=== Amaranth_ is now known as Amaranth [05:21] Oh, balls. Nouveau's gained DRI2 vsync support, and it looks like it's going to crash in *exactly* the same way intel's and then radeon's did. [05:22] intel's vsync crashed? [05:23] Yeah, on client close. [05:25] Because DRI2 buffers aren't reference counted, and get destroyed on client close, and the vsync event can occur after the request is queued and the client shutdown... === Bernardo is now known as Bernardo|Away === njpatel is now known as njpatel_ [08:41] RAOF, heh phoronix thinks you should enable gallium 3d for nouveau [09:33] wait what? phoronix thinks? [10:25] jcristau, I'm trying to figure out how to merge the words "phoronix" and "paparazzi" [11:15] paphonorazzix! [11:15] "phoronazzi" ?-) [11:15] if in doubt, smush things together with as much force as possible ;) [11:15] tjaalton: sounds a bit too much like nazi [11:15] heh [11:16] jcristau: meh :) [11:23] it's harder than it seems it should be [11:26] phopronozzi is best I've come up with but it has prono in it [11:27] maybe that's a good thing. pr0n is good for page hits. [11:27] lol === yofel_ is now known as yofel [16:12] apw: ok there's more to it, every other boot with gfxpayload=keep is screwed in grub space still, it's stuck before the kernel selection screen [16:13] Sarvatt, i don't see that behaviour on any of my kit, i always get to a kernel before things go wrong [16:13] not that it supprised me there is some h/w which gets broken [16:32] sheesh, moving drm.so out of the way isn't helping anymore either, I guess it's just really racy and I got lucky because it was doing a fsck that one time. almost done building an i386 kernel with Sanitize modesetting registers.(v2) [16:36] htop is fun on tangerine.buildd, guess you need 2560x1600 to use it with 64 cores :) [16:56] argh [16:56] apw: If I cold boot Sanitize modesetting registers.(v2) fixes it every time, 11 boots that way worked perfect [16:57] apw: but if I reboot instead, it hangs at GRUB loading. [16:57] hrm [16:57] (v2) ? [16:57] yeah from your bug https://bugs.freedesktop.org/show_bug.cgi?id=32078 [16:57] Freedesktop bug 32078 in DRM/Intel "Intel N10 graphics does not initialise correctly if GRUB hands off the display in a graphics mode" [Major,New] [16:58] Sarvatt, damn i'd started the previous one building and been on the release meeting ... and [16:59] * apw restarts his build ... gah [17:00] can you grab it from /home/sarvatt on tangerine? [17:05] Sarvatt, its nearly built again now anyhow ... but thanks === Bernardo is now known as Bernardo|Away [17:18] i'm totally stumped now, cold boot with that patch for sure works right but a reboot hangs in grub every time === Bernardo|Away is now known as Bernardo === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [17:49] using vesafb sure does make VT's suck on the blobs [18:07] apw: I haven't heard any complaining about radeon yet, with intel fixed maybe this will work this cycle :) [18:07] Sarvatt, the vesafb stuff looooks _sooo_ bad :) [18:07] but yeah there is some hope perhaps [18:08] yeah wish we could kill it in a fire, I know I'll still be disabling it on all my systems by booting with vesafb.sucks=1 :) [18:09] it's not like it helps that much even with the blobs since they turn off the display when X starts unconditionally, I see the splash for all of 1 second [18:11] text plugin progress dots loaded in the initramfs early for everyone looks best to me :) it's just the Ubuntu 10.10 words that look like crap in text mode [18:13] well the Ubuntu 10.10 text looks nice on a drmfb at native res, but crappy on 80x25 [18:19] Sarvatt, hehe yeah sometimes i wonder if that would be a better plan overall [18:34] hmm, if i boot an old kernel (2.6.35-23) then grub works after a reboot, if i boot 2.6.37-7-generic with or without that patch grub hangs after a reboot every time. [18:34] 2.6.37-5 hangs === Bernardo is now known as Bernardo|Away [18:38] 2.6.36-1 is the last good kernel that reboots right, fun === Bernardo|Away is now known as Bernardo [18:42] Sarvatt, have you tried any backports of -intel to maverick? Is it at all feasible? [18:43] yeah just needs a newer xutils-dev, lowered of the xserver-xorg-dev build-dep to 1.9.0 and a natty libdrm backport [18:43] want me to put it in x-updates while i'm bisecting? [18:45] Sarvatt, yes that would be great [18:45] Sarvatt, I'm