On Fri, Sep 2, 2011 at 16:46, Jesse Gross <je...@nicira.com> wrote: > Even before when OVS was less picky about the input port, it didn't > implement the behavior that you're describing in terms self forwarding > checks on virtual ports. For most of the virtual ports, it's fairly > non-nonsensical and spec actually specifies that you should use > OFPP_NONE if you don't have a real port to give. >
Sure, I'm just arguing that OFPP_CONTROLLER *is* a real port.. :-) (Ok, not in the sense that it's a physical port - well, you could even make that argument for OOB and certain in-band fixed-port systems (indigo) - but it's a well defined virtual port). -- Nick
_______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss