From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [RFC PATCH] net: add additional lock to qdisc to increase
 enqueue/dequeue fairness
Date: Wed, 02 Jun 2010 11:48:28 +0200
Message-ID: <1275472108.2725.142.camel@edumazet-laptop>
References: <20100323202553.21598.10754.stgit@gitlad.jf.intel.com>
	 <1269377667.2915.25.camel@edumazet-laptop>
	 <20100323.144512.140757007.davem@davemloft.net>
	 <1269382380.2915.40.camel@edumazet-laptop>
	 <20100521084349.0d6f8f9a@nehalam>
	 <1274460275.2439.469.camel@edumazet-laptop>
	 <1274463643.2439.473.camel@edumazet-laptop>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: alexander.h.duyck@intel.com, netdev@vger.kernel.org
To: Stephen Hemminger <shemminger@vyatta.com>,
	David Miller <davem@davemloft.net>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-fx0-f46.google.com ([209.85.161.46]:59947 "EHLO
	mail-fx0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1754928Ab0FBJse (ORCPT
	<rfc822;netdev@vger.kernel.org>); Wed, 2 Jun 2010 05:48:34 -0400
Received: by fxm8 with SMTP id 8so1680493fxm.19
        for <netdev@vger.kernel.org>; Wed, 02 Jun 2010 02:48:33 -0700 (PDT)
In-Reply-To: <1274463643.2439.473.camel@edumazet-laptop>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

Le vendredi 21 mai 2010 =C3=A0 19:40 +0200, Eric Dumazet a =C3=A9crit :
> Le vendredi 21 mai 2010 =C3=A0 18:44 +0200, Eric Dumazet a =C3=A9crit=
 :
>=20
> > We could use cmpxchg() and manipulate several bits at once in fast =
path.
> > ( __QDISC_STATE_RUNNING, __QDISC_STATE_LOCKED ... ) but making the =
crowd
> > of cpus spin on the same bits/cacheline than dequeue worker would
> > definitely slowdown the worker.
> >=20
> >=20
>=20
> Maybe I am missing something, but __QDISC_STATE_RUNNING is always
> manipulated with the lock held...
>=20
> We might avoid two atomic ops when changing this state (if moved to a
> separate container) in fast path (when a cpu sends only one packet an=
d
> returns)
>=20
>=20

Here are two patches to implement this idea.

=46irst patch to abstract QDISC_STATE_RUNNING access.

Second patch to add a __qstate container and remove the atomic ops.