From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [PATCH] ifb: add multi-queue support
Date: Sat, 14 Nov 2009 13:53:04 +0100
Message-ID: <4AFEA830.3000101@gmail.com>
References: <412e6f7f0911122216u6880e855g6a15dac29ad6a100@mail.gmail.com>	 <20091113074508.GA6605@ff.dom.local>	 <412e6f7f0911130054i7a508a6ah16368f11bdc7353d@mail.gmail.com>	 <20091113091825.GA7449@ff.dom.local>	 <412e6f7f0911130138td181935w36cab3119972753e@mail.gmail.com>	 <20091113081553.0568296c@s6510>	 <412e6f7f0911131528i3105648y49213143e12f2522@mail.gmail.com>	 <20091113153229.408c889f@nehalam> <412e6f7f0911131542w1029893cmd07599660b810ed5@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Stephen Hemminger <shemminger@vyatta.com>,
	Jarek Poplawski <jarkao2@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Patrick McHardy <kaber@trash.net>,
	Tom Herbert <therbert@google.com>, netdev@vger.kernel.org
To: Changli Gao <xiaosuo@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from gw1.cosmosbay.com ([212.99.114.194]:37383 "EHLO
	gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1751362AbZKNMxH (ORCPT
	<rfc822;netdev@vger.kernel.org>); Sat, 14 Nov 2009 07:53:07 -0500
In-Reply-To: <412e6f7f0911131542w1029893cmd07599660b810ed5@mail.gmail.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

Changli Gao a =C3=A9crit :
> Yea, the overhead of SoftIRQ is less than kernel threads, and I'll tr=
y
> to find a way to solve both flexibility and efficiency. Maybe I need
> some real NIC drivers as examples. Is there a standard API to bind RQ=
s
> of NIC to CPUs, such as ioctl or setsockopt?

Thats a good question... napi is bound to a cpu, and you'll need
things that were done by Tom Herbert in its RPS patch, to
be able to deleguate work to remote cpus napi contexts.

But if we consider RPS being close to be committed, shouldnt
IFB use its, in order to not duplicate changes ?

Or, if you prefer, once RPS is in, we dont need to change IFB, since
RPS will already split the load to multiple cpus before entering IFB.