>    I have been able to get /bin/csh loaded (the ignominy!) from init's
>    singleuser prompt, and can even to type a few characters to csh, but
>    the kernel vm_faults seconds afterwards with this:
>    ----------------------
>    vm_fault(f8190474,0,1,0) -> 5
>    kernel: pagefault trap, code=0
>    ----------------------
>    which sends me to ddb. When I do a `ps' from ddb, I get:
> 
> If you're in DDB, why don't you type `t' and send me a stack trace?
> 
> BTW, did you do `make clean' before building a new kernel?

Here it is:

_db_read_bytes(a08,1,f7bffc34,a08,0) at _db_read_bytes+0x10
_db_get_value(a08,1,0,a08,a08) at _db_get_value+0x17
_db_disasm(a08,0,f8102f1e,a08,2) at _db_disasm+0x16
_db_print_loc_and_inst(a08,0,f7bffd7c,0,f7bffcf0) at _db_print_loc_and_inst+0x21
_db_trap(6,0,1,6,f8197a28) at _db_trap+0x08
_kdb_trap(6,0,f7bffd3c) at _kdb_trap+0x94
_trap() at _trap+0x174
--- trap (number 6) ---
vm_fault(f8190474,0,1,0) -> 5
kernel: page fault trap, code=0
Faulted in DDB; continuing...

And yes, I have done a make clean. I'm at a loss - if I don't type anything
at the alternate-shell prompt in init, there is no hang. But within a
couple of seconds of loading /bin/csh, I'll get a vm_fault (with the
above).

Thanks for your help.

-- 
Brian de Alwis - University of Waterloo: bsdealwis@undergrad.math.uwaterloo.ca
 "Democracy means simply the bludgeoning of the people | NetBSD - Try it!
  by the people for the people." -- Oscar Wilde        |  You'll like it!