John Valdes wrote:

>Perhaps creating the NetBSD filesystem using newfs w/o the "-O" flag
>(so it creates a 4.4BSD ffs, correct?) prevents MacOS X from
>recognizing it?  What newfs options does sysinst use when it creates
>the root filesystem?

The last time I checked sysinst didn't use any options on newfs.

>  (Perhaps too this is why the kernel panics when
>it tries to mount it?

I guess this is possible, but I doubt it.

>  I know the NetBSD/m68k root has to be 4.3BSD
>ffs (ie, "newfs -O"), but this is because the MacOS Booter can't read
>4.4BSD ffs and not a kernel limitation, correct?)

This isn't exactly the case.  The m68k partition can be a 4.4BSD one - it 
is on my system.  It's ture the Booter can't completely handle this type 
of filesystem, but I've been told that the new and old style are similar 
enough in the beginning that the Booter usually works if the file's inode 
number isn't to large.  Since sysinst creates both a root and usr 
partition (at least it used to and still does on the mac68k port), this 
usually gets around the Booter problems.

>  Is OS X's UFS
>identical to the 4.4BSD ffs, or is it the 4.3BSD ffs or its own
>variant?

Don't know, but as Bill suggested maybe it's time for a NetBSD type of 
partition.  We might also be able to just use something other that slice 
0, but I'm not sure what OS X will do with non-zero slices.  I think 
NetBSD/mac68k (and probably NetBSD/macppc) recognizes slice 1 as well as 
zero.  Allen Briggs mentioned something about supporting both when they 
were originally using AU/X to get to a point where MacBSD would boot and 
run.  MacBSD was later merged into NetBSD 0.9 to become the first 
NetBSD/mac68k port.

-bob