[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v2 3/3] mm: MEMF_node should handle changes in nodeid_t size

>>> On 25.02.15 at 14:30, <boris.ostrovsky@xxxxxxxxxx> wrote:
> On 02/25/2015 05:35 AM, Jan Beulich wrote:
>>>>> On 24.02.15 at 20:11, <boris.ostrovsky@xxxxxxxxxx> wrote:
>>> @@ -121,10 +123,12 @@ struct npfec {
>>>   #define _MEMF_exact_node  4
>>>   #define  MEMF_exact_node  (1U<<_MEMF_exact_node)
>>>   #define _MEMF_node        8
>>> -#define  MEMF_node(n)     ((((n)+1)&0xff)<<_MEMF_node)
>>> +#define  MEMF_node(n)     ((((n)+1) & MEMF_node_mask) << _MEMF_node)
>>>   #define _MEMF_bits        24
>>>   #define  MEMF_bits(n)     ((n)<<_MEMF_bits)
>>> +#define MEMF2NODE(memflags) (MASK_EXTR(memflags, MEMF_node_mask) - 1)
>> As this is being used just once, I don't really see a need for it to be
>> exposed globally. Perhaps do away with the macro, but at the very
>> least confine it to page_alloc.c.
> I intentionally put it here: we have macro to encode ('put', so to 
> speak) nodeID in memflags in this file (by adding one to it) and so I 
> felt that we don't need to expose this encoding outside of this file by 
> providing a "get" macro.

Hmm, don't know. The encoding one obviously is needed in various
places, while the decoding one is not supposed to be needed
elsewhere. But if you strongly feel so, I can live with that macro
staying here.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.