[PATCH v2 1/2] mm/gup: Remove dead pgmap refcounting code

[PATCH v2 1/2] mm/gup: Remove dead pgmap refcounting code

[Alistair Popple]

Prior to commit aed877c2b425 ("device/dax: properly refcount device dax
pages when mapping") ZONE_DEVICE pages were not fully reference counted
when mapped into user page tables. Instead GUP would take a reference
on the associated pgmap to ensure the results of pfn_to_page() remained
valid.

This is no longer required and most of the code was removed by commit
fd2825b0760a ("mm/gup: remove pXX_devmap usage from get_user_pages()").
Finish cleaning this up by removing the dead calls to put_dev_pagemap()
and the temporary context struct.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
---
 mm/gup.c | 67 ++++++++++++++++++++++----------------------------------
 1 file changed, 26 insertions(+), 41 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index adffe663594d..be5791a43c73 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -28,11 +28,6 @@
 #include "internal.h"
 #include "swap.h"
 
-struct follow_page_context {
-	struct dev_pagemap *pgmap;
-	unsigned int page_mask;
-};
-
 static inline void sanity_check_pinned_pages(struct page **pages,
 					     unsigned long npages)
 {
@@ -661,7 +656,7 @@ static inline bool can_follow_write_pud(pud_t pud, struct page *page,
 
 static struct page *follow_huge_pud(struct vm_area_struct *vma,
 				    unsigned long addr, pud_t *pudp,
-				    int flags, struct follow_page_context *ctx)
+				    int flags, unsigned long *page_mask)
 {
 	struct mm_struct *mm = vma->vm_mm;
 	struct page *page;
@@ -688,7 +683,7 @@ static struct page *follow_huge_pud(struct vm_area_struct *vma,
 	if (ret)
 		page = ERR_PTR(ret);
 	else
-		ctx->page_mask = HPAGE_PUD_NR - 1;
+		*page_mask = HPAGE_PUD_NR - 1;
 
 	return page;
 }
@@ -714,7 +709,7 @@ static inline bool can_follow_write_pmd(pmd_t pmd, struct page *page,
 static struct page *follow_huge_pmd(struct vm_area_struct *vma,
 				    unsigned long addr, pmd_t *pmd,
 				    unsigned int flags,
-				    struct follow_page_context *ctx)
+				    unsigned long *page_mask)
 {
 	struct mm_struct *mm = vma->vm_mm;
 	pmd_t pmdval = *pmd;
@@ -751,7 +746,7 @@ static struct page *follow_huge_pmd(struct vm_area_struct *vma,
 #endif	/* CONFIG_TRANSPARENT_HUGEPAGE */
 
 	page += (addr & ~HPAGE_PMD_MASK) >> PAGE_SHIFT;
-	ctx->page_mask = HPAGE_PMD_NR - 1;
+	*page_mask = HPAGE_PMD_NR - 1;
 
 	return page;
 }
@@ -759,7 +754,7 @@ static struct page *follow_huge_pmd(struct vm_area_struct *vma,
 #else  /* CONFIG_PGTABLE_HAS_HUGE_LEAVES */
 static struct page *follow_huge_pud(struct vm_area_struct *vma,
 				    unsigned long addr, pud_t *pudp,
-				    int flags, struct follow_page_context *ctx)
+				    int flags, unsigned long *page_mask)
 {
 	return NULL;
 }
@@ -767,7 +762,7 @@ static struct page *follow_huge_pud(struct vm_area_struct *vma,
 static struct page *follow_huge_pmd(struct vm_area_struct *vma,
 				    unsigned long addr, pmd_t *pmd,
 				    unsigned int flags,
-				    struct follow_page_context *ctx)
+				    unsigned long *page_mask)
 {
 	return NULL;
 }
@@ -813,8 +808,7 @@ static inline bool can_follow_write_pte(pte_t pte, struct page *page,
 }
 
