| === Amaranth_ is now known as Amaranth | ||
| Sarvatt | tjaalton: where did you look? I wonder if there has been any mplayer rebuilds since xorg 7.5 was pulled in, bet that will have a problem with the new header layout | 02:12 |
|---|---|---|
| Sarvatt | ugh, yeah, mplayer hasnt built since xorg 7.5 was pulled in, something tells me its gonna have problems | 02:35 |
| Sarvatt | i'll try it now | 02:36 |
| Sarvatt | odd | 02:37 |
| Sarvatt | they packed the vdpau headers directly into the mplayer source, but it has a nvidia-185-libvdpau-dev build-dep | 02:37 |
| superm1 | debian was not doing it right fo ra while | 05:17 |
| superm1 | they were trying to find a way to have libvdpau be in contrib when it should have been in main so stuff could build dep on it | 05:18 |
| Sarvatt | yeah I was just reading all about that, that was a silly bug | 05:24 |
| Sarvatt | seemed like they were saying it didnt need to be a build dep because the only thing that used it had the headers bundled in with it already | 05:24 |
| Sarvatt | (which was a hack to get it to build with vdpau support before libvdpau-dev existed) | 05:24 |
| superm1 | yup | 05:27 |
| superm1 | well now more ffmpeg stuff can build dep on it too | 05:29 |
| Sarvatt | Well, since I have installed 195.30 on my gentoo box, I experience a constantly high performance with kwin and plasmadesktop. There must be something seriously wrong. | 06:01 |
| Sarvatt | I like that bug report :D | 06:02 |
| Sarvatt | the difference is really noticeable here in gnome too | 06:03 |
| Sarvatt | sheesh this xrenderbenchmark is a little biased to these new nvidia drivers | 06:48 |
| Sarvatt | Transformation/Bilinear filter...................6000 frames in 2.01387 seconds = 2979.338 FPS vs Transformation/Bilinear filter...................1322810 frames in 2.0001 seconds = 661372.593 FPS | 06:49 |
| tjaalton | Sarvatt: http://qa.ubuntuwire.com/ftbfs/ | 06:49 |
| Sarvatt | ahhhh thanks | 06:50 |
| Sarvatt | evdev synaptics and mouse are the only input drivers on there? odd | 06:56 |
| Sarvatt | doesn't that mean keyboard mouse and such built with the old input abi? | 06:59 |
| Sarvatt | ah some of the others arent in main | 07:02 |
| Sarvatt | and it looks like keyboards dropped :D | 07:02 |
| tjaalton | Sarvatt: no, it only lists packages that failed on some arch | 07:27 |
| tjaalton | sparc/powerpc still don't have the new xserver | 07:27 |
| tjaalton | and -keyboard was dropped, yes | 07:27 |
| tjaalton | Sarvatt: have you bisected which libdrm commit breaks your setup? | 08:14 |
| Sarvatt | no I haven't, this is my main system I've got in the car with me all day while I was working 14 hour days because of the holidays so I've needed things to be stable | 08:33 |
| Sarvatt | but I'll actually try to start that now | 08:34 |
| Sarvatt | busy times all over :D | 08:34 |
| Sarvatt | it would be so much easier if i could just install it system wide, hmm | 08:35 |
| Sarvatt | hmm, should i start at a checkout of 2.4.16? they merged in a few branches just after 2.4.16 released and the commits went back | 08:36 |
| Sarvatt | not sure if it would include those commits if i start with master and mark 2.4.16 good | 08:37 |
| Sarvatt | odd thing with nvidia just now.. purged 195, installed 190, rebooted and got failsafe and the xorg.0.log says the nvidia module wasnt found, but i manually modprobed it fine | 08:39 |
| Sarvatt | when i exited to console login | 08:39 |
| Duke` | which bug(s) are you hunting? | 08:40 |
| Sarvatt | oh thats the x module nvidia failing to load | 08:41 |
| Sarvatt | the one you get too Duke` | 08:41 |
| Duke` | ok | 08:41 |
| Duke` | cool :D | 08:41 |
| Sarvatt | that started between 11-30 and 12-03 | 08:41 |
| Sarvatt | oh wow that package was screwed up, it didnt install nvidia-190-glx :D | 08:43 |
| Sarvatt | oh, can just bisect it off of origin/ubuntu | 08:48 |
| Sarvatt | wish it didnt take almost 24 hours to trigger a crash sometimes though :( how long have you managed without a crash Duke`? | 08:50 |
| Duke` | hum I think 2 days ago, but I'm on hollidays so I use my computer a bit less, and since I hate this bug I try to not provoque it :D | 08:52 |
