On Tue, 2022-02-22 at 11:47 -0300, Marcelo Tosatti wrote:
> @@ -918,14 +917,23 @@ atomic_t lru_disable_count = ATOMIC_INIT
> void lru_cache_disable(void)
> {
> atomic_inc(&lru_disable_count);
> + synchronize_rcu();
> #ifdef CONFIG_SMP
> /*
> - * lru_add_drain_all in the force mode will schedule draining on
> - * all online CPUs so any calls of lru_cache_disabled wrapped by
> - * local_lock or preemption disabled would be ordered by that.
> - * The atomic operation doesn't need to have stronger ordering
> - * requirements because that is enforced by the scheduling
> - * guarantees.
> + * synchronize_rcu() waits for preemption disabled
> + * and RCU read side critical sections
> + * For the users of lru_disable_count:
> + *
> + * preempt_disable, local_irq_disable() [bh_lru_lock()]
> + * rcu_read_lock [lru_pvecs CONFIG_PREEMPT_RT]
> + * preempt_disable [lru_pvecs !CONFIG_PREEMPT_RT]
> + *
> + *
> + * so any calls of lru_cache_disabled wrapped by
> + * local_lock+rcu_read_lock or preemption disabled would be
> + * ordered by that. The atomic operation doesn't need to have
> + * stronger ordering requirements because that is enforced
> + * by the scheduling guarantees.
"The atomic operation doesn't need to have stronger ordering requirements
because that is enforced by the scheduling guarantees."
This is no longer needed.
Regards,
--
Nicolás Sáenz
