Re: Question on the TCP aliasing implementation.

From: Carlos Morgado (chbm@chbm.nu)
Date: Wed May 17 2000 - 15:39:15 EST


On Wed, May 17, 2000 at 11:41:38PM -0400, Daniel Shane wrote:
> Hello,
>
> Maybe this has been addressed before but I was wondering if it was
> possible to create a TCP alias without using the ethX:0..1..2..3
> interface. The problem is that I need these interfaces to behave like a
> real eth0. For example, it is not possible to execute a dhcpd server
> with eth0:0 as an argument, it doesnt work.
>

The correct semantics would be to specify an IP address not an interface on
the dhcpd. eth0:0 is 2.0 semantics. (fix dhcpd :))

> I know there is a dummy interface but if I recall all it does is eat
> packets to simulate a link. Is there a way to redirect a dummy
> interfafce towards a particular eth.. interface?

you can kludge that with policy based routing (routing per source not per
destiny), but then i guess you wouldn't need it cause you'd just number
dummy0 with the ip you want.

> There is also the vmnet modules from vmware that simulates a real
> network card, but I dont think that these drivers are GPL'ed.
>

they're ugly.

> Has enyone worked on redesigning the aliasing code so that it behaves
> more like a real network interface?
>

the alias code isn't supposed to work as a real interface. and imho programs
that expect to bind to a device not an ip are broken.

-- 
Carlos Morgado - chbm(at)chbm(dot)nu - http://chbm.nu/ -- gpgkey: 0x1FC57F0A 
http://wwwkeys.pgp.net/ FP:0A27 35D3 C448 3641 0573 6876 2A37 4BB2 1FC5 7F0A
Software is like sex; it's better when it's free. - Linus Torvalds 

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:13 EST