From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753649AbbJNSDl (ORCPT ); Wed, 14 Oct 2015 14:03:41 -0400 Received: from resqmta-ch2-12v.sys.comcast.net ([69.252.207.44]:59065 "EHLO resqmta-ch2-12v.sys.comcast.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752946AbbJNSDk (ORCPT ); Wed, 14 Oct 2015 14:03:40 -0400 Date: Wed, 14 Oct 2015 13:03:38 -0500 (CDT) From: Christoph Lameter X-X-Sender: cl@east.gentwo.org To: Linus Torvalds cc: Tejun Heo , Andrew Morton , Michal Hocko , Linux Kernel Mailing List , Lai Jiangshan , Shaohua Li , linux-mm Subject: Re: [GIT PULL] workqueue fixes for v4.3-rc5 In-Reply-To: Message-ID: References: <20151013214952.GB23106@mtj.duckdns.org> Content-Type: text/plain; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 14 Oct 2015, Linus Torvalds wrote: > On Tue, Oct 13, 2015 at 2:49 PM, Tejun Heo wrote: > > > > Single patch to fix delayed work being queued on the wrong CPU. This > > has been broken forever (v2.6.31+) but obviously doesn't trigger in > > most configurations. > > So why is this a bugfix? If cpu == WORK_CPU_UNBOUND, then things > _shouldn't_ care which cpu it gets run on. UNBOUND means not fixed to a processor. The system should have freedom to schedule unbound work requests anywhere it wants. This is something we also want for the NOHZ work in order to move things like these workqueue items to processors that are not supposed to be low latency.