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

Bug#390630: With xserver-xorg-video-savage 1:2.1.2-5 and EXA X.org shows garbled display



Martin Steigerwald wrote:
> Yes. It also hangs with XAA disabled.
>   

You mean EXA disabled and XAA enabled? Or XAA and EXA disabled (ie NoAccel)?

> KDM after reboot is working correctly. Its about KDM after logout. It 
> doesn't display a login screen then.
>
> I was logged into the machine via SSH to be able to do some diagnostics 
> and I tried /etc/init.d/kdm restart in order to get a login screen. Then 
> it told me that KDM was not running before it tried to start it again. 
> Starting it again didn't work also. I have to reboot the machine twice 
> cause of that.
>   

You should try to start X without kdm in this case. Sometimes, the
display manager has trouble when trying to restart because a problem
occurred during the previous stop. Something like 'startx
/usr/bin/xterm' is good way to check to X server.

> I am using VESA framebuffer driver, not the Savage one, cause I read von 
> savage mailinglist that the framebuffer driver and the X.org one do not 
> handle resetting registers gracefully.
>
> I am using this self compiled kernel
>
> martin@deepdance:~> cat /proc/version
> Linux version 2.6.21.1-tp23-cfs-v11-sws2-2.2.9.13 (martin@deepdance) (gcc 
> version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)) #1 PREEMPT Thu May 
> 10 14:40:21 CEST 2007
>
> with these kernel options:
>
> root=/dev/sda5 ro resume2=swap:/dev/sda7 resume=/dev/sda7 vga=792 
> atapi_enabled=1
>   

You could possibly try with a regular kernel without any framebuffer
driver, but I am not sure it will help here.

> I just could use a bit of guidance on how to proceed with testing and 
> reporting and I will take the time to do so as could as I can without 
> neglecting higher priority stuff.
>   

It's hard to distinguish all the problems here. EXA should probably be
avoided for the beginning. And the respawn problem after logout debugged
first.

If you want to debug all this, it would probably be easier to discuss
directly with upstream, assuming you find a maintainer that really wants
to work on it...

Brice




Reply to: