Not too sure how to word this but - here goes:
Given the existing (and possible future) IEEE function processing of the u & g bits in a v6 address, am I correct in thinking that IF (with long masks for instance):-

The 9th octet's last nibble is even-ACE (71st = 0)
AND
The 10th octet's first nibble is 0 through to 7 (72nd = 0)

This won't interfere with any existing (or future)  multcast function and implies override of manufacturer IID?

(odd-BDF and 8-F being the reverse and not good)
Regards