 static struct page *follow_page_pte(struct vm_area_struct *vma,
-		unsigned long address, pmd_t *pmd, unsigned int flags,
-		struct dev_pagemap **pgmap)
+		unsigned long address, pmd_t *pmd, unsigned int flags)
 {
 	struct mm_struct *mm = vma->vm_mm;
 	struct folio *folio;
@@ -912,7 +906,7 @@ static struct page *follow_page_pte(struct vm_area_struct *vma,
 static struct page *follow_pmd_mask(struct vm_area_struct *vma,
 				    unsigned long address, pud_t *pudp,
 				    unsigned int flags,
-				    struct follow_page_context *ctx)
+				    unsigned long *page_mask)
 {
 	pmd_t *pmd, pmdval;
 	spinlock_t *ptl;
@@ -926,7 +920,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
 	if (!pmd_present(pmdval))
 		return no_page_table(vma, flags, address);
 	if (likely(!pmd_leaf(pmdval)))
-		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
+		return follow_page_pte(vma, address, pmd, flags);
 
 	if (pmd_protnone(pmdval) && !gup_can_follow_protnone(vma, flags))
 		return no_page_table(vma, flags, address);
@@ -939,16 +933,16 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
 	}
 	if (unlikely(!pmd_leaf(pmdval))) {
 		spin_unlock(ptl);
-		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
+		return follow_page_pte(vma, address, pmd, flags);
 	}
 	if (pmd_trans_huge(pmdval) && (flags & FOLL_SPLIT_PMD)) {
 		spin_unlock(ptl);
 		split_huge_pmd(vma, pmd, address);
 		/* If pmd was left empty, stuff a page table in there quickly */
 		return pte_alloc(mm, pmd) ? ERR_PTR(-ENOMEM) :
-			follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
+			follow_page_pte(vma, address, pmd, flags);
 	}
-	page = follow_huge_pmd(vma, address, pmd, flags, ctx);
+	page = follow_huge_pmd(vma, address, pmd, flags, page_mask);
 	spin_unlock(ptl);
 	return page;
 }
@@ -956,7 +950,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
 static struct page *follow_pud_mask(struct vm_area_struct *vma,
 				    unsigned long address, p4d_t *p4dp,
 				    unsigned int flags,
-				    struct follow_page_context *ctx)
+				    unsigned long *page_mask)
 {
 	pud_t *pudp, pud;
 	spinlock_t *ptl;
@@ -969,7 +963,7 @@ static struct page *follow_pud_mask(struct vm_area_struct *vma,
 		return no_page_table(vma, flags, address);
 	if (pud_leaf(pud)) {
 		ptl = pud_lock(mm, pudp);
-		page = follow_huge_pud(vma, address, pudp, flags, ctx);
+		page = follow_huge_pud(vma, address, pudp, flags, page_mask);
 		spin_unlock(ptl);
 		if (page)
 			return page;
@@ -978,13 +972,13 @@ static struct page *follow_pud_mask(struct vm_area_struct *vma,
 	if (unlikely(pud_bad(pud)))
 		return no_page_table(vma, flags, address);
 
-	return follow_pmd_mask(vma, address, pudp, flags, ctx);
+	return follow_pmd_mask(vma, address, pudp, flags, page_mask);
 }
 
 static struct page *follow_p4d_mask(struct vm_area_struct *vma,
 				    unsigned long address, pgd_t *pgdp,
 				    unsigned int flags,
-				    struct follow_page_context *ctx)
+				    unsigned long *page_mask)
 {
 	p4d_t *p4dp, p4d;
 
@@ -995,7 +989,7 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
 	if (!p4d_present(p4d) || p4d_bad(p4d))
 		return no_page_table(vma, flags, address);
 
-	return follow_pud_mask(vma, address, p4dp, flags, ctx);
+	return follow_pud_mask(vma, address, p4dp, flags, page_mask);
 }
 
