mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
To: Linux Kernel <linux-kernel@vger.kernel.org>,
	Suresh B Siddha <suresh.b.siddha@intel.com>,
	Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Ingo Molnar <mingo@elte.hu>, Dipankar Sarma <dipankar@in.ibm.com>,
	Balbir Singh <balbir@linux.vnet.ibm.com>,
	Vatsa <vatsa@linux.vnet.ibm.com>,
	Gautham R Shenoy <ego@in.ibm.com>,
	Andi Kleen <andi@firstfloor.org>,
	David Collier-Brown <davecb@sun.com>,
	Tim Connors <tconnors@astro.swin.edu.au>,
	Max Krasnyansky <maxk@qualcomm.com>,
	Gregory Haskins <gregory.haskins@gmail.com>,
	Pavel Machek <pavel@suse.cz>,
	Andrew Morton <akpm@linux-foundation.org>,
	Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
Subject: [PATCH v7 6/8] sched: activate active load balancing in new idle cpus
Date: Thu, 18 Dec 2008 23:26:36 +0530	[thread overview]
Message-ID: <20081218175635.29812.82242.stgit@drishya.in.ibm.com> (raw)
In-Reply-To: <20081218175313.29812.4781.stgit@drishya.in.ibm.com>

Active load balancing is a process by which migration thread
is woken up on the target CPU in order to pull current
running task on another package into this newly idle
package.

This method is already in use with normal load_balance(),
this patch introduces this method to new idle cpus when
sched_mc is set to POWERSAVINGS_BALANCE_WAKEUP.

This logic provides effective consolidation of short running
daemon jobs in a almost idle system

The side effect of this patch may be ping-ponging of tasks
if the system is moderately utilised. May need to adjust the
iterations before triggering.

Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
---

 kernel/sched.c |   54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 54 insertions(+), 0 deletions(-)

diff --git a/kernel/sched.c b/kernel/sched.c
index 3415fa3..f82a7a0 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -3692,10 +3692,64 @@ redo:
 	}
 
 	if (!ld_moved) {
+		int active_balance;
+
 		schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]);
 		if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
 		    !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
 			return -1;
+
+		if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP)
+			return -1;
+
+		if (sd->nr_balance_failed++ < 2)
+			return -1;
+
+		/*
+		 * The only task running in a non-idle cpu can be moved to this
+		 * cpu in an attempt to completely freeup the other CPU
+		 * package. The same method used to move task in load_balance()
+		 * have been extended for load_balance_newidle() to speedup
+		 * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2)
+		 *
+		 * The package power saving logic comes from
+		 * find_busiest_group().  If there are no imbalance, then
+		 * f_b_g() will return NULL.  However when sched_mc={1,2} then
+		 * f_b_g() will select a group from which a running task may be
+		 * pulled to this cpu in order to make the other package idle.
+		 * If there is no opportunity to make a package idle and if
+		 * there are no imbalance, then f_b_g() will return NULL and no
+		 * action will be taken in load_balance_newidle().
+		 *
+		 * Under normal task pull operation due to imbalance, there
+		 * will be more than one task in the source run queue and
+		 * move_tasks() will succeed.  ld_moved will be true and this
+		 * active balance code will not be triggered.
+		 */
+
+		/* Lock busiest in correct order while this_rq is held */
+		double_lock_balance(this_rq, busiest);
+
+		/*
+		 * don't kick the migration_thread, if the curr
+		 * task on busiest cpu can't be moved to this_cpu
+		 */
+		if (!cpu_isset(this_cpu, busiest->curr->cpus_allowed)) {
+			double_unlock_balance(this_rq, busiest);
+			all_pinned = 1;
+			return ld_moved;
+		}
+
+		if (!busiest->active_balance) {
+			busiest->active_balance = 1;
+			busiest->push_cpu = this_cpu;
+			active_balance = 1;
+		}
+
+		double_unlock_balance(this_rq, busiest);
+		if (active_balance)
+			wake_up_process(busiest->migration_thread);
+
 	} else
 		sd->nr_balance_failed = 0;
 


  parent reply	other threads:[~2008-12-18 17:54 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-12-18 17:55 [PATCH v7 0/8] Tunable sched_mc_power_savings=n Vaidyanathan Srinivasan
2008-12-18 17:56 ` [PATCH v7 1/8] sched: convert BALANCE_FOR_xx_POWER to inline functions Vaidyanathan Srinivasan
2008-12-18 17:56 ` [PATCH v7 2/8] sched: Framework for sched_mc/smt_power_savings=N Vaidyanathan Srinivasan
2008-12-18 17:56 ` [PATCH v7 3/8] sched: favour lower logical cpu number for sched_mc balance Vaidyanathan Srinivasan
2008-12-18 17:56 ` [PATCH v7 4/8] sched: nominate preferred wakeup cpu Vaidyanathan Srinivasan
2008-12-18 18:12   ` Balbir Singh
2008-12-19 21:55   ` Andrew Morton
2008-12-19 22:19     ` Andrew Morton
2008-12-19 22:27       ` Ingo Molnar
2008-12-19 22:31         ` Ingo Molnar
2008-12-19 22:38           ` Andrew Morton
2008-12-19 22:54             ` Ingo Molnar
2008-12-20  4:36     ` Vaidyanathan Srinivasan
2008-12-20  4:44       ` Andrew Morton
2008-12-20  7:54         ` Ingo Molnar
2008-12-20 10:02         ` Vaidyanathan Srinivasan
2008-12-20 10:36           ` Vaidyanathan Srinivasan
2008-12-20 10:56             ` Vaidyanathan Srinivasan
2008-12-21  8:46               ` Ingo Molnar
2008-12-18 17:56 ` [PATCH v7 5/8] sched: bias task wakeups to preferred semi-idle packages Vaidyanathan Srinivasan
2008-12-18 18:11   ` Balbir Singh
2008-12-18 17:56 ` Vaidyanathan Srinivasan [this message]
2008-12-18 17:56 ` [PATCH v7 7/8] sched: add SD_BALANCE_NEWIDLE at MC and CPU level for