Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 6f05c315 authored by Se Wang (Patrick) Oh's avatar Se Wang (Patrick) Oh
Browse files

workqueue: Fix workqueue stall issue after cpu down failure 



When the hotplug notifier call chain with CPU_DOWN_PREPARE
is broken before reaching workqueue_cpu_down_callback(),
rebind_workers() adds WORKER_REBOUND flag for running workers.
Hence, the nr_running of the pool is not increased when scheduler
wakes up the worker. The fix is skipping adding WORKER_REBOUND
flag when the worker doesn't have WORKER_UNBOUND flag in
CPU_DOWN_FAILED path.

Change-Id: I2528e9154f4913d9ec14b63adbcbcd1eaa8a8452
Signed-off-by: default avatarSe Wang (Patrick) Oh <sewango@codeaurora.org>
parent 75d530af

kernel/workqueue.c

+13 −6
Original line number Diff line number Diff line
@@ -4494,10 +4494,13 @@ static void wq_unbind_fn(struct work_struct *work) 
/**

 * rebind_workers - rebind all workers of a pool to the associated CPU

 * @pool: pool of interest

 * @force: if it is true, replace WORKER_UNBOUND with WORKER_REBOUND

 * irrespective of flags of workers. Otherwise, replace the flags only

 * when workers have WORKER_UNBOUND flag.

 *

 * @pool->cpu is coming online.  Rebind all workers to the CPU.

 */

static void rebind_workers(struct worker_pool *pool)

static void rebind_workers(struct worker_pool *pool, bool force)

{

	struct worker *worker;
@@ -4546,11 +4549,13 @@ static void rebind_workers(struct worker_pool *pool) 
		 * fail incorrectly leading to premature concurrency

		 * management operations.

		 */

		if (force || (worker_flags & WORKER_UNBOUND)) {

			WARN_ON_ONCE(!(worker_flags & WORKER_UNBOUND));

			worker_flags |= WORKER_REBOUND;

			worker_flags &= ~WORKER_UNBOUND;

			ACCESS_ONCE(worker->flags) = worker_flags;

		}

	}



	spin_unlock_irq(&pool->lock);

}
@@ -4618,7 +4623,9 @@ static int workqueue_cpu_up_callback(struct notifier_block *nfb, 
			mutex_lock(&pool->attach_mutex);



			if (pool->cpu == cpu)

				rebind_workers(pool);

				rebind_workers(pool,

					(action & ~CPU_TASKS_FROZEN)

						!= CPU_DOWN_FAILED);

			else if (pool->cpu < 0)

				restore_unbound_workers_cpumask(pool, cpu);