On Wed, Mar 07, 2001 at 10:50:01AM -0800, Benjamin Kite wrote:
> Solution:
>
> 1. If anybody else is still having trouble with the HFS boot disk
> locking the keyboard (or not recognizing ADB input) on an old-
> world macintosh, there is a bootable floppy floating around that
> fixes this. You can e-mail me if you can't find a copy. I think
> it was available from http://www.netfall.com/.
yup its broken, no there isn't a fix. though Andrew resurected an
older 2.2.17 boot floppy that still works. hopefully some solution
will be found for potato r3 but it isn't looking hopeful.
> 2. I can't seem to boot my linux partition on /dev/hdb2 even though
> I have (supposedly) both quik and yaboot installed. I've even
> made sure the bootstrap partition exists and has yaboot files on
> it. I don't know for sure how to test if quik worked or not.
don't bother with yaboot, and don't do anything with the bootstrap
partition. they aren't used on OldWorlds only newworlds. its still
good to have the bootstrap partition created on an oldworld in case
you ever move the disk into a newworld box you will be able to make it
bootable. (since newworlds are the exact opposite, they can't use
quik or any bootblock based loader, only yaboot and an Apple_Bootstrap
partition)
oldworld OF requires varying ammounts of black magic and witchcraft
depending on what machine you have. to start debian's installer will
install quik on the disk, but it doesn't bother to attempt fixing the
boot-device OF variable, this is the first required step to quik
booting.
> 3. Are there any success stories out there?
> What is the most explicit device name for one of these drives?
>
> ATA-Disk@x,y:z
i have not seen such a OF device path before... try ofpath /dev/hdb
to get an idea what the device path is. for boot-device this will
work (assuming ofpath /dev/hdb worked):
nvsetenv boot-device "$(ofpath /dev/hdb)0"
or
nvsetenv boot-device "$(ofpath /dev/hdb3)"
on some oldworlds you need to change the load-base and other
variables, im not sure what and which, the netbsd FAQ is helpful
here. (don't have the url handy)
> I've seen a comma in the device tree, but I wonder if that's
> not a red herring. It seems clear that it's supposed to be
> x=unit and z=partition number. Is partition 0 the same as the
> first partition after the partition map? That makes /dev/hdb2
> come out to "ata/ATA-Disk@1:0". That's not working for me.
what i have usually seen is something like ata/disk@1:0
you may also need OF patches to make OF read the disk properly.
> Detail:
>
> I have a powerbase 160 and would love to get Linux on it without having to
> find a CD-ROM of MacOS that will work with it. They're getting pretty rare.
it can be done, its just a tad tricky.
> Secondly, I've tried creating an Apple_Bootstrap partition for yaboot and
> a quik bootblock header on my root filesystem, but still don't seem to be
> able to boot into the OS without the install system.
don't even try to boot yaboot, just leave the bootstrap partition
alone just in case you move the disk to a newworld box. for now
ignore it.
> I've tried a variety of boot-device setting, including the obvious choices:
>
> ata/ATA-Disk@0:0 (For some reason I thought that this might boot the first
> bootable partition)
yes partition 0 should find the bootable partition.
> ata/ATA-Disk@0:2 (since the root partition is on /dev/hdb2)
> ata/ATA-Disk@0:8 (since the yaboot bootstrap partition is on /dev/hdb8)
don't bother here, ignore the bootstrap partition oldworld OF cannot
use it.
> ata/ATA-Disk@1:2 (since it's /dev/hdb2 and not /dev/hdba2 -- who knows)
> ata/ATA-Disk@1:8 (" ")
> (responds with: "DEFAULT CATCH!, code=FFF00300 at %SRR0: FF80BFD8 %SRR1: 00008070")
> (I assume this means there's no drive there.)
if you booted via auto-boot it could mean the disk was not spun up
when OF tried to read from it, Oldworlds did not boot via OF, OF just
kicked in the MacOSROM which had its own methods for reading the
disk. thus apple did not bother to make OF read disks correctly.
there is a small forth script to use as the boot-command around that
adds a delay to fix this. otherwise it probably means some variables
need to be changed, such as load-base. you also might need OF patches
but i don't know if any of apple's patches work on this machine.
--
Ethan Benson
http://www.alaska.net/~erbenson/
Attachment:
pgpUvIswnld_P.pgp
Description: PGP signature