RAOF | Hello, x86-64 assembly reference. Do you have 400 pages worth of instruction definitions? You do? Guess it's lucky there's code already written! | 00:00 |
---|---|---|
bryceh_ | dang sun, get out of my eyes and off my monitors | 00:05 |
bryceh_ | RAOF, did you see bug 732809? | 00:20 |
ubot4` | Launchpad bug 732809 in mesa (Debian) (and 1 other project) "mesa update to 7.10.1-0ubuntu1 breaks ipython matplotlib (dup-of: 259219)" [Unknown,New] https://launchpad.net/bugs/732809 | 00:20 |
ubot4` | Launchpad bug 259219 in software-center (Ubuntu Natty) (and 3 other projects) "Broken TLS support in libGL.so (affects: 121) (dups: 28) (heat: 300)" [High,Triaged] https://launchpad.net/bugs/259219 | 00:20 |
bryceh_ | ah that's just the TLS bug, nevermind :-) | 00:21 |
RAOF | bryceh_: I did indeed ;) | 00:24 |
RAOF | And, barring assembly, I've got it done locally. | 00:24 |
bryceh_ | pity no feedback yet on bug 702090. did I go overboard with the paint-by-numbers directions and scare everyone off? | 00:34 |
ubot4` | Launchpad bug 702090 in xserver-xorg-video-intel (Ubuntu Natty) (and 6 other projects) "i965gm GPU lockup if vesafb is left loaded (EIR: 0x00000010 PGTBL_ER: 0x00000100) - *ERROR* EIR stuck: 0x00000010, masking (affects: 72) (dups: 57) (heat: 520)" [High,Incomplete] https://launchpad.net/bugs/702090 | 00:35 |
bryceh_ | apw, from my limited testing the vesafb kernel doesn't appear to *cause* issues near as I can tell. If you wanted to just roll it out, we get enough automatic gpu error reports I think it would become pretty clear within a couple days whether it has eliminated the issue. We should notice a flattening of the curve and no further bug reports filed with these codes. | 00:40 |
RAOF | I WIN! (At causing glxinfo to segfault) | 01:32 |
RAOF | And if you don't clobber registers mesa assumes aren't clobbered, glxinfo doesn't segfault either! | 01:42 |
RAOF | Ahem. And if you actually preserve the other registers, glxgears works, too. | 02:16 |
bdrung | hi, my new sandbybridge system has frequent gpu lockups (bug #733006). is there something that i can do (test)? | 02:26 |
ubot4` | Launchpad bug 733006 in xserver-xorg-video-intel (Ubuntu) "[sandybridge] GPU lockup 1a28f5a9 () (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/733006 | 02:26 |
RAOF | bdrung: Is that before or after mesa 7.10.1? There were lots of sandybridge fxxups in that. | 02:43 |
bryceh_ | looks like it - version.libgl1-mesa-glx: libgl1-mesa-glx 7.10.1-0ubuntu1 | 02:44 |
bryceh_ | bdrung, unfortunately your gpu dump has a 00's for the error codes | 02:45 |
bryceh_ | [ 49.302005] [drm:init_ring_common] *ERROR* render ring initialization failed ctl 00000000 head 00000000 tail 00000000 start 00000000 | 02:45 |
bryceh_ | "render ring initialization", haven't heard of that one before | 02:45 |
bryceh_ | bdrung, so here are the things I'd suggest testing, in no particular order: | 02:46 |
bryceh_ | * boot an older kernel if you still have one installed. That can help identify if it's a drm regression | 02:46 |
bryceh_ | * downgrade mesa. Like raof says it has some changes for sandybridge. Conceivably those changes could carry bugs too | 02:47 |
bryceh_ | bdrung, the above two are especially useful if this is a regression from it used to work right in the past | 02:47 |
RAOF | * upgrade mesa; I think xorg-edgers is usable at the moment, and upstream's always more interested in bugs not fixed in the bleeding-edge. | 02:48 |
bryceh_ | * try the intel-drm-next kernel - http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/ - which will either identify if there might be a fix upstream, or else demonstrate it's an upstream bug | 02:48 |
bryceh_ | ok I am told I have a tired demanding pregnant wife needing some attention, so gtg | 02:51 |
bryceh_ | bdrung, hopefully the above helps. Also keep filing apport bugs, sometimes it'll catch error codes when you're getting a string of 0's. | 02:52 |
