Not a review, only one vey small comment below, Jarno
On Aug 15, 2014, at 12:43 PM, Daniele Di Proietto <ddiproie...@vmware.com> wrote: > > + * A miniflow with a given hash can be in one of EM_FLOW_HASH_SEGS different > + * entries. Given its hash (h), the miniflow can be in the entries whose > index > + * is: > + * > + * h & EM_FLOW_HASH_MASK > + * h >> EM_FLOW_HASH_SHIFT & EM_FLOW_HASH_MASK > + * h >> 2 * EM_FLOW_HASH_SHIFT & EM_FLOW_HASH_MASK > + * h >> 3 * EM_FLOW_HASH_SHIFT & EM_FLOW_HASH_MASK > + * … The last one has only two hash bits left, maybe it would be better to drop that? I noticed only later that EM_FLOW_HASH_SEGS was defined as 2 (and not 4 as I implied from the above... > + * > + * Thread-safety > + * ============= > + * > + * Each pmd_thread has its own private exact match cache. > + * If dp_netdev_input is not called from a pmd thread, a mutex is used. > + */ > + > +#define EM_FLOW_HASH_SHIFT 10 > +#define EM_FLOW_HASH_ENTRIES (1u << EM_FLOW_HASH_SHIFT) > +#define EM_FLOW_HASH_MASK (EM_FLOW_HASH_ENTRIES - 1) > +#define EM_FLOW_HASH_SEGS 2 > + _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