 /**
@@ -1003,20 +997,16 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
  * @vma: vm_area_struct mapping @address
  * @address: virtual address to look up
  * @flags: flags modifying lookup behaviour
- * @ctx: contains dev_pagemap for %ZONE_DEVICE memory pinning and a
- *       pointer to output page_mask
+ * @page_mask: a pointer to output page_mask
  *
  * @flags can have FOLL_ flags set, defined in <linux/mm.h>
  *
- * When getting pages from ZONE_DEVICE memory, the @ctx->pgmap caches
- * the device's dev_pagemap metadata to avoid repeating expensive lookups.
- *
  * When getting an anonymous page and the caller has to trigger unsharing
  * of a shared anonymous page first, -EMLINK is returned. The caller should
  * trigger a fault with FAULT_FLAG_UNSHARE set. Note that unsharing is only
  * relevant with FOLL_PIN and !FOLL_WRITE.
  *
- * On output, the @ctx->page_mask is set according to the size of the page.
+ * On output, @page_mask is set according to the size of the page.
  *
  * Return: the mapped (struct page *), %NULL if no mapping exists, or
  * an error pointer if there is a mapping to something not represented
@@ -1024,7 +1014,7 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
  */
 static struct page *follow_page_mask(struct vm_area_struct *vma,
 			      unsigned long address, unsigned int flags,
-			      struct follow_page_context *ctx)
+			      unsigned long *page_mask)
 {
 	pgd_t *pgd;
 	struct mm_struct *mm = vma->vm_mm;
@@ -1032,13 +1022,13 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,
 
 	vma_pgtable_walk_begin(vma);
 
-	ctx->page_mask = 0;
+	*page_mask = 0;
 	pgd = pgd_offset(mm, address);
 
 	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
 		page = no_page_table(vma, flags, address);
 	else
-		page = follow_p4d_mask(vma, address, pgd, flags, ctx);
+		page = follow_p4d_mask(vma, address, pgd, flags, page_mask);
 
 	vma_pgtable_walk_end(vma);
 
@@ -1376,7 +1366,7 @@ static long __get_user_pages(struct mm_struct *mm,
 {
 	long ret = 0, i = 0;
 	struct vm_area_struct *vma = NULL;
-	struct follow_page_context ctx = { NULL };
+	unsigned long page_mask = 0;
 
 	if (!nr_pages)
 		return 0;
@@ -1418,7 +1408,7 @@ static long __get_user_pages(struct mm_struct *mm,
 						pages ? &page : NULL);
 				if (ret)
 					goto out;
-				ctx.page_mask = 0;
+				page_mask = 0;
 				goto next_page;
 			}
 
@@ -1441,7 +1431,7 @@ static long __get_user_pages(struct mm_struct *mm,
 		}
 		cond_resched();
 
-		page = follow_page_mask(vma, start, gup_flags, &ctx);
+		page = follow_page_mask(vma, start, gup_flags, &page_mask);
 		if (!page || PTR_ERR(page) == -EMLINK) {
 			ret = faultin_page(vma, start, gup_flags,
 					   PTR_ERR(page) == -EMLINK, locked);
@@ -1474,7 +1464,7 @@ static long __get_user_pages(struct mm_struct *mm,
 			goto out;
 		}
 next_page:
-		page_increm = 1 + (~(start >> PAGE_SHIFT) & ctx.page_mask);
+		page_increm = 1 + (~(start >> PAGE_SHIFT) & page_mask);
 		if (page_increm > nr_pages)
 			page_increm = nr_pages;
 
@@ -1524,8 +1514,6 @@ static long __get_user_pages(struct mm_struct *mm,
 		nr_pages -= page_increm;
 	} while (nr_pages);
 out:
-	if (ctx.pgmap)
-		put_dev_pagemap(ctx.pgmap);
 	return i ? i : ret;
 }
 
@@ -2853,7 +2841,6 @@ static int gup_fast_pte_range(pmd_t pmd, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, unsigned int flags, struct page **pages,
 		int *nr)
 {
-	struct dev_pagemap *pgmap = NULL;
 	int ret = 0;
 	pte_t *ptep, *ptem;
 