bryceh_ | bdrung, there are a couple other people with similar error=0x00000 types of bugs for various chips, it's possible they're related. I hesistate to send them upstream without more specifics but if we get enough reported will do so | 02:53 |
bryceh_ | l8r | 02:53 |
Sarvatt | bdrung: try booting with with drm_kms_helper.poll=0 additionally please :) | 05:04 |
Sarvatt | hmm, I guess updating lucid/maverick in xorg-edgers wont be that big a deal in the end, I just didn't want to backport the libdrm-nouveau1a rename since I'd have to rebuild crap like plymouth in there when the one built against 2.4.17 nouveau abi works fine but its just plymouth | 05:20 |
RAOF | Bah. When are those darned new Sony notebooks coming out? | 06:02 |
bryceh_ | check it --> http://www.bryceharrington.org/cgi-bin/upstreamer.cgi?bug=728823 | 06:07 |
bryceh_ | attachment forwarding is partly in there to but it's sloooowwww so left it disabled for now | 06:08 |
RAOF | Funky! | 06:11 |
RAOF | Doh! | 06:14 |
RAOF | Stupid IA32 assembler. | 06:14 |
bryceh_ | apw, https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/733096 | 07:17 |
ubot4` | Launchpad bug 733096 in xserver-xorg-video-intel (Ubuntu) "[q45] GPU lockup 32b1ab02 (EIR: 0x00000010 ESR: 0x00000010 PGTBL_ER: 0x00000002 IPEHR: 0x01000000) (affects: 1) (heat: 6)" [Undecided,New] | 07:17 |
ricotz | RAOF, hi | 09:12 |
ricotz | RAOF, did you see issues like these yet? nvidia-blob: http://enyo.zakx.de/~angelos/docky.png -- intel: http://img713.imageshack.us/i/docky.png/ | 09:14 |
bdrung | good morning. i will try all your suggestions and report back. | 09:51 |
bdrung | bryceh_: the drm-intel-next kernel crashes with an ops while booting | 10:38 |
bryceh_ | bdrung, the drm-next branch might be a bit more stable | 10:44 |
bdrung | bryceh_: that crashes too (general protection fault) | 10:53 |
bryceh_ | bdrung, the fact that it's crashing is probably interesting and may be relevant on your original issue. hope you're taking notes | 10:59 |
bdrung | bryceh_: nope, but screenshots ;) | 10:59 |
bdrung | now i am trying xorg-edgers | 11:00 |
bryceh_ | have I mentioned today how much I hate the intel video driver's bugginess? | 11:02 |
bdrung | no and i have to agree | 11:02 |
bdrung | i had to start Windows to verify that the hardware is ok and that it's only a software problem | 11:03 |
bdrung | bryceh_: xorg-edgers brought no improvement | 11:04 |
bdrung | still gpu lockups | 11:05 |
bdrung | right after login | 11:05 |
* bryceh_ nods | 11:05 | |
bryceh_ | well, xorg-edgers won't make a difference if it's a kernel drm issue, which is what it's starting to sound like | 11:05 |
bryceh_ | I'm curious about the oops and the gpf | 11:06 |
bryceh_ | if those are unrelated that would be seriously unfortunate | 11:06 |
bryceh_ | more likely though, they are different reactions to the same or similar fundamental issue | 11:06 |
bdrung | bryceh_: are they stored somewhere on the disk? will apport report them? | 11:06 |
bryceh_ | you could look in /var/crash - if that's empty apport won't report them | 11:07 |
bryceh_ | I *think* that apport might neglect to report them since they're not stock kernels, but I'm not sure | 11:07 |
bryceh_ | you can look in /var/log/kern.log* to see if those files show the errors | 11:08 |
bryceh_ | if not, I'd suggest rebooting to those kernels and capture dmesg | 11:08 |
bdrung | how should i capture dmesg if they crash on booting? | 11:08 |
bdrung | and the second question is: what workaround can i use to avoid those crashed to have a chance to run apport? | 11:10 |
jcristau | netconsole | 11:10 |
bdrung | what command do i need to report these problems? | 11:11 |
bryceh_ | bdrung, a photo to capture the dmesg output showing the OOPS or fault would be a starting point | 11:12 |
bryceh_ | sometimes you can ssh in and grab it | 11:12 |
bryceh_ | if you can't ssh in, that would suggest a deeper kernel issue | 11:13 |
bryceh_ | like, ssh with an ethernet cable, if wireless hasn't kicked in yet | 11:13 |
