framebuffer device drivers

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

Hello,

one new feature in NetBSD 4.0 will be support for hardware
acceleration 
in framebuffer console drivers - there's one problem though,
I can't 
possibly test them all on all possible hardware.
The following drivers need testing:

chipsfb - supports the Chips & Technologies 65550 found
in the 
PowerBook 2400, 3400c and original G3. Can't set video modes
so far and 
XFree86's chips driver doesn't restore enough status for us
to make the 
console usable again. Besides that it should Just Work.
Question remains - is there any other Mac hardware which
uses this 
chip? Does it work properly?

voodoofb - supports the 3Dfx Voodoo3. This driver can switch
video 
modes so getting back from X should work fine. I don't
expect much 
trouble here.

machfb - supports lots of ATI Mach64-based chips. There are
exceptions 
though, mostly because there are so many different mach64
chips and 
many need special treatment of some sort. So, if you have
this kind of 
hardware please report:
   - does machfb attach?
   - does machfb work properly?
   - pcictl dump output for the chip in question
XFree86's atimisc driver behaves well so getting out of X
should give a 
usable console

radeonfb - should support most Radeons up to the R300
family. Tested 
mostly on R2xx chips though, mainly because that's the
hardware I have. 
So, just as with machfb please report success or failure and
include a 
pcictl dump. This driver should re-initialize the console
when X exits.

The main purpose of this exercise is to keep drivers from
attaching to 
chips we don't support properly - these should still work
with ofb - 
and to see if there are chips out there we should attach to
but don't.

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRqJHp8pnzkX8Yg2nAQJQLAf9GChRGkEwuxSkcQjzWeXKFiquH5t1
9FV9
KsHu/pSmHYhae5GwqxallPvDtxSnRW+pSev5UQ1qIIbAFGs/KNcTZ7cyND+a
n9/G
jnPI0sAh4LtfUQjSYTLDp6ncor1yvu76FmIOjqRKuvIMtHGmnQOQUUcWAx1d
BR2G
hpPeeTTSMr2GfGMfjcLICwLkXVVA4tN/ifPcgnK6e03PzJv0eni7tg7LUBrC
5FRL
4VnCa97j07PiMWhlyfHpf5xcSyM1vcTkzQoXRULArvWn554yTaXW5mzYVqfI
qIRS
2ITQU4kOYs6a+WB3lq6XUUz8mLZ5Ll/jtFdyBlLPnHp0QhPIA9wqjw==
=dFwN
-----END PGP SIGNATURE-----

 	Has anyone put together a NetBSD/macppc 'live' CD image?
 	It could be ideal for getting people to test situations
 	like this without messing with their installed system.

-- 
 		David/absolute       -- www.NetBSD.org: No hype required
--

On 21-Jul-07, at 1:51 PM, Michael Lorenz wrote:

> radeonfb - should support most Radeons up to the R300
family.  
> Tested mostly on R2xx chips though, mainly because
that's the  
> hardware I have. So, just as with machfb please report
success or  
> failure and include a pcictl dump. This driver should
re-initialize  
> the console when X exits.

AFAIK this still has an issue with using certain widescreen
modes  
over DVI -- I ran into this on my Mac Mini G4 1.5GHz with a
DVI  
monitor at 1680x1050. I believe there's an XXX about it in
the driver  
itself.

Cheers,
Jared