From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] KVM: IA64: enable external interrupt in vmm Date: Wed, 25 Mar 2009 13:54:03 +0200 Message-ID: <49CA1B5B.6090107@redhat.com> References: <10C63FAD690C13458F0B32BCED571F140E29AF79@pdsmsx502.ccr.corp.intel.com> <49CA0900.2040404@redhat.com> <10C63FAD690C13458F0B32BCED571F140E29AF98@pdsmsx502.ccr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: "kvm-ia64@vger.kernel.org" , "kvm@vger.kernel.org" , "Zhang, Xiantao" To: "Zhang, Yang" Return-path: Received: from mx2.redhat.com ([66.187.237.31]:36842 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751474AbZCYLyG (ORCPT ); Wed, 25 Mar 2009 07:54:06 -0400 In-Reply-To: <10C63FAD690C13458F0B32BCED571F140E29AF98@pdsmsx502.ccr.corp.intel.com> Sender: kvm-owner@vger.kernel.org List-ID: Zhang, Yang wrote: > Avi Kivity wrote: > >> I'm surprised it doesn't work this way already. What happens now if >> the guest disables interrupts and spins? >> > when vmm disables interrupt, all the external interrupts are pending > until return back to the guest. and this will degrade the i/o performance. Oh I misunderstood. When the guest us running interrupts are enabled, this was true before. The change enables interrupts in the monitor, so it improves responsiveness. -- error compiling committee.c: too many arguments to function