From mboxrd@z Thu Jan 1 00:00:00 1970 From: "H. Peter Anvin" Subject: Re: [PATCH] do not keep interrupt window closed by sti in real mode Date: Tue, 07 Apr 2009 23:25:48 -0700 Message-ID: <49DC436C.2050309@zytor.com> References: <1239161017-7398-1-git-send-email-glommer@redhat.com> <49DC24C2.7000608@zytor.com> <49DC39EE.1070606@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Glauber Costa , kvm@vger.kernel.org, linux-kernel@vger.kernel.org To: Avi Kivity Return-path: Received: from terminus.zytor.com ([198.137.202.10]:52574 "EHLO terminus.zytor.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751606AbZDHG0I (ORCPT ); Wed, 8 Apr 2009 02:26:08 -0400 In-Reply-To: <49DC39EE.1070606@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: Avi Kivity wrote: > > I'm guessing the problem is due to the second instruction. We don't > clear the 'blocked by interrupt shadow' flag when we emulate, which > extends interrupt shadow by one more instruction. If the instruction > sequence is 'sti hlt' we end in an inconsistent state. > Ah, and since we're in real mode, we have to emulate everything (at least on some hardware), right? So we really do need to clear the interrupt shadow bit in the interpreter... I don't see a way around that. Otherwise not just STI but MOV SS shadows will break, and in real mode MOV SS shadow is crucial. -hpa -- H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf.