I am using WindowMaker on my Pentium based Gentoo Linux system. The whole system is built with the ~x86 keyword. After doing a large update, Xorg is hanging some time after startx. I have noticed that doing a lot of graphic intensive tasks expedites the hang. For example, if I just do startx and let it sit, it didn't hang for 10 minutes. But if I open firefox and start browsing websites or if I rapidly move windows around, the hang happens fairly quickly. When the system hangs, I can move the mouse cursor, but the keyboard isn't working and the visual elements of WindowMaker don't respond to mouse clicks. I have rebuilt the xf86-input-keyboard, xf86-input-mouse and xf86-video-intel drivers after upgrading xorg-server. I have also upgraded the kernel to 2.6.30-r4 and have adjusted the /usr/src/linux symbolic link before building these. I have attached files that contain information about my system. the Xorg.log might not be a good log, because I am doing a hard reset when the system hangs and sometimes the log ends up containing binary garbage. Reproducible: Always Steps to Reproduce: 1. Reboot 2. startx 3. open firefox or move any window around rapidly Actual Results: The system hangs. Mouse cursor moves, but keyboard is dead. The Ctrl-Alt-Backspace combination does not work to kill the xorg-server. Expected Results: No hang I am attaching all relevant system information to this bug. Let me know if additional information is required.
Created attachment 200506 [details] emerge --info output
Created attachment 200507 [details] Xorg log file
Created attachment 200508 [details] xorg.conf file
Created attachment 200510 [details] Linux boot log
Created attachment 200512 [details] lspci output
Created attachment 200513 [details] cat /proc/cpuinfo output
The version of the intel video driver is 2.8.0.
The xorg.conf file is generated using Xorg -configure. I know that this system is a ~x86 system and that bugs such as these are to be expected. But I want to work with the developer to fix this problem. I can provide any additional information required and I can try test releases when available.
Test: Section "InputDevice" Identifier "Keyboard0" # Driver "kbd" -- change to xorg > Driver "xorg" EndSection
or set INPUT_DEVICES to "keyboard mouse evdev" and xorg-server emerge with hal flag , and try remove xorg.conf
(In reply to comment #9) > Section "InputDevice" > Identifier "Keyboard0" > # Driver "kbd" -- change to xorg > > Driver "xorg" > EndSection there's no "xorg" driver, you must be confusing things. @Timur, there's little I can do to help you here. Please file a bug in FreeDesktop's bugzilla [1] where Intel developers will be able to figure this out better than me. Once you do, please paste the url here so I can track the issue too. Thanks [1] http://intellinuxgraphics.org/how_to_report_bug.html
Just for the record. I tried changing "kbd" to "xorg", but this made the keyboard stop entirely. Before, the keyboard was working up to the hang.
@Remi, I have searched the freedesktop bug database and indeed, there is one bug that seems to be my bug: https://bugs.freedesktop.org/show_bug.cgi?id=21470 I will continue there and once there is a resolution, I will let everybody know.
A few tests that I have done today: I have tried downgrading the intel video driver to version 2.7.1 and the problem still happened the same way. Then I tried booting linux 2.6.30-r1 (I kept the intel video driver at 2.7.1) and this time the problem hasn't happened. So the problem must have been introduced between kernel 2.6.30-r1 and 2.6.30-r4. I will write about this in the freedesktop bugzilla, but can anybody working closely with the kernel sources tell me what the significant differences are between these two versions of the kernel? For example, drm related changes etc...
(In reply to comment #12) > Just for the record. I tried changing "kbd" to "xorg", but this made the > keyboard stop entirely. Before, the keyboard was working up to the hang. The "xorg" driver doesn't exist. What Ondrej said in comment #9 was wrong, no wonder it didn't work :) (In reply to comment #14) > So the problem must have been introduced between kernel 2.6.30-r1 and > 2.6.30-r4. I will write about this in the freedesktop bugzilla, but can anybody > working closely with the kernel sources tell me what the significant > differences are between these two versions of the kernel? For example, drm > related changes etc... You might want to use git to get a full copy of the kernel sources and do tests on your own. The "git bisect" command is extremely useful for such purposes. Thanks
And reassigning to the kernel team as they're the ones in charge of - you guessed it - the kernel. :) I'll CC myself on the upstream bug to make sure only the kernel is responsible for the hang. Thanks
After letting the system run longer (3 hours!), it came out that the problem happens even with kernel version 2.6.30-r1. Here is another bug entry which explains the same problem: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=527349#27
Can you try booting the latest 2.6.31-rc kernel, and report if the bug persists?
