mve> gif and gre cache the route to the destination when configured

    dy> somebody suggested that we use sequence numbers for the
    dy> purpose.

yeah, working, robust tunnels would be very useful.  I have so many
plans on hold waiting for 2.0_BETA to stop crashing all the time.  It
is ironic that the most vulnerable piece of software in my whole
network is ipfilter.

FWIW, I could live without that weird obsession in gre of sending
packets between the tunnel endpoints through the tunnel.  I don't
understand why that is there, and even gre(4) points out if the other
end is Cisco, they want nothing to do with it---traffic between
outer-header addresses always goes outside the tunnel.  I don't know
if this obsession is encouraging the tunnel implementation to not
support dynamic routing, or if it is really all about routing loops in
the kernel and you get both the obsession and no routing loops at
once.  but uh, yeah, if someone feels like fixing it, please post an
announcement so I can start using them right away.

--
Le fascisme est la dictature ouverte de la bourgeoisie.
		-- Georg Dimitrov