=== cup is now known as LLStarks | ||
LLStarks | re bug 714719, no lockups on 945gm, so the problem shouldn't reside in shared code | 02:48 |
---|---|---|
ubot4 | Launchpad bug 714719 in xserver-xorg-video-intel (Ubuntu) (and 2 other projects) "[i915gm] GPU lockup (ESR: 0x00000001 IPEHR: 0x02000004) (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/714719 | 02:49 |
tjaalton | duh, some sort of dependency loop prevents upgrade to maverick | 06:57 |
tjaalton | between nouveau and libdrm | 06:57 |
tjaalton | weird | 06:58 |
tjaalton | had to remove nouveau and video-all | 07:03 |
RAOF | tjaalton: Upgrade to maverick, or from maverick to natty? | 07:35 |
tjaalton | RAOF: to maverick | 07:38 |
tjaalton | from lucid | 07:38 |
RAOF | Weird. I would have thought there would have been much complaining about that. Perhaps you've hit an uncommon case? | 07:39 |
tjaalton | no idea, they both are the lucid versions | 07:39 |
mvo | good moring! I used to be able to write Identifier "trackpoint"\nMatchIsPointer "on"\nOption "Emulate3Buttons" "on" (in the inputclass section). but with the recent xorg upgrade that does no longer seem to work after resume - worth a bugreport? | 09:02 |
tjaalton | mvo: probably so.. | 09:04 |
* bryceh waves to RAOF, tjaalton, mvo | 09:05 | |
tjaalton | hey | 09:05 |
RAOF | Aloha! | 09:05 |
bryceh | RAOF, btw just now noticed we're getting a lot of auto-dupes of bug #712785. Might be something interesting going on there (haven't looked at the bug at all) | 09:05 |
ubot4 | bryceh: Bug 712785 on http://launchpad.net/bugs/712785 is private | 09:05 |
mvo | thanks tjaalton | 09:05 |
tjaalton | upgrade to maverick almost done, then reboot to check that it works, and then to natty alpha2 | 09:05 |
* mvo waves back to bryceh | 09:05 | |
* bryceh unprivates | 09:06 | |
bryceh | ahem, bug #712785 | 09:06 |
ubot4 | Launchpad bug 712785 in xorg-server (Ubuntu) "Xorg assert failure: *** glibc detected *** /usr/bin/X: corrupted double-linked list: 0x0a435968 *** (affects: 9) (dups: 8) (heat: 74)" [Medium,New] https://launchpad.net/bugs/712785 | 09:06 |
tjaalton | mvo: btw, I had a weird conflict on upgrade from lucid->maverick, caused by nouveau and libdrm. solved it by removing x-x-v-nouveau, and all the logs should be on the disk if you're interested | 09:07 |
mvo | tjaalton: please mail them to me, can't hurt to check them | 09:07 |
bryceh | bug #711422 looks like a variant with it's own collection of dupes | 09:07 |
ubot4 | Launchpad bug 711422 in xorg-server (Ubuntu) (and 1 other project) "Xorg assert failure: X: double free or corruption (!prev): 0x089f5b20 *** (affects: 13) (dups: 12) (heat: 106)" [High,Triaged] https://launchpad.net/bugs/711422 | 09:07 |
tjaalton | mvo: ok, will do, after lunch :) | 09:07 |
tjaalton | though it's really weird that it happened | 09:08 |
RAOF | Ah, hello GLX resource tracking. How much fun are you! | 09:09 |
RAOF | Yeah, we should investigate and fix/send upstream. | 09:10 |
bryceh | RAOF, can you try to reproduce it? I tried but I haven't seen it at all on my own hardware. | 09:16 |
bryceh | RAOF, if not, then yeah maybe bumping it upstream is the right next step | 09:16 |
RAOF | bryceh: Yeah, I don't think I've reproduced it. | 09:16 |
RAOF | Or maybe I have - it's probably another one of those on-shutdown crashes. | 09:17 |
mvo | tjaalton: sure, no rush | 09:18 |
RAOF | Hm, no, some people seem to have that bug kill their sessions. | 09:19 |
jcristau | you need to play more 3d games | 09:20 |
jcristau | for work. | 09:20 |
RAOF | Time to fire up Civ V… FOR SCIENCE! | 09:23 |
bryceh | RAOF, I sent half of the gpu lockup bug reports upstream today. The other half seemed less common; I'll send them up once we get some action on the first half. That double free bug looks worth a bit more study before sending it upstream. | 09:23 |
RAOF | I've got a patch series in mind to investigate there. | 09:23 |
