[09:06] <apw> tseliot, i hear you have bryces mantle this month ... just in time for a nasty intel gpu hang to appear
[09:07] <tseliot> apw: you're right. And I don't seem to be able to reproduce the problem here
[09:07] <tseliot> apw: did you find a way to reproduce the lockups?
[09:07] <apw> damn i had two different machines do it in 5 minutes last night.  the second of wihch i was using to debug the first
[09:07] <apw> i _think_ but am not sure that a suspend/resume in the past may be required to trigger it
[09:08] <tseliot> ok
[09:08] <apw> cirtainly two of the ones i've seen on my GM45E were instantly on resume
[09:08] <apw> i did maybe 5-6 s/r cycles last night and got one
[09:08] <tseliot> ah
[09:08] <apw> it was definatly a gpu hang, sequence number thingy not going up any more
[09:09] <apw> jk mentioned that he had had one and indicated he had noted that he wasn't gettting interrupts any more
[09:09] <apw> i don't know if he meant they were disabled or not ...
[09:09] <tseliot> ok, what bug number are you referring to? (just to be sure)
[09:09]  * apw is looking for the 'how to debug GPU hangs page' to see if there are any preparetory steps to get more debug
[09:10]  * apw gets the numbers
[09:11] <apw> the original was filed by mdz
[09:13] <apw> bug #424055
[09:13] <apw> i have bug #429199 and bug #429191 for my two occurances
[09:14] <tseliot> great, I have a Dell Mini 10v so I should be able to reproduce the 1st bug
[09:14] <apw> mdz mentioned two other people who have seen it
[09:14] <apw> my gme945 is also a mini10v, so you are in the right area there, though that has been the one  i have reprod'd it less often on
[09:14] <apw> mine are both with the -10.32 karmic kernel
[09:15] <tseliot> ok
[09:15] <apw> is there any steps i should take (or a script to folllow) to get the most info from the next reproduce by
[09:17] <apw> tseliot, is this more likely to be kernel or mesa level issue do you think?  should i open kernel tasks on these bugs?
[09:17] <tseliot> apw: you can either try mdz's script or follow amitk's instructions: http://smackerelofopinion.blogspot.com/2009/06/looking-at-intel-x-hangs.html
[09:18] <apw> smackerel of opinion is ckings blog
[09:19] <tseliot> apw: right, wrong person ;)
[09:19]  * apw notes OSD has lost its mind and is using its second slot only ... sigh ... what else can break
[09:19] <tseliot> as regards the bug, I wouldn't know, I have yet to look into this problem
[09:20] <tseliot> this is worse (with nvidia): https://bugs.launchpad.net/ubuntu/+bug/429003
[09:21] <apw> tseliot, to add to your woes, my ATI based T30 with stock karmic on it, does not repaint the screen properly at all
[09:21] <apw> so i'd say thats all three major graphics in the crapper one way or anohter.  though that does at least make the intel look good
[09:22] <tseliot> apw: with and/or without compiz?
[09:22] <apw> with compiz
[09:22] <tseliot> have you filed a bug report about it?
[09:23] <apw> nope not yet ... damn i should do that now ...
[09:23] <apw> i'll get it updated to 'now' and make sure its still there and get it filed too ... 
[09:23]  * tseliot installs the new libc6 for testing. Wish me good luck
[09:24]  * apw has that, so that i think will work :)
[09:25] <apw> tseliot, i note that the gdm login is ok on this machine ... i assume that is not compositied, it goes to poop as i login
[09:26] <tseliot> apw: try disabling compiz and see if you notice the same problem
[09:26] <apw> will do
[09:27] <tseliot> thanks
[09:28]  * tseliot restarts X
[09:40] <apw> tseliot, ok just did 4 s/r cycles on my GM45 and its come back up ok, but i am seeing screen corruption now, and all of my window manager wigets are invisible.  still there but invisible
[09:40] <apw> oh and now, hitting print screen, has taken it out
[09:41] <tseliot> apw: what do you mean by window manager widget? Can you take a screen shot, please?
[09:41] <tseliot> oh
[09:42] <apw> as in the decorations the wm normally adds, the top bar with the title etc ... all invisible
[09:42] <apw> you can click (could) on them and move the windows, but they are 'not there'
[09:42] <tseliot> are you using compiz?
[09:43] <apw> yep using compiz
[09:44] <apw> apw@dm$ cat i915_gem_seqno
[09:44] <apw> Current sequence: 1807468
[09:44] <apw> Waiter sequence:  1807469
[09:44] <apw> IRQ sequence:     1807447
[09:44] <apw> i assume that that is bad?
[09:44] <apw> tseliot, is there anything else i can gather to help understand this one?
[09:45] <tseliot> ok so obviously something wrong happens when suspending/resuming
[09:45] <tseliot> can you add these details to the bug report?
