Sarvatt | http://xkcd.com/963/ | 04:59 |
---|---|---|
ginggs | hi | 09:03 |
ginggs | I have a lab full of natty machines that broke their nvidia drivers upgrading from 2.6.38-8 to 2.6.38-11, anyone able to assist please? | 09:05 |
tjaalton | define "broke" | 09:07 |
tjaalton | assuming the kernel driver didn't build; check the dkms logs | 09:08 |
ginggs | from what i can tell the nvidia driver is built, but if the machines boot up on 2.6.38-11 nvidia encounters a general protection fault, if i select the old 2.6.38-8 kernel machine boots fine | 09:09 |
bryceh | uninstall nvidia; boot to 2.6.38-11; reinstall nvidia; reboot | 09:09 |
ginggs | i have tried purging nvidia-common and installing again, and this broke the nvidia driver in 2.6.38-8 as well! | 09:10 |
* bryceh waves to tjaalton | 09:10 | |
tjaalton | hey bryceh | 09:10 |
ginggs | so i'm guessing it's not so much the kernel, it's more to do with compiling the driver or building initrd? | 09:11 |
bryceh | ginggs, 99% of the time, yep | 09:12 |
ginggs | /var/lib/dkms/nvidia-current/270.41.06/2.6.38-11-generic/x86_64/log/make.log and /var/lib/dkms/nvidia-current/270.41.06/2.6.38-8-generic/x86_64/log/make.log look much the same to me | 09:29 |
bryceh | pastebinit | 09:30 |
ginggs | the 2.6.38-11 make.log? | 09:31 |
bryceh | anything you want our feedback on | 09:31 |
ginggs | 2.6.38-11 dkms make.log http://pastebin.com/FKgzN8uS | 10:18 |
tseliot | ginggs: nvidia-common is only used for driver detection in Jockey, it's not part of the driver | 11:24 |
ginggs | sorry - meant nvidia-current | 11:44 |
ginggs | just got back from the lab, found the cause of the problem | 11:44 |
ginggs | i had gcc-4.4 installed and had overridden gcc-4.5 in /etc/alternatives for CUDA SDK | 11:46 |
ginggs | removing gcc-4.4 and reinstalling gcc-4.5 solved the problem of the nvidia driver general protection fault | 11:47 |
ginggs | now just need to find a solution for CUDA | 11:47 |
gnomefreak | i no longer have this issue, but after i installed fglrx drivers i lost X. I tried to revert back to default drivers but i couldnt find a way to do it since there is no *xorg.conf. any idea how i would do that, in case the next time i install them and they fail? | 12:38 |
gnomefreak | also do the fglrx drivers work yet? | 12:39 |
jcristau | uninstall fglrx? | 12:41 |
gnomefreak | jcristau: didnt help | 12:45 |
gnomefreak | i couldnt find a setting in updte-alternatives either | 12:45 |
=== chrisccoulson_ is now known as chrisccoulson | ||
=== yofel_ is now known as yofel | ||
bdmurray | bryceh: "ExtraDebuggingInterest: Yes, whatever it takes to get this fixed in Ubuntu" | 21:13 |
bdmurray | that's kind of funny | 21:13 |
bdmurray | I noticed it in bug 857816 which seems rather important | 21:13 |
ubot4 | Launchpad bug 857816 in xserver-xorg-video-intel (Ubuntu) "X fails to start on boot (affects: 2) (heat: 10)" [Undecided,Confirmed] https://launchpad.net/bugs/857816 | 21:13 |
Sarvatt | bryceh: shouldn't lightdm logs be getting attached? | 21:20 |
Sarvatt | ExtraDebuggingInterest: Yes, whatever it takes to get this fixed in Ubuntu | 21:24 |
Sarvatt | doesn't matter much for this bug, I've only hit it twice since the beginning of september and haven't been able to in the past 2 weeks trying to | 21:24 |
bryceh | Sarvatt, yes it should include the lightdm logs if they're available, but there's a prompt (since it's owned by root), so it depends on the user whether they're actually included | 21:26 |
Sarvatt | ahh gotcha, was worried logs were gone completely since GDM was gone and I relied on gdm logs to fix a lot of bugs in the past | 21:26 |
Sarvatt | hmm lightdm doesn't keep historical logs like gdm did though so nevermind | 21:27 |
Sarvatt | mainly people filing bugs after the fact with nothing wrong in Xorg.0.log but a gdm log from 3 boots ago had it and was attached | 21:27 |
bjsnider | Sarvatt, what do i do with the latest blob in x-updates, do i package it normally or as the -updates version? | 22:05 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!