Fwd: xserver-xorg-video-ati git20071220 bug (and solutoin)

[Michel Dänzer]

On Fri, 2007-12-21 at 19:43 -0200, Rodrigo Kassick wrote:
> Hi Michel,
> 
> I'm writing directly to you since I saw in the changelog that you've
> commited the change that most likely caused the problem.

Even so, it's always better to file a bug or post to the xorg-driver-ati
mailing list, so the response to and tracking of an issue doesn't depend
on any one individual. I'm CC'ing the list now.

> I'm using Tormod Volden's version of the ati driver package in Gutsy
> (pulled from git) but since version git20071215 the driver was not
> working properly (version git20071214 was ok). 
> 
> X started, but only to show a black screen and some horizontal lines.
> Mouse moved, but I was unable to leave X to a console (can't tell if
> the rest of the system was working ok, since I had no other machine
> nearby to do an ssh to my notebook). 
> 
> Turns out that the driver was forcing AGP 1x and
> (Flying-Spaghetti-Monster knows why) this does not work for my radeon.
> I forced AGPMode to 4 in my xorg.conf and it's working now.

Sigh. A little background:

The driver originally always defaulted to the lowest rate supported by
the setup. Then we got reports of some AGPv3 setups only working with 8x
but not 4x when 8x was set in the BIOS setup. So we changed the default
to leave the rate unchanged. Then we got reports of that breaking setups
that work with 1x, so we changed the default back to 1x for setups that
support it. Now we got your report... seems like we're finally
officially screwed wrt AGP defaults, or at least missing something
important.

Maybe we should just default to PCI mode. :}


-- 
Earthling Michel Dänzer           |          http://tungstengraphics.com
Libre software enthusiast         |          Debian, X and DRI developer