From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rick Jones Subject: Re: [PATCH net-next] net: gro: add a per device gro flush timer Date: Wed, 05 Nov 2014 17:38:35 -0800 Message-ID: <545AD11B.5050603@hp.com> References: <1415235320.13896.51.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: netdev , Or Gerlitz , Willem de Bruijn To: Eric Dumazet , David Miller Return-path: Received: from g2t1383g.austin.hp.com ([15.217.136.92]:3618 "EHLO g2t1383g.austin.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750877AbaKFBii (ORCPT ); Wed, 5 Nov 2014 20:38:38 -0500 Received: from g2t2352.austin.hp.com (g2t2352.austin.hp.com [15.217.128.51]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by g2t1383g.austin.hp.com (Postfix) with ESMTPS id 64A3DBAD for ; Thu, 6 Nov 2014 01:38:38 +0000 (UTC) In-Reply-To: <1415235320.13896.51.camel@edumazet-glaptop2.roam.corp.google.com> Sender: netdev-owner@vger.kernel.org List-ID: On 11/05/2014 04:55 PM, Eric Dumazet wrote: > Tested: > Ran 200 netperf TCP_STREAM from A to B (10Gbe link, 8 RX queues) > > Without this feature, we send back about 305,000 ACK per second. > > GRO aggregation ratio is low (811/305 = 2.65 segments per GRO packet) > > Setting a timer of 2000 nsec is enough to increase GRO packet sizes > and reduce number of ACK packets. (811/19.2 = 42) > > Receiver performs less calls to upper stacks, less wakes up. > This also reduces cpu usage on the sender, as it receives less ACK > packets. > > Note that reducing number of wakes up increases cpu efficiency, but can > decrease QPS, as applications wont have the chance to warmup cpu caches > doing a partial read of RPC requests/answers if they fit in one skb. Speaking of QPS, what happens to 200 TCP_RR tests when the feature is enabled? rick jones