On Wed, 2011-03-23 at 11:56 -0400, Chuck Lever wrote: > On Mar 23, 2011, at 11:26 AM, Andy Adamson wrote: > > > Are you reviving the "RPC Upcall" that I implemented 10 years ago? > > This patchset doesn't specifically revive that idea, as its narrow purpose is to replace kernel rpcbind registration over network loopback. > > But the work could be a pre-requisite for a generic upcall mechanism over RPC. An AF_LOCAL transport can be used as a generic upcall mechanism if user space services set up an AF_LOCAL listener. > > I'm guessing ten years ago there wasn't built-in standardized support for AF_LOCAL RPC transports in the glibc RPC implementation, which may have impeded the success of such a proposal back then. libtirpc makes AF_LOCAL a little more practical, as the incremental cost of supporting AF_LOCAL for any user space service is now minimal. No. The idea was rejected simply because there is no need for all that XDR overhead when we're talking to a process on the same machine. Trond -- Trond Myklebust Linux NFS client maintainer NetApp Trond.Myklebust@xxxxxxxxxx www.netapp.com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
