Re: [Xen-devel] [PATCH 1/8] x86/p2m: make p2m_change_type() behavior match p2m_change_entry_type_global()'s

[Andres Lagar-Cavilla <andreslc@xxxxxxxxxxxxxx>]

> At 15:31 +0000 on 26 Mar (1395844319), Jan Beulich wrote:
>> - don't reset access permissions
>> - don't shatter super pages
>> 
>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> Reviewed-by: Tim Deegan <tim@xxxxxxx>
> 
>> ---
>> Should p2m_change_type() also behave consistently?
> 
> Well, I suppose all these operations ought to match.  It swould be
> nice to hear from anyone who's actualy using the access types whether
> it owuld be better to reset on type change or not.  My instinct is to
> leave the access permission alone.

As a lightweight user of access, I foresee reseting access bits resulting in 
convoluted control logic (must check which pfns changed type to get them back 
to the access that I wanted).

As long as access and type are independent, their life cycles should also be. 
IMHO

Andres
> 
> Cheers,
> 
> Tim.
> 


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