Thorsten Frueauf writes:
> 1) Is it possible that recent changes to fsck_ffs (lite 2 integration) messed
>    something up? I had those lock ups before, but fsck was always able
>    to clean it up properly.

Yes, it appears that this was the case.
I accidentally reintroduced some code (that was still in Lite 2) that 
Mycroft removed when he optimised the propagate() function.

That code has now been fixed. The correct version of fsck_ffs/pass2.c is:
	$NetBSD: pass2.c,v 1.23 1997/09/21 03:51:34 lukem Exp $

If you have a recent copy of fsck_ffs with versions 1.20 or 1.21 of
pass2.c, please update your sources immediately!