Re: [RFC PATCH 03/23] sched/core: Initialize the class of a new task

[Ricardo Neri]

On Mon, Sep 26, 2022 at 02:57:29PM +0000, Joel Fernandes wrote:
> Hi Ricardo,
> 
> On Fri, Sep 09, 2022 at 04:11:45PM -0700, Ricardo Neri wrote:
> > New tasks shall start life as unclassified. They will be classified by
> > hardware when they run.
> > 
> > Cc: Ben Segall <bsegall@xxxxxxxxxx>
> > Cc: Daniel Bristot de Oliveira <bristot@xxxxxxxxxx>
> > Cc: Dietmar Eggemann <dietmar.eggemann@xxxxxxx>
> > Cc: Len Brown <len.brown@xxxxxxxxx>
> > Cc: Mel Gorman <mgorman@xxxxxxx>
> > Cc: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> > Cc: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
> > Cc: Tim C. Chen <tim.c.chen@xxxxxxxxx>
> > Cc: Valentin Schneider <vschneid@xxxxxxxxxx>
> > Cc: x86@xxxxxxxxxx
> > Cc: linux-kernel@xxxxxxxxxxxxxxx
> > Signed-off-by: Ricardo Neri <ricardo.neri-calderon@xxxxxxxxxxxxxxx>
> > ---
> >  kernel/sched/core.c | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index ee28253c9ac0..db548c1a25ef 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -4336,6 +4336,9 @@ static void __sched_fork(unsigned long clone_flags, struct task_struct *p)
> >  	p->se.prev_sum_exec_runtime	= 0;
> >  	p->se.nr_migrations		= 0;
> >  	p->se.vruntime			= 0;
> > +#ifdef CONFIG_SCHED_TASK_CLASSES
> > +	p->class			= TASK_CLASS_UNCLASSIFIED;
> > +#endif
> 
> I find the term 'class' very broad and unclear what kind of class (without
> further reading). So I am worried about how this generic term usage plays
> with Linux source code in the long-term (like what if someone else comes up
> with a usage of term 'class' that is unrelated to IPC.)

Thank you very much for your review Joel! Yes, class seems too generic. It is
meant to read, for instance, task_struct::class or p->class, or rq->current->
class. This should imply that we are referring to the class of a task. But yes,
I agree that it is too generic.

> 
> To that end, I was wondering if it could be renamed to p->ipc_class, and
> CONFIG_SCHED_TASK_IPC_CLASSES, or something.

This is a good suggestion. I will take it, unless others disagree.

BR,
Ricardo