Re: [PATCH] mm/hugetlb: fix swap entry corruption when clearing uffd-wp at fork()

[David Hildenbrand (Arm)]

On 7/3/26 18:18, Kiryl Shutsemau wrote:
> From: "Kiryl Shutsemau (Meta)" <kas@xxxxxxxxxx>
> 
> copy_hugetlb_page_range() clears the uffd-wp bit of hwpoison and
> migration entries with huge_pte_clear_uffd_wp(), which operates on the
> present-PTE bit position. Swap entries keep the uffd-wp state elsewhere
> -- the same branches read and set it with pte_swp_uffd_wp() and
> pte_swp_mkuffd_wp() -- and the present-PTE position falls into the swap
> payload. On x86-64 it lands in the inverted swap offset, where a
> naturally-aligned hugetlb PFN always has the affected bit set, so the
> clear advances the encoded PFN by two pages.
> 
> No userfaultfd needs to be involved: the clear is guarded only by the
> child VMA not being uffd-wp registered, so a plain fork() with an
> in-flight hugetlb migration entry (or a poisoned hugetlb page) corrupts
> the entry copied into the child. Instrumenting the hwpoison branch and
> forking after MADV_HWPOISON on a 2MB anon hugetlb page shows:
> 
>   offset before=120e00
>   offset after =120e02
> 
> The fallout is mostly latent: rmap walks match migration entries by
> folio range and remove_migration_pte() rebuilds the PTE from the folio,
> so a within-folio PFN skew heals once migration completes. But any path
> that re-encodes the corrupted offset -- e.g. hugetlb_change_protection()
> rewriting a writable migration entry via
> make_readable_migration_entry(swp_offset(entry)) -- propagates it, and
> an hwpoison entry misidentifies which page is poisoned.
> 
> Use pte_swp_clear_uffd_wp(), matching copy_nonpresent_pte() and
> move_huge_pte().
> 
> Reported-by: Sashiko AI review <sashiko-bot@xxxxxxxxxx>
> Closes: https://lore.kernel.org/all/20260703140011.99E601F000E9@xxxxxxxxxxxxxxx/
> Fixes: bc70fbf269fd ("mm/hugetlb: handle uffd-wp during fork()")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Kiryl Shutsemau <kas@xxxxxxxxxx>
> Assisted-by: Claude:claude-fable-5
> ---
>  mm/hugetlb.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 571212b80835..a4e6dd3a82f4 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -4918,7 +4918,7 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
>  		softleaf = softleaf_from_pte(entry);
>  		if (unlikely(softleaf_is_hwpoison(softleaf))) {
>  			if (!userfaultfd_wp(dst_vma))
> -				entry = huge_pte_clear_uffd_wp(entry);
> +				entry = pte_swp_clear_uffd_wp(entry);

I think installing a hwpoison pte will actually drop the uffd marker.

hugetlb_change_protection() does nothing on hwpoison entrues.

So how could be possibly get a hwpoison entry with an uffd-wp bit set here?

If we indeed can't, Id assume there is nothing to clear here at all.

>  			set_huge_pte_at(dst, addr, dst_pte, entry, sz);
>  		} else if (unlikely(softleaf_is_migration(softleaf))) {
>  			bool uffd_wp = pte_swp_uffd_wp(entry);
> @@ -4936,7 +4936,7 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
>  				set_huge_pte_at(src, addr, src_pte, entry, sz);
>  			}
>  			if (!userfaultfd_wp(dst_vma))
> -				entry = huge_pte_clear_uffd_wp(entry);
> +				entry = pte_swp_clear_uffd_wp(entry);


That looks correct.


-- 
Cheers,

David