bdrung | wow, launching the drm-next kernel in recovery mode works. i have a console and can run dmesg | 11:14 |
bryceh_ | for workarounds, if you snag the file from /var/crash/ that corresponds to the crash and copy it to another machine, there are ways to file it from the other machine | 11:14 |
bdrung | here's the dmesg output of drm-intel-next: http://paste.ubuntu.com/578782/ | 11:15 |
bryceh_ | [ 11.620556] Process plymouthd (pid: 346, threadinfo ffff88041edfa000, task ffff88041b2ec440)\ | 11:16 |
bdrung | so plymouthd died? | 11:17 |
bryceh_ | looks like it | 11:17 |
bryceh_ | could try booting with splash disabled, see if that makes the crash go away | 11:17 |
bryceh_ | or uninstall plymouth (if that still works) | 11:18 |
bdrung | how is the splash disabling parameter called? | 11:18 |
bryceh_ | I think you just delete 'splash' from the kernel cmdline | 11:19 |
bryceh_ | linux/boot/vmlinuz-2.6.38-4-generic root=UUID=2b238928-f3f1-4074-b551-2cceff54c6a4 ro quiet splash vt.handoff=7 | 11:19 |
bdrung | i copied all crash files from /var/crash to this machine. | 11:20 |
bryceh_ | so you'd just go into grub (hold down left shift during boot), trim out 'splash', and boot | 11:20 |
bryceh_ | apport-cli [ --save file ] symptom | pid | package | program path | .apport/.crash file | 11:20 |
bryceh_ | APPORT_IGNORE_OBSOLETE_PACKAGES=true apport-cli <foo.crash> | 11:21 |
bryceh_ | ^ that should be the magic to get the report to file from your other machine | 11:21 |
bdrung | apport-cli complains that the crash doesn't belong to an not installed package | 11:24 |
bryceh_ | hrm, yeah I should have added I'd never gotten that magic to work right myself | 11:28 |
bryceh_ | fwiw, the .crash file is largely text... you can upload it to one of your bug reports for reference. I think I should be able to extract the errors from the file myself. | 11:29 |
bryceh_ | I don't like doing that in general, but for you... ;-) | 11:29 |
bdrung | removing plymouth let the drm-intel-next kernel boot and it's currently running without locking up | 11:31 |
bdrung | so i can report the bug on the main machin | 11:31 |
bdrung | e | 11:31 |
bryceh_ | sweet | 11:33 |
bryceh_ | bdrung, ok, so sounding like a kernel/plymouth type of bug. | 11:33 |
bdrung | but i can't report the crashes because apport complained that the package are not ubuntu packages | 11:33 |
bryceh_ | yeah, like I mentioned earlier | 11:35 |
bryceh_ | just attach the .crash files to the bug report, and I'll sort it from there | 11:36 |
bdrung | bryceh_: i could pastebin them | 11:37 |
bryceh_ | it's getting a bit late for me, I'd prefer if you attached them to one of your bug reports, and I'll look at it tomorrow | 11:38 |
bryceh_ | (3:30am... need to hit the sack soon) | 11:38 |
bdrung | bryceh_: should i file it against the kernel package or intel? | 11:38 |
bryceh_ | kernel | 11:41 |
bryceh_ | make sure to tag it 'natty' | 11:41 |
bryceh_ | actually, if you know how, I'd suggest filing it against both 'linux' and 'plymouth' | 11:42 |
bdrung | k, will do that | 11:42 |
=== evilvish is now known as vish | ||
bdrung | bryceh_: bug #733321 reported and crash logs attached | 15:32 |
ubot4` | Launchpad bug 733321 in plymouth (Ubuntu) (and 1 other project) "[sandybridge] GPU lockups and other crashes (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/733321 | 15:32 |
=== bdmurray_ is now known as bdmurray | ||
Sarvatt | bdrung: go figure you have the screwed up board mentioned here too http://www.phoronix.com/scan.php?page=article&item=intel_sandy_speed&num=1 :( | 16:45 |
bdrung | Sarvatt: can you give me the pointer to the phoronix command that triggers the issue? | 16:52 |
bdrung | Sarvatt: i have the same board (with an b3 stepping) | 16:52 |
Sarvatt | just booting was screwed up for them, i'm trying to find any discussion on what the actual problems were | 16:53 |
bdrung | Sarvatt: my system runs fine with the drm-intel-next kernel (after only one plymouth related boot crash -> bug #733321) | 16:55 |
