Re: [Patch v3 1/7] sched/pelt.c: Add support to track thermal pressure

[Vincent Guittot]

Hi Thara,

On Mon, 14 Oct 2019 at 02:58, Thara Gopinath <thara.gopinath@xxxxxxxxxx> wrote:
>
> Extrapolating on the exisitng framework to track rt/dl utilization using

s/exisitng/existing/

> pelt signals, add a similar mechanism to track thermal pressue. The

s/pessure/pressure/

> difference here from rt/dl utilization tracking is that, instead of
> tracking time spent by a cpu running a rt/dl task through util_avg,
> the average thermal pressure is tracked through load_avg.

It would be good to mention why you use load_avg field instead of
util_avg field: because the signal is weighted with the capped
capacity and is not binary
And also explained a bit what capacity refer to

> In order to track average thermal pressure, a new sched_avg variable
> avg_thermal is introduced. Function update_thermal_avg can be called
> to do the periodic bookeeping (accumulate, decay and average)
> of the thermal pressure.
>
> Signed-off-by: Thara Gopinath <thara.gopinath@xxxxxxxxxx>
> ---
>  kernel/sched/pelt.c  | 13 +++++++++++++
>  kernel/sched/pelt.h  |  7 +++++++
>  kernel/sched/sched.h |  1 +
>  3 files changed, 21 insertions(+)
>
> diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c
> index a96db50..f06aae3 100644
> --- a/kernel/sched/pelt.c
> +++ b/kernel/sched/pelt.c
> @@ -353,6 +353,19 @@ int update_dl_rq_load_avg(u64 now, struct rq *rq, int running)
>         return 0;
>  }
>
> +int update_thermal_avg(u64 now, struct rq *rq, u64 capacity)

All the other functions are named :
update_cfs_rq/rt_rq/dl_rq/irq_load_avg

might be good to keep similar name with update_thermal_load_avg

> +{
> +       if (___update_load_sum(now, &rq->avg_thermal,
> +                              capacity,
> +                              capacity,
> +                              capacity)) {
> +               ___update_load_avg(&rq->avg_thermal, 1, 1);
> +               return 1;
> +       }
> +
> +       return 0;
> +}
> +
>  #ifdef CONFIG_HAVE_SCHED_AVG_IRQ
>  /*
>   * irq:
> diff --git a/kernel/sched/pelt.h b/kernel/sched/pelt.h
> index afff644..01c5436 100644
> --- a/kernel/sched/pelt.h
> +++ b/kernel/sched/pelt.h
> @@ -6,6 +6,7 @@ int __update_load_avg_se(u64 now, struct cfs_rq *cfs_rq, struct sched_entity *se
>  int __update_load_avg_cfs_rq(u64 now, struct cfs_rq *cfs_rq);
>  int update_rt_rq_load_avg(u64 now, struct rq *rq, int running);
>  int update_dl_rq_load_avg(u64 now, struct rq *rq, int running);
> +int update_thermal_avg(u64 now, struct rq *rq, u64 capacity);
>
>  #ifdef CONFIG_HAVE_SCHED_AVG_IRQ
>  int update_irq_load_avg(struct rq *rq, u64 running);
> @@ -175,6 +176,12 @@ update_rq_clock_pelt(struct rq *rq, s64 delta) { }
>  static inline void
>  update_idle_rq_clock_pelt(struct rq *rq) { }
>
> +static inline int

can you keep some function ordering as above and move
update_thermal_avg() just after update_dl_rq_load_avg


> +update_thermal_avg(u64 now, struct rq *rq, u64 capacity)
> +{
> +       return 0;
> +}
> +
>  #endif
>
>
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index 0db2c1b..d5d82c8 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -944,6 +944,7 @@ struct rq {
>  #ifdef CONFIG_HAVE_SCHED_AVG_IRQ
>         struct sched_avg        avg_irq;
>  #endif
> +       struct sched_avg        avg_thermal;
>         u64                     idle_stamp;
>         u64                     avg_idle;
>
> --
> 2.1.4
>