rcu: Fix a performance regression.

Commit "rcu: Create RCU-specific workqueues with rescuers" switched RCU
to using local workqueses and removed power efficiency flag from them.

This caused a performance regression that can be observed in Geekbench 5
after enabling CONFIG_WQ_POWER_EFFICIENT_DEFAULT: score went down from
760/2500 to 620/2300 (single/multi core respectively).

Add WQ_POWER_EFFICIENT flag to avoid this regression.

Change-Id: I2c4f41faa55548f9e81a1c0cbe10703948062d89
This commit is contained in:
Yaroslav Furman 2019-12-11 03:06:26 +03:00 committed by Ksawlii
parent 5380adeb80
commit 9c41dd9383

View file

@ -4607,12 +4607,11 @@ void __init rcu_init(void)
}
/* Create workqueue for expedited GPs and for Tree SRCU. */
rcu_gp_wq = alloc_workqueue("rcu_gp", WQ_POWER_EFFICIENT | WQ_MEM_RECLAIM | WQ_UNBOUND, 0);
rcu_gp_wq = alloc_workqueue("rcu_gp", WQ_POWER_EFFICIENT | WQ_MEM_RECLAIM, 0);
WARN_ON(!rcu_gp_wq);
rcu_par_gp_wq = alloc_workqueue("rcu_par_gp", WQ_POWER_EFFICIENT | WQ_MEM_RECLAIM | WQ_UNBOUND, 0);
rcu_par_gp_wq = alloc_workqueue("rcu_par_gp", WQ_POWER_EFFICIENT | WQ_MEM_RECLAIM, 0);
WARN_ON(!rcu_par_gp_wq);
srcu_init();
/* Fill in default value for rcutree.qovld boot parameter. */
/* -After- the rcu_node ->lock fields are initialized! */
if (qovld < 0)