[279] in linux-net channel archive
Re: multiple IP# for one IF ?
daemon@ATHENA.MIT.EDU (Robert Sanders)
Sat May 6 18:11:21 1995
Date: Sat, 6 May 1995 17:23:03 -0400
From: Robert Sanders <rsanders@interbev.mindspring.com>
To: linux-net@vger.rutgers.edu
In-Reply-To: <m0s7pKJ-00013lC@iiit.swan.ac.uk>
In article <m0s7pKJ-00013lC@iiit.swan.ac.uk> you write:
>NEVER NEVER give world read to kmem!
I already posted a correction to that. Russ was quoting a late-night
fumble. Apologies to those whose systems I broke wide open ;-)
>Dummy devices do all ipalias can do btw 8). I may well add ipalias as a compile
Really? Quite a few people complain about dummy devices not responding to
ARP requests. I don't think dummy devices correctly deal with the case
of having two networks on one wire; the ipalias patches go out of their
way to choose the mose comfortable source address for outgoing packets,
etc.
Ipaliases are just easier to deal with. With dummy devices, you
either have to choose how many devices you have at compile time or insmod
multiple copies. Finally, I don't like having to create an artificial
correspondance with dummyXX devices when creating aliases. It complicates
scripts, for example.
I'm not saying dummy devices aren't a solution, but there seems to
be a definite need for ipaliases. Even if not implemented the way I
did, many users will be needlessly confused by the "dummy" way of doing
things. All the www server documents will soon document "ifconfig alias",
and there's no need for Linux to be gratuitously different in yet another
area.
Regards,
-- Robert