@@ -2926,8 +2913,6 @@ static int gup_fast_pte_range(pmd_t pmd, pmd_t *pmdp, unsigned long addr,
 	ret = 1;
 
 pte_unmap:
-	if (pgmap)
-		put_dev_pagemap(pgmap);
 	pte_unmap(ptem);
 	return ret;
 }
-- 
2.50.1

[PATCH v2 2/2] mm/memremap: Remove unused get_dev_pagemap() parameter

[Alistair Popple]

GUP no longer uses get_dev_pagemap(). As it was the only user of the
get_dev_pagemap() pgmap caching feature it can be removed.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
---
 include/linux/memremap.h |  6 ++----
 mm/memory-failure.c      |  2 +-
 mm/memory_hotplug.c      |  2 +-
 mm/memremap.c            | 22 ++++------------------
 4 files changed, 8 insertions(+), 24 deletions(-)

diff --git a/include/linux/memremap.h b/include/linux/memremap.h
index 4aa151914eab..5cfc2ae6c98b 100644
--- a/include/linux/memremap.h
+++ b/include/linux/memremap.h
@@ -204,8 +204,7 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid);
 void memunmap_pages(struct dev_pagemap *pgmap);
 void *devm_memremap_pages(struct device *dev, struct dev_pagemap *pgmap);
 void devm_memunmap_pages(struct device *dev, struct dev_pagemap *pgmap);
-struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
-		struct dev_pagemap *pgmap);
+struct dev_pagemap *get_dev_pagemap(unsigned long pfn);
 bool pgmap_pfn_valid(struct dev_pagemap *pgmap, unsigned long pfn);
 
 unsigned long memremap_compat_align(void);
@@ -227,8 +226,7 @@ static inline void devm_memunmap_pages(struct device *dev,
 {
 }
 
-static inline struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
-		struct dev_pagemap *pgmap)
+static inline struct dev_pagemap *get_dev_pagemap(unsigned long pfn)
 {
 	return NULL;
 }
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index fc30ca4804bf..21298f861a5a 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -2266,7 +2266,7 @@ int memory_failure(unsigned long pfn, int flags)
 			goto unlock_mutex;
 
 		if (pfn_valid(pfn)) {
-			pgmap = get_dev_pagemap(pfn, NULL);
+			pgmap = get_dev_pagemap(pfn);
 			put_ref_page(pfn, flags);
 			if (pgmap) {
 				res = memory_failure_dev_pagemap(pfn, flags,
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 1f15af712bc3..14e98c89f5b4 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -375,7 +375,7 @@ struct page *pfn_to_online_page(unsigned long pfn)
 	 * the section may be 'offline' but 'valid'. Only
 	 * get_dev_pagemap() can determine sub-section online status.
 	 */
-	pgmap = get_dev_pagemap(pfn, NULL);
+	pgmap = get_dev_pagemap(pfn);
 	put_dev_pagemap(pgmap);
 
 	/* The presence of a pgmap indicates ZONE_DEVICE offline pfn */
diff --git a/mm/memremap.c b/mm/memremap.c
index b0ce0d8254bd..3b4717cf3a6b 100644
--- a/mm/memremap.c
+++ b/mm/memremap.c
@@ -153,14 +153,14 @@ static int pagemap_range(struct dev_pagemap *pgmap, struct mhp_params *params,
 				"altmap not supported for multiple ranges\n"))
 		return -EINVAL;
 
-	conflict_pgmap = get_dev_pagemap(PHYS_PFN(range->start), NULL);
+	conflict_pgmap = get_dev_pagemap(PHYS_PFN(range->start));
 	if (conflict_pgmap) {
 		WARN(1, "Conflicting mapping in same section\n");
 		put_dev_pagemap(conflict_pgmap);
 		return -ENOMEM;
 	}
 
