kthread_worker: check all delayed works when destroy kthread worker

When destroying a kthread worker warn if there are still some pending
delayed works.  This indicates that the caller should clear all pending
delayed works before destroying the kthread worker.

Link: https://lkml.kernel.org/r/20230104144230.938521-1-qiang1.zhang@intel.com
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/kernel/kthread.c b/kernel/kthread.c
index f97fd01a29325f6d7f2e6d04ad5f87cb9d3577f3..7e6751b29101e5ca4050aa512b922ab05ddfc5e2 100644 (file)
--- a/kernel/kthread.c
+++ b/kernel/kthread.c
@@ -1382,6 +1382,10 @@ EXPORT_SYMBOL_GPL(kthread_flush_worker);
  * Flush and destroy @worker.  The simple flush is enough because the kthread
  * worker API is used only in trivial scenarios.  There are no multi-step state
  * machines needed.
+ *
+ * Note that this function is not responsible for handling delayed work, so
+ * caller should be responsible for queuing or canceling all delayed work items
+ * before invoke this function.
  */
 void kthread_destroy_worker(struct kthread_worker *worker)
 {
@@ -1393,6 +1397,7 @@ void kthread_destroy_worker(struct kthread_worker *worker)
 
        kthread_flush_worker(worker);
        kthread_stop(task);
+       WARN_ON(!list_empty(&worker->delayed_work_list));
        WARN_ON(!list_empty(&worker->work_list));
        kfree(worker);
 }