Re: [PATCH mm-new v3 1/3] mm/khugepaged: optimize PTE scanning with if-else-if-else-if chain

[Lorenzo Stoakes <lorenzo.stoakes@oracle.com>]

On Wed, Oct 08, 2025 at 12:37:46PM +0800, Lance Yang wrote:
> From: Lance Yang <lance.yang@linux.dev>
>
> As pointed out by Dev, the PTE checks for disjoint conditions in the
> scanning loops can be optimized. is_swap_pte, (pte_none && is_zero_pfn),
> and pte_uffd_wp are mutually exclusive.

But you're not using is_swap_pte anywhere :) This comes back to my review
quesiotn on the series this is dependent upon.

>
> This patch refactors the loops in both __collapse_huge_page_isolate() and
> hpage_collapse_scan_pmd() to use a continuous if-else-if-else-if chain
> instead of separate if blocks. While at it, the redundant pte_present()
> check before is_zero_pfn() is also removed.

I mean see review below, I don't see why you're doing this and I am
unconvinced by how redundant that check is.

Also this just feels like it should be part of the series where you change
these? I'm not sure why this is separate?

>
> Also, this is a preparatory step to make it easier to merge the
> almost-duplicated scanning logic in these two functions, as suggested
> by David.
>
> Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
> Reviewed-by: Dev Jain <dev.jain@arm.com>
> Reviewed-by: Zi Yan <ziy@nvidia.com>
> Suggested-by: Dev Jain <dev.jain@arm.com>
> Suggested-by: David Hildenbrand <david@redhat.com>
> Signed-off-by: Lance Yang <lance.yang@linux.dev>
> ---
>  mm/khugepaged.c | 12 ++++--------
>  1 file changed, 4 insertions(+), 8 deletions(-)
>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index bec3e268dc76..e3e27223137a 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -548,8 +548,7 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
>  	for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
>  	     _pte++, addr += PAGE_SIZE) {
>  		pte_t pteval = ptep_get(_pte);
> -		if (pte_none(pteval) || (pte_present(pteval) &&
> -				is_zero_pfn(pte_pfn(pteval)))) {
> +		if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {

You can have non-pte_none() non-present entries no? Isn't pte_present() a
prerequisite for pfe_pfn() to be valid? If it's a swap entry couldn't you
end up accidentally (unlikely but still) hitting this?

Seems like this is required isn't it? I may be missing something here...

>  			++none_or_zero;
>  			if (!userfaultfd_armed(vma) &&
>  			    (!cc->is_khugepaged ||
> @@ -560,12 +559,10 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
>  				count_vm_event(THP_SCAN_EXCEED_NONE_PTE);
>  				goto out;
>  			}
> -		}
> -		if (!pte_present(pteval)) {
> +		} else if (!pte_present(pteval)) {

This seems pointless, since either the above logic will continue or goto
out right?

>  			result = SCAN_PTE_NON_PRESENT;
>  			goto out;
> -		}
> -		if (pte_uffd_wp(pteval)) {
> +		} else if (pte_uffd_wp(pteval)) {

Again, what is the point of an else when the if() branch unconditionally
->out?

>  			result = SCAN_PTE_UFFD_WP;
>  			goto out;
>  		}
> @@ -1321,8 +1318,7 @@ static int hpage_collapse_scan_pmd(struct mm_struct *mm,
>  				count_vm_event(THP_SCAN_EXCEED_SWAP_PTE);
>  				goto out_unmap;
>  			}
> -		}
> -		if (pte_uffd_wp(pteval)) {
> +		} else if (pte_uffd_wp(pteval)) {

Same comment as above, I'm really confused about the purpose of this logic?


>  			/*
>  			 * Don't collapse the page if any of the small
>  			 * PTEs are armed with uffd write protection.
> --
> 2.49.0
>