[PATCH 1/2 V4] workqueue: rescuer_thread() processes all pwqs before exit

[Lai Jiangshan]

Before the rescuer is picked to running, the works of the @pwq
may be processed by some other workers, and destroy_workqueue()
may called at the same time. This may result a nasty situation
that rescuer may exit with non-empty mayday list.

It is no harm currently, destroy_workqueue() can safely to free
them all(workqueue&pwqs) togerther, since the rescuer is stopped.
No rescuer nor mayday-timer can access the mayday list.

But it is nasty and error-prone in future development. Fix it.

Signed-off-by: Lai Jiangshan <laijs@xxxxxxxxxxxxxx>
---
 kernel/workqueue.c |   21 +++++++++++++++------
 1 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 0ee63af..7539244 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -2398,6 +2398,7 @@ static int rescuer_thread(void *__rescuer)
 	struct worker *rescuer = __rescuer;
 	struct workqueue_struct *wq = rescuer->rescue_wq;
 	struct list_head *scheduled = &rescuer->scheduled;
+	bool should_stop;
 
 	set_user_nice(current, RESCUER_NICE_LEVEL);
 
@@ -2408,12 +2409,14 @@ static int rescuer_thread(void *__rescuer)
 	rescuer->task->flags |= PF_WQ_WORKER;
 repeat:
 	set_current_state(TASK_INTERRUPTIBLE);
-
-	if (kthread_should_stop()) {
-		__set_current_state(TASK_RUNNING);
-		rescuer->task->flags &= ~PF_WQ_WORKER;
-		return 0;
-	}
+	/*
+	 * When the rescuer is requested to stop, the workqueue has no
+	 * work pending, but wq->maydays may still have pwq(s) queued.
+	 * This can happend when some other workers process all works
+	 * before this rescuer is scheduled. The rescuer must process
+	 * all pwq(s) before exit.
+	 */
+	should_stop = kthread_should_stop();
 
 	/* see whether any pwq is asking for help */
 	spin_lock_irq(&wq_mayday_lock);
@@ -2459,6 +2462,12 @@ repeat:
 
 	spin_unlock_irq(&wq_mayday_lock);
 
+	if (should_stop) {
+		__set_current_state(TASK_RUNNING);
+		rescuer->task->flags &= ~PF_WQ_WORKER;
+		return 0;
+	}
+
 	/* rescuers should never participate in concurrency management */
 	WARN_ON_ONCE(!(rescuer->flags & WORKER_NOT_RUNNING));
 	schedule();
-- 
1.7.4.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/