[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: yabootconfig fails on B&W G3



Kent West wrote:
Chris Tillman wrote:

On Wed, Jul 10, 2002 at 04:02:10PM -0500, Kent West wrote:


Then ybin complained quite a bit:

/proc filesystem is not mounted; nvram will not be updated
unable to find OpenFirmware path for boot=/dev/hde6
and similar message for other partitions.

Do I need to crack the case and check out jumpers and cables on the hard drives or something, like I'd do on the i386 architecture?



Dang. No, you've got bugs in your advice.

I was trying to tell you ow to do it the 'right' way, but since proc
isn't mounted in /target, it doesn't work. Here's how I always do it, which is kind of a cheap workaround but works.
cp /target/etc/yaboot.conf /etc/yaboot.conf
ybin -v


The cp was successful, but the "ybin -v" produced the same errors.

Kent


So I started the installation over, and this time, during partitioning, wiped the Apple Partition which gave me an empty drive, then created a bootstrap partition (with the "b" command), created a 256MB swap partition, and used the rest of the drive for a single / partition.

When I got to the step to "Make System Bootable", I had the same failures: yabootconfig failed. When I tried to run "yabootconfig" manually on the second vt, it complained that it couldn't determine the root partition.

I'm really thinking there's something screwy about this particular G3's hardware setup. I opened the case and took a look (didn't change anything). The Maxtor hard drive is the only device plugged into the IDE port on the mobo, and it's set to Master. The CDROM and zip drives are plugged into the Ultra ATA port on the mobo.

I've been trying to figure out how the aliases work, and it's not making sense to me.

According to "printenv", the "bootdevice" is set to "hd:,//:txbi", and according to "devalias", "hd" and "ultra0" both are set to "/pci/@d/pci-ata@1/ata-4@0/disk@0". BIG ASSUMPTION: I assumed that the command "mac-boot" is equivalent to "boot hd:,//:txbi", so I ran that command, and got the (by now, familiar) error "can't OPEN: hd:,//:txbi". I tried it also with "ultra0" instead of "hd", and with the long name (appropriately adapted).

And since I've been able to boot the installation by referring to "ide0" instead of "hd", it seems to me that "hd" doesn't refer to the hard disk at all. "devalias" refers to "cd" and "zip" and "ide0" as being on the same controller ("mac-io/ata-3@20000", with the cd adding "disk@0" and the zip adding "disk@1"). So "cd" is essentially the same as "ide0/disk@0", but that can't be because when I boot the Debian installation using a reference to "ide0", the installation starts from the hard disk instead of the CDROM.

So it seems to me that the aliases in OpenFirmware are messed up. I tried a "set-defaults" command (or something similar; I found it on the web yesterday but don't remember the exact command name), followed by a "reset-all", which rebooted the Mac, but nothing changed.

Related, sort of: Does setting the "Startup Disk" in MacOS simply change the "boot-device" setting in OpenFirmware? In other words, could I simulate changing the Startup Disk in MacOS with a command similar to "setenv boot-device=[whatever makes CDROM vs Hard Drive, etc)?

Kent



--
To UNSUBSCRIBE, email to debian-powerpc-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org



Reply to: