Joakim - The note is for the original users of Cumulus's unnumbered implementation. We would like them to continue to think about the problem in the same manner that they have been, while we changed the underlying solution.
As for why path->unnumbered I wanted to make sure that the unnumbered interface got sent like we wanted it to be. dopnald On Tue, Nov 24, 2015 at 3:37 PM, Joakim Tjernlund < [email protected]> wrote: > On Mon, 2015-11-23 at 09:03 -0500, Donald Sharp wrote: > > Joakim - > > > > It doesn't need to magically(I disagree with this term :) ) connect to > any loopback anymore. If an > > I don't get it then(note I haven't read the patch properly yet), why is > the loopback stuff included > here? > > I think you could also drop the ZEBRA_IFA_UNNUMBERED flag and just test > for a /32 mask in the few places you > need it. > > Just noted this: > > > > > --- a/ospfd/ospf_zebra.c > > > > +++ b/ospfd/ospf_zebra.c > > > > @@ -373,6 +373,27 @@ ospf_zebra_add (struct prefix_ipv4 *p, struct > ospf_route *or) > > > > /* Nexthop, ifindex, distance and metric information. */ > > > > for (ALL_LIST_ELEMENTS_RO (or->paths, node, path)) > > > > { > > > > + if (path->unnumbered || > > Why a path->unnumbered? I think there is no need to touch this function at > all. > > > > > + (path->nexthop.s_addr != INADDR_ANY && > > > > + path->ifindex != 0)) >
_______________________________________________ Quagga-dev mailing list [email protected] https://lists.quagga.net/mailman/listinfo/quagga-dev