bryceh | RAOF, you rock. :-) Ok, bedtime for me, cya tomorrow. | 09:32 |
RAOF | Catch you! | 09:32 |
tjaalton | upgrade to natty complete, still no compiz on nouveau, unity or not | 13:24 |
tjaalton | I'll start digging if it might be fixed somewhere already | 13:24 |
mvo | I have no copiz on nouveau too (fwiw). the unity checker claims I should be able to run it though | 13:40 |
mvo | but no texture_from_pixmap iirc | 13:40 |
tjaalton | i get the same error as on bug 710588 | 13:43 |
ubot4 | Launchpad bug 710588 in mesa (Ubuntu) (and 1 other project) "compiz assert failure: compiz: nv50_pc_emit.c:863: emit_flop: Assertion `STYPE(i, 0) == 0x09' failed. (affects: 16) (dups: 5) (heat: 100)" [Medium,Confirmed] https://launchpad.net/bugs/710588 | 13:43 |
tjaalton | I'll try mesa from edgers next | 13:43 |
tjaalton | still the same | 13:55 |
tjaalton | even with libdrm/nouveau etc from there | 13:55 |
tjaalton | oh well, best to get a proper backtrace and see what to make of it | 14:21 |
=== alf_ is now known as alf__ | ||
=== thade_w is now known as Alexia_Death | ||
apw | tjaalton, for that nv0 or whatever it was, i added xorg-edgers and dist-upgraded, is that sufficient to get support? | 15:07 |
tjaalton | apw: I think so yes | 15:07 |
tjaalton | nvc0 | 15:07 |
apw | hrm, then i think it doesn't work | 15:07 |
tjaalton | duh | 15:08 |
apw | tjaalton, it is still saying 'sw renderer' ... am fully rebooting to be sure | 15:08 |
tjaalton | check for the (EE) lines of the logfile | 15:08 |
tjaalton | see if they are the same | 15:08 |
tjaalton | or something else | 15:09 |
tjaalton | though could be that the kernel drm from .38 isn't enough | 15:09 |
apw | yeah says the same 'Erorr creating GPU channel: -19 | 15:12 |
tjaalton | ok, then maybe the drm code isn't fully baked yet | 15:13 |
tjaalton | though 2d works | 15:13 |
apw | yeah 2d is passable indeed, shame though | 15:14 |
apw | i assume the official soln is binary drivers for these | 15:15 |
tjaalton | actually, you might be missing the firmware | 15:15 |
tjaalton | suggested on #dri-devel | 15:16 |
tjaalton | though I've no idea where those are | 15:16 |
apw | ahh yes, it says failed to load fuc409d | 15:16 |
apw | which is ammusing to read | 15:16 |
tjaalton | "extract it from a blob trace" | 15:16 |
tjaalton | they'll write a free one later | 15:17 |
apw | do what? | 15:17 |
tjaalton | dunno :) | 15:17 |
apw | that sounds like running windows or something | 15:17 |
tjaalton | no just the blob | 15:17 |
tjaalton | I think | 15:17 |
apw | oh i see | 15:17 |
apw | hmmm | 15:17 |
tjaalton | ok, so compiz on nouveau works after all.. | 16:27 |
tjaalton | after I figured out that the profile had to be changed to default | 16:30 |
tjaalton | tomorrow I'll try if unity works without the assert | 16:31 |
jaytaoko | bryceh: ping | 16:55 |
bryceh | hi jaytaoko | 17:06 |
jaytaoko | bryceh: hello | 17:11 |
jaytaoko | bryceh: I was wondering if the fglrx can be installed right now and work with the latest release of XOrg 1.10 | 17:12 |
bryceh | jaytaoko, I wouldn't think so, unless amd has a new one for private testing available | 17:13 |
jaytaoko | bryceh: I was told that it wouldn't be released before march... Just checking with you about it | 17:15 |
bryceh | jaytaoko, yeah sorry, I'm fairly sure the options are either run -ati or downgrade xserver (and/or kernel) | 17:16 |
bryceh | jaytaoko, the updated driver will first be available via their internal testing program, but I think you're already subscribed to that? | 17:17 |
jaytaoko | bryceh: yes, I am | 17:17 |
bryceh | jaytaoko, ok yeah then keep an eye there for a fixed up driver | 17:17 |
jaytaoko | bryceh: ok, I will... In the mean time, I wonder how they are going to help us solve Unity on fglrx | 17:18 |
jaytaoko | bryceh: but, we have identified a few places that could help us resolve the issue... | 17:20 |
