On my box X dies about once a week (black screen, but the machine is still alive and you can switch to another text console) during normal operation (Web browsing). This is on a ThinkPad X60s. I'll attach more information and am happy to provide whatever is needed. Relevant software version: CFLAGS="-O2 -march=prescott -pipe" sys-kernel/tuxonice-sources-2.6.24-r2 x11-base/xorg-server-1.4.0.90-r3 x11-base/xorg-x11-7.3 x11-drivers/xf86-video-i810-2.2.1 x11-libs/libdrm-2.3.0 Last part of Xorg.0.log: Error in I830WaitLpRing(), timeout for 2 seconds pgetbl_ctl: 0x7ffc0001getbl_err: 0x0 ipeir: 0 iphdr: 2000011 LP ring tail: 12ba8 head: 11bac len: 1f001 start 0 eir: 0 esr: 0 emr: ffff instdone: fa41 instpm: 0 memmode: 306 instps: 800f00c4 hwstam: fffe ier: 2 imr: 0 iir: 20 Ring at virtual 0xa7a50000 head 0x11bac tail 0x12ba8 count 1023 [...] Ring end space: 126972 wanted 131064 (II) intel(0): [drm] removed 1 reserved context for kernel (II) intel(0): [drm] unmapping 8192 bytes of SAREA 0xf8d79000 at 0xb7b82000 (II) intel(0): [drm] Closed DRM master. Fatal server error: lockup
Created attachment 145397 [details] Installed packages
Created attachment 145399 [details] Kernel configuration (tuxonice-sources-2.6.24-r2)
Created attachment 145401 [details] output of lspci -v
Created attachment 145403 [details] xorg.conf
Created attachment 145405 [details] Log of a crashing X session
Julian, Your bug report is pretty much perfect in the sense that there's nothing more I can ask from you but to open a bug in FreeDesktop's bugzilla with all the info you've provided here. Intel engineers have created this document when writing bug reports : http://intellinuxgraphics.org/how_to_report_bug.html If you do open a bug there, please add me CC there so I can keep track of it (for potential patches and whatnot). Thanks
I've been harassed by this issue as well (also Thinkpad X60). Let me know too, how the upstream bug report goes.
Andri, nobody has reported this bug upstream. Feel free to open it like I said in comment #6. Thanks
Just out of curiosity---do you also run the tp_smapi module? I've stopped using that, and have not had any X "crashes" when resuming from hibernate-ram since. I'll monitor the issue for a week or so, perhaps it's just a coincidence---I've never checked what tp_smapi's code exactly does.
On a second thought, the other possible culprit might be the EnableVbetool setting in /etc/hibernate/*.conf files. Julian doesn't specify whether his X dies while running or while resuming---but his tuxonice kernel makes me think it might be while resuming from hibernate.
Julian? Ping? If you want this issue to be fixed, you'll have to open a bug in FreeDesktop's bugzilla as only they have the necessary knowledge to fix this. I strongly advise you to take the time to do it :) Thanks
Somehow reproducible using the .3 as well (I don't own such card so I'll push the owner to get an upstream bug done tomorrow)
Ping, *anyone* ? Bugs won't get fixed if upstream doesn't know about them. Thanks
Since switching to xorg-server-1.4.0.90 and always unloading 'hdaps thinkpad_ec' before suspending, my machine hasn't crashed. Unfortunately, Julian hasn't responded to any questions, so I'm thinking this bug should soon be closed due to inactivity.
Alright. Julian, please try with the latest intel driver and don't hesitate to reopen this bug (along with a bug on freedesktop's bugzilla) if the issue is still there. Thanks
It's sad, because I'm experiencing this issue myself now (also x60s). I may have to check upstream soon if not loading tp_smapi fixes the problem (I was loading it prior to reading these comments). I'll keep everyone here posted...
Erm, if not loading tp_smapi *doesn't* fix the problem. Sorry. ;]
Then please open a bug on freedesktop's bugzilla. It's the _only_ thing that will get this bug fixed. Thanks
I encountered this error as well. My laptop, X60s with i945, recently started to exhibit this behaviour, regularly on window swaps or workspace switches in Gnome, and especially in Firefox (at least 4-5 times per day during regular work). Since Devin did not report back on the tp_smapi issue Andri hinted at, I hereby announce that this was apparently indeed the problem for me! I had quite recently put those modules in the autoload since I was playing a bit with reading the HDAPS sensors with the tp_smapi version of hdaps. I removed all related modules (tp_smapi, thinkpad_ec, hdaps) from the /etc/conf.d/modules autoload line and have not had a single "i830 crash" with those modules not loaded (stable like it was till around the New Year's.. I just could not pinpoint retrospectively what changes I had made to the system - THANKS :-D )