Subject: Re: Disklabel(5)/(8) ??
To: The Great Mr. Kurtz \[David A. Gatwood\] <davagatw@Mars.utm.edU>
From: Scott Reynolds <scottr@edsi.org>
List: port-mac68k
Date: 10/11/1996 00:54:11
On Tue, 8 Oct 1996, The Great Mr. Kurtz [David A. Gatwood] wrote:

> On Tue, 8 Oct 1996, Hauke Fath wrote:
> > sd230|mo230|mo230-iso|Fujitsu M2512A Magneto-optical:\
> >         :ty=optical:dt=SCSI:\
> >         :ns#40:nt#8:nc#1394:su#446324:se#512:\
> >         :pc#446324:oc#0:bc#8192:fc#1024:tc=unused:\
> >         :pg#446323:og#1:bg#8192:fg#1024:tg=4.2BSD:
>                                              ^^^^^^
> I assume that the above tg is to define the label format, which for disks
> formatted normally would default to... Macintosh?  Mac4.2BSD?  AUX4.2BSD?

Actually, that "tg" means "type for partition `g' is 4.2BSD" -- i.e. the
normal BSD FFS.  This has nothing to do with "label format."

> The point I was trying to make (poorly) was that if writing to the MacOS
> partition table were supported in the future, some means would be required
> to tell disklabel whether it was writing a BSD-type disklabel or a Mac
> partition table label, and more importantly, be able to override whatever
> was on the disk initially (since the read would use an auto-check).

This is more or less how the i386 port works, though that code is a little
weird sometimes.  (Now that I've been bitten by it and subsequently picked
apart the code, I understand why... but there's certainly room for
improvement.)  At a bare minimum, I'd like to be able to support a BSD
label _inside_ the AU/X Root slice 0 partition -- this has potential to
give us both a quicker install process and to simplify resizing of
filesystems.  This really has little to do with being able to write a
native Mac partition table.

> Which brings up another question: anybody know anything about (or have
> sources of info about) "Foreign File Access"?

Foreign File Access is the common layer for CD-ROM access.  The various
CD-ROM filesystems (ISO9660, audio, Photo CD) are all implemented on top
of FFA.

> Which brings up yet another question: is there any chance of adding code
> to distinguish an ext2fs from a ffs, and perhaps even making MkLinux
> partitions mountable under NetBSD (and vice versa)?  Ideas?

If there is, I would expect it to be a machine-independent solution.
Hopefully someone with access and time will work on this.

--scott