From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lawrence Brakmo Subject: Re: [PATCH net-next] tcp: ack immediately when a cwr packet arrives Date: Tue, 24 Jul 2018 17:42:46 +0000 Message-ID: <36385AF4-4EE4-4237-84B8-1978B091E28B@fb.com> References: <20180724004939.2874202-1-brakmo@fb.com> <3c785042-0b9a-69cd-9415-e8a10ea0baf7@iogearbox.net> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-7" Content-Transfer-Encoding: quoted-printable Cc: Daniel Borkmann , Netdev , Kernel Team , Alexei Starovoitov , "Eric Dumazet" To: Neal Cardwell , Yuchung Cheng Return-path: Received: from mx0b-00082601.pphosted.com ([67.231.153.30]:41408 "EHLO mx0b-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S2388414AbeGXSuu (ORCPT ); Tue, 24 Jul 2018 14:50:50 -0400 In-Reply-To: Content-Language: en-US Content-ID: Sender: netdev-owner@vger.kernel.org List-ID: Note that without this fix the 99+ACU- latencies when doing 10KB RPCs in a = congested network using DCTCP are 40ms vs. 190us with the patch. Also note = that these 40ms high tail latencies started after commit 3759824da87b30ce7a= 35b4873b62b0ba38905ef5 in Jul 2015, which triggered the bugs/features we ar= e fixing/adding. I agree it is a debatable whether it is a bug fix or a fea= ture improvement and I am fine either way. Lawrence +/v8-On 7/24/18, 10:13 AM, +ACI-Neal Cardwell+ACI- +ADw-ncardwell+AEA-googl= e.com+AD4- wrote: On Tue, Jul 24, 2018 at 1:07 PM Yuchung Cheng +ADw-ycheng+AEA-google.co= m+AD4- wrote: +AD4- +AD4- On Mon, Jul 23, 2018 at 7:23 PM, Daniel Borkmann +ADw-daniel+AEA-= iogearbox.net+AD4- wrote: +AD4- +AD4- Should this go to net tree instead where all the other fixe= s went? +AD4- I am neutral but this feels more like a feature improvement =20 I agree this feels like a feature improvement rather than a bug fix. =20 neal =20