Subject: Re: Imaging.
To: Allen Briggs <briggs@wasabisystems.com>
From: Nyef <nyef@softhome.net>
List: port-mac68k
Date: 01/02/2002 17:35:21
On Wed, 2 Jan 2002, Allen Briggs wrote:

> On Wed, Jan 02, 2002 at 10:31:46AM -0700, Don Yuniskis wrote:
> > So, when is the "MacOS partition" requirement going to go away?
> > Is anyone working on that?  Is it a problem of lack of information?
> > Or, just lack of resources (man hours)?
>
> Basically, when someone gets to figuring out how to do this.  It seems
> quite possible from the documentation, but it's hard to know what bogons
> will be there until it's attempted.  The possible pitfalls that I'm
> aware of:
>
> 	1. Things may not work as documented (this could be on a per-
> 	   machine basis; i.e., some machines work one way while others
> 	   don't work, or work differently).

Then the systems that work differently can use the MacOS booter until we
figure out what's going on with them and work around it.  Even saying
"yes, you can boot without MacOS, but only on an SE/30" is an improvement
over the current situation.

> 	2. Things may work differently depending on which disk driver
> 	   is installed on the boot volume.

On a related subject, would it be feasable to use a BSD-type partition
table if MacOS isn't going to be installed on the system? (the boot code
would have to be modified to use the SCSI Manager directly, but that
shouldn't be too hard...)

> 	3. It may be difficult to get all of the information that the
> 	   booter does about the system at boot time.

Is there an actual list somewhere of what this information is, or must it
be grubbed out of the booter source?

> 	4. Early machines may not be in 32-bit mode and will have to be
> 	   switched, if possible.

There's an API for this (SetMMUMode), which is known to be used by some
NuBus video card drivers. It probably won't be too much of a problem.

Actually, I've been wondering. Why is it that the Booter requires the
system to be in 32-bit mode anyway, given the existance of this API?

> 	5. Some system patches may be required for some aspects of the
> 	   hardware to behave well.

Mmm... And while these patches might load normally, if we use
SetOSDefault(), things won't be normal.

On the other paw, if this is the case, we could insert our own patch into
the partition map, and have it load all the MacOS patches that failed to
load. Or we could add a boot menu...

> Anyway, it would be really nifty if someone could get it working.  I
> know at least one person at least started on it, but I don't recall
> seeing anything recently (not that I've been paying close attention ;-).

I'm still here, and still bash on it occasionally, but haven't made much
progress over the past six months, which explains why you haven't seen
anything about it recently.  (This is also why you haven't heard anything
about a SWIM driver, I haven't managed to get it to find an address mark,
don't know why not, and haven't had enough time/patience to track it
down.)

> We won't know if any of the above are really an issue until someone
> tries it.

Hopefully this year.

> -allen

---------------------------
All programming can be viewed as an exercise.
---------------------------
Alastair Bridgewater
e-mail: nyef@softhome.net