[00:24] oh no, more ash clouds.. [00:28] really? [00:33] Sarvatt, got a link? only stuff I see shows it mostly just affecting Ireland [00:34] bryceh: http://news.yahoo.com/s/ap/20100507/ap_on_bi_ge/eu_iceland_volcano [00:34] great timing huh [00:36] yep [00:38] well maybe they can fly around [00:38] my guess is they'll be able to get us into europe but coming home will be a different story [00:39] just found out the hotel was overbooked too, now I'm really getting worried about going. been working my butt off for 3 weeks to afford just the one week off of work, i'm really screwed if i get stuck there [00:40] if you end up stuck in paris we can get some $beverage :) [00:44] i dont mind a longer flight there honestly, was arriving at 7:45 am and can't check in until 3 pm anyhow :) [00:52] Sarvatt, I wouldn't worry about it, there's not anything we can do, and chances are that it'll probably all work out fine [00:53] I seem to recall there were various potential disasters looming around the past couple UDS's that never really materialized. Swine flu epidemics and so on [00:55] yeah i'm just nervous because i'm leaving in the morning and finding out a lot of things at the last minute like the hotel situation, it's no biggie :) [00:56] interesting, my flight goes up north over iceland, so appears to miss the plume entirely that way [00:56] how'd you check the way the flight went? [00:56] superm1, went to delta.com and typed in the flight #, and it showed a google map [00:56] http://www.delta.com/flifo/servlet/DeltaFlifo?airline_code=DL&flight_number=178&flight_date=Today&request=main [00:56] oh neat. i doubt AA would be as nice for me [00:57] I bet if you find a delta flight from/to the same cities, the airlines probably use the same flight paths [00:57] good call, i'll see [00:58] regarding the hotel switchup, yeah that happened last UDS too. Was annoying but worked out okay [00:58] well they can already tell me my flight is an hour late taking off [00:58] weird, that's like 24 hours from now [00:59] Sarvatt, sometimes they let you check in early, 3pm is sort of ballpark [01:00] Sarvatt, in a couple cases they've been so booked that people couldn't check in until after that. But it's not a huge deal, just hang out in the lobby with everyone else and have a beer or a nap [01:01] a bingo http://flightaware.com/live/flight/AAL88 [01:01] superm1, hey you're going north of iceland too :-) [01:02] yeah.. i'll cross my fingers this doesn't throw things off for me too then [01:11] funny, i'm going back the way you're coming [01:12] brussels - o'hare - reagan national [01:13] wouldnt it make more sense to go through jfk or something for that ? [01:13] feels like running half way across the country to ord feels out of the way [01:13] thats how i'm getting to there :) [01:14] ah :) [01:14] http://flightaware.com/live/flight/AAL172 [01:14] extra 500 miles going back but its only 40 minutes longer [01:15] weird === furan is now known as furan_oot [02:18] Updated for meerkat: http://www2.bryceharrington.org:8080/X/Reports/ubuntu-x-swat/versions-current.html [05:21] Please point me in the right direction if this is the wrong channel, i was pointed here. I am trying to setup my some defaults for my touchpad that aren't configurable via the mouse program in the system menu. I can configure these via synclient, but they do not last though a reboot or even suspend. [05:22] I believe this needs to be done in udev, and I have found the synaptics section, which seems to point to other files for specific netbooks. I would like to create one for this netbook but cant find those other files anywhere. I would appreciate any feedback or pointing me in the right direction [06:22] goor morning [06:22] good morning [06:27] as soon as I attach gbd to the xorg process and continue, xorg goes to 100% cpu and apparently hangs [06:27] starting xorg from inside gdb doesn't work, as the psb module won't load === furan_oot is now known as furan_nostream [06:29] I'm following https://wiki.ubuntu.com/X/Backtracing any