On Fri, 2005-07-01 at 17:41, Sean Hefty wrote: > Looking at the kernel RMPP implementation, I think that we want to ensure > that an application doesn't send a non-DATA RMPP MAD. The RMPP code would > interpret such a MAD as being internally generated when it wasn't.
Right; sounds like a good safety check to me. There should be no need for an application to generate anything other than DATA. ACKs, etc. are all hidden and handled rather nicely by your implementation. If there were some reason to terminate an in progress RMPP send, would that be handled by a cancel operation ? -- Hal _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
