Re: [PATCH 1/1 v2] KVM: MMU: Use ptep_user for cmpxchg_gpte()

[Avi Kivity <avi@redhat.com>]

On 05/04/2011 05:00 PM, Takuya Yoshikawa wrote:
> >
> >  We do get_user() in read_gpte().  That is equivalent to
> >  get_user_pages().  So we already broke that layer of abstraction.
>
> At first, I broke hva_to_pfn() into two functions:
>    hva_to_page
>    hva_to_pfn
> and used the former to get the page.

Ouch, what a complicated function.

> But after making that patch, I thought it might be a bit extra to do such
> things in the function which is doing low level page manipulations like
> kmapping.
>
> Actually, we are already assuming that the page returned by gfn_to_page is
> always a usual page which contains gptes without extra checks.

When gfn_to_page() returns an error, it is actually a real page in host 
memory that can be scribbled on.  So no further checks are needed.

> Which way do you like the best?

I think it should work fine as is.  The question is whether we're doing 
a layering violation here (but in any case, that was introduced by 
ptep_user, not this patch).

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