suggestion on alternatives? I'm going to try crashing a generating a core, but being able to run xorg inside gdb would be more useful [07:11] warning: the debug information found in "/usr/lib/debug/usr/bin/Xorg" does not match "/usr/bin/Xorg" (CRC mismatch). [07:11] This on plain lucid [07:12] should I use a ppa to have valid debug symbols for Xorg? === Bernardo is now known as Bernardo|away === Bernardo|away is now known as Bernardo [11:18] hi lucazade [11:18] hi Bernardo [11:19] I'm fighting now xorg-core debug symbols.... Tried to build it on my netbook, but the resulting packages had some problem and it wouldn't start dri (again) [11:20] and the ones in ubuntu have some problem with the debug information, CRC mismatch [11:20] too bad, dri issue again! i've read your post on forum [11:21] is there a bug for the crc issue on launchpad [11:21] ? [11:21] without that information and with gdb hanging, I'm finding it hard to find where does it die, what parameters are being passed wrong [11:21] haven't checked yet [11:21] lucazade: yes [11:21] jcristau: thanks [11:22] jcristau: do you have the bug #? [11:22] not off hand, trying to look it up.. [11:23] just googled for it, I think I found it, thanks anyway [11:23] possibly 423748 [11:25] but i think there was another one [11:25] I'm looking at 562418 right now [11:26] yeah that's the one [11:28] Unfortunately that one states we'll have to wait for the new packages to show up on updates... And since building localy seems to break something on my config, I'll have to wait or find a ppa with working packages [11:31] yes, i believe we should wait for a fix [11:34] Bernardo, what package has this problem? [12:05] ricotz: xorg-xserver-core [12:05] I mean xserver-xorg-core [12:12] Bernardo, you are speaking of the gdb symbol crc mismatch? you shouldnt have this problem using a ppa package of xorg-xserver [12:13] ricotz: yes, that is the one. I'm trying now with a ppa package, bryceh's "purple" ppa, which should also have some exa related fix [12:13] you are using the xorg-edgers package? [12:15] not that one, but maybe I should. Bryce Harrington had posted a ppa to try and solve a problem with miDoCopy and the ati proprietary driver [12:15] as the problems we have with psb seem to point also to miDoCopy, I decided to try that one, but it still crashes [12:16] and as I just fond out, the backtrace is also very inconclusive, maybe I need to reboot the machine and test again [12:18] Bernardo, this crc mismatch bug was solved on 10-05-04 [12:19] Bernardo, so try the xorg-edgers/ppa package, which was built more recently [12:20] https://bugs.edge.launchpad.net/ubuntu/+source/xorg-server/+bug/562418 [12:20] Launchpad bug 562418 in pkg-create-dbgsym "empty ddebs when dh_strip is called twice" [High,Fix released] [12:21] just wondering if close to us issue (even if is about uxa) http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=5009127de7d9527ae329d1c2fbc7283648bde2e6 [12:21] *our issue [12:23] ricotz: that was the bug I was looking at [12:23] but for our issue with miDoCopy I was looking here: [12:23] https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/539772 [12:23] bbl [12:23] Launchpad bug 539772 in xorg-server "Lucid 2.6.32-16 crashed to login screen - miCopyRegion" [High,Fix committed] === Bernardo is now known as Bernardo|away [13:20] hah, looks like i'm on your flight superm1, they shifted me around.. [13:46] superm1: apparently AA88 is canceled now too? [13:47] hope he isn't already on the plane to chicago :( [14:04] Sarvatt, i just got the email that 88 was canceled too... [14:04] i'm supposed to head to chicago in like 2 or 3 hours [14:04] same here [14:04] so i'm on hold right now to figure out's up [14:04] what'd they do for you? [14:05] nothing yet, they just switched me to AA88 since the one out of NY got cancelled and they haven't caught up, scott ritchie just said AA88 was canceled in #ubuntu-devel though and said they booked him on a flight for tuesday [14:05] but he's taking a flight to paris and taking the train