From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751730AbdHQN0k (ORCPT ); Thu, 17 Aug 2017 09:26:40 -0400 Received: from mail-qt0-f194.google.com ([209.85.216.194]:33630 "EHLO mail-qt0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750857AbdHQN0j (ORCPT ); Thu, 17 Aug 2017 09:26:39 -0400 Date: Thu, 17 Aug 2017 06:26:35 -0700 From: Tejun Heo To: Boqun Feng Cc: linux-kernel@vger.kernel.org, Byungchul Park , Peter Zijlstra , Ingo Molnar , Lai Jiangshan Subject: Re: [PATCH] workqueue/lockdep: Explicitly initialize wq_barrier::done::map Message-ID: <20170817132635.GE3238792@devbig577.frc2.facebook.com> References: <20170817094622.12915-1-boqun.feng@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170817094622.12915-1-boqun.feng@gmail.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Aug 17, 2017 at 05:46:12PM +0800, Boqun Feng wrote: > With CROSSRELEASE feature introduced, such a potential deadlock is > reported: > > > Worker A : acquired of wfc.work -> wait for cpu_hotplug_lock to be released > > Task B : acquired of cpu_hotplug_lock -> wait for lock#3 to be released > > Task C : acquired of lock#3 -> wait for completion of barr->done > > (Task C is in lru_add_drain_all_cpuslocked()) > > Worker D : wait for wfc.work to be released -> will complete barr->done > > However, given this very case, such a dead lock could not happen because > Task C's barr->done and Worker D's barr->done can not be the same > instance. So this is a false positive. > > The reason of this false positive is we initialize all wq_barrier::done > at insert_wq_barrier() via init_completion(), which makes them belong to > the same lock class, therefore, impossible circles are reported. > > To fix this, explicitly initialize the lockdep map for wq_barrier::done > in insert_wq_barrier(), so that the lock class key of wq_barrier::done > is a subkey of the corresponding work_struct, as a result we won't build > a dependency between a wq_barrier with a unrelated work, and we can > differ wq barriers based on the related works, so the false positive > above is avoided. > > Also define the empty lockdep_init_map_crosslock() for !CROSSRELEASE > to make the code simple and away from unnecessary #ifdefs. > > Reported-by: Ingo Molnar > Signed-off-by: Boqun Feng > Cc: Byungchul Park Acked-by: Tejun Heo Thanks. -- tejun