From mboxrd@z Thu Jan 1 00:00:00 1970 From: John Dykstra Subject: Re: kernel dies if loopback device not intialized Date: Wed, 17 Jun 2009 21:51:22 -0500 Message-ID: <1245293482.10073.5.camel@merlyn> References: <20090610203536.4357d309@nehalam> <1245080498.7134.15.camel@Maple> <20090616215331.2a1bd276@nehalam> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: Stephen Hemminger Return-path: Received: from mail-yx0-f198.google.com ([209.85.210.198]:46486 "EHLO mail-yx0-f198.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752518AbZFRC42 (ORCPT ); Wed, 17 Jun 2009 22:56:28 -0400 Received: by yxe36 with SMTP id 36so503707yxe.33 for ; Wed, 17 Jun 2009 19:56:30 -0700 (PDT) In-Reply-To: <20090616215331.2a1bd276@nehalam> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 2009-06-16 at 21:53 -0700, Stephen Hemminger wrote: > The packets come from SNMP smux which opens a TCP socket over 127.1 > but the loopback interface is in admin down. Hmmm...the stack is doing the ARP because it doesn't have a valid route for 127.0.0.1. But I would expect it to be ARPing a default router, and getting a response (modulo your network state). What am I missing? > I suspect the packet > bouncing from ARP is the TCP SYN packet which does not have a > mac_header() > because it is going over the (down) loopback. Loopback packets have a L2 header--it's just not very interesting. Even if they never got an L2 header added, skb->mac_header shouldn't be garbage. -- John