Parallel routing code patch #3

[Matthew Dillon]

    Here's the third patch:

    fetch http://apollo.backplane.com/DFlyMisc/routing03.patch

    This should fix the reported panic that occurs when an attempt is made
    to delete a cloned route.  The panic was occuring because cloned routes
    are not replicated.  We now allow cloned routes to not be replicated and
    forward RTM_DELETE requests for cloned routes even if a particular cpu
    does not have the route.

    I'm still seeing a discrepancy in the ARP table for IPV6 which I am
    not sure about.

						-Matt