--+QahgC5+KEYLbs62
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Sep 26, 2003 at 12:23:54AM -0400, Allen Briggs wrote:
> On Thu, Sep 25, 2003 at 09:59:29PM -0500, Zach Lowry wrote:
> > I got to thinking the other day about writing support for NuBus PPC
> > machines. Since we're in need of a bootloader, what about using the
> > mac68k booter app? Would major changes need to be made? It seems to
> > be a better choice than using BootX, since it already understands UFS
> > and etc, plus we have the formatter and installer utilities.=20
>=20
> It's a neat idea, but there are a lot of changes that would have to be
> made.  It wouldn't really save any time on the BSD side, and large parts
> would have to be rewritten on the Mac side.  The boot process itself
> would have to be entirely different, and there would have to be some way
> to get into supervisor mode (not an issue on MacOS/m68k).
>=20
> With BootX, I think we'd be looking at a multi-stage boot, not too
> unlike what we do now on a number of platforms, including OF-based
> PPC.  We'd point BootX at the beginning of a partition, teach it how
> to load a few blocks of code to a known location, make sure caches
> are flushed, and jump into the loaded code.  That code would understand
> the UFS/FFS/LFS/RAIDframe/whatever, and really be the equivalent of
> ofwboot.{xcf,elf}.

I'd also vote for BootX, from what I know of it. One of the main
issues for NuBus macppc will be that we can't call into OF; our
kernel loves to.  I think everything we'd need can be obtained from
a static OF tree, which BootX will pass in. It of course will be
faked in a known way for NuBus machines.

Take care,

Bill

--+QahgC5+KEYLbs62
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQE/dMBbWz+3JHUci9cRAnQyAJ9rIjP87W+fGDNLceQzLnF+utbtlwCff1cV
to+RJ5Ckdkkeu0odsdG81+0=
=Dwrz
-----END PGP SIGNATURE-----

--+QahgC5+KEYLbs62--