ubot4` | Launchpad bug 733321 in plymouth (Ubuntu) (and 1 other project) "[sandybridge] GPU lockups and other crashes (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/733321 | 16:55 |
bdmurray | bryceh_: I'm still seeing bug 726807 | 16:56 |
Sarvatt | bdrung: even without drm_kms_helper.poll=0 ? | 16:56 |
ubot4` | Launchpad bug 726807 in xserver-xorg-video-ati (Ubuntu) "drawing of drop down boxes produces artifacts temporarily (affects: 1) (heat: 479)" [High,Triaged] https://launchpad.net/bugs/726807 | 16:56 |
bdrung | Sarvatt: not tested yet | 16:56 |
Sarvatt | it looks like 2.6.38-rc8 should work if drm-intel-next works | 16:59 |
ricotz | Sarvatt, hi | 19:40 |
ricotz | Sarvatt, did you hear about some cairo breakages (with transformation matrizes) lately? | 19:41 |
Sarvatt | ricotz: afraid not, whats up? | 19:46 |
ricotz | Sarvatt, http://enyo.zakx.de/~angelos/docky.png | 19:47 |
evilvish | Sarvatt: hey, i recall you mentioned to use for the x freeze: $ echo 0x001ee62d | eu-addr2line -e /usr/lib/debug/lib/libdrm_radeon.so.1.0.0 | 19:47 |
evilvish | and echo 0x001ee5da ; echo 0x001d35da ; echo 0x001d362d | 19:47 |
evilvish | Sarvatt: should that be before the freeze or after? | 19:48 |
* evilvish forgot.. :s | 19:48 | |
ricotz | Sarvatt, this happens for people on archlinux and gentoo lately, which seems to be cairos fault or the mono 2.10 bindings | 19:48 |
evilvish | i got this without those echo lines » http://launchpadlibrarian.net/66142834/gdb-Xorg-2.6.35-6.7-generic.txt and after echo » http://launchpadlibrarian.net/66142892/gdb-Xorg-2.6.35-6.7-generic-echo.txt | 19:48 |
ricotz | Sarvatt, but i suspected cairo, mesa or xserver | 19:48 |
Sarvatt | real fix for jcastro's pageflip problem - http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commit;h=fdc315a19a2c33da29dd87d4ca88f4e4407bd42d | 21:19 |
bryceh_ | Sarvatt, what was jcastro's bug #? | 21:21 |
bryceh_ | hm, bug 715330 maybe? | 21:22 |
ubot4` | Launchpad bug 715330 in xserver-xorg-video-ati (Ubuntu) (and 2 other projects) "Freeze after login with KMS enabled on Radeon HD6310 (affects: 2) (heat: 124)" [High,Confirmed] https://launchpad.net/bugs/715330 | 21:22 |
bryceh_ | yep | 21:22 |
Sarvatt | yep, beat me to it (was digging through mail) | 21:23 |
Sarvatt | also he's working on a fix for the AGP pageflip problem https://bugs.freedesktop.org/show_bug.cgi?id=35183 \o/ | 21:23 |
ubot4` | Freedesktop bug 35183 in DRM/Radeon "system freezes with 2.6.38-rc kernel and kms pageflip enabled" [Normal,New] | 21:23 |
bryceh_ | Sarvatt, ok flagged it for JFo | 21:25 |
Sarvatt | awwh he changed it from Not-reported-usefully-by: Michael Larabel @ phoronix to Article-written-by: Michael Larabel @ phoronix | 21:25 |
bryceh_ | hehe | 21:26 |
bryceh_ | Sarvatt, feel free to close the -ati task if you think there's nothing more needed to be done on our end (unless you think it's worth leaving open just for tracking it) | 21:26 |
cnd | bryceh_, heads up: I have asked for a FFe for XInput 2.1 support in xinput (the utility) (bug 733536) | 21:49 |
ubot4` | Launchpad bug 733536 in xinput (Ubuntu) "Add XInput 2.1 support to xinput utility (affects: 1) (heat: 10)" [Undecided,New] https://launchpad.net/bugs/733536 | 21:49 |
cnd | I'm outta here now though :) | 21:50 |
cnd | see you later | 21:50 |
bryceh_ | cnd, ok thanks for the notice | 21:53 |
Sarvatt | ricotz: please dont tell me it's https://bugs.launchpad.net/bugs/259219 too :) | 22:09 |
ubot4` | Launchpad bug 259219 in software-center (Ubuntu Natty) (and 3 other projects) "Broken TLS support in libGL.so (affects: 202) (dups: 66) (heat: 688)" [High,Triaged] | 22:09 |
Sarvatt | ricotz: so whats needed to reproduce that docky problem? | 22:15 |
ricotz | Sarvatt, unlikely that bug :P -- you would need mono 2.10, so not reproduceable on ubuntu/debian that easy | 22:21 |
ricotz | looks like it is a mono/docky problem perhaps a change in the cairo bindings, pretty weird though | 22:22 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!