Date: 04 Sep 1996 14:47:26 +0100 From: Paul Richards <p.richards@elsevier.co.uk> To: Gary Palmer <gpalmer@freebsd.org> Cc: Garrett Wollman <wollman@lcs.mit.edu>, Ulf Zimmermann <ulf@lamb.net>, bugs@freebsd.org Subject: Re: 2 bugs, I think ;-) Message-ID: <57u3te8k9d.fsf@elsevier.co.uk> In-Reply-To: Gary Palmer's message of Tue, 03 Sep 1996 16:50:00 -0400 References: <8119.841783800@orion.webspan.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Gary Palmer <gpalmer@freebsd.org> writes: > Ulf has a 4 ``port'' Zynx ethernet card, based around the 2104x > chipset. (I was talking to him on IRC about this). If he took a > machine connected to one of the interfaces (say de0, which has aliases > defined on it) and tracerouted ACROSS the machine with the Zynx card > (to another net on one of the other interfaces, since this is a > router), traceroute returned the address of the last alias added to > the de0 interface as being the first hop, not the ``primary'' address > of the interface, which IMHO (and his too) should be the address > returned. There's no concept of "primary" address in the kernel, they just sit on a linked list. The kernel does a match of those addresses with the destination address to determine the "source" address. I forget the function that does this but it's likely it's just matching the address on the end of the list straight away. At least, it's something like that, Garrett probably knows the details. I ran into this problem with the virtual hosting box I set up. I couldn't control the "source" address of the packets that were originating from programs running in the virtual host environments. -- Paul Richards. Originative Solutions Ltd. (Netcraft Ltd. contractor) Elsevier Science TIS online journal project. Email: p.richards@elsevier.co.uk Phone: 0370 462071 (Mobile), +44 (0)1865 843155
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?57u3te8k9d.fsf>