mm/migration: fix possible do_pages_stat_array racing with memory offline

When follow_page peeks a page, the page could be migrated and then be
offlined while it's still being used by the do_pages_stat_array().  Use
FOLL_GET to hold the page refcnt to fix this potential race.

Link: https://lkml.kernel.org/r/20220318111709.60311-12-linmiaohe@huawei.com
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Acked-by: "Huang, Ying" <ying.huang@intel.com>
Reviewed-by: Muchun Song <songmuchun@bytedance.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/mm/migrate.c b/mm/migrate.c
index cb3fa7c24081388e3ab777e5ad7651fd8159a9c6..35f6416d29534dc8182ab07f542f13a019554aaf 100644 (file)
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1796,13 +1796,18 @@ static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages,
                        goto set_status;
 
                /* FOLL_DUMP to ignore special (like zero) pages */
-               page = follow_page(vma, addr, FOLL_DUMP);
+               page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP);
 
                err = PTR_ERR(page);
                if (IS_ERR(page))
                        goto set_status;
 
-               err = page ? page_to_nid(page) : -ENOENT;
+               if (page) {
+                       err = page_to_nid(page);
+                       put_page(page);
+               } else {
+                       err = -ENOENT;
+               }
 set_status:
                *status = err;