[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220318111709.60311-12-linmiaohe@huawei.com>
Date: Fri, 18 Mar 2022 19:17:09 +0800
From: Miaohe Lin <linmiaohe@...wei.com>
To: <akpm@...ux-foundation.org>
CC: <ziy@...dia.com>, <baolin.wang@...ux.alibaba.com>,
<ying.huang@...el.com>, <songmuchun@...edance.com>,
<apopple@...dia.com>, <linux-mm@...ck.org>,
<linux-kernel@...r.kernel.org>, <linmiaohe@...wei.com>
Subject: [PATCH v2 11/11] 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.
Signed-off-by: Miaohe Lin <linmiaohe@...wei.com>
---
mm/migrate.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index dbd91fbdb127..cd85ba0ab592 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1807,13 +1807,18 @@ 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;
--
2.23.0
Powered by blists - more mailing lists