tjaalton | bah, bedtime -> | 00:17 |
---|---|---|
tjaalton | ok, the moment of truth.. restarting X | 08:11 |
tjaalton | no drama, things just work | 08:14 |
tjaalton | except compiz, hmm :) | 08:15 |
tjaalton | huh, "Failed to load module "glx" (module does not exist, 0)" | 08:16 |
tjaalton | ah, the path of the object has changed | 08:18 |
tjaalton | mvo: soon you'll be able to test your r5xx ;) | 08:18 |
mvo | hey tjaalton! that is great, looking forward to it :) | 08:21 |
tjaalton | mvo: need to fix the xserver build first, seems that tormod was right after all. gl.pc needs to find a home or no glx | 08:22 |
* mvo nods | 08:22 | |
tjaalton | er, mesa & xserver build | 08:22 |
tjaalton | everything else seems to be working though | 08:23 |
tjaalton | mvo: glx works now, but compiz gives me a white screen. Any ideas how to debug? | 09:18 |
tjaalton | (intel) | 09:18 |
mvo | tjaalton: white screen may mean that compiz is confused about direct vs. indirect rendering | 09:19 |
mvo | what does it print on the terminal when it starts? or what got captured in ~/.xsession-errors? | 09:19 |
mvo | it may have other causes too | 09:19 |
tjaalton | "failed to create drawable" | 09:20 |
tjaalton | a number of times | 09:20 |
tjaalton | mvo: with LIBGL_ALWAYS_INDIRECT=1 I can get something, but mostly some corrupt window borders and the rest is black | 09:25 |
tjaalton | mvo: apparently it's a known bug | 10:06 |
mvo | aha, a driver bug? | 10:09 |
tjaalton | 12:08 < MrCooper> the white screen is another reported bug, GLX_EXT_texture_from_pixmap being incorrectly advertised with direct rendering with DRI1 | 10:10 |
tjaalton | 12:09 < MrCooper> sadly those bugs seem to be getting ignored | 10:10 |
tjaalton | in mesa, yes | 10:10 |
tjaalton | so, since we really need to get these uploaded within the next week, compiz might need to blacklist intel for now | 10:12 |
seb128 | so anybody has an idea of what changed between hardy and intrepid which makes compiz unusably slow? | 10:19 |
tjaalton | seb128: have you tried the older kernel? | 10:19 |
tjaalton | hardy kernel.. | 10:19 |
seb128 | tjaalton: no, why would that makes a difference? just curious | 10:19 |
tjaalton | drm | 10:19 |
seb128 | I tried to downgrade the xorg server, compiz and libx11 | 10:20 |
seb128 | ok, will try that next then | 10:20 |
tjaalton | mesa could also be the culprit, but since it's the stable branch now in intrepid I find it hard to believe it would break like this | 10:21 |
tjaalton | ok, the intel bugs are https://bugs.freedesktop.org/show_bug.cgi?id=14441 and https://bugs.freedesktop.org/show_bug.cgi?id=15477 | 10:38 |
ubottu | Freedesktop bug 14441 in Drivers/DRI/i965 "Compiz shows only black screen on i965" [Normal,Assigned] | 10:38 |
seb128 | tjaalton: ok, that's neither linux nor mesa, any other candidate? ;-) | 10:45 |
tjaalton | seb128: nothing comes to mind :/ | 10:47 |
tjaalton | tormod might have an idea since playing alot with ati AIUI | 10:47 |
seb128 | I'm installing sysprof to see where it's spending cpu | 10:49 |
seb128 | 59.68% ressources use in r300_dri.so | 10:50 |
tjaalton | try the 32bit alpha1 livecd | 10:51 |
tjaalton | if the hardy installation was 32bit? | 10:52 |
seb128 | it is | 10:52 |
tjaalton | so to rule that out try a 32bit intrepid... | 10:52 |
seb128 | but the new installation was hardy amd64 which worked fine | 10:52 |
tjaalton | ah | 10:53 |
seb128 | and I upgraded to intrepid | 10:53 |
seb128 | and the trouble started | 10:53 |
seb128 | ok, tried again without ia32-libs which has a r300_dri.so | 10:54 |
seb128 | still as slow | 10:54 |
seb128 | sysprofing a dialog resize shows it's spending 95.71 time in X | 10:54 |
tjaalton | strange... | 10:59 |
tjaalton | all I can say ;( | 11:00 |
tjaalton | but, you'll get new mesa/xserver soon anyway :) | 11:00 |
seb128 | I downgraded those to the hardy version and still the same issue so I doubt it's due to those | 11:01 |
seb128 | trying to install some dbgsym to get details in the sysprof logs now | 11:01 |
tjaalton | sigh, my laptop doesn't stay suspended.. when I close the lid it resumes | 16:00 |
tjaalton | at least hibernate works :) | 16:01 |
tjaalton | gone -> | 16:01 |
pwnguin | my laptop has an annoying habit of having the display come back on after five minutes or so | 17:49 |
pwnguin | close the lid, leave, come back to a closed lid with a display on =( | 17:49 |
bryce | tjaalton: great news on the xserver front | 20:13 |
tjaalton | bryce: well, the intel dri driver is buggy so compiz doesn't work | 20:50 |
tjaalton | maybe you can try to make intel do something about it. the bugs are well known for some time now | 20:51 |
tjaalton | I'm also making mesa to use the new autotoolized build system.. it's getting better but still some work to be done | 20:53 |
bryce | tjaalton: for now maybe just list the known issues on the X/Drivers page that we note as new with xserver 1.5 | 21:11 |
tjaalton | bryce: yep, and probably blacklist intel from compiz until the bugs are fixed | 21:22 |
bryce | tjaalton: how long until all the new components will be in the archive? | 21:24 |
tjaalton | I need to finish mesa, which might take a day.. | 21:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!