Re: [PATCH 1 of 7] x86: add _PAGE_IOMAP pte flag for IO mappings

[Avi Kivity <avi@qumranet.com>]

Jeremy Fitzhardinge wrote:
> Avi Kivity wrote:
>   
>> Jeremy Fitzhardinge wrote:
>>     
>>> Use one of the software-defined PTE bits to indicate that a mapping is
>>> intended for an IO address.  On native hardware this is irrelevent,
>>> since a physical address is a physical address.  But in a virtual
>>> environment, physical addresses are also virtualized, so there needs
>>> to be some way to distinguish between pseudo-physical addresses and
>>> actual hardware addresses; _PAGE_IOMAP indicates this intent.
>>>
>>> By default, __supported_pte_mask masks out _PAGE_IOMAP, so it doesn't
>>> even appear in the final pagetable.
>>>   
>>>       
>> Could PTE_SPECIAL, added for get_user_pages_really_fast(), be reused
>> for this?
>>
>>     
>
> I'm not sure; I still don't really understand how _PAGE_SPECIAL gets
> used, other than being user-mode mapping only.  But in principle,
> _PAGE_IOMAP could be set on both kernel and user mappings (if you direct
> map a device into a process address space), so I think they would
> conflict then?
>
>   

It's a "don't refcount me" flag, which is not sematically the same as 
I/O, but may be close enough.

> Is _PAGE_IOMAP at all useful for device passthrough in kvm?
>   

No.  We don't care if a page is an I/O page or RAM (other than 
refcounting correctness and page attributes, which are handled by other 
means).

-- 
error compiling committee.c: too many arguments to function