[00:29] byebye klingon? :D [00:30] * Sarvatt slaps forehead [00:30] mistook the libXi sync for libX11 :) [00:36] RAOF: savage DRI sucks huh? [00:36] Sarvatt: Looks like it :) [00:37] That said, in hindsight I should probably have pushed that bug on the DDX; running with swrast kills X. [00:37] every time I see bug mail from you for desktop app crashes it's always savage DRI :D [00:37] Yes! :{ [00:46] the crashes fixed by http://lists.x.org/archives/xorg-devel/2010-March/006436.html seem to be triggered by kmail mostly [00:46] https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/401045 [00:46] Ubuntu bug 401045 in xserver-xorg-video-ati "X crashes in FindGlyphByHash" [Undecided,Confirmed] [00:46] https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/442144 [00:46] Ubuntu bug 442144 in xserver-xorg-video-intel "Xorg crashes with SIG 11 in FindGlyphRef" [Undecided,Incomplete] [00:46] https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-180/+bug/479031 [00:46] Ubuntu bug 479031 in nvidia-graphics-drivers-180 "Ubuntu 9.10 Crash in /usr/bin/X(FindGlyphRef+0x2c) [0x5277cc] on nvidia nvidia-glx-185, SMP quadcore" [Undecided,Confirmed] [00:46] lots :) [00:48] will give it a few days and see if anyone reviews it before building test packages for people, not sure if its correct. the commit its referencing is from 2007 so its been around awhile [01:35] ok, updated the wiki some more with these new symptoms [01:38] https://wiki.ubuntu.com/X/Bugs that is [02:22] Balls. It looks like just *loading* vga16fb can annoy the precious flower that in nouveau. [03:10] huh RAOF? vga16fb isn't loading at all in that guys blacklist=vga16fb dmesg [03:11] Right. And in that dmesg you'll note that nouveau complains a lot less? :) [03:11] oh i didnt look at the one with it, the blacklisted one still locks up as soon as it starts [03:11] * Sarvatt has WAY too many tabs open [03:12] :) [03:12] https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/447159 [03:12] Ubuntu bug 447159 in xserver-xorg-video-intel "[i945gm] Xorg assert failure: X: ../../src/i830_batchbuffer.h:79: intel_batch_emit_dword: Assertion `pI830->batch_ptr != ((void *)0)' failed." [Medium,Confirmed] [03:12] needs some serious love [03:12] * Sarvatt todo's [03:14] RAOF whats the bug# for the nouveau guy? [03:14] i just had the dmesg tab open [03:14] nevermind I see it in #nouveau [03:14] 542950 [03:14] Yah. [03:15] hmm, first though, that an AGP card? [03:15] * Sarvatt looks [03:16] oh 6100 IGP.. [03:16] * Sarvatt has a 6150IGP booted up right next to me with nouveau [03:18] sure wish it was easier to see what commits we have in our kernel [03:19] I recently learnt that “git log nouveau/master...” is git for “the commits in nouveau/master that aren't in your current HEAD” [03:21] all of nouveau was added in one commit [03:21] That's quite true. Why was that command useful to me again?... [03:23] Oh! Because we know that lucid has drm from 2.6.33, so you can find the commits in nouveau/master that aren't in 2.6.33 [03:24] Acutally, we now have drm from 2.6.33.1 I believe, but there weren't any nouveau commits of interest. [03:26] dont know what i'm gonna do with lbm-nouveau now that its 2.6.34 [03:27] theres more changes that touch nouveau outside of drm, some acpi change [03:27] How much does drivers/gpu/drm interact with stuff that's out of that tree? [03:28] Ah. [03:28] Maybe it's time to switch to providing full-on kernels? [03:31] That shouldn't be *that* much more effort than updating lbm-nouveau. [03:32] “lshw causes display to freak out”. That's my sort of bug title. [03:48] yeaaah drivers/gpu/drm/nouveau/nouveau_irq.c where the messages are happening is a pretty rough read [03:49] thats happening for alot of people on intel right now for some reason [03:49] the lshw -C video making the screen trippy, doesn't happen here [04:02] I think there are rather a lot of duplicates of bug #535640 wandering in at the moment. [04:02] Launchpad bug 535640 in xserver-xorg-video-intel "[gm45] GPU lockup de05bf80bf83cd22541cb55f1a2ee99e (xorg crash when opening the laptop lid)" [Unknown,Confirmed] https://launchpad.net/bugs/535640 [04:18] yeah which is odd since we had a patch to fix that in -15 [04:19] something tells me it got dropped with the drm backport and its not in there [04:20] is EIR: 00000000 in those? batchbuffer I/O errors? [04:20] * Sarvatt looks [04:20] [ 74.800222] render error detected, EIR: 0x00000000 [04:20] [ 74.800252] [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -5 (awaiting 2128 at 2125) [04:20] yup we are missing the patch, looking into it [04:21] it wasn't upstream yet and I think it got dropped [04:24] i'll ping smb since apw is on vacation, we did drop the patch [04:33] that affected *alot* of machines, theres probably dozens of other dupes of that now [04:37] Yup. I'm wandering through them. [04:37] * RAOF sings the “...upon closing the laptop lid...” song. [04:47] ...and that's the last of the intel bugmail. Has the nouveaufixes kernel built yet? :) [04:50] nouveaufixes kernel? [04:52] Just pulling a couple of patches from nouveau/linux-2.6 [04:53] Particularly: mark LVDS as unknown if lid-closed, the slight monitor timing fix for VGA outputs, and fixing framebuffer fonts with width%8 != 0 [04:54] I'm going to guess that bug #540017 is *also* that missing patch :) [04:54] Launchpad bug 540017 in xserver-xorg-video-intel "i915 crashes after resume (not powersave mode problem)" [Undecided,Incomplete] https://launchpad.net/bugs/540017 [04:57] theres a few grctx update commits too [04:57] yeah but that guy is on a 945, i haven't seen that issue on <965 so I didn't dupe it yet but it does look the same [04:58] Yeah. [05:11] Sarvatt: Have you seen a number of intel bugs where the drm bails because agpgart loads *after* drm tries to initialise? [05:13] no actually, I haven't come across any of those yet in ubuntu [05:14] bug #542251 seems to be one, but I think I saw some more during my mail-purge. [05:14] Launchpad bug 542251 in xserver-xorg-video-intel "Compiz broken in Lucid Beta 1" [Undecided,Confirmed] https://launchpad.net/bugs/542251 [05:14] I wonder if it's reproducible? [05:17] with these 5 second SSD boots I bet it is a big problem.. [05:24] RAOF: if you see an error like agp loading after drm in dmesg would ya mind pasting that chunk into the bug description so its searchable? [05:25] Ok. [05:26] helps *alot* later on finding dupes like https://bugs.edge.launchpad.net/ubuntu/+bugs?field.searchtext=*ERROR*+Cannot+initialize+the+agpgart+module&orderby=-importance&search=Search&field.status:list=NEW&field.status:list=INCOMPLETE_WITH_RESPONSE&field.status:list=INCOMPLETE_WITHOUT_RESPONSE&field.status:list=CONFIRMED&field.status:list=TRIAGED&field.status:list=INPROGRESS&field.status:list=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field. [05:26] omit_dupes=on&field.has_patch=&field.has_no_package= [05:27] err [05:27] https://bugs.edge.launchpad.net/ubuntu/+bugs?field.searchtext=*ERROR*+Cannot+initialize+the+agpgart+module [05:28] upstream people really want all agp modules built in :( [05:29] Really? Why? [05:31] i'd have to dig through mails to find reasons, i've just seen it brought up by the radeon and intel people a ton of times over the past year [05:32] think there was a recent thread on intel-gfx about it, lets see.. [05:37] here we go [05:37] http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=1f7a6e372e9cb4d749f34c0738d832e6cadb4071 [05:37] Ok, so that's probably not in our tree. [05:37] nope was post .33 [05:38] thought i saw it recently [05:38] hmm it looks like it just fails to load if agp isn't already loaded, not sure that it tries again? [05:39] It uses the magic of linkage. [05:40] You'll notice that intel-agp.c exports the symbol that i915 consumes? That should ensure that intel-agp is loaded before i915. [05:41] At least, that's my understanding of it. By explicitly depending on a symbol in intel-agp, i915 will be guaranteed to load after intel-agp, which will have done its init. [05:42] Are you going to attach that patch to the bug? [05:42] ok so thats 3 patches now I'll see if stable will take [05:42] yepyep [06:00] * RAOF :note to self. Don't try to rsync /proc [06:13] airlied mentioned it once that he's probably going to make the agp modules mandatory (built-in) [06:15] It seems like a rich seam of bugs is possible if drm only depends on the agp core, which won't actually provide the necessary stuff until the platform-specific module is loaded. [06:28] i've seen it happen on radeon too and they just say agp should be built into the kernel [06:29] sheesh so many dupes for that 965 hang [06:32] at least its super easy to spot, EIR is 0000000 and do_wait_request returns -5 (which is -EIO) in current dmesg [06:33] found a bunch more dupes for the agp one too [06:53] bryceh, hello, still around? [07:20] thats a new one - intel_gpu_dump: page allocation