if pwq_adjust_max_active() changes max_active from 0 to saved_max_active,
it needs to wakeup worker. This action is already done by thaw_workqueues().
if pwq_adjust_max_active() increase the max_active for unbound wq,
it also needs to wakeup worker. This action is missing.
To make these two cases happy, we move kicking workers code from
thaw_workqueues() to pwq_adjust_max_active().
It also makes thaw_workqueues() simpler.
Signed-off-by: Lai Jiangshan <la...@cn.fujitsu.com>
---
kernel/workqueue.c | 15 ++++++++-------
1 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 8c882ae..fb81159 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -3600,6 +3600,14 @@ static void pwq_adjust_max_active(struct pool_workqueue
*pwq)
while (!list_empty(&pwq->delayed_works) &&
pwq->nr_active < pwq->max_active)
pwq_activate_first_delayed(pwq);
+
+ /*
+ * Need to wake up worker in any of these cases:
+ * wq is just thawed
+ * unbound wq's max_active is just increased
+ * But this function is slowpath, wake up worker unconditionally
+ */
+ wake_up_worker(pwq->pool);
}
spin_unlock(&pwq->pool->lock);
@@ -4401,13 +4409,6 @@ void thaw_workqueues(void)
}
spin_unlock_irq(&pwq_lock);
- /* kick workers */
- for_each_pool(pool, pi) {
- spin_lock_irq(&pool->lock);
- wake_up_worker(pool);
- spin_unlock_irq(&pool->lock);
- }
-
workqueue_freezing = false;
out_unlock:
mutex_unlock(&wq_mutex);
--
1.7.7.6
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