in instead [14:05] oh man [14:06] did AA pay for the train for him? [14:06] i've only got about $700 in the bank so even if it's paid back later I can't afford to change things like that [14:06] don't know [14:08] okay just got through [14:08] they rebooked me for a flight into heathrow tomorrow which goes into bruseels afterward [14:09] airports in the north of Spain & Portugal and in the south of France have been closed today [14:46] superm1: I'm on flight 88 tomorrow [14:47] I'm an idiot and thought it was today they switched me to that was canceled, but it's tomorrow flight 88 (which will probably be canceled too..) [14:49] oh lucky though to be on 88 tomorrow, that was full for me :) === Bernardo|away is now known as Bernardo [15:00] hi [15:01] what time do you get in monday? 8 am here, going to miss some of the early stuff [15:08] ricotz: I can't use the xorg-edgers ppa - it is based on xorg 1.8, not 1.7 [15:09] anyway, brycehs purple ppa didn't add much of debug symbols, the backtrace is still inconclusive [15:10] Bernardo: what problem are you having? [15:11] Sarvatt: the missing debug symbols in xorg are making very difficult to find out why the psb driver is crashing in miDoCopy [15:12] they aren't missing if you're using the PPA one though [15:12] Bernardo: are you dumping a core to play with or just looking at the backtrace in the log? [15:13] I've been unable to dump a core [15:13] gdb makes xorg take 100% cpu so I've also been unlucky there [15:13] you may want to do this from a VT as root - gdb -p $(pidof X) -batch -ex 'handle all nostop' -ex 'handle all pass' -ex 'handle 11 stop' -ex 'cont' -ex 'bt full' -ex 'cont' [15:13] ok, let me see if that won't hang as it is hanging now [15:14] probably because its stuck in a signal waiting for you to continue, are you using gdb from another pc? [15:14] yes [15:14] over ssh [15:14] Sarvatt, 11 am :( [15:14] assuming no other problems arise [15:14] oh yeah you want to redirect that to a log if you dont have gdb logging by default too [15:15] well, it doesn't help also, 100%cpu and xorg is hung. And gdb doesn't answer ctrl-c [15:19] Sarvatt: which ppa would you recommend I'd try for xorg? [15:20] bryce's [15:21] i thought that was in lucid-proposed already but he's probably waiting until after UDS to upload it [15:22] Well, I'm already using one of his PPAs [15:22] and I just found that at least apport seems to have created something in /var/crash [15:23] really? apport actually caught the crash?! [15:27] 8 hours ago, I enabled apport while I was trying to play with gdb, so I don't even know it is valid or it was generated one of the times I killed gdb and X as they had hung [15:27] can ya file a bug based on it? [15:28] possibly, but I'd like to validate it first, is there any way I can check it caught the crash or something unrelated? [15:29] well, opening it seems to have caught something like the backtrace we are seeing so it might be valid [15:30] apport-bug /var/crash/whatever.crash will show you the info before submitting it [15:30] you can use apport-unpack too [15:32] Bernardo: where's the most current packaging you guys are using so I can take a look at it since I have a bit of free time thanks to my flight getting delayed? [15:33] I've dumped my binaries in the google code, let me update that, it will take a couple minutes [15:34] don't want to reinvent the wheel if you have already done a lot of the work like it sounds like, but a bunch of things will need changing if you didn't do it already. xpsb-glx packaging needs a huge overhaul with all of the alternatives stuff and libdrm needs a bunch of changes as well [15:34] will give me something to do on the 8 hour flight :) [15:35] :) [15:35] yep, both of those need redoing [15:36] ok, the latest version of my binaries is already at http://code.google.com/p/gma500/ [15:36] inside bernardo-debs [15:36] ah, of course apport refuses to open a bug as xorg-core isn't the original ubuntu one [15:38] seems I have a core in the report, so I'll check that now [15:38] does APPORT_IGNORE_OBSOLETE_PACKAGES=1 apport-bug /var/crash/whatever work? [15:38] doubt it but it might [15:40] going to install libdrm-psb to another directory and fix things up so they can coexist, diverting libdrm2 stinks :) [15:40] I've opened the core with gdb and it didn't find symbols for psb_drv.o and Xpsb.so - I guess it would be expecting too much to hope for those having debug symbols [15:41] I know, it breaks the vesa driver, among other stuff [15:41] Bernardo: install pkg-create-dbgsym and build them locally [15:41] it'll create -dbgsym packages for them [15:41] it just doesn't automatically create debug packages for everything for ppa builds [15:43] do the rules for those two packages handle nostrip/noopt? [15:44] let me check, I had been hacking them to try to generate debug packages, but it seems I was unable to do it properly [15:44] just grep debian/rules for noopt [15:45] xpsb-glx handles noopt [15:45] ifneq (,$(findstring noopt,$(DEB_BUILD_OPTIONS))) [15:45] Bernardo: actually ignore what I said, just build with DEB_BUILD_OPTIONS="nostrip" debuild -uc -us -b [15:46] ok [15:46] that will ignore even the dh_strip in the rules file? [15:46] xserver-xorg-video-psb should only take a minute tops to compile but xpsb-glx will probably take ages :) [15:47] yeah [15:47] yes, with all the mesa stuff inside [15:47] does your backtrace involve mesa though? [15:47] no, only exa, fb, libpixmap [15:47] i'd be *really* surprised if xpsb-glx even works [15:48] its diverting libs that dont exist and not installing alternatives so it wouldn't be used anyway [15:49] i really want to redo all this packaging from scratch, ugh [15:49] not here, it is installing its own libGL, and only diverting the existing one [15:50] Bernardo: can you ldd /usr/bin/glxgears | pastebinit [15:50] It does need cleaning up, separating the binary crap into one or two packages, and all the open stuff around it needs to be easier to maintain [15:50] when does X crash? right after starting? [15:50] no, at least with kde [15:50] The way I can get it to crash reliably is to do a "top" in a konsole, the quit it [15:50] the other part of me doesn't want to redo it because it's a dead end anyway and is pretty much guaranteed it's not going to work with 10.10 :) [15:51] crashes before I can see the prompt [15:51] yes, xorg 1.8 [15:51] ah so can you do that ldd before it crashes? [15:51] i would love to see your Xorg.0.log and dmesg after booting before it crashes [15:51] sure [15:52] * Sarvatt wishes he had a psb machine, hah [15:52] my Xorg.0.org hasn't changed much from what I pasted on the ubuntuforums thread [15:52] what thread? [15:52] * Sarvatt doesn't go to the forums much [15:53] http://ubuntuforums.org/showthread.php?p=9259980 [15:54] http://pastebin.com/GrKJJ0Pc [15:55] dmesg is at slajhAxS [15:55] psb really is a dead end, intel has said they aren't going to update the drivers and are only supporting gma600/moorestown in meego [15:55] thanks Bernardo [15:55] so yeah you are using mesa and not xpsb-glx [15:55] xorg log at VLUmNSY5 [15:56] bbl, got to run now, wife is calling === Bernardo is now known as Bernardo|away [15:57] http://pastebin.com/slajhAxS doesn't exist [16:00] s1ajhAxS [16:00] sorry [16:00] bbl [16:10] thanks for putting the binaries in svn too, will be sure i have them all on the plane :D ok so i'll assume you guys have psb-kernel-source right and start with libdrm, that needs to install things to another directory, patch a few things to account for the different location and have ldconfig set up to find it. xpsb-glx needs to install libGL to another directory and set up alternatives for it with a higher priority than mesa, install libdri.so [16:10] and Xpsb.so which looks to be the libglx replacement to an extra modules directory with alternatives setting that to /usr/lib/xorg/extra-modules, psb_drv_video.so is installed somewhere its not getting used at all currently as well [16:10] the psb-modules package in psb-kernel-source isn't needed [16:13] doesn't look so bad, that looks like all