failure. order:8, mode:0x40d0 [07:21] getting close to hitting 50 dupes tonight on https://launchpad.net/bugs/535640 :D [07:21] Ubuntu bug 535640 in xserver-xorg-video-intel "[gm45] GPU lockup de05bf80bf83cd22541cb55f1a2ee99e (xorg crash when opening the laptop lid)" [Unknown,Confirmed] [07:22] oh only 27 so far, i was going by the total bug count on -intel and someone else must be working it :D [07:46] ahhh so these two bugs are intertwined [07:47] we've got DPMS events hanging the GPU, then batchbuffer I/O errors, then in comes the "X: ../../src/i830_batchbuffer.h:79: intel_batch_emit_dword: Assertion `pI830->batch_ptr != ((void *)0)' failed." from the other bug with tons of dupes [07:49] found one of the dupes that showed both bugs happening in one log and it all makes sense now :) http://launchpadlibrarian.net/41051827/GdmLog2.txt [09:01] Sarvatt: ewww order 8 allocation? that's... bad. [09:06] Sarvatt: I've wandered through and duped a bunch of bugs that one. :) [09:13] ara, hi [10:47] could you have a look at bug 544781 . nvidia-96 failed to upgrade from 9.10 to 10.04. Thanks. [10:47] Launchpad bug 544781 in nvidia-graphics-drivers-96 "nvidia-96 96.43.14-0ubuntu11 failed to upgrade from 9.10 to 10.04: Error! Application of patch fall_back_on_mtrr_if_no_pat.patch failed." [High,Incomplete] https://launchpad.net/bugs/544781 [10:51] Well, that's pretty cool. We've managed to russtle up a guy with some hardware where the -nv driver can only drive the LVDS and the nvidia driver can only drive an external monitor. [10:53] %-) [12:14] * tormod curses xchat-gnome [12:43] tormod, just use xchat [12:44] xchat-gnome is notable for being functionally challenged [12:46] bjsnider, thanks I'll try it, this time connections were acting up after resume [15:59] bryceh: So, any ideas what is causing gnome-session to flip my screen (Bug #544813). Any testing/troubleshooting you would like me to preform? [15:59] Launchpad bug 544813 in xserver-xorg-video-intel "[Lucid Beta] after gdm, screen is backwards, upside down." [Undecided,Confirmed] https://launchpad.net/bugs/544813 [16:00] komputes: bogus stuff in ~/.config/monitors.xml (or whatever it's called) [16:01] jcristau: would you like me to remoce and test again (I think I have already tried this) [16:01] create a new user and try with it [16:02] jcristau: the monitor is not flipped upside down as gnome-display-properties proposes it, if you look carefully, the items are flipped inside out [16:02] tjaalton: I have done this as well, but will try it again now [16:03] nice effect [16:05] so just the upper panel flipped over [16:05] both actually [16:06] komputes: I guess your problem is that you're using the wrong libraries for your driver [16:06] komputes: how did you switch between drivers? [16:07] this isn't a driver issue [16:07] the panels are mirrored somehow [16:07] tseliot: jockey-gtk [16:07] tjaalton: which is why I didn't call it driver issue ;) [16:07] yes, but when entering and exiting gnome-session, the windows/screen is notmal for a second [16:08] tseliot: wrong libGL could do that? [16:08] tjaalton: yep [16:08] tseliot: however I did reproduce it to see if it constantly fails after installing the nvidia driver and it does seem to be the root cause [16:09] alrighty [16:09] komputes: let me rephrase my question, what did you do to switch between drivers in Jockey? [16:10] and did you restart your computer? [16:10] I suspect that you're using Nvidia's libGL with intel [16:11] tseliot: i plugged in the usb key, ran kockey, downloaded nvidia driver, rebooted, changed the resolution and wrote to xorg.conf, moved usb key to mini 9, removed xorg.conf from tty, restarted, and screen was flipped. [16:11] tseliot: shouldn't bulletproof-X know better? [16:11] komputes: you should have disabled nvidia from Jockey first [16:12] tseliot: I don't remember having to do this before [16:12] tseliot: how can I disable it from a TTY? [16:12] yes, provided that X fails to start [16:12] komputes: sudo update-alternatives --config gl_conf [16:12] sudo ldconfig [16:13] tseliot: so i need to diable nvidia from jockey and remove the xorg.conf every time i move my USB key to another computer? [16:13] sudo update-initramfs -u [16:13] and reboot [16:13] komputes: yep, or you can do it manually [16:14] disabling nvidia shouldn't remove the package though, I'll talk again to pitti about this [16:14] tseliot: any way around this, that it could be made to auto-detect? [16:14] (as