In reply to your message of Wed, 14 Jul 2004 10:42:16 EDT:

Unfortunately, I didn't record my boot syntax attempts in my notes when
working on a beige G3. If your OFW requires a boot-zero method, you can
just ignore everything I'm going to say (it's been suggested that may be a
good policy in general ;-) So, keeping in mind that these are the
instructions for a New World G4/AGP OFW 3.X machine, here's the sequence I
use for a hybrid HFS+/4.2BSD XCOFF boot:

boot-device: hd:9,\ofwboot.xcf
boot-file: hd:10/netbsd
boot-command: boot

This is for the following geometry on a 4G EIDE disk:

This is all Drive Setup information:
Part#	Type		Size
8	HFS+		32 MB
9	A/UX Root	499 MB
10	A/UX Swap	758 MB
11	A/UX Usr	1016 MB
12	A/UX Free	Remainder (~1.75 MB)

Here is the NetBSD disklabel (not written with disklabel per the
installation instructions since that would clobber the HFS+ partition):

7 partitions:
#        size    offset     fstype  [fsize bsize cpg/sgs]
 a:   1022646     67080     4.2BSD      0     0     0   # (Cyl.   70*- 1153*)
 b:   1551766   1089726       swap                      # (Cyl. 1153*- 2795*)
 c:   8421840         0     unused      0     0         # (Cyl.    0 - 8911)
 d:     65536      1544        HFS                      # (Cyl.    1*- 70*)
 e:   3699473   4722357     4.2BSD      0     0     0   # (Cyl. 4997*- 8911*)
 f:        10   8421830    unknown                      # (Cyl. 8911*- 8911)
 g:   2080865   2641492     4.2BSD      0     0     0   # (Cyl. 2795*- 4997*)
disklabel: boot block size 0
disklabel: super block size 0

Now, you will notice a number of inconsistencies between the boot commands
and the actual Drive Setup partition numbers. I put this down to the joys
of OFW, and the need for some iterative trial-and-error when getting
those parameters worked out is why I'm passing them on in the hopes
they may be of some help. OFW already contains the devalias for hd:, so 
I was able to use that rather than hand-coding the full OFW device path; I
suspect you'll need either use the long-form path or create a devalias
since you've got a 3rd party IDE controller.

Also note that I've repeatedly encountered a problem in certain versions
of 3.X OFW where the above parameters work fine from the command line but
fail when used in nvram variables. Historically, I've put that down to an OFW 
bug of some sort and just changed CPU's (which fortunately I have the luxury 
to do) and moved on, but I'd love to know what's going awry there...

Also FWIW, I've found the dir command invaluable when trying to sort all
this out by testing device paths, and to that end having an HFS partition on 
my startup disk turned out to be a nice bonus...

With no truer words than YMMV where Mac OFW is concerned,
Best of luck,
	Paul


| the only thing left to do is figure out how to boot off of the local disk.  i
| t's
| an IDE disk on an ACard PCI IDE controller.  how would i go about getting thi
| s
| to boot from that disk?
| 
| thanks!!!
| 
| -brian


                          Paul "Corwin" Frommeyer
             Work         Internet Engineer, CCIE               Play
     Senior Network Engineer                          Network Sorcerer At Large
   Deskey Integrated Branding                            Paul's Fone Company
     pfrommeyer@deskey.com                                corwin@palas.com
          *** Speaking solely for myself unless otherwise noted ***