[05:51] libdrm, 2.7.99.1+git uploaded [09:08] compiz is still disabled for i965.. mvo? [09:10] tjaalton: yes, is the driver all good now so that this can be changed? [09:11] mvo: I've been using it for a couple of weeks in karmic with no issues [09:12] tjaalton: thanks, I fix it in bzr [09:12] mvo: cool, thanks [11:04] Is "xset force dpms off" supposed to work with KMS? [12:21] damn .... xorg sigsegv, not handy [12:28] jbarnes, is in the UK? [12:29] jbarnes_LHR, even [12:31] bryce, seeing coredumps on the standard xorg server (not kms) on an intel deskside [12:47] yep in the UK [12:47] london near the tower bridge [13:09] heh, for long? having fun? [13:14] apparently not ;) [13:26] lool: not sure dpms is hooked up yet [14:21] tjaalton, jcristau: is xserver-xorg the only package that calls dexconf? [14:22] tseliot: if it's not, i'd love to know [14:23] should be, yes :) [14:24] jcristau: I built an image with a package that modifies the xorg.conf (for the OEM team) and I predepend on xserver-xorg but it looks like my modified xorg.conf is renamed as xorg.conf.$date and dexconf is called [14:24] my package modifies the xorg.conf in the postinst and I'm wondering why xserver-xorg modifies the xorg.conf after my package [14:25] despite the predepends [14:25] maybe debconf is delayed? [14:26] dpkg tells you when it configures packages [14:26] so check what gets configured after your package? [14:27] This happens before my package is installed: Setting up xserver-xorg (1:7.4~5ubuntu18) ... xserver-xorg postinst warning: not updating configuration as per $XORG_CUSTOM [14:29] then my package is configured. The rest is just a standard jaunty installation [14:31] jcristau: these are configured after my package: [14:31] Setting up xinit (1.0.9-2) ... Setting up xinput (1.4.0-1) ... Setting up xorg (1:7.4~5ubuntu18) ... Setting up xscreensaver-gl (5.07-0ubuntu3) ... [14:31] none of those touch xorg.conf.. [14:32] yes, I know [14:33] but it looks like the xorg.conf that my package modifies is still empty in spite of the fact that xserver-xorg has already been configured [14:33] the preinst of xserver-xorg creates an empty file [14:33] which is filled in the postinst of xserver-xorg [14:34] default jaunty installation results in an empty xorg.conf [14:35] tjaalton: not here. It has Device, Screen and Monitor sections [14:35] tseliot: ah right. i should fix that. [14:35] jcristau: ? What do you need to fix (and how)? [14:36] make sure i don't create an empty xorg.conf [14:39] ok [14:42] tseliot: ok, maybe happens only when doing a preseeded installation [14:43] tjaalton: and how can I get around the problem? Shall I predepend on the "xorg" metapackage? [14:57] 'not updating configuration as per $XORG_CONFIG' means you set XORG_CONFIG=custom in /etc/default/xorg, so xserver-xorg.postinst exits without doing anything [14:58] jcristau: but there's no such file in the generated image [15:00] *shrug* [15:01] and dexconf is being called. Maybe something reconfigures xserver-xorg at a certain point === You're now known as ubuntulog [15:12] jcristau: shouldn't this: http://pastebin.ubuntu.com/187401/ be more like this? http://pastebin.ubuntu.com/187404/ [15:12] dunno, what's the difference? [15:13] i mean, why not allow XORG_CONFIG from the environment rather than that default file? [15:14] where can that variable be defined apart from the default file? [15:14] your environment [15:17] ok, so somewhere in this system has that variable set to custom... [15:17] without sourcing? [15:23] morning [15:23] hey bryce [15:28] tseliot: heya, all done with demos and such? hope all went well :-) [15:29] bryce: I'm still fighting with xserver-xorg which is overwriting my customised xorg.conf even though I predepend on it :-( [15:37] apw: if you can reproduce the crash and get a full backtrace (process documented at http://wiki.ubuntu.com/X/) it could help [15:39] bryce, yep 100% reproucible [15:40] jcristau: removing the postinst/preinst, etc. of xserver-xorg shouldn't cause problems, right? (just to be sure) [15:42] dunno. [15:50] wow, now this looks interesting http://git.kernel.org/?p=linux/kernel/git/anholt/drm-intel.git;a=commit;h=5112f5941c34ea49155a02672647c61afb1e6bf5 [16:08] bryce: could you look at a bug for me, it seems to match an upstream link for 2 other bugs you were workign on [16:08] bug 374398 [16:08] Launchpad