I believe this has worked in previous releases) [16:14] komputes: things were quite different in previous releases [16:15] and they didn't work [16:15] like you suggest [16:15] komputes: ^ [16:15] thanks Alberto [16:15] tjaalton: will test to make sure, cheers [16:17] np [16:45] hmm, newest blob is failing to resume about 1 in 5 times and not having any luck finding out why unlike before === radoe_ is now known as radoe [18:42] hmm, think this would work? I know its a horrible hack, not sure if it's correct and wont be around a machine I can try it on until tonight - http://sarvatt.com/downloads/patches/0001-hw-xfree86-common-xf86AutoConfig.c-Append-Module-sec.patch [18:45] nvidia works without an xorg.conf but it loads libglx.so from /usr/lib/xorg/modules/extensions/ instead of /usr/lib/extra-modules/ where nvidia-current puts it unless there's a module section explicitly loading glx [18:47] so it loads SGI-GLX stuff instead of NV-GLX, thats the only thing stopping us from having no a xorg.conf nvidia blob. not sure about fglrx though [18:50] fglrx looks strange, its not installing stuff to /usr/lib/xorg/extra-modules like nvidia is and still has an xsession script changing LIBGL_DRIVERS_PATH and LD_LIBRARY_PATH? [18:53] still, if you have a conffile it'll only try the one on the top of the list [18:53] so if you put a blob in there it'll fail if you don't have it installed... [19:46] i dont know for sure but it looks like disabling nvidia-current will unset the extra-modules alternative link so the libglx from there wouldn't be used anyway and wouldn't hurt? [20:30] i'll put it on a ppa with a change to prefer nvidia first then nouveau and test it out on a bunch of configurations [21:18] oops, thought disabling FBC didn't actually work afterall since I started getting hangs again today but I didnt notice we got a -17 kernel update that overwrote my old one :D [21:21] heh [21:39] hmm, anyone else having problems with gnome-terminal after todays updates? its blacking out my screen except a sliver of the panel until i move the cursor out of the gnome-terminal window, this has been happening for awhile but its happened at least 10 times today vs once a week for the past few weeks [21:44] ok *thats* odd, someone reporting lshw corrupts their display on my exact machine [21:51] xserver merge pushed to git, not build-tested though. wondering if more of the udeb support needs to be dropped [21:54] hmm, I wonder if lshw screwing up the display might be related to having vga16fb on fb1, only difference i can see between me and this bug report with my machine [21:54] or bdeps relaxed [21:54] Sarvatt: That's an excellent guess. Let's see... [21:55] rebooting now to find out [21:56] Well, doesn't kill *this* laptop. [21:56] hmm, i should use a livecd incase its different, all these bugs are [22:19] I'm truly disappointed... installed Lucid on a recently new Desktop PC, running on a 64GB SSD, with Win7 and VirtualBox. install time inside the VM: 2-3min. boottime: 2-3 sec till gdm. after installing guest additions X would not start :( [22:26] yay [22:26] livecd + lshw -C network = trippy colors [22:28] ahaha [22:28] oh man [22:28] I remember gusty [22:28] I had so many colors on my screen back then [22:38] BUGabundo: hmm? you have a computer with win7? [22:38] no [22:38] it wasn't mine [22:38] from a friend [22:39] he was showing off how fast the SSD was [22:39] so I installed Ubuntu on it [22:40] inside vm? [22:40] yep [22:40] virtual box [22:40] it was SOOOOOOOOOOOOOOO fast [22:40] I was planning buying a 32GB ssd. now I definitely want one [22:41] :) [22:45] * baptistemm have a 160 GB ssd disk :p [22:45] you did??? [22:45] and yes this blazing fast to boot [22:45] how much did that baby cost [22:45] and even for testing vm this really really valuable [22:46] BUGabundo, too much [22:46] I bet [22:46] at current cost its around 400-500€ [22:46] for a 128GB [22:46] so a 160 who whooo [22:52] I paid mine 350€ [22:52] coming from taiwan [22:54] how long is life expentancy for ssd disks? [22:58] ahh that's _cheap_ [22:58] kklimonda: depends *very* much on manufacture and series [22:59] And correlates well with cost - the best ones will be single-level cells, and will cost a bomb. [22:59] on those without realiners you can count with about 10k re-writes for sector [23:00] on the most recent, dynamic reallocation , and extra sectors, from high brand manufactures (intel, coz, etc) something like 500k-1M re-writes [23:05] you got entry level kingston for 150€ for 32GBs [23:05] and intel is launching a new series of their famous X25 for entry price [23:07] BUGabundo, mine is a x25 g2 [23:08] don't know the g2 series [23:14] RAOF, btw when you get a chance look at #545493 - this is from a friend of mine here in Portland [23:16] Hm. Someone complaining about poor compiz performance with nouveau. [23:17] lol [23:18] Actually, it seems to be just mis-filed against nouveau; the user doesn't seem to be using the PPA, so... [23:19] bryceh, now that Benjamin has triaged bug 544904, is anything else needed or can I trust the archive managers to get to it? [23:19] Launchpad bug 544904 in xserver-xorg-video-sis "please sync xserver-xorg-video-sis 1:0.10.2-2 from Debian unstable main" [Wishlist,Triaged] https://launchpad.net/bugs/544904 [23:20] not a single 965 GPU hang report that isn't a dupe of the DPMS problem [23:20] Sarvatt, wow [23:21] tormod, should be good. Note that the sync processing crew is a bit slow so might be a few days [23:21] is anyone gonna merge -ati? I can volunteer but not today [23:22] * tormod reboots into 2.6.32-17 [23:23] not sure its a good idea :D [23:25] Sarvatt, why? [23:25] https://bugs.freedesktop.org/show_bug.cgi?id=27186 [23:25] Freedesktop bug 27186 in Driver/Radeon "Visual corruption with new r6xx/r7xx accel code" [Normal,Resolved: fixed] [23:25] oh yeah [23:25] Sarvatt: There's no lbm-nouveau for -17 yet, is there? [23:26] fixed just today but the fix isnt in 6.12.192 [23:26] Sarvatt, but it's marked fixed [23:26] Sorry, should just look myself :/ [23:26] well, we can pull the patch in [23:26] RAOF: i uploaded it like 8 hours ago but its mozilla-ed I think [23:26] tormod, if you can get to it tomorrow I'll be happy to sponsor it [23:27] otherwise it's on my todo list to do this week [23:27] looking through the lshw code to see how to get around it turning the drmfb into the vga16fb.. [23:28] Can we just make vga16fb fail to load when we've got a real framebuffer? [23:28] As I say, then nouveau guys are paranoid about it messing things up. [23:30] we've got a silly patch to vga16fb making it bind to all video devices [23:30] bryceh, ok I'll either try tomorrow or not at all :) [23:30] Sarvatt: Why? Can we drop that patch? [23:30] * Sarvatt wishes plymouth could be deferred to MM.. [23:31] Sarvatt, what's that patch against? kernel? [23:31] so there's a FB for plymouth to use for any video card always [23:32] yeah its in ubuntu-lucid.git [23:32] they added that back in november or december [23:32] Can't we just have that keybuk's swanky text-boot? That looks awesome :) [23:32] swanky text-boot? [23:32] you mean the splash you see when you dont have KMS? [23:32] Yeah; the one you'll get currently if you don't have kms. [23:32] because that needs a color framebuffer [23:33] bryceh: #545493 dealt with; nouveau kernel ABI bump in xorg-edgers strikes again! [23:34] Now, coffee. [23:34] heyo tormod! :) [23:36] hey sarvatt :) [23:41] tormod: you have an rv515 don't you? do you have any problems with lucid at the moment? [23:42] Sarvatt, I haven't got to test it with lucid yet. maybe friday. [23:42] ahh ok, there are alot of complaints from people on RV515 and RV530 with 2.6.33's drm [23:43] radeon.new_pll=0 is fixing it for RV530 people but haven't figured out the RV515 flickering problems yet [23:46] the irc log mentions savage DRI trouble? I have to test that again also, now that the X crasher got fixed [23:48] yeah savage is horrible at the moment with so many clutter based apps, clutter completely falls apart on savage because it only has 16 bit visuals and breaks some assumptions it has [23:52] I think savage can do 24 bit, but 16 bit is default because it works better and many cards have little VRAM [23:52] will clutter be fixed? [23:55] Isn't clutter 1.2 supposed to be better about that? That's getting merged in soon. [23:55] sweet, it is?! [23:56] Yup. GNOME 2.30 requires it, apparently. [23:56] 1.2 is a lot better about it but it still needs fixes [23:56] oh mutter counts for gnome 2.30? thought that was the only thing requiring 1.2 [23:56] one sec, I'll find the bug where some guy posted some patches for clutter 1.2 branch to work on savage [23:56] it was that netbook launcher bug you were working on RAOF [23:57] https://bugs.launchpad.net/bugs/467474 [23:57] Ubuntu bug 467474 in mesa "netbook-launcher crashed with SIGSEGV in glGetString()" [Undecided,Confirmed]