freebsd 2.2.5 and arla-0.13

[Eric Hoffman]

  That's weird.  Doesn't the_fbuf get initialized to {NULL, -1, 0, 0} ?

right, thus the segv...making the branch be unconditional means I get
back ENODEV when trying to opendir /afs/cs.washington.edu instead of
crashing arlad, but the kernel trace looks the same (bogus fileid)

  The fileid being zero is bad and is probably a result of fcache_fhget
  failing. 

yeah. I've been looking at it further, the handle passed to fhopen is
zero...tracing it back the entire queue that create_nodes traverses
during startup seems to have zeroed handles

  Did you get any debug output from xfs while arlad was
  creating the cache files?

you mean orginally? no, just:

Oct  6 15:18:10 tesuji /kernel: Loading mod...
Oct  6 15:18:10 tesuji /kernel: syscall 210
Oct  6 15:18:10 tesuji /kernel: Using VFS slot 17

do you think it would help if I blew away my cache?