Re: Why no printk for duplicate IPs?

David Lang (dlang@diginsite.com)
Thu, 30 Sep 1999 12:28:46 -0700 (PDT)


OK, I am misunderstanding what a gratuitus arp is supposed to do. I
understood that is was sending an arp RESPONSE for your address not an arp
REQUEST for your address. Sending a request for your address that you will
be responding to will only do some good if you have a machine set to
broadcast all arp responses (there was a comment that some machines not
unicast the arp response) and seems like additional overhead compared to
just sending the one packet.

David Lang

On Thu, 30 Sep 1999, B. James Phillippe wrote:

> On Thu, 30 Sep 1999, David Lang wrote:
>
> > On Wed, 29 Sep 1999, Christopher McCrory wrote:
> >
> > > John Fulmer wrote:
> > > >
> > > > The way I've seen this done (on BSDI boxen, I think) is that the
> > > > system does an arp for it's own IP address before/as the interface is
> > > > brought up. It it gets a response, it assumes that there is another
> > > > box out there with the same IP address (not a BAD assumption).
> > >
> > > This IIRC is a gratuitous arp. In the kernel source I have seen
> ...
> > no a gratuttous arp is a broadcast ARP response, telling eveyone who you
> > are before they ask. arping to see if someone else has your address is a
> > different ballgame.
>
> Well, actually the original comment _does_ describe a gratuitous ARP, which
> is defined as an ARP request sent for your own protocol address. All ARP
> requests are broadcast. There are no responses to a gratuitous ARP, unless
> there is an IP conflict. However, you're right in that the gratuitous ARP
> doesn't itself have anything to do with address conflict detection. That
> would be an additional (rather simple) step. Gratuitous ARP's alone are
> handy for hot-swapping routers or NIC's.
>
> -bp
> --
> # bryan at terran dot org
> # http://www.terran.org/~bryan
>
>

-
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/