From mboxrd@z Thu Jan 1 00:00:00 1970 From: Glauber Costa Subject: Re: [PATCH 8/9] coalesce mmio regions with an explicit call Date: Wed, 24 Sep 2008 18:59:14 -0300 Message-ID: <20080924215914.GA5049@poweredge.glommer> References: <1221840506-22996-1-git-send-email-glommer@redhat.com> <1221840506-22996-9-git-send-email-glommer@redhat.com> <48D54370.7090709@redhat.com> <20080922135645.GF3618@poweredge.glommer> <48D89AEC.5090102@redhat.com> <20080923162233.GD3583@poweredge.glommer> <48DA2022.8020807@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: kvm@vger.kernel.org, aliguori@us.ibm.com To: Avi Kivity Return-path: Received: from mx2.redhat.com ([66.187.237.31]:57986 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752804AbYIXWAg (ORCPT ); Wed, 24 Sep 2008 18:00:36 -0400 Content-Disposition: inline In-Reply-To: <48DA2022.8020807@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Wed, Sep 24, 2008 at 02:10:26PM +0300, Avi Kivity wrote: > Glauber Costa wrote: >>> You can't coalesce the registers which trigger device action. You'll >>> destroy latency and/or functionality. >>> >> >> which kills the goal of getting rid of explicit kvm code. >> >> > > It's a fact that coalescing helps kvm but not qemu. > >> So maybe the solution here is to add calls in qemu to a memory >> coalescing function that in the raw qemu / kqemu case just don't >> do anything? >> > > That's just word games. s/kvm/qemu/ won't change the fact that this is > a kvm specific hook. Any ideas about what's up for the other hypervisors that may (we hope) be integrated in the future? Xen?