From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: netfilter conntrack tcp lock Date: Wed, 22 Apr 2009 06:57:21 +0200 Message-ID: <49EEA3B1.9030803@cosmosbay.com> References: <20090421210412.268fc4e7@nehalam> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "Paul E. McKenney" , netfilter-devel@vger.kernel.org To: Stephen Hemminger Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:60042 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750821AbZDVE52 convert rfc822-to-8bit (ORCPT ); Wed, 22 Apr 2009 00:57:28 -0400 In-Reply-To: <20090421210412.268fc4e7@nehalam> Sender: netfilter-devel-owner@vger.kernel.org List-ID: Stephen Hemminger a =E9crit : > Not sure what the performance impact would be but simply changing tcp= _lock > (in nf_conntrack_proto_tcp) to a spin_lock might get a performance bo= ost. > I thought I heard Paul say read_locks are way slower than spin_lock. >=20 > Alternatively, going to some form of bit based hash lock might work. >=20 >=20 I thought Patrick was OK with adding a spinlock into each nf_conn struc= t ? (We shrunk this struct with the rcu_head removal) A straight forward locking schem. We have spinlocks everywhere in kerne= l afterall, and conntracking deserves this. =20 Just resubmit one of your prior patch perhaps ? I can dig it in a coupl= e of hours eventually. Thanks -- To unsubscribe from this list: send the line "unsubscribe netfilter-dev= el" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html