workqueue: UNBOUND -> REBIND morphing in rebind_workers() should be atomic

The compiler may compile the following code into TWO write/modify
instructions.

	worker->flags &= ~WORKER_UNBOUND;
	worker->flags |= WORKER_REBIND;

so the other CPU may temporarily see worker->flags which doesn't have
either WORKER_UNBOUND or WORKER_REBIND set and perform local wakeup
prematurely.

Fix it by using single explicit assignment via ACCESS_ONCE().

Because idle workers have another WORKER_NOT_RUNNING flag, this bug
doesn't exist for them; however, update it to use the same pattern for
consistency.

tj: Applied the change to idle workers too and updated comments and
    patch description a bit.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@vger.kernel.org

kernel/workqueue.c

@@ -1396,12 +1396,15 @@ static void rebind_workers(struct global_cwq *gcwq)
 	/* set REBIND and kick idle ones, we'll wait for these later */
 	for_each_worker_pool(pool, gcwq) {
 		list_for_each_entry(worker, &pool->idle_list, entry) {
+			unsigned long worker_flags = worker->flags;
+
 			if (worker->flags & WORKER_REBIND)
 				continue;
 
-			/* morph UNBOUND to REBIND */
-			worker->flags &= ~WORKER_UNBOUND;
-			worker->flags |= WORKER_REBIND;
+			/* morph UNBOUND to REBIND atomically */
+			worker_flags &= ~WORKER_UNBOUND;
+			worker_flags |= WORKER_REBIND;
+			ACCESS_ONCE(worker->flags) = worker_flags;
 
 			idle_rebind.cnt++;
 			worker->idle_rebind = &idle_rebind;
@@ -1434,10 +1437,12 @@ static void rebind_workers(struct global_cwq *gcwq)
 	/* rebind busy workers */
 	for_each_busy_worker(worker, i, pos, gcwq) {
 		struct work_struct *rebind_work = &worker->rebind_work;
+		unsigned long worker_flags = worker->flags;
 
-		/* morph UNBOUND to REBIND */
-		worker->flags &= ~WORKER_UNBOUND;
-		worker->flags |= WORKER_REBIND;
+		/* morph UNBOUND to REBIND atomically */
+		worker_flags &= ~WORKER_UNBOUND;
+		worker_flags |= WORKER_REBIND;
+		ACCESS_ONCE(worker->flags) = worker_flags;
 
 		if (test_and_set_bit(WORK_STRUCT_PENDING_BIT,
 				     work_data_bits(rebind_work)))