[PATCH v2 0/5] close various race windows for swap

From: Miaohe Lin <linmiaohe@huawei.com>
To: <akpm@linux-foundation.org>
Cc: <ying.huang@intel.com>, <dennis@kernel.org>,
	<tim.c.chen@linux.intel.com>, <hughd@google.com>,
	<hannes@cmpxchg.org>, <mhocko@suse.com>, <iamjoonsoo.kim@lge.com>,
	<alexs@kernel.org>, <david@redhat.com>, <minchan@kernel.org>,
	<richard.weiyang@gmail.com>, <linux-kernel@vger.kernel.org>,
	<linux-mm@kvack.org>, <linmiaohe@huawei.com>
Subject: [PATCH v2 0/5] close various race windows for swap
Date: Sat, 17 Apr 2021 05:40:34 -0400	[thread overview]
Message-ID: <20210417094039.51711-1-linmiaohe@huawei.com> (raw)

Hi all,
When I was investigating the swap code, I found some possible race
windows. This series aims to fix all these races. But using current
get/put_swap_device() to guard against concurrent swapoff for
swap_readpage() looks terrible because swap_readpage() may take really
long time. And to reduce the performance overhead on the hot-path as
much as possible, it appears we can use the percpu_ref to close this
race window(as suggested by Huang, Ying). The patch 1 adds percpu_ref
support for swap and most of the remaining patches try to use this to
close various race windows. More details can be found in the respective
changelogs. Thanks!

v1->v2:
  reorganize the patch-2/5
  various enhance and fixup per Huang, Ying
  Many thanks for the comments of Huang, Ying, Dennis Zhou and Tim Chen.

Miaohe Lin (5):
  mm/swapfile: add percpu_ref support for swap
  mm/swapfile: use percpu_ref to serialize against concurrent swapoff
  swap: fix do_swap_page() race with swapoff
  mm/swap: remove confusing checking for non_swap_entry() in
    swap_ra_info()
  mm/shmem: fix shmem_swapin() race with swapoff

 include/linux/swap.h | 15 +++++++--
 mm/memory.c          |  9 ++++++
 mm/shmem.c           |  6 ++++
 mm/swap_state.c      |  6 ----
 mm/swapfile.c        | 74 +++++++++++++++++++++++++++-----------------
 5 files changed, 73 insertions(+), 37 deletions(-)

-- 
2.19.1