From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965086AbVHPDdl (ORCPT ); Mon, 15 Aug 2005 23:33:41 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S965087AbVHPDdl (ORCPT ); Mon, 15 Aug 2005 23:33:41 -0400 Received: from mailout1.vmware.com ([65.113.40.130]:14340 "EHLO mailout1.vmware.com") by vger.kernel.org with ESMTP id S965086AbVHPDdk (ORCPT ); Mon, 15 Aug 2005 23:33:40 -0400 Message-ID: <43015E8E.5020105@vmware.com> Date: Mon, 15 Aug 2005 20:33:34 -0700 From: Zachary Amsden User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.2) Gecko/20040803 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Brian Gerst Cc: akpm@osdl.org, chrisw@osdl.org, linux-kernel@vger.kernel.org, virtualization@lists.osdl.org Subject: Re: [PATCH 0/6] i386 virtualization patches, Set 3 References: <200508152258.j7FMw9p8005295@zach-dev.vmware.com> <43015894.1090307@didntduck.org> In-Reply-To: <43015894.1090307@didntduck.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 16 Aug 2005 03:33:35.0164 (UTC) FILETIME=[485697C0:01C5A213] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Brian Gerst wrote: > If you really want to test the math emu code, you can hack check_x87 > in head.S to always leave the fpu disabled. Then you can test it on > any cpu, not just a 386. That is a good idea, and while a valid point, it actually still requires writing the code to actually test the FPU, specifically, using weird prefixes, LDT based segments, and other oddities that don't get generated from "normal" C code. I'm pretty sure the existing code works for the 99% cases or else it wouldn't have gotten there in the first place. But testing the corner cases here is even nastier than testing the LDT corner cases - you would basically need to write a lot of hand coded i387 assembler. Perhaps such a test might exist, but in all honesty, without a comprehensive test, it is simply far too easy to introduce a bug here, and far too likely it will either not be noticed until it has caused someone a possibly undetected numerical error, or I'm just wasting my time because noone is using this code anyways. Fortunately, the Hubble telescope has been upgraded to a 486 ;) Zach