FS#18319 - [xf86-video-intel] segfaults

Attached to Project: Arch Linux
Opened by none (foxdie) - Saturday, 13 February 2010, 15:19 GMT
Last edited by Jan de Groot (JGC) - Thursday, 17 February 2011, 15:22 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Jan de Groot (JGC)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 4
Private No

Details

Description:

I've done a fresh install of Archlinux on a laptop with an Intel 4500MHD. Hal and KMS are turned on, and Xorg is configured by Hal, but, sometimes, when initialising KDM the resolution isn't correct (it should be 1280x800), and other times, Xserver segfaults (when Arch boots or when I switch from tty to X). The attached file is the log of Xorg with a segfault when kdm starts at boot.
And, if I type:
'cat /sys/module/i915/parameters/modeset' the result if 1
but, when I do an 'Xorg -configure' there is an (EE) with intel(0): no kernel modesetting driver detected.

Additional info:
I'm using:
KDE 4.4
xorg-server 1.7.4.901-1
xf86-video-intel 2.10.0-1
hal 0.5.14-1


Steps to reproduce:
This task depends upon

Closed by  Jan de Groot (JGC)
Thursday, 17 February 2011, 15:22 GMT
Reason for closing:  Fixed
Comment by Adria Arrufat (swiftscythe) - Saturday, 13 February 2010, 21:01 GMT
I have a similar problem, which I think is the same bug.
Xconfigure -a complains about the same thing.
If I enable KMS the screen goes black at boot when loading modules, which makes the computer (tablet pc) unusable.
If I disable KMS, X crashes
I can only startx with the vesa driver instead of intel.
Additional info here:
http://bbs.archlinux.org/viewtopic.php?id=90914&p=1
Comment by Jan de Groot (JGC) - Saturday, 13 February 2010, 22:26 GMT
Disabling KMS is not supported. KMS bugs should go here: http://bugs.archlinux.org/task/18298
Only comment if you actually have anything useful to add.

@reporter: can you attach dmesg output when KMS fails to initialize correctly? Looks like the kernel can't detect any outputs and fails to setup KMS correctly, after which the driver will crash or segfault when it fails to find KMS support. You might want to add boot options to force outputs and resolutions.
Comment by none (foxdie) - Sunday, 14 February 2010, 04:32 GMT
@Adria, what are you talking about happens to the laptop of my girlfriend, with the same graphics card (and configuration) as me (Intel 4500 MHD).
Comment by none (foxdie) - Sunday, 14 February 2010, 04:47 GMT
Xorg and dmesg when X segfaults. Jan de Groot, I know that this is not Arch related, because I switched from a gentoo testing where the same thing was happening. This behaviour started some days after kde 4.3.98 was released in addition of new kernel fixes (gentoo-sources-2.6.32-r4 which contains kernel 2.6.32.8 patches), but I don't know if this is related. Tomorrow, (here are 5:43 AM), I will test with an old kernel version and I will post the results.
   dmesg (31.3 KiB)
   xorg.log (8.6 KiB)
Comment by none (foxdie) - Sunday, 14 February 2010, 13:08 GMT
I think I've solved it. I've spent the last 30 minutes rebooting the computer and switching from tty to KDM with kernel 2.6.32.7 and the behaviour described here doesn't happen anymore. Should I notify kernel guys? Thanks.
Comment by Jan de Groot (JGC) - Sunday, 14 February 2010, 13:19 GMT
You might want to have a look at the stable queue upstream:
http://git.kernel.org/?p=linux/kernel/git/stable/stable-queue.git;a=tree;f=queue-2.6.32

There's several i915 patches there, if they fix your problem this will be fixed when 2.6.32.9 is released. If not, you should check which of the 3 i915-related patches in 2.6.32.8 broke your setup.
Comment by Paul Mattal (paul) - Friday, 05 March 2010, 13:01 GMT
Did 2.6.32.9 resolve any/all of these issues?
Comment by Adria Arrufat (swiftscythe) - Friday, 05 March 2010, 13:48 GMT
Hi! Kernel 2.6.32.9 did not solve anything, but I installed 2.6.33 as soon as it reached the testing repo, and I've been using my tablet PC with that kernel, and works perfectly. In the 2.6.33 changelog, it was explicitly said that that bug was fixed.

Thanks :)
Comment by none (foxdie) - Saturday, 06 March 2010, 01:15 GMT
  • Field changed: Percent Complete (100% → 0%)
This is not closed yet. The bug which Adria is talking about isn't the same that I was talking about. I've opened a bug request on bugzilla.kernel.org [http://bugzilla.kernel.org/show_bug.cgi?id=15417]. Also, as i've posted on bugzilla, kernel 2.6.33 partially closed this bug.
Comment by Thomas Dziedzic (tomd123) - Monday, 28 June 2010, 20:50 GMT
Upstream bug report http://bugzilla.kernel.org/show_bug.cgi?id=15417 says it's fixed with 2.6.35 rc2.
Comment by Thomas Dziedzic (tomd123) - Monday, 16 August 2010, 22:20 GMT
status?

Loading...