panic: bad dir: mangled entry, fsck: missing dot/dotdot

[Edgar Fuß <ef%math.uni-bonn.de@localhost>]

On a FFSv2/WAPBL file system successfully fsck -f'd less than two hours 
before, I got a "bad dir: mangled entry" panic.
I fsck'd again, finding missing dot/dotdot entries and a bunch of unconnected
files.
Just within minutes after going live again, I experienced the same panic, 
and fsck found missing dot/dotdot in exactly the same directory.
I rmdir'd the directory in question. Now I'm getting "Bad file descriptor" 
in a different place during security's setuid/device find (no panic so far).

I'm running a 23-hour RAIDframe reconstruction, so I'm not going to touch 
anything the next 39 minutes, but after that, what can I do?

I had not yet re-started the reconstruction when the second panic occured.

Can there be some weird file system inconsistency fsck doesn't spot?
If the cause was a hardware or software error, I wouldn't expect the panic 
to happen again in exactly the same dir that was fixed before.
I'm thinking of something like multiple (non-dot/dotdot) dir entries pointng 
to the same inode, dot/dotdot pointing to a non-directory entry or the like.

Before the first (error-free) fsck mentioned above, I had a RAIDframe 
component failure (hardware error/internal target failure), tstile-stuck 
nfsd processes, hang after "done running shutdown hooks" and failing cache 
flushes upon ddb reboot -- just in case that gives a clue.