From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [BUG,REGRESSION?] 3.11.6+,3.12: GbE iface rate drops to few KB/s Date: Mon, 18 Nov 2013 09:13:28 -0800 Message-ID: <1384794808.8604.76.camel@edumazet-glaptop2.roam.corp.google.com> References: <8761s0cqhh.fsf@natisbad.org> <87y54u59zq.fsf@natisbad.org> <20131112083633.GB10318@1wt.eu> <87a9hagex1.fsf@natisbad.org> <20131112100126.GB23981@1wt.eu> <87vbzxd473.fsf@natisbad.org> <20131113072257.GB10591@1wt.eu> <20131117141940.GA18569@1wt.eu> <20131118112601.65feb708@skate> <20131118104448.GI16823@kw.sim.vm.gnt> <20131118085407.25dadcaf@samsung-9> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Simon Guinot , Thomas Petazzoni , Willy Tarreau , Arnaud Ebalard , Cong Wang , edumazet@google.com, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, Vincent Donnefort To: Stephen Hemminger Return-path: Received: from mail-pb0-f43.google.com ([209.85.160.43]:45431 "EHLO mail-pb0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751375Ab3KRRNc (ORCPT ); Mon, 18 Nov 2013 12:13:32 -0500 Received: by mail-pb0-f43.google.com with SMTP id rq2so4910311pbb.2 for ; Mon, 18 Nov 2013 09:13:31 -0800 (PST) In-Reply-To: <20131118085407.25dadcaf@samsung-9> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, 2013-11-18 at 08:54 -0800, Stephen Hemminger wrote: > On Mon, 18 Nov 2013 11:44:48 +0100 > Simon Guinot wrote: > > c9eeec26 tcp: TSQ can use a dynamic limit > > > > But without that patch there was a performance regression for high speed > interfaces whihc was caused by TSQ. 10G performance dropped to 8G Yes, this made sure we could feed more than 2 TSO packets on TX ring. But decreasing minimal amount of queuing from 128KB to ~1ms of the current rate did not please NIC which can have a big delay between ndo_start_xmit() and actual skb freeing (TX completion) So http://git.kernel.org/cgit/linux/kernel/git/davem/net.git/commit/?id=98e09386c0ef4dfd48af7ba60ff908f0d525cdee restored this minimal amount of buffering, and let the bigger amount for 40Gb NICs ;) From mboxrd@z Thu Jan 1 00:00:00 1970 From: eric.dumazet@gmail.com (Eric Dumazet) Date: Mon, 18 Nov 2013 09:13:28 -0800 Subject: [BUG,REGRESSION?] 3.11.6+,3.12: GbE iface rate drops to few KB/s In-Reply-To: <20131118085407.25dadcaf@samsung-9> References: <8761s0cqhh.fsf@natisbad.org> <87y54u59zq.fsf@natisbad.org> <20131112083633.GB10318@1wt.eu> <87a9hagex1.fsf@natisbad.org> <20131112100126.GB23981@1wt.eu> <87vbzxd473.fsf@natisbad.org> <20131113072257.GB10591@1wt.eu> <20131117141940.GA18569@1wt.eu> <20131118112601.65feb708@skate> <20131118104448.GI16823@kw.sim.vm.gnt> <20131118085407.25dadcaf@samsung-9> Message-ID: <1384794808.8604.76.camel@edumazet-glaptop2.roam.corp.google.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Mon, 2013-11-18 at 08:54 -0800, Stephen Hemminger wrote: > On Mon, 18 Nov 2013 11:44:48 +0100 > Simon Guinot wrote: > > c9eeec26 tcp: TSQ can use a dynamic limit > > > > But without that patch there was a performance regression for high speed > interfaces whihc was caused by TSQ. 10G performance dropped to 8G Yes, this made sure we could feed more than 2 TSO packets on TX ring. But decreasing minimal amount of queuing from 128KB to ~1ms of the current rate did not please NIC which can have a big delay between ndo_start_xmit() and actual skb freeing (TX completion) So http://git.kernel.org/cgit/linux/kernel/git/davem/net.git/commit/?id=98e09386c0ef4dfd48af7ba60ff908f0d525cdee restored this minimal amount of buffering, and let the bigger amount for 40Gb NICs ;)