| Sarvatt | yeah, i figured out ways to prolong it too :D | 08:53 |
| Duke` | (navigating on the map of this site: http://france.meteofrance.com/ makes me crash sometimes) | 08:53 |
| Sarvatt | scrolling with pageup/pagedown instead of the trackpad, avoiding flash, turning off thumbnails | 08:53 |
| Duke` | yeah | 08:54 |
| Sarvatt | thanks i'll try crashing it with that site :) | 08:54 |
| Duke` | I don't know if bisecting is really the proper way for this sort of bug we can't reproduce easily | 08:54 |
| Sarvatt | its only like 8 commits | 08:55 |
| Duke` | maybe it would be better to start with a broken version and decrease until the bug doesn't appear | 08:55 |
| Sarvatt | only reason i'm going to try | 08:55 |
| Sarvatt | yeah i'm starting with 2.4.16 in origin/ubuntu on debian git | 08:55 |
| Duke` | because you can run during a week and still not be sure it's OK :D | 08:55 |
| Sarvatt | maybe i'll make some kind of script that opens nautilus, browses to a folder with about 6k videos in it, scrolls, deletes video thumbnails and repeats :D | 08:56 |
| Sarvatt | ACTUALLY | 08:57 |
| Sarvatt | its another bug but its kind of related and i can trigger it faster | 08:57 |
| Duke` | ah? | 08:57 |
| Sarvatt | i'm having problems with flickering after resume from suspend on 2.6.32+ kernels | 08:57 |
| Duke` | I never had problem with thumbnails, but well I don't generate them everyday | 08:57 |
| Sarvatt | and if i compound it with libdrm 2.4.16+ I can trigger the execbuff while wedged error right after resume | 08:58 |
| Duke` | do you know if someone at intel is actively hunting this bug too? | 08:58 |
| tjaalton | are they aware of it? | 08:59 |
| Sarvatt | doesn't look like it, lots of bugs assigned to people and it seems like they're gonna release 2.10 any time now | 08:59 |
| Sarvatt | the guy who commited those libdrm commits that are causing us problems is on vacation | 08:59 |
| Sarvatt | http://bugs.freedesktop.org/show_bug.cgi?id=25475 | 09:00 |
| ubottu | Freedesktop bug 25475 in Driver/intel "[i915] Xorg crash / Execbuf while wedged" [Critical,New] | 09:00 |
| tjaalton | ah, great | 09:01 |
| Sarvatt | whoa, check out the speed difference between 195.30 and 190.42 nvidia blobs -- http://sarvatt.com/downloads/xrenderbenchmark-results.txt | 09:06 |
| Duke` | o_O | 09:07 |
| Sarvatt | i cant even use AccelerateTrapeziods on this 8400M GS | 09:07 |
| Sarvatt | smells like nvidia benchmark fudging from windows leaking over to linux :D | 09:08 |
| Sarvatt | (not that optimizing it for glxgears fps wasnt already) | 09:09 |
| Duke` | in the bug report Tobias says he can trigger the bug in 5 mins or half an hour. Lucky guy. Or not. | 09:09 |
| Sarvatt | me too, it's called set up your online banking account | 09:10 |
| Sarvatt | it crashed on me right when it popped up some info i needed to save :) | 09:10 |
| Duke` | Murphy law ftw! | 09:11 |
| Sarvatt | LOL | 09:13 |
| Sarvatt | last time I crashed -- I was writing a bug report response about the crash saying it was stable | 09:14 |
| Sarvatt | guess I lost the bookmark I saved to that bug right before it crashed | 09:14 |
| Sarvatt | [264301.566382] (WW) intel(0): i830_uxa_pixmap_swap_bo_with_image: bo map failed | 09:16 |
| Sarvatt | that line shows up in xorg.0.log when it does the execbuff while wedged crash right after resume | 09:17 |
| Duke` | graaahh I've been bitten by the rabbit x_x | 09:17 |
| Sarvatt | so its a different bug but maybe that'll be easier to bisect | 09:17 |
| Sarvatt | https://bugs.freedesktop.org/show_bug.cgi?id=25598 | 09:18 |
| ubottu | Freedesktop bug 25598 in Driver/intel "[965GM] Corruption on resume from hibernation with xf86-video-intel-git" [Critical,New] | 09:18 |
| Duke` | what does *-dbg packages bring exactly? do they replace the non-debug package, or do they just "add" symbols and still need the non-debug package to work? | 09:19 |
| tjaalton | add symbols | 09:20 |
| Sarvatt | actually I'm wondering that too, i know it installs another copy but i'm not sure if gdb is just set up to read it from there or if it sets up an alternative to the files installed in the normal one | 09:25 |
