Re: [PATCH 1/6] mm/gup: move gup_can_follow_protnone() to gup.c

[David Hildenbrand (Arm)]

On 8/23/26 14:17, Mike Rapoport (Microsoft) wrote:
> gup_can_follow_protnone() is defined in include/linux/mm.h but only used
> by mm/gup.c.
> 
> First, there is no reason to have it in already gigantic header.

Once upon a time there was a user in mm/huge_memory.c, in a beautifully named
function called follow_trans_huge_pmd().

> 
> Next, the upcoming refactoring of userfaultfd flags will make
> gup_can_follow_protnone() depend on userfaultfd_k.h which would cause a
> cyclic header dependency.
> 
> Move gup_can_follow_protnone() to mm/gup.c.
> 
> No functional change.
> 
> Assisted-by: copilot:claude-opus-5
> Signed-off-by: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>
> ---


[...]

>  typedef int (*pte_fn_t)(pte_t *pte, unsigned long addr, void *data);
>  extern int apply_to_page_range(struct mm_struct *mm, unsigned long address,
>  			       unsigned long size, pte_fn_t fn, void *data);
> diff --git a/mm/gup.c b/mm/gup.c
> index eb898ea1ee22..500e2aa99e48 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -633,6 +633,44 @@ static struct page *no_page_table(struct vm_area_struct *vma,
>  	return NULL;
>  }
>  
> +/*
> + * Indicates whether GUP can follow a PROT_NONE mapped page, or whether
> + * a (NUMA hinting or userfaultfd RWP) fault is required.
> + */
> +static inline bool gup_can_follow_protnone(const struct vm_area_struct *vma,
> +					   unsigned int flags)

While at it, could switch to double-tab here.

Acked-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>

-- 
Cheers,

David