[00:01] bryce: doubt there will be any difference in it under KMS if thats what you're after, the KMS support is in jglisse's xserver-xorg-video-ati git tree seperate from the main one :( you can grab the drm 2.4.4 and -ati from here and build them locally if you want to test it out though, i dont know if tormod got the 2.4.11 patch working but if so that'd let you build radeon-rewrite thats on here also https://edge.launchpad.net/~sarvatt/+a [00:01] rchive/ati [00:02] Sarvatt: no, this is just to resolve a few bug reports [00:02] oh gotcha, sorry [00:02] clean up patches, &tc. so when we pull a git snapshot it'll be less involved [00:16] just looked and all patches (besides 01_gen_pci_ids.diff of course) are upstream so that'll help :D [00:17] well everything here at least http://patches.ubuntu.com/x/xserver-xorg-video-ati/extracted/ [00:18] in xf86-video-ati master that is [00:25] the exa patch was not in the 6.12 branch so I've kept it [00:25] I see debian dropped 01_gen_pci_ids.diff [00:26] I'm going to test once without that to see if it works without it. I'm guessing we can only drop it once we've revved xserver up [00:26] yep works [00:26] hmmm [00:27] maybe I had a pre-existing pci file tho [00:28] I'm going to chicken out and leave patch 01 in for now [00:28] it should work, that default to vesa patch was screwing things up in that regard while you were on vacation though [00:29] yeah i was doing that too for awhile on edgers after debian dropped it from everything :D [00:29] when was I on vacation? [00:30] oh, allhands/uds right [00:35] when we got updated to 2.7.1 on intel a few weeks ago, xserver wasn't working right with the default to vesa patch and not having the pciids so alot of people were getting shot to vesa. hmm just noticed debian added two new patches regarding pci device detection in xorg-server since our 2:1.6.1.901-2ubuntu1 [01:23] well that was fun. We're having a major thunderstorm, and a power dip took out my desktop (WTF UPS?) then fsck decided there was a problem with my disk, ...... [01:24] how my sister's cat deals with the heat: http://www2.bryceharrington.org:8080/Photos/Cats/Leah/ [01:32] lol! [01:35] smart cat :) mine run for their lives away from the tub [01:39] anyone have a rough idea of what would be lost using --disable-config-dbus for xserver? 1.7 builds fine disabling it and I noticed debian disables but ubuntu enables it. [02:51] seems to work fine and even loaded up a little faster, strange [07:49] tomod yeah i know it has a bad changelog. it wasn't worth the 20 minute upload time to fix it as it was meant for bryce t touch test. i'll make sure its documented correctly for the next one :) [07:52] heya, good morning === jcristau_ is now known as jcristau === halfline_ is now known as halfline [15:54] * apw waves to bryce ... got any nvidia hardware? [16:12] apw: indeed [16:13] cool. i am just getting some test builds of a kerenl with nouveau in it [17:26] apw: where is the kernel development for that at? I could test that out on a few machines as well even though i'm really happy with the binary driver on those [17:51] apw's kernel requires newer libdrm-nouveau1 and ddx than whats on edgers to use though (api bumped this morning). i dont know much about how the nouveau packages are set up, will try to get those updated on edgers today if noone else does it :) [18:12] Sarvatt: cool thanks [18:12] I'm setting up a box to test the nouveau kms today [18:15] Sarvatt, more new behaviour: this last time I tried suspend, it came with everything working except that the mouse cursor is invisible, but it still interacts with the desktop [18:15] now thats odd [18:16] bryce, re: https://bugs.freedesktop.org/show_bug.cgi?id=22039 , how useful is it to add additional intel-gpu-dump's when I see 'different' broken behaviour? [18:16] Freedesktop bug 22039 in Driver/intel "Xorg hangs after resume from suspend when KMS enabled on GM945" [Major,New] [18:16] I mean, "symptom of the day" effect I have probably means it's just random corruption of some sort, but is it still useful to have more raw data? [18:17] cwillu: it can be useful, however if it's at all likely to be a different bug, it can be better to file a new bug [18:17] yes more data is usually useful [18:17] anything interesting in /var/log/pm-suspend.log cwillu? [18:17] it's like a 1 time in ten that suspend works, although there's always something else broken [18:17] oooo, mouse cursor just came back [18:18] ...and moments later X locked up [18:18] one