Subject: Re: RC1 and framebuffer
To: Allen Briggs <email@example.com>
From: Michael Lorenz <firstname.lastname@example.org>
Date: 09/07/2007 01:08:38
-----BEGIN PGP SIGNED MESSAGE-----
On Sep 6, 2007, at 16:25, Allen Briggs wrote:
> On Thu, Sep 06, 2007 at 08:59:38PM +0100, Robert Swindells wrote:
>> I have some local diffs that improve Q950 ADB support, but not by
>> enough to make it usable. Probing all the ADB devices and shuffling
>> their addresses around works, I just can't get auto-polling to work.
>> I had been meaning to ask you whether I ought to commit them anyway
>> so someone else could take a look.
> If they work well enough for you and shouldn't affect others, then I'd
> say to go ahead.
Yeah, makes it easier for others to continue the work and avoid
> I think it would also be worth looking at the work that Macallan has
> done with ADB on port-macppc.
Maybe - basically I separated the ADB controller parts from bus
enumeration and actual ADB device drivers. The stuff in dev/adb/
shouldn't contain any macppc-isms, all it needs is a way to send and
receive ADB messages.
So we end up with
cuda* at obio?
pmu* at obio?
... which are more or less macppc-specific, and
nadb* at adb_bus?
adbkbd* at nadb?
adbms* at nadb?
adbbt* at nadb?
... which are machine independent.
Main advantages over the existing code are support for more than one
instance of each ADB device, support for controlling keyboard LEDs,
proper handling of things like caps lock, num lock etc., configurable
mouse button emulation, and a lot less cruft.
There are a few things missing though, like apm emulation and support
for smart batteries but at least the latter wouldn't matter on mac68k
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)
-----END PGP SIGNATURE-----