From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jose Abreu Subject: Re: [PATCH] Revert "net: stmmac: Do not keep rearming the coalesce timer in stmmac_xmit" Date: Tue, 28 Aug 2018 09:12:33 +0100 Message-ID: References: <20180824090440.13411-1-jbrunet@baylibre.com> Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit Cc: , , Joao Pinto , Vitor Soares , Corentin Labbe To: Jerome Brunet , Giuseppe Cavallaro , Alexandre Torgue , "Jose Abreu" , Return-path: Received: from us01smtprelay-2.synopsys.com ([198.182.47.9]:54818 "EHLO smtprelay.synopsys.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726923AbeH1MDJ (ORCPT ); Tue, 28 Aug 2018 08:03:09 -0400 In-Reply-To: <20180824090440.13411-1-jbrunet@baylibre.com> Sender: netdev-owner@vger.kernel.org List-ID: Hi Jerome, On 24-08-2018 10:04, Jerome Brunet wrote: > This reverts commit 4ae0169fd1b3c792b66be58995b7e6b629919ecf. > > This change in the handling of the coalesce timer is causing regression on > (at least) amlogic platforms. > > Network will break down very quickly (a few seconds) after starting > a download. This can easily be reproduced using iperf3 for example. > > The problem has been reported on the S805, S905, S912 and A113 SoCs > (Realtek and Micrel PHYs) and it is likely impacting all Amlogics > platforms using Gbit ethernet > > No problem was seen with the platform using 10/100 only PHYs (GXL internal) > > Reverting change brings things back to normal and allows to use network > again until we better understand the problem with the coalesce timer. > > Apologies for the delayed answer but I was in FTO. I'm not sure what can be causing this but I have some questions for you: - What do you mean by "network will break down"? Do you see queue timeout? - What do you see in ethtool/ifconfig stats? Can you send me the stats before and after network break? - Is your setup multi-queue/channel? - Can you point me to the DT bindings of your setup? Thanks and Best Regards, Jose Miguel Abreu