thats needed [16:18] probably should patch xserver to load psb by default though, and i noticed it's only matching 1 psb pci id to skip intel and use vesa instead, it tries to load intel for people with the 8109 pci id [16:26] oh sheesh xpsb-glx is precompiled, you aren't going to get debug symbols for it Bernardo|away === Bernardo|away is now known as Bernardo [17:17] Sarvatt: psb_drv_video.so is used by libva, but I remember that in karmic I had to symlink it to another directory [17:18] xpsb-glx has the closed source crap [17:19] what directory? [17:19] did you have to symlink it to I mean [17:20] i dont get why its got this in install - dri/psb_drv_video.* usr/X11R6/lib/modules/dri [17:20] let me check [17:21] got it [17:21] /usr/lib/va/drivers/ [17:22] Yep - I found it here: http://ubuntuforums.org/showthread.php?t=1234502 [17:24] Do you think we still need the mesa from xpsb-glx? [17:26] wow, managed to capture a crash with gdb now [17:27] no, sorry [17:27] hang [17:49] hmm why does xpsb-glx even build libGL [17:50] gotta pick tseliot's brain at UDS, I think he added that because of a problem with jaunty's mesa [17:52] I remember that the first packages I used with karmic I had full video acceleration with kde's dragon player [17:52] and then xpsb-glx was installing libGL [17:53] later with another xpsb-glx I had to remove redirection as it didn't include libGL [17:53] don't remember where I got either [17:55] well all of the libgl stuff isn't doing you any good right now since it's not even used from what you showed me, and i dont see it even being installed by the other distros [17:56] right now things are so broken I didn't even think of it... :) [17:57] if you want to divert it you'd need to edit the maintainer scripts to divert libGL from /usr/lib/mesa/ instead of /usr/lib, but thats just a hack until alternatives would get added [17:57] yes [17:58] do you know if ubuntu's xpsb-glx is in git anywhere? [17:58] hi [17:59] hi lucazade [17:59] Sarvatt: I have no idea [17:59] Bernardo: have you tried the xorg.conf? [18:00] lucazade: not yet [18:00] I've been trying to extract meaning from the crashes [18:01] i'd imagine [18:01] i don't suppose anyone with a psb machine with lucid on it would be willing to give me ssh access to it? :D [18:02] it seems stable.. no crash yet [18:03] obviously is some sw-rendering :) [18:03] atm [18:03] My psb machine is beyond two routers... Playing with nat and port redirection is a bit of a mess, but I can do it (I think) [18:05] i have not played enough with xorg.conf so i don't know exactly *every* options used... need to investigate better [18:06] Bernardo: that'd be a big help, there is a lot of info I'd like to see from a running system so I'm sure things are right [18:35] Hi guys [19:28] i'll try to install karmic and drivers, after this i'll update all packages except xorg-server and their dependences while lucazade, jbernado and company fix the problem with psb [19:40] Bernardo: you may want to disable use_udev.dpatch from psb-kernel-source, it's already there in the other patches === BUGabundo is now known as BUGabundo_DrWho [20:23] woohoo can view it in a sane VCS now - http://sarvatt.com/git/cgit.cgi/gma500/ [20:40] Sarvatt: you're right, with vga16fb blacklisted we fb framebuffer works! [20:42] it's because psb.ko is in updates/ instead of drm/ [20:43] interesting! [20:43] that and vga16fb is just the devil [20:43] ok [20:43] so we need to change that in psb-kernel-source [20:44] I already tried, but it still got installed to updates by dkms [20:44] i'm not sure its a good idea to install things from dkms outside of updates/ ? [20:45] I'm not sure if it is possible... :) Why should it be a bad idea? It is a module, built with the same kernel headers as the other modules distributed with the kernel [20:47] well mandriva does it, guess it's ok :) [20:47] lucazade: your xorg.conf means I can't crash the system (at least as I was able to crash it until now) [20:48] i wasn't sure if it screws up things like if the kernel gets removed