Is there still a known locking against self bug in the FFS
code?  None of the open PRs appeared relevant/similar to my
panic.

  My machine with kernels from two days ago repeatedly
panics at the same point while running sup, and has done so
since August, and with at least a half-dozen kernels since
then.  (I have to compile the kernels on a remote machine, then
scp them over the modem to the machine that can't sup.)

  Also, I created a debugging kernel, but when I enter
gdb on the core dump using netbsd.gdb, the `where' command
doesn't work -- it only says 0x0 (??).  Is there some tricky
thing I'm missing?  The steps (from memory) that I did:
  1.  blow away old kernel compile directory.
  2.  Enable -g option in config file, and reconfig.
  3.  Build kernel
  4.  scp new netbsd and netbsd.gdb to flaky machine.
  5.  Reboot using new kernel.  Do a sup to force the panic.
  6.  Reboot, and have savecore save the core file.
  7.  Invoke "gdb netbsd.gdb" (which is the sym version of netbsd.4)
  8.  Type "target kcore netbsd.4.core".  It prints 
"panic: lockmgr: locking against myself
#0  0x0 in ?? ()

  I wrote down an abbreviated version of the trace a few weeks
ago, if this helps anyone:
_lockmgr
_ufs_lock
_vn_lock
_lookup
_namei
_sys___lstat13
_syscall

  Let me know if I can help further.  If the full backtrace would
help, I can do that, or provide the core file and netbsd.gdb
file, etc.  The machine in question is i386, running userland
from Jul 12, 1998.  Thanks in advance!

  Brian
-- 
"Old programmers never die.  They just branch to a new address."
						-Anonymous