On Sat, 30 Jul 2005, Michael wrote:

> Wait a second. There's one thing I didn't try yet. Your XFree86.0.log is
> full of messages like this:
> (II) PCI Memory resource overlap reduced 0x00004000 from 0x00007fff to
> 0x00005fff
> - I don't see a single one of them. When XFree messes with PCI BARs
> machfb won't notice and probably mmap wrong address ranges. I'll mail
> you a patch for machfb to test this.

Cool, please do so! Btw, this X and NetBSD is compiled from the sources of 
-current of about July 20th.

Unfortunately, xorg-servers will not build as various things require 
"machine/fbio.h" (e.g. in the xc/programs/Xserver/hw/sun directory) which 
is not present. I havent look at commenting sun out from the makefiles 
yet.

>> Until I got the Pro, perhaps having a look at ATIMach64Probe() might
>> help  if I at least add some debug print output to see where exactly
>> it fails...
>
> This would indeed help a lot - I can't do anything here. I had a look at
> the code, it sure looks funky but without being able to reproduce the
> problem that's rather pointless.

I have two of them and I could donate one ... well ... but guess sending 
it is more expensive than grabbing one for EUR 2 or so :)

-Florian