From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1162473Ab2GLW5p (ORCPT ); Thu, 12 Jul 2012 18:57:45 -0400 Received: from cantor2.suse.de ([195.135.220.15]:41629 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1162095Ab2GLWnl (ORCPT ); Thu, 12 Jul 2012 18:43:41 -0400 Date: Fri, 13 Jul 2012 00:43:34 +0200 From: Jiri Bohac To: John Stultz Cc: Linux Kernel , Ingo Molnar , Peter Zijlstra , Prarit Bhargava , Thomas Gleixner , stable@vger.kernel.org Subject: Re: [PATCH 0/6] Fix for leapsecond caused hrtimer/futex issue (updated) Message-ID: <20120712224334.GA32620@midget.suse.cz> References: <1341960205-56738-1-git-send-email-johnstul@us.ibm.com> <4FFCB287.80701@us.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4FFCB287.80701@us.ibm.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 Tue, Jul 10, 2012 at 03:53:59PM -0700, John Stultz wrote: > On 07/10/2012 03:43 PM, John Stultz wrote: > >Over the weekend, Thomas got a chance to review the leap second fix > >in more detail and had a few additional changes he wanted to make > >to improve performance as well as style. > > > >So this iteration includes his modifications. > > > >Once merged, I'll be working to get the backports finished as quickly > >as I can and sent to -stable. looking at the proposed 2.6.32.y stable patch at: http://git.linaro.org/gitweb?p=people/jstultz/linux.git;a=commitdiff;h=18d208632bf17aed56c581b882868b2be44be71e;hp=6d224606bb8eec78027522d6dd5abfea8108c41a Is this the final version you are about to send to -stable? In 2.6.32 timekeeping_leap_insert() is not called from the timer interrupt, but from the leap_timer hrtimer. I think the new clock_was_set_timer will thus not be called by irq_exit() because TIMER_SOFTIRQ has not been raised. Unless TIMER_SOFTIRQ is raised, clock_was_set() will not be called until the next periodic timer interrupt, correct? Wouldn't the original schedule_work() approach work better for 2.6.32? Or do you plan backporting the most recent version to 2.6.32? Thanks, -- Jiri Bohac SUSE Labs, SUSE CZ