From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH v3 7/8] ARM: KVM: Handle guest faults in KVM Date: Sun, 12 Jun 2011 11:24:07 +0300 Message-ID: <4DF477A7.30300@redhat.com> References: <20110603150318.17011.82777.stgit@ubuntu> <20110603150414.17011.72525.stgit@ubuntu> <4DEB7B10.2070304@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: catalin.marinas@arm.com, android-virt@lists.cs.columbia.edu, s.raho@virtualopensystems.com, a.motakis@virtualopensystems.com, kvm@vger.kernel.org, a.costa@virtualopensystems.com To: Christoffer Dall Return-path: Received: from mx1.redhat.com ([209.132.183.28]:16379 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751897Ab1FLIYb (ORCPT ); Sun, 12 Jun 2011 04:24:31 -0400 In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On 06/11/2011 01:37 PM, Christoffer Dall wrote: > > > > Okay, this is about a zillion times simpler than x86. Congratulations. > > Well, I need to handle the I/O aborts, but it's quite simple. What > makes it much more complicated on x86? - lack of nested paging on earlier processors - 97 different paging modes - lots of extra bits bringing in wierd functionality - lots of optimizations > > > > What are your thoughts about mmu notifier support? > > For what purpose? There is no swapping on ARM, so only case that jumps > to my mind is for KSM. And I'm not quite there yet :) Really? I imaging swapping will be needed for server workloads. mmu notifiers are also useful for transparent hugepages and page migrations. I imagine these will all follow if ARM servers take off. -- error compiling committee.c: too many arguments to function