Messages in this thread | | | Subject | Re: [PATCH 6/8] mm, hugetlb: remove vma_has_reserves | From | Davidlohr Bueso <> | Date | Wed, 29 Jan 2014 11:24:16 -0800 |
| |
On Mon, 2014-01-27 at 16:04 -0500, Naoya Horiguchi wrote: > On Sun, Jan 26, 2014 at 07:52:24PM -0800, Davidlohr Bueso wrote: > > From: Joonsoo Kim <iamjoonsoo.kim@lge.com> > > > > vma_has_reserves() can be substituted by using return value of > > vma_needs_reservation(). If chg returned by vma_needs_reservation() > > is 0, it means that vma has reserves. Otherwise, it means that vma don't > > have reserves and need a hugepage outside of reserve pool. This definition > > is perfectly same as vma_has_reserves(), so remove vma_has_reserves(). > > I'm concerned that this patch doesn't work when VM_NORESERVE is set. > vma_needs_reservation() doesn't check VM_NORESERVE and this patch changes > dequeue_huge_page_vma() not to check it. So no one seems to check it any more.
Good catch. I agree, this is new behavior and quite frankly not worth changing just for a cleanup - the code is subtle enough as it is. I'm dropping this patch and #7 which depends on this one, if Joonsoo wants to later pursue this, he can.
Thanks, Davidlohr
| |