-	conflict_pgmap = get_dev_pagemap(PHYS_PFN(range->end), NULL);
+	conflict_pgmap = get_dev_pagemap(PHYS_PFN(range->end));
 	if (conflict_pgmap) {
 		WARN(1, "Conflicting mapping in same section\n");
 		put_dev_pagemap(conflict_pgmap);
@@ -394,26 +394,12 @@ EXPORT_SYMBOL_GPL(devm_memunmap_pages);
 /**
  * get_dev_pagemap() - take a new live reference on the dev_pagemap for @pfn
  * @pfn: page frame number to lookup page_map
- * @pgmap: optional known pgmap that already has a reference
- *
- * If @pgmap is non-NULL and covers @pfn it will be returned as-is.  If @pgmap
- * is non-NULL but does not cover @pfn the reference to it will be released.
  */
-struct dev_pagemap *get_dev_pagemap(unsigned long pfn,
-		struct dev_pagemap *pgmap)
+struct dev_pagemap *get_dev_pagemap(unsigned long pfn)
 {
+	struct dev_pagemap *pgmap;
 	resource_size_t phys = PFN_PHYS(pfn);
 
-	/*
-	 * In the cached case we're already holding a live reference.
-	 */
-	if (pgmap) {
-		if (phys >= pgmap->range.start && phys <= pgmap->range.end)
-			return pgmap;
-		put_dev_pagemap(pgmap);
-	}
-
-	/* fall back to slow path lookup */
 	rcu_read_lock();
 	pgmap = xa_load(&pgmap_array, PHYS_PFN(phys));
 	if (pgmap && !percpu_ref_tryget_live_rcu(&pgmap->ref))
-- 
2.50.1

Re: [PATCH v2 1/2] mm/gup: Remove dead pgmap refcounting code

[John Hubbard]

On 9/3/25 3:59 PM, Alistair Popple wrote:
> Prior to commit aed877c2b425 ("device/dax: properly refcount device dax
> pages when mapping") ZONE_DEVICE pages were not fully reference counted
> when mapped into user page tables. Instead GUP would take a reference
> on the associated pgmap to ensure the results of pfn_to_page() remained
> valid.
> 
> This is no longer required and most of the code was removed by commit
> fd2825b0760a ("mm/gup: remove pXX_devmap usage from get_user_pages()").
> Finish cleaning this up by removing the dead calls to put_dev_pagemap()
> and the temporary context struct.
> 
> Signed-off-by: Alistair Popple <apopple@nvidia.com>
> Acked-by: David Hildenbrand <david@redhat.com>
> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
> Reviewed-by: Dan Williams <dan.j.williams@intel.com>
> ---
>  mm/gup.c | 67 ++++++++++++++++++++++----------------------------------
>  1 file changed, 26 insertions(+), 41 deletions(-)
> 

Wow, nice culmination of a lot of work. So good to remove this
complication!

Maybe I'll post a quick follow-up, in fact, to remove page_mask
as well. No need to pass that around all over the place, I think
it can instead be computed by the single place that actually
consumes it.

