| Sarvatt | actually, it would be best if you updated your system first to see if it still happens, a new major mesa upgrade just went in | 00:00 |
|---|---|---|
| Dink | I guess that is a no it is not enabled. Says I am running mutter | 00:00 |
| Dink | maverick netbook edition | 00:00 |
| Sarvatt | ah so you're running UNE, it could be something else.. definitely try updating to get the new mesa and see if it still happens there | 00:01 |
| Dink | what package can I check via apt-cache policy to see if its the latest you are talking about | 00:01 |
| Sarvatt | libgl1-mesa-dri, version 7.9~git20100909 | 00:02 |
| Dink | k | 00:02 |
| Sarvatt | it might not be accepted yet actually | 00:02 |
| Dink | 7.8.2-2ubuntu2 here | 00:03 |
| Dink | So yeah guess it is not. Using the pdx repo which I think is current | 00:03 |
| Sarvatt | your problem is probably that UNE is freezing for you, those errors are pretty much harmless and if it was the server crashing you'd get more info or if it was a GPU hang you wouldn't be able to just restart X like that on that machine. I think there's a pretty high chance that the new mesa will fix that for you when its up | 00:06 |
| Dink | Ohh and another thing.. Sometimes when I click on a window to move it I see the hand to move but it moves another window | 00:06 |
| Dink | Sarvatt, ok will keep a look out for it. | 00:07 |
| Sarvatt | I know that backtrace in your xorg.0.log.old is caused by the copy-fb patch to x-x-video-intel though for sure and it happens on server shutdown pretty much all the time, haven't been able to figure that one out yet :( | 00:07 |
| Dink | Ahh. So is there anytihng I can run from vt1 when all this happens to help figure out what is going on ? | 00:08 |
| Dink | To figure out whatever package is causing it. | 00:09 |
| Sarvatt | maybe check ~/.xsession-errors from VT1 while X is still running to see if anything stands out when it happens? | 00:10 |
| Dink | Yeah lots of stuff there. I just reproduced it | 00:12 |
| Dink | is there a pastebin I can sent it to ? | 00:13 |
| Sarvatt | have pastebinit installed? just pastebinit ~/.xsession-errors if so, would probably be better to save it somewhere and attach it to the bug though | 00:13 |
| Dink | http://pastebin.com/DM1gaNFG | 00:15 |
| Dink | Not pretty huh | 00:18 |
| Sarvatt | pretty standard for unity.. definitely not pretty :) | 00:20 |
| Dink | Well at least I know I can reproduce it on the fly :P | 00:21 |
| Dink | Just want to get the bug to the correct people and with the correct info. | 00:22 |
| Sarvatt | well, i'd check if the mesa update fixes things, if not i'd move the bug over to the unity package or ping me and I'll do it. just responded that those errors are red herrings and the real problem is somewhere higher up in the stack. you might want to look through bugs against unity for people having freezing problems if you want to try to find something similar that might have a workaround in it to get it working for now, I'm about to head out | 00:25 |
| Sarvatt | for the night though so can't dig into it more at the moment | 00:25 |
| Dink | Sarvatt, cool. Thanks for the time and effort you put in. I will look at the unity bugs and see if that mesa package fixes things. Thanks again !!! | 00:26 |
| Sarvatt | updated the title, can you maybe explain what you do to reproduce it a bit more in the bug? is it always when you alt-tab for instance or clicking between windows? | 00:27 |
| Sarvatt | that'd be a big help if so :) | 00:28 |
| Dink | ok will do | 00:28 |
| Sarvatt | guess I need to force myself to start using unity one of these days, been quite happy with just docky + appmenu-gtk + go-home-applet + compiz/metacity + indicator-applet-complete on my netbook.. | 00:40 |
| Dink | Does those mesa packages get renamed ? | 00:42 |
| Dink | or could I try installing them by hand and it should replace existing. | 00:42 |
| Dink | libgl1-mesa-glx_7.9~git20100909-0ubuntu1_i386.deb | 00:42 |
| Dink | I probably shouldn't be using the une on this netbook anyways since I mostly have it plugged into my monitor. Its basically my primary computer. | 00:44 |
| User1 | Hello, the latest libgl1-mesa packages causes white screen after login. | 04:40 |
| User1 | It basically breaks mutter/unity | 04:56 |
| Dink | Sarvatt, the new mesa package actually broke unity/mutter even more. Sorry man did not fix my issue. I reverted back to the previous version. | 05:18 |
| apw | RAOF, poke | 08:52 |
| apw | Sarvatt, about ? | 09:31 |
| gord | hey guys, seems the latest updates have broken all clutter based applications, including unity. we suspect intel drivers | 10:57 |
| njpatel | hi, with latest updates from maverick gl seems to be partially broken on my Intel i965 chip, and mostly broken on i945 | 10:59 |
| gord | njpatel, i just said that here :P | 10:59 |
| njpatel | heh, woops :) | 10:59 |
| sabdfl | hi folks | 13:24 |
| sabdfl | is there a bug # for the white-screen-on-login issue, with the current intel drivers, mesa and unity? | 13:25 |
| tjaalton | sabdfl: some recent gpu lockup issues, but no bugs that mention whitescreen symptoms | 13:34 |
| tjaalton | some have mentioned the issue here though | 13:35 |
| tjaalton | raof is in xds, sarvatt not awake yet and iäm at a conference, so there haven't been people to look into the issue | 13:36 |
| ara | davmor2 filed one against unity: bug 638921 | 13:42 |
| ubot4 | Launchpad bug 638921 in unity (Ubuntu) "Unity displays only a white desktop on 20100915's updates (affects: 1) (dups: 1) (heat: 12)" [Undecided,New] https://launchpad.net/bugs/638921 | 13:42 |
| tjaalton | ok, i only looked at x bugs that have appeared on my mail folders.. | 13:48 |
| tjaalton | the real package in this case would be mesa | 13:48 |
| sabdfl | ok, thanks, the mesa link is useful | 13:54 |
| sabdfl | was the mesa change a big one? easy to bisect for the culprit? | 13:55 |
| tjaalton | huge :) | 13:55 |
| tjaalton | but if it affects only intel, then it's easier to bisect | 13:56 |
| gord | and it affects any clutter application, seems like intel + clutter = fail | 13:56 |
| tjaalton | maybe clutter uses some more advanced features? is compiz affected as well? | 13:58 |
| sabdfl | compiz isn't affected, if i correctly followed some of the chatter | 13:59 |
| gord | compiz works fine, so does glxgears | 14:00 |
| tjaalton | so that narrows it down even more | 14:00 |
| gord | so it seems that clutter is getting BadDrawable X errors when making its glx window | 14:02 |
| Sarvatt | (unity:1405): ClutterGLX-CRITICAL **: Unable to make the stage window 0x480004f the current GLX drawable | 14:10 |
| Sarvatt | Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 | 14:10 |
| tjaalton | the cavalry arrived :) | 14:11 |
| Sarvatt | this is pretty nasty, not having any luck debugging it and am going to have to bisect across the 9 months of changes and submit a bug upstream. the cavalry is at XDS! :) | 15:27 |
| kenvandine | hey Sarvatt | 15:27 |
| kenvandine | Sarvatt, i assume you are aware of https://bugs.edge.launchpad.net/mesa/+bug/638808 | 15:28 |
| kenvandine | ? | 15:28 |
| ubot4 | Launchpad bug 638808 in unity (and 1 other project) "No unity board but White screen after GDM (affects: 2) (heat: 10)" [Undecided,New] | 15:28 |
| Sarvatt | yep | 15:28 |
| kenvandine | was there another bug filed on that already? | 15:28 |
| Sarvatt | https://launchpad.net/bugs/638921 but no info on that one either | 15:28 |
| ubot4 | Launchpad bug 638921 in unity (Ubuntu) "Unity displays only a white desktop on 20100915's updates (affects: 2) (dups: 1) (heat: 16)" [Undecided,New] | 15:28 |
| kenvandine | ok, which one should we dupe ? :) | 15:29 |
| kenvandine | i guess the newer one, i'll do that | 15:29 |
| Sarvatt | Intel FramebufferRenderbuffer 2 1 | 15:31 |
| Sarvatt | Render to A8 texture OK | 15:31 |
| Sarvatt | Render to A8 texture OK | 15:31 |
| Sarvatt | Begin render texture tid b77d3870 tex=32 w=256 h=256 refcount=1 | 15:31 |
| Sarvatt | Render to RGBA8 texture OK | 15:31 |
| Sarvatt | Begin render texture tid b77d3870 tex=41 w=1024 h=600 refcount=1 | 15:31 |
| Sarvatt | Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 | 15:31 |
| Sarvatt | I wish these INTEL_DEBUG env vars were documented somewhere :) | 15:32 |
| kenvandine | Sarvatt, hope you don't mind, i confirmed bug 638808, set milestone and importance and duped the other one | 15:32 |
| kenvandine | :) | 15:32 |
| ubot4 | Launchpad bug 638808 in mesa (Ubuntu) (and 2 other projects) "No unity board but White screen after GDM (affects: 4) (dups: 2) (heat: 24)" [Critical,Confirmed] https://launchpad.net/bugs/638808 | 15:32 |
| vish | kenvandine: oops! i just noticed my one comment regarding the white screen is getting you more Unity bugs.. ;) | 15:34 |
| Sarvatt | intel users are now getting the white screens ATI users have gotten all along, fun! | 15:35 |
| njpatel_ | heh | 15:35 |
| === njpatel_ is now known as njpatel | ||
| kenvandine | Sarvatt, spreading the love i see :) | 15:36 |
| kenvandine | i have been suffering with that on my old r300 wishing i could run unity | 15:37 |
| kenvandine | now my main laptop and test boxes are all hosed (intel) :) | 15:37 |
| njpatel | let's all buy nvidia | 15:39 |
| === ara_ is now known as ara | ||
| Sarvatt | kenvandine: what GPU's are you using that are broken? | 15:45 |
| Sarvatt | anything 965+? | 15:45 |
| kenvandine | GM45 | 15:46 |
| Sarvatt | i only have a 945 and a sandybridge, 3D isn't working on sandybridge yet so can only confirm it on 945 and there are no logs on any of these bug reports to see if its happening on newer ones | 15:46 |
| kenvandine | whatever that is :) | 15:46 |
| Sarvatt | thanks | 15:46 |
| kenvandine | Sarvatt, feel free to use me as a guinea pig... | 15:48 |
| kenvandine | :) | 15:48 |
| njpatel | Sarvatt, , 965 is broken too, fwiw | 15:51 |
| njpatel | oh, I can test too | 15:51 |
| Sarvatt | just wanted to know if it was happening across multiple generations, I can reproduce it on 945 luckily. bisecting this is going to be rough, writing up an upstream bug report at the moment but most everyone is at XDS :( | 15:51 |
| njpatel | GM45 | 15:51 |
| Sarvatt | do you guys get a white unity launching it in a normal desktop session also? | 15:57 |
| Sarvatt | ok the results are totally different using unity -p | 15:57 |
| Sarvatt | I get our friend (unity:2933): Clutk-WARNING **: [CheckGLError] GL_INVALID_OPERATION error in File ./ctk-render-target.c at line: 128 | 15:57 |
| Sarvatt | (unity:2933): Clutk-WARNING **: [CheckGLError] OpenGL Error 1282 in File ./ctk-render-target.c at line: 128 | 15:57 |
| Sarvatt | launching it with -p | 15:57 |
| Sarvatt | it segfaults with just unity | 15:58 |
| Sarvatt | well aborts, not segfault | 15:58 |
| Sarvatt | gotta be something racy there, every few unity -p invocations crashes with Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 but the rest all die with the above | 16:01 |
| Sarvatt | which is incidentally the same thing ati has gotten all along | 16:04 |
| kenvandine | Sarvatt, let me try | 16:15 |
| apw | Sarvatt, Wubbbi is reporting the white screen only in netbook too | 16:27 |
| Wubbbi | yes ... I realy do. You wanna know something? | 16:28 |
| apw | Sarvatt, its been suggested its a mesa issue, have we confirmed that, does someone need to do some testing? | 16:28 |
| kenvandine | (unity:32356): ClutterGLX-CRITICAL **: Unable to make the stage window 0x5e0004f the current GLX drawable | 16:28 |
| Sarvatt | it's in progress, i can reproduce it locally | 16:29 |
| kenvandine | sarvatt: that is what i am getting from unity -p | 16:29 |
| Sarvatt | apw, Wubbbi https://bugs.edge.launchpad.net/mesa/+bug/638808 | 16:29 |
| ubot4 | Launchpad bug 638808 in mesa (Ubuntu) (and 2 other projects) "No unity board but White screen after GDM (affects: 5) (dups: 2) (heat: 32)" [Critical,Confirmed] | 16:29 |
| Sarvatt | kenvandine: what after that? | 16:29 |
| Sarvatt | because that isn't really a problem | 16:29 |
| kenvandine | core dump | 16:29 |
| Sarvatt | kenvandine: try it a few more times? | 16:29 |
| Sarvatt | does it change? | 16:30 |
| kenvandine | nope | 16:30 |
| kenvandine | 4 times, same | 16:30 |
| apw | Wubbbi, sounds like its in hand | 16:30 |
| Sarvatt | using mutter or compiz? | 16:30 |
| kenvandine | compiz | 16:30 |
| kenvandine | and running unity -p | 16:31 |
| Dink | Sarvatt, Hello | 16:31 |
| Dink | Busy morning huh ;) | 16:31 |
| kenvandine | Sarvatt: i was just talking to rickspencer, he said last week RAOF knew about this problem and thought he had a fix | 16:32 |
| Sarvatt | yeah, so much for my hardware enablement work :) | 16:32 |
| Sarvatt | oh? | 16:32 |
| Sarvatt | he's at a conference until next week :( | 16:33 |
| kenvandine | said he planned on having it in for this upload | 16:33 |
| kenvandine | yeah... :/ | 16:33 |
| Wubbbi | Sarvatt: do you need any help? I have time atm! ^^ | 16:33 |
| Sarvatt | i'm bisecting it in mesa now but it's going to take quite some time, mesa doesn't exactly build fast even with a minimal config | 16:34 |
| JanC | Sarvatt: RAOF just has to offer free beer at XDS to everyone who helps fix this ;) | 16:34 |
| Wubbbi | Sarvatt: just take your time. But please do it the most fast way. I cant use me netbook atm xD | 16:35 |
| jcristau | JanC: interesting.. maybe i should get on it then. | 16:35 |
| Sarvatt | Wubbbi: just disable autologin in /etc/gdm/custom.conf and log in a gnome session | 16:35 |
| kenvandine | sarvatt, is rolling back the mesa update an option? | 16:35 |
| Sarvatt | if we want to rebreak KDE | 16:35 |
| kenvandine | sarvatt, well is the problem with kde as sever as this? | 16:36 |
| kenvandine | UNE won't even load... | 16:36 |
| Wubbbi | Sarvatt: I take a look. thx | 16:38 |
| Dink | Sarvatt, or whomever let me know if you need a guinea pig to test the white screen issue on UNE. | 16:38 |
| Sarvatt | want me to prepare a revert to upload? I'm not confident this can be fixed in a reasonable timeframe | 16:41 |
| Sarvatt | I think we should wait for RAOF to pop on though | 16:42 |
| === yofel_ is now known as yofel | ||
| kenvandine | sarvatt: any idea when he lands? | 16:44 |
| bjsnider | Sarvatt, how does one go about bisecting something in git? | 16:46 |
| bjsnider | i hear about this all the time | 16:46 |
| Sarvatt | kenvandine: no I don't, he said he was getting on the plane yesterday though so I imagine he'll be checking in sometime today | 16:47 |
| kenvandine | long flight though... | 16:48 |
| Sarvatt | I'm not sure why he didn't commit a fix if he had one though, maybe it was outside of mesa? | 16:48 |
| nigelb | bjsnider: man git-bisect :) | 16:48 |
| Sarvatt | or google git bisect guide :) | 16:49 |
| nigelb | http://progit.org/book/ch6-5.html#binary_search | 16:50 |
| kenvandine | sarvatt: i don't know he had one... but rick said he would have it fixed before the upload | 16:50 |
| nigelb | bjsnider: this might help ^^ | 16:50 |
| Sarvatt | kenvandine: he pushed a bunch of changes and was ready for the upload, asked me to try to get it uploaded while he was gone | 16:51 |
| Dink | Sarvatt, sorry for throwing this at you while that is going on... The bug I initially reported about the "freezing" I have not been able to reproduce it without an external monitor,keyboard,mouse present. It seems to only happen when those things are connected. | 16:57 |
| Dink | Should I still move it to Unity ? | 16:58 |
| rickspencer3 | Sarvatt, kenvandine can someone give me a quick run down of the situation wrt mesa and unity? | 16:58 |
| kenvandine | Sarvatt, ^^ please | 16:58 |
| kenvandine | i told rickspencer3 everything i know already :) | 16:59 |
| * Sarvatt explodes :) | 17:00 | |
| kenvandine | afaict if we don't have a lead to a fix by now we should just revert | 17:00 |
| kenvandine | we can't afford to have UNE completely hosed | 17:01 |
| kenvandine | that's my $0.02 | 17:01 |
| kenvandine | although i would love to see 7.9 make it in | 17:01 |
| kenvandine | as an ati user | 17:01 |
| Sarvatt | the mesa update broke unity, RAOF apparently had it fixed but I don't see any fix and he's at a conference, I can reproduce it on 945 and am in the middle of bisecting where it started | 17:01 |
| kenvandine | sarvatt: apparently not just unity... anything clutter | 17:02 |
| kenvandine | at least that is what i heard | 17:02 |
| rickspencer3 | Sarvatt should we not revert the change until we know what is going on? | 17:02 |
