On Thu, Aug 12, 2004 at 09:26:02AM -0400, Ralph Droms wrote: > Seems to me it would be useful to allow both M and O flag on. > > While, in theory, the support for the subset of DHCPv6 indicated > by the O bit is implied by the support for all of DHCPv6 > indicated by the M bit, it seems there would be little harm > in advertising both. > > Some hosts may choose to use both the Request message exchange > for address assignment and the Information-Request message > exchange for other information. For example, an application > may need other configuration information that was not supplied > during the initial Request message exchange, requiring a > subsequent Informaton-Request message exchange for that > information. There are DHCPv4 hosts that work this way today.
Agreed, but wouldn't it always be ok for the host to do Information-Request message exchange for other information when M is on? My thinking is: M=0, O=0 stateless autoconf of addresses M=0, O=1 stateless autoconf of addresses + information-request M=1, O=0 stateful autoconf of addresses where the latter includes information-requests for other config if client wants. I believe in the latter case a client only implementing the Information-Request message exchange or with manually configured addresses, may possibly only use information-request. Stig -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------