From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264641AbUEaOzz (ORCPT ); Mon, 31 May 2004 10:55:55 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264659AbUEaOzz (ORCPT ); Mon, 31 May 2004 10:55:55 -0400 Received: from lvs00-fl-n04.valueweb.net ([216.219.253.138]:7822 "EHLO ams004.ftl.affinity.com") by vger.kernel.org with ESMTP id S264641AbUEaOzy (ORCPT ); Mon, 31 May 2004 10:55:54 -0400 Message-ID: <40BB474F.9050503@coyotegulch.com> Date: Mon, 31 May 2004 10:55:11 -0400 From: Scott Robert Ladd User-Agent: Mozilla Thunderbird 0.6 (X11/20040522) X-Accept-Language: en-us, en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Re: How to use floating point in a module? References: <200405310152.i4V1qNk03732@mailout.despammed.com> In-Reply-To: <200405310152.i4V1qNk03732@mailout.despammed.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org ndiamond@despammed.com wrote: > And does anyone know a really safe method? I recently built a 64-bit fixed-point library for a client, which they used in a device driver. I'd suggest following a similar pattern, thus avoiding the entire issues of "floating-point" in the kernel. ..Scott -- Scott Robert Ladd Coyote Gulch Productions (http://www.coyotegulch.com) Software Invention for High-Performance Computing