From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265800AbUIWAff (ORCPT ); Wed, 22 Sep 2004 20:35:35 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266009AbUIWAff (ORCPT ); Wed, 22 Sep 2004 20:35:35 -0400 Received: from rproxy.gmail.com ([64.233.170.205]:37394 "EHLO mproxy.gmail.com") by vger.kernel.org with ESMTP id S265800AbUIWAfb (ORCPT ); Wed, 22 Sep 2004 20:35:31 -0400 Message-ID: <516c87e80409221735b1b6a92@mail.gmail.com> Date: Thu, 23 Sep 2004 08:35:31 +0800 From: Long Pu Reply-To: Long Pu To: linux-kernel@vger.kernel.org Subject: Why is clear_user_highpage necessary in do_anonymous_page()? In-Reply-To: <516c87e8040922022973701fff@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit References: <516c87e8040922022973701fff@mail.gmail.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I wonder why clear_user_highpage is necessary if we don't concern security. When we write to a anonymous page the very first time, we don't care whatever the original contents of that page is. And clear_user_highpage() takes a lot of cpu time under certain workload such as compiling kernel. Is there any reason I don't know to reserve it?