Reviewed-by: John Hubbard <jhubbard@nvidia.com>

thanks,
-- 
John Hubbard

> diff --git a/mm/gup.c b/mm/gup.c
> index adffe663594d..be5791a43c73 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -28,11 +28,6 @@
>  #include "internal.h"
>  #include "swap.h"
>  
> -struct follow_page_context {
> -	struct dev_pagemap *pgmap;
> -	unsigned int page_mask;
> -};
> -
>  static inline void sanity_check_pinned_pages(struct page **pages,
>  					     unsigned long npages)
>  {
> @@ -661,7 +656,7 @@ static inline bool can_follow_write_pud(pud_t pud, struct page *page,
>  
>  static struct page *follow_huge_pud(struct vm_area_struct *vma,
>  				    unsigned long addr, pud_t *pudp,
> -				    int flags, struct follow_page_context *ctx)
> +				    int flags, unsigned long *page_mask)
>  {
>  	struct mm_struct *mm = vma->vm_mm;
>  	struct page *page;
> @@ -688,7 +683,7 @@ static struct page *follow_huge_pud(struct vm_area_struct *vma,
>  	if (ret)
>  		page = ERR_PTR(ret);
>  	else
> -		ctx->page_mask = HPAGE_PUD_NR - 1;
> +		*page_mask = HPAGE_PUD_NR - 1;
>  
>  	return page;
>  }
> @@ -714,7 +709,7 @@ static inline bool can_follow_write_pmd(pmd_t pmd, struct page *page,
>  static struct page *follow_huge_pmd(struct vm_area_struct *vma,
>  				    unsigned long addr, pmd_t *pmd,
>  				    unsigned int flags,
> -				    struct follow_page_context *ctx)
> +				    unsigned long *page_mask)
>  {
>  	struct mm_struct *mm = vma->vm_mm;
>  	pmd_t pmdval = *pmd;
> @@ -751,7 +746,7 @@ static struct page *follow_huge_pmd(struct vm_area_struct *vma,
>  #endif	/* CONFIG_TRANSPARENT_HUGEPAGE */
>  
>  	page += (addr & ~HPAGE_PMD_MASK) >> PAGE_SHIFT;
> -	ctx->page_mask = HPAGE_PMD_NR - 1;
> +	*page_mask = HPAGE_PMD_NR - 1;
>  
>  	return page;
>  }
> @@ -759,7 +754,7 @@ static struct page *follow_huge_pmd(struct vm_area_struct *vma,
>  #else  /* CONFIG_PGTABLE_HAS_HUGE_LEAVES */
>  static struct page *follow_huge_pud(struct vm_area_struct *vma,
>  				    unsigned long addr, pud_t *pudp,
> -				    int flags, struct follow_page_context *ctx)
> +				    int flags, unsigned long *page_mask)
>  {
>  	return NULL;
>  }
> @@ -767,7 +762,7 @@ static struct page *follow_huge_pud(struct vm_area_struct *vma,
>  static struct page *follow_huge_pmd(struct vm_area_struct *vma,
>  				    unsigned long addr, pmd_t *pmd,
>  				    unsigned int flags,
> -				    struct follow_page_context *ctx)
> +				    unsigned long *page_mask)
>  {
>  	return NULL;
>  }
> @@ -813,8 +808,7 @@ static inline bool can_follow_write_pte(pte_t pte, struct page *page,
>  }
>  
>  static struct page *follow_page_pte(struct vm_area_struct *vma,
> -		unsigned long address, pmd_t *pmd, unsigned int flags,
> -		struct dev_pagemap **pgmap)
> +		unsigned long address, pmd_t *pmd, unsigned int flags)
>  {
>  	struct mm_struct *mm = vma->vm_mm;
>  	struct folio *folio;
> @@ -912,7 +906,7 @@ static struct page *follow_page_pte(struct vm_area_struct *vma,
>  static struct page *follow_pmd_mask(struct vm_area_struct *vma,
>  				    unsigned long address, pud_t *pudp,
>  				    unsigned int flags,
> -				    struct follow_page_context *ctx)
> +				    unsigned long *page_mask)
>  {
>  	pmd_t *pmd, pmdval;
>  	spinlock_t *ptl;
> @@ -926,7 +920,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
>  	if (!pmd_present(pmdval))
>  		return no_page_table(vma, flags, address);
>  	if (likely(!pmd_leaf(pmdval)))
> -		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
> +		return follow_page_pte(vma, address, pmd, flags);
>  
