From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265494AbUATNdl (ORCPT ); Tue, 20 Jan 2004 08:33:41 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S265500AbUATNdl (ORCPT ); Tue, 20 Jan 2004 08:33:41 -0500 Received: from grex.cyberspace.org ([216.93.104.34]:15890 "HELO grex.cyberspace.org") by vger.kernel.org with SMTP id S265494AbUATNdk (ORCPT ); Tue, 20 Jan 2004 08:33:40 -0500 From: Andrew Halliwell Message-Id: <200401201334.IAA01060@grex.cyberspace.org> Subject: AMD64/3ware/8Gig RAM and iommu=fullflush To: linux-kernel@vger.kernel.org Date: Tue, 20 Jan 2004 08:34:17 -0500 (EST) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I've been trying to find information about the iommu=fullflush kernel switch and was wondering if anyone had any information on what the actual performance hit is, forcing it to write out its disk cache every write, or whatever it does? It must slow things down, but I was wondering how much. Also, as this was admittedly a kludge, how far off is a more permanent solution to the 3ware 4+gig/32bit only on a 64bit system problem? Thanks