On 5 April 2017 at 02:00, Paul Blakey <[email protected]> wrote: > Actually we've based the code on odp_flow_key_from_flow__, we are doing the > same thing, and looking at it again, it isn't foolproof as well and in the > same way. We will do the said check and printing a generic msg in case we > did (maybe with some DBG info, like offset) something.
OK. I think the difference in odp_flow_key_from_flow__() is that that function is always updated with the latest features and so on, given that the primary push for new features and so on is in the OVS netlink API. For TC flower, it's playing catchup so the lack of support for, eg, connection tracking needs to be handled carefully. I look forward to the next version. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
