Re: [PATCH 19/23] sched,rcuperf: Convert to sched_set_fifo_low()

[Paul E. McKenney]

On Wed, Apr 22, 2020 at 01:27:38PM +0200, Peter Zijlstra wrote:
> Because SCHED_FIFO is a broken scheduler model (see previous patches)
> take away the priority field, the kernel can't possibly make an
> informed decision.
> 
> Effectively no change.
> 
> Cc: paulmck@xxxxxxxxxx
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
> Reviewed-by: Ingo Molnar <mingo@xxxxxxxxxx>

Reviewed-by: Paul E. McKenney <paulmck@xxxxxxxxxx>

> ---
>  kernel/rcu/rcuperf.c |    8 ++------
>  1 file changed, 2 insertions(+), 6 deletions(-)
> 
> --- a/kernel/rcu/rcuperf.c
> +++ b/kernel/rcu/rcuperf.c
> @@ -353,7 +353,6 @@ rcu_perf_writer(void *arg)
>  	int i_max;
>  	long me = (long)arg;
>  	struct rcu_head *rhp = NULL;
> -	struct sched_param sp;
>  	bool started = false, done = false, alldone = false;
>  	u64 t;
>  	u64 *wdp;
> @@ -362,8 +361,7 @@ rcu_perf_writer(void *arg)
>  	VERBOSE_PERFOUT_STRING("rcu_perf_writer task started");
>  	WARN_ON(!wdpp);
>  	set_cpus_allowed_ptr(current, cpumask_of(me % nr_cpu_ids));
> -	sp.sched_priority = 1;
> -	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
> +	sched_set_fifo_low(current);
>  
>  	if (holdoff)
>  		schedule_timeout_uninterruptible(holdoff * HZ);
> @@ -419,9 +417,7 @@ rcu_perf_writer(void *arg)
>  			started = true;
>  		if (!done && i >= MIN_MEAS) {
>  			done = true;
> -			sp.sched_priority = 0;
> -			sched_setscheduler_nocheck(current,
> -						   SCHED_NORMAL, &sp);
> +			sched_set_normal(current, 0);
>  			pr_alert("%s%s rcu_perf_writer %ld has %d measurements\n",
>  				 perf_type, PERF_FLAG, me, MIN_MEAS);
>  			if (atomic_inc_return(&n_rcu_perf_writer_finished) >=
> 
>