[Devel] Re: [PATCH v3 05/28] memcg: Reclaim when more than one page needed.

Glauber Costa glommer at parallels.com
Tue May 29 08:45:42 PDT 2012


On 05/29/2012 06:20 PM, Christoph Lameter wrote:
> On Tue, 29 May 2012, Christoph Lameter wrote:
>
>>>   	* unlikely to succeed so close to the limit, and we fall back
>>>   	 * to regular pages anyway in case of failure.
>>>   	 */
>>> -	if (nr_pages == 1&&  ret)
>>> +	if (nr_pages<= (PAGE_SIZE<<  PAGE_ALLOC_COSTLY_ORDER)&&  ret) {
>
> Should this not be
>
> 	 nr_pages<= 1<<  PAGE_ALLOC_COSTLY_ORDER

I myself believe you are right.

Not sure if Suleiman had anything in mind that we're not seeing when he 
wrote this code?




More information about the Devel mailing list