On Tue, Apr 17, 2012 at 02:11:43PM -0700, Ethan Jackson wrote: > I'm not sure that the change to dpif_assert_class() is necessary.
Oops, that's an intentional change but I didn't mean it for this patch series. Thanks, removed. > There's a redundant newline added after dpif_flow_del(). Thanks, removed. > Out of curiosity, where did EXPIRE_MAX_BATCH = 50 come from? Seems > like as good a number as any, just curious if the value has a > significant impact on performance. It's just a number significantly larger than 1 but not big enough that multiplying by EXPIRE_MAX_BATCH is a huge amount of memory. I didn't try other numbers, perhaps some tuning is warranted. > In expire_subfacets(), "struct subfacet *batch[50]" should probably be > defined in terms of EXPIRE_MAX_BATCH. Thanks, I changed this to EXPIRE_MAX_BATCH. > Everything else looks good, thanks. Thank you for the review. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
