On Tue, 23 Feb 1999, Dr. Bill Studenmund wrote:

> On Tue, 23 Feb 1999, Mike Frager wrote:
> 
> > The advantage of the imposter Apple_Driver is that it is loaded very early
> > and would be simple to implement since the documentation for writing an
> > Apple_Driver is easily available. The downside is that we would have to
> > create a user space utility in NetBSD to manage it.
> 
> I think this method would be the best of the two. It's been a
> while since I read IM, but I seem to recall that Apple added hooks 
> to boot other OS's at about this stage. I thought it was a little
> different in that an Apple_Driver is trying to be a MacOS disk driver,
> while the other-os thingie is trying to be an OS-booter. The difference
> being that if you choose to boot MacOS, with the booter different from the
> disk driver, you could still have a macOS disk driver.

Even if Booter-Partition support proves incomplete, the driver should be
hackable enough. Perhaps it could just walk the partition table for a
bootable NetBSD partition, and if it doesn't find one, bow out and unload
itself. You could boot MacOS by unsetting the "bootable" flag on the
NetBSD partition. There are already several MacOS tools which can do that.
 
> Probably the best thing to do is to add two partitions, one containing
> code the other preferences.

FWB tools already does this. The problem with that is that it cuts into
the (practical) limit of seven partitions per disk. That seems to be a bug
that appeared in 1.3. Prior, the extra partitions were simply ignored.

I'd also really like to see NetBSD disklabels supported. There could be
just enough room for one in block 0. The Mac firmware ignores block 0
entirely--the bottom few bytes need to be left for the scsi mode pages,
and LaCie writes it's settings and drive icon into the top. NetBSD can
already write a raw label, it just doesn't read it. Once you have the
flexibility to use a custom disklabel, many of the limitations vanish.