Re: [PATCH 1/2] x86/mm: Use mfn_t for new_guest_cr3()

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 30/08/17 16:45, Jan Beulich wrote:
>>>> On 30.08.17 at 14:19, <andrew.cooper3@citrix.com> wrote:
>> @@ -2772,23 +2772,23 @@ int vcpu_destroy_pagetables(struct vcpu *v)
>>      return rc != -EINTR ? rc : -ERESTART;
>>  }
>>  
>> -int new_guest_cr3(unsigned long mfn)
>> +int new_guest_cr3(mfn_t mfn)
>>  {
>>      struct vcpu *curr = current;
>>      struct domain *d = curr->domain;
>>      int rc;
>> -    unsigned long old_base_mfn;
>> +    mfn_t old_base_mfn;
>>  
>>      if ( is_pv_32bit_domain(d) )
>>      {
>> -        unsigned long gt_mfn = pagetable_get_pfn(curr->arch.guest_table);
>> -        l4_pgentry_t *pl4e = map_domain_page(_mfn(gt_mfn));
>> +        mfn_t mmfn = pagetable_get_mfn(curr->arch.guest_table);
>> +        l4_pgentry_t *pl4e = map_domain_page(mmfn);
> What was wrong with "gt_mfn" for "guest table MFN"? I can't help
> thinking mmfn is a typo, where you've hit the m key one too many
> times. What is that first m supposed to stand for?

mmfn is the shadow code nomenclature for the monitor mfn.  I can move
back to gt_mfn if you think thats clearer.

~Andrew

>
> Other than this the patch is certainly fine.
>
> Jan
>


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