Hello,

First I must say I'm impressed how far this project has come from its days at
Virginia Tech!  My roommate was CS at VT with Chris and Allen and Brad, and I
was always hearing about one cool hack or another.  The power of perseverance.

I managed to get the booter to start the current -29 kernel on my PB180, so I
thought I'd try to get a small system working on an external 105M HD.  Got
the drive partitioned, made filesystems, installed base + etc, made devices,
etc.  I kept the kernel on an HFS partition on another drive, just so I could
easily try more than one if I wanted.

First problem: (current) GENERIC-29 had problems with SCSI.  Used SBC-29, and
the drives + geometries seemed to be id'd correctly.

Second problem: couldn't successfully boot to single user mode--didn't get
keyboard access when asked for a shell.  Switched to a multi-user boot and
got past this.

Third (big) problem:  got as far as
  Automatic boot ...
  /dev/rsd1a: FREE BLK COUNT(S) WRONG...
  /dev/rsd1a: 1862 files...
and I got a bus error in sbc_pdma_out() in fsck_ffs.  But the worst part is
whatever happens on the way down corrupts the filesystem so that only a
reinstall will fix it.

I should add that the same problem occurred again after another from-scratch
partition and install.

I wonder if this is related to an earilier post by Tim Hoffman "Booting
problems" in which he laments an "error = 79" on a PB180??  After the above
problem, the filesystem is corrupt and this error is the result.

I recorded some more info from the debugger, in case anyone's interested.
I could prol'y also repeat the install.

Any comments.

shawn.