering over it and dehovering restores notify-osd's popups [15:42] Sarvatt: then banshee's cover art gets graphical corruption.. [16:13] hyperair: lucid? [16:13] Sarvatt: karmic [16:14] what kind of graphic corruption on notify osd? [16:14] no-one cares about karmic anymore :) [16:14] tjaalton: heh. sorry, but i still kinda need this machine to work. i don't have enough time to spend fixing issues yet [16:14] oh hyperair I'm sure its that old crappy mesa checkout in edgers for karmic :D [16:14] Sarvatt: ._. it might be.. [16:15] Sarvatt: how do you describe what kind of graphic corruption? [16:15] Sarvatt: i mean it's got colours. [16:15] er [16:15] weird colours [16:15] and it is completely obscured [16:15] o hey my cursor's messed up from time to time too [16:15] background missing just around icons was the last one i saw [16:16] not quite it [16:16] my I cursor seems to be a little strange too [16:16] like i'm seeing double [16:16] with some random colours at the side [16:18] http://www.ubuntu-pics.de/bild/47853/screenshot_01_XQ8b5f.jpg [16:18] see thunderbird's icons? [16:18] darnit [16:18] i'm just upgrading to mesa 7.9 for karmic, sorry tormod if you read this :) [16:19] where would a user specific setting that breaks the touchpad be living? [16:20] describe breaks? [16:21] it works in gdm, and stops working in my normal user account, but works in a guest account [16:21] a plugged in mouse works fine in all accounts [16:21] hyperair: i'll upload a new mesa for karmic now, just going to bump it to 7.9 since it takes 5 minutes to do karmic by hand if I have lucid's and just replace debian/ vs an hour or so to do a whole new mesa and make sure its right [16:21] Sarvatt: cool. what should i restart then? [16:22] entire X? [16:22] you're using compiz right? is the corruption still there without it? [16:22] that's the first thing i'd check [16:22] hmm [16:22] lemme try.. [16:22] mesa in edgers for karmic needs updating really bad though anyway even if thats not the problem :D [16:23] <965 is screwed in there right now unless people manually disable texture tiling support in driconf [16:23] ouch. [16:23] why? [16:24] was enabled by default and hitting assertions for a few weeks [16:24] just got fixed 2 days ago [16:25] Sarvatt: you're right. it's not happening in metacity. [16:26] doh, forgot to upload libdrm to karmic too yesterday even though I did it, go figure I uploaded mesa right before [16:27] er. [16:27] mesa will fail to build but i'll retry it [16:27] now that compiz has restarted itself, i suddenly can't reproduce it [16:27] * hyperair grumbles [16:34] apw: 2 hours strong with no FBC post resume and powersave=1 [16:35] Sarvatt, with the kernel with just your fbc disable patch ? [16:35] yeah [16:35] your fbc-broken -17 kernel [16:35] thats awsome, lets see what other people say [16:35] as that is a smaller cleaner fix [16:37] yeah it helps narrow down where the problem is too even if it does eventually hang since theres not much else done with powersave. you said you edited it some? can ya send me what you applied to this kernel so I can bring it up upstream? [16:41] ah, yet another gconf-schema-not-installed-for-whatever-reason problem; gconf-schemas --register-all brought the touchpad back [16:48] hyperair: as soon as libdrm builds on i386 i'll retry those mesa's, they needed the libdrm update due to the annoying nouveau_class.h changes in libdrm.. [16:57] btw mesa 7.7.1 is going to be released next friday [16:58] and wacom was released today [16:58] wacom 0.10.5 === yofel_ is now known as yofel [17:04] ricotz: ya should bring up your issue in #dri-devel or #nouveau, give them a log with the trace messages when you try to start compiz [17:11] Sarvatt, thanks, have to looked into it on your box? [17:12] i am getting different segfault positions with mutter and compiz [17:13] Sarvatt, i have used gdb, but i am not that familiar yet, i only get the positon not the whole trace [17:16] try installing at least libgl1-mesa-dri-dbg and libgl1-mesa-glx-dbg then gdb --args mutter --replace then run then bt full once it segfaults? [17:19] Sarvatt, ok, thanks [17:34] Sarvatt, http://paste.ubuntu.com/397910/ [17:37] Sarvatt, and with compiz http://paste.ubuntu.com/397911/ [17:39] * Ng wonders how much pain will be involved in upgrading an 855GM laptop to lucid [17:40] +Sarvatt: yeah the first one is what I just noticed and what I am investigating now [17:40] +and second one looks different but is the sampler view code we suspect [17:40] ricotz: ^ [17:58] RAOF: did the apport-collect thing,hope that helps... [18:06] RAOF: yay tseliot has a nva3 that the nouveau guys have been looking for a dump of for months [18:13] Sarvatt: any idea if 3D support will make it till the final release? [18:13] shadeslayer_: for what? [18:13] nouveau? guaranteed no if so [18:13] of Lucid.. [18:13] yes nouveau [18:13] theres 3D support in lucid, i dont know what GPU you have [18:14] yeah there wont be 3D support [18:14] its *wildly* unstable [18:14] hm.. and what about nvidia+plymouth? [18:14] Sarvatt: for NV50 too? [18:15] it'll be ugly and low resolution as well as hardly any colors but it should work, there is no smooth transition from the splash to the desktop like there are with KMS drivers so its only displayed for about a second at the moment.. [18:15] yeah theres a bug against it [18:15] what for NV50 too? there wont be 3D support for anything nouveau until +1 [18:16] hmm... too bad :( [18:17] i guess ill have to dump the splash then [18:17] and I dont know if I'd call it support, planning on having an experimental dri package people can install seperately if they want to break things and I doubt it'll be really stable in that timeframe [18:18] Sarvatt: you mean : libdrm-nouveau1? [18:18] nah libgl1-mesa-dri-experimental or something [18:18] whats that for? [18:19] its most likely not going to be stable enough to globally install with libgl1-mesa-dri [18:21] what does it do tho?' [18:22] Sarvatt: have a look https://bugs.launchpad.net/bugs/526892 [18:22] Ubuntu bug 526892 in plymouth "No graphical splash on VGA16fb (e.g., nvidia binary drivers), plymouth uses text plugin ("Ubuntu 10.04" in text)" [Medium,In progress] [18:23] how is it WIP?? :P [18:24] the graphics devices can load sooner, they used to [18:24] hmmm [18:26] you'll see the splash longer if it has to fsck, or if you have to enter a password for an encrypted home directory, otherwise theres not really anything that can be done besides loading the devices earlier and it'll still only be there for a few seconds [18:27] would you rather the boot take longer so you can see it? [18:27] lol no [18:28] i would rather have the login screen bought up and then load the processes in the background till i enter my pass [18:30] thats exactly what happens [18:30] nice [18:35] theres just no way to transition from the splash and leave it on the screen while the login screen is loading when KMS isn't used, especially with a binary driver we have no control over :D the only option I see would be to load plymouth on another VT initially when the drm renderer isnt used and not switch to the login screen's VT until it signals that its done loading maybe? === radoe_ is now known as radoe === maxb_ is now known as maxb [19:14] apw: pushed it upstream - http://lists.freedesktop.org/archives/intel-gfx/2010-March/006292.html [19:15] excellent ... i am off next week ... if you need anything smb should have the branches available with those one [19:55] Sarvatt: do we have a 3d enabled nouveau (with mesa I assume) in some ppa? [20:38] tseliot: yeah but upstream mesa nouveau is pretty busted right now - https://edge.launchpad.net/~xorg-edgers/+archive/ppa [20:41] Sarvatt: ok, I asked as, with that patch that they gave me, they told me that I can even try 3D :-) [20:41] sweet, can you pass along your patch? I can add it to linux-backports-modules-nouveau in xorg-edgers for you [20:44] Sarvatt: I'm still rebuilding my kernel to test it but here it is: http://0x04.net/~mwk/0001-drm-nv50-Add-NVA3-support-in-ctxprog-ctxvals-generat.patch [20:44] the kernel nouveau wont work for 3D though [20:45] oh, well, 2D acceleration should be enough for now [20:48] uploading a new lbm-nouveau with that patch added to xorg-edgers now, should be up within an hour [20:48] fantastic, thanks [22:00] *whoa* we got a fglrx now? [22:06] * tseliot whistles [22:06] we do ?!? [22:06] really? [22:06] can you begin to imagine HOW MANY ppl we would make happy ? [22:07] :-) [22:07] * BUGabundo blows calculator stack trying the math [22:07] i would be 100% happy with R600 -radeon, its not like the nvidia situation :D [22:10] talking about nvidia [22:10] what's going on with nouveau? [22:10] #+1 is filled with ppl complaning it fails to boot [22:12] i went back 20 pages of +1 backscroll and i only see one person with a nvidia xorg.conf trying to use nouveau [22:13] pages? [22:13] there was a guy who said it resulted in a black screen and expressed a desire to use nv [22:14] thats all my bouncer forwards to me [22:14] 1000 lines not including joins/parts [22:15] i think his hardware was too new or whatever [22:15] bjsnider: I read that one [22:16] humm [22:16] no, I think that was another one [22:17] Seq at 2:36? [22:17] and nasso_ at ten to 1 [22:18] yep [22:18] and 3 or 4 more this morning [22:30] Sarvatt: what do I need to do to enable lbm nouveau? Shall I set an alias? [22:31] just activating the PPA and deactivating nvidia-current will work [22:31] i've got it blacklisting nouveau and vga16fb [22:32] Sarvatt: my system begs to differ ;) [22:32] aah [22:32] do you still have /etc/modprobe.d/nvidia-whatever? [22:33] maybe the udev and plymouth support things for lbm-nouveau were just backed out, eek [22:33] no nvidia-$flavour in modprobe.d here [22:34] lbm-nouveau loading in dmesg? [22:36] nope [22:36] wait, are you sure you didnt boot that 2.6.32-17 kernel? [22:36] lbm-nouveau is only for 2.6.32-16 [22:36] yay corruption doesn't seem to happen any more =D [22:36] thanks Sarvatt! [22:36] no worries, sorry i forgot to hit rebuild :D [22:37] $ uname -r [22:37] 2.6.32-16-generic [22:38] this is with a blacklisted nouveau and vga16fb [22:38] :~$ dmesg | grep nouveau [22:38] [ 9.505893] Error: Driver 'nouveau' is already registered, aborting... [22:39] :~$ cat /etc/modprobe.d/blacklist-alberto.conf [22:39] blacklist nouveau [22:39] blacklist vga16fb [22:39] and I updated the initramfs with sudo update-initramfs -u -k 2.6.32-16-generic [22:40] Sarvatt: I installed only linux-backports-modules-nouveau-2.6.32-16-generic [23:20] ok I have no clue whats wrong there, sounds like the lbm-nouveau support was dropped from plymouth though [23:20] anyone know where the ubuntu plymouth bzr branch is? i cant find the darn thing anywhere [23:21] lp:ubuntu/plymouth.. go figure [23:21] what the heck is going on with plymouth right now though, it gave me the fugly text logo on intel that boot [23:23] sorry, I shouldn't say fugly because its *much* better than the tricolor progress bar text had before, just wasn't expected since I got the real one this morning [23:25] my nvidia blob laptop suspends \o/ \o/ [23:29] lbm-nouveau support is still in plymouth so its not that, udev rules still there too [23:34] oh wait, tseliot's problem was he only installed lbm-nouveau, no wonder :D [23:50] hi , Somebody help me about gma 500 on ubuntu 10.0.4 (acer 751/h) please (driver) [23:51] there is none as far as I know [23:51] besides vesa which it should use automatically