Re: [PATCH v6 3/4] sched, psi: Don't account irq time if sched_clock_irqtime is disabled

[Yafang Shao]

On Wed, Dec 11, 2024 at 9:56 PM Michal Koutný <mkoutny@xxxxxxxx> wrote:
>
> On Wed, Dec 11, 2024 at 09:17:28PM GMT, Yafang Shao <laoar.shao@xxxxxxxxx> wrote:
> > @@ -1286,7 +1286,7 @@ struct psi_trigger *psi_trigger_create(struct psi_group *group, char *buf,
> >       bool privileged;
> >       u32 window_us;
> >
> > -     if (static_branch_likely(&psi_disabled))
> > +     if (static_branch_likely(&psi_disabled) || !irqtime_enabled())
> >               return ERR_PTR(-EOPNOTSUPP);
>
> Beware this jumps out for _any_ PSI metric when only irq is disabled.
> I meant to add a guard to psi_show() (this is psi_trigger_create()).

My apologies, I'll fix it in the next version. How about the following
change instead?

diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index 7341d33d9118..263c26a36511 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -1233,6 +1233,9 @@ int psi_show(struct seq_file *m, struct
psi_group *group, enum psi_res res)
        if (static_branch_likely(&psi_disabled))
                return -EOPNOTSUPP;

+       if (!irqtime_enabled() && res == PSI_IRQ)
+               return -EOPNOTSUPP;
+
        /* Update averages before reporting them */
        mutex_lock(&group->avgs_lock);
        now = sched_clock();



--
Regards
Yafang