sec, I'll check [18:21] success on every line [18:23] Sarvatt, ^^^ [18:25] ok drm for nouveau is up on edgers, gotta look into the ddx since it needed some hook changes last time i tried it [18:29] Sarvatt: shall we set up a 'Nouveau crack - kernel modesetting' PPA like was done for Radeon? [18:30] oh also I wanted to ask how you're doing the drm-snapshot packages, and vs. rolling them as 'libdrm' package snapshots? [18:32] * cwillu didn't think to check hibernation before [18:32] 5 hibernation cycles so far, X hasn't hung yet. [18:33] -ati 6.12.2-2ubuntu1 uploaded. [18:36] bryce, a nouveau crack PPA would be handy as i could then not make a combined kernel for testing and would be easier to manage. sounds like you have management issues trying to shoehorn both into the same package too (ati and nouveau) [18:36] shouldnt be needed as far as I can see, nouveau works straight on mesa/drm master vs ati having a crapload of specific branches for different things that arent compatable with each other and need different DDX depending which you go off of.. but I think we might have to package gallium for 3d support for nouveau in mesa, i havent looked into that at all yet [18:37] well let me know if you have any luck with my kernel, if so i have a strategy for a combo job there [18:39] might actually be able to enable KMS with the nouveau module in drm-snapshot.. [18:43] ok, guess we can always add it later if we absolutely need it [18:43] shame you can't *delete* ppa's [19:26] bryce, yeah i could do with a purge too [19:26] bryce: you can, by poking the launchpad people [19:28] hyperair: what I have in mind is being able to set up quickie ppa's for specific short-term objectives (like this kms stuff) and then destroy them once that experiment is complete or merged in [19:28] bryce: yeah, that would be really awesome. [19:28] hyperair: bugging the launchpad guys each time an experiment is complete would probably annoy them too much ;-) [19:28] agreed =p [19:28] well just reuse them ppas [19:29] but it's a good point; maybe once a lot of cruft has built up they could be pinged to do cleanup [19:29] delete all the packages [19:29] * bryce nods [19:29] I've been using that approach with some of my personal ppa's. I ended up just giving them abstract names so I could use them for multiple purposes... 'blue', 'green', etc. [19:30] but that's a pretty ugly hack ;-) [19:32] bryce: if you annoy them enough they'll end up implementing it [19:32] lol [19:32] =p [19:32] i have one called "bugfix". i just dump anything and everything there. [19:33] i even have a prototype uswsusp package with usplash support [19:33] i just can't figure out how to make uswsusp *not* hang when resuming with usplash [19:33] it seems alt+sysrq+e is a quick workaround to the hang [19:33] probably usplash hanging [19:33] ah well [19:39] yeah I just started naming PPA's I add with more generic names so I can wipe it all and reuse it for something else, they are starting to add up lol [19:41] heya [19:42] do you need some report about kms working with intel? [19:43] i think there's one somewhere around.. [19:43] in the wiki [19:43] say... i haven't noticed the cursor corruption for a while O_o [19:43] https://wiki.ubuntu.com/X/KernelModeSetting [19:43] could it be that it's fixed? O_O [19:44] not here [19:44] ya do something different recently? [19:44] hmm wait [19:44] is there a sure fire way to reproduce the cursor corruption thing? [19:45] i havent ever seen an animated cursor come up and not get glitchs within 2 seconds :D [19:45] my resize cursor is animated [19:45] no glitches [19:45] =\ [19:45] its funny, the glitches I get with the cursor actually happen to the whole screen sporadically on windows though [19:46] is UXA activated in Karmic, or should I keep my configuration file? [19:46] hmm mine was confined [19:46] Declares that file is expected in the directory defined above. In Autoconf proper, this macro does nothing: its sole purpose is to be traced by third-party tools to produce a list of expected auxiliary files. For instance it is called by macros like AC_PROG_INSTALL (see Particular Programs) or AC_CANONICAL_BUILD (see Canonicalizing) to register the auxiliary files they need. [19:47] oh hell [19:47] #$%^&*( compiz [19:47] wonder if theres something related to how it displays the cursor and how windows display video overlays [19:47] my super+middleclick ends up middle clicking into urxvt. [19:48] keep your configuration file crevette, it is activated by default but also