Re: [PATCH v4 4/9] xen: introduce XEN_DOMCTL_devour

[Julien Grall <julien.grall@linaro.org>]

On 04/12/14 15:12, Vitaly Kuznetsov wrote:
> Julien Grall <julien.grall@linaro.org> writes:
> 
>> Hi Vitaly,
>>
>> On 03/12/2014 17:16, Vitaly Kuznetsov wrote:
>>> New operation sets the 'recipient' domain which will recieve all
>>
>> s/recieve/receive/
>>
>>> memory pages from a particular domain and kills the original domain.
>>>
>>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>>> ---
>>> @@ -1764,13 +1765,32 @@ void free_domheap_pages(struct page_info *pg, unsigned int order)
>>
>> [..]
>>
>>> +        else
>>> +        {
>>> +            mfn = page_to_mfn(pg);
>>> +            gmfn = mfn_to_gmfn(d, mfn);
>>> +
>>> +            page_set_owner(pg, NULL);
>>> +            if ( assign_pages(d->recipient, pg, order, 0) )
>>> +                /* assign_pages reports the error by itself */
>>> +                goto out;
>>> +
>>> +            if ( guest_physmap_add_page(d->recipient, gmfn, mfn, order) )
>>
>> On ARM, mfn_to_gmfn will always return the mfn. This would result to
>> add a 1:1 mapping in the recipient domain.
>>
>> But ... only DOM0 has its memory mapped 1:1. So this code may blow up
>> the P2M of the recipient domain.
> 
> I know almost nothing about ARM so please bear with me. So, for a guest
> domain the mapping is not 1:1 (so guest sees different addresses) but
> mfn_to_gmfn() doesn't returs these addresses?

Yes. I guess it's because this macro is not really used (see why below).
Ian, Stefano: any idea why mfn_to_gfmn is not correctly implemented on ARM?

> I was under an impression
> it is not x86-specific and I can see its usage in e.g. getdomaininfo(),
> memory_exchange(),..

I can find two places in the common code:
	- getdomaininfo: The return is obviously buggy. Though, it doesn't seem
to be used in the toolstack side for ARM
	- memory_exchange: AFAIK this is not supported right now.

> How can one figure out the mapping then?

AFAIK, there is no way on ARM to get a GMFN from an MFN in Xen.

Maybe we should implement it correctly mfn_to_gmfn on ARM? Ian, Stefano,
any though?

> Anyway, what I want to do here is: when this page is freed I want to
> reassign it to our newly-created guest at the exact same address it was
> mapped in the original domain. 

> BTW, what's the current state of affairs with kexec and ARM guest?

AFAIK, nobody has worked on it for Xen ARM. I don't even know the status
for Kexec on ARM.

> I
> suppose we should have similar problems: vcpu_info, event channels,
> .. 

ARM guest is very similar to PVH guest. All the problems you will fixed
now means less work when we will add support for ARM ;).

Regards,

-- 
Julien Grall