on 3/23/01 1:35 AM, gabriel rosenkoetter at gr@eclipsed.net wrote something
like:

> On Thu, Mar 22, 2001 at 09:40:02PM -0500, Chris Tribo wrote:
<snip>
> Hrm.
> 
> I'm not sure splitting the macppc floppy is even a possibility.
> 
> Since we're incapable of dealing with the majority of fdcs in these
> machines (mostly swim3), would we be able to recognize and load
> anything off of a second inserted disk?
> 
> OpenFirmware isn't around to do the 'boot fd:0' any more at that
> point, is it?

    Maybe, but if a PC can boot that way, why couldn't an OFW machine? I
would think it possible, if OFWBOOT was re-written, but that's much larger
project. Doesn't OFW load OFWBOOT, and then OFWBOOT uses OFW to load the
kernel into a certain memory segment and then execute it? OFW has an "eject"
word that will eject the floppy, I would think you could just take the two
kernel segments, put them "together" and then execute it with OFW. On the
flip side, it's a nearly pointless project, as new-world machines don't have
floppies and old-world ones don't need USB necessarily.

> (Granted, the solution is to get around to writing a swim3 driver,
> but...)

    Well, I've been told that MkLinux (and now Linux-NuBus PPC on
SourceForge?) has a swim3 driver that would be mostly drag and drop into
NetBSD, provided the author(s) don't object. However that still wouldn't
help us load the kernel, unless there was a two stage booter. A two stage
booter would be nice, because it could pave the way to booting off of a
CCD/RAID/CD-ROM on MS-DOS, HFS/HFS+, ISO, UDF
 
> I'll try pulling VERBOSE_ and sound stuff out of the kernel config
> tomorrow in the interest of getting a new snapshot out.

    Snapshot = good   Heh, well macppc isn't as far behind as pmax. I don't
think there has been a pmax snapshot since December 6th of '99 (1.4P).



    Chris
-- 

Food = GOOD
Beer = GOOD
Napster = GOOD
Money = GOOD
True = False