On Fri, 28 Jun 2002, Bernd Eckenfels wrote:
> > I think you misunderstood my problem. I stated that V1 and V2 donot
> > generate ARP (NOARP flag is set). Instead they use eth0 to xmit
> > their packets. So the ethernet header carries the source and
> > destination mac addresses of machines 1 and 2 rather than V1 & V2.
>
> I thought u just saied u generate them by eth0, why should they have 2
> different MACs? Have you hardwired the source mac and are you sure
> that the computer actually send them that way?
The pkt generated by the application (on m/c 1) passes thru the virtual
device V1 and then piggybacks over eth0 to be sent to another machine.
On the other machine (m/c 2), the packet is recieved over its eth0 and
then passed on to virtual device V2.
Yes, the mac addresses are hardwired. When a pkt is sent from m/c 1, is
has src mac address of m/c 1 eth0 and destination mac addr of m/c 2
eth0.
> > BTW, for this experiment, I only have unidirectional UDP traffic. So
> > I dont expect any packets back on the reverse path. Not that it
> > should really matter, should it?
>
> Sure it should. How should the switch learn it if it does not receive
> packets?
I guess that is the main drawback in my experimental stategy: I send
traffic only in one direction. In that case as everybody has pointed
out, I need a way to ensure that the switch knows & remembers the mac
addresses of all connected interfaces.
Thanks everybody in helping out on this one. It was a subtle one, since
I *assumed* that the hardware 'just works'.
ciao,
Amit
--
I'm an angel!!! Honest!
The horns are just there to hold the halo up straight.
^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^
Amit Kucheria
EECS Grad. Research Assistant
University of Kansas @ Lawrence
(R): +1-785-830-8521 ||| (C): +1-785-760-2871
____________________________________________________
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
