Received-SPF: pass (google.com: domain of linux-kernel+bounces-94339-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Precedence: bulk
MIME-Version: 1.0
References: <20240306102454.341014-1-sshegde@linux.ibm.com> <20240306102454.341014-3-sshegde@linux.ibm.com>
In-Reply-To: <20240306102454.341014-3-sshegde@linux.ibm.com>
From: Vincent Guittot <vincent.guittot@linaro.org>
Date: Wed, 6 Mar 2024 18:40:18 +0100
Message-ID: <CAKfTPtCyw7KLzwW6KEhva89hKF-yC91nARvx87wcP6GKGk_YKQ@mail.gmail.com>
Subject: Re: [PATCH v5 2/3] sched/fair: Use helper function to access rd->overutilized
To: Shrikanth Hegde <sshegde@linux.ibm.com>
Cc: mingo@kernel.org, peterz@infradead.org, yu.c.chen@intel.com, 
	dietmar.eggemann@arm.com, linux-kernel@vger.kernel.org, nysal@linux.ibm.com, 
	aboorvad@linux.ibm.com, srikar@linux.ibm.com, vschneid@redhat.com, 
	pierre.gondois@arm.com, qyousef@layalina.io
Content-Type: text/plain; charset="UTF-8"

On Wed, 6 Mar 2024 at 11:25, Shrikanth Hegde <sshegde@linux.ibm.com> wrote:
>
> Overutilized field is accessed directly in multiple places.
> So it could use a helper function. That way one might be more
> informed that it needs to be used only in case of EAS.
>
> No change in functionality intended.
>
> Reviewed-by: Qais Yousef <qyousef@layalina.io>
> Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
> ---
>  kernel/sched/fair.c | 20 +++++++++++++++++---
>  1 file changed, 17 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 997e570d9423..e3086c8930ea 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -6675,6 +6675,15 @@ static inline bool cpu_overutilized(int cpu)
>         return !util_fits_cpu(cpu_util_cfs(cpu), rq_util_min, rq_util_max, cpu);
>  }
>
> +/*
> + * Ensure that caller can do EAS. overutilized value
> + * make sense only if EAS is enabled
> + */
> +static inline int is_rd_overutilized(struct root_domain *rd)
> +{
> +       return READ_ONCE(rd->overutilized);
> +}
> +
>  static inline void set_rd_overutilized_status(struct root_domain *rd,
>                                               unsigned int status)
>  {
> @@ -6694,7 +6703,7 @@ static inline void check_update_overutilized_status(struct rq *rq)
>         if (!sched_energy_enabled())
>                 return;
>
> -       if (!READ_ONCE(rq->rd->overutilized) && cpu_overutilized(rq->cpu))
> +       if (!is_rd_overutilized(rq->rd) && cpu_overutilized(rq->cpu))
>                 set_rd_overutilized_status(rq->rd, SG_OVERUTILIZED);
>  }
>  #else
> @@ -6708,6 +6717,11 @@ static inline void set_rd_overutilized_status(struct root_domain *rd,
>  {
>         return 0;
>  }
> +
> +static inline int is_rd_overutilized(struct root_domain *rd)

I don't think that is_rd_overutilized() is called outside #ifdef
CONFIG_SMP so you can remove it.


> +{
> +       return 0;
> +}
>  #endif
>
>  /* Runqueue only has SCHED_IDLE tasks enqueued */
> @@ -7989,7 +8003,7 @@ static int find_energy_efficient_cpu(struct task_struct *p, int prev_cpu)
>
>         rcu_read_lock();
>         pd = rcu_dereference(rd->pd);
> -       if (!pd || READ_ONCE(rd->overutilized))
> +       if (!pd || is_rd_overutilized(rd))
>                 goto unlock;
>
>         /*
> @@ -10872,7 +10886,7 @@ static struct sched_group *find_busiest_group(struct lb_env *env)
>         if (sched_energy_enabled()) {
>                 struct root_domain *rd = env->dst_rq->rd;
>
> -               if (rcu_dereference(rd->pd) && !READ_ONCE(rd->overutilized))
> +               if (rcu_dereference(rd->pd) && !is_rd_overutilized(rd))
>                         goto out_balanced;
>         }
>
> --
> 2.39.3
>