> PS: I've tried one day the 2*64 definition on a Sparc (ie. with 64 bit
> alignment). It didn't work because of some details, for instance the
> route structure (pointer + sockaddr) or IPv6 over ATM (32 bit difference
> in alignment between unicast and multicast). I can't see a good reason
> to add a new source of problems...
We tried the same thing a while back (thinking it would be trivial).
Turned out that the embedded in6_addr structures in various places
(ancillary data which gets mapping into TPI option buffers, socket
options, etc) made the introduction of 64 bit alignment non-trivial.
So we punted for the time being.
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]
--------------------------------------------------------------------
