From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net 00/11] netpoll: second round of fixes. Date: Fri, 28 Sep 2018 11:14:18 -0700 (PDT) Message-ID: <20180928.111418.1921695439745155316.davem@davemloft.net> References: <20180927163201.56609-1-edumazet@google.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, michael.chan@broadcom.com, aviad.krawczyk@huawei.com, songliubraving@fb.com, dougmill@linux.vnet.ibm.com, yisen.zhuang@huawei.com, mst@redhat.com, jasowang@redhat.com, harish.patil@cavium.com, manish.chopra@cavium.com, netanel@amazon.com, linux-net-drivers@solarflare.com, tlfalcon@linux.vnet.ibm.com To: edumazet@google.com Return-path: Received: from shards.monkeyblade.net ([23.128.96.9]:39860 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726150AbeI2AjS (ORCPT ); Fri, 28 Sep 2018 20:39:18 -0400 In-Reply-To: <20180927163201.56609-1-edumazet@google.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Eric Dumazet Date: Thu, 27 Sep 2018 09:31:50 -0700 > As diagnosed by Song Liu, ndo_poll_controller() can > be very dangerous on loaded hosts, since the cpu > calling ndo_poll_controller() might steal all NAPI > contexts (for all RX/TX queues of the NIC). > > This capture, showing one ksoftirqd eating all cycles > can last for unlimited amount of time, since one > cpu is generally not able to drain all the queues under load. > > It seems that all networking drivers that do use NAPI > for their TX completions, should not provide a ndo_poll_controller() : > > Most NAPI drivers have netpoll support already handled > in core networking stack, since netpoll_poll_dev() > uses poll_napi(dev) to iterate through registered > NAPI contexts for a device. > > First patch is a fix in poll_one_napi(). > > Then following patches take care of ten drivers. Series applied, thanks Eric.