From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: RPS and forwarding Date: Mon, 26 Apr 2010 09:25:16 +0200 Message-ID: <1272266716.2346.5.camel@edumazet-laptop> References: <20100426045316.GA21810@gondor.apana.org.au> <20100425.232857.145290769.davem@davemloft.net> <20100426063136.GA22401@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , therbert@google.com, netdev@vger.kernel.org To: Herbert Xu Return-path: Received: from mail-bw0-f219.google.com ([209.85.218.219]:37685 "EHLO mail-bw0-f219.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754533Ab0DZHZV (ORCPT ); Mon, 26 Apr 2010 03:25:21 -0400 Received: by bwz19 with SMTP id 19so181811bwz.21 for ; Mon, 26 Apr 2010 00:25:20 -0700 (PDT) In-Reply-To: <20100426063136.GA22401@gondor.apana.org.au> Sender: netdev-owner@vger.kernel.org List-ID: Le lundi 26 avril 2010 =C3=A0 14:31 +0800, Herbert Xu a =C3=A9crit : > On Sun, Apr 25, 2010 at 11:28:57PM -0700, David Miller wrote: >=20 > > It might be possible to somehow make the sock table get bypassed fo= r > > forwarded traffic, but I can't think of a cheap way to do that at > > the moment. >=20 > Yeah I thought about that too but as we need to go through the > routing table before we know whether something is forwarded or > not, it certainly seems non-trivial. >=20 > Hmm, maybe if we used a routing cache keyed by rxhash we could > make an appromixation? After all, we only need to make sure that > no forwarded traffic is redirected, and not the reverse. IOW, > it's OK if we incorrectly classify some local traffic as forwarded. >=20 RFS is already working like that, if RPS is not used in conjonction. =46orwarded traffic will see a non tagged rfs flow entry, so this skb w= ill be processed by this cpu, not a remote one. Currently, only an application can set a RFS flow entry.