From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 6 Mar 2002 15:21:52 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 6 Mar 2002 15:21:41 -0500 Received: from ns01.netrox.net ([64.118.231.130]:21169 "EHLO smtp01.netrox.net") by vger.kernel.org with ESMTP id ; Wed, 6 Mar 2002 15:21:28 -0500 Subject: Re: [PATCH] 2.5: preemptive kernel on UP From: Robert Love To: george anzinger Cc: linux-kernel@vger.kernel.org In-Reply-To: <3C866B56.AE2F63BD@mvista.com> In-Reply-To: <1015287099.865.3.camel@phantasy> <3C866B56.AE2F63BD@mvista.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/1.0.2 Date: 06 Mar 2002 15:21:20 -0500 Message-Id: <1015446092.1482.6.camel@icbm> Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 2002-03-06 at 14:17, george anzinger wrote: > With out a lot of looking, wouldn't it be easier to just change fork? Don't we need to unlock the rq regardless of what fork does? Also, the alternative would be to set preempt_count to 0 if CONFIG_PREEMPT and 1 if CONFIG_PREEMPT && CONFIG_SMP. I figured Linus would balk at that sort #if/#else in do_fork. And, even then, is it safe to start the task with a preempt_count of 0? Robert Love