[PATCH V13 2/6] mm: userfaultfd: Add pgtable_supports_uffd_wp()

Chunyan Zhang zhang.lyra at gmail.com
Wed Sep 17 02:20:14 PDT 2025


On Wed, 17 Sept 2025 at 15:25, David Hildenbrand <david at redhat.com> wrote:
>
> On 17.09.25 05:36, Chunyan Zhang wrote:
> > Some platforms can customize the PTE/PMD entry uffd-wp bit making
> > it unavailable even if the architecture provides the resource.
> > This patch adds a macro API that allows architectures to define their
> > specific implementations to check if the uffd-wp bit is available
> > on which device the kernel is running.
> >
> > Also this patch is removing "ifdef CONFIG_HAVE_ARCH_USERFAULTFD_WP" and
> > "ifdef CONFIG_PTE_MARKER_UFFD_WP" in favor of pgtable_supports_uffd_wp()
> > and uffd_supports_wp_marker() checks respectively that default to
> > IS_ENABLED(CONFIG_HAVE_ARCH_USERFAULTFD_WP) and
> > "IS_ENABLED(CONFIG_HAVE_ARCH_USERFAULTFD_WP) && IS_ENABLED(CONFIG_PTE_MARKER_UFFD_WP)"
> > if not overridden by the architecture, no change in behavior is expected.
> >
> > Acked-by: David Hildenbrand <david at redhat.com>
> > Signed-off-by: Chunyan Zhang <zhangchunyan at iscas.ac.cn>
> > ---
>
> [...]
>
> Taking another look.
>
> >   /* mm helpers */
> > @@ -415,68 +475,24 @@ static inline bool vma_has_uffd_without_event_remap(struct vm_area_struct *vma)
> >       return false;
> >   }
> >
> > -#endif /* CONFIG_USERFAULTFD */
> > -
> >   static inline bool userfaultfd_wp_use_markers(struct vm_area_struct *vma)
> >   {
> > -     /* Only wr-protect mode uses pte markers */
> > -     if (!userfaultfd_wp(vma))
> >               return false;
>
> Isn't this indented one level too deep?

Oh right, I will fix these.

Thanks to you spotting them out!

Chunyan

>
> > -
> > -     /* File-based uffd-wp always need markers */
> > -     if (!vma_is_anonymous(vma))
> > -             return true;
> > -
> > -     /*
> > -      * Anonymous uffd-wp only needs the markers if WP_UNPOPULATED
> > -      * enabled (to apply markers on zero pages).
> > -      */
> > -     return userfaultfd_wp_unpopulated(vma);
> >   }
> >
> >   static inline bool pte_marker_entry_uffd_wp(swp_entry_t entry)
> >   {
> > -#ifdef CONFIG_PTE_MARKER_UFFD_WP
> > -     return is_pte_marker_entry(entry) &&
> > -         (pte_marker_get(entry) & PTE_MARKER_UFFD_WP);
> > -#else
> > -     return false;
> > -#endif
> > +             return false;
>
> Same here.
>
> >   }
> >
> >   static inline bool pte_marker_uffd_wp(pte_t pte)
> >   {
> > -#ifdef CONFIG_PTE_MARKER_UFFD_WP
> > -     swp_entry_t entry;
> > -
> > -     if (!is_swap_pte(pte))
> >               return false;
>
> Same here.
>
> > -
> > -     entry = pte_to_swp_entry(pte);
> > -
> > -     return pte_marker_entry_uffd_wp(entry);
> > -#else
> > -     return false;
> > -#endif
> >   }
>
>
> --
> Cheers
>
> David / dhildenb
>



More information about the linux-riscv mailing list