[01:28] <qdb> hello. netbook with intel gma 3600 loads to black screen, but  recovery - resume normal and recovery - root - restart fixes it temporarily ( http://ubuntuforums.org/showthread.php?p=12055698#post12055698 ) . i have looked at logs  - found nothing. i have tried to add "nomodeset" to grub menu item , has not helped... but at next start after starting with nomodeset it loads nirmally, as i remember... any ideas? we need gma 3600 driver just for correct resoluti
[01:28] <qdb>  on 1024 600, and there  it  is, just there is black screen appear
[07:42] <tjaalton> RAOF: is 1.12 ready to enter quantal?
[07:43] <tjaalton> or is it best left for after the weekend..
[07:44] <RAOF> tjaalton: You might want to try “apt-cache policy xserver-xorg-core” ☺
[07:44] <tjaalton> RAOF: ah :)
[07:44] <tjaalton> and I might want to try to actually upgrade to quantal too..
[07:44] <RAOF> It got shunted out after the A2 freeze ended.
[07:44] <tjaalton> oh
[07:45] <tjaalton> hmm, I uploaded -synaptics to proposed
[07:45] <tjaalton> probably should've pushed it straight to quantal then, especially since it b-depends on 1.12
[07:45] <tjaalton> so it would've just waited if it wasn't available
[07:46] <mlankhorst> morning P)
[08:15] <jcristau> heya
[08:16] <jcristau> https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/785368 seems to affect lucid, would you consider fixing it there?
[08:20]  * jcristau LD_PRELOADs the squeeze libGL for now.
[08:23] <tjaalton> don't see why not. have you tested the patch?
[08:24] <jcristau> well i fixed it in squeeze last september
[08:26] <tjaalton> ok
[08:27] <jcristau> http://anonscm.debian.org/gitweb/?p=pkg-xorg/lib/mesa.git;a=blob;f=debian/patches/08_glx-ignore-BadRequest-errors-from-DRI2Connect.diff;h=ab0f6701c550c0bd9453b68284624507e070bbbe;hb=40f332c65bf0ae454c9c0a07d3e87c671422cde0 should apply to lucid without too much trouble
[08:27] <tjaalton> i've added a lucid task
[08:27] <jcristau> since they're both on 7.7.1
[08:30] <mlankhorst> lucid still supported? :(
[08:30] <tjaalton> a year to go
[08:30] <tjaalton> or 10 months
[08:31] <mlankhorst> guess it makes sense for transitioning over to precise
[08:31] <tjaalton> in this case the fix is known, but for bugs where you'd need to chase down commits.. nah
[08:32] <tjaalton> nearly closed bug 1018706 due to that
[08:33] <tjaalton> hmm, for cases like this tags would be nice
[08:34] <tjaalton> guess I'll just pull the source package
[08:34] <mlankhorst> well fortunately we would have the backported stack next time we could point to
[08:34] <tjaalton> right
[08:38] <mlankhorst> although I hope we won't port the lts+1 to previous lts right?
[08:38] <tjaalton> nono :)
[08:38] <tjaalton> 46.3 GiB will be downloaded into archive.
[08:38] <tjaalton> Downloading 37877 archive files using 10 threads...
[08:39] <tjaalton> mirroring sid takes time..
[08:39] <tjaalton> just for sbuild :)
[08:53] <tjaalton> anyone got 'parallel=n' working with sbuild?
[08:53] <tjaalton> it doesn't honor DEB_BUILD_OPTIONS here
[08:56] <RAOF> It worked last time I tried it?
[08:56] <RAOF> Oh!
[08:56] <RAOF> sbuild -j5
[08:56] <mlankhorst> morning
[08:56] <tjaalton> RAOF: oh.. thanks
[08:57] <tjaalton> the build environment does get DEB_BUILD_OPIONS but they're not used
[08:57] <tjaalton> anyway, glad if this works
[08:58] <mlankhorst> so mess up TTM today or not? :)
[08:59] <mlankhorst> deadlock problem was solved by delayed fput interestingly enough, so next kernel
[09:00] <tjaalton> jcristau: uploaded
[09:00] <tjaalton> build-tested
[09:00] <jcristau> tjaalton: awesome, thanks a lot
[09:01] <jcristau> i can glxinfo-test an amd64 libgl1-mesa-glx deb once it's built
[09:02] <mlankhorst> RAOF: it's still useful to work on the prime stuff right? Feels like I'm at my third attempt from scratch now
[09:02] <tjaalton> yeah, needs someone (*cough* RAOF *cough*) to accept it to proposed
[09:03] <jcristau> sneaky :)
[09:05] <mlankhorst> but at this point I learned all the requirements to prevent deadlocks, the locking one will be solved in 3.6, none of the dma-buf operations except import/export/destroy may take another lock to prevent deadlocks.
