From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lance Richardson Subject: Re: [net v3] veth: advertise peer link once both links are tied together Date: Fri, 10 Jun 2016 09:20:58 -0400 (EDT) Message-ID: <1764704950.51133004.1465564858359.JavaMail.zimbra@redhat.com> References: <574C6095.9050804@6wind.com> <1464623917-11536-1-git-send-email-vincent@bernat.im> <87bn3n358j.fsf@zoro.exoscale.ch> <574C69DE.3050305@6wind.com> <574D56A0.3090606@6wind.com> <694175272.49661350.1465417849878.JavaMail.zimbra@redhat.com> <575ABD55.7020202@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Vincent Bernat , "David S. Miller" , Vijay Pandurangan , Paolo Abeni , netdev@vger.kernel.org To: nicolas dichtel Return-path: Received: from mx4-phx2.redhat.com ([209.132.183.25]:50626 "EHLO mx4-phx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932783AbcFJNVE convert rfc822-to-8bit (ORCPT ); Fri, 10 Jun 2016 09:21:04 -0400 In-Reply-To: <575ABD55.7020202@6wind.com> Sender: netdev-owner@vger.kernel.org List-ID: ----- Original Message ----- > From: "Nicolas Dichtel" > To: "Lance Richardson" , "Vincent Bernat" > Cc: "David S. Miller" , "Vijay Pandurangan" , "Paolo Abeni" > , netdev@vger.kernel.org > Sent: Friday, June 10, 2016 9:15:01 AM > Subject: Re: [net v3] veth: advertise peer link once both links are t= ied together >=20 > Le 08/06/2016 22:30, Lance Richardson a =C3=A9crit : > [snip] > > I've been pondering how to fix this very problem off-and-on for a f= ew > > months > > now, without having arrived at any solution that was particularly > > satisfying. > >=20 > > The main constraints I've been trying to meet are: > > - User-space should be informed of veth pairing for both peers. > > - RTM_NEWLINK messages should not refer to interfaces that haven= 't > > been announced to user-space via previous RTM_NEWLINK messages= =2E > > - The first (and only the first) RTM_NEWLINK message for a given > > interface should have ifi_changes set to ~0U, subsequent RTM_N= EWLINK > > messages should have ifi_changes set to reflect the flags that > > have changed. > >=20 > > This is the closest I've come to a satisfactory solution, it does m= eet > > the above constraints but still seems a little unnatural to me: > The patch looks good to me. Can you submit it formally? >=20 Will post in a bit, thanks! Lance