As I read it, if you want to embed v4 addresses in v6, here's where
you do it, but in fact nobody does.
Either you don't use dual-stack Linux boxen or there's something I'm
missing.
I have a fair number of network programs that I've gone to some pains
to make v6-ready. When I build and use them on Linux - any Linux, in
my admittedly quite limited experience - I see behaviour indicating
that tcp4 connections to port N in the presence of a process listening
on tcp6 port N get turned into tcp6 connections with a mapped-v4
address for the (v4-speaking) peer. This is particularly annoying
because it means that, after binding a v6 socket listening on port N,
an attempt to bind a v4 socket listening on port N gets an
address-already-in-use error. (There's really nothing like
compatability - and that's nothing like it.)
I forget whether v4 gets mapped into v6 this particular way
(::ffff:x.y.z.w) or some other way.
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse(_at_)rodents-montreal(_dot_)org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/asrg