Some while ago, we were discussing the proper design for the networking components of the hurd, i.e., the "replacing pfinet" thing.
If anybody here is interested in practical design of such stuff, and you haven't seen it already, I'm pretty sure you want to check out the slides at http://www.lemis.com/grog/Documentation/vj/ One part of the solution is moving almost all of the processing needed for a tcp/ip "stack" to a user space library. The main argument is performance, in particular for scalability for multi/cpu machines. Regards, /Niels PS. And I'm not actively subscribed to this list, so please cc me any replies you want me to see. _______________________________________________ L4-hurd mailing list [email protected] http://lists.gnu.org/mailman/listinfo/l4-hurd
