From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751583Ab2GSM5J (ORCPT ); Thu, 19 Jul 2012 08:57:09 -0400 Received: from 1wt.eu ([62.212.114.60]:5039 "EHLO 1wt.eu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751011Ab2GSM5G (ORCPT ); Thu, 19 Jul 2012 08:57:06 -0400 Date: Thu, 19 Jul 2012 14:56:53 +0200 From: Willy Tarreau To: John Stultz Cc: stable@vger.kernel.org, Prarit Bhargava , Thomas Gleixner , Linux Kernel Subject: Re: [PATCH 00/11] 2.6.32-stable: Fix for leapsecond deadlock & hrtimer/futex issue Message-ID: <20120719125653.GK16208@1wt.eu> References: <1342562735-55870-1-git-send-email-johnstul@us.ibm.com> <50060496.2090909@us.ibm.com> <20120719095311.GE16208@1wt.eu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20120719095311.GE16208@1wt.eu> User-Agent: Mutt/1.4.2.3i Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Jul 19, 2012 at 11:53:11AM +0200, Willy Tarreau wrote: > On Tue, Jul 17, 2012 at 05:34:30PM -0700, John Stultz wrote: > > On 07/17/2012 03:05 PM, John Stultz wrote: > > >This is the last backport I have ready at this time. I realize there's > > >still the 2.6.27 tree, but if I recall correctly I had some trouble > > >getting that booted in my test environment. I can take a swing at > > >backporting it, but I may need a hand with testing. > > > > And if anyone is interested in helping test, I've got a first pass > > (builds & mostly boots) of the 2.6.27-stable backports here: > > git://git.linaro.org/people/jstultz/linux.git leapfix/2.6.27.62 > > > > http://git.linaro.org/gitweb?p=people/jstultz/linux.git;a=shortlog;h=refs/heads/leapfix/2.6.27.62 > > > > Unfortunately I run into udev issues trying to boot this on my more > > recent test environment. > > John, don't waste your time trying to boot older kernels, I'm well > equipped for this task and it's much easier for me, so let's split > the work efficiently :-) OK so first, I managed to crash latest 2.6.32.x in less than one minute using leapcrash and I cannot crash it anymore after an hour with your patches applied. I'll try to develop a small futex-based PoC to test the other reproducer, then I'll duplicate the tests on 2.6.27. Looks good so far. Thanks, Willy