[047/192] writeback, cgroup: add smp_mb() to cgroup_writeback_umount()

Commit Message

Andrew Morton June 29, 2021, 2:35 a.m. UTC

From: Roman Gushchin <[e-mail]>
Subject: writeback, cgroup: add smp_mb() to cgroup_writeback_umount()

A full memory barrier is required between clearing SB_ACTIVE flag in
generic_shutdown_super() and checking isw_nr_in_flight in
cgroup_writeback_umount(), otherwise a new switch operation might be
scheduled after atomic_read(&isw_nr_in_flight) returned 0.  This would
result in a non-flushed isw_wq, and a potential crash.

The problem hasn't yet been seen in the real life and was discovered by
Jan Kara by looking into the code.

Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Roman Gushchin <[e-mail]>
Suggested-by: Jan Kara <[e-mail]>
Reviewed-by: Jan Kara <[e-mail]>
Cc: Alexander Viro <[e-mail]>
Cc: Dave Chinner <[e-mail]>
Cc: Dennis Zhou <[e-mail]>
Cc: Jan Kara <[e-mail]>
Cc: Tejun Heo <[e-mail]>
Cc: Jens Axboe <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>
---

 fs/fs-writeback.c |    6 ++++++
 1 file changed, 6 insertions(+)

Patch

--- a/fs/fs-writeback.c~writeback-cgroup-add-smp_mb-to-cgroup_writeback_umount
+++ a/fs/fs-writeback.c
@@ -1000,6 +1000,12 @@  out_bdi_put:
  */
 void cgroup_writeback_umount(void)
 {
+	/*
+	 * SB_ACTIVE should be reliably cleared before checking
+	 * isw_nr_in_flight, see generic_shutdown_super().
+	 */
+	smp_mb();
+
 	if (atomic_read(&isw_nr_in_flight)) {
 		/*
 		 * Use rcu_barrier() to wait for all pending callbacks to