before dkms deletes the module [20:48] but i guess it's ok, was just being cautious :) [20:48] Bernardo no crash until now.. stable [20:49] lucazade: with your xorg.conf gl gets broken, but apart from that we seem to have 2d working. [20:49] yes [20:49] now we need to clean up everythinh [20:49] is all about: Option "ShadowFB" "true" [20:49] i've updated the xorg.conf on forum [20:49] Sarvatt: I've had situations already when the kernel dir doesn't get removed because of non-empty dirs [20:49] so it will still happen [20:50] yeah same here [20:50] but those were all kernels I made with make-kpkg as far as I can remember [20:50] I had that even with "regular" kernels [20:51] /dev/dri/card0 wasn't in the video group on your system btw Bernardo [20:52] and still isn't [20:52] another thing to check [20:53] might as well drop the psb-module package from psb-kernel-source, its kind of useless [20:53] that'll fi [20:53] err [20:53] fix one of the errors i saw you guys having with the build [20:54] trying to write to extras? [20:54] installing to the PPA's kernel [20:56] Bernardo: blacklisted vga16fb... ok [20:56] just need to comment out these lines in rules # $(MAKE) LINUXDIR=/lib/modules/$(KERNEL_VER)/build DRM_MODULES="psb" [20:56] # mkdir -p $(CURDIR)/debian/psb-modules/lib/modules/$(KERNEL_VER)/updates/char/drm [20:56] # cp psb.ko drm-psb.ko $(CURDIR)/debian/psb-modules/lib/modules/$(KERNEL_VER)/updates/char/drm [20:56] and the psb-modules package section in control [20:56] and i think there were some maintainer scripts for it but i deleted them if so and dont remember [20:57] lol [20:57] i've to go... i'll check for updates if any! bye [20:57] I was commenting them now and wondering if they were the right ones [20:57] bye [20:57] will go soon too [20:58] I also dropped the psb-kernel-headers from control [20:58] all that is in libdrm-poulsbo1 now [21:02] dkms.conf.in specifies the modules should go in gpu/drm [21:02] looks like dkms doesn't honour that [21:06] damn svn [21:09] bye, have to go now === Bernardo is now known as Bernardo|away [21:12] weird - http://netbook-remix.archive.canonical.com/updates/pool/public/libd/libdrm/ [21:13] * Sarvatt wonders what the heck that repo is and how his packages got there :D [21:13] lol [21:13] :) I'm still here, fighting with the dkms error "can't find source dir" [21:15] No idea why that happens and why I have to remove psb-kernel-source and reinstall to be able to build the modules again [21:16] really gone now [21:56] Sarvatt, there are some lucid patches relating to vaapi hardware acceleration for intel here: http://www.splitted-desktop.com/~gbeauchesne/libva/ironlake.patches/ubuntu.lucid/ [21:56] they're meant to patch the kernel module and libdrm [21:56] yeah I know, they've been going around the lists for months [21:57] really [21:57] i wondered if they'd been applied to the xorg-edgers ppa [21:57] nope [21:58] ain't that a kick in the pants [21:59] not trying to regress things I care about and they have been getting rewritten just about weekly for months.. when it goes upstream it'll be there :D [22:00] i think the kernel patches are in drm-intel-next now though? [22:00] nope doesn't look like it [22:02] ah looks like they keep posting them based off of released kernels and it doesn't apply to drm-intel-next where it'll actually go [22:02] well, i think it's important because i'm sure adobe is going to provide vaapi to flash 10.1 [22:02] so they haven't been accepted yet [22:03] wait, no they aren't. that's crazy [22:03] pigs are flying now? [22:03] haha [22:04] what i meant to say is i'm sure adobe will continue to whine about not being able to fix flash [22:04] theres already like 80 packages in xorg-edgers now, would like to keep it to X/mesa stuff anyhow :) [22:05] well, these are x driver patches. they have to go somewhere [22:05] maybe the graphics driver shouldn't be in the kernel [22:26] it's just one libdrm patch i'd be adding whenever it hits drm-intel-next so people can use the mainline drm-intel-next kernel with it