From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1767351AbXDEVuo (ORCPT ); Thu, 5 Apr 2007 17:50:44 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1767352AbXDEVuo (ORCPT ); Thu, 5 Apr 2007 17:50:44 -0400 Received: from e34.co.us.ibm.com ([32.97.110.152]:49540 "EHLO e34.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1767351AbXDEVun (ORCPT ); Thu, 5 Apr 2007 17:50:43 -0400 Subject: Re: [PATCH] timekeeping: drop irq-context clocksource polling From: john stultz To: Andrew Morton Cc: Daniel Walker , linux-kernel@vger.kernel.org, mingo@elte.hu In-Reply-To: <20070405143641.a751f43e.akpm@linux-foundation.org> References: <20070405210316.785839431@mvista.com> <1175808319.28526.26.camel@localhost> <20070405143641.a751f43e.akpm@linux-foundation.org> Content-Type: text/plain Date: Thu, 05 Apr 2007 14:50:34 -0700 Message-Id: <1175809834.28526.34.camel@localhost> Mime-Version: 1.0 X-Mailer: Evolution 2.8.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 2007-04-05 at 14:36 -0700, Andrew Morton wrote: > On Thu, 05 Apr 2007 14:25:19 -0700 > john stultz wrote: > > > On Thu, 2007-04-05 at 14:03 -0700, Daniel Walker wrote: > > > Before this change the timekeeping code would poll the clocksource > > > list every interrupt. This changes that so the clocksource list is > > > only checked when there has been and update, and no longer checks > > > in interrupt context. > > > > > > This also has a few small space and line cleanups. > > > > > > Boot tested on i386, compile tested on x86_64 .. However, I couldn't > > > find a !GENERIC_TIME that compiled without this change so it's untested.. > > > > > > Signed-Off-By: Daniel Walker > > > > Err.. I think you need to be holding a write on the xtime_lock (as is > > done before calling update_wall_time()) when changing the clocksource. > > The patch does add the appropriate locking to change_clocksource(), > doesn't it? Yep. Sorry for the confusion, reading too hastily. > It looks like a good change to me - we avoid taking the kernel-wide > clocksource_lock every tick? Agreed. Acked-by: John Stultz -john