Hi,

> OK.  The crash point in net/radix.c:rn_match() is at least plausible.
> I don't know why the stack looks hosed, though.  My initial guess is
> some kind of memory corruption.  My second is that something has
> corrupted the radix tree.
>
> Is this with -current or 2.0 or other?

2.0.

> Would you be willing to try to get a crash dump if I supply you with
> experimental code for that?  I understand that this is a production
> system and that's a lot to ask...

If it could possibly help remove a problem that keeps my production system 
less in production, sure! I have a serial console on the machine already.

Thanks,
John