Received-SPF: pass (google.com: domain of linux-kernel+bounces-45052-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Tue, 30 Jan 2024 18:38:22 +0100
From: Vincent Guittot <vincent.guittot@linaro.org>
To: Qais Yousef <qyousef@layalina.io>
Cc: Ingo Molnar <mingo@kernel.org>, Peter Zijlstra <peterz@infradead.org>,
	"Rafael J. Wysocki" <rafael@kernel.org>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
	Lukasz Luba <lukasz.luba@arm.com>, Wei Wang <wvw@google.com>,
	Rick Yiu <rickyiu@google.com>, Chung-Kai Mei <chungkai@google.com>
Subject: Re: [PATCH v2 8/8] sched/pelt: Introduce PELT multiplier
Message-ID: <Zbk0DhibX0oDLk1s@vingu-book>
References: <20231208002342.367117-1-qyousef@layalina.io>
 <20231208002342.367117-9-qyousef@layalina.io>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20231208002342.367117-9-qyousef@layalina.io>

Le vendredi 08 d?c. 2023 ? 00:23:42 (+0000), Qais Yousef a ?crit :
> From: Vincent Donnefort <vincent.donnefort@arm.com>
> 
> The new sched_pelt_multiplier boot param allows a user to set a clock
> multiplier to x2 or x4 (x1 being the default). This clock multiplier
> artificially speeds up PELT ramp up/down similarly to use a faster
> half-life than the default 32ms.
> 
>   - x1: 32ms half-life
>   - x2: 16ms half-life
>   - x4: 8ms  half-life
> 
> Internally, a new clock is created: rq->clock_task_mult. It sits in the
> clock hierarchy between rq->clock_task and rq->clock_pelt.
> 
> The param is set as read only and can only be changed at boot time via
> 
> 	kernel.sched_pelt_multiplier=[1, 2, 4]
> 
> PELT has a big impact on the overall system response and reactiveness to
> change. Smaller PELT HF means it'll require less time to reach the
> maximum performance point of the system when the system become fully
> busy; and equally shorter time to go back to lowest performance point
> when the system goes back to idle.
> 
> This faster reaction impacts both dvfs response and migration time
> between clusters in HMP system.
> 
> Smaller PELT values are expected to give better performance at the cost
> of more power. Under powered systems can particularly benefit from
> smaller values. Powerful systems can still benefit from smaller values
> if they want to be tuned towards perf more and power is not the major
> concern for them.
> 
> This combined with respone_time_ms from schedutil should give the user
> and sysadmin a deterministic way to control the triangular power, perf
> and thermals for their system. The default response_time_ms will half
> as PELT HF halves.
> 
> Update approximate_{util_avg, runtime}() to take into account the PELT
> HALFLIFE multiplier.
> 
> Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
> Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
> [Converted from sysctl to boot param and updated commit message]
> Signed-off-by: Qais Yousef (Google) <qyousef@layalina.io>
> ---
>  kernel/sched/core.c  |  2 +-
>  kernel/sched/pelt.c  | 52 ++++++++++++++++++++++++++++++++++++++++++--
>  kernel/sched/pelt.h  | 42 +++++++++++++++++++++++++++++++----
>  kernel/sched/sched.h |  1 +
>  4 files changed, 90 insertions(+), 7 deletions(-)
> 

..

> +__read_mostly unsigned int sched_pelt_lshift;
> +static unsigned int sched_pelt_multiplier = 1;
> +
> +static int set_sched_pelt_multiplier(const char *val, const struct kernel_param *kp)
> +{
> +	int ret;
> +
> +	ret = param_set_int(val, kp);
> +	if (ret)
> +		goto error;
> +
> +	switch (sched_pelt_multiplier)  {
> +	case 1:
> +		fallthrough;
> +	case 2:
> +		fallthrough;
> +	case 4:
> +		WRITE_ONCE(sched_pelt_lshift,
> +			   sched_pelt_multiplier >> 1);
> +		break;
> +	default:
> +		ret = -EINVAL;
> +		goto error;
> +	}
> +
> +	return 0;
> +
> +error:
> +	sched_pelt_multiplier = 1;
> +	return ret;
> +}
> +
> +static const struct kernel_param_ops sched_pelt_multiplier_ops = {
> +	.set = set_sched_pelt_multiplier,
> +	.get = param_get_int,
> +};
> +
> +#ifdef MODULE_PARAM_PREFIX
> +#undef MODULE_PARAM_PREFIX
> +#endif
> +/* XXX: should we use sched as prefix? */
> +#define MODULE_PARAM_PREFIX "kernel."
> +module_param_cb(sched_pelt_multiplier, &sched_pelt_multiplier_ops, &sched_pelt_multiplier, 0444);
> +MODULE_PARM_DESC(sched_pelt_multiplier, "PELT HALFLIFE helps control the responsiveness of the system.");
> +MODULE_PARM_DESC(sched_pelt_multiplier, "Accepted value: 1 32ms PELT HALIFE - roughly 200ms to go from 0 to max performance point (default).");
> +MODULE_PARM_DESC(sched_pelt_multiplier, "                2 16ms PELT HALIFE - roughly 100ms to go from 0 to max performance point.");
> +MODULE_PARM_DESC(sched_pelt_multiplier, "                4  8ms PELT HALIFE - roughly  50ms to go from 0 to max performance point.");
> +
>  /*
>   * Approximate the new util_avg value assuming an entity has continued to run
>   * for @delta us.

..

> +
>  static inline void
> -update_rq_clock_pelt(struct rq *rq, s64 delta) { }
> +update_rq_clock_task_mult(struct rq *rq, s64 delta) { }
>  
>  static inline void
>  update_idle_rq_clock_pelt(struct rq *rq) { }
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index bbece0eb053a..a7c89c623250 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -1029,6 +1029,7 @@ struct rq {
>  	u64			clock;
>  	/* Ensure that all clocks are in the same cache line */
>  	u64			clock_task ____cacheline_aligned;
> +	u64			clock_task_mult;

I'm not sure that we want yet another clock and this doesn't apply for irq_avg.

What about the below is simpler and I think cover all cases ?

diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c
index f951c44f1d52..5cdd147b7abe 100644
--- a/kernel/sched/pelt.c
+++ b/kernel/sched/pelt.c
@@ -180,6 +180,7 @@ static __always_inline int
 ___update_load_sum(u64 now, struct sched_avg *sa,
 		  unsigned long load, unsigned long runnable, int running)
 {
+	int time_shift;
 	u64 delta;

 	delta = now - sa->last_update_time;
@@ -195,12 +196,17 @@ ___update_load_sum(u64 now, struct sched_avg *sa,
 	/*
 	 * Use 1024ns as the unit of measurement since it's a reasonable
 	 * approximation of 1us and fast to compute.
+	 * On top of this, we can change the half-time period from the default
+	 * 32ms to a shorter value. This is equivalent to left shifting the
+	 * time.
+	 * Merge both right and left shifts in one single right shift
 	 */
-	delta >>= 10;
+	time_shift = 10 - sched_pelt_lshift;
+	delta >>= time_shift;
 	if (!delta)
 		return 0;

-	sa->last_update_time += delta << 10;
+	sa->last_update_time += delta << time_shift;

 	/*
 	 * running is a subset of runnable (weight) so running can't be set if


>  	u64			clock_pelt;
>  	unsigned long		lost_idle_time;
>  	u64			clock_pelt_idle;
> -- 
> 2.34.1
>