> In lacp_unref(), I think we need to move list_remove(&lacp->node); > inside the lock/unlock. Otherwise we have a race here where > lacp_find() can grab an entry that is being destroyed.
Good catch. I fixed this, and also systematically added lock annotations to all of the function definitions as this would have caught bugs in other patches. Thanks for the review. I'll merge this once the thread safety annotation patch is in. Ethan X-CudaMail-Whitelist-To: dev@openvswitch.org _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev