[v9] khugepaged: mTHP support

[PATCH v9 08/14] khugepaged: skip collapsing mTHP to smaller orders

Posted by Nico Pache 2 months, 3 weeks ago

khugepaged may try to collapse a mTHP to a smaller mTHP, resulting in
some pages being unmapped. Skip these cases until we have a way to check
if its ok to collapse to a smaller mTHP size (like in the case of a
partially mapped folio).

This patch is inspired by Dev Jain's work on khugepaged mTHP support [1].

[1] https://lore.kernel.org/lkml/20241216165105.56185-11-dev.jain@arm.com/

Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Co-developed-by: Dev Jain <dev.jain@arm.com>
Signed-off-by: Dev Jain <dev.jain@arm.com>
Signed-off-by: Nico Pache <npache@redhat.com>
---
 mm/khugepaged.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 5d7c5be9097e..a701d9f0f158 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -612,7 +612,12 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
 		folio = page_folio(page);
 		VM_BUG_ON_FOLIO(!folio_test_anon(folio), folio);
 
-		/* See hpage_collapse_scan_pmd(). */
+		if (order != HPAGE_PMD_ORDER && folio_order(folio) >= order) {
+			result = SCAN_PTE_MAPPED_HUGEPAGE;
+			goto out;
+		}
+
+		/* See khugepaged_scan_pmd(). */
 		if (folio_maybe_mapped_shared(folio)) {
 			++shared;
 			if (order != HPAGE_PMD_ORDER || (cc->is_khugepaged &&
-- 
2.50.0

Re: [PATCH v9 08/14] khugepaged: skip collapsing mTHP to smaller orders

Posted by David Hildenbrand 2 months, 3 weeks ago

On 14.07.25 02:32, Nico Pache wrote:
> khugepaged may try to collapse a mTHP to a smaller mTHP, resulting in
> some pages being unmapped. Skip these cases until we have a way to check
> if its ok to collapse to a smaller mTHP size (like in the case of a
> partially mapped folio).
> 
> This patch is inspired by Dev Jain's work on khugepaged mTHP support [1].
> 
> [1] https://lore.kernel.org/lkml/20241216165105.56185-11-dev.jain@arm.com/
> 
> Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
> Co-developed-by: Dev Jain <dev.jain@arm.com>
> Signed-off-by: Dev Jain <dev.jain@arm.com>
> Signed-off-by: Nico Pache <npache@redhat.com>
> ---
>   mm/khugepaged.c | 7 ++++++-
>   1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 5d7c5be9097e..a701d9f0f158 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -612,7 +612,12 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
>   		folio = page_folio(page);
>   		VM_BUG_ON_FOLIO(!folio_test_anon(folio), folio);
>   
> -		/* See hpage_collapse_scan_pmd(). */
> +		if (order != HPAGE_PMD_ORDER && folio_order(folio) >= order) {
> +			result = SCAN_PTE_MAPPED_HUGEPAGE;
> +			goto out;
> +		}

Probably worth adding a TODO in the code like

/*
  * TODO: In some cases of partially-mapped folios, we'd actually
  * want to collapse.
  */

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers,

David / dhildenb

Re: [PATCH v9 08/14] khugepaged: skip collapsing mTHP to smaller orders

Posted by Nico Pache 2 months, 3 weeks ago

On Wed, Jul 16, 2025 at 8:32 AM David Hildenbrand <david@redhat.com> wrote:
>
> On 14.07.25 02:32, Nico Pache wrote:
> > khugepaged may try to collapse a mTHP to a smaller mTHP, resulting in
> > some pages being unmapped. Skip these cases until we have a way to check
> > if its ok to collapse to a smaller mTHP size (like in the case of a
> > partially mapped folio).
> >
> > This patch is inspired by Dev Jain's work on khugepaged mTHP support [1].
> >
> > [1] https://lore.kernel.org/lkml/20241216165105.56185-11-dev.jain@arm.com/
> >
> > Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
> > Co-developed-by: Dev Jain <dev.jain@arm.com>
> > Signed-off-by: Dev Jain <dev.jain@arm.com>
> > Signed-off-by: Nico Pache <npache@redhat.com>
> > ---
> >   mm/khugepaged.c | 7 ++++++-
> >   1 file changed, 6 insertions(+), 1 deletion(-)
> >
> > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > index 5d7c5be9097e..a701d9f0f158 100644
> > --- a/mm/khugepaged.c
> > +++ b/mm/khugepaged.c
> > @@ -612,7 +612,12 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> >               folio = page_folio(page);
> >               VM_BUG_ON_FOLIO(!folio_test_anon(folio), folio);
> >
> > -             /* See hpage_collapse_scan_pmd(). */
> > +             if (order != HPAGE_PMD_ORDER && folio_order(folio) >= order) {
> > +                     result = SCAN_PTE_MAPPED_HUGEPAGE;
> > +                     goto out;
> > +             }
>
> Probably worth adding a TODO in the code like
>
> /*
>   * TODO: In some cases of partially-mapped folios, we'd actually
>   * want to collapse.
>   */
Done! Good idea with these TODOs!
>
> Acked-by: David Hildenbrand <david@redhat.com>
Thank you :)
>
> --
> Cheers,
>
> David / dhildenb
>