thinking it might help on bug 682712 [18:45] Launchpad bug 682712 in xserver-xorg-video-intel (Ubuntu) (and 3 other projects) "x server crashes with: [drm:i915_gem_mmap_gtt_ioctl] *ERROR* Attempting to mmap a purgeable buffer (affects: 3) (heat: 18)" [High,Triaged] https://launchpad.net/bugs/682712 [18:46] but thats lucid? [18:46] some of the reporters mention maverick [18:46] although they may be seeing different bugs [18:47] it fixes all of these https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/626967 [18:47] Launchpad bug 626967 in xserver-xorg-video-intel (Ubuntu) "[arrandale] GPU lockup 85a8ad2d (IPEHR: 0x01820000) (affects: 8) (dups: 20) (heat: 120)" [Medium,Confirmed] [18:47] Sarvatt, that's a real good sign [18:47] which was fixed by http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=a44a63d2ff6c01c3dc61de6f736dd441ddd25e52 [18:47] ooh [18:48] Sarvatt, well, a maverick build of that driver would probably help move things forward [18:48] if that works for the maverick folks, then if apw still needs it on lucid maybe we can look at building it for that too [18:52] Sarvatt, is that worth considering backporting to maverick? [18:54] yes very much so just haven't had a chance to do it [18:59] apw: oh fun, it looks like some ubuntu specific patch is breaking the boot on this netbook. mainline 2.6.37-rc3 can reboot fine, 2.6.36-1 is the last ubuntu kernel that can reboot === Bernardo is now known as Bernardo|Away [19:01] ahh didn't try 2.6.37-2 or 2.6.37-3, whoops. 2.6.36-1.7 is good though [19:16] bingo, https://lists.ubuntu.com/archives/natty-changes/2010-November/001282.html broke reboot for me [19:19] ah darnit, 2.6.37-4.12 was broken that time, booted the wrong one [19:20] yep 2.6.37-3.11 is bad, 2.6.37-2.10 is good === Bernardo|Away is now known as Bernardo === Bernardo is now known as Bernardo|Away [19:42] hey sarvatt, is alpha broken on the stable intel/mesa bits for natty? === Bernardo|Away is now known as Bernardo [19:44] not on mine, booting is broken without disabling gfxpayload=keep in grub here though [19:49] sarvatt: http://paste.ubuntu.com/539524/ [19:49] white boxes and everything [19:49] i945gm [19:50] have you ever had the touhou game working in wine? [19:50] because i never have had any of them working [19:50] yes. [19:50] as far back as jaunty [19:50] if you have dx9 libs installed [19:52] where are glx settings stored? [19:53] any idea where it stopped working? did it work in maverick? [19:53] yeah [19:53] could i downpin to maverick for drivers? [19:54] have you tried changing the orm= options in winetricks to see if it works in one of the others by any chance? [19:55] yeah [19:55] backbuffer loads faster [19:55] still white boxes [19:56] no textures whatsoever [19:56] wine 1.3.8 [19:57] glxgears is fine though [19:57] lemme try 1.2 === Bernardo is now known as Bernardo|Away [20:04] what gpu do you have? I can try it with a few mesa versions if its a 945 [20:04] [#Comiket](C78) Yousei Daisensou - Touhou 12.8 Sangessei - The Great Fairy Wars.zip ? using an english patch? [20:04] yeah [20:06] 00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03) (prog-if 00 [VGA controller]) [20:06] need vendor id? [20:07] nope, perfect [20:07] grabbing Great Fairy Wars English Patch v1.1 [20:10] japanese version is messed up too [20:10] doesn't really matter [20:13] btw, brb restarting x === jewsucanuse is now known as erappleman [20:15] jewsucanuse: well, works out of the box with wine 1.3.8 and mesa 7.10 from git, hm [20:16] heh outside of unity being screwed up http://sarvatt.com/downloads/Screenshot.png [20:17] oh I take that back, I'm using mesa from the archives [20:17] unity is a mess right now. it can't handle video player windows right. [20:17] 7.9+repack-1ubuntu3 [20:17] they creep upwards. [20:17] this is a 945GSE, it's the same gpu to mesa's eyes [20:18] i have no dx overrides and no non standard config changes [20:18] do I have to go into the actual game to have it screw up? [20:18] no [20:19] lemme screenie [20:19] menus are fine, i'd put it down to a wine config problem [20:19] i cleared the configs [20:19] winetricks and wine [20:20] erappleman: do you have vertex shader support on none and unchecked allow pixel shader too? [20:20] mine's set to vista also === Bernardo|Away is now known as Bernardo [20:21] everything's default [20:21] http://img593.imageshack.us/img593/5980/screenshotyouseidaisens.png [20:21] http://pastebin.com/fCibRJ76 [20:22] hmm [20:22] i've got an idea [20:23] wget -O ~/.drirc http://sarvatt.com/downloads/drirc.txt [20:23] driconf it? [20:23] driconf cant adjust vblank_mode for dri2, gotta do it manualy [20:24] vblank_mode enabled screwed up with white screens like that with clutter apps for me sometimes and i'm using that, just a long shot [20:24] dri2 on screen 0? [20:25] or should it be i915? [20:25] needs to be dri2 [20:25] driconf only adjusts the i915 driver and it doesn't work unless its for the dri2 driver [20:26] Driver "dri2" is not installed or does not support configuration. [20:26] save your ~/.drirc and overwrite it with that one and move it back later if anything [20:27] you adjust other things in there? that could screw it up too [20:27] * Sarvatt tries to screw it up [20:28] hmm it doesn't fullscreen for you? [20:29] it does [20:30] dri2 is my only drirc entry. doesn't fix it. [20:30] this is slow as heck on this netbook [20:30] you have window borders in that screenshot is why i was asking, theres a checkbox I can't read because no jp support in wine on this thing, maybe that toggles fullscreen [20:31] 1st option [20:36] lemme try edgers === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [20:38] do i sudo driconf? [20:40] nope that'd only change settings for root and running wine as root screws everything up === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [20:47] darn gnome classic session is all kinds of screwed up, gnome-panel segfaulting on start [20:48] unity: fixing what isn't broken [20:49] Sarvatt, kees was complaining about that yesterday === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [20:54] erappleman: yeah outside of being horribly slow on an atom cpu it's working fine here with stock natty packages outside of wine 1.3.8 even windowed, I'm stumped [20:55] the error spam from wine is a red herring, I get it all even though it works === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [20:56] well, what i'm seeing is what i see whenever alpha channel is broken in the drivers yet doesn't always affect the gnome environment [20:59] ahh enabling stub ARB_occlusion_query and ARB_fragment_shader support trying to break it is what killed the speed [20:59] lemme try the windows world of goo demo [20:59] doesn't that force glsl on 945gm? [20:59] yea [20:59] was trying to screw with options to get it to mess up [21:00] thanks for giving me an excuse to play touhou at work btw :) [21:00] world of goo is screwed up too [21:00] lemme see native linu [21:01] world of goo native is fine here [21:01] native is fine here too [21:02] goo on wine isn't [21:02] hmm, you on amd64? [21:02] hmm, the wayland mesa fails to build on natty [21:03] i686 [21:03] just a thought that you have old mesa in ia32-libs [21:03] darn [21:03] nouveau_context.c:132:9: error: too many arguments to function 'nouveau_channel_alloc' [21:03] /usr/include/nouveau/nouveau_channel.h:51:1: note: declared here [21:03] make[7]: *** [nouveau_context.o] Error 1 [21:03] http://launchpadlibrarian.net/60063332/buildlog_ubuntu-natty-i386.mesa_7.10.0%2Bgit20101118.3dcc3153-0ubuntu1~bryce9~natty_FAILEDTOBUILD.txt.gz [21:03] yeah need git libdrm [21:03] look familiar to anyone? [21:03] i need the git? === Bernardo is now known as Bernardo|Away [21:03] natty libdrm doesn't have the nouveau stuff from post 2.4.22 === Bernardo|Away is now known as Bernardo [21:03] 2.4.22+git20101127.1443bea4-0ubuntu0sarvatt new enough? [21:03] erappleman: was talking to bryce sorry [21:03] ah [21:04] bryceh: yeah [21:04] that already in there? [21:04] ok thanks [21:04] I happened to copy it in while waiting for mesa to build :-P [21:04] what's making nouveau 3d still a no-go? i know it wasn't ready for lucid or maverick, but why not natty? [21:06] erappleman, upstream is not yet supporting it (no bug reports) [21:07] erappleman, as well, it feels like something that's going to have a lot of weird random bugs for oddball hardware or strange corner cases, and we're not sure the group of us have the skillz/time for solving them [21:07] loldebian [21:07] erappleman, huh? [21:07] isn't debian our upstream? [21:08] erappleman, nouveau upstream at fdo [21:08] oh [21:09] for Xorg the debian guys prefer us to work directly with X.org upstream if it's issues