> >	as for if_loop.c:
> >	your logic makes sense, but we may want to use tags to packet loop
> >	prevention within the kernel. [...]
> I'm not convinced, but willing to listen.  The point is, the semantics
> need to be _documented_.  Particularly now the KAME code is no longer
> the only, or fastest, game in town.

	m_tag is not just for ipsec.  yes, we need more documentation.
	i will check if openbsd has some docs.  Angelos wrote a paper on m_tag
	for this coming Freenix IIRC.
	to be perfectly clear, loop prevention is not related to ipsec.

itojun