From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH v4 0/10] bql: Byte Queue Limits Date: Tue, 29 Nov 2011 09:43:54 +0100 Message-ID: <1322556234.2970.84.camel@edumazet-laptop> References: <1322550138.2970.70.camel@edumazet-laptop> <4ED4885F.8060309@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: John Fastabend , Tom Herbert , "davem@davemloft.net" , "netdev@vger.kernel.org" To: Dave Taht Return-path: Received: from mail-fx0-f46.google.com ([209.85.161.46]:36758 "EHLO mail-fx0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752604Ab1K2IoA (ORCPT ); Tue, 29 Nov 2011 03:44:00 -0500 Received: by faaq16 with SMTP id q16so26011faa.19 for ; Tue, 29 Nov 2011 00:43:59 -0800 (PST) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: Le mardi 29 novembre 2011 =C3=A0 09:37 +0100, Dave Taht a =C3=A9crit : > On Tue, Nov 29, 2011 at 8:23 AM, John Fastabend > wrote: > > > > I wonder if we should consider enabling TSO/GSO per queue or per tr= affic > > class on devices that support this. At least in devices that suppor= t > > multiple traffic classes it seems to be a common usage case to put = bulk > > storage traffic (iSCSI) on a traffic class and low latency traffic = on a > > separate traffic class, VoIP for example. >=20 >=20 > VOIP is a drop in the bucket. >=20 > Turning TSO off on TCP exiting the datacenter (or more specifically), > destined anywhere there is potential tx/rx bandwidth disparity > would be goooooood. >=20 If your cpu is fast enough (and they are most of the time), this makes no difference at all. Instead of consuming 3% of cpu with TSO, you'll consume 10% or 15% and no difference seen on the wire. Really, if you want to avoid bursts, TSO has litle to do with them.