From mboxrd@z Thu Jan 1 00:00:00 1970 From: George Dunlap Subject: Re: Please review the key 4.4 release docs Date: Fri, 14 Mar 2014 17:43:03 +0000 Message-ID: <53233FA7.5090304@eu.citrix.com> References: <55E78A57290FB64FA0D3CF672F9F3DA21A7000@SJCPEX01CL03.citrite.net> <531D8BE5.2070607@eu.citrix.com> <53233B59.8080208@eu.citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <53233B59.8080208@eu.citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: "Zhang, Yang Z" , Russell Pavlicek , "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 03/14/2014 05:24 PM, George Dunlap wrote: > On 03/11/2014 01:23 AM, Zhang, Yang Z wrote: >> George Dunlap wrote on 2014-03-10: >>> On 03/10/2014 04:27 AM, Zhang, Yang Z wrote: >>>> Russell Pavlicek wrote on 2014-03-10: >>>>> I have done my best to compile the various key release documents >>>>> for the 4.4 release. >>>>> >>>>> All have hyperlinks from the download page: >>>>> http://www.xenproject.org/downloads/xen-archives/supported-xen-44-s >>>>> er >>>>> i >>>>> es/ xen-440.html >>>> I thought that we have agreed that we will move the nested >>>> virtualization >>> from experimental to 1.0 or something else in Xen 4.4 release note. >>> But it seems it still in experimental state and there is no mention of >>> it in the release note. >>> >>> Well we discussed it, but there were too many things still missing to >>> call it a properly supported feature: in particular, doubts about how >>> well shadow-on-hap would work, which would be a potential security >> Actually, I'd like to know all potential nested issues and I will try >> to solve it if possible. But the problem is that I am not clear about >> those issues that you guys mentioned. Is there any thread talk about >> them? > > I think I was thinking along the lines of what I wrote in this thread: > > Msg-ID <52E28EFB.3020008@eu.citrix.com> > > It seems there the minimum thing for a "1.0" release is that an L1 > admin must not be able to do anything to affect an L0; and thinks at > the moment likely to do so are enabling PoD (and probably also paging) > for L2 guests. It's OK if enabling PoD crashes the *L1* hypervisor > (because that's under the L1 admin's control); but it must not be > allowed to crash / DoS the L0 hypervisor. BTW, I take it that you're only supporting 64-bit hypervisors at the moment? A few weeks ago I tried Win7's "XP Compatibility Mode" with a 32-bit guest and it just gave me weird errors when trying to start an L2 guest. 64-bit worked just fine, however. I don't think 32-bit L1's are particularly important, but it's officially not supported, that should probably be documented on the wiki page. -George