From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Poirier Subject: Re: notifications of interfaces going down/up/being created/destroyed? Date: Mon, 30 Mar 2015 13:59:34 -0700 Message-ID: <20150330205934.GA29542@f1.synalogic.ca> References: Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "netdev@vger.kernel.org" To: Dave Taht Return-path: Received: from mail-pd0-f176.google.com ([209.85.192.176]:33946 "EHLO mail-pd0-f176.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753202AbbC3U7i (ORCPT ); Mon, 30 Mar 2015 16:59:38 -0400 Received: by pdbni2 with SMTP id ni2so187147465pdb.1 for ; Mon, 30 Mar 2015 13:59:38 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On 2015/03/28 09:37, Dave Taht wrote: > is there an asynchronous way to register for a notification of an > interface going down/up? Look at what `ip monitor` is doing. >=20 > ---------- Forwarded message ---------- > From: Juliusz Chroboczek > Date: Sat, Mar 28, 2015 at 9:30 AM > Subject: Re: [Babel-users] Looping in EAGAIN > To: Dave Taht > Cc: Simon Kelley , > "babel-users@lists.alioth.debian.org" > >=20 >=20 > > Say wlan0 vanishes. All the routes going out that interface are no > > longer valid, but from what I understood of this patch, it will loo= p for > > a while, then give up. >=20 > If wlan0 vanishes, this will be recognised the next time check_interf= aces > is run, and all neighbours visible through wlan0 will be flushed. >=20 > The issue we're having is a race condition -- if wlan0 goes down and = then > back up before we run check_interfaces, and the IP addresses don't ch= ange, > then check_interfaces will not notice the transition, and Babel will = think > that its routes through wlan0 are still up -- end you end up with a F= IB > that is not a subset of the RIB. Ouch. >=20 > So it might be a good idea to run check_interfaces early when we get > EAGAIN, but I'm not sure what consequences it might have -- EAGAIN ca= n > also happen when we're under load, and we'd rather not be repeatedly > scanning our interfaces in that case. >=20 > It would be better to get async notifications from the kernel about > interfaces going down. >=20 > > Not clear to me if this would happen for 4 hello intervals before t= he > > interface is recognised as gone? >=20 > No, the hellos are used to notice vanishing neighbours, not vanishing > interfaces. That's a completely different mechanism. >=20 > -- Juliusz >=20 >=20 > --=20 > Dave T=E4ht > Let's make wifi fast, less jittery and reliable again! >=20 > https://plus.google.com/u/0/107942175615993706558/posts/TVX3o84jjmb > -- > To unsubscribe from this list: send the line "unsubscribe netdev" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html