From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933142AbcETKpG (ORCPT ); Fri, 20 May 2016 06:45:06 -0400 Received: from outbound-smtp12.blacknight.com ([46.22.139.17]:45711 "EHLO outbound-smtp12.blacknight.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932232AbcETKpF (ORCPT ); Fri, 20 May 2016 06:45:05 -0400 Date: Fri, 20 May 2016 11:45:00 +0100 From: Mel Gorman To: Ingo Molnar Cc: Peter Zijlstra , Davidlohr Bueso , manfred@colorfullife.com, Waiman.Long@hpe.com, torvalds@linux-foundation.org, ggherdovich@suse.com, linux-kernel@vger.kernel.org Subject: Re: sem_lock() vs qspinlocks Message-ID: <20160520104500.GL2527@techsingularity.net> References: <20160520053926.GC31084@linux-uzut.site> <20160520083008.GD3193@twins.programming.kicks-ass.net> <20160520090034.GO3205@twins.programming.kicks-ass.net> <20160520100901.GA20016@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <20160520100901.GA20016@gmail.com> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, May 20, 2016 at 12:09:01PM +0200, Ingo Molnar wrote: > > * Peter Zijlstra wrote: > > > On Fri, May 20, 2016 at 10:30:08AM +0200, Peter Zijlstra wrote: > > > On Thu, May 19, 2016 at 10:39:26PM -0700, Davidlohr Bueso wrote: > > > > Specifically > > > > for the 'cascade_cond' and 'cascade_flock' programs, which exhibit hangs in libc's > > > > semop() blocked waiting for zero. > > > > > > OK; so I've been running: > > > > > > while :; do > > > bin/cascade_cond -E -C 200 -L -S -W -T 200 -I 2000000 ; > > > bin/cascade_flock -E -C 200 -L -S -W -P 200 -I 5000000 ; > > > done > > > > > > for a few minutes now and its not stuck and my machine didn't splat. > > > > > > Am I not doing it right? > > > > Hooray, it went *bang*.. > > I suspect a required step was to post about failure to reproduce! > It is known that the bug is both intermittent and not all machines can reproduce the problem. If it fails to reproduce, it's not necessarily a methodology error and can simply be a function of luck. -- Mel Gorman SUSE Labs