Date:   Mon, 6 Mar 2023 12:06:09 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Zqiang <qiang1.zhang@intel.com>
Cc:     mingo@redhat.com, vincent.guittot@linaro.org,
        juri.lelli@redhat.com, dietmar.eggemann@arm.com,
        rostedt@goodmis.org, vschneid@redhat.com,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] sched/fair: Make hrtimer callback execute in the hard
 interrupt context
Message-ID: <20230306110609.GA1267364@hirez.programming.kicks-ass.net>
References: <20230306070938.3099273-1-qiang1.zhang@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230306070938.3099273-1-qiang1.zhang@intel.com>
Precedence: bulk

On Mon, Mar 06, 2023 at 03:09:38PM +0800, Zqiang wrote:
> The scheduler related hrtimers callback will be executed in softirqd
> context for PREEMPT_RT enabled kernel, this commit therefore mark
> hrtimers as harded even on PREEMPT_RT enabled kernels, there is no
> functional change.

There very much is a functional change, also your reasoning is
non-existant.

Also:

  https://lkml.kernel.org/r/Y1F5sDVGen7ZVW+U@hirez.programming.kicks-ass.net

> Signed-off-by: Zqiang <qiang1.zhang@intel.com>
> ---
>  kernel/sched/fair.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index c36aa54ae071..98c48d144089 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -5610,7 +5610,7 @@ static void start_cfs_slack_bandwidth(struct cfs_bandwidth *cfs_b)
>  
>  	hrtimer_start(&cfs_b->slack_timer,
>  			ns_to_ktime(cfs_bandwidth_slack_period),
> -			HRTIMER_MODE_REL);
> +			HRTIMER_MODE_REL_HARD);
>  }
>  
>  /* we know any runtime found here is valid as update_curr() precedes return */
> @@ -5813,9 +5813,9 @@ void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
>  	cfs_b->burst = 0;
>  
>  	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
> -	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
> +	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED_HARD);
>  	cfs_b->period_timer.function = sched_cfs_period_timer;
> -	hrtimer_init(&cfs_b->slack_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
> +	hrtimer_init(&cfs_b->slack_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
>  	cfs_b->slack_timer.function = sched_cfs_slack_timer;
>  	cfs_b->slack_started = false;
>  }
> @@ -5835,7 +5835,7 @@ void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
>  
>  	cfs_b->period_active = 1;
>  	hrtimer_forward_now(&cfs_b->period_timer, cfs_b->period);
> -	hrtimer_start_expires(&cfs_b->period_timer, HRTIMER_MODE_ABS_PINNED);
> +	hrtimer_start_expires(&cfs_b->period_timer, HRTIMER_MODE_ABS_PINNED_HARD);
>  }
>  
>  static void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
> -- 
> 2.25.1
>