| Sarvatt | they usually package in debugging aids in the dbg packages too, like intel_reg_dumper goes in the dbg package (or did at least) | 09:26 |
| Sarvatt | i seem to remember tormod mentioning having made an extension to auto-xorg-git to aid bisecting at some point.. | 09:28 |
| Sarvatt | not really much of a vdpau difference between 195.30 and 190.43 http://sarvatt.com/downloads/vdpau-benchmark.txt | 09:51 |
| Sarvatt | time to start crashing :D | 10:12 |
| Sarvatt | wow I'm impatient at 5 am trying to crash this box. only been about 20 minutes and i'm thinking about starting it a few days later than 2.4.16 just incase | 10:30 |
| Duke` | have you crashed yet? | 10:34 |
| Sarvatt | if you see me quit, I crashed :) | 10:42 |
| Sarvatt | nope though, havent yet | 10:42 |
| Duke` | ok :) | 10:42 |
| Sarvatt | wooonderful :D | 10:55 |
| Sarvatt | first crash, nothing in the logs | 10:56 |
| Sarvatt | http://upload.wikimedia.org/wikipedia/commons/b/b7/Singapore_port_panorama.jpg | 10:57 |
| Sarvatt | that pic works great, it used to crash me reliably almost a year ago and I guess its back to doing it again | 10:57 |
| Sarvatt | only on 2.4.16 though! | 10:58 |
| Sarvatt | dont know if i should mark 2.4.16 bad or try to get the execbuff while wedged error crash first | 10:58 |
| Duke` | I should not have loaded your crashpic >_< | 11:04 |
| Sarvatt | crashed you just loading it? | 11:08 |
| Sarvatt | i opened it in about 20 tabs in 3 versions of firefox at the same time | 11:09 |
| Sarvatt | took about 30 minutes of that though | 11:09 |
| Sarvatt | trying the next step now halfway between 2.4.16 and Correctly set DRM_MAX_MINOR for all platforms. that worked right | 11:19 |
| tjaalton | whee, kernel crash while typing in oo.org | 11:30 |
| Duke` | Sarvatt, it crashed during loading (I don't know if it was totally loaded or not) | 11:31 |
| tjaalton | wow, a reproducible crash too | 11:33 |
| Sarvatt | yuck :) | 11:44 |
| Sarvatt | X frozen using 100% cpu at that commit | 11:45 |
| Sarvatt | happened again just using gnome-terminal after a reboot, i just marked it bad and moved on but i guess i cant be sure about it now | 11:46 |
| Sarvatt | it draws that panorama picture a heck of alot faster at intel: Clear bo->used_as_reloc_target flag on destroy than 2.4.16, thats for sure | 11:51 |
| Sarvatt | git bisect doesn't really work when everything between good and the middle has different problems :D | 12:19 |
| tjaalton | probably true.. | 12:37 |
| Sarvatt | well, tried everything i could bisecting it, too many problems needing later commits to fix it those few days.. now i'm trying 2.4.16 and reverting individual commits to see if it fares any better | 13:32 |
| Sarvatt | started by reverting http://cgit.freedesktop.org/mesa/drm/commit/?id=acb4aa671507aa181b3ff50ccf26a1c0d705a309 | 13:36 |
| Sarvatt | well, no luck, went through all the bisect which didnt work then started reverting 1 intel patch at a time backwards from 2.4.16 and i'm at 3 reverts | 15:21 |
| Duke` | yeah it's hard to hunt | 15:34 |
| _Narc_ | Hello everyone, sorry to bother, but I think I'm experiencing and important bug and nobody's able to help me on #ubuntu or ubuntu-bugs. Someone there told me to ask you : I recently updated Karmic and now my left click is almost dead. I can't focus, close, click buttons or tabs, etc. I disabled Compiz effects, but it still doesn't work. Thanks a lot. | 16:16 |
| === \vish is now known as mac_v | ||
| === mac_v is now known as \vish | ||
| dyek | Hi! Is Ubuntu 9.10 Karmic using X11R7? Is libx11-6 considered old, or it continues in its role to provide the xlib interface for X11R7 while xcb is introduced? | 18:17 |
| jcristau | yes. | 18:18 |
| dyek | jcristau: Thanks. | 18:20 |
| Sarvatt | looks pretty promising, 6 hours of firefox huge image torture and no crashes reverting "intel: Repeat execbuffer after EINTR" and "intel: Check and propagate errors from building reloc-tree" from 2.4.16 | 20:11 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
