From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [PATCH 3/4] net ipv4: When possible test for IFF_LOOPBACK and not dev == loopback_dev Date: Thu, 27 Sep 2007 10:10:27 -0600 Message-ID: References: <46FB873C.6060200@fr.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: David Miller , Linux Containers , netdev@vger.kernel.org, urs@isnogud.escape.de To: Daniel Lezcano Return-path: Received: from ebiederm.dsl.xmission.com ([166.70.28.69]:36136 "EHLO ebiederm.dsl.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753992AbXI0QLN (ORCPT ); Thu, 27 Sep 2007 12:11:13 -0400 In-Reply-To: <46FB873C.6060200@fr.ibm.com> (Daniel Lezcano's message of "Thu, 27 Sep 2007 12:34:36 +0200") Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Daniel Lezcano writes: > Eric W. Biederman wrote: >> Now that multiple loopback devices are becoming possible it makes >> the code a little cleaner and more maintainable to test if a deivice >> is th a loopback device by testing dev->flags & IFF_LOOPBACK instead >> of dev == loopback_dev. >> >> Signed-off-by: Eric W. Biederman > > > Urs Thuermann posted the patch: > > [PATCH 5/7] CAN: Add virtual CAN netdevice driver > > This network driver set its flag to IFF_LOOPBACK for testing. > Is it possible this can be a collision with your patch ? I have brought it up on that thread. As best as I tell the CAN usage of IFF_LOOPBACK will be a problem even without my patch. Assuming something other then the CAN layer will see the CAN devices. The CAN documentations IFF_LOOPBACK should be set on all CAN devices. It seems that the people who want high performance predictable CAN don't want this and the people who want something they can trace easily want this. It sounds to me like CAN routers don't exist. Anyway hopefully that usage can be resolved as that code is reviewed, and made ready to merge. Eric