* Vlastimil Babka [off-list ref] [211209 10:54]:

On 12/1/21 15:29, Liam Howlett wrote:

quoted

From: "Matthew Wilcox (Oracle)" <willy@infradead.org>

This simplifies the implementation and is faster than using the linked
list.

You mean faster because more cache friendly?
OTOH you do max(addr, vma->vm_start) which wasn't the case.

I doubt it matters though and yeah it's much simpler...

quoted

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>

Acked-by: Vlastimil Babka <redacted>

Btw, no signed-off-by Liam?

I now also notice some other patches (e.g. patch 4) have From: and s-o-b by
Liam followed by Matthew, despite it's Liam sending them. I guess that's
less of an issue than the missing one.

Anyway something for you to check and make consistent in the whole series, I
won't point it out individually.

Thanks for catching this.  I will clean up the s-o-b for the next
revision.

quoted

---
 mm/mmap.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)

diff --git a/mm/mmap.c b/mm/mmap.c
index 9fee6e6b276f..de78fc0ce809 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c

@@ -669,29 +669,19 @@ munmap_vma_range(struct mm_struct *mm, unsigned long start, unsigned long len,
 
 	return 0;
 }
+
 static unsigned long count_vma_pages_range(struct mm_struct *mm,
 		unsigned long addr, unsigned long end)
 {
-	unsigned long nr_pages = 0;
+	VMA_ITERATOR(vmi, mm, addr);
 	struct vm_area_struct *vma;
+	unsigned long nr_pages = 0;
 
-	/* Find first overlapping mapping */
-	vma = find_vma_intersection(mm, addr, end);
-	if (!vma)
-		return 0;
-
-	nr_pages = (min(end, vma->vm_end) -
-		max(addr, vma->vm_start)) >> PAGE_SHIFT;
-
-	/* Iterate over the rest of the overlaps */
-	for (vma = vma->vm_next; vma; vma = vma->vm_next) {
-		unsigned long overlap_len;
-
-		if (vma->vm_start > end)
-			break;
+	for_each_vma_range(vmi, vma, end) {
+		unsigned long vm_start = max(addr, vma->vm_start);
+		unsigned long vm_end = min(end, vma->vm_end);
 
-		overlap_len = min(end, vma->vm_end) - vma->vm_start;
-		nr_pages += overlap_len >> PAGE_SHIFT;
+		nr_pages += (vm_end - vm_start) / PAGE_SIZE;
 	}
 
 	return nr_pages;