Re: [PATCH 4/4] mm/mempolicy: skip zone device folios when queueing folios

On Thu, Aug 06, 2026 at 04:11:21PM +0200, David Hildenbrand (Arm) wrote:
> On 8/6/26 15:36, Lorenzo Stoakes (ARM) wrote:
> > On Thu, Aug 06, 2026 at 03:25:50PM +0200, David Hildenbrand (Arm) wrote:
> >> On 7/28/26 21:47, Gregory Price wrote:
> >>>  mm/mempolicy.c | 4 ++++
> >>> @@ -797,6 +799,8 @@ static int queue_folios_hugetlb(pte_t *pte, unsigned long hmask,
> >>>  	folio = pfn_folio(pte_pfn(ptep));
> >>>  	if (!queue_folio_required(folio, qp))
> >>>  		goto unlock;
> >>> +	if (folio_is_zone_device(folio))
> >>> +		goto unlock;
> >>
> >> Leave the hugetlb part alone. hugetlb folios will never be ZONE_DEVICE.
> >
> > <Insert leave Britney alone meme here> :P
> >
> > But dear God we have got to find ways to stop hugetlb being this big stupid
> > nightmare that every single path of mm has to put an asterix next to itself
> > about!
>
> Yeah, possibly unifying the code (possibly, because it's a mess) but not
> replicating checks that don't make any sense for hugetlb.

Yup agreed.

>
> --
> Cheers,
>
> David

--
Cheers, Lorenzo