Re: [Xen-devel] [PATCH RFC v13 08/20] pvh: PVH access to hypercalls

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 27.09.13 at 23:15, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx> wrote:
> On Thu, 26 Sep 2013 16:33:07 +0100
> "Jan Beulich" <JBeulich@xxxxxxxx> wrote:
> 
>> >>> On 23.09.13 at 18:49, George Dunlap <george.dunlap@xxxxxxxxxxxxx>
>> >>> wrote:
>> > --- a/xen/arch/x86/hvm/mtrr.c
>> > +++ b/xen/arch/x86/hvm/mtrr.c
>> > @@ -578,6 +578,7 @@ int32_t hvm_set_mem_pinned_cacheattr(
>> >  {
>> >      struct hvm_mem_pinned_cacheattr_range *range;
>> >  
>> > +    /* Side note: A PVH guest writes to MSR_IA32_CR_PAT natively.
>> > */ if ( !((type == PAT_TYPE_UNCACHABLE) ||
>> >             (type == PAT_TYPE_WRCOMB) ||
>> >             (type == PAT_TYPE_WRTHROUGH) ||
>> 
>> I always wondered what this comment is trying to tell the
>> reader.
> 
> Original code:
> 
>     /* Side note: A PVH guest writes to MSR_IA32_CR_PAT natively. */
>     if ( is_pvh_domain(d) )
>         return -EOPNOTSUPP;
> 
> We don't support this because the guest writes PAT natively, was 
> my intention. Either it was your comment to return EOPNOTSUPP
> or just my thought - it will return EINVAL now, I guess ok too.

That wasn't the question (albeit nevertheless good you comment
on it). I was really after understanding what "writes natively" is
supposed to mean. _If_ there's any behavioral difference to how
"normal" HVM accesses the PAT MSR, this surely requires (in the
commit message at least) a more detailed explanation. And if
there was no difference, the comment would seem rather
pointless.

Jan


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