[PATCH V1] sched/debug: change "runnable tasks" to "Tasks on cpu" on sched debug

[PATCH V1] sched/debug: change "runnable tasks" to "Tasks on cpu" on sched debug

[zihan zhou]

In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
tasks" table, but not all tasks in the table are runnable.
It is inappropriate to refer to this table as "runnable tasks", so here it
is changed to "Tasks on CPU %d", like:

Tasks on cpu 31:
 S            task   PID       vruntime   eligible    deadline      
--------------------------------------------------------------------
 S        cpuhp/31   173         0.803286   E           2.153245    
 S    migration/31   174         8.167468   E          11.167468    

PS For the sake of clarity in reality, some table information has been
omitted here.


Signed-off-by: zihan zhou <15645113830zzh@gmail.com>
Signed-off-by: yaozhenguo <yaozhenguo@jd.com>
Signed-off-by: yaowenchao1 <yaowenchao@jd.com>
---
 kernel/sched/debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index a1be00a988bf..687487128400 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -771,7 +771,7 @@ static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu)
 	struct task_struct *g, *p;
 
 	SEQ_printf(m, "\n");
-	SEQ_printf(m, "runnable tasks:\n");
+	SEQ_printf(m, "Tasks on cpu %d:\n", rq_cpu);
 	SEQ_printf(m, " S            task   PID       vruntime   eligible    "
 		   "deadline             slice          sum-exec      switches  "
 		   "prio         wait-time        sum-sleep       sum-block"
-- 
2.33.0

Re: [PATCH V1] sched/debug: change "runnable tasks" to "Tasks on cpu" on sched debug

[Vincent Guittot]

On Fri, 17 Jan 2025 at 03:23, zihan zhou <15645113830zzh@gmail.com> wrote:
>
> In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
> tasks" table, but not all tasks in the table are runnable.
> It is inappropriate to refer to this table as "runnable tasks", so here it
> is changed to "Tasks on CPU %d", like:

We have used replaced runnable by queued in fair scheduler

>
> Tasks on cpu 31:
>  S            task   PID       vruntime   eligible    deadline
> --------------------------------------------------------------------
>  S        cpuhp/31   173         0.803286   E           2.153245
>  S    migration/31   174         8.167468   E          11.167468
>
> PS For the sake of clarity in reality, some table information has been
> omitted here.
>
>
> Signed-off-by: zihan zhou <15645113830zzh@gmail.com>
> Signed-off-by: yaozhenguo <yaozhenguo@jd.com>
> Signed-off-by: yaowenchao1 <yaowenchao@jd.com>
> ---
>  kernel/sched/debug.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
> index a1be00a988bf..687487128400 100644
> --- a/kernel/sched/debug.c
> +++ b/kernel/sched/debug.c
> @@ -771,7 +771,7 @@ static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu)
>         struct task_struct *g, *p;
>
>         SEQ_printf(m, "\n");
> -       SEQ_printf(m, "runnable tasks:\n");
> +       SEQ_printf(m, "Tasks on cpu %d:\n", rq_cpu);
>         SEQ_printf(m, " S            task   PID       vruntime   eligible    "
>                    "deadline             slice          sum-exec      switches  "
>                    "prio         wait-time        sum-sleep       sum-block"
> --
> 2.33.0
>

Re: [PATCH V1] sched/debug: change "runnable tasks" to "Tasks on cpu" on sched debug

[Peter Zijlstra]

On Fri, Jan 17, 2025 at 09:32:43AM +0100, Vincent Guittot wrote:
> On Fri, 17 Jan 2025 at 03:23, zihan zhou <15645113830zzh@gmail.com> wrote:
> >
> > In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
> > tasks" table, but not all tasks in the table are runnable.
> > It is inappropriate to refer to this table as "runnable tasks", so here it
> > is changed to "Tasks on CPU %d", like:
> 
> We have used replaced runnable by queued in fair scheduler

Right, but also 'tasks on cpu' is equally wrong -- but really, if you're
looking at sched/debug you have to know what you're doing anyway, so why
bother with trivial stuff like this?

> > Tasks on cpu 31:
> >  S            task   PID       vruntime   eligible    deadline
> > --------------------------------------------------------------------
> >  S        cpuhp/31   173         0.803286   E           2.153245
> >  S    migration/31   174         8.167468   E          11.167468
> >
> > PS For the sake of clarity in reality, some table information has been
> > omitted here.
> >
> >
> > Signed-off-by: zihan zhou <15645113830zzh@gmail.com>
> > Signed-off-by: yaozhenguo <yaozhenguo@jd.com>
> > Signed-off-by: yaowenchao1 <yaowenchao@jd.com>

And that's a broken SoB chain.

Re: [PATCH V2] sched: Forward deadline for early tick

[zihan zhou]

Thanks for your reply.

> > >
> > > In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
> > > tasks" table, but not all tasks in the table are runnable.
> > > It is inappropriate to refer to this table as "runnable tasks", so here it
> > > is changed to "Tasks on CPU %d", like:
> > 
> > We have used replaced runnable by queued in fair scheduler
> 
> Right, but also 'tasks on cpu' is equally wrong -- but really, if you're
> looking at sched/debug you have to know what you're doing anyway, so why
> bother with trivial stuff like this?

Sorry, I just wanted to participate more in the kernel community. While
debugging the scheduler, I found this wording error. It doesn't really
affect anything, it just looks a bit awkward.

By the way, I am confused why 'tasks on cpu' is wrong. I think that
the tasks in the table are on a certain CPU, maybe they are sleeping,
but at least they were running on this CPU at the last moment. 

Looking forward to your answer. Of course, this matter doesn't seem
important, and I will try to understand it myself in the future.

Re: [PATCH V1] sched/debug: change "runnable tasks" to "Tasks on cpu" on sched debug

[zihan zhou]

Thanks for your reply.

> > >
> > > In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
> > > tasks" table, but not all tasks in the table are runnable.
> > > It is inappropriate to refer to this table as "runnable tasks", so here it
> > > is changed to "Tasks on CPU %d", like:
> > 
> > We have used replaced runnable by queued in fair scheduler
> 
> Right, but also 'tasks on cpu' is equally wrong -- but really, if you're
> looking at sched/debug you have to know what you're doing anyway, so why
> bother with trivial stuff like this?

Sorry, I just wanted to participate more in the kernel community. While
debugging the scheduler, I found this wording error. It doesn't really
affect anything, it just looks a bit awkward.

By the way, I am confused why 'tasks on cpu' is wrong. I think that
the tasks in the table are on a certain CPU, maybe they are sleeping,
but at least they were running on this CPU at the last moment. 

Looking forward to your answer. Of course, this matter doesn't seem
important, and I will try to understand it myself in the future.

Re: [PATCH V2] sched: Forward deadline for early tick

[zihan zhou]

I'm sorry, this email was an incorrect one. Please ignore this email.

> Thanks for your reply.
>
> > > >
> > > > In sched debug file /sys/kernel/debug/sched/debug, there is a "runnable
> > > > tasks" table, but not all tasks in the table are runnable.
> > > > It is inappropriate to refer to this table as "runnable tasks", so here it
> > > > is changed to "Tasks on CPU %d", like:
> > > 
> > > We have used replaced runnable by queued in fair scheduler
> > 
> > Right, but also 'tasks on cpu' is equally wrong -- but really, if you're
> > looking at sched/debug you have to know what you're doing anyway, so why
> > bother with trivial stuff like this?
> 
> Sorry, I just wanted to participate more in the kernel community. While
> debugging the scheduler, I found this wording error. It doesn't really
> affect anything, it just looks a bit awkward.
> 
> By the way, I am confused why 'tasks on cpu' is wrong. I think that
> the tasks in the table are on a certain CPU, maybe they are sleeping,
> but at least they were running on this CPU at the last moment. 
> 
> Looking forward to your answer. Of course, this matter doesn't seem
> important, and I will try to understand it myself in the future.