From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sun, 23 Sep 2001 14:28:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sun, 23 Sep 2001 14:27:56 -0400 Received: from host154.207-175-42.redhat.com ([207.175.42.154]:33789 "EHLO lacrosse.corp.redhat.com") by vger.kernel.org with ESMTP id ; Sun, 23 Sep 2001 14:27:44 -0400 Date: Sun, 23 Sep 2001 14:28:09 -0400 From: Benjamin LaHaise To: Manfred Spraul Cc: Andrea Arcangeli , linux-kernel@vger.kernel.org Subject: Re: 2.4.10pre13aa1 Message-ID: <20010923142809.A11346@redhat.com> In-Reply-To: <3BADAF6A.8090400@colorfullife.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <3BADAF6A.8090400@colorfullife.com>; from manfred@colorfullife.com on Sun, Sep 23, 2001 at 11:46:18AM +0200 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sun, Sep 23, 2001 at 11:46:18AM +0200, Manfred Spraul wrote: > I don't think that this is a valid argument: > you are testing on i386 and make design decisions for the architecture > independant part. The hook is only needed for x86-like architectures, as all of the risc like cpus perform the dirty state modification in the page fault handler while holding the page_table_lock. > I'd prefer ptep_get_and_clear_and_flush(), then the arch part can do > what's needed to get the final pte value. (if a single page is modified, > otherwise the arch can define a suitable mmu_gather) mmu_gather is the hook. ia64 for instance can create a gather that uses the hardware tlb flush instruction to shootdown the entry across all cpus and eliminates the race. -ben