From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: QEMU bumping memory bug analysis Date: Mon, 8 Jun 2015 14:44:05 +0100 Message-ID: <55759C25.4070402@citrix.com> References: <20150605164354.GK29102@zion.uk.xensource.com> <20150605181014.GM29102@zion.uk.xensource.com> <20150608131000.GC29102@zion.uk.xensource.com> <20150608133220.GE29102@zion.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Stefano Stabellini , Wei Liu Cc: George Dunlap , Ian Jackson , Ian Campbell , dslutz@verizon.com, xen-devel@lists.xen.org List-Id: xen-devel@lists.xenproject.org On 08/06/15 14:38, Stefano Stabellini wrote: >> Also device-mode/$domid/state is writable by QEMU so we can't trust >> > the content as indicator either. > We can because the write happens before we unpause the guest Only when creating the domain fresh. On resume, the guest has possibly had the chance to code-inject via the qemu save format. There are many CVEs in this area, and I am not willing to be all of them are fixed. In XenServer, even loading VM state from the save file happens in the deprivilelged environment. ~Andrew