[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190729085711.GQ25636@localhost.localdomain>
Date: Mon, 29 Jul 2019 10:57:11 +0200
From: Juri Lelli <juri.lelli@...hat.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: mingo@...nel.org, linux-kernel@...r.kernel.org,
dietmar.eggemann@....com, luca.abeni@...tannapisa.it,
bristot@...hat.com, balsini@...roid.com, dvyukov@...gle.com,
tglx@...utronix.de, vpillai@...italocean.com, rostedt@...dmis.org
Subject: Re: [RFC][PATCH 01/13] sched/deadline: Impose global limits on
sched_attr::sched_period
Hi,
On 26/07/19 16:54, Peter Zijlstra wrote:
>
> Cc: Daniel Bristot de Oliveira <bristot@...hat.com>
> Cc: Luca Abeni <luca.abeni@...tannapisa.it>
> Cc: Juri Lelli <juri.lelli@...hat.com>
> Cc: Dmitry Vyukov <dvyukov@...gle.com>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
> ---
> include/linux/sched/sysctl.h | 3 +++
> kernel/sched/deadline.c | 23 +++++++++++++++++++++--
> kernel/sysctl.c | 14 ++++++++++++++
> 3 files changed, 38 insertions(+), 2 deletions(-)
>
> --- a/include/linux/sched/sysctl.h
> +++ b/include/linux/sched/sysctl.h
> @@ -56,6 +56,9 @@ int sched_proc_update_handler(struct ctl
> extern unsigned int sysctl_sched_rt_period;
> extern int sysctl_sched_rt_runtime;
>
> +extern unsigned int sysctl_sched_dl_period_max;
> +extern unsigned int sysctl_sched_dl_period_min;
> +
> #ifdef CONFIG_UCLAMP_TASK
> extern unsigned int sysctl_sched_uclamp_util_min;
> extern unsigned int sysctl_sched_uclamp_util_max;
> --- a/kernel/sched/deadline.c
> +++ b/kernel/sched/deadline.c
> @@ -2597,6 +2597,14 @@ void __getparam_dl(struct task_struct *p
> }
>
> /*
> + * Default limits for DL period; on the top end we guard against small util
> + * tasks still getting rediculous long effective runtimes, on the bottom end we
s/rediculous/ridiculous/
> + * guard against timer DoS.
> + */
> +unsigned int sysctl_sched_dl_period_max = 1 << 22; /* ~4 seconds */
> +unsigned int sysctl_sched_dl_period_min = 100; /* 100 us */
These limits look sane to me. I've actually been experimenting with 10us
period tasks and throttling seemed to behave fine, but I guess 100us is
a saner default.
So, (with a few lines of changelog :)
Acked-by: Juri Lelli <juri.lelli@...hat.com>
Thanks for doing this.
Best,
Juri
Powered by blists - more mailing lists