[09:06] <mlankhorst> destroy will potentially deadlock on current kernels but the work from https://lwn.net/Articles/494158/ will cause it to be fixed without needing any additional code :)
[09:33] <tjaalton> meh, -j isn't settable in .sbuildrc
[09:34] <tjaalton> time for an alias then
[09:46] <mlankhorst> tjaalton: or fix sbuild?
[09:47] <jcristau> incidentally:
[09:47] <jcristau> 00:31 < rleigh> I'm working an sbuild upload prior to the freeze.  I've pushed fixes for #629109, #663089, #665215, #665372, #673676, #675861 and #676475 
[09:47] <jcristau>                 tonight.
[09:47] <jcristau> 00:32 < rleigh> I'll look at #529281, #637931, #641065, #662785, #673009, #674669 and #676834 tomorrow evening.
[09:47] <jcristau> 00:32 < rleigh> If there's anything else you'd like looking into for wheezy, please let me know!
[09:47] <jcristau> :)
[09:51] <tjaalton> jcristau: oh :)
[09:57] <tjaalton> though looks like -j doesn't work either
[09:57] <tjaalton> wth
[10:00] <tjaalton> oh it's the package that's faulty
[10:01] <tjaalton> still, wth :)
[10:02] <tjaalton> hmm so the package needs the NUMJOBS/MAKEFLAGS stuff, even if using dh?
[10:04] <tjaalton> oh, --parallel
[10:08] <tjaalton> finally :)
[10:12] <tjaalton> though in this case libtool wasn't happy in the end
[13:25] <bjsnider> tjaalton, unfortunately, the .27 kernel doesn't fix 999910 for me. still crashes about once a day
[13:26] <tjaalton> bjsnider: oh..
[13:27] <tjaalton> you still get a hard hang and can't get to the machine remotely?
[13:28] <bjsnider> right
[13:28] <tjaalton> has any mainline kernel worked?
[13:28] <tjaalton> and is it a laptop or desktop?
[13:28] <bjsnider> desktop
[13:28] <tjaalton> ok
[13:29] <bjsnider> i just built it a week ago
[13:29] <tjaalton> and using just the igp?
[13:29] <bjsnider> i am going to have to try the latest kernel
[13:29] <bjsnider> yep, just the ivybridge chip
[13:29] <tjaalton> ok, cool
[13:30] <bjsnider> what kernel were you pulling code from?
[13:30] <tjaalton> the workaround was from 3.3-rc1
[13:31] <tjaalton> but 3.3.x wasn't stable, so it regressed there somehow
[13:31] <tjaalton> 3.4-rc1 was stable again
[13:31] <bjsnider> alrigth i'll just grab whatever linus released most recently
[13:31] <bjsnider> and report whether it fixed the issue
[13:31] <bjsnider> later today that is
[13:31] <tjaalton> cool
[13:35] <tjaalton> well, give it some time. also, can you reproduce it faster by running the silly gzip/glxgears -script?
[13:35] <johanbr> Hi. The linux-image-3.5.0-2-generic kernel in xorg-edgers for precise breaks sound for me (pulseaudio hangs). Is this a known problem?
[13:36] <tjaalton> johanbr: no idea, but check http://www.theorangenotebook.com/2012/06/call-for-testing-1210-kernel-on-1204.html
[13:37] <tjaalton> probably would need to purge edgers first
[13:38] <johanbr> tjaalton, ahh, thank you, I'll see if that kernel works better
[13:38] <tjaalton> it's the same kernel
[13:39] <tjaalton> the point was that ther's a procedure for reporting issues..
[13:39] <tjaalton> +e
[13:40] <johanbr> oh :)
[13:40] <johanbr> then I'll try reporting the issue :)
[13:41] <tjaalton> good :)
[14:07] <bjsnider> tjaalton, you said in the bug report that you were git bisecting looking for the commit that stabilizes the intel driver, but isn't that git bisect in reverse? isn't it usually that y ou're looking backwards to find the regression?
