On Thu, 17 Aug 2000, Kurt J. Lidl wrote:

> On Thu, Aug 17, 2000 at 09:37:13AM -0700, Eduardo Horvath wrote:
> > On Mon, 14 Aug 2000, Gerald Richter wrote:
> > 
> > > I wasn't thinking of something for major production, just a niche
> > > product for those of use that wanna play with more bits in the graphics
> > > card... Kinda like that G4 slave core I wanted to build for my 4/110 if
> > > I ever got that going...
> > 
> > Let's ignore the cost of producing a one-off board for now.  (You could
> > probably buy a new U5 for less).  You would need to get XFree86 working on
> > SPARC.  Do you really want to mess with that can of worms?
> 
> Why do you think this is so hard?
> 
> I had Xfree86 3.3.6 running on the BSD/OS 4.01 (sparc) version -- it
> only took me a little bit of dinking to make it work.  Mostly educating
> the system that __bsdi__ != little endian!
> 
> The guys at Redhat that have been doing the Sparc work have
> contributed drivers for the following sparc board under XFree86
> 4.0[.1].  Now, these drivers would require some work to get working
> under BSD-ish systems, as they have gone down the path towards the
> Linux fbcon code, but I don't think they are a lost cause.
> 
> There are currently drivers for the following sparc graphics boards:
> 	sunbw0
> 	suncg14
> 	suncg3
> 	suncg6
> 	sunggb
> 	sunleo
> 	suntcx

Those are *NOT* using XFree86.  Those are all supported by Xsun or Xsun24.

If you take a PCI graphics card and stick it in a PCI slot in an
UltraSPARC workstation then you would need to use XFree86 instead of
Xsun.  That is a huge can of worms, especially 4.0 with the screwey
dynamic linker.


Eduardo Horvath