From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1K8wXN-00015w-7z for qemu-devel@nongnu.org; Wed, 18 Jun 2008 08:13:17 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1K8wXM-00015H-KQ for qemu-devel@nongnu.org; Wed, 18 Jun 2008 08:13:16 -0400 Received: from [199.232.76.173] (port=44564 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1K8wXM-000158-Cl for qemu-devel@nongnu.org; Wed, 18 Jun 2008 08:13:16 -0400 Received: from mail2.shareable.org ([80.68.89.115]:39789) by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1K8wXM-0006Au-6d for qemu-devel@nongnu.org; Wed, 18 Jun 2008 08:13:16 -0400 Received: from jamie by mail2.shareable.org with local (Exim 4.63) (envelope-from ) id 1K8wXJ-0000Bq-21 for qemu-devel@nongnu.org; Wed, 18 Jun 2008 13:13:13 +0100 Date: Wed, 18 Jun 2008 13:13:12 +0100 From: Jamie Lokier Subject: Re: [Qemu-devel] [RESEND PATCH v3 0/5] Add UUID support. Message-ID: <20080618121312.GA305@shareable.org> References: <20080616130135.24174.93474.stgit@gleb-debian.qumranet.com.qumranet.com> <48583502.6090803@bellard.org> <20080618060713.GD32664@minantech.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080618060713.GD32664@minantech.com> Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Gleb Natapov wrote: > > I don't like this vmport "backdoor" and I would like to disable it by > > default, so I consider it is not the solution to pass the UUID > > information to the BIOS. > > Do you have other interface in mind for host/guest communication? > VMware uses the I/O port backdoor interface. Microsoft Virtual PC > uses "Invalid Opcode" mechanism as a backdoor. > > We need this interface for other things too. For instance there are OEM > Windows versions that require specific vendor's ACPI table to be present > for installation. I have a patch that pass additional ACPI tables to BIOS > using backdoor interface since it's not practical to compile different > BIOSes for different Windows OEM versions. > > > BTW, what is the use of this UUID ? > > It is used for system management to tell two identical computers apart. > Something like MAC but more stable. In Windows WMI can be used to obtain > UUID and we have vendors that use VB scripts to configure machines > differently based on their UUIDs. Also, if you're porting an exiting Windows VM image from some other VM system, the existing image will expect certain things which are UUIDs (BIOS asset tags etc.) to have particular values. If they are not present, or the values are different, then the VM image may refuse to run and ask you to "activate" it with a phone call to Microsoft. I don't like this, but being able to port existing images makes a difference when considering whether to use a Linux VM host or a Windows VM host, to host Windows guests. (Admittedly some other things also need to be emulated to seamlessly take a Windows image from MS Virtual PC into QEMU.) ... On the other hand, VM backdoors are a way in which OSes can refuse to run if they detect they are in particular VMs, or behave differently. This could be ugly: imagine an OS which runs fine in VMware and MS Virtual PC, but refuses to run in QEMU/KVM _only_ because it detects what it's running in, and not for technical reasons. So it's important than VM backdoors are emulated as carefully as other hardware, and don't expose too much real information about the host, just the minimum required and under user control. -- Jamie