[00:35] RAOF, lead on a mesa patch for a gpu hang - https://bugs.launchpad.net/bugs/807440 [00:35] Launchpad bug 807440 in xserver-xorg-video-intel (Ubuntu) "[i965gm] GPU lockup (affects: 1) (heat: 6)" [Undecided,New] [00:36] RAOF, might check if that could/should be backported to natty too [00:38] Cool. [00:38] What is everyone doing working on Sunday afternoon, anyway? You're here poking mesa bugs, Didier uploaded unity-2d, … === soreau_ is now known as soreau [04:17] RAOF, guilty :-) [12:57] Hi all, is there a command I can give that will give the EDID output from my monitor ? [12:57] xrandr --prop [12:59] thanks [13:02] jcristau: double thanks. thats _exactly_ what I was after :) [17:13] Sarvatt, what's the bug number for the no keyboard/mouse thing? [17:18] bug 807306 [17:18] Launchpad bug 807306 in udev (Ubuntu Oneiric) (and 1 other project) "[oneiric] Keyboard & mouse not working in X (affects: 24) (dups: 3) (heat: 136)" [High,Confirmed] https://launchpad.net/bugs/807306 [17:18] bjsnider: ^ [17:23] thanks === yofel_ is now known as yofel [20:33] so digging through some old debian bugs from may and looks like sysvinit needs to be 2.88 to populate /run and have things work, having this screwed up /run with just motd in it around from the latest base-files might be causing it? cnd: did you try sudo mv /run{,.bak} and reboot? [20:33] Sarvatt, to fix which bug(s)? [20:34] no mouse/keyboard [20:34] Sarvatt, ok, I'll try that in a bit [20:34] thanks! [20:36] cnd: thanks a ton, i tried desperately to reproduce it and gave up.. 6 machines with fresh alpha 2 then updated, and 2 of those doing a natty to oneiric dist upgrade and all were fine :( [20:36] initramfs-tools 0.99 is what mounts /run as tmpfs and crap too and we dont even have that [20:36] hmm [20:44] cnd: jackpot [20:44] kenvandine tried it, removing /run made it work again [20:46] err, bryceh already pointed it out on the bug, only /run/udev is needed apparently :P [20:46] cool [21:05] cnd, Sarvatt, yeah root caused on #807306 [21:05] downgrading base-files might undo the damage [21:05] slangasek plans to tackle the sysvinit merge ~today [21:07] nope downgrading wont clean up /run from what I can see [21:07] [18:20:08] https://lists.ubuntu.com/archives/oneiric-changes/2011-July/004255.html hrm [21:08] that makes me want a drink, that was friday.. wish i followed up on it more :) [21:12] Sarvatt, deleting it? [21:13] er, +deleting it? [21:19] bet i can reproduce it on all the sytems after this udev 172 update that creates /run/udev [21:20] you have to have updated udev after base-files was installed to reproduce it before, there wasn't a udev update for a month.. [21:31] Sarvatt, seen any bugs with flash video causing large X cpu usage (>75%)? [21:37] haven't seen any bugs, thats what flash does so not surprised :P [21:58] Sarvatt, if you can reproduce it, mind testing this fix? https://code.launchpad.net/~rsalveti/ubuntu/oneiric/base-files/run-transition/+merge/67444 [21:58] if it is confirmed as solving it, I can upload it as the fix [21:58] well, as temporary fix [22:32] Sarvatt, plunked it into a PPA too - https://launchpad.net/~bryce/+archive/lime [22:39] bryceh: sorry I went out at EOD and just got back, lets see.. [22:39] bryceh: i dont think this is going to have any effect, the udev package itself is trying to use /run if it exists even if things arent already migrated to it, and this isnt going to remove /run so it'll be the same? [22:40] i dont even know if i can reproduce it, let me move /run/udev back and see if it happens [22:42] udev 167-2 from debian had the fix we'd need to fix it, udev (167-2) unstable; urgency=high * Only use /run if it is a tmpfs. (Closes: #621036, #620995) [22:42] afaik [22:47] adding /run is what broken it in other words, not removing the creation of /var/run because that already existed here anyway [22:47] brb rebooting into it [22:50] bryceh: no luck [22:50] your base-files + /run/udev existing = broken [22:50] everyone that updates oneiric is broken atm because udev was updated [22:51] if /run exists udev package updates install crap to /run/udev instead of /dev/.udev, but the rest of the world is still expecting to use /dev/.udev [22:56] hrm [22:58] bryceh: the libc6 bug is different [22:58] that might indeed fix that one [22:59] udev will still be broken though [23:00] base-files isn't creating the base file system layout on a normal apt-get upgrade, /var/run already exists here [23:34] Wheee! No udev love for me! [23:35] RAOF, move aside /run/udev [23:42] Or just plug/unplug the keyboard and mouse :) [23:43] real easy to do on a laptop and all :) [23:43] it's our Bug of the Day [23:43] (see scrollback in just about any channel) [23:44] Sarvatt: I've gone all standing-desky, so I do have an external keyboard/mouse to prod. ;) [23:45] You could probably also trigger udev events to fix it, too; you can get to a VT by switching the keyboard into raw mode. [23:46] While we're on Bug Of The Day, has anyone else's /dev/dri permissions been wrong? [23:47] That might also be udev, I guess, but could be consolekit failing to set the acl properly.