[09:45] <apw> i'd say its a trigger for sure ... i logged out and in, then did 4 s/r and thats the state i ended up in
[09:45] <tseliot> or maybe file a new one
[09:46] <apw> whats your preference, can do either
[09:46] <apw> i have some pictures of the mess just before it locked too
[09:46] <tseliot> maybe file a separate bug so that I can upstream it
[09:47] <apw> ack
[09:47] <tseliot> if they are really the same bug then we'll merge the 2 reports
[09:49] <apw> yep understand its easier to merge than split every time
[10:05] <apw> tseliot, bug #429241
[10:09] <hyperair> is there a known boot-hang with the karmic kernel and ati cards?
[10:09] <hyperair> boot-hang issue i mean
[10:16] <tseliot> apw: thanks
[10:18] <apw> hyperair, not that i am aware of as yet
[10:18] <apw> tseliot, confirmed my ati experience is fine with compiz disabled
[10:19] <apw> hyperair, i assume your hang is before login?
[10:19] <tseliot> hyperair: what do you mean by boot-hang exactly?
[10:20] <hyperair> apw, tseliot: i'm not sure of the specifics, as i've never actually looked at it yet. this is from descriptions a friend gave me (he went and reinstalled 8.10 due to network issues in jaunty and boot-hang issues in karmic)
[10:20] <tseliot> apw: ok, can you file a bug report about it (including this detail), please?
[10:20] <hyperair> apw: it definitely doesn't reach the login screen, at least.
[10:21] <hyperair> apw: from what i've heard, it's like the usplash times out and then falls back onto a black screen with a blinking cursor.
[10:21] <tseliot> hyperair: I would need him to reproduce the problem, reboot, boot in Recovery mode and collect /var/log/Xorg.0.log and /var/log/Xorg.0.log.old
[10:22]  * apw turns into bug file monster
[10:22] <hyperair> i'll see if i can grab hold of him sometime =\
[10:23] <tseliot> heh
[10:32]  * tseliot reboots into jaunty hoping that ubuntu startup disk creator works there
[10:37] <hyperair> it works for me on karmic =\
[10:37] <hyperair> it just takes up loads of i/o and hangs my entire machine for a good 10 minutes or so
[11:07] <apw> tseliot bug #429251
[11:07] <tseliot> apw: thanks again
[11:08] <apw> heh ... a world of pain shoved over the wall and he thanks me :)
[11:09] <tseliot> a world of pain without useful information would be much worse ;)
[11:10] <apw> i shoved a GPU dump into that GM45E hang no idea what that says mind, hopefully someone can read it
[11:17] <_B00> good morning :D
[11:18] <_B00> anyone about?
[11:21] <tseliot> apw: that's definitely useful. I'll try to get a GPU dump on the Dell Mini 10v too. Upstream developers such as jbarnes should be able to read them (unfortunately I'm not)
[11:21] <tseliot> _B00: what's up?
[11:22] <_B00> hi tseliot, just wondering what this is.
[11:22] <_B00> is it updates for the x-server?
[11:22] <_B00> as I'm having trouble getting gfx drivers to work on me TNT2 (snigger)
[11:22] <tseliot> _B00: "this" being what? This channel?
[11:23] <_B00> yea adn the webby, also the selection in the 3rd party software on ubuntu tweak
[11:23] <tseliot> _B00: are you using Nvidia's proprietary driver?
[11:23] <_B00> I've tried loads of variations and settings from various forums
[11:23] <_B00> even the drivers on nvidia web
[11:24] <_B00> I read there was a known bug for my card with this release of ubuntu (jaunty 9.04)
[11:25] <_B00> I think all I have is nv at the moment, after cleaning the drivers out i chose
[11:25] <_B00> which did not work
[11:25] <tseliot> _B00: that might be a problem. Installing all kinds of drivers (especially the ones from Nvidia's website) can cause problems with the packages in Ubuntu
[11:26] <_B00> VGA compatible controller: nVidia Corporation NV5M64 [RIVA TNT2 Model 64/Model 64 Pro] (rev 15)
[11:26] <_B00> I installed a set of drivers one at a time tseliot 
[11:26] <_B00> then purged the installs after no success
[11:27] <_B00> "no nvidia driver present nvidia_something_so
[11:27] <_B00> "
[11:28] <_B00> I went to the directory and there it was. So no idea personally
[11:28] <tseliot> _B00: did you uninstall the one from nvidia's website?
[11:28] <_B00> yea
[11:28] <_B00> i do a whereis nvidia and nothing shows
[11:28] <_B00> also gui search :D no results
[11:30] <tseliot> please try installing the driver (nvidia-glx-71) with either jockey or envyng and, if it doesn't work, please file a bug report on launchpad about nvidia-glx-71
[11:30]  * tseliot restarts X
[11:55] <tseliot> apw: can you add the version of mesa that you're using with the -ati driver to the bug report, please?
[11:56] <apw> tseliot, doesn't that get picked up with the ubuntu-bug ...-ati ... shame on us
[11:57] <tseliot> apw: my bad. This is what happens when I look at different bugs at the same time...
[11:57] <tseliot> ;)
[11:58] <apw> tseliot, is it already in there?
