Fill receive-side message handling gaps in Ken support

[FUDCo]

Supporting the Ken protocol (see, for example, HP Labs tech report 2010-155 or Terence Kelly et al's Usenix paper) would ensure that our handling of distributed persistent data is mutually consistent across communicating peers. The remaining gaps in being fully Ken compliant are on the message receive side of our inter-kernel network communications support. A more detailed analysis of what needs to be done can be found here, but the remaining work is relatively modest compared to the other stuff we've already done in this regard. Finishing this should complete the work needed to close issue #621.