| Sarvatt | <Sarvatt> want me to prepare a revert to upload? I'm not confident this can be fixed in a reasonable timeframe | 17:03 |
| Sarvatt | <Sarvatt> <Sarvatt> I think we should wait for RAOF to pop on though | 17:03 |
| Sarvatt | incase he has a fix he forgot to commit and he should be on sometime today | 17:03 |
| rickspencer3 | Sarvatt, well, perhaps we should give it a few hours, and then revert if we can't contact him | 17:04 |
| * Sarvatt nods | 17:04 | |
| rickspencer3 | skaet, here's the situation as I understand it: | 17:04 |
| rickspencer3 | 1. several teams need mesa 7.9 (a component of xorg-xserver) | 17:04 |
| rickspencer3 | 2. mesa 7.9 has not been working for Intel and AMD hardware, or nvidia hardware running free drivers | 17:05 |
| rickspencer3 | 3. RAOF may or may not have thought that he fixed the issue | 17:05 |
| Sarvatt | ^ with clutter based apps | 17:05 |
| rickspencer3 | 4. RAOF pushed an upload yesterday, and then left for a conference | 17:05 |
| Sarvatt | it works significantly better otherwise | 17:05 |
| rickspencer3 | 5. Unity is clutter based, and the bug makes clutter display just a white screen | 17:06 |
| rickspencer3 | 6. So anyone running unity without an proprietary nvidia driver gets a white screen | 17:06 |
| Sarvatt | unity has no fallback if there is a problem, and I believe it is auto logging in so people can't just pick a gnome session that will work at the moment | 17:06 |
| rickspencer3 | so, if we revert, Kubutu breaks again, Linaro is screwed, etc... | 17:07 |
| rickspencer3 | but if we don't, well Unity is unusable | 17:07 |
| rickspencer3 | so we have to decide ... | 17:07 |
| rickspencer3 | keep Unity broken until we fix it, or ... | 17:07 |
| rickspencer3 | roll back now and leave Kubuntu and others broken | 17:08 |
| Sarvatt | also no hope of the new intel GPU's coming out having working 3D in 10.10 | 17:08 |
| kenvandine | rock meet hard place | 17:08 |
| skaet | urk. not nice at all. | 17:08 |
| rickspencer3 | so I am proposing that we sit tight until we can get RAOF engaged | 17:08 |
| skaet | thanks for the summary, it helps put the bugs I was seeing in context. | 17:08 |
| rickspencer3 | I suspect we are missing some critical information | 17:08 |
| rickspencer3 | let me check with the Unity team and see if they are blocked today | 17:09 |
| kenvandine | i fear he has no fix... :/ | 17:09 |
| rickspencer3 | because they have lots of bugs I want fixed | 17:09 |
| rickspencer3 | kenvandine, I can't imagine that he would have pushed that change without a fix | 17:09 |
| kenvandine | rickspencer3, kind of... they need to manually revert | 17:09 |
| Sarvatt | I haven't had any luck debugging it but am in the middle of bisecting, it could end up being something simple but its 9 months of developement to bisect through | 17:09 |
| Sarvatt | ironic that this upload fixed unity on r600+, also 3D works on ati HD5000 series cards | 17:10 |
| rickspencer3 | Sarvatt the last time I talked to RAOF, he ws working on it | 17:10 |
| rickspencer3 | so he must have some data, and I can't imagine that he would have intentionally busted Unity like this | 17:11 |
| skaet | rickspencer3, several unity bugs were closed between Friday and yesterday, which ones are you concerned about? | 17:11 |
| rickspencer3 | skaet, there is are a slew of issues, no way I could list them out | 17:12 |
| gord | FYI, this isn't a unity problem, its a clutter problem | 17:13 |
| rickspencer3 | but njpatel and his team are very prductive | 17:13 |
| gord | no clutter application works on the latest mesa | 17:13 |
| rickspencer3 | gord, right, but it impacts the productivity of the unity team | 17:13 |
| rickspencer3 | or it might be | 17:13 |
| gord | rickspencer3, that would be me ;) | 17:13 |
| rickspencer3 | heh | 17:13 |
| rickspencer3 | okay, I was just asking elsewhere | 17:13 |
| Sarvatt | i've tried the latest 1.2 branch clutter git and it's no different there at least | 17:13 |
| rickspencer3 | gord, are you currently blocked? | 17:13 |
| gord | rickspencer3, no, but i have nvidia, anyone on intel graphics will be | 17:13 |
| rickspencer3 | Sarvatt I am certain that RAOF did all this analysis already | 17:13 |
| Sarvatt | other clutter apps not working is interesting because I didn't experience the same thing when I'm on xorg-edgers with the same mesa version, worth checking if xorg-edgers is broken in the same way | 17:18 |
| bryceh | heya, mesa troubles? | 17:19 |
| Sarvatt | because that'd put the problem outside of mesa | 17:19 |
| Sarvatt | heyo bryceh, yeah its nuts today | 17:19 |
| rickspencer3 | bryceh, | 17:19 |
| rickspencer3 | <rickspencer3> 1. several teams need mesa 7.9 (a component of xorg-xserver) | 17:19 |
| rickspencer3 | 2. mesa 7.9 has not been working for Intel and AMD hardware, or nvidia hardware running free drivers | 17:19 |
| rickspencer3 | 3. RAOF may or may not have thought that he fixed the issue | 17:19 |
| rickspencer3 | <Sarvatt> ^ with clutter based apps | 17:19 |
| rickspencer3 | <rickspencer3> 4. RAOF pushed an upload yesterday, and then left for a conference | 17:19 |
| rickspencer3 | <Sarvatt> it works significantly better otherwise | 17:19 |
| rickspencer3 | <rickspencer3> 5. Unity is clutter based, and the bug makes clutter display just a white screen | 17:19 |
| rickspencer3 | 6. So anyone running unity without an proprietary nvidia driver gets a white screen | 17:19 |
| rickspencer3 | <Sarvatt> unity has no fallback if there is a problem, and I believe it is auto logging in so people can't just pick a gnome session that will work at the moment | 17:19 |
| rickspencer3 | <rickspencer3> so, if we revert, Kubutu breaks again, Linaro is screwed, etc... | 17:19 |
| rickspencer3 | but if we don't, well Unity is unusable | 17:19 |
| rickspencer3 | so we have to decide ... | 17:19 |
| rickspencer3 | keep Unity broken until we fix it, or ... | 17:19 |
| rickspencer3 | roll back now and leave Kubuntu and others broken | 17:19 |
| rickspencer3 | bryceh, did RAOF discuss any of this with you? | 17:19 |
| njpatel | mutter without the unity plugin (mutter --replace --mutter-plugins="") is broken too | 17:19 |
| rickspencer3 | njpatel, yeah, RAOF knew about this before | 17:20 |
| Sarvatt | njpatel: white screen or just freezes? | 17:20 |
| rickspencer3 | so I don't understand why it was uploaded without a fix in place | 17:20 |
| Sarvatt | njpatel: if its freezes does CLUTTER_VBLANK=none fix it for you too? | 17:20 |
| rickspencer3 | Sarvatt I think our current focus should be on figuring out what happened with the uploads rather than trying to diagnose the problem right now | 17:21 |
| rickspencer3 | I don't think we'll get until we know if RAOF thinks it was fixed | 17:22 |
| kenvandine | if edgers works... we can try to bisect the difference there | 17:22 |
| njpatel | Sarvatt, white screen | 17:23 |
| Sarvatt | well we're stalled waiting to contact RAOF in that case, if it turns out he doesn't have a fix the bisect will help :) | 17:23 |
| rickspencer3 | good idea, see if there a patch in there | 17:23 |
| Sarvatt | njpatel: hmm, CLUTTER_VBLANK=none mutter --replace works fine here | 17:23 |
| njpatel | Sarvatt, both mutter with unity and mutter without unity, the process is still running and handling requests | 17:23 |
| Sarvatt | on 945 | 17:23 |
| rickspencer3 | Sarvatt fair enough, but I'm certain that you are on well trod ground | 17:23 |
| rickspencer3 | the last time I talked to RAOF about this, he knew about the problem and was working on a fix | 17:23 |
| njpatel | Sarvatt, let me try cluttervblank=none, hold up | 17:23 |
| rickspencer3 | Sarvatt what exactly did he ask you to upload? | 17:24 |
| njpatel | Sarvatt, woohoo, CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter & works! | 17:25 |
| rickspencer3 | woah | 17:25 |
| rickspencer3 | back in business! | 17:25 |
| njpatel | Sarvatt, it wasn't working before I restarted :/ | 17:25 |
| njpatel | yessir, awesome, I don't need to setup my desktop now | 17:25 |
| njpatel | i mean, vblanking would be very much appreciated, but at least we're not blocked now | 17:25 |
| Sarvatt | rickspencer3: we both have been working on it in debian git for the FFe, he merged in his work a few days ago and once the FFe was granted he asked if I would upload what was there since he would be gone | 17:26 |
| njpatel | that crash wasn't related, but it works, so thanks | 17:27 |
| rickspencer3 | Sarvatt so do you think RAOF believed that it was fixed? | 17:27 |
| bryceh | sorry, was in middle of another meeting, but now over | 17:27 |
| rickspencer3 | bryceh, understood | 17:27 |
| bryceh | rickspencer3, no I hadn't discussed mesa with raof previously | 17:27 |
| rickspencer3 | dang it | 17:27 |
| rickspencer3 | man, I bungled this badly :( | 17:27 |
| Sarvatt | njpatel: whoa, you're right! | 17:28 |
| Sarvatt | CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter works here | 17:28 |
| rickspencer3 | njpatel, so I take it we can't ship without vblanking in Unity? | 17:28 |
| njpatel | rickspencer3, it would give artefacts and bad rendering of videos etc, it's not ideal. I can speak to the team about the consequences tomorrow to be certain, though | 17:29 |
| Sarvatt | now why does that work but manually invoking unity not? do clutter env variables not pass through to unity? | 17:29 |
| rickspencer3 | njpatel, hmmm | 17:30 |
| rickspencer3 | can we apply it as a work around until we get this mesa thing properly sorted? | 17:30 |
| bryceh | rickspencer3, ok so how can I help? need me to upload some bits, or revert mesa, or do diagnostics or...? | 17:30 |
| rickspencer3 | bryceh, well, I was hoping that you could sniff around and see what RAOF was up to | 17:32 |
| rickspencer3 | I suspect that we are pissing a piece of the puzzle | 17:32 |
| bryceh | hmm, artifacts and bad video rendering sounds less severe than white screens | 17:32 |
| bryceh | rickspencer3, ok | 17:32 |
| bryceh | rickspencer3, fwiw he's likely on a plane at the moment, maybe in a day or two he'll be online at the conference | 17:32 |
| rickspencer3 | bryceh, if you have some bandwidth to help Sarvatt find the root cause, that would be really useful as well | 17:32 |
| rickspencer3 | *sigh* | 17:33 |
| bryceh | although X conferences notoriously have had bad wireless so who knows | 17:33 |
| bryceh | ok | 17:33 |
| rickspencer3 | I can call the airline and ask them to turn the plane around | 17:33 |
| kenvandine | hehe | 17:33 |
| bryceh | rickspencer3, can you drop a quick note to deryck.hodge@canonical.com ? | 17:33 |
| Sarvatt | I'm almost positive he marked down that he was arriving on wednesday france time, let me check the XDS wiki | 17:34 |
| Sarvatt | not that he'll jump on IRC first thing or anything :) | 17:34 |
| vish | bug with : 00:02.0 VGA compatible controller [0300]: Intel Corporation N10 Family Integrated Graphics Controller [8086:a011] (prog-if 00 [VGA controller]) | 17:34 |
| vish | should be tagged in the title as? | 17:34 |
| rickspencer3 | Sarvatt well, his cell phone did not pick up, and it's middle of the night, so I can't call his house | 17:35 |
| njpatel | rickspencer3, Sarvatt you can add it to /etc/environment to get it picked up | 17:35 |
| rickspencer3 | kenvandine, can you do something like njpatel suggested to get Unity working for the next few days until this is sorted? | 17:36 |
| bryceh | sarvatt, are there bug #'s for the issue(s)? | 17:36 |
| Sarvatt | yeah he didn't mark dates on the wiki - http://www.x.org/wiki/Events/XDS2010/Attendees | 17:36 |
| Sarvatt | bryceh: https://bugs.edge.launchpad.net/mesa/+bug/638808 | 17:37 |
| ubot4 | Launchpad bug 638808 in mesa (Ubuntu) (and 2 other projects) "No unity board but White screen after GDM (affects: 5) (dups: 2) (heat: 32)" [Critical,Confirmed] | 17:37 |
| kenvandine | yeah | 17:37 |
| Sarvatt | bryceh: if you have an intel on maverick its easier to test in a gnome session by just launching unity | 17:37 |
| Sarvatt | no quick access to a terminal in unity?! | 17:40 |
| bryceh | njpatel, from the irc backlog it sounds like you have a workaround now by turning off vblanking... how practical is that as a global workaround in general until this gets fixed? Can you control that param in the packaging sufficiently? | 17:44 |
| Sarvatt | it's been done many times in the past to fix UNE | 17:45 |
| kenvandine | njpatel, i should be able to set that in /etc/X11/Xsession.d/ right? | 17:45 |
| Sarvatt | it's always the first thing I check when something clutter related is broken and has fixed it all but one time (that memory leak) | 17:46 |
| kenvandine | bryceh, ^^ | 17:46 |
| bryceh | kenvandine, bet you're right, can you easily give it a test? | 17:47 |
| Wubbbi | Sarvatt: How is your status atm? Can we get a fix today? Just a question | 17:47 |
| Wubbbi | ^^ | 17:48 |
| bryceh | (I'm not conversant in how mutter is launched) | 17:48 |
| kenvandine | bryceh, just did, that worked | 17:48 |
| kenvandine | i'll stuff that in the package for now | 17:48 |
| jcristau | would seem easier to change an if in clutter than adding a random conffile | 17:49 |
| bryceh | jcristau, except for testing purposes, something folks can fiddle in a conf file might be more convenient | 17:52 |
| bryceh | but whatever works | 17:53 |
| kenvandine | i'd rather the conf file... much easier for people to test fixes... since i would really love to see a proper fix :) | 17:54 |
| bryceh | https://bugs.edge.launchpad.net/ubuntu/+source/mesa/+bug/616997 looks like a dupe of this issue, yes? | 17:55 |
| ubot4 | Launchpad bug 616997 in unity (Ubuntu) (and 3 other projects) "[RV515] Unity keeps reloading with a white background (affects: 8) (dups: 2) (heat: 157)" [Low,Triaged] | 17:55 |
| bryceh | although ati not intel, maybe similar but different issue | 17:56 |
| Sarvatt | vish: can you try CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter in a gnome session on your rv515? | 17:58 |
| vish | Sarvatt: on the ati.. it dint work for unity , but VBLANK worked for mutter.. comment 63 | 17:59 |
| vish | i'll try again , from today's daily if you want .. :) | 17:59 |
| Sarvatt | vish: specifically the command I said though, launching unity doesn't seem to honor the clutter env variables | 18:00 |
| vish | ah.. cool.. just a min then.. | 18:01 |
| Sarvatt | sweet, thanks! | 18:01 |
| Sarvatt | if that "fixes" it on ATI too.. ugh | 18:01 |
| Sarvatt | bryceh: http://paste.ubuntu.com/494295/ | 18:05 |
| Sarvatt | oddly I get one of two different crashes starting it up, that implementation error happens less often than the clutk one | 18:09 |
| bryceh | (unity:10067): ClutterGLX-CRITICAL **: Unable to make the stage window 0x460004f the current GLX drawable | 18:09 |
| bryceh | is that the relevant error? | 18:09 |
| Sarvatt | nope | 18:09 |
| Sarvatt | Begin render texture tid b784f890 tex=43 w=1022 h=548 refcount=1 | 18:10 |
| Sarvatt | Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 | 18:10 |
| Sarvatt | Please report at bugzilla.freedesktop.org | 18:10 |
| bryceh | ok | 18:10 |
| Sarvatt | i've gotten ClutterGLX-CRITICAL **: Unable to make the stage window 0x460004f the current GLX drawable at the startup of every clutter app for over a year now.. | 18:11 |
| bryceh | interesting, well on a side note if that's just an innocuous warning, maybe it doesn't deserve being labeled CRITICIAL ;-) | 18:13 |
| Sarvatt | quadrapassel dies in the same place in the backtrace | 18:13 |
| Sarvatt | #4 0x007a7aed in DRI2GetMSC (dpy=0x896c320, drawable=75497477, ust=0xbfe6ab28, msc=0xbfe6ab20, sbc=0xbfe6ab18) at dri2.c:573 | 18:13 |
| Sarvatt | #5 0x007a6db7 in dri2DrawableGetMSC (psc=0x89779e8, pdraw=0x89666e0, ust=0xbfe6ab88, msc=0xbfe6ab80, sbc=0xbfe6ab78) at dri2_glx.c:301 | 18:13 |
| Sarvatt | #6 0x0077ffbc in __glXGetVideoSyncSGI (count=0xbfe6ac3c) at glxcmds.c:1733 | 18:13 |
| Sarvatt | #7 0x00cacc15 in _clutter_stage_glx_redraw (stage_glx=0x8950090, stage=0x8cc41d0) at ../../../../clutter/glx/clutter-stage-glx.c:571 | 18:13 |
| Sarvatt | #8 0x00cabdcc in clutter_backend_glx_redraw (backend=0x890c070, stage=0x8cc41d0) at ../../../../clutter/glx/clutter-backend-glx.c:777 | 18:13 |
| bryceh | sarvatt, you had mentioned you'd narrowed it down a bit via bisecting? what is the most recent version that does not have the issue? | 18:14 |
| Sarvatt | it depends on what the issue is, I was looking for the bad renderbuffer format error but the first step halfway between 7.8.2 and what we have doesn't have that but still crashes the same way | 18:15 |
