Can we come to a conclusion about the path forward for this patch-set? I still maintain that the design is reasonable within the limits of what the POSIX/Linux makes available. However, if the design is considered to be wrong, I would appreciate alternative suggestions.