Re: [Dri-devel] Re: Recent r128 2D in trunk

[Kostas G. <gew...@im...>]

On Fri, 29 Sep 2000, Michel [iso-8859-1] Dδnzer wrote:

> > There was a patch very recently to the XFree86 CVS tree that looked like
> > it would fix the DDC failure in the r128 driver.  Try that patch and see
> > if you get past this problem.
> 
> I will, I just don't let the driver fail if the DDC init function fails for
> now.

The DDC code will never work in ppc as it is now , it depends in vbe/int10
which requires a x86 BIOS in the card. Note that for now it will fail for
cards with x86 BIOS as well since ISA IO/MEM accesses don't work for
linux/ppc.
 
> > The new Rage 128 Mobility code is on the trunk.  The driver now reads
> > the display mode parameters from the video BIOS mode tables since those
> > are the only parameters that are guaranteed to work with the graphics
> > chip and flat panel in your system (this info is directly from ATI).
> > I'm not sure how this will affect the PPC code.  If it cannot access the
> > video BIOS, then I'm not sure if/how it will work.
> 
> Apparently there is no video BIOS at all, we can only determine the PLL values
> from what OF set them to. Do you think this could be done for the FP
> parameters as well?

  The OF method wasn't working perfectly we were still leaving some of the
PLL values at the default setting :(. It will be better to check if there is
anything we can use in the OF rom, the values might be there as well either
in a table similar to the x86 one or in lost somewhere in forth code ... 

  Kostas