[PATCH v2] sched: pick_next_rt_entity(): checked list_entry

[PATCH v2] sched: pick_next_rt_entity(): checked list_entry

[Pietro Borrello]

Commit 326587b84078 ("sched: fix goto retry in pick_next_task_rt()")
removed any path which could make pick_next_rt_entity() return NULL.
However, BUG_ON(!rt_se) in _pick_next_task_rt() (the only caller of
pick_next_rt_entity()) still checks the error condition, which can
never happen, since list_entry() never returns NULL.
Remove the BUG_ON check, and instead emit a warning in the only
possible error condition here: the queue being empty which should
never happen.

Fixes: 326587b84078 ("sched: fix goto retry in pick_next_task_rt()")
Signed-off-by: Pietro Borrello <borrello@diag.uniroma1.it>
---
Changes in v2:
- pick_next_rt_entity(): emit warning instead of crashing
- Link to v1: https://lore.kernel.org/r/20230128-list-entry-null-check-sched-v1-1-c93085ee0055@diag.uniroma1.it
---
 kernel/sched/rt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index ed2a47e4ddae..c024529d8416 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1777,6 +1777,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rt_rq *rt_rq)
 	BUG_ON(idx >= MAX_RT_PRIO);
 
 	queue = array->queue + idx;
+	SCHED_WARN_ON(list_empty(queue));
 	next = list_entry(queue->next, struct sched_rt_entity, run_list);
 
 	return next;
@@ -1789,7 +1790,6 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq)
 
 	do {
 		rt_se = pick_next_rt_entity(rt_rq);
-		BUG_ON(!rt_se);
 		rt_rq = group_rt_rq(rt_se);
 	} while (rt_rq);
 

---
base-commit: 2241ab53cbb5cdb08a6b2d4688feb13971058f65
change-id: 20230128-list-entry-null-check-sched-a3f3dfd6d468

Best regards,
-- 
Pietro Borrello <borrello@diag.uniroma1.it>

Re: [PATCH v2] sched: pick_next_rt_entity(): checked list_entry

[Steven Rostedt]

On Tue, 31 Jan 2023 13:01:16 +0000
Pietro Borrello <borrello@diag.uniroma1.it> wrote:

> index ed2a47e4ddae..c024529d8416 100644
> --- a/kernel/sched/rt.c
> +++ b/kernel/sched/rt.c
> @@ -1777,6 +1777,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rt_rq *rt_rq)
>  	BUG_ON(idx >= MAX_RT_PRIO);
>  
>  	queue = array->queue + idx;
> +	SCHED_WARN_ON(list_empty(queue));

I wonder if we should make this:

	if (SCHED_WARN_ON(list_empty(queue)))
		return NULL;

>  	next = list_entry(queue->next, struct sched_rt_entity, run_list);
>  
>  	return next;
> @@ -1789,7 +1790,6 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq)
>  
>  	do {
>  		rt_se = pick_next_rt_entity(rt_rq);
> -		BUG_ON(!rt_se);

		if (unlikely(!rt_se))
			return NULL;

-- Steve

>  		rt_rq = group_rt_rq(rt_se);
>  	} while (rt_rq);
>  

Re: [PATCH v2] sched: pick_next_rt_entity(): checked list_entry

[Phil Auld]

On Mon, Feb 06, 2023 at 11:23:42AM -0500 Steven Rostedt wrote:
> On Tue, 31 Jan 2023 13:01:16 +0000
> Pietro Borrello <borrello@diag.uniroma1.it> wrote:
> 
> > index ed2a47e4ddae..c024529d8416 100644
> > --- a/kernel/sched/rt.c
> > +++ b/kernel/sched/rt.c
> > @@ -1777,6 +1777,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rt_rq *rt_rq)
> >  	BUG_ON(idx >= MAX_RT_PRIO);
> >  
> >  	queue = array->queue + idx;
> > +	SCHED_WARN_ON(list_empty(queue));
> 
> I wonder if we should make this:
> 
> 	if (SCHED_WARN_ON(list_empty(queue)))
> 		return NULL;
> 
> >  	next = list_entry(queue->next, struct sched_rt_entity, run_list);
> >  
> >  	return next;
> > @@ -1789,7 +1790,6 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq)
> >  
> >  	do {
> >  		rt_se = pick_next_rt_entity(rt_rq);
> > -		BUG_ON(!rt_se);
> 
> 		if (unlikely(!rt_se))
> 			return NULL;

I think that's better than taking a digger in one of the subsequent macros.


Cheers,
Phil


> 
> -- Steve
> 
> >  		rt_rq = group_rt_rq(rt_se);
> >  	} while (rt_rq);
> >  
>



-- 

Re: [PATCH v2] sched: pick_next_rt_entity(): checked list_entry

[Pietro Borrello]

On Mon, 6 Feb 2023 at 17:57, Phil Auld <pauld@redhat.com> wrote:
>
> On Mon, Feb 06, 2023 at 11:23:42AM -0500 Steven Rostedt wrote:
> > On Tue, 31 Jan 2023 13:01:16 +0000
> > Pietro Borrello <borrello@diag.uniroma1.it> wrote:
> >
> > > index ed2a47e4ddae..c024529d8416 100644
> > > --- a/kernel/sched/rt.c
> > > +++ b/kernel/sched/rt.c
> > > @@ -1777,6 +1777,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rt_rq *rt_rq)
> > >     BUG_ON(idx >= MAX_RT_PRIO);
> > >
> > >     queue = array->queue + idx;
> > > +   SCHED_WARN_ON(list_empty(queue));
> >
> > I wonder if we should make this:
> >
> >       if (SCHED_WARN_ON(list_empty(queue)))
> >               return NULL;
> >
> > >     next = list_entry(queue->next, struct sched_rt_entity, run_list);
> > >
> > >     return next;
> > > @@ -1789,7 +1790,6 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq)
> > >
> > >     do {
> > >             rt_se = pick_next_rt_entity(rt_rq);
> > > -           BUG_ON(!rt_se);
> >
> >               if (unlikely(!rt_se))
> >                       return NULL;
>
> I think that's better than taking a digger in one of the subsequent macros.
>

Thanks for the feedback.
Fixed in v3: https://lore.kernel.org/all/20230128-list-entry-null-check-sched-v3-1-b1a71bd1ac6b@diag.uniroma1.it/T/#u

Best regards,
Pietro