>I'm not sure this can be considered a bug in the radeon driver then,
>unless maybe there's a way for it to detect the unusable AGP aperture
>size.
Prior to upgrading to etch this system ran sarge with X.org packages
(6.9) from backports.org. Those were usable even with the aperture
size set to 32MB in the BIOS. So, this seems to be either a new issue
in X.org 7.1, or a change was made that exposes some underlying
problem that was present before. A 2.6.18 kernel was used in both
cases.
But I must confess that I'm rather ignorant about what the APG
aperture setting really does. The motherboard manual doesn't say much,
and googling for the term seems to indicate it's a performance
optimization. If there are unusable aperture size, ideally either the
agpgart driver or the X server should refuse them. (Unless it's
hardware and/or BIOS bugs.)
Can you suggest a good test case to verify that DRI really works, and
it's not just a matter of postponing the lockup? I've run glxgears
without any problems.
--
Andreas Johansson [EMAIL PROTECTED]
http://www.berget.nu/~andjo/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