>  	if (pmd_protnone(pmdval) && !gup_can_follow_protnone(vma, flags))
>  		return no_page_table(vma, flags, address);
> @@ -939,16 +933,16 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
>  	}
>  	if (unlikely(!pmd_leaf(pmdval))) {
>  		spin_unlock(ptl);
> -		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
> +		return follow_page_pte(vma, address, pmd, flags);
>  	}
>  	if (pmd_trans_huge(pmdval) && (flags & FOLL_SPLIT_PMD)) {
>  		spin_unlock(ptl);
>  		split_huge_pmd(vma, pmd, address);
>  		/* If pmd was left empty, stuff a page table in there quickly */
>  		return pte_alloc(mm, pmd) ? ERR_PTR(-ENOMEM) :
> -			follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
> +			follow_page_pte(vma, address, pmd, flags);
>  	}
> -	page = follow_huge_pmd(vma, address, pmd, flags, ctx);
> +	page = follow_huge_pmd(vma, address, pmd, flags, page_mask);
>  	spin_unlock(ptl);
>  	return page;
>  }
> @@ -956,7 +950,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
>  static struct page *follow_pud_mask(struct vm_area_struct *vma,
>  				    unsigned long address, p4d_t *p4dp,
>  				    unsigned int flags,
> -				    struct follow_page_context *ctx)
> +				    unsigned long *page_mask)
>  {
>  	pud_t *pudp, pud;
>  	spinlock_t *ptl;
> @@ -969,7 +963,7 @@ static struct page *follow_pud_mask(struct vm_area_struct *vma,
>  		return no_page_table(vma, flags, address);
>  	if (pud_leaf(pud)) {
>  		ptl = pud_lock(mm, pudp);
> -		page = follow_huge_pud(vma, address, pudp, flags, ctx);
> +		page = follow_huge_pud(vma, address, pudp, flags, page_mask);
>  		spin_unlock(ptl);
>  		if (page)
>  			return page;
> @@ -978,13 +972,13 @@ static struct page *follow_pud_mask(struct vm_area_struct *vma,
>  	if (unlikely(pud_bad(pud)))
>  		return no_page_table(vma, flags, address);
>  
> -	return follow_pmd_mask(vma, address, pudp, flags, ctx);
> +	return follow_pmd_mask(vma, address, pudp, flags, page_mask);
>  }
>  
>  static struct page *follow_p4d_mask(struct vm_area_struct *vma,
>  				    unsigned long address, pgd_t *pgdp,
>  				    unsigned int flags,
> -				    struct follow_page_context *ctx)
> +				    unsigned long *page_mask)
>  {
>  	p4d_t *p4dp, p4d;
>  
> @@ -995,7 +989,7 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
>  	if (!p4d_present(p4d) || p4d_bad(p4d))
>  		return no_page_table(vma, flags, address);
>  
> -	return follow_pud_mask(vma, address, p4dp, flags, ctx);
> +	return follow_pud_mask(vma, address, p4dp, flags, page_mask);
>  }
>  
>  /**
> @@ -1003,20 +997,16 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
>   * @vma: vm_area_struct mapping @address
>   * @address: virtual address to look up
>   * @flags: flags modifying lookup behaviour
> - * @ctx: contains dev_pagemap for %ZONE_DEVICE memory pinning and a
> - *       pointer to output page_mask
> + * @page_mask: a pointer to output page_mask
>   *
>   * @flags can have FOLL_ flags set, defined in <linux/mm.h>
>   *
> - * When getting pages from ZONE_DEVICE memory, the @ctx->pgmap caches
> - * the device's dev_pagemap metadata to avoid repeating expensive lookups.
> - *
>   * When getting an anonymous page and the caller has to trigger unsharing
>   * of a shared anonymous page first, -EMLINK is returned. The caller should
>   * trigger a fault with FAULT_FLAG_UNSHARE set. Note that unsharing is only
>   * relevant with FOLL_PIN and !FOLL_WRITE.
>   *
> - * On output, the @ctx->page_mask is set according to the size of the page.
> + * On output, @page_mask is set according to the size of the page.