unrelated to just packaging [21:18] ah. sarvatt. [21:18] dx7 works [21:18] not dx8 or dx9 [21:19] how are you changing dx modes? [21:19] dxdiag in wine [21:21] BINGO [21:21] downgraded wine to 1.0.x [21:21] nice. === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [21:28] bryceh: ok intel in x-updates, sorry that took so long [21:31] Sarvatt, awesome thanks [21:33] thanks sarvatt === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo === Bernardo is now known as Bernardo|Away === Bernardo|Away is now known as Bernardo [22:29] Hi I'm having a problem with X crashes, I recently replaced my nvidia 6600GT with a 9500GT, now X has started crashing. I followed the Obtaining backtraces instructions and got a backtrace, which is here http://azelphur.com/crashlog.txt.zip [22:29] I had to zip it because it's a pretty large backtrace (17MB) [22:30] It looks like libarecord is doing something crazy and recursive that's causing X to crash :s [22:42] Azelphur: guessing you were using nvidia-173 for the 6600GT and left it installed when you switched? [22:42] I did originally, then I uninstalled anything to do with nvidia and reinstalled using jockey to ensure I had the right driver [22:43] can ya pastebin the /var/log/Xorg.0.log of it trying to start up? [22:43] oh hey, I do have nvidia-173-modaliases [22:43] Sarvatt: sorry, it does start up, it just crashes after a few hours [22:43] thats ok, modaliases are needed for jockey to work [22:43] ah [22:44] what version is the nvidia driver and what ubuntu release are you using? [22:44] even if the machine is left idle it'll crash all by itself, so I don't think there's anything I'm doing to trigger the crash [22:44] it could be a ton of things, there should be some trace of the crash in /var/log/Xorg.0.log.old [22:44] I'm using maverick (specifically mythbuntu) and I'm using nvidia-glx-185 [22:45] Sarvatt: yea, I checked there and there's literally nothing, which makes it weird [22:45] I can only guess that X is crashing before it has a chance to write anything to the log file [22:45] well not literally nothing, but nothing related to the crash [22:46] how does it crash? [22:46] screen go blank and not come back? does X restart? [22:46] X restarts, so I end up at the login screen [22:47] there should definitely be something in /var/log/Xorg.0.log.old when that happens, hmm [22:48] http://pastebin.com/WFiVW0eM [22:48] I don't see anything :( [22:48] can ya cat /var/log/Xorg.*.log | pastebinit [22:48] err wait [22:48] haha, beat you to it [22:48] cat /var/log/Xorg* | pastebinit [22:49] pastebin above is that [22:49] maybe its /var/log/Xorg.1.log.old or something [22:49] yea, I did Xorg* [22:49] oh hmm [22:49] look in /var/log/gdm [22:50] bryceh: there's a lot of files in there, anything in particular? [22:50] sudo cat /var/log/gdm/:0* | pastebinit ? [22:51] http://pastebin.com/j29tsTbQ [22:51] Azelphur, usually can pick it out by last change timestamp on the file [22:52] i'm stumped, no traces of you even logging out ever [22:52] D: [22:53] dmesg? [22:53] has it happened since you booted last? [22:53] bryceh: nothing, checked that :( [22:53] only 4 messages for the entire day, none of them relevant [22:53] Sarvatt: yes [22:53] can ya pastebin the dmesg just in case? [22:54] ~/.xsession-errors might be handy too but you might not want to pastebin that in case there is anything sensitive in it [22:55] http://pastebin.com/WM8HaB0S got a bit longer since I checked it at 4 lines, but nothing interesting I can see [22:55] that's all of today, and it's crashed a few times today [22:55] I don't have a ~/.xsession-errors [22:55] oh wait derp, logged in as root when I was looking at the gdm folder [22:56] http://pastebin.com/dNs0sm53 there's xsession-errors [22:58] Azelphur, have you tried reinstalling the driver? [22:58] bryceh: yes [22:59] this bit's interesting: [22:59] # [22:59] 03/12/2010 19:09:42 Scroll Detection: -scrollcopyrect mode is in effect to [22:59] # [22:59] 03/12/2010 19:09:42 use RECORD extension to try to detect scrolling windows [22:59] # [22:59] 03/12/2010 19:09:42 (induced by either user keystroke or mouse input). [22:59] # [22:59] 03/12/2010 19:09:42 If this yields undesired behavior (poor response, painting [22:59] # [22:59] 03/12/2010 19:09:42 errors, etc) it may be disabled via: '-noscr' [22:59] # [22:59] 03/12/2010 19:09:42 Also see the -help