| kenvandine | if it works in edgers, can we just compare the ppa to maverick? | 18:16 |
| Sarvatt | haven't tried edgers yet, been stuck in gdb, will check now | 18:16 |
| bryceh | ok, so 7.8.2 is the most recent known-good? would that be 7.8.2-2ubuntu2 ? | 18:16 |
| bryceh | also, I'm not seeing the 7.9 bits in the ubuntu mesa git branch, is it committed somewhere else? | 18:17 |
| Sarvatt_ | ubuntu-maverick branch, sorry | 18:17 |
| bryceh | aha | 18:17 |
| vish_ | Sarvatt_: no go with that command: http://paste.ubuntu.com/494303/ and xsession errors: http://paste.ubuntu.com/494300/ | 18:18 |
| vish_ | Sarvatt_: but mutter works with VBLANK=none | 18:18 |
| vish_ | need any debugging? | 18:19 |
| Sarvatt_ | vish_: can you try a few more times? | 18:19 |
| vish_ | ok.. :) | 18:19 |
| Sarvatt_ | thats the *other* error I get | 18:19 |
| Sarvatt_ | though I haven't had that with CLUTTER_VBLANK=none, I get that clutk error about 90% of the time without it and the mesa implementation error the other times | 18:20 |
| Sarvatt_ | edgers is broken too | 18:20 |
| bryceh | mmm | 18:20 |
| vish_ | oh wow! if i keep trying that unity gets angry ;p | 18:21 |
| vish_ | it crashes the terminal itself ;p | 18:21 |
| Sarvatt_ | hah | 18:21 |
| Sarvatt_ | I think that meant it loaded | 18:21 |
| Sarvatt_ | loading mutter closes gnome-terminal here | 18:21 |
| Sarvatt_ | i do it from a VT with DISPLAY=:0 | 18:21 |
| kenvandine | vish_, was the terminal maximized? | 18:22 |
| kenvandine | there is a bug related to windows maximizing :) | 18:23 |
| kenvandine | makes this even more fun | 18:23 |
| vish_ | kenvandine: nope.. | 18:23 |
| vish_ | yeah not that one.. :) | 18:23 |
| vish_ | Sarvatt_: ok , i tried to add an exec file with that command and launched it, unity does not load .. so the terminal shutting down seems to not be the issue? | 18:24 |
| vish_ | i just have no WM when i try like that | 18:25 |
| bryceh | Sarvatt_, you tested 7.9.0+git20100913.e7eff0cf-0ubuntu0sarvatt ? Noting it in the bug | 18:25 |
| Sarvatt_ | yeah20100911 3ad3cbfb | 18:26 |
| Sarvatt_ | err minus the yeah part :) | 18:26 |
| vish_ | Sarvatt_: from vt , its : DISPLAY=:0 CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter | 18:28 |
| vish_ | or the other way around? | 18:28 |
| Sarvatt_ | 0913 failed to build and i didn't notice it | 18:28 |
| Sarvatt_ | vish_: yeah | 18:28 |
| vish_ | Sarvatt: so i did a $ DISPLAY=:0 CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter 2>&1 | tee unity.log , as soon as i get back to the session no WM > http://paste.ubuntu.com/494310/ | 18:36 |
| vish_ | and no unity either ;) | 18:36 |
| * vish_ waits for further instructions.. :) | 18:41 | |
| kenvandine | vish_, that is the error i get on my old ati laptop... been getting it for ages | 18:42 |
| kenvandine | haven't tried today | 18:42 |
| vish_ | kenvandine: yeah.. unity doesnt work yet on this ati.. but i noticed this white screen when i was testing Sarvatt's mesa ppa iteslf | 18:43 |
| Sarvatt | launch unity 5 times get 4 different errors, ugh! | 18:43 |
| bryceh | o_O | 18:44 |
| * vish_ still needs to update AOA | 18:44 | |
| Sarvatt | mesa implementation error is a red herring for whatever the real problem is though, says that when it works fine too | 18:46 |
| Sarvatt | yet another new one that launch - Bug in window manager: Unexpected X error: BadWindow (invalid Window parameter) serial 2154 error_code 3 request_code 148 minor_code 1) | 18:47 |
| bryceh | ok, this maverick upgrade has 35 min to go, I'm going to go grab food and stuff and bbiab | 18:59 |
| vish | ok , Unity is broken in AOA too! :) | 19:06 |
| Sarvatt_ | vish: thats the problem we're looking at :) | 19:06 |
| vish | cool! glad that i dint update :) | 19:07 |
| Sarvatt | should be worked around by https://lists.ubuntu.com/archives/maverick-changes/2010-September/007682.html | 19:08 |
| vish | yup , CLUTTER_VBLANK=none mutter --replace --mutter-plugins=libunity-mutter | 19:13 |
| vish | works , | 19:13 |
| vish | but not CLUTTER_VBLANK=none unity or CLUTTER_VBLANK=none unity -p | 19:13 |
| Sarvatt | yeah that threw me off too that CLUTTER_VBLANK is dropped launching unity directly | 19:13 |
| vish | sabdfl: and in AOA too , mutter does not work with mesa 7.9 , it needs to be run as CLUTTER_VBLANK=none mutter --replace only | 19:14 |
| vish | oops, Sarvatt ^ | 19:14 |
| vish | ;) | 19:14 |
| sabdfl | vish: i'll test that now, if I drop off IRC, you know why :-) | 19:15 |
| Sarvatt | vish: that'll be in https://edge.launchpad.net/ubuntu/maverick/+source/ubuntu-netbook-default-settings/0.8.6 :) | 19:15 |
| sabdfl | phwoar! | 19:16 |
| sabdfl | worked! | 19:16 |
| vish | sabdfl: hehe, actually it was a tab fail.. dint want to ping you again with an unping.. ;) but glad it works for you there :) | 19:17 |
| sabdfl | modulo some glitches :-) | 19:19 |
| Wubbbi | Sarvatt: Thank you ;D only 32 minutes left until I can use the Netbook-Desktop again :D | 19:20 |
| * Dink awaits for the new package as well | 19:21 | |
| Dink | its in 0.8.6 right ? | 19:23 |
| Sarvatt | yeah | 19:23 |
| Wubbbi | yes | 19:23 |
| Dink | got happy for a sec when I saw it in my update list... checked policy and it was 0.8.5 | 19:24 |
| Wubbbi | Sarvatt: So this is just a Workaround. The real Bug is still there? | 19:24 |
| Sarvatt | yep, indeed it's just a workaround | 19:24 |
| Wubbbi | k | 19:24 |
| vish | just the unity part gets fixed , the mesa task can still be open i guess.. | 19:25 |
| Wubbbi | When you deactivate Clutter, does this have any funcionality we lose with it? | 19:25 |
| Sarvatt | you can echo "export CLUTTER_VBLANK=none" | sudo tee /etc/X11/Xsession.d/80clutter_vblank_none if you dont want to wait for the package | 19:27 |
| Sarvatt | it doesn't deactivate clutter, it deactivates clutter's perpetually broken vsync implementation | 19:27 |
| Dink | Sarvatt, I added a note to my "freezing" bug. It only happens when I have external mouse,keyboard,monitor connected. It works fine when using just the netbook. | 19:27 |
| Sarvatt | oh? thats interesting | 19:28 |
| Dink | yeah very | 19:28 |
| Wubbbi | Sarvatt: Aha .. k ... hehe sounds funny ... broken vsync. Seems like we have to trow a book to the devs of clutter xD | 19:28 |
| Dink | Wonder if its related to dual monitor support something that sorts | 19:29 |
| Dink | lol, love it when I click on an opened window and another window resizes itself automatically | 19:30 |
| vish | heh , time to build now is 1hr ;) | 19:33 |
| Dink | Where you get that info ? | 19:34 |
| JanC | on LP probably | 19:35 |
| Wubbbi | vish: ufff ... I wanna build my own. Where can I DL the source code? | 19:35 |
| Dink | ahh I see it now | 19:35 |
| Dink | https://edge.launchpad.net/ubuntu/+source/ubuntu-netbook-default-settings/0.8.6 | 19:35 |
| vish | Wubbbi: ^ or simpler if you use Sarvat_t's echo.. | 19:36 |
| Wubbbi | Is it possible to build Ubuntu Packages on Windows ( I dont mean Vbox with it ) ? ... just a question | 19:36 |
| Wubbbi | vish: Well ... But when I build my own ( taking less then 1 minuten because it is that small ) its like the same. ^^ | 19:37 |
| Wubbbi | ubuntu-netbook-default-settings_0.8.6.tar.gz .. SC? | 19:38 |
| Dink | Wubbbi, yeah | 19:40 |
| Dink | should contain the debian dir and everything else in there | 19:40 |
| Wubbbi | k | 19:40 |
| bryceh | Sarvatt, have you confirmed that downgrading mesa does indeed make the regression go away? | 19:47 |
| Sarvatt | I don't use unity and was taking the fact it worked before and not now at face value, but I just tested 7.8.2-2ubuntu2 and it works | 19:52 |
| kenvandine | bryceh, i haven't downgraded... but i can confirm that unity worked fine... i did an upgrade that brought in mesa an mono... nothing else | 19:52 |
| kenvandine | and boom | 19:52 |
| kenvandine | in fact i think mono was earlier in the evening... this update was just the mesa packages | 19:53 |
| bryceh | kenvandine, do you happen to have an earlier kernel you could try booting? | 19:53 |
| Sarvatt | it's fun hitting other unrelated bugs every bisect step :( | 19:53 |
| kenvandine | bryceh, sure... | 19:53 |
| kenvandine | although i am pretty certain i had reboot before the mesa upgrade | 19:54 |
| kenvandine | brb | 19:54 |