bug 374398 in xserver-xorg-video-ati "Glitches around checkboxes in Firefox" [Undecided,New] https://launchpad.net/bugs/374398 [16:10] hi micahg, what aspect needs looked at? [16:10] if it should be merged with bug 291053 [16:10] Launchpad bug 291053 in xserver-xorg-driver-ati "[R500 x1400] Occasional mouse cursor corruption [EXA enabled]" [Unknown,Confirmed] https://launchpad.net/bugs/291053 [16:10] fwiw, there's a slew of known glitches with -ati since we moved to EXA. [16:11] that's good to know [16:12] yes, it could well be a dupe [16:12] do you want to change what's appropriate in the title in the original? [16:12] or does more triage need to happen [16:13] it can probably be changed to "Small pixmap corruption [EXA enabled]" to match upstream's bug title [16:13] ok, and I should just mark as dupe then? [16:13] micahg: at this stage what needs done is to test patches alex provides and give feedback [16:13] micahg: yes that sounds good [16:14] are the patches upstream? [16:14] they get posted on the upstream bug report [16:14] alex likes to have folks test them before he puts them in the upstream tree [16:15] Where should people comment about the patches, upstream, right? [16:16] yes, on the upstream bug [16:17] ok [16:17] done [16:17] thanks micahg :-) [16:17] no problem, I just wasn't sure what to do as I normally work on Firefox bugs [16:18] now I know where to go for X bugs :) [16:19] micahg: what do you work on normally? [16:19] Firefox bugs [16:19] ah kewl [16:20] bryce: may I PM you [16:20] yeah, oftentimes the firefox X bugs are display issues rather than freezes/crashes, so since they're cosmetic often aren't given as much attention [16:21] so if you can take them upstream directly, it could help out a lot [16:21] micahg: if it is about something private; otherwise if it's just X questions this channel is most appropriate [16:21] the problem is usually figuring out which other component is causing the trouble [16:26] micahg: btw there's some info at http://wiki.ubuntu.com/X/ that could be useful [16:27] covers how to triage X bugs, how to troubleshoot issues, and so forth. It should help answer the question about what component causes what problem [16:27] I should probably read that [16:27] as a general rule, display issues are almost always caused either by the video driver the user is using [16:27] bryce, added the full stack trace to the bug [16:27] s/either// [16:28] apw: excellent thanks [16:28] I830DRI2CopyRegion() goes pop :) have fun [16:28] micahg: er I should say, if the problem is an X bug, it's usually in the video driver. Sometimes with firefox the issues have been further up the stack in cairo or an intermediate library [16:29] apw: was this while viewing a really large jpg? [16:29] nope. this is while engaging the screen saver [16:29] hmm [16:29] it fades to black then dumps [16:29] gl screensaver? [16:29] i guess its making a big black square to hide the screen [16:29] ahh [16:29] default gnome screen saver [16:30] apw: the 2.7.99.1 upload I did yesterday had a fix for a crash bug involving I830DRI2CopyRegion() when viewing large jpg images [16:30] hrm when might that hit the archives. i updated within the last two hours [16:30] yikes, there is a fricken huge spider on the ceiling above me. [16:30] apw: should have hit last night some time [16:30] it wants you for lunch [16:30] apt-cache madison xserver-xorg-video-intel [16:31] apw@chloe:~$ apt-cache madison xserver-xorg-video-intel [16:31] xserver-xorg-video-intel | 2:2.7.99.1+git20090602.ec2fde7c-0ubuntu1 | http://gb.archive.ubuntu.com karmic/main Packages [16:31] xserver-xorg-video-intel | 2:2.7.99.1+git20090602.ec2fde7c-0ubuntu1 | http://gb.archive.ubuntu.com karmic/main Sources [16:33] apt-cache policy xserver-xorg-video-intel ? [16:36] (spider got away, crazy fast mfker) [16:41] famous last words [16:41] bryce: this relevant? http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=b8e360bf2b77d28559d15a7c0f9c766848eb6ced [16:42] regarding apw's backtrace [16:42] also http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=5901a67fc85ac80fabfa98b78202a388445275c3 [16:43] apw@chloe:~$ apt-cache policy xserver-xorg-video-intel [16:43] xserver-xorg-video-intel: [16:43] Installed: 2:2.7.99.1+git20090602.ec2fde7c-0ubuntu1 [16:43] Candidate: 2:2.7.99.1+git20090602.ec2fde7c-0ubuntu1 [16:43] Version table: [16:43] *** 2:2.7.99.1+git20090602.ec2fde7c-0ubuntu1 0 [16:43] 500 http://gb.archive.ubuntu.com karmic/main Packages [16:43] 100 /var/lib/dpkg/status [16:43] updating edgers now [16:45] bryce, need anything else? [16:45] pDestBuffer = (DRI2BufferPtr) 0x0 [16:45] apw: nope, looks like a good trace I can take it from here [16:46] ouch ... [16:46] apw: later we may have some patches to test [16:46] sure thing. ping me, i may be here still if not shove them somewhere and i'll get to them first thing [16:46] i am guessing most of us are using the edgers things to get kms and not testing what the basic users get! [16:46] ok, may be a few hours, I have an AMD confcall in a few hours [16:47] i don't need the machine its on for interactive use so it can hang around waiting for you [16:47] ie. no rush, whenever suits [16:50] apw, re: the 2.6.30-999 resume swapper crash I had recently, I wasn't at home, so I didn't have netconsole running. I'll try to duplicate it again today. I've got a different issue when I resume with kms enabled, so I'm trying to split my time between the two ways I can make it crash :) [16:51] heheh fun for all the family for sure [16:52] re: trace, you just want the console output, right? [16:54] I'll try to work off 2.6.30-7 #8 with kms off, if that dupes easily, I'll move up to the latest daily and dupe it there too [17:31] apw: https://edge.launchpad.net/~sarvatt/+archive/ppa/+sourcepub/643434/+listing-archive-extra [17:35] Sarvatt, looking good here [17:37] good to hear, there was a last minute commit to enable vsync for DRI2 CopyRegion (to fix tearing problems) that made it into the 06-02 and it got fixed up by a bunch of commits this morning [17:39] and your test there was just picking up those fixes too [17:39] Sarvatt, this is logged under bug #383129 [17:39] Launchpad bug 383129 in xorg "x server dies with a SIGSEGV when gnome screen saver blanks the display" [Undecided,New] https://launchpad.net/bugs/383129 [18:18] holy hell, anholt's latest commit in for-review on drm-intel sped up 3d performance in UXA an amazing amount for me [18:19] sorry, wrong channel yet again [18:30] relevant here though, I'm seeing a 50% 3D performance increase in everything i've run so far on intel, not just openarena at high settings like the commit says. i put a kernel here incase anyone else wants to test it out http://sarvatt.com/downloads/LATEST/ [18:40] Sarvatt: kewl [18:44] wow, kms is wonderful, instant switching between VT [18:44] and full resolution in terminal [18:53] :-) [18:53] and no vt switch on s/r [18:54] I rarely suspend my laptop [18:57] jcristau: do you think we could --disable-gallium until the packaging issues are sorted out? [18:58] of mesa [18:59] tjaalton: i was going to wait for a reply from michel [19:00] jcristau: ok [19:00] tjaalton: maybe you can ping him about it on irc? [19:00] sure, could do that [20:45] Sarvatt: did i mention yet that the cursor gets hopelessly corrupted after hibernating and resuming? [20:48] just hibernate or suspend/resume too? why dont you file bug reports on fd.o about the problems you have? :D [20:49] if its suspend/resume too -- http://git.kernel.org/?p=linux/kernel/git/anholt/drm-intel.git;a=commit;h=596652e1094a9c7bebcd4d36a7f3d3a2d9698950 can try that kernel i just pasted to see if its fixed by that [23:15] is the intel driver crashing Xorg (sig 11) when full-screening things a known problem? [23:16] yup [23:16] can grab https://edge.launchpad.net/%7Esarvatt/+archive/ppa/+sourcepub/643434/+listing-archive-extra or disable compiz until its updated [23:17] why is it a compiz problem? [23:17] ah, not a compiz probelm; compiz-triggered [23:17] yeah compiz triggered, sorry [23:17] is ok; just saw a compiz update come through [23:19] any idea when it'll get fixed in the ppa or karmic? [23:19] just linked a PPA with the fix in it :) [23:24] it's not the ubuntu-x-swat ppa [23:25] * solarion is leery of installing from unknown repos [23:25] the one i linked is drop in replaceable with karmic, its also in xorg-edgers but you'd have to upgrade drm too to use that one [23:25] ah understandable [23:25] is there any point to having the jaunty x-swat repo in? [23:26] it's too bad you can't limit what can come out of a repo [23:26] the best I can figure atm is not including the keys and then watching what comes unsigned, but it's unsigned.