mm/vmscan: don't try to reclaim freed folios

If folios were freed from under us, there's no need to reclaim them.  Skip
these folios to save lots of cpu cycles and avoid possible unnecessary
disk I/O.

Link: https://lkml.kernel.org/r/20220608141432.23258-1-linmiaohe@huawei.com
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 97ac6c6c026dcd9d28650dcad424719cd83a2f5d..cd2201ba6d7c822f2c8cd67373803479769615a9 100644 (file)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1610,13 +1610,19 @@ retry:
                folio = lru_to_folio(page_list);
                list_del(&folio->lru);
 
+               nr_pages = folio_nr_pages(folio);
+
+               if (folio_ref_count(folio) == 1 &&
+                   folio_ref_freeze(folio, 1)) {
+                       /* folio was freed from under us. So we are done. */
+                       goto free_it;
+               }
+
                if (!folio_trylock(folio))
                        goto keep;
 
                VM_BUG_ON_FOLIO(folio_test_active(folio), folio);
 
-               nr_pages = folio_nr_pages(folio);
-
                /* Account the number of base pages */
                sc->nr_scanned += nr_pages;