From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] KVM: Qemu: Flush i-cache after ide-dma operation in IA64 Date: Thu, 02 Apr 2009 18:59:32 +0300 Message-ID: <49D4E0E4.90606@redhat.com> References: <10C63FAD690C13458F0B32BCED571F140F98ED4B@pdsmsx502.ccr.corp.intel.com> <49D47D7F.2040904@redhat.com> <1238686878.49d4dc9e051aa@imp.free.fr> <49D4DF6A.7080108@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: "Zhang, Yang" , "kvm-ia64@vger.kernel.org" , "kvm@vger.kernel.org" , "Zhang, Xiantao" To: tgingold@free.fr Return-path: Received: from mx2.redhat.com ([66.187.237.31]:37622 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756788AbZDBP7k (ORCPT ); Thu, 2 Apr 2009 11:59:40 -0400 In-Reply-To: <49D4DF6A.7080108@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: Avi Kivity wrote: >>> >> >> It doesn't had to do it. The PCI transaction will automatically >> invalidate >> caches - but qemu doesn't emulate this (and doesn't need to do on x86). >> > > So any DMA on ia64 will flush the instruction caches?! > Or maybe, the host kernel will do it after the transaction completes? In our case the lack of zero-copy means the host is invalidating the wrong addresses (memcpy source) and leaving the real destination intact. -- error compiling committee.c: too many arguments to function