Re: [PATCH] nested vmx: Fix the booting of L2 PAE guest

[George Dunlap <george.dunlap@eu.citrix.com>]

On 27/06/13 09:40, Jan Beulich wrote:
>>>> On 27.06.13 at 03:14, "Xu, Dongxiao" <dongxiao.xu@intel.com> wrote:
>> Hi stakeholders,
>>
>> I saw the patch is not merged yet. Do you have any other comment about this
>> patch? I think it is a critical fix for 4.3 release in nested virtualization
>> side.
> Irrespective of Keir's ack I was hoping for an ack from one of the
> VMX maintainers. Even more so as they are, just like you, working
> for Intel I think it would be appropriate for you to get in touch
> with them to fulfill their maintainer task here. In fact you should
> have Cc-ed them with your initial patch submission.
>
> Independently of that, you should have also Cc-ed George if you
> want this to go in for 4.3. In the absence of this, I had simply put
> this on my post-4.3 queue...

Well normally I think I would have said "no" to this change.  You guys 
haven't done a very good job of engaging with the release process we've 
been trying to develop -- you didn't report this bug to me so that I 
could track it and make informed decisions regarding the release.

However, I still consider nested VMX as "experimental" -- the fact that 
in the current tree, Win7 won't boot on Xen-on-Xen kind of confirms that 
status to me. :-)  Since it is experimental, like the ARM port, it has a 
slightly different release criteria: basically, as long as you don't 
touch code in the non-experimental path, you can take your own risks 
regarding breaking functionality.

Given that this patch only touches code inside vvmx.c, I'm assuming that 
this code *cannot* be touched by anyone who is *not* running in nested 
virt mode.

If that's correct, I'm inclined to say we can accept it.  But I'd be 
interested in hearing other people's opinions.

  -George