jaytaoko | bryceh: I will talk to them about it | 17:20 |
bryceh | jaytaoko, not to sound heartless but this sort of sounds like a problem of their own making ;-) | 17:20 |
jaytaoko | bryceh: yes, they knew XOrg 1.10 was coming | 17:21 |
bryceh | again though, they could use a snapshot of ubuntu from prior to the xserver and kernel bumps if they need to do the dev work against older versions, or they could manually downgrade X and the kernel on a current snapshot if they prefer | 17:21 |
bryceh | or they could do the development using their development drivers that do support 1.10 | 17:21 |
bryceh | seems like they have many more options than we do... ;-) | 17:21 |
bryceh | well, at least one more | 17:22 |
jaytaoko | bryceh: in the meantime, we can work on the opensource driver and fix the issues that are there... | 17:23 |
bryceh | sounds good | 17:24 |
bryceh | jaytaoko, are you aware of issues with unity and -ati? | 17:24 |
bryceh | I've mostly been focusing on -intel since the updates; it's the one with all the bug reports | 17:24 |
jaytaoko | bryceh: yes, there are some bugs such as random screens bugs.. things that you see only on the ati driver | 17:25 |
jaytaoko | bryceh: you don't get them on nvidia or intel... | 17:25 |
jaytaoko | bryceh: it looks more like low level driver issues | 17:25 |
bryceh | jaytaoko, ok, point me to bugs that look like driver issues; I should get those upstream asap | 17:26 |
jaytaoko | bryceh: I am searching in the bug list | 17:28 |
jaytaoko | bryceh: https://bugs.launchpad.net/unity/+bug/684992 | 17:30 |
ubot4 | Launchpad bug 684992 in unity (Ubuntu) (and 1 other project) "Artefact in launcher tooltip and icons (affects: 4) (dups: 2) (heat: 64)" [Medium,Triaged] | 17:30 |
jaytaoko | bryceh: I am still getting this one | 17:30 |
bryceh | ok | 17:30 |
jaytaoko | bryceh: it is random, and only happens with the ati radeon driver. Never saw it on other GPUs | 17:30 |
bryceh | jaytaoko, yeah rendering artifacts like that are sort of a distinguishing characteristic of -ati ;-) | 17:32 |
bryceh | I'll forward it upstream. I recall them discussing a rendering fix patch recently, perhaps they have something we can test | 17:33 |
jaytaoko | bryceh: thanks... I knew there was nothing in our rendering that could cause something like that... glad that you confirms it! | 17:35 |
bryceh | jaytaoko, yeah so if you can post X logs to the bug I'll forward it upstream today | 17:42 |
jaytaoko | bryceh: alright | 17:42 |
seb128 | bryceh, hey | 18:09 |
seb128 | bryceh, just for info debian landed a cairo build with the gl backend enabled and I did merge that in natty | 18:10 |
bryceh | seb128, oh, finally | 18:11 |
bryceh | seb128, yeah I know while I was blocked that debian has been working on getting wayland in themselves | 18:11 |
seb128 | bryceh, yeah sorry about the delay | 18:11 |
seb128 | bryceh, but didn't you work in ppa for it? | 18:12 |
bryceh | seb128, yes | 18:12 |
seb128 | ok, so "blocked" was for landing in the official archive only | 18:13 |
bryceh | seb128, but I couldn't upload the packages to natty until the dependency was available | 18:13 |
seb128 | anyway it's done now | 18:13 |
seb128 | sorry about the delay again | 18:13 |
seb128 | bryceh, right, but is there any interest to have wayland in the distro? like is it usuable for anything else than testing (which could be done in the ppa as well)? | 18:13 |
bryceh | seb128, I think I saw one blog post from someone interested in having wayland in the distro ;-) | 18:14 |
seb128 | :-) | 18:15 |
bryceh | but no it won't be usable until GNOME or other window managers support it | 18:15 |
seb128 | bryceh, right, I'm just saying why it was not on top of my todo but no point arguing over upload delay since that's over and uploaded | 18:16 |
bryceh | so the motivation for getting it in is to make life easier for those folks that will be doing that development work | 18:16 |
