Hi guys. 

I've bitten the bullet and decided to install the current (1.1 Beta)
version of RiscBSD. I installed an early Alpha version around the
end of last year (for the hell of it more than anything :) so I've
a vague idea what I'm doing, but don't count on it. Because I wanted to
increase the partition size (and because my alpha setup was a real mess)
I've installed from scratch rather than upgrading. 

Anyway, everything (just about) went sweetly all the way through Mark's
installation documents, until I came to do a multi-boot after using RCM. 
Here's what I've done. Can anyone spot a stupid mistake?

* Partitioned my master CFS420A so as to have 0-31 cyls for RISC OS,
  with head parked at cyl 825. Slapped the RISCOS bootstrap stuff back on.
* Ran bb_riscbsd, setting BSD start at cylinder 32 with ADFS_DiscOp.
* Booted from floppy/ramdisc non-natively with kernel 3598 (one step at
  a time - I'll upgrade the kernel later)
* Set disc size to 826 cyls, 16 heads, 63 secs
* BSD partition is 792 cyls starting at offset 32 (Last sector is head park)
* Root and /var = 60 cyls
  Swap          = 60 cyls
  usr           = 672 cyls mounted on /usr
* All OK at this stage. Lots of disklabelling and base binary installations
  going on.

* Created /usr/distrib and made it world-writeable
* halt, and return to RISC OS where I drag the sets into /usr/distrib
  using UnixFS_5 (I have the old filecore)
* Reboot BSD, single user, with root as /dev/wd0a this time
* Run /usr/local/sbin/inst and validate & install 
    base, config, etc, local, man, misc, and shells sets.
* Run /usr/local/sbin/rcm and give myself a user in the wheel group.
  (Plus other configurations like timezone, hostname etc.)
* chmod 1777 /tmp
* explicitly put my username in the wheel group
* Reboot

* Now, the trouble starts... booting as a single user and and typing
  
    mount -t ffs /dev/wd0a /                (This is OK. Root is now r/w)
    mount -t ffs /dev/wd0e /usr
  
  gives the error:
  
    ffs: /dev/wd0e on /usr : Operation not supported by device
    
  The disc passed fsck OK.
  
* Booting into multi-user mode produces the same error so unsurprising
  none of the daemons can be found, including getty, which is where it
  all grinds to a halt. Sending a kernel panic seems to be the only way
  to stop the search for getty, but this just causes a kernel deadlock,
  requiring a reboot.

The weird thing is that I can access the same wd0e if I boot off the boot
disc and mount it without any trouble

Anyone come across this problem before? Should I start playing around
with disklabels and newfs, or should I call it a day and just slap
/usr in the same partition as / and 'have done'? Without any access the
/usr directories I'm pretty limited to what I can do ;-)

This is on an original RISC OS 3.50 ARM 610 24MB+2 Risc PC with:

o The standard 420MB Connor drive on which I'm putting BSD.
o Cumana SCSI2
o Partitioned internal gig SCSI for my RISCOS and DOS stuff 
o SCSI EZ135 (ext)
o Cumana Indigo CDROM (which RiscBSD knows nothing about, obviously)
o 16 bit Minnie sound card. 486SX/33 thingumy-bob.

TIA for any help.

Cheers
-- 
Andy Ward