From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH][RESEND] x86 emulator: Add in/out instructions (opcodes 0xe4-0xe7, 0xec-0xef) Date: Wed, 20 Aug 2008 15:58:55 +0300 Message-ID: <48AC150F.50407@qumranet.com> References: <20080818190746.GA6296@mohd-laptop> <48ABD8FC.2070309@qumranet.com> <52d4a3890808200457y24411b2bw748d09781061dcb@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: kvm@vger.kernel.org, riel@surriel.com To: Mohammed Gamal Return-path: Received: from il.qumranet.com ([212.179.150.194]:34923 "EHLO il.qumranet.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752760AbYHTM65 (ORCPT ); Wed, 20 Aug 2008 08:58:57 -0400 In-Reply-To: <52d4a3890808200457y24411b2bw748d09781061dcb@mail.gmail.com> Sender: kvm-owner@vger.kernel.org List-ID: Mohammed Gamal wrote: > On Wed, Aug 20, 2008 at 11:42 AM, Avi Kivity wrote: > >> Mohammed Gamal wrote: >> >>> The patch adds in/out instructions to the x86 emulator. >>> >>> The instruction was encountered while running the BIOS while using >>> the invalid guest state emulation patch. >>> >>> >> We still don't have a testcase for this. >> >> > To be clear, do you want an emulator test case (in > user/test/x86/emulator.c) for every instruction added to the emulator? > Yes. We've been bitten by rarely-used instructions being broken. I've now added a simple harness to test these instructions. It works by jumping into big real mode, forcing kvm to emulate (if emulate_invalid_guest_state=1) and returning. Please review it as well as the few fixes I've made in kvm to support it. It's in the 'realmode' branch of both kvm.git and kvm-userspace.git. Once it passes regression testing, I'll merge it into master. -- error compiling committee.c: too many arguments to function