On Tue, Mar 20, 2012 at 04:05:20PM -0700, Ethan Jackson wrote:
> It seems odd to me that we call it 'may_add_flows' in facet_account,
> and 'do_learn_action' in action_xlate_ctx. I'd be inclined to call
> both of them the same thing. It strikes me, that an appropriate name
> for the flag may be "may_flow_mod" as that's the behavior we actually
> care about. If we add actions in the future which flow_mod, this flag
> would still be applicable.
Fair enough. I changed them both to "may_flow_mod". I also changed
"do_learn_macs" to "may_learn_macs".
> Am I correct that this patch prevents leftover packets in the datapath
> from updating the timeouts in learned actions in some cases? e.g. if
> you change the actions of a rule, packets which applied to the old
> rule may not be properly accounted? This seems like a fine trade-off,
> just wanted to make sure I understand it.
Yes, that's correct. I added this to the commit message:
This commit has a side effect that leftover unaccounted packets no longer
update the timeouts in MAC learning actions in some cases, when the facets
that cause updates are deleted. At most one second of updates should be
lost.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