disabled by default (confusing huh) [19:48] uh? [19:49] when you build the kernel you get an option to enable KMS by default or not, enabling it by default links all the other modules in the correct order with it, ubuntu enables it by default there so those things work right but makes the default modeset option for the i915 module 0 so its off [19:50] doing it that way makes it work properly by just adding it to a modprobe option in /etc/modprobe.d/ [19:53] oh shoot crevette, I misread your question as "is KMS activated" lol [19:54] sheesh [19:54] there's only UXA in karmic right now, dont need to keep your config [19:54] okay sweet [19:55] thanks Sarvatt [19:55] no worries, sorry about the confusion :D [19:56] "(note that UXA is enabled by default). " /me should have read better the wiki [20:03] does anyone know of a reason why xserver is built with --enable-conf-dbus on ubuntu and what might be hurt disabling it? It's defaulted to no upstream, disabled in debian and moblin and seems to boot a little faster disabled. I had to disable it for xserver 1.7 to build post xi2 merge and I can't find anything broken by it [20:04] Sarvatt: some hack involving wacom [20:04] before wacom got fixed to work with hal [20:04] ahh I see [20:10] thanks jcristau [20:11] I have the message paste in http://dpaste.com/51959/ in the Xorg log, I know this is armless. Should I submit this to any upstream / downstream BTS? [20:11] submit what? [20:13] the EDID [20:20] probably not, if you don't have a problem with it? [20:21] no, no, I was just reading my xorg log file, to check there is no error after kms activation [20:21] my point is just if yo see a warning, report it :) [20:30] * crevette tries suspend [20:48] ugh. what's this la_AU locale that got added to ubuntu libx11? [20:55] Sarvatt: just hibernated and resumed. no cursor corruption. [20:55] Sarvatt: however, i had corruption on my terminal upon resuming [20:56] Sarvatt: as in, rxvt, which was configured to use argb. [20:56] resizing the window solved the issue [20:57] unrelated but just curious, are you using the older mesa with those patches or the newer one hyperair? [20:57] wondering if removing the patches fixed the problem you bisected down [20:57] Sarvatt: custom compiled mesa, with the commit 700something reverted. [20:57] the patches are in place [20:58] ah [20:58] but i reverted the commit that caused my crash [20:58] i'm using the rc8 kernel from kernel.ubutnu.com [20:58] ubuntu* [20:58] why not karmic's 2.6.30-8.9? [20:58] er. [20:58] is it the same? [20:58] its rc8 yeah [20:58] bryce, Sarvatt, those kernels built finally [20:59] its not the same. the rc8 mainline kernel has no ubuntu delta [20:59] except its got all the ubuntu patches and changes instead of being based on jaunty and dropping the cust...... apw beat me to it [20:59] i see [20:59] by same i meant rc8 based, sorry [21:00] i'm going to try using the -8 kernel then [21:00] off i go to reboot. [21:00] yeah its mostly the same for sure, karmic's delta is smaller [21:07] ooh. -8-generic has no issues either! [21:07] cursor's 100% okay [21:07] hyperair: yeah it's running fine here [21:08] woohoo [21:08] =D [21:09] Sarvatt: 086ecea179ed572c89aa77c5f465671a5cef87a7 <-- my mesa is of this particular snapshot, with your patches enabled, and commit 7f8000db8bd45bb95bda4a4f8535c49b8ef74254 reverted. [21:09] Sarvatt: why did you drop the uxamemleak patches though =\ [21:09] Sarvatt: won't that cause the uxa memory leak to regress back to its previous state? [21:27] if its the right solution it'll be upstream soon, rather not possibly cause problems for everyone using it in that PPA :) [21:28] hmm [21:28] why wait? =p [21:29] is the commit i reverted that important, anyway? [22:02] heyo tormod [22:02] heya sarvatt [22:03] did they fix the buildds? [22:03] yup [22:17] guess i'll fudge the date to switch to libdrm over drm-snapshot, failing to upload because of the stale older version i deleted 4 hours ago [22:18] its 0606 somewhere in the world :D [22:24] Sarvatt: I sometimes jump timezones to get the right datestamp :) [22:25] but more often the other way, to be proactive [23:56] ugh, mesa failed to build in general just now [23:56] what the hell, it built fine in my xorg-testing PPA [23:57] /usr/bin/install: cannot change permissions of `/build/buildd/mesa-7.6.0~git20090605.b7aa5b1d/debian/tmp/usr/include/GL/glxext.h': No such file or directory [23:57] its happening on all arches now [23:58] maybe its the new mktemp from coreutils having problems