| bryce | I've fleshed out https://wiki.ubuntu.com/X/Bugs/IntelDriver a bit more | 00:06 |
|---|---|---|
| tseliot | I've added another I830WaitLpRing() bug to the wiki | 00:21 |
| bryce | it'd be nice to update that LpRing table to remove the fixed bugs | 00:39 |
| * bryce reorders by chip id | 00:42 | |
| cwillu | 22741 is a dupe of 28326 | 00:44 |
| cwillu | which is odd, seeing as it's not the same chip? | 00:44 |
| cwillu | oh, nvm, just notice the other line | 00:45 |
| bryce | let's try sorting by symptom... | 00:46 |
| bryce | there we go | 00:48 |
| bryce | hmm | 00:49 |
| bryce | well maybe that'll give some ideas on further triaging... bugs with similar symptoms should test each other's workarounds and steps to reproduce | 00:50 |
| bryce | hah, I've reproduced the freeze on i965 | 07:33 |
| bryce | heya RAOF | 08:34 |
| RAOF | Howdie bryce. | 08:34 |
| RAOF | What's up? | 08:34 |
| bryce | ah, well I was finally able to reproduce that i965 freeze bug that's been going around | 08:35 |
| bryce | I'm poking around collecting any data that looks interesting | 08:35 |
| bryce | not finding strong clues, but am finding some "interesting evidence" | 08:35 |
| RAOF | I think I'll rebuild a kernel with the MMIO tracer enabled, and provide someone with a log of the suspend/resume cycle with the nvidia driver. | 08:36 |
| andresmujica | ping bryce | 20:25 |
| === maco_ is now known as maco | ||
| bryce | andresmujica: just ask | 22:36 |
| bryce | tjaalton: mesa 7.4 brought a freeze regression. Downgraded to 7.3 and it seems to have gone away. | 22:37 |
| superm1 | bryce, it also brought a regression in myth causing segfaults | 22:38 |
| superm1 | i've been bisecting all day to find the commit | 22:39 |
| superm1 | (bug 355242) | 22:39 |
| ubottu | Launchpad bug 355242 in mesa "mythfrontend.real crashed with SIGSEGV in QGLWidget::resizeEvent()" [Undecided,Confirmed] https://launchpad.net/bugs/355242 | 22:39 |
| bryce | superm1: aha, yeah I wondered if that could be the case, I saw that bug too | 22:39 |
| bryce | superm1: in looking through the mesa 7.4 changelog, it seems the majority of the changes for -intel were for DRI2 issues, we probably don't care so much about | 22:40 |
| superm1 | bryce, i dont know what drivers are affected, but i'm able to reproduce it with vesa, so virtualbox has been my friend | 22:40 |
| bryce | hmm | 22:45 |
| bryce | yeah I can't tell from the bug report whether it's driver-specific or in the genera mesa code | 22:45 |
| superm1 | well i'm down to like 2 commits left, so i should know soon enough | 22:45 |
| andresmujica | hi bryce, i just was going to ask about the freeze | 22:45 |
| superm1 | just annoying that it takes like 25min to do mesa builds | 22:45 |
| andresmujica | it's only related to x? mesa and dri? | 22:46 |
| andresmujica | or is kernel involved too? | 22:46 |
| bryce | hrm, seems this mesa 7.4 upgrade is kind of poor quality, if we've already found two serious regressions with it | 22:46 |
| bryce | superm1: I also looked in the mailing list for recent discussions of regressions but didn't see anything obvious, maybe the problems are some weird conflict with this mesa and ubuntu-specific kernel/xserver junk | 22:47 |
| bryce | hopefully timo can weigh in | 22:47 |
| bryce | andresmujica: heya, thanks for looking at the freeze bugs :-) | 22:47 |
| andresmujica | hehe | 22:48 |
| bryce | andresmujica: from what I suspect, there is definitely one freeze bug, which I believe started with the mesa 7.4 upgrade on 4/3, which only affects i965 | 22:48 |
| andresmujica | yeap, i'm reading about them.. | 22:49 |
| andresmujica | ok. | 22:49 |
| bryce | andresmujica: however there are also plenty of other freeze bugs that still exist | 22:49 |
| andresmujica | gonna test it here downgrading | 22:49 |
| andresmujica | yeap | 22:49 |
| andresmujica | that's what i've noticed.. | 22:49 |
| bryce | andresmujica: so we have to distinguish between them | 22:49 |
| andresmujica | which factor are u using to distinguish them? | 22:49 |
| bryce | they're really, really hard to debug though - no error messages, and all have identical symptoms, and a lot seem to come on randomly | 22:49 |
| bryce | so far, the most useful thing is if you can pin it to a particular date when it started happening, and from that discern which package changed at that time, that brought the freeze into existence | 22:50 |
| bryce | unfortunately, when people first noticed after upgrading from intrepid -> jaunty, the number of changes is too great to sort through, so that's kind of a shame | 22:51 |
| bryce | anyway, what you've been posting to the bugs is exactly the right info and questions to present to reporters, so keep it up, hopefully that can get us better knowledge for what to do with the bugs. | 22:52 |
| andresmujica | ok. perfect then! | 22:52 |
| bryce | if you get answers that show that they're freezes in -intel that still exist in jaunty, please add them to that table so we can keep track of them for later | 22:53 |
| andresmujica | ok | 22:53 |
| bryce | ok, wife is kicking me out in the shop to get some house projects done... bbiaw | 22:58 |
| superm1 | well git bisect claims fe5328bfad063f2ccbcafa2cebe6971d7e8e1db7 is first bad commit | 23:00 |
| superm1 | i'll do a test build with just that commit reverted to check tho | 23:01 |
| superm1 | well that would have been wishful thinking i guess. there must be some more (possibly related) commits that are also a factor | 23:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!