[EMAIL PROTECTED] wrote:
Yes, agreed. Going further, maybe A=0 could signal this?
But the A flag is per prefix, not per RA.
And so far we haven't assigned any semantics to a flag in the prefix
being zero; the semantics are associated with the flags being set to one.
That model seems to be useful when extending the flags in the prefixes
since it provide orthogonal flags. For instance, we added the 'R' flag
and with that we can have Prefix option which set only the R flag (as
well as the R flag in combination with one of A and L) and the addrconf
and on-link pieces don't get confused by a prefix with R=1, A=0, L=0;
those functions just ignore prefixes that have A=0 and L=0, respectively.
Erik
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