[11:58] <tseliot> apw: yep
[11:58]  * apw finds another corruption bug in ati
[11:58] <tseliot> libgl1-mesa-glx 7.6.0~git20090817.7c422387-0ubuntu3
[11:58] <apw> seems that the OSD gets shows crapola like an old fashioned TV off tune when connecting the network!
[11:58] <apw> without compiz
[11:59] <tseliot> :-/
[11:59]  * apw files more bugs ... boo
[12:00] <apw> now is that an OSD bug, or a grpahics bug
[12:00] <tseliot> so xrender is broken too... nice
[12:01] <apw> only seems to be affecting OSD ...
[12:01] <apw> nothing else i've tried is ill
[12:01] <apw> oh xrender, not everyone uses that i guess
[12:03]  * tseliot nods
[12:03] <tseliot> can you reproduce the problem with -intel (without compiz)?
[12:04]  * apw tried
[12:04] <apw> tries
[12:06] <apw> tseliot, seems ok on this intel GM945E
[12:10] <tseliot> apw: ok, let me see if upstream is aware of this
[12:10] <apw> tseliot, bug #429295
[12:11] <tseliot> apw: great, I'll try to upstream this too
[12:14] <jcristau> tseliot: so you're the new bryce? :)
[12:14] <tseliot> jcristau: only for this month
[12:14] <apw> i think thats a compliment :)
[12:15] <tseliot> apw: it surely is ;)
[12:18]  * apw fills his digital camera with pictures of broken ubuntu :/
[12:20] <apw> i've marked that OSD one for osd as well, so its visible when regular users try and file a duplicate
[12:20] <tseliot> apw: good move. Thanks
[12:21]  * tseliot is waiting on the Dell Mini 10 v to finish its endless dist-upgrade...
[12:22] <apw> heh nasty
[12:22] <apw> tseliot, anything more i can specifically do on any of these?  else i'll wander off and tread the fine line between them
[12:23] <apw> obviously if you think any of them are kernel, do yell at me
[12:24] <tseliot> apw: ok, I'll let you know. Thanks
[12:24] <apw> the intel stuff is likely to be a major pain point till its fixed 
[12:26]  * tseliot (sadly) agrees
[12:26]  * tseliot -> lunch
[12:37] <tjaalton> isn't it the one where downgrading mesa helps?
[13:06] <apw> tjaalton, it ?
[13:07] <apw> i915 hangs, ati corruption, or osd corruption, or something else?
[13:13] <tjaalton> apw: i915 hangs with compiz
[13:13] <apw> tjaalton, is there a recommended downlevel for testing there?
[13:14] <apw> i seem to be able to trigger it pretty quickly if i suspend/resume a lot
[13:14] <tjaalton> apw: the last 7.5.x version should work
[13:14] <tjaalton> before the snapshot
[13:14] <tjaalton> and I see it too
[13:15] <apw> its the suspend/resume contribution to the issue which worries me, is mesa aware of such things?
[13:17] <hyperair> re suspend/resume compiz hanging, is it the one where it doesn't redraw, but continues to accept input?
[13:17] <hyperair> it seems to have disappeared for me recently
[13:17] <hyperair> but before that, i use a hotkey that restarts compiz
[13:17] <apw> i've separatly seen that one today as well
[13:17] <tjaalton> apw: guess the driver state gets broken on susp/resume
[13:17] <apw> where you could shift desktops using the keys but the screen was blank
[13:18] <apw> i think it regressed to death pretty soon after that
[13:18] <apw> bah too many similar issues ... hard to know which one one has
[13:18] <hyperair> heh yeah
[13:18] <hyperair> if you're on KMS, switching to a VT works
[13:19] <hyperair> i'm not sure about without
[15:47] <tseliot> apw: as regards the problem with osd and ati. Can you try adding the following options (one at the time) in the "Device" section of your xorg.conf? Option "EXANoComposite" Option "EXANoUploadToScreen" Option "EXANoDownloadFromScreen"
[15:52] <apw> tseliot, one at time as in 'each on its own'
[15:53] <tseliot> apw: yes, that would be a better explanation of what I meant to say ;)
[15:55] <apw> tseliot, do those take a parameter?
[15:56] <apw> ie are they all 'true' in your thing there
[15:56] <jcristau> no parameter is the same as true
[15:56] <tseliot> right
[16:02] <apw> tseliot, nope none of those help
[16:03] <tseliot> apw: ok, thanks
[16:05] <apw> tseliot, bug updated
[16:06] <tseliot> apw: thanks again
[16:19]  * tseliot is trying hard to make its GPU hang after 5 suspend/resume cycles
[16:40] <tseliot> apw: any suggestions on how to make the GPU lock up on suspend/resume (other than beating the netbook with a stick). It hasn't locked up after 11 cycles
[16:52] <apw> tseliot, not really ... i found it tended to happen on resume that was all
[16:52] <tseliot> ok
[16:52] <apw> i've not see any other definative trigger
[16:52] <tseliot> this GPU seems to be bullet-proof...