[ewg] Re: [PATCH v3] mlx4_ib: Optimize hugetlab pages support

[Yossi Etigin]

Eli Cohen wrote:
> On Thu, Apr 02, 2009 at 11:31:42PM +0300, Yossi Etigin wrote:
>> Hi Eli,
>>
>> I've placed a printk in the new hugetlb function to print n and j.
>> While running the mckey test (attached in bugzilla), I got j=0, n=1.
>>
>> Why do you say that the number of pages must cover HUGE_PAGES?
>> In ib_umem_get, hugetlb is set to 0 only if any of the pages is  
>> not-hugetlb - otherwise it's 1. Am I missing something?
>>
> 
> Only if ALL of the registered area is huge pages, then umem->hugetlb
> will remain 1; otherwise it is cleared and we don't execute
> handle_hugetlb_user_mr(). In case we do execute
> handle_hugetlb_user_mr(), I expect the number of PAGE_SIZEed pages to
> fill full huge pages.

I don't understand - if all area is huge pages, it does not mean that
it fills full huge pages - I can have just 4096 bytes in huge page memory 
and umem->hugetlb will remain 1, right?