Re: [PATCH] x86: add an extra check when validating a huge pv L2 entry

[Michael J Coss <mjcoss@alcatel-lucent.com>]

Jan Beulich wrote:
>>>> Keir Fraser <keir.fraser@eu.citrix.com> 09.09.09 17:35 >>>
>>>>         
>> On 09/09/2009 16:19, "Jan Beulich" <JBeulich@novell.com> wrote:
>>
>>     
>>> -            rc = get_data_page(mfn_to_page(m), d, writeable);
>>> +            if ( mfn_valid(m) )
>>> +                rc = get_data_page(mfn_to_page(m), d, writeable);
>>> +            else
>>> +                rc = -EINVAL;
>>>       
>> 'else rc = 0' would be more like it, eh?
>>     
>
> Oh, yes - I got confused (again) by the inconsistencies in what return
> values mean for the various functions.
>
> Jan
>   
I don't think that that's correct.  The else clause is for the 
mfn_valid(m) failure case, and it seems appropriate to set the rc to 
some non-zero value.

---Michael J Coss