seb128 | sorry if that blocked other tasks for you | 18:16 |
bryceh | ok, yeah I've had enough other stuff to keep me busy, and you're right it seems few actual people have interest in wayland | 18:17 |
bryceh | hopefully I'm not wasting my time, although probably I am ;-) | 18:17 |
bryceh | seb128, anyway thanks for letting me know; I was going to ask about that. Glad to have it done and I can wrap up work on this stuff. | 18:23 |
tjaalton | meh, so nouveau dri2 is crashing, but the backtrace is useless even though I have all (?) the dbg-packages | 19:48 |
tjaalton | http://pastebin.com/fVnw771q | 19:48 |
tjaalton | wat | 19:52 |
tjaalton | has the X wrapper changed... there is no Xorg process anymore | 19:52 |
bryceh | look for an 'X' process | 19:53 |
tjaalton | yes, but the path for Xorg | 19:54 |
tjaalton | now we got something :) | 19:54 |
tjaalton | http://pastebin.com/HB6PN3bc | 19:55 |
tjaalton | but no time to debug further atm :) | 19:56 |
jaytaoko | bryceh: How do I generate an X log | 20:06 |
bryceh | tjaalton, hmm, "nouveau_bo_handle_get(nouveau_pixmap(ppix)->bo, &nvbuf->base.name);" has quite a few derefs in there | 20:07 |
bryceh | jaytaoko, how do you mean? The Xorg.0.log should generate automatically from the server | 20:08 |
jaytaoko | bryceh: ok I have it... I though it was something else | 20:09 |
jaytaoko | bryceh: apport-collect won't let me add my system info to 684992 | 20:32 |
bryceh | jaytaoko, hrm, what's it say? | 20:32 |
bryceh | jaytaoko, work around just attach your dmesg, /var/log/Xorg.0.log, /etc/X11/xorg.conf (if any), and output of 'lspci -vvnn' | 20:33 |
bryceh | I think that should be enough | 20:33 |
jaytaoko | bryceh: You are not the reporter or subscriber of this problem report, or the report is a duplicate or already closed. Please create a new report using "apport-bug". | 20:33 |
bryceh | ohh | 20:34 |
bryceh | jaytaoko, ok that makes sense | 20:35 |
bryceh | jaytaoko, you could file a new bug using 'ubuntu-bug xserver-xorg-video-ati' then, and link to it from the original bug | 20:35 |
jaytaoko | bryceh: will do | 20:36 |
bryceh | jaytaoko, also if you can take a photo of the screen corruption you reproduce for comparison, that might be helpful too | 20:36 |
jaytaoko | bryceh: yes, I just took one | 20:36 |
jaytaoko | bryceh: here https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/715405 | 20:54 |
ubot4 | Launchpad bug 715405 in xserver-xorg-video-ati (Ubuntu) "Graphics artifact in Unity, around the tooltips (dup-of: 684992)" [Undecided,New] | 20:54 |
ubot4 | Launchpad bug 684992 in xserver-xorg-video-ati (Ubuntu) (and 2 other projects) "Artefact in launcher tooltip and icons (affects: 5) (dups: 3) (heat: 34)" [Undecided,Incomplete] | 20:54 |
bryceh | jaytaoko, ok upstreamed at https://bugs.freedesktop.org//show_bug.cgi?id=34051 | 21:08 |
ubot4 | Freedesktop bug 34051 in Driver/Radeon "Graphics artifact in Unity, around the tooltips" [Normal,New] | 21:08 |
jaytaoko | bryceh: great! | 21:09 |
jaytaoko | bryceh: thanks | 21:09 |
bryceh | hopefully we will get a helpful response | 21:11 |
tjaalton | bryceh: ookay | 21:51 |
bryceh | tjaalton, I would set a break in that first routine and inspect the state of the variables | 21:52 |
bryceh | tjaalton, however given that the next call is from -nouveau to libdrm I would wonder if there's abi breakage or something | 21:52 |
tjaalton | bryceh: hmm ok, I could try edgers to see if I can reproduce it | 21:52 |
bryceh | it doesn't look like the function signature in libdrm changed recently though | 21:53 |
tjaalton | but that's for tomorrow | 21:53 |
bryceh | tjaalton, do you only see the crash when running edgers? | 21:53 |
tjaalton | the assert with unity should just be removed "and it should be fine" | 21:53 |
tjaalton | bryceh: no this is with stock natty | 21:53 |
