Coda _has_ worked fine on BSD for a long time.  There was a change in
NetBSD vfs locking rules for lookup recently (in 1.6 but not 1.5), and
that introduced a problem in the coda code.  With my fix, I think the
NetBSD kernel code is OK - this is the only problem I've had that is
kernel-related.

I have not looked at the Darwin code; Coda worked fine (the kernel
part) on FreeBSD 3.x, so it is probably either ok or reasonably easy
to port.

Someone will have to update the kernel code for the interface changes
from the realms code.  The changes aren't hard, though, unless one
wants to have a kernel that supports both the old and the new way at
the same time.  I'm leaning towards

options	 CODA_NOREALMS

to get the old behavior.