Rich Draves wrote:
> The mobility draft has at various points or times talked about swapping
> the
> source address in the IPv6 header and the address in the home address
> option. This is OK conceptually and I understand that some implementations
> actually work that way. But at least in our implementation, we treat
> packet
> data as read-only. I would not like to see specifications that assumed the
> swapping really does happen.
Fancis Dupont wrote:
> => I understand your concern but:
> - routing header processing already specifies a real swap (then if home
> agent option has the symmetrical effect it may use the same way to do it)
> - current specifications don't say if I (real swap) or you (don't swap
> data but only swap pointers) do the right thing and only one is correct
> (because AH won't give the same message digest in both cases).
> This must be fixed!
In terms of the API (and IP_RECVDSTOPTS in particular) we can presumably
satisfy both of you by saying that the content of the IP_RECVDSTOPTS ancillary
data item, when it contains is a home address option, has the address
field in the home address option set to the care of address i.e. the
source address of the packet when it arrived.
This doesn't require modifying the packet when it arrived (Rich's concern
as I understand it) - instead it requires some care when constructing
the IP_RECVDSTOPTS ancillary data item.
Of course, implementations that process the home address option by
doing a swap have less work to perform when constructing the ancillary data
item.
Should we put this in rfc2292bis?
Erik
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
