Re: [PATCH] rcu: Fix race in set and clear TICK_DEP_BIT_RCU_EXP bitmask

[Paul E. McKenney]

On Sat, Dec 31, 2022 at 07:25:08PM +0100, Frederic Weisbecker wrote:
> On Wed, Dec 21, 2022 at 12:08:49PM -0800, Paul E. McKenney wrote:
> > diff --git a/kernel/rcu/tree_exp.h b/kernel/rcu/tree_exp.h
> > index 249c2967d9e6c..7cc4856da0817 100644
> > --- a/kernel/rcu/tree_exp.h
> > +++ b/kernel/rcu/tree_exp.h
> > @@ -594,6 +594,7 @@ static void synchronize_rcu_expedited_wait(void)
> >  	struct rcu_data *rdp;
> >  	struct rcu_node *rnp;
> >  	struct rcu_node *rnp_root = rcu_get_root();
> > +	unsigned long flags;
> >  
> >  	trace_rcu_exp_grace_period(rcu_state.name, rcu_exp_gp_seq_endval(), TPS("startwait"));
> >  	jiffies_stall = rcu_exp_jiffies_till_stall_check();
> > @@ -602,17 +603,17 @@ static void synchronize_rcu_expedited_wait(void)
> >  		if (synchronize_rcu_expedited_wait_once(1))
> >  			return;
> >  		rcu_for_each_leaf_node(rnp) {
> > +			raw_spin_lock_irqsave_rcu_node(rnp, flags);
> >  			mask = READ_ONCE(rnp->expmask);
> >  			for_each_leaf_node_cpu_mask(rnp, cpu, mask) {
> >  				rdp = per_cpu_ptr(&rcu_data, cpu);
> >  				if (rdp->rcu_forced_tick_exp)
> >  					continue;
> >  				rdp->rcu_forced_tick_exp = true;
> > -				preempt_disable();
> >  				if (cpu_online(cpu))
> >  					tick_dep_set_cpu(cpu, TICK_DEP_BIT_RCU_EXP);
> > -				preempt_enable();
> >  			}
> > +			raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
> >  		}
> >  		j = READ_ONCE(jiffies_till_first_fqs);
> >  		if (synchronize_rcu_expedited_wait_once(j + HZ))
> 
> Reviewed-by: Frederic Weisbecker <frederic@xxxxxxxxxx>

Thank you!

> BTW why are we forcing the tick on the whole node?

Now that you mention it, that would be more precise.

> And shouldn't we set the tick dependency from rcu_exp_handler() instead?

Because it never occurred to me to check whether this could be invoked
from an interrupt handler?  ;-)

But that does sound like it might be a better approach.

Zqiang, would you be willing to look into this?

							Thanx, Paul