bryceh | hm | 21:54 |
tjaalton | looks like my suspend issue should be fixed in .38rc4 | 21:54 |
tjaalton | sweet | 21:54 |
bryceh | nice | 21:56 |
bryceh | RAOF, what are your thoughts on the double free bug? | 22:12 |
RAOF | I recall a series of patches on xorg-devel about GLX resource cleanup, one of which was to avoid a double free. | 22:15 |
bryceh | that does sound promising | 22:15 |
bryceh | RAOF, are you able to reproduce the bug locally? | 22:16 |
RAOF | I haven't managed to yet, but it's early in the day :) | 22:16 |
RAOF | I don't think they got applied, or perhaps there's another codepath that needs fixing. | 22:16 |
bryceh | ok, you thinking maybe just roll the patch(s) into a ppa and solicit testers? | 22:16 |
RAOF | Yup. | 22:16 |
bryceh | sounds like a good plan | 22:17 |
RAOF | I'll see whether I can reproduce locally first, and then make that happen. | 22:17 |
bryceh | btw I've accumulated a few random -intel patches I'm going to put in | 22:17 |
bryceh | nothing earth shattering, just odds and ends and minor fixes that look worth including | 22:18 |
RAOF | Cool. | 22:18 |
bryceh | oh there was one bug I was puzzled about the response on, which might make more sense to you | 22:19 |
bryceh | forwarded upstream at https://bugs.freedesktop.org//show_bug.cgi?id=34017 | 22:19 |
ubot4 | Freedesktop bug 34017 in Driver/intel "[i965gm] GPU lockup during login" [Major,Reopened] | 22:19 |
bryceh | ickle pointed to a commit which is already included in 2.14.0 that the user reported against; am I missing something obvious? | 22:20 |
RAOF | I can't *see* anything obvious there, no. | 22:20 |
bryceh | ok, well I pushed back to ickle maybe he'll flame me more sensibly ;-) | 22:21 |
RAOF | :) | 22:23 |
bryceh | I think we're on the verge of bringing http://www.bryceharrington.org/X/Reports/ubuntu-x-swat/totals-natty-workqueue.svg back down. Looks like we're able to dupe/close/upstream bugs as fast as they're coming in, | 22:23 |
LLStarks | /ns identify robert | 22:30 |
RAOF | Time for a new throwaway password! | 22:31 |
LLStarks | yes, i fear the righteous ubuntu-x team stealing my password | 22:45 |
LLStarks | btw, yay dithering | 22:45 |
RAOF | Fonud the problem? | 22:46 |
LLStarks | it's fixed | 22:54 |
LLStarks | in a custom apw kernel | 22:54 |
RAOF | Hurray! | 22:55 |
RAOF | Let's see if I can trigger that double-free… | 22:57 |
LLStarks | Observations, measurements | 23:05 |
LLStarks | wrong cp | 23:05 |
LLStarks | raof, you mentioned occasional graphical corruption on 945, what bug was that? | 23:16 |
bryceh | bug #710961 ? | 23:23 |
ubot4 | Launchpad bug 710961 in xserver-xorg-video-intel (Ubuntu) "[i945gm] Screen Corruption with new Xorg stack (affects: 3) (heat: 505)" [Undecided,New] https://launchpad.net/bugs/710961 | 23:23 |
LLStarks | yeah, seeing that from time to time | 23:23 |
LLStarks | firefox too | 23:23 |
bryceh | RAOF, do you think we should accept i8xx bug reports or just close them and redirect people upstream? | 23:34 |
RAOF | I think close them and point them upstream, for the moment. | 23:34 |
bryceh | e.g. bug #714634 (which looks like a kernel regression in any case) | 23:34 |
ubot4 | Launchpad bug 714634 in xserver-xorg-video-intel (Ubuntu) "intel 82852/82855 cannot display login screen (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/714634 | 23:34 |
bryceh | alright | 23:34 |
RAOF | That long-awaited hope of the kernel working around those hardware problems is still brewing; maybe in Natty+1 we'll be able to turn things back on :) | 23:35 |
bryceh | bit of an essay, but here's my spiel for 8xx - https://bugs.launchpad.net/ubuntu/+source/linux/+bug/714634/comments/2 | 23:48 |
ubot4 | Launchpad bug 714634 in linux (Ubuntu) "intel 82852/82855 cannot display login screen (affects: 1) (heat: 6)" [Undecided,Won't fix] | 23:48 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!