From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Madden Subject: Re: Oops in filter add Date: Tue, 20 Mar 2007 12:27:21 -0400 Message-ID: <46000B69.8000401@reflexsecurity.com> References: <45FEEE35.6090606@reflexsecurity.com> <20070319.192206.21926062.davem@davemloft.net> <1174373645.4895.15.camel@localhost> <45FFEC78.2090708@reflexsecurity.com> <45FFF455.80003@trash.net> <45FFF675.8040501@trash.net> <20070320151151.GM521@postel.suug.ch> <45FFFA36.7080404@trash.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Cc: Thomas Graf , hadi@cyberus.ca, netdev@vger.kernel.org To: Patrick McHardy Return-path: Received: from crown.reflexsecurity.com ([72.54.139.163]:46912 "EHLO crown.reflexsecurity.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751784AbXCTQ1Y (ORCPT ); Tue, 20 Mar 2007 12:27:24 -0400 In-Reply-To: <45FFFA36.7080404@trash.net> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Patrick McHardy wrote: > > There was a small bug in my patch (broken return value on memory > allocation failure, not relevant for testing though). I'll push > the fixed patch to Dave once Chris confirms that it fixes the > problem he's seeing. > Looks like that may have done it. I ramped up the traffic on it and did thousands of inserts/removals and the box seems happy. I had ing_filter still locking dev->ingress_lock ( instead of the queue_lock suggestion earlier ). So while that race may still be in there, I haven't been able to hit the locking problem. chris