From mboxrd@z Thu Jan 1 00:00:00 1970 From: George Dunlap Subject: Re: some questions about the HVM Date: Tue, 30 Mar 2010 10:57:37 +0100 Message-ID: References: <20100326165856.GA4975@whitby.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Return-path: In-Reply-To: <20100326165856.GA4975@whitby.uk.xensource.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Tim Deegan Cc: "xen-devel@lists.xensource.com" , WangWei???? List-Id: xen-devel@lists.xenproject.org On Fri, Mar 26, 2010 at 5:58 PM, Tim Deegan wrote: >> I have a question about the VMEXIT in HVM. >> Does the behavior of modifying the pte of guest OS cause a vmexit. > > When using shadow pagetables, yes (more precisely, modifying a page that > is shadowed causes a VMEXIT) ...but it's complicated. We have an optimization called "out-of-sync" such that modifying PTE may not cause a VMEXIT; instead, the VMEXIT may happen when the virtual address the PTE controls is dereferenced (akin to a TLB miss). As Tim said, this is only in shadow mode, if HAP (EPT or NPT) is not present / disabled. -George