[RFC PATCH v3 1/3] workqueue: Hold wq_pool_mutex while affining tasks to wq_unbound_cpumask

[Valentin Schneider]

When unbind_workers() reads wq_unbound_cpumask to set the affinity of
freshly-unbound kworkers, it only holds wq_pool_attach_mutex. This isn't
sufficient as wq_unbound_cpumask is only protected by wq_pool_mutex.

This is made more obvious as of recent commit

  46a4d679ef88 ("workqueue: Avoid a false warning in unbind_workers()")

e.g.

unbind_workers()                             workqueue_set_unbound_cpumask()
  kthread_set_per_cpu(p, -1);
  if (cpumask_intersects(wq_unbound_cpumask, cpu_active_mask))
					       cpumask_copy(wq_unbound_cpumask, cpumask);
    WARN_ON_ONCE(set_cpus_allowed_ptr(worker->task, wq_unbound_cpumask) < 0);

Make workqueue_offline_cpu() invoke unbind_workers() with wq_pool_mutex
held.

Fixes: 10a5a651e3af ("workqueue: Restrict kworker in the offline CPU pool running on housekeeping CPUs")
Signed-off-by: Valentin Schneider <vschneid@xxxxxxxxxx>
---
 kernel/workqueue.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index aa8a82bc6738..97cc41430a76 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -5143,14 +5143,15 @@ int workqueue_offline_cpu(unsigned int cpu)
 	if (WARN_ON(cpu != smp_processor_id()))
 		return -1;
 
+	mutex_lock(&wq_pool_mutex);
+
 	unbind_workers(cpu);
 
 	/* update NUMA affinity of unbound workqueues */
-	mutex_lock(&wq_pool_mutex);
 	list_for_each_entry(wq, &workqueues, list)
 		wq_update_unbound_numa(wq, cpu, false);
-	mutex_unlock(&wq_pool_mutex);
 
+	mutex_unlock(&wq_pool_mutex);
 	return 0;
 }
 
-- 
2.31.1