On Sun, May 25, 2014 at 5:22 PM, Simon Horman <ho...@verge.net.au> wrote: > diff --git a/datapath/flow_netlink.c b/datapath/flow_netlink.c > index 803a94c..8ce596c 100644 > --- a/datapath/flow_netlink.c > +++ b/datapath/flow_netlink.c > + case OVS_ACTION_ATTR_POP_MPLS: > + if (!eth_p_mpls(eth_type)) > + return -EINVAL;
Should this also take into account the VLAN tag? It's really part of the EtherType although it has been stripped out here. Actually, maybe it's better to not track the vlan_tci separately at all during validation but just fold it into the EtherType. The practical implication of this is that you wouldn't be able to pop out from underneath a VLAN. This may be a good thing if we are trying to avoid tag order issues - after all, you can't push underneath a VLAN either. I'm not sure what effects this has on the need to track mac_len, if any. Otherwise, I'm happy with this. I think that we need to conclude the discussion on the other patch and update this appropriately first. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
