Subject: Re: Reserved port range patches
To: Jun-ichiro itojun Hagino <firstname.lastname@example.org>
From: Robert Elz <kre@munnari.OZ.AU>
Date: 08/27/2000 10:21:35
Date: Fri, 25 Aug 2000 11:10:00 +0900
From: Jun-ichiro itojun Hagino <email@example.com>
| i'll look at IPv6 counterpart.
This really ought to be quite difficult. Neither IPv4 nor IPv6
know anything about ports, so there should be no v4 or v6 specific
code that has anything to do with this at all ... this stuff is
really all TCP/UDP code, that (in the *BSD implementation) just
happens to have been dumped in the ip layer code (and ioctl/sysctl)
to avoid associating it with either TCP or UDP and hence confusing
the other one of them...
Whether the actual implementation is sharing that code or not I am
not sure (netinet6 duplicates quite a lot that one day I expect will
get merged back again, once it is truly stable ... which partly means
once the IETF stops fiddling around the edges).
Of course, it would probably be better to have separate ranges of these
vars for TCP and UDP, there's no particular reason why they should be
compelled to share the same reserved/low/high/... ranges of port numbers.