On Wed, Apr 04, 2012 at 02:43:04PM +0200, Rimi Philippe wrote:
> Thanks Claudio.
> The way I see it is that RD are only local, they identify the VRF (or
> rdomain) locally on the router, then the RT import / export handles
> the way the routes are distributed. This permits the hub & spoke
> approach for example.
> 
> If you need help on the testing side feel free to send me the code,
> I'll give it a try.
> 

Give this a try.

-- 
:wq Claudio

Index: bgpd.conf.5
===================================================================
RCS file: /cvs/src/usr.sbin/bgpd/bgpd.conf.5,v
retrieving revision 1.116
diff -u -p -r1.116 bgpd.conf.5
--- bgpd.conf.5 17 Sep 2011 16:29:44 -0000      1.116
+++ bgpd.conf.5 4 Apr 2012 18:46:54 -0000
@@ -494,13 +494,13 @@ for further information about the argume
 .Pp
 .It Ic rd Ar as-number Ns Li : Ns Ar local
 .It Ic rd Ar IP Ns Li : Ns Ar local
-The Route Distinguishers uniquely identifies a set of VPN prefixes.
-Only prefixes matching the
+The sole purpose of the Route Distinguisher
 .Ic rd
-will be imported into the routing domain.
-The purpose of the
+is to ensure that possible common prefixes are destinct between VPNs.
+The
 .Ic rd
-is solely to allow one to create distinct routes to a common address prefix.
+is neither used to identify the origin of the prefix nor to control into
+which VPNs the prefix is distributed to.
 The
 .Ar as-number
 or
Index: rde.c
===================================================================
RCS file: /cvs/src/usr.sbin/bgpd/rde.c,v
retrieving revision 1.312
diff -u -p -r1.312 rde.c
--- rde.c       27 Mar 2012 18:22:07 -0000      1.312
+++ rde.c       4 Apr 2012 15:40:41 -0000
@@ -2414,8 +2414,6 @@ rde_send_kroute(struct prefix *new, stru
                        break;
 
                SIMPLEQ_FOREACH(rd, rdomains_l, entry) {
-                       if (addr.vpn4.rd != rd->rd)
-                               continue;
                        if (!rde_rdomain_import(p->aspath, rd))
                                continue;
                        /* must send exit_nexthop so that correct MPLS tunnel

Reply via email to