On Thu, Sep 12, 2013 at 04:39:01PM -0700, Ethan Jackson wrote: > Well for the first suggestion, I think it simply comes down to doing a > check if the reval_seq changed during handle_miss_upcalls() like we > did before. If it changed then we skip the guarded_list_push_back(). > Of course there's a race window, but it's small and doesn't matter.
I didn't realize you were OK with a small race. It's trivial then. I'll change that. > For the second suggestion, I think we just need to change the infinite > for loop to loop 50 times and give up if there's nothing there. No problem, will do. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
