[v2,1/2] mm, hugetlbfs: introduce ->pagesize() to vm_operations_struct

Commit Message

Dave Jiang Feb. 10, 2018, 12:38 a.m. UTC

From: Dan Williams <[e-mail]>

When device-dax is operating in huge-page mode we want it to behave like
hugetlbfs and report the MMU page mapping size that is being enforced by
the vma. Similar to commit 31383c6865a5 "mm, hugetlbfs: introduce
->split() to vm_operations_struct" it would be messy to teach
vma_mmu_pagesize() about device-dax page mapping sizes in the same
(hstate) way that hugetlbfs communicates this attribute.  Instead, these
patches introduce a new ->pagesize() vm operation.

Cc: Benjamin Herrenschmidt <[e-mail]>
Cc: Paul Mackerras <[e-mail]>
Cc: Michael Ellerman <[e-mail]>
Reported-by: Jane Chu <[e-mail]>
Signed-off-by: Dan Williams <[e-mail]>
Signed-off-by: Dave Jiang <[e-mail]>
---
 arch/powerpc/mm/hugetlbpage.c |    5 +----
 include/linux/mm.h            |    1 +
 mm/hugetlb.c                  |   23 ++++++++++++-----------
 3 files changed, 14 insertions(+), 15 deletions(-)

Patch

diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c
index a9b9083c5e49..c6a2e577e842 100644
--- a/arch/powerpc/mm/hugetlbpage.c
+++ b/arch/powerpc/mm/hugetlbpage.c
@@ -568,10 +568,7 @@  unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
 	if (!radix_enabled())
 		return 1UL << mmu_psize_to_shift(psize);
 #endif
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	return huge_page_size(hstate_vma(vma));
+	return vma_kernel_pagesize(vma);
 }
 
 static inline bool is_power_of_4(unsigned long x)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index ea818ff739cd..37b9aef91ec7 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -383,6 +383,7 @@  struct vm_operations_struct {
 	int (*huge_fault)(struct vm_fault *vmf, enum page_entry_size pe_size);
 	void (*map_pages)(struct vm_fault *vmf,
 			pgoff_t start_pgoff, pgoff_t end_pgoff);
+	unsigned long (*pagesize)(struct vm_area_struct * area);
 
 	/* notification that a previously read-only page is about to become
 	 * writable, if an error is returned it will cause a SIGBUS */
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 9a334f5fb730..8fa069b5cb4d 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -637,14 +637,9 @@  EXPORT_SYMBOL_GPL(linear_hugepage_index);
  */
 unsigned long vma_kernel_pagesize(struct vm_area_struct *vma)
 {
-	struct hstate *hstate;
-
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	hstate = hstate_vma(vma);
-
-	return 1UL << huge_page_shift(hstate);
+	if (vma->vm_ops && vma->vm_ops->pagesize)
+		return vma->vm_ops->pagesize(vma);
+	return PAGE_SIZE;
 }
 EXPORT_SYMBOL_GPL(vma_kernel_pagesize);
 
@@ -654,12 +649,10 @@  EXPORT_SYMBOL_GPL(vma_kernel_pagesize);
  * architectures where it differs, an architecture-specific version of this
  * function is required.
  */
-#ifndef vma_mmu_pagesize
-unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
+__weak unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
 {
 	return vma_kernel_pagesize(vma);
 }
-#endif
 
 /*
  * Flags for MAP_PRIVATE reservations.  These are stored in the bottom
@@ -3132,6 +3125,13 @@  static int hugetlb_vm_op_split(struct vm_area_struct *vma, unsigned long addr)
 	return 0;
 }
 
+static unsigned long hugetlb_vm_op_pagesize(struct vm_area_struct *vma)
+{
+	struct hstate *hstate = hstate_vma(vma);
+
+	return 1UL << huge_page_shift(hstate);
+}
+
 /*
  * We cannot handle pagefaults against hugetlb pages at all.  They cause
  * handle_mm_fault() to try to instantiate regular-sized pages in the
@@ -3149,6 +3149,7 @@  const struct vm_operations_struct hugetlb_vm_ops = {
 	.open = hugetlb_vm_op_open,
 	.close = hugetlb_vm_op_close,
 	.split = hugetlb_vm_op_split,
+	.pagesize = hugetlb_vm_op_pagesize,
 };
 
 static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page,