[PATCH V1 2/5] mm/hugetlb: fix memfd_pin_folios free_huge_pages leak

[Steve Sistare <steven.sistare@xxxxxxxxxx>]

memfd_pin_folios followed by unpin_folios fails to restore
free_huge_pages if the pages were not already faulted in, because the
folio refcount for pages created by memfd_alloc_folio never goes to 0.
memfd_pin_folios needs another folio_put to undo the folio_try_get
below:

memfd_alloc_folio()
  alloc_hugetlb_folio_nodemask()
    dequeue_hugetlb_folio_nodemask()
      dequeue_hugetlb_folio_node_exact()
        folio_ref_unfreeze(folio, 1);    ; adds 1 refcount
  folio_try_get()                        ; adds 1 refcount
  hugetlb_add_to_page_cache()            ; adds 512 refcount (on x86)

With the fix, after memfd_pin_folios + unpin_folios, the refcount for
the (unfaulted) page is 512, which is correct, as the refcount for a
faulted unpinned page is 513.

Fixes: 89c1905d9c14 ("mm/gup: introduce memfd_pin_folios() for pinning memfd folios")

Signed-off-by: Steve Sistare <steven.sistare@xxxxxxxxxx>
---
 mm/gup.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/mm/gup.c b/mm/gup.c
index 54d0dc3..5b92f1d 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -3618,7 +3618,7 @@ long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end,
 	pgoff_t start_idx, end_idx, next_idx;
 	struct folio *folio = NULL;
 	struct folio_batch fbatch;
-	struct hstate *h;
+	struct hstate *h = NULL;
 	long ret = -EINVAL;
 
 	if (start < 0 || start > end || !max_folios)
@@ -3662,6 +3662,8 @@ long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end,
 							     &fbatch);
 			if (folio) {
 				folio_put(folio);
+				if (h)
+					folio_put(folio);
 				folio = NULL;
 			}
 
-- 
1.8.3.1