Re: [PATCH 2/8] rcu: remove irq/softirq context check in rcu_read_unlock_special()

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Wed, Aug 07, 2013 at 06:24:58PM +0800, Lai Jiangshan wrote:
> After patch 10f39bb1, "special & RCU_READ_UNLOCK_BLOCKED" can't be true
> in irq nor softirq.(due to RCU_READ_UNLOCK_BLOCKED can only be set
> when preemption)
> 
> Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
> ---
>  kernel/rcutree_plugin.h |    6 ------
>  1 files changed, 0 insertions(+), 6 deletions(-)
> 
> diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
> index 8fd947e..54f7e45 100644
> --- a/kernel/rcutree_plugin.h
> +++ b/kernel/rcutree_plugin.h
> @@ -361,12 +361,6 @@ void rcu_read_unlock_special(struct task_struct *t)
>  		rcu_preempt_qs(smp_processor_id());
>  	}
> 
> -	/* Hardware IRQ handlers cannot block. */
> -	if (in_irq() || in_serving_softirq()) {
> -		local_irq_restore(flags);
> -		return;
> -	}
> -

Good point, it is time to relax the redundant checking.  Paranoid that
I am, I took an intermediate position, wrapping a WARN_ON_ONCE() around
the check as follows:

	if (WARN_ON_ONCE(in_irq() || in_serving_softirq())) {
		local_irq_restore(flags);
		return;
	}

If this warning never triggers over a period of some time, we can remove
it entirely.

I have queued this for 3.14 with your Signed-off-by.  Please let me know
if you have any objections.

							Thanx, Paul

>  	/* Clean up if blocked during RCU read-side critical section. */
>  	if (special & RCU_READ_UNLOCK_BLOCKED) {
>  		t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED;
> -- 
> 1.7.4.4
>