Re: [Xen-devel] [PATCH v7 07/10] xen/common: move the memory_mapping DOMCTL hypercall to common code

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 10.05.14 at 02:26, <avanzini.arianna@xxxxxxxxx> wrote:
> On 05/07/2014 01:09 PM, Ian Campbell wrote:
>> On Tue, 2014-05-06 at 09:40 +0100, Jan Beulich wrote:
>>>>>> On 05.05.14 at 17:54, <avanzini.arianna@xxxxxxxxx> wrote:
>>>>     v7:
>>>>         - Change the name of tmp_rc to rc. Also, do not uselessly 
>>>> initialize it.
>>>>         - ifdef out the invocation of memory_type_changed() to be called 
>>>> only
>>>>           if the architecture is x86 instead of adding an useless empty 
>>>> stub
>>>>           for ARM.
>>>
>>> Was this requested by one of the ARM maintainers?
>> 
>> I don't think it was me.
>> 
>>> Do memory types not exist at all on ARM?
>> 
>> This is types in "MTRR" sense rather than p2m type sense I think? These
>> are part of the PT mappings, so I would expect something like this to
>> probably be needed for any IOMMU stuff at some point.
>> 
>>>  I would have expected the function to be
>>> empty only until someone would get to implement it properly...
>> 
>> Me too.
>> 
> 
> Yes, in the previous version I had defined an empty stub for a
> memory_type_changed() function for ARM. While trying to move its prototype 
> to a
> common header according to Jan Beulich's suggestion, however, I have noticed
> that the invocation of that same function, performed from the common code
> handling the iomem_permission domctl, is ifdef'd out to be compiled only on 
> an
> x86 machine. I therefore thought to do the same for the memory_mapping 
> domctl.
> 
> Do you prefer that an empty stub is added? If so, would you like also the
> handling of memory_type_changed() in iomem_permission to be modified in this
> patchset?

I think that would make sense.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel