On Wed, Dec 25, 2013 at 03:43:52PM +0900, Simon Horman wrote: > This is my interpretation of a > missing piece of "Alternate approach to MPLS" > > Signed-off-by: Simon Horman <ho...@verge.net.au>
I think that flow_count_mpls_labels() counts the innermost LSEs in common, not the outermost, so I changed its comment to reflect that. I think that commit_mpls_action() needs to pop off any existing VLAN header before it pushes on a new MPLS label (otherwise the MPLS label would get pushed in front of the VLAN tag), so I made that change too. (The VLAN tag will get pushed back on when we commit VLANs.) With those changes, I folded this in. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
