From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: 3.3-rc3+ Crash in __neigh_for_each_release Date: Tue, 21 Feb 2012 20:15:56 +0100 Message-ID: <1329851756.18384.47.camel@edumazet-laptop> References: <1329851004.18384.45.camel@edumazet-laptop> <20120221.140716.1396286389544946379.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: mroos@linux.ee, netdev@vger.kernel.org To: David Miller Return-path: Received: from mail-ww0-f44.google.com ([74.125.82.44]:61846 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755818Ab2BUTQB (ORCPT ); Tue, 21 Feb 2012 14:16:01 -0500 Received: by wgbdt10 with SMTP id dt10so5989396wgb.1 for ; Tue, 21 Feb 2012 11:16:00 -0800 (PST) In-Reply-To: <20120221.140716.1396286389544946379.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: Le mardi 21 f=C3=A9vrier 2012 =C3=A0 14:07 -0500, David Miller a =C3=A9= crit : > From: Eric Dumazet > Date: Tue, 21 Feb 2012 20:03:24 +0100 >=20 > > But I dont know enough this code to know if the following patch is = the > > way to fix this. (and __neigh_for_each_release() can also be delete= d if > > no users left in tree) >=20 > I think instead of removing the code, we need to have it iterate over > "arp_tbl" but only invoke the callback for devices which are of type > ATM. That makes sense... Or invoke callback for all entries, and filter in callback non ATM ones= =2E