[ltp] KDE crashes but not Gnome after suspend

[Ivarsson, Torbjorn (T)]

> -----Original Message-----
> From: Tod Harter [mailto:tharter@giantelectronicbrain.com]
> Sent: Friday, March 26, 2004 4:26 PM
> 
> Well, my desktop box reports those same type of atkbd.c and 
> mtrr overlap 
> errors too. ATI's fglrxconfig script makes mention of some 

I did some research and the "atkbd.c: Unknown key released. This is an XFree86 bug. It shouldn't access hardware directly."-bug is a XFree86 bug. See
  http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=115769
for more info. I don't think that's the cause of my problem.

I am currently investigating as to why KDE crashes (KDM/XFree goes into a restart-loop, where I occasionally get XDM and can shutdown) after an apm suspend/resume, whereas Gnome does not crash. I have done some test-cases and are now going through a bunch of log-files...

One thing I've noticed is in my /var/log/XFree86.?.log is that for the log-files generated by the crashing KDE/XFree, I see the following two lines:

  (II) RADEON(0): [drm] drmSetBusid failed (7, PCI:1:0:0), Device or resource busy
  (EE) RADEON(0): [dri] DRIScreenInit failed.  Disabling DRI.

In a successful KDE/XFree start, the log-file has more [drm] and [dri] (as well as [agp]) lines.

I have a sneaky suspicion that somehow DRI get messed up when suspending and can't be reestablished during a resume. I've seen on another R40 web page that someone is using is using the ATI DRI driver from http://dri.sourceforge.net/.

This may be a stupid question, but, is there is a difference between this DRI driver and the radeon driver my system currently has (kernel 2.6.3-7mdk)?

If there is a diff, should I try to install the Sourceforge DRI? (My successful XFree86 log says "RADEON(0): Direct rendering enabled") For a newbie, are there any quick pointer on how to do the compilation and installation if I have to install it?

I'll spend the weekend going through the rest of the log-files... :)

Any help or suggestions are welcome.

Thanks,
T.