At Fri, 29 Jun 2007 12:06:21 -0400, Ralph Droms <[EMAIL PROTECTED]> wrote:
> Stig - you wrote "At least as a sysadmin/user I would find it > confusing if the prefix length I configured would not be used for on- > link determination. I think it's more bad than good to try to > separate the two. I'm happy the way it currently is on the systems > I've seen." I can understand that it should be possible to manually > configure on-link prefix information. I question whether that > configuration should be conceptually tied with address assignment > because of the design of IPv6, or do we mix address assignment with > prefix information because that's the way it was done in IPv4? I believe so. Even though the concepts of address/prefix are different for IPv4 and for IPv6 in some points, I'm pretty sure that an administrator would expect when they execute # ifconfig ed0 inet6 2001:db8:1234::abcd prefixlen 64 the system treats 2001:db8:1234::/64 as on-link, just like they would expect when they execute # ifconfig ed0 inet 192.0.2.1 netmask 0xffffff00 the system treats 192.0.2.0/24 as on-link. > Looking back to the definition of DHCPv6 - we received a lot of input > that DHCPv6 should *not* include information about default routers > and on-link prefixes, because that information comes from RAs. That > argument made sense to me at the time; makes sense to me in the case > of manual address assignment, too... IMO the key difference is "automatic" (whether stateless or DHCPv6) vs "manual". In the case of RA (for on-link prefix) + DHCPv6, everything will be set up automatically, so I'd agree that it makes more sense not to have the redundant information in DHCPv6. In the case of manual configuration of addresses, the administrator should be able to control everything, so I think it's reasonable for the administrator to allow to specify both the address and on-link prefix information even when the node is also receiving RAs. BTW: I personally think this situation (manual address assignment with RA for on-link info) is very atypical and is not really appropriate to compare with the common case for DHCPv6(address) + RA(on-link) in the first place. But if the administrator really wants to do it, they can achieve the result by specifying 128 as the prefix length: # ifconfig ed0 inet6 2001:db8:1234::abcd prefixlen 128 at least on BSDs. Finally, I'd like to note that prefix information option with L=0 does not mean the correspondent prefix is off-link. [...] Note, however, that a Prefix Information option with the on-link flag set to zero conveys no information concerning on-link determination and MUST NOT be interpreted to mean that addresses covered by the prefix are off-link. The only way to cancel (Section 6.3.4 of RFC2461) So, by executing # ifconfig ed0 inet6 2001:db8:1234::abcd prefixlen 64 (while receiving RAs with the prefix information for 2001:db8:1234::/0 and L=0) the administrator is not "overriding" the information advertised by the RA; they are simply "specifying" the information which has not been provided. JINMEI, Tatuya Communication Platform Lab. Corporate R&D Center, Toshiba Corp. [EMAIL PROTECTED] -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------