>   *
>   * Return: the mapped (struct page *), %NULL if no mapping exists, or
>   * an error pointer if there is a mapping to something not represented
> @@ -1024,7 +1014,7 @@ static struct page *follow_p4d_mask(struct vm_area_struct *vma,
>   */
>  static struct page *follow_page_mask(struct vm_area_struct *vma,
>  			      unsigned long address, unsigned int flags,
> -			      struct follow_page_context *ctx)
> +			      unsigned long *page_mask)
>  {
>  	pgd_t *pgd;
>  	struct mm_struct *mm = vma->vm_mm;
> @@ -1032,13 +1022,13 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,
>  
>  	vma_pgtable_walk_begin(vma);
>  
> -	ctx->page_mask = 0;
> +	*page_mask = 0;
>  	pgd = pgd_offset(mm, address);
>  
>  	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
>  		page = no_page_table(vma, flags, address);
>  	else
> -		page = follow_p4d_mask(vma, address, pgd, flags, ctx);
> +		page = follow_p4d_mask(vma, address, pgd, flags, page_mask);
>  
>  	vma_pgtable_walk_end(vma);
>  
> @@ -1376,7 +1366,7 @@ static long __get_user_pages(struct mm_struct *mm,
>  {
>  	long ret = 0, i = 0;
>  	struct vm_area_struct *vma = NULL;
> -	struct follow_page_context ctx = { NULL };
> +	unsigned long page_mask = 0;
>  
>  	if (!nr_pages)
>  		return 0;
> @@ -1418,7 +1408,7 @@ static long __get_user_pages(struct mm_struct *mm,
>  						pages ? &page : NULL);
>  				if (ret)
>  					goto out;
> -				ctx.page_mask = 0;
> +				page_mask = 0;
>  				goto next_page;
>  			}
>  
> @@ -1441,7 +1431,7 @@ static long __get_user_pages(struct mm_struct *mm,
>  		}
>  		cond_resched();
>  
> -		page = follow_page_mask(vma, start, gup_flags, &ctx);
> +		page = follow_page_mask(vma, start, gup_flags, &page_mask);
>  		if (!page || PTR_ERR(page) == -EMLINK) {
>  			ret = faultin_page(vma, start, gup_flags,
>  					   PTR_ERR(page) == -EMLINK, locked);
> @@ -1474,7 +1464,7 @@ static long __get_user_pages(struct mm_struct *mm,
>  			goto out;
>  		}
>  next_page:
> -		page_increm = 1 + (~(start >> PAGE_SHIFT) & ctx.page_mask);
> +		page_increm = 1 + (~(start >> PAGE_SHIFT) & page_mask);
>  		if (page_increm > nr_pages)
>  			page_increm = nr_pages;
>  
> @@ -1524,8 +1514,6 @@ static long __get_user_pages(struct mm_struct *mm,
>  		nr_pages -= page_increm;
>  	} while (nr_pages);
>  out:
> -	if (ctx.pgmap)
> -		put_dev_pagemap(ctx.pgmap);
>  	return i ? i : ret;
>  }
>  
> @@ -2853,7 +2841,6 @@ static int gup_fast_pte_range(pmd_t pmd, pmd_t *pmdp, unsigned long addr,
>  		unsigned long end, unsigned int flags, struct page **pages,
>  		int *nr)
>  {
> -	struct dev_pagemap *pgmap = NULL;
>  	int ret = 0;
>  	pte_t *ptep, *ptem;
>  
> @@ -2926,8 +2913,6 @@ static int gup_fast_pte_range(pmd_t pmd, pmd_t *pmdp, unsigned long addr,
>  	ret = 1;
>  
>  pte_unmap:
> -	if (pgmap)
> -		put_dev_pagemap(pgmap);
>  	pte_unmap(ptem);
>  	return ret;
>  }

Re: [PATCH v2 1/2] mm/gup: Remove dead pgmap refcounting code

[John Hubbard]

On 9/5/25 5:56 PM, John Hubbard wrote:
...> Maybe I'll post a quick follow-up, in fact, to remove page_mask
> as well. No need to pass that around all over the place, I think
> it can instead be computed by the single place that actually
> consumes it.
> 

Almost, but not quite. The problem is in the -EEXIST return code
case for follow_page_mask(). In that case, we also need folio
size (PMD, PUD, or normal) to be returned. The page_mask argument
does that now, and as long as we are in C, it's awkward to do
anything else. 

oh well.

thanks,
-- 
John Hubbard