diff mbox series

[1/6] rcu/tasks: Check processor-ID assumptions

Message ID 20240802003426.4134196-1-paulmck@kernel.org (mailing list archive)
State Accepted
Commit c3554bba31640269e5aa4536809a71e525a260b3
Headers show
Series Add RCU Tasks diagnostics and documentation | expand

Commit Message

Paul E. McKenney Aug. 2, 2024, 12:34 a.m. UTC
The current mapping of smp_processor_id() to a CPU processing Tasks-RCU
callbacks makes some assumptions about layout.  This commit therefore
adds a WARN_ON() to check these assumptions.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
---
 kernel/rcu/tasks.h | 1 +
 1 file changed, 1 insertion(+)

Comments

Neeraj Upadhyay Aug. 14, 2024, 12:16 p.m. UTC | #1
Hi Paul,

On Thu, Aug 01, 2024 at 05:34:21PM -0700, Paul E. McKenney wrote:
> The current mapping of smp_processor_id() to a CPU processing Tasks-RCU
> callbacks makes some assumptions about layout.  This commit therefore
> adds a WARN_ON() to check these assumptions.
> 
> Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
> ---
>  kernel/rcu/tasks.h | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
> index b6fcf744af75d..d473846a572aa 100644
> --- a/kernel/rcu/tasks.h
> +++ b/kernel/rcu/tasks.h
> @@ -355,6 +355,7 @@ static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
>  	rcu_read_lock();
>  	ideal_cpu = smp_processor_id() >> READ_ONCE(rtp->percpu_enqueue_shift);
>  	chosen_cpu = cpumask_next(ideal_cpu - 1, cpu_possible_mask);
> +	WARN_ON_ONCE(chosen_cpu >= nr_cpu_ids);

I have changed this s/nr_cpu_ids/rcu_task_cpu_ids/ , as there is a
another commit 	fd70e9f1d85f "rcu-tasks: Fix access non-existent percpu
rtpcp variable in rcu_tasks_need_gpcb()" which is included in the tree
here [1]. Please let me know if something looks incorrect.


[1] https://git.kernel.org/pub/scm/linux/kernel/git/neeraj.upadhyay/linux-rcu.git/log/?h=next.14.08.24a


- Neeraj

>  	rtpcp = per_cpu_ptr(rtp->rtpcpu, chosen_cpu);
>  	if (!raw_spin_trylock_rcu_node(rtpcp)) { // irqs already disabled.
>  		raw_spin_lock_rcu_node(rtpcp); // irqs already disabled.
> -- 
> 2.40.1
>
Paul E. McKenney Aug. 14, 2024, 4:13 p.m. UTC | #2
On Wed, Aug 14, 2024 at 05:46:54PM +0530, Neeraj Upadhyay wrote:
> Hi Paul,
> 
> On Thu, Aug 01, 2024 at 05:34:21PM -0700, Paul E. McKenney wrote:
> > The current mapping of smp_processor_id() to a CPU processing Tasks-RCU
> > callbacks makes some assumptions about layout.  This commit therefore
> > adds a WARN_ON() to check these assumptions.
> > 
> > Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
> > ---
> >  kernel/rcu/tasks.h | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
> > index b6fcf744af75d..d473846a572aa 100644
> > --- a/kernel/rcu/tasks.h
> > +++ b/kernel/rcu/tasks.h
> > @@ -355,6 +355,7 @@ static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
> >  	rcu_read_lock();
> >  	ideal_cpu = smp_processor_id() >> READ_ONCE(rtp->percpu_enqueue_shift);
> >  	chosen_cpu = cpumask_next(ideal_cpu - 1, cpu_possible_mask);
> > +	WARN_ON_ONCE(chosen_cpu >= nr_cpu_ids);
> 
> I have changed this s/nr_cpu_ids/rcu_task_cpu_ids/ , as there is a
> another commit 	fd70e9f1d85f "rcu-tasks: Fix access non-existent percpu
> rtpcp variable in rcu_tasks_need_gpcb()" which is included in the tree
> here [1]. Please let me know if something looks incorrect.
> 
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/neeraj.upadhyay/linux-rcu.git/log/?h=next.14.08.24a

Good catch, looks good, and thank you!

							Thanx, Paul

> - Neeraj
> 
> >  	rtpcp = per_cpu_ptr(rtp->rtpcpu, chosen_cpu);
> >  	if (!raw_spin_trylock_rcu_node(rtpcp)) { // irqs already disabled.
> >  		raw_spin_lock_rcu_node(rtpcp); // irqs already disabled.
> > -- 
> > 2.40.1
> >
diff mbox series

Patch

diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
index b6fcf744af75d..d473846a572aa 100644
--- a/kernel/rcu/tasks.h
+++ b/kernel/rcu/tasks.h
@@ -355,6 +355,7 @@  static void call_rcu_tasks_generic(struct rcu_head *rhp, rcu_callback_t func,
 	rcu_read_lock();
 	ideal_cpu = smp_processor_id() >> READ_ONCE(rtp->percpu_enqueue_shift);
 	chosen_cpu = cpumask_next(ideal_cpu - 1, cpu_possible_mask);
+	WARN_ON_ONCE(chosen_cpu >= nr_cpu_ids);
 	rtpcp = per_cpu_ptr(rtp->rtpcpu, chosen_cpu);
 	if (!raw_spin_trylock_rcu_node(rtpcp)) { // irqs already disabled.
 		raw_spin_lock_rcu_node(rtpcp); // irqs already disabled.