>    It seems to have died somewhere in dead_ something or other,
>    but the message flashed by while I was just walking in.
> 
> dead_badop(), and it's already been fixed.  Update your kernel sources
> (specifically dead_vnops.c).

Ouch.  Having just merged the Feb-03 tar_file sys tree, I noted that
there were no differences in dead_vnops.c, so I went to my local
NetBSD mirror (could not reach ftp.netbsd.org), and grabbed the file
which was also dated Feb-03, and still no diffs.

All I was able to glean so far from the crash dump was that dead_badop
was called from runtime().  After that bt just said can't read memory
at 0x0.  Not much help, sorry.  

Does anyone have an FAQ about poking around crash dumps - eg, what
process was running at the time etc.

I'd really like to avoid having to restore 95.08.18 from tape.

--sjg