Re: [RFC PATCH v4 4/6] sched/cpufreq: Introduce sugov_cpu_ramp_boost

[Peter Zijlstra <peterz@infradead.org>]

On Wed, Jan 22, 2020 at 05:35:36PM +0000, Douglas RAILLARD wrote:

> +static unsigned long sugov_cpu_ramp_boost_update(struct sugov_cpu *sg_cpu)
> +{
> +	struct rq *rq = cpu_rq(sg_cpu->cpu);
> +	unsigned long util_est_enqueued;
> +	unsigned long util_avg;
> +	unsigned long boost = 0;
> +

Should we NO-OP this function when !sched_feat(UTIL_EST) ?

> +	util_est_enqueued = READ_ONCE(rq->cfs.avg.util_est.enqueued);

Otherwise you're reading garbage here, no?

> +	util_avg = READ_ONCE(rq->cfs.avg.util_avg);
> +
> +	/*
> +	 * Boost when util_avg becomes higher than the previous stable
> +	 * knowledge of the enqueued tasks' set util, which is CPU's
> +	 * util_est_enqueued.
> +	 *
> +	 * We try to spot changes in the workload itself, so we want to
> +	 * avoid the noise of tasks being enqueued/dequeued. To do that,
> +	 * we only trigger boosting when the "amount of work" enqueued
> +	 * is stable.
> +	 */
> +	if (util_est_enqueued == sg_cpu->util_est_enqueued &&
> +	    util_avg >= sg_cpu->util_avg &&
> +	    util_avg > util_est_enqueued)
> +		boost = util_avg - util_est_enqueued;
> +
> +	sg_cpu->util_est_enqueued = util_est_enqueued;
> +	sg_cpu->util_avg = util_avg;
> +	WRITE_ONCE(sg_cpu->ramp_boost, boost);
> +	return boost;
> +}