From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH v2 net-next] net: use jump_label to shortcut RPS if not setup Date: Thu, 17 Nov 2011 23:19:16 +0100 Message-ID: <1321568356.2444.7.camel@edumazet-laptop> References: <1321535606.2751.35.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> <1321565689.2444.5.camel@edumazet-laptop> <20111117.170442.1731100546322591134.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, therbert@google.com To: David Miller Return-path: Received: from mail-ww0-f44.google.com ([74.125.82.44]:60207 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753496Ab1KQWTV (ORCPT ); Thu, 17 Nov 2011 17:19:21 -0500 Received: by wwe5 with SMTP id 5so3864330wwe.1 for ; Thu, 17 Nov 2011 14:19:20 -0800 (PST) In-Reply-To: <20111117.170442.1731100546322591134.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: Le jeudi 17 novembre 2011 =C3=A0 17:04 -0500, David Miller a =C3=A9crit= : > From: Eric Dumazet > Date: Thu, 17 Nov 2011 22:34:49 +0100 >=20 > > I wonder if we could use this jump_label infrastructure to shortcut > > expensive netfilter nf_hook_slow() / nf_iterate on empty > > tables/chains... >=20 > That would be nice, I can't see why it wouldn't be possible. Almost done :) Only problem is arch_static_branch() has a "i" (key), so some calls won= t be optimized (hooknum in a variable, not a constant) for example in nf_ct_frag6_output() Not a big deal, most calls can be optimized away.