Merging SM501 driver from old xfree86 into xf86-video-siliconmotion

[Clemens Koller]

Hello, Alex!

Alex Deucher schrieb:
> smi501 support really doesn't belong in the same driver as the older
> hardware since the hardware is substantially different. The original
> patch and the code in xfree86 cvs is a mess.

True. The classic "siliconmotion" chips are VGA compatible, whereas
the SM501/SM502 are not.

The SM501 (and it's hw/sw compatible successor, the SM502) chips can
be connected as PCI device (my case) or to some localbus (which will
be more difficult to support because of the different implementations).
The chips assemble a so called MSOC (Multimedia System On Chip) device
and come with USB and AC97 audio support.

The combined xfree86 driver is a mess and there are some run-time
decisions being made due to the incompatible architecture.

>> My questions:
>> 1. Is there already anybody working workin on the SM501 code / pulling it
>> over from the old driver?
> 
> I was working on it a while ago, but at the time I didn't have
> hardware, and since I got hardware, I haven't had time to mess with
> it.  I have a driver available, but I wanted to test a bit more before
> putting it in git.  Also, dualhead support wasn't working and I had
> hoped to get that fixed.  However, I suppose it probably makes sense
> to just put up.  single head works fine IIRC.

I am fine with single head for the beginning.

I am willing to offer an ssh account on at least one machine with
the SM501 development board on PCI, if anybody wants to test
something. (I am planning to do regression tests of the whole X system
on an embedded MPC8540 PowerPC).

> modular:
> http://www.botchco.com/alex/xorg/smi501.tar.gz
> monolith:
> http://www.botchco.com/alex/xorg/smi501-monolith.tar.gz

Very good! Thank you!
I will just give it a try.
(The chip is actually called SM501 and not SMI501.)

>> 2. Is this the right list to get help with the migration?
>> (I was hacking lots of code for the SM501 fb driver, but not for
>> the native X one.)
> 
> yes.  If you want to hack on the driver, let me know and I'll get the
> current code in git and you can submit patches or have commit access
> to the tree.

Yes, that would be fine! I think it's good to have everything published,
even if it's not in perfect condition. A README containing the missing/broken
things is sufficient. Nobody wants to ride dead horses.

Best greets,
-- 
Clemens Koller
_______________________________
R&D Imaging Devices
Anagramm GmbH
Rupert-Mayer-Str. 45/1
81379 Muenchen
Germany

http://www.anagramm-technology.com
Phone: +49-89-741518-50
Fax: +49-89-741518-19