From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xiao Guangrong Subject: Re: [PATCH v4 5/6] KVM: MMU: combine guest pte read between walk and pte prefetch Date: Mon, 05 Jul 2010 17:09:19 +0800 Message-ID: <4C31A13F.9030408@cn.fujitsu.com> References: <4C2C9DC0.8050607@cn.fujitsu.com> <4C2C9E6C.2040803@cn.fujitsu.com> <20100702170303.GC25969@amt.cnet> <4C2F117C.2000006@cn.fujitsu.com> <4C2F2835.5060508@redhat.com> <4C2F2A0C.90704@cn.fujitsu.com> <4C2F2C5B.9020503@redhat.com> <4C2F2DBB.50904@cn.fujitsu.com> <4C2F30BD.7050702@redhat.com> <4C2F31D9.5010104@redhat.com> <4C2F3527.3020307@cn.fujitsu.com> <4C309B23.9060808@redhat.com> <4C3148FF.3030209@cn.fujitsu.com> <4C319699.9000104@redhat.com> <4C319BBB.5020408@cn.fujitsu.com> <4C31A069.6090806@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , LKML , KVM list To: Avi Kivity Return-path: In-Reply-To: <4C31A069.6090806@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org Avi Kivity wrote: > > I'm not convinced we can bypass the checks. Consider: > > > VCPU0 VCPU1 > > #PF > walk_addr > -> gpml4e0,gpdpe0,gpde0,gpte0 > > replace gpdpe0 with gpdpe1 > #PF > walk_addr > -> gpml4e0,gpdpe1,gpde1,gpte1 > fetch > -> establish hpml4e0,hpdpte1,hpde0,hpte1 > fetch > read hpdpe1 > if (present(hpdpe1)) > continue; > ... > write hpte0 using shadow hieratchy for hpte1 > Ah, i missed this case, thanks for you point it out, i'll fix it in the next version.