On Thu, Nov 19, 2015 at 01:33:18PM -0800, Jarno Rajahalme wrote: > Currently ovs-ofctl replace-flows and diff-flows commands only support > flows in table 0. Extend this to cover all possible tables. > > Signed-off-by: Jarno Rajahalme <ja...@ovn.org>
There's one oddity that may deserve consideration. It depends on how careful we want to be. OpenFlow 1.0 does not define a way to add a flow to a particular table. The switch is responsible for deciding which table is most appropriate for a given flow. For example, a switch might have one table that supports wildcards and another one that is exact-match (this is in fact specifically envisioned by OF1.0 through its insistence that exact-match flows have the highest priority). This means that when talking to an OF1.0 switch, "ovs-ofctl replace-flows" (and friends) should ignore the table number. If a flow on the switch is in table 1, but the input file says it is in table 0 (probably because it doesn't specify a table at all), ovs-ofctl should do nothing, because that's the desired state. However, for practically forever, OVS has had special extensions to allow control over the table in which a flow lives. This means that if ovs-ofctl is talking to OVS, even in OpenFlow 1.0, it should place flows where the user requested and should not ignore the table numbers. This distinction is reflected through ofputil_protocol values. If a switch supports OFPUTIL_P_OF10_STD_TID or OFPUTIL_P_OF10_NXM_TID, then ovs-ofctl can place flows arbitrarily; if it only supports OFPUTIL_P_OF10_STD (or, theoretically, only OFPUTIL_P_OF10_NXM), then it is just a plain OF1.0 switch and all of the tables should be treated alike. OF1.1+ all support placing flows where the user requests. It's probably not too hard to support this, and possibly it is worthwhile. What do you think? _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