entry for tuning parameters. [22:59] well I see it crashing there but nothing about why it happened.. are you sure you don't have logout mapped to a button on the remote or something maybe? [22:59] # [22:59] 03/12/2010 19:09:42 You can press 3 Alt_L's (Left "Alt" key) in a row to [22:59] # [22:59] 03/12/2010 19:09:42 repaint the screen, also see the -fixscreen option for [23:00] # [23:00] 03/12/2010 19:09:42 periodic repaints. [23:00] # [23:00] 03/12/2010 19:09:42 X FBPM extension not supported. [23:00] Sarvatt: nope, as I said I've seen it do it while all input devices where a good 5 meters are from any people :) [23:02] it'd help to have symbols installed [23:03] bryceh: what's the package for the symbols? [23:04] guessing librecord is capturing some event from a window, then triggering a callback via CallCallbacks, then when that's written to the client it triggers a librecord event [23:04] but hard to guess without line numbers [23:05] Azelphur, xserver-xorg-core-dbg maybe [23:06] Azelphur: what's the version of your xserver-xorg-core package? [23:06] got that installed, next time it crashes I'll get another backtrace in here :) [23:07] Sarvatt: 2:1.9.0-0ubuntu7 [23:07] i'm going to go out on a limb and say some part of mythtv uses qt somewhere and you're hitting the bug fixed in maverick-updates [23:07] because you're using xinerama too [23:07] mythtv is pretty much entirely qt, so yea... [23:07] Azelphur: ahhh enable maverick-updates [23:07] shiny :D [23:07] or is xorg-server not in -updates yet? :( [23:08] xorg-server (2:1.9.0-0ubuntu7.1) maverick-proposed; urgency=low [23:08] * debian/patches/207_Xext_panoramiXprocs_fix_typo.patch: [23:08] - This prevents Qt applications from crashing when using [23:08] Xinerama multi-head with drivers such as nvidia (LP: #650539). [23:08] I'm not using xinerama though [23:08] maybe when mythtv crashes because of that it restarts the session in mythbuntu [23:08] at least I shouldn't be, it's a single display [23:08] your log said you were I thought [23:09] that's odd then, there is only one display [23:09] 03/12/2010 19:09:42 Xinerama is present and active (e.g. multi-head). [23:09] 03/12/2010 19:09:42 Xinerama: number of sub-screens: 1 [23:09] 03/12/2010 19:09:42 Xinerama: no blackouts needed (only one sub-screen) [23:09] fun, maybe it's an s-video out thing [23:09] wonky [23:09] I have maverick-updates enabled, did you mean maverick-proposed or backports? [23:10] Azelphur: ahh darn, I was hoping it migrated to -updates already :( [23:10] yeah it's in proposed for sure if its not there [23:11] ok, got myself on -updates and running an upgrade now, will keep an eye on it and see if it does it again, if it does hopefully I'll have a better backtrace with symbols this time :) [23:11] i'd enable proposed, apt-get update && sudo apt-get dist-upgrade (just to see what packages are offered, hit n) then upgrade the ones you need. should be xserver-xorg-core xserver-xorg-core-dbg xserver-common [23:12] I'm cool with -proposed, I use -proposed on my PC anyway [23:18] anyone happen to have a machine using fglrx on natty or maverick handy that could try xorg-edgers? need to know if fglrx works without an xorg.conf with it before trying to push that into natty :) [23:18] nvidia has been working fine for the past week of me trying to break things with no xorg.conf [23:21] would be nice to not have things broken because of the old xorg.conf when people remove nvidia or fglrx outside of jockey [23:28] ugh yeah we need a fixed up libvdpau for 32 bit flash acceleration to work on amd64, it works in debian [23:29] bjsnider: install http://ftp.us.debian.org/debian/pool/main/libv/libvdpau/lib32vdpau1_0.4.1-2_amd64.deb :) [23:29] we can't build that package because ia32-libs is in universe and libvdpau in main.. [23:40] yay, finally no more 8080's in my urls [23:40] http://www.bryceharrington.org/X/Reports/ubuntu-x-swat/versions-current.html [23:47] huh, how does wine manage to build-dep on ia32-libs yet be on the livecd? [23:48] wine1.2 in universe, wine dummy package in main, hmm [23:48] bryceh: nice, did they open port 80 or did ya fix it another way? [23:59] Sarvatt, would that work for 64-bit flash on amd64? [23:59] bjsnider: yea [23:59] err [23:59] no [23:59] no acceleration in the 64 bit flash release [23:59] works for 32 bit flash on amd64 thogh [23:59] i don't use that