-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


I have two in my B&W G3 that I use, but I could loan one or purchase  
one on your/NetBSDs behalf. I also have the Apple OEM 2x Rage 128  
from my Sawtooth hanging around that I can swap back in for testing  
if need be. (Right now it has the Apple GeForce 2 MX in it)

The Rage128y I have used in my Beige G3 as well, so I know it works  
with at least OF 2 and 3.

ATY,Rage128:

   Name:	ATY,Rage128n
   Type:	display
   Bus:	PCI
   Slot:	J11
   Vendor ID:	0x1002
   Device ID:	0x5245
   Subsystem Vendor ID:	0xb530
   Subsystem ID:	0x0408
   Revision ID:	0x0000

ATY,Rage128:

   Name:	ATY,Rage128y
   Type:	display
   Bus:	PCI
   Slot:	J12
   Vendor ID:	0x1002
   Device ID:	0x5245
   Subsystem Vendor ID:	0xb530
   Subsystem ID:	0x0408
   Revision ID:	0x0000


On Dec 6, 2005, at 7:58 AM, Michael wrote:

> Hello,
>
>> Yep, but the driver should be very similar.
>
> Not /that/ similar. There's a reason why X has separate drivers for
> mach64, Rage 128 and Radeon. At least the R128 code in XFree seems
> readable enough to serve as documentation.
>
>> but I'll mail Yun Young to know more on this framebuffer.
>
> He didn't touch machfb for a while ( and I wrote the acceleration  
> code )
> To get this supported we need:
> - some changes in macppc's console handling to allow other drivers at
>   all. Has been done but it's sort of a hack, currently the kernel is
>   hardwired to start up ofb very early, before probing any PCI bus.
> - some guy who writes the driver. I can't do it since I don't have any
>   Rage 128 that would work in any of my Macs ( there's a PCish AGP  
> R128
>   sitting here but that won't help since we need OF console support  
> and
>   I don't have an AGP Mac anyway. If for some reason a PCI R128  
> with Mac
>   firmware would suddenly materialise on my doorstep that would be a
>   different story and we'd probably have a driver a few days later )
> - you could do it yourself. For macppc you can ignore all the mode
>   switching stuff since OF will set up some graphics mode, usually in
>   8bit ( with newer firmware it also supports 24bit ) and tells us
>   exactly where we find the framebuffer and what it looks like, so all
>   you'd have to  do is map the registers, probably map the whole
>   framebuffer ( OF would map only what it needs to display the console
>   ), write some basic functions to identify and set up the chip,  
> copy or
>   fill rectangles, probably one to draw characters using the  
> blitter and
>   steal the virtual console code from some other driver ( since many
>   drivers use very similar code here it should probably be moved out )
>   and that's more or less it.
>
> have fun
> Michael
> !DSPAM:43958b0022451174472460!

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)

iD8DBQFDla2HBZaZ8/mXbSkRAgQ8AKDczOvn7GURhp/OwQ/YWex9ArQpFgCg7phc
rK8TSI8G9uPBhElGv2IV16c=
=UPmp
-----END PGP SIGNATURE-----