Well, it would seem to me then that the lacp_compose and parse functions, ethernet destination address, and protocol data structures should move into the lacp header file. Does that seem reasonable to you? If so I will go ahead and reorganize this series, resend it, and add some patches which pull the lacp stuff out of packets.
Ethan On Wed, Mar 23, 2011 at 11:41 AM, Ben Pfaff <[email protected]> wrote: > The purpose of the packets library is not carved in stone, but I've > been regarding it mainly as a grab bag for stuff that doesn't have a > clear place elsewhere in the tree. "benign" packets fall into that > category I think, but for LACP and CCM it seems less clear--we have > LACP and CFM files. Anything that is esoteric and protocol specific > would seem well suited to go into protocol-specific .c and .h files > when we already have them. Not sure why we'd banish it to some > generic place. > > On Wed, Mar 23, 2011 at 11:29:54AM -0700, Ethan Jackson wrote: >> I guess I'm not fully understanding the purpose of the packets >> library. The definition of ccm messages is in its header file. It >> would seem natural to me for composition and parsing to be there as >> well. Furthermore, this is how it's done for lacp and "benign" >> packets. It seems like a good place to put esoteric protocol specific >> logic that is only interesting in creating the packets freeing up the >> caller to deal with the logic of what actually belongs in there. >> >> On Wed, Mar 23, 2011 at 10:29 AM, Ben Pfaff <[email protected]> wrote: >> > On Tue, Mar 22, 2011 at 06:32:56PM -0700, Ethan Jackson wrote: >> >> This cleans up the CFM library a bit by pushing ccm composition >> >> details into the packets library. >> > >> > I don't yet see the benefit to moving composing (patch 4) and parsing >> > (patch 5) CCM packets from cfm into packets. ?Can you explain? >> > > _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
