Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757400Ab2KIBvi (ORCPT ); Thu, 8 Nov 2012 20:51:38 -0500 Received: from mx1.redhat.com ([209.132.183.28]:37155 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757066Ab2KIBvh (ORCPT ); Thu, 8 Nov 2012 20:51:37 -0500 Date: Thu, 8 Nov 2012 19:51:13 -0600 From: Clark Williams To: Peter Zijlstra Cc: Thomas Gleixner , Ingo Molnar , LKML , Steven Rostedt Subject: [PATCH v2] sched: add a tuning knob to allow changing RR tmeslice Message-ID: <20121108195113.2dd51d43@redhat.com> Organization: Red Hat, Inc Mime-Version: 1.0 Content-Type: multipart/signed; micalg=PGP-SHA1; boundary="Sig_/pO7WSUl+BFUNZ_qvNRWh/6y"; protocol="application/pgp-signature" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5100 Lines: 169 --Sig_/pO7WSUl+BFUNZ_qvNRWh/6y Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable This version stores the user-input value in a separate location from the jiffies values used by the scheduler, to prevent a race condition. Subject: [PATCH v2] sched: add a tuning knob to allow changing RR timeslice User wanted a facility simliar to the ability on Solaris to adjust the SCHED_RR timeslice value. Add a /proc/sys/kernel scheduler knob named sched_rr_timeslice_ms which allows global changing of the SCHED_RR timeslice value. User visable value is in milliseconds but is stored as jiffies. Setting to 0 (zero) resets to the default (currently 100ms). Patch against tip/master, currently 3.7-rc3. Signed-off-by: Clark Williams --- include/linux/sched.h | 7 ++++++- kernel/sched/core.c | 30 ++++++++++++++++++++++++++++++ kernel/sched/rt.c | 4 ++-- kernel/sysctl.c | 7 +++++++ 4 files changed, 45 insertions(+), 3 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 03be150..1e2f38a 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2071,11 +2071,16 @@ static inline unsigned int get_sysctl_timer_migration(void) #endif /* CONFIG_SCHED_DEBUG */ extern unsigned int sysctl_sched_rt_period; extern int sysctl_sched_rt_runtime; - +extern int sysctl_sched_rr_timeslice_ms; +extern int sched_rr_timeslice; int sched_rt_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos); =20 +int sched_rr_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos); + #ifdef CONFIG_SCHED_AUTOGROUP extern unsigned int sysctl_sched_autogroup_enabled; =20 diff --git a/kernel/sched/core.c b/kernel/sched/core.c index c2e077c..318f617 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -288,6 +288,17 @@ __read_mostly int scheduler_running; int sysctl_sched_rt_runtime =3D 950000; =20 =20 +/* + * SCHED_RR timeslice in jiffies + * + */ +int sched_rr_timeslice =3D RR_TIMESLICE; +int sysctl_sched_rr_timeslice_ms; + +static inline void init_rr_timeslice() +{ + sysctl_sched_rr_timeslice_ms =3D msecs_to_jiffies(RR_TIMESLICE); +} =20 /* * __task_rq_lock - lock the rq @p resides on. @@ -6863,6 +6874,8 @@ void __init sched_init(void) init_rt_bandwidth(&def_rt_bandwidth, global_rt_period(), global_rt_runtime()); =20 + init_rr_timeslice(); + #ifdef CONFIG_RT_GROUP_SCHED init_rt_bandwidth(&root_task_group.rt_bandwidth, global_rt_period(), global_rt_runtime()); @@ -7543,6 +7556,23 @@ int sched_rt_handler(struct ctl_table *table, int write, return ret; } =20 +int sched_rr_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos) +{ + int ret; + static DEFINE_MUTEX(mutex); + + mutex_lock(&mutex); + ret =3D proc_dointvec(table, write, buffer, lenp, ppos); + /* make sure we maintain jiffies internally */ + if (!ret && write) + sched_rr_timeslice =3D (sysctl_sched_rr_timeslice_ms <=3D 0) ? + RR_TIMESLICE : msecs_to_jiffies(sysctl_sched_rr_timeslice_ms); + mutex_unlock(&mutex); + return ret; +} + #ifdef CONFIG_CGROUP_SCHED =20 /* return corresponding task_group object of a cgroup */ diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 418feb0..71aa6d0 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -2010,7 +2010,7 @@ static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued) if (--p->rt.time_slice) return; =20 - p->rt.time_slice =3D RR_TIMESLICE; + p->rt.time_slice =3D sched_rr_timeslice; =20 /* * Requeue to the end of queue if we (and all of our ancestors) are the @@ -2041,7 +2041,7 @@ static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) * Time slice is 0 for SCHED_FIFO tasks */ if (task->policy =3D=3D SCHED_RR) - return RR_TIMESLICE; + return sched_rr_timeslice; else return 0; } diff --git a/kernel/sysctl.c b/kernel/sysctl.c index b769d25..9fa0885 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -403,6 +403,13 @@ static struct ctl_table kern_table[] =3D { .mode =3D 0644, .proc_handler =3D sched_rt_handler, }, + { + .procname =3D "sched_rr_timeslice_ms", + .data =3D &sysctl_sched_rr_timeslice_ms, + .maxlen =3D sizeof(int), + .mode =3D 0644, + .proc_handler =3D sched_rr_handler, + }, #ifdef CONFIG_SCHED_AUTOGROUP { .procname =3D "sched_autogroup_enabled", --=20 1.7.11.7 --Sig_/pO7WSUl+BFUNZ_qvNRWh/6y Content-Type: application/pgp-signature; name=signature.asc Content-Disposition: attachment; filename=signature.asc -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) iEYEARECAAYFAlCcYZEACgkQHyuj/+TTEp1MdgCgofISpWr2WOryOfc3bKxieO2s fHEAnjHLU72id97bZI4BCt9Xo/Nte8C8 =h2cg -----END PGP SIGNATURE----- --Sig_/pO7WSUl+BFUNZ_qvNRWh/6y-- -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/