Re: [PATCH v2 00/12] Alternate p2m: support multiple copies of host p2m

[Ed White <edmund.h.white@intel.com>]

On 06/23/2015 02:27 PM, Lengyel, Tamas wrote:
>> Testability is still a potential issue. We have offered to make our
>> internal
>> Windows test binaries available for intra-domain testing. Tamas has
>> been working on toolstack support for cross-domain testing with a slightly
>> earlier patch series, and we hope he will submit that support.
>>
> 
> Hi Ed,
> the toolstack support for externel monitoring seems to be functioning now.
> I can post it separately but IMHO it would make most sense to just append
> it to the series (if you plan to submit it again), or wait till your side
> gets merged. My branch can be found at
> https://github.com/tklengyel/xen/tree/altp2m_mine.
> 
> I've extended xen-access to exercise this new feature taking into account
> some of the current limitations. Using the altp2m_write|exec options we
> create a duplicate view of the default hostp2m, and instead of relaxing the
> mem_access permissions when we encounter a violation, we swap the view on
> the violating vCPU while also enabling MTF singlestepping. When the
> singlestep event fires, we use the response to that event to swap the view
> back to the restricted altp2m view.
> 
> # ./xen-access 6 altp2m_write
> xenaccess init
> max_gpfn = ff000
> starting altp2m_write 6
> altp2m view created with id 1
> Setting altp2m mem_access permissions.. done! Permissions set on 260171
> pages.
> Got event from Xen
> Got event from Xen
> PAGE ACCESS: rw- for GFN 272e (offset 000b98) gla 000000008272eb98 (valid:
> y; fault in gpt: n; fault with gla: y) (vcpu 0, altp2m view 1)
>     Switching back to hostp2m default view!
> Got event from Xen
> Singlestep: rip=0000000082a1a634, vcpu 0
>     Switching altp2m to view 1!
> Got event from Xen
> PAGE ACCESS: rw- for GFN 272e (offset 000b8c) gla 000000008272eb8c (valid:
> y; fault in gpt: n; fault with gla: y) (vcpu 0, altp2m view 1)
>     Switching back to hostp2m default view!
> 
> Some of the more exotic features, such as the gfn remapping, is left as
> future work for now. We definitely have plans on utilizing it in the near
> future though and it is exposed via libxc but no toolside test exercises it
> at the moment.
> 
> Cheers!
> 

That's great news! Thanks for doing this. I agree that it makes sense to
wait and see how the patch series progresses before deciding how best to
submit your changes. I'm pleased but slightly surprised that you didn't
find any show-stopping bugs, since not one line of the cross-domain code
had ever been tested.

Incidentally, the algorithm you've implemented is very similar to what our
Windows tests are currently using.

Ed