[14:13] <tjaalton> bjsnider: i think it works both ways. and I bumped to this commit by accident, by going through the commit list
[14:13] <tjaalton> didn't get to git-bisect at all, which was a relief
[16:35] <qdb> hello. netbook with intel gma 3600 loads to black screen, but  recovery - resume normal and recovery - root - restart fixes it temporarily ( http://ubuntuforums.org/showthread.php?p=12055698#post12055698 ) . i have looked at logs  - found nothing. i have tried to add "nomodeset" to grub menu item , has not helped... but at next start after starting with nomodeset it loads nirmally, as i remember... any ideas? we need gma 3600 driver just for correct resoluti
[16:35] <qdb>  on 1024 600, and there  it  is, just there is black screen appear
[16:36] <tjaalton> qdb: file a bug with 'ubuntu-bug xorg'
[16:36] <tjaalton> ahh
[16:36] <mlankhorst> powervr? Ouch..
[16:37] <tjaalton> yeah don't bother, i guess :/
[16:37] <qdb> wait , this is about ubuntu quantal
[16:37] <qdb> i copied this from #ubuntu+1
[16:37] <tjaalton> does precise work?
[16:37] <qdb> older versions do not have gma 3600 driver enabled
[16:38] <tjaalton> then you'd be better off reporting it upstream i think
[16:38] <tjaalton> is it cedartrail?
[16:38] <mlankhorst> do they still ship new machines with powervr? 
[16:38] <tjaalton> yeah
[16:38] <tjaalton> CDV..
[16:38] <mlankhorst> :S
[16:39] <tjaalton> ah, so it is cedartrail
[16:39] <tjaalton> if wikipedia is to be trusteed
[16:39] <tjaalton> -e
[16:44] <Sarvatt> qdb: try adding GRUB_GFXPAYLOAD_LINUX=text to /etc/default/grub, sudo update-grub and reboot, that'll get rid of vt.handoff=7
[16:46] <Sarvatt> (get rid of nomodeset if you still have it there too at the same time)
[17:09] <bryceh> Sarvatt, hey did you get a chance to turn on SNA in edgers?
[17:11] <Sarvatt> bryceh: by default? its on so you can pick it in xorg.conf, didn't make it default
[17:11] <Sarvatt> i'll turn it on by default now
[17:12] <Sarvatt> and put a big note on the ppa page about how to change it back (didn't work right here yesterday)
[17:13] <Sarvatt> ok its up now in quantal
[17:32] <bryceh> thanks
[18:09]  * mlankhorst prepares for incoming complaints ;)
[18:13] <mlankhorst> http://littlebigdetails.com/post/20286226116/mac-osx-the-world-map-displayed-within-the-data
[18:29] <mlankhorst> bryceh: so erm what is the official solution for powervr machines, don't support? :S
[18:30] <qdb> thank you , i ll try
[18:30] <bryceh> mlankhorst, right no support
[18:32] <tjaalton> there's some modesetting support in the kernel but that's it
[18:35] <mlankhorst> tjaalton: yeah if pandaboard is powervr too I can imagine the whole no support thing, corrupts and crashes without doing anything
[18:38] <mlankhorst> pretty sure that changing resolution twice caused a kernel panic.
[18:42] <qdb> Sarvatt, it works, thank you
[20:20] <tjaalton> wha.. managed to get a soft cpu lock, audio on a loop, display frozen, could ssh in
[20:20] <tjaalton> then after a minute or so business as usual
[20:20] <tjaalton> no gpu reset
[20:45] <bjsnider> tjaalton, with the ivybridge chip?
[20:48] <tjaalton> bjsnider: yes
[20:48] <bjsnider> i'll try waiting a minute next time
[20:49] <tjaalton> i had accidentally minimized all windows via alt-tabbing, and then I restored them by going through all the workspaces, and after restoring one fullscreen window it froze
[20:49] <bjsnider> you're using the .27 kernel?
[20:51] <tjaalton> no, mine (-26 plus the three commits)
[20:52] <tjaalton> aka -90
[21:08] <bryceh> mlankhorst, did you see https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/1019131 ?
[21:32] <mlankhorst> shouldn't that be done by compiz?
[21:33] <bryceh> dunno
[21:35] <Sarvatt> ddx it looks like, GLXVblank wasn't enabled by default because it needed a newish kernel in 2010 but safe to safe it could be now?
[21:36] <bjsnider> it needs to be done in xorg.conf for nouveau
[21:36] <Sarvatt> http://cgit.freedesktop.org/nouveau/xf86-video-nouveau/commit/?id=70f0d2c886ceaa965ca4864788f4dd8e8f757a92
[21:36] <mlankhorst> might be
[21:37] <mlankhorst> but im not sure i think calim was working on nvc0 for that
[21:37] <Sarvatt> safe to say it could be rather
[22:03] <mlankhorst> but I'm all for enabling vsync, real 3d is going to have to wait for a bit