On Tue, 1 Jan 2002, erik vogan wrote:
> the other bit that might help is this: when the screen gets into this
> state, everytime i try to move the cursor to the vesa driven screen, the
> corrupted bits change in time with cursor movement, almost as if the vesa
> driver now thinks the display memory is mapped over the nvidia screen.
Well, if it's doing legacy vga (planar) or banked linear access
the vesa driver would be rendering to the 64K vga aperture. There's
only one such aperture and it would point to either the S3 or the
NVIDIA card depending on how the RAC code set it. It's possible
that there was a bug related to this back then. I suppose Egbert
would say you should be using 4.1.0.
I suppose this could even be a bug in the nvidia driver. If
the nvidia driver forced the aperture to point to its own framebuffer
(I don't know why it would do that), it could mess up RAC's accounting
and the vesa driver could very well be rendering to the nvidia
framebuffer. You should try it with the "nv" driver for a sanity
check, just to make sure that's not happening. I know the "nv" driver
isn't touching the pci configuration space (except in one place
on nForce where it needs to read it to figure out how much memory
there is).
Mark.
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
