booting 3400 help (Re: booting CD ?)
charlie allom
im installing onto a 3400c. it has OF version 2.0.1, so it can't use
ELF. it also has no floppy drive bay so i've created a bootable CD ..
i have the netbsd kernel, ofwboot.elf and ofwboot.xcf (last one from
/pub/NetBSD/misc/wrstuden cause i couldn't get it from boot.fs - no
other unix boxes with floppy's)
im having major trouble with OF, can anyone else running a 3400 help
me out and tell me how to boot from a CDROM? (which boot-device alias
i should be using..)
im getting error's like 'cannot LOAD' and 'formatstate not valid'
the alias 'cd' and 'enet' arent there when i 'devalias' so hmmmm =\
muchly appreciated. im not a NetBSD -newbie- exactly. ive gotten my
SE/30 running quite nicely it's just this OF im lost in =)
cheers. charlie.
--
cha...@rubberduck.com
Melbourne, Australia
http://rubberduck.com/ - PGP available
Henry B. Hotz
>im having major trouble with OF, can anyone else running a 3400 help
>me out and tell me how to boot from a CDROM? (which boot-device
>alias i should be using..)
>
>im getting error's like 'cannot LOAD' and 'formatstate not valid'
>
>the alias 'cd' and 'enet' arent there when i 'devalias' so hmmmm =\
'formatstate not valid' sounds like you have found something partly
valid, but are trying ofwboot.elf instead of ofwboot.xcf.
Presuming there is a devalias for scsi and your CD-ROM drive is ID 3,
then "boot scsi/sd@3:2,ofwboot.xcf -a" should get you started. I'm
not sure how your CD is formatted so you may want to try partitions
0-3, but I suggest you power cycle the machine after a failed attempt
since I don't think OF always cleans up after an error properly.
If there is no obvious devalias for the scsi bus then you will need
to search the whole device tree for the exact path to use. Maybe
something like /bandit/gc/mesh/sd@3...
Signature failed Preliminary Design Review.
Feasibility of a new signature is currently being evaluated.
h.b....@jpl.nasa.gov, or hbh...@oxy.edu
charlie allom
i swear i will write a nice little HOWTO once this is over
i have found that the alias cd (which isn't there) == atapi
the cdrom spins up (but doesn't flash it's light) when i 'boot atapi'
i get 'unrecognized Client Program formatstate not valid'
HenryB thought this might have to do with the fact that im trying to
load an ELF rather than an XCOFF. but i'm not =\
last try is to have a go with an HFS and an MSDOS partitioned CDROM.
how this will help i dont know but i have to explore all possibilites.
has anyone found some weird happenings in OF 2.0.1 that might help me?
Henry B. Hotz
>last try is to have a go with an HFS and an MSDOS partitioned CDROM.
>how this will help i dont know but i have to explore all
>possibilites.
Certainly worth a try. FWIW Bob Nestor reports that a number of CD
images have slightly messed up Apple Partition Maps which probably
accounts for a lot of people's failure to boot from CD.
I know I was able to boot ofwboot.xcf directly from a CD a long time
ago. I believe it was an ISO CD and it seemed to matter that I put
the name in ALL CAPS in the boot command to get it to work. At any
rate I think that's why I got it to work on a 7500/604 but not on an
8500.
Bob Nestor
>At 12:19 AM +1000 7/21/00, charlie allom wrote:
>>last try is to have a go with an HFS and an MSDOS partitioned CDROM.
>>how this will help i dont know but i have to explore all
>>possibilites.
>
>Certainly worth a try. FWIW Bob Nestor reports that a number of CD
>images have slightly messed up Apple Partition Maps which probably
>accounts for a lot of people's failure to boot from CD.
As Henry said, I've been looking at a lot of Partition Maps on various
disks and CDs, mainly on the mac68k port as part of my effort to get
sysinst running there. In the process I've generated a modified version
of the MkLinux pdisk utility that formats and dumps the Partition Map. I
built a version for both mac68k and macppc systems, and I've run it on
both types of systems. I also did a similar utility that runs under
NetBSD and generates the same format dump of the Partition Map. I've run
them on both NetBSD/mac68k and NetBSD/macppc. It could be made to run on
other ports, but will require some minor code changes to do so (location
of disklabel.h and byte order details). If anyone is interested in
getting copies (source and/or binary) of these they can find them at:
ftp://murphy.dyndns.org/pub/map
Here is the simple README file:
The MacOS utility is called "pdisk" and is based on the MkLinux 0.7a3
code which has some sort of bug in accessing the Master IDE disk.
Sometimes it finds it and sometimes it doesn't. The 0.7a4 code fixes
this but I can't find the sources for it. If you know where they are I'd
be happy to make the changes to it. There is a 68k executable and a PPC
one that I pre-built in the archive (pdisk.sea.hqx) in case you don't
have CodeWarrior. The
option I added is the "V" (upper case).
The NetBSD utility is called "pdump". You can build it and run it on
either a NetBSD/mac68k or NetBSD/macppc system. It won't build on other
NetBSD systems because of the location of disklabel.h and probably won't
run correctly on little-endian systems. Just invoke it with the name of
the raw disk you want to dump or the name of an ISO Image file, i.e.
"./pdump /dev/sd0c" or "./pdump macppc.iso". You must run it with root
privs if you want to use it on a raw disk. Sorry it's not very well
documented but I hadn't intended on releasing it this soon. ;-)
The Partition Map dump in both cases is not 100% complete yet. It
doesn't incorporate some of the new items that Apple introduced with the
newer Macs such as more flags and the meaning of the Driver types. I'll
be adding that in the next few days, but the raw data is still dumped
it's just not identified as to what it means yet. If you have questions
about it I think I can answer them now that I've read all the new
documentation.
One last word of warning: Most disk formatters, CDs, utilities and such
that muck with the Partition Map are pretty free about how they interpret
the Apple Specs. This includes Apple. If you don't believe me try
dumping the Disk Partition on an Apple manufactured CD! Just because
these programs report errors in the format doesn't mean the disk or CD
isn't usable under MacOS. In writing these programs I wanted to be as
accurate as possible to insure that NetBSD wouldn't have any problems
with the disks/CDs. In particular the order disk partitions are seen is
important to NetBSD/mac68k because of the way it searches for root and
usr partitions. Because of that I always try to present the Map in
disk-order, or the way the partitions are stored on the disk physically.
The sysinst/mac68k code goes _way_ out of it's way to insure this order
is maintained to make damn sure NetBSD/mac68k finds the partitions in the
right order.
-bob
7/18/2000