Re: Live migration with MMIO pages

[Keir Fraser <keir@xensource.com>]

On 2/11/07 09:31, "Kieran Mansley" <kmansley@solarflare.com> wrote:

>> You could do that.  I think the correct test is shadow_mode_refcounts();
>> as Keir pointed out, guests whose refcounting isn't done by the shadow
>> code will have had their pagetables validated bby the PV MMU interface.
>> 
>> The test around the log-dirty code can just be on mfn_valid().
> 
> Attached is a new patch with those changes.  Seems much more sensible to
> me that way too, and it works just as well.

I changed the final chunk test from !shadow_mode_translate() to
shadow_mode_refcounts(). I think refcounts is a better test than translate
here, and also I think your test was the wrong way round!

I also note that guarding the mark-dirty-or-zap-writable-bit with
mfn_valid() is not really correct. mfn_valid() only checks whether the mfn <
max_page. I bet this would not work if you migrate on a machine with 4GB of
RAM, as the MMIO hole will be below max_page. Really mfn_valid needs to
handle such MMIO holes, or the shadow code needs to be using a test other
than mfn_valid in many places (e.g., the function iomem_page_test() that you
added before).

 -- Keir