Sorry forget to reply all, On Wed, Aug 14, 2013 at 1:57 PM, Alex Wang <al...@nicira.com> wrote:
> > > On Wed, Aug 14, 2013 at 1:49 PM, Ben Pfaff <b...@nicira.com> wrote: > >> I don't understand bfd well enough to understand this comment. Do you >> mean that this patch fixes such a bug or that it introduces such a >> bug? >> > > It will introduce a bug in my bfd patch. Consider if bfd is in decay, and > we set bfd:min_rx and bdf:decay_min_rx=0 (disable bfd decay) together, the > change to bfd:min_rx will trigger the poll and the bfd:decay_min_rx > configuration will not be included in the poll. So, the decay is not > stopped. > > Also, I didn't see in the BFD protocol that each poll sequence can only > include one parameter change. > To be more clear, in current implementation, once bfd_poll() is called. Before the poll sequence finishes, the following calls to bfd_poll() will have no effects. This is why when we configure multiple parameters together, we will see many poll sequence logs.
_______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
