[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#604897: xserver-xorg-core: Segmentation fault on Squeeze with radeon



Michel Dänzer wrote:
> On Don, 2010-11-25 at 09:11 +0100, Tobias Nissen wrote: 
>> After my routine `aptitude safe-upgrade` two days ago I noticed,
>> that I was unable to playback flash videos in fullscreen. Since
>> some X-component had been upgraded I restarted GDM. After that X
>> didn't start anymore at all. The log then showed this:
[Segfault]
>> I then copied all files from my backup
>> at /usr/lib/xorg/modules/extensions to the running system and
>> restarted X -- now it works again. I think something is wrong with
>> libglx.so.
> [...]
>> Versions of packages xserver-xorg-core recommends:
>> ii  libgl1-mesa-dri             7.9+repack-1 A free implementation
>> of the OpenG
> [...]
>> [UPGRADE] libgl1-mesa-dri 7.8.2-2 -> 7.9+repack-1
> 
> Does downgrading libgl1-mesa-dri to sid work around either of the
> problems?

I didn't try that because now I also get the segfault with the old
version...

> Though I'm not sure why restoring the older extension modules would
> make a difference, the xorg-server changelog doesn't seem to mention
> any GLX/DRI2 related changes.

...sorry for the confusion, but as it seems the behaviour is triggered
by my modprobe-configuration. With KMS enabled, resuming from
hibernation often doesn't work with my Radeon Mobility X1400. The
screen is distorted sometimes too. So I disabled KMS with this
configuration in /etc/modprobe.d/radeon-kms.conf:

  options radeon modeset=0
  options radeon agpmode=-1

If I boot with KMS enabled (which seems to be the default for my
system¹) I get a working X system (with the occasional distorted
screen, as before :-/ )

However, a segfaulting X isn't good. Can I do something to narrow down
the problem for you?

¹ I just read in `aptitde changelog xserver-xorg-video-radeon` that KMS
  has been disabled for architectures other than x86. Well, at least on
  my system, which is x86, I have to disable it because otherwise I get
  the problems mentioned above.

Attachment: signature.asc
Description: PGP signature


Reply to: