From: Naoya Horiguchi on
On Fri, Jul 02, 2010 at 11:08:54AM +0200, Andi Kleen wrote:
> On Fri, Jul 02, 2010 at 02:47:22PM +0900, Naoya Horiguchi wrote:
> > We can't use existing hugepage allocation functions to allocate hugepage
> > for page migration, because page migration can happen asynchronously with
> > the running processes and page migration users should call the allocation
> > function with physical addresses (not virtual addresses) as arguments.
> I looked through this patch and didn't see anything bad. Some more
> eyes familiar with hugepages would be good though.


> Since there are now so many different allocation functions some
> comments on when they should be used may be useful too

OK. How about this?

+ * This allocation function is useful in the context where vma is irrelevant.
+ * E.g. soft-offlining uses this function because it only cares physical
+ * address of error page.
+ */
+struct page *alloc_huge_page_node(struct hstate *h, int nid)

BTW, I don't like this function name very much.
Since the most significant difference of this function to alloc_huge_page()
is lack of vma argument, so I'm going to change the name to
alloc_huge_page_no_vma_node() in the next version if it is no problem.

Or, since the postfix like "_no_vma" is verbose, I think it might be
a good idea to rename present alloc_huge_page() to alloc_huge_page_vma().
Is this worthwhile?

Naoya Horiguchi
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)
More majordomo info at
Please read the FAQ at