| bryceh | we've seen a couple times in the past with last minute mesa bugs like this where the actual root bug was not an error in mesa but rather something in the kernel which the newer mesa simply exposed better | 19:54 |
| kenvandine | is it esc to get the grub menu? | 19:54 |
| kenvandine | i always forget :) | 19:54 |
| kenvandine | or tab? | 19:55 |
| Dink | shift I think it is now | 19:55 |
| kenvandine | ok, brb | 19:55 |
| Dink | same here,brb going to try latest mesa and netbook-default packages on my aod250. | 19:57 |
| Wubbbi | Sarvatt: I have done building Ubuntu-Netbook-default-settings. Is it normal that I have 2 .deb files now? One calls "Ubuntu-netbook-unity-default-settings_0.8.6_all.deb" and the other calls "Ubuntu-netbook-default-settings_0.8.6_all.deb". What to install now? both? | 19:57 |
| Dink | Wubbbi, I think its just default, well it was for me... you can check to see what is currently installed | 19:57 |
| Dink | I don't think you can hve both | 19:58 |
| Wubbbi | k ... so the one without "unity"?! | 19:58 |
| Dink | It was for me..... | 19:59 |
| Dink | Going to reboot as soon as my updates are done to test it all | 19:59 |
| Dink | speaking of which, wish me luck... bbiab | 19:59 |
| kenvandine | bryceh, confirmed... previous 2 kernels busted the same way | 19:59 |
| Wubbbi | I reboot to now to see if it works ;D | 19:59 |
| bryceh | kenvandine, ok thanks | 19:59 |
| kenvandine | np | 19:59 |
| bryceh | kenvandine, one other thing - can you check if there is a backtrace at the end of your Xorg.0.log? | 20:00 |
| Sarvatt_ | darnit, unity crashes the server on every bisect step so far, intel invalidate stuff | 20:00 |
| kenvandine | bryceh, sure | 20:00 |
| bryceh | I notice the log attached to the bug report has a trace, but not clear if that corresponds directly to the issue | 20:00 |
| bryceh | Sarvatt_, do you get backtraces? | 20:00 |
| Sarvatt_ | does it mention FreeAllClientResources? if so its not | 20:00 |
| bryceh | yep | 20:00 |
| bryceh | https://bugs.edge.launchpad.net/mesa/+bug/638808/comments/31 | 20:00 |
| ubot4 | Launchpad bug 638808 in ubuntu-netbook-default-settings (Ubuntu) (and 3 other projects) "No unity board but White screen after GDM (affects: 5) (dups: 2) (heat: 32)" [Low,Fix committed] | 20:00 |
| Sarvatt_ | thats the copy-fb patch in x-x-v-intel causing that on server shutdown | 20:01 |
| bryceh | oh | 20:01 |
| Sarvatt_ | yep thats the one | 20:01 |
| bryceh | well that sounds crufty | 20:01 |
| kenvandine | bryceh, nope | 20:01 |
| bryceh | darn | 20:01 |
| bryceh | backtraces would make this so much easier :-) | 20:01 |
| kenvandine | bryceh, my current session worked... but Xorg.0.log.old doesn't either | 20:02 |
| Wubbbi | Sarvatt_: Good News: It works now! Thank you. ... Bad News: Unity is still extrem slow on animating things. Is it possible to turn it off? | 20:02 |
| kenvandine | that would have been the previous attempt right? | 20:02 |
| bryceh | yep | 20:02 |
| Sarvatt_ | the .old was probably from just shutting down normally | 20:02 |
| kenvandine | let me make it fail again and look :) | 20:03 |
| Sarvatt_ | Wubbbi: I have no idea, best ask someone who knows something about unity :) | 20:03 |
| Dink | Woot, everything came up fine | 20:03 |
| Dink | no white screen of death | 20:03 |
| Sarvatt_ | if anything CLUTTER_VBLANK=none should have sped it up a good deal | 20:04 |
| Dink | going to test to see if the "freeze" still occurs | 20:04 |
| Wubbbi | Dink: on me too ... only unity sucks a bit ... well its usable ;) | 20:04 |
| kenvandine | bryceh, ok... confirmed for sure... no trace | 20:05 |
| Wubbbi | This fu***** plymouthd is crashing every time on my PC ... I hate it -.- | 20:05 |
| Dink | nice so far no "freeze" | 20:05 |
| kenvandine | Dink, unity is quite speeding on my intel GM45 | 20:06 |
| kenvandine | s/speeding/speedy | 20:06 |
| Dink | dang spoke too soon, the "freeze" still happens | 20:06 |
| Wubbbi | kenvandine: Ouhh ... Try to use it on a netbook with Intel GMA ... good night! I even dont knoiw why I bought this Intel shit. ... this GMA is the most horrible thing ever! The worst! | 20:07 |
| kenvandine | it works great on my old intel classmate | 20:07 |
| Wubbbi | I was young, stupid and had the money xD | 20:08 |
| slangasek | Wubbbi: you probably mean that plymouth is being killed every time, fwiw; there's a bug in ureadahead that triggers the kernel's OOM killer, not much plymouthd can do to help that | 20:08 |
| kenvandine | i thnk it's a 945 | 20:08 |
| Dink | Yeah works fine other than this weird freeze issue I am havin gon 945GME | 20:08 |
| bryceh | Wubbbi, gma500? | 20:08 |
| Wubbbi | slangasek: ohhh k ... thx 4 info | 20:08 |
| slangasek | (the same problem would have affected usplash back when we were using that, assuming we could even keep usplash working) | 20:08 |
| Wubbbi | bryceh: let me take a look fast | 20:08 |
| Sarvatt_ | he wouldn't be able to use unity if it was, probably is 945 | 20:09 |
| Wubbbi | Jep ... its 945 ... its like a modem! xD | 20:09 |
| Dink | But yeah it does seem faster now on my 945 | 20:10 |
| Wubbbi | I cant even whatch Youtube Videos in 360p without stucks! -.- | 20:10 |
| bryceh | 945/965 typically have been the best supported for intel | 20:10 |
| bryceh | not to say they've not been without tons and tons of bugs ;-) | 20:11 |
| Dink | Wubbbi, I haven't had any issues until recently with my weird freezing bug | 20:11 |
| kenvandine | my 945 on the classmate is great | 20:11 |
| Wubbbi | "for intel" seems not to be such a high categorie xD | 20:11 |
| Sarvatt_ | Dink: if it only happens with an external monitor/keyboard plugged in I doubt it's unity where the problem is, dunno if you moved that bug over or not but I wouldn't now :) | 20:11 |
| Dink | Sarvatt_, doh I moved it... Should I move it back ? | 20:11 |
| Sarvatt_ | yeah | 20:11 |
| Dink | k | 20:11 |
| slangasek | Sarvatt_: so did this mesa bug affect unity on all systems, or is it hardware-specific? I'm wondering how this got missed in the FFe vetting | 20:11 |
| Sarvatt_ | just intel, but non intel had problems with it before already | 20:12 |
| slangasek | /all/ intel, or specific intel cards? | 20:12 |
| slangasek | because the FFe did say new mesa had been tested on intel hardware | 20:12 |
| Sarvatt_ | all intel apparently | 20:12 |
| slangasek | ok | 20:12 |
| Dink | Sarvatt_, mesa or xorg-intel ? | 20:12 |
| slangasek | I'll go beat up the folks who were asking for this FFe then :) | 20:13 |
| * Sarvatt_ hides from slangasek! | 20:13 | |
| Sarvatt_ | we're waiting to hear back from RAOF who is at a conference, thinking he might have missed a fix in it because he mentioned to rickspencer3 specifically that he had a fix for this issue | 20:13 |
| bryceh | Sarvatt_, did he say he definitely had a fix? | 20:15 |
| bryceh | Sarvatt_, I'm poking around but not locating any evidence that there was a patch on hand... and if there was wouldn't he have included it? | 20:15 |
| Wubbbi | I have a Intel 945GME ... is it the same as you Dink | 20:15 |
| Wubbbi | ? | 20:15 |
| Dink | Wubbbi, yeah | 20:15 |
| rickspencer3 | he did not say that he had a fix | 20:15 |
| rickspencer3 | he said he was working on it, and that's the last I heard | 20:16 |
| rickspencer3 | Sarvatt_, bryceh ^ | 20:16 |
| Dink | I have it connected to external monitor,keyboard,mouse | 20:16 |
| Wubbbi | Dink: k ... and do you can see Videos well? ... I only see pictures trying to come after and after xD | 20:16 |
| bryceh | rickspencer3, ok | 20:16 |
| rickspencer3 | bryceh, Sarvatt_ I just find it hard to believe that he would have pushed it in if he didn't have a plan | 20:16 |
| Dink | Wubbbi, yeah I watch cnn video, metacafe, youtube fine | 20:16 |
| slangasek | Sarvatt_: I meant RAOF / ScottK, but I can include you in the beatings if you wish :) | 20:17 |
| Wubbbi | Dink: CPU? | 20:17 |
| Dink | Intel(R) Atom(TM) CPU N270 @ 1.60GHz | 20:17 |
| Dink | Its an acer aod250 | 20:17 |
| slangasek | Sarvatt_: it's possible I misunderstood the bug summary; I thought "testing done: unity" meant unity had been regression-tested on intel | 20:18 |
| Wubbbi | Dink: duh ... the CPU is the same as mine ... o.O ... how could there be such a diffrent? | 20:18 |
| Dink | got that keyboard,mouse,external dvd, speakers, 2gb ram, and usb hub for 225$ on tigerdirect | 20:18 |
| bryceh | rickspencer3, true. I do have to admit that the final weeks of release the xorg hot seat has a lot, lot, lot of balls to juggle, it's easy to miss something | 20:18 |
| Dink | was a black friday deal | 20:18 |
| Dink | Wubbbi, umm | 20:18 |
| bryceh | slangasek, bug#? | 20:19 |
| slangasek | bryceh: the FFe bug; 631413 | 20:19 |
| Dink | Wubbbi, using flushplugin-installer | 20:19 |
| Dink | and firefox | 20:19 |
| Dink | and sun java1.6 plugin | 20:19 |
| slangasek | had I understood that unity hadn't actually been regression-tested on intel, I would have blocked the FFe on that | 20:19 |
| slangasek | because yeah, mesa regression testsuites are *not* evidence that anyone will be able to see their UI :) | 20:20 |
| kenvandine | i doubt any enjoyed that white screen of death this morning :/ | 20:23 |
| Dink | I got it right when it was released. Was hoping it fixed my issue :P | 20:24 |
| Dink | Was waiting for the new mesa package | 20:24 |
| Wubbbi | Dink: hmm ... well ... can you confirm a very slow Wlan connection? ( I bet you have broadcom as I have ) | 20:24 |
| Dink | Wubbbi, umm I can try sec let me enable wifi | 20:24 |
| Wubbbi | k | 20:24 |
| vish | kenvandine: oh no! you just dint! " white screen of death " , trying to start a trend, eh? ;) | 20:25 |
| Wubbbi | I download with 3,4kb/s and it dont matter what and where o.O | 20:25 |
| vish | WSOD! | 20:25 |
| kenvandine | vish, i woke up the guy next to me on the plane going to prague with that white screen :) | 20:25 |
| kenvandine | there was like 4 or 5 seconds of that when mutter loaded, and on a dark plane, wow that is bright | 20:26 |
| kenvandine | he woke up and gave me an evil eye | 20:26 |
| vish | hehe! | 20:26 |
| Dink | Wubbbi, got both wifi and eth0 on now. Going to shutdown eth0 and run speed test | 20:26 |
| Wubbbi | Dink: thank you! :) | 20:26 |
| bryceh | kenvandine, heh | 20:29 |
| bryceh | kenvandine, that's awesome | 20:29 |
| === Dink|W is now known as Dink | ||
| Dink | Wubbbi, what site you using ? | 20:30 |
| kenvandine | that screen is bright in a dark room! | 20:30 |
| Dink | Last Result: | 20:30 |
| Dink | Download Speed: 13916 kbps (1739.5 KB/sec transfer rate) | 20:30 |
| Dink | Upload Speed: 4514 kbps (564.3 KB/sec transfer rate) | 20:30 |
| Wubbbi | Dink: where can I test my connection? | 20:30 |
| Dink | http://www.speakeasy.net/speedtest/ | 20:30 |
| Dink | That is where I just tested from | 20:30 |
| bryceh | ok guys, network testing is definitely off topic for this channel ;-) | 20:30 |
| Wubbbi | Dink: It dont matter. I start google.com and I takes 20 sec. to load it | 20:30 |
| Dink | ok let me rephrase that... so its "on topic". I used that site to test how my video card rendered the test result | 20:32 |
| Dink | ;) | 20:32 |
| Wubbbi | hehe | 20:32 |
| bryceh | Dink, aha better ;-) | 20:33 |
| bryceh | Sarvatt, looking at the source, "Mesa %s implementation error" appears to be just a generic "something went wrong" error | 20:34 |
| bryceh | wait, nevermind | 20:36 |
| bryceh | ok | 20:39 |
| bryceh | Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 | 20:39 |
| bryceh | this is occurring (I guess) in i915_set_draw_region() in a switch statement that examines irb->Base.Format | 20:40 |
| bryceh | format 21 appears to be MESA_FORMAT_A8 if I count my enums correctly | 20:41 |
| bryceh | that's certainly not covered in the switch | 20:41 |
| bryceh | but A8 sounds like it's two bytes of just alpha info | 20:41 |
| bryceh | kenvandine, this making any sense to you? | 20:42 |
| kenvandine | bryceh, no... anything X related it greek to me | 20:43 |
| kenvandine | njpatel, did you see that? | 20:43 |
| bryceh | well, it gives me some nice google fodder at least :-) | 20:44 |
| kenvandine | hehe | 20:44 |
| njpatel | no, it's not making sense, but I'm coming off the back of a 12hr day | 20:44 |
| njpatel | :) | 20:44 |
| kenvandine | slangasek, could you rescore https://edge.launchpad.net/ubuntu/+source/ubuntu-netbook-default-settings/0.8.6/+build/1961558 ? | 20:44 |
| kenvandine | that feels like it is never going to build... and has the work around for this mesa bug | 20:45 |
| bryceh | well, what I *think* it means here is that unity is passing in a request for some sort of texture that's in a format mesa isn't prepared to deal with... like some texture that is pure alpha channel? | 20:45 |
| bryceh | I'll keep poking | 20:46 |
| njpatel | mutter is passing, mutter | 20:46 |
| njpatel | actually, clutter is passing, clutter | 20:46 |
| bryceh | mutter clutter sputter | 20:46 |
| bryceh | "some X client is passing..." *grin* | 20:46 |
| njpatel | lol | 20:47 |
| bryceh | hmm, I see MESA_FORMAT_A8 referenced in the i965 code but not the i915 (aka 945) code... has anyone confirmed seeing this problem on i965 or newer hardware? | 20:48 |
| kenvandine | bryceh, is GM45 the same as 965? | 20:48 |
| kenvandine | i see it on GM45 | 20:48 |
| slangasek | kenvandine: I'm not a buildd admin, sorry | 20:48 |
| kenvandine | ah, sorry | 20:48 |
| bryceh | hmm I think so lemme check | 20:48 |
| * kenvandine looks for someone else :) | 20:49 | |
| Sarvatt | the renderbuffer error shouldn't be there on 965 | 20:49 |
| Sarvatt | http://cgit.freedesktop.org/mesa/mesa/commit/?id=638342858894293246400d95a90d153c7f66719a | 20:49 |
| slangasek | GM45 is newer than 965 | 20:49 |
| slangasek | er, as a chip - maybe you're asking if the code is the same :) | 20:50 |
| Sarvatt | yeah its the same kenvandine | 20:50 |
| JanC | According to intel's site that chipset contains a 4500MHD | 20:51 |
| bryceh | yeah, cantiga right, yeah that's >=965 | 20:52 |
| DBO | dont mind me, just here to watch the train wreck | 20:53 |
| kenvandine | hey DBO | 20:54 |
| kenvandine | your new identity? | 20:54 |
| kenvandine | new/old | 20:54 |
| kenvandine | :) | 20:54 |
| bryceh | sassy DBO? | 20:54 |
| DBO | maybe I can get Jason on freenode | 20:54 |
| * DBO checks | 20:54 | |
| DBO | bryceh, a little, had a rough day, I figure humor is the only solution | 20:54 |
| * kenvandine rather likes DBO | 20:54 | |
| kenvandine | :) | 20:54 |
| DBO | also I like watching clutter fail | 20:55 |
| Wubbbi | seems like our Mesa friends didn't a good job :/ ... Have we told them that Clutter is working fine? | 21:04 |
| bryceh | tbh it's probably Intel rather than mesa | 21:04 |
| bryceh | I mean, the i9xx code in mesa, which is maintained by Intel, not the mesa folks | 21:04 |
| Wubbbi | ha! I always knew that Intel is not perfect! xD ... Wel rather I like AMD more :P | 21:05 |
| KnifeySpooney | Hi, earlier my gpu hung and I had to go to a console to reboot. I saw that 'intel_gpu_dump' was running, then it finished, but I don't know where it logged to. Does anyone know where the intel log file is? | 21:17 |
| bryceh | /var/crash probably | 21:17 |
| Dink | Sarvatt_, I have attached a short video of some of the things that occur with my bug. Hope it helps visualizing what I wrote. | 21:18 |
| Wubbbi | Dink: I wanna see it to. Have you got a link? | 21:18 |
| Dink | https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/638496 | 21:19 |
| ubot4 | Launchpad bug 638496 in xserver-xorg-video-intel (Ubuntu) "[i945GME] Unity session freezing. (affects: 1) (heat: 8)" [Undecided,New] | 21:19 |
| Dink | The video is what happens right before the freeze. It is when things start working funky. Then after a while of that BAM it freezes | 21:20 |
| Dink | Wubbbi, do you experience any of that ? | 21:25 |
| Dink | On an external monitor, windows dont move, get put in front when you click on it, and/or freezes everything up? | 21:25 |
| bryceh | kenvandine, on your system do you also see the error message sarvatt sees? Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 | 21:27 |
| bryceh | from what I can tell, that exact error message would not occur on i965 or newer, so it seems like you wouldn't be seeing that particular error message | 21:27 |
| bryceh | http://paste.ubuntu.com/494295/ | 21:28 |
| kenvandine | <kenvandine> (unity:32356): ClutterGLX-CRITICAL **: Unable to make the stage window 0x5e0004f the current GLX drawable | 21:28 |
| kenvandine | that is what i see | 21:28 |
| kenvandine | and crash | 21:28 |
| bryceh | hrm | 21:29 |
| bryceh | kenvandine, can you pastebin your full output when running unity? | 21:30 |
| bryceh | boy it'd suck if there are multiple bugs going on here | 21:30 |
| kenvandine | not right now | 21:30 |
| kenvandine | gotta get a few uploads done... soon :) | 21:30 |
| bryceh | ok, well that error message appears not to be an X error message (it's not in mesa's source code anyway), so isn't giving me much to go on | 21:31 |
| kenvandine | yeah, sorry... will try in a few | 21:33 |
| Wubbbi | Dink: I dont understand whats the issue in your video? seems to work fine | 21:39 |
| Dink | Sometimes when I click on open windows it does not bring it to the top, when I try to move the window it moves another window. After that eventually everything freezes and I cant click, move, edit anything and have to hup mutter to get it work again. | 21:40 |
| Dink | And the strange part is this only happens when external monitor is connected | 21:41 |
| Dink | If you look at the end of the video shows the exaple where I clicked on FF to move it and it moved xchat | 21:43 |
| Dink | Before that it also showed where I click on xchat and it did not bring it to the top | 21:44 |
| Wubbbi | Seems to be a variable mess in the source code. Sarvatt_ did you take a lot on it? | 21:44 |
| Dink | and jsut now I have a pidgin window open and when I click on it to type, it types it in here and not there. I just alt-tab to it and it works. | 21:45 |
| Wubbbi | look | 21:45 |
| Dink | Wubbbi, so you don't experience any of that ? | 21:46 |
| Wubbbi | Dink: not yet ... | 21:46 |
| Dink | Do you use your netbook with external monitor ? | 21:46 |
| Wubbbi | Dink: nope ... maybe thats the point ;) | 21:53 |
| Dink | Ahh yeah to me it only happens when I do. | 21:54 |
| Dink | Works fine when external monitor is not connected. | 21:54 |
| Sarvatt | CLUTTER_VBLANK=dri also works | 21:56 |
| Dink | Is it better setting it to dri vs none ? | 22:00 |
| bryceh | Sarvatt, I've reproduced the unity output you get, with the Bad renderbuffer format 21 error | 22:00 |
| vish | bryceh: oh... i'm not very familiar with unity.. i just reported my bug .. :) | 22:08 |
| vish | re: the comment on the RV515 bug. | 22:08 |
| bryceh | vish, ok well I took a look but looks to me like more analysis needs done before it can be taken as an X bug | 22:09 |
| bryceh | hard to tell what the issue is | 22:09 |
| vish | bryceh: dbarth mentioned it as a "Related to the disable-effects flag requirement" , he might know more. i guess | 22:12 |
| bryceh | huh I've no idea what that means | 22:15 |
| njpatel | bryceh, vish, that only would be the issue if unity was effected, however mutter is, without unity plugin, so it can't be the effects framework | 22:18 |
| njpatel | if only unity was....* | 22:18 |
| vish | njpatel: nutter works if the CLUTTER_VBLANK=none is used | 22:18 |
| vish | only unity fails | 22:18 |
| vish | err , mutter ;) | 22:18 |
| njpatel | vish, I'm using Unity now with CLUTTER_VBLANK=none | 22:19 |
| vish | njpatel: but it doesnt help in this ATI RV515 | 22:19 |
| njpatel | vish, the unity failure then is a different one (which is to do with effects), but that's not related to this issue with mesa | 22:19 |
| njpatel | vish, sorry, I thought you were talking about the issue currently with Clutter and the latest update to mesa | 22:19 |
| vish | njpatel: nah , not related to today's WSOD! [ kenvandine ;) ] | 22:20 |
| njpatel | heh | 22:20 |
| bryceh | "white screen on _____" is a really generic symptom with 3D / mesa / X bugs, so just because the two cases have similar symptoms is no indication it's the same bug | 22:20 |
| bryceh | indeed, different hardware, different timings when the issue was first noticed, different workarounds... all point to these being two separate bugs | 22:21 |
| vish | njpatel: it was a problem even before , previously it would just keep reloading , and with mesa7.9 its just a blank white screen. its about Bug 616997 btw | 22:21 |
| ubot4 | Launchpad bug 616997 in unity (Ubuntu) (and 3 other projects) "[RV515] Unity keeps reloading with a white background (affects: 8) (dups: 2) (heat: 54)" [Low,Triaged] https://launchpad.net/bugs/616997 | 22:21 |
| njpatel | vish, yep, we got stalled on a workaround, I pinged the team for that this morning, hopefully we'll have an answer soon. Sorry it's taken so long, but it's harder to switchs tuff off then switch it on it seems :) | 22:22 |
| vish | \o/ unityyyyyyyyy soon! | 22:22 |
| * njpatel wishes drivers just worked | 22:23 | |
| njpatel | right now we have an issue that, only on intel it seems, maximising windows will cause them to close | 22:23 |
| njpatel | anyway, it's late here and time to go to sleep. have a good night all | 22:24 |
| Sarvatt | dingding, regression happened between 0825 and 0909 | 22:27 |
| Sarvatt | unity and clutter apps are fine on the 0825 snapshot we tested the hell out of | 22:28 |
| * Sarvatt starts eyeing glx changes between those two dates | 22:29 | |
| bryceh | sarvatt, how can I best help? | 22:30 |
| Sarvatt | yow, quite a lot of glx changes between them :( | 22:32 |
| jcristau | krh got taken down the glx rabbit hole | 22:35 |
| bryceh | sarvatt, yeah pages and pages | 22:36 |
| bryceh | what was the commit id for 825? | 22:36 |
| Sarvatt | b2872ea353efd117fcc4d22f0ca66a26f95a14c4 | 22:37 |
| Sarvatt | everything from Revert "Enable nouveau-vieux classic mesa driver." http://git.debian.org/?p=pkg-xorg/lib/mesa.git;a=shortlog;h=refs/heads/ubuntu-maverick to my commit here http://git.debian.org/?p=pkg-xorg/lib/mesa.git;a=shortlog;h=refs/heads/ubuntu-maverick;pg=5 | 22:38 |
| Sarvatt | wait, let me double check that hash | 22:39 |
| Sarvatt | f81cec52a | 22:40 |
| Sarvatt | yeesh, that was the id of the g.d.o commit :( | 22:41 |
| bryceh | Sarvatt, I see there's several build config changes that RAOF did, have we ruled them out? | 22:42 |
| bryceh | e.g. http://git.debian.org/?p=pkg-xorg/lib/mesa.git;a=commitdiff;h=d8ac3c6a5befcfb832cbcee30f165c087cc05d54;hp=d7ae1936eee920273af36a39edee6797f5bf4c79 | 22:43 |
| Sarvatt | yeah those are in this 0825 checkout, he's a day ahead :) | 22:44 |
| bryceh | ok this one looks the most interesting to me - http://git.debian.org/?p=pkg-xorg/lib/mesa.git;a=blobdiff;f=src/mesa/drivers/dri/intel/intel_fbo.c;h=2693b5fa72e5ec9eb1cbe406dcd4761b7497c492;hp=6435857f33020470c7eb48f22a789fcfd1812fd8;hb=bda941e1b895547d680b68eaf28ae2db11e6149f;hpb=0599509fc4e21a69bcbf121deacbc631c38ccfa2 | 22:46 |
| bryceh | wait, no | 22:46 |
| bryceh | http://git.debian.org/?p=pkg-xorg/lib/mesa.git;a=commitdiff;h=86ddb356e8278423ef41125be627e57e073471d1;hp=7b07674667e8e34c936c417ab6f1c2e33093205b | 22:47 |
| Sarvatt | http://sarvatt.com/downloads/unity-xtrace.txt | 22:56 |
| bryceh | feh, I wish mesa didn't take so fricken long to build | 23:17 |
| Sarvatt | got the 0825 debs at http://sarvatt.com/downloads/foobar/ | 23:18 |
| bryceh | sarvatt, what are you working on next? | 23:19 |
| Sarvatt | bryceh: git20100903 a5fd0396 = good | 23:49 |
| bryceh | well that narrows it down nicely | 23:51 |
| Sarvatt | git20100904.20bf5037 good | 23:52 |
| Dink | Sarvatt, hmmm would any of these packages affect the problem I had.... gnome-shell, mesa-utils, gir1.0-* <-- removed libqt4-*, firefox upgraded ?? Things seem to work a lot better now. Haven't have any of my issues yet. | 23:52 |
| Dink | gnome-shell, mesa-utils, girl1.0-* was removed. libqt4-* and firefox was upgraded | 23:53 |
| bryceh | Dink, come back later, we're a bit tied up on this release critical issue | 23:53 |
| bryceh | and I suspect sarvatt's wearing out :-) | 23:54 |
| Dink | ohh my bad sorry was just excited that things work now with no issue. | 23:54 |
| Sarvatt | git20100907.60fce154 good | 23:55 |
| Sarvatt | git20100908.5ecd9c70 bad | 23:55 |
| bryceh | SWEET | 23:55 |
| bryceh | wow, you're testing these fast | 23:55 |
| bryceh | can you post a debdiff between those two dsc's? | 23:56 |
| Sarvatt | one sec | 23:56 |
| Sarvatt | takes a bit on this atom :) | 23:58 |
| Sarvatt | http://sarvatt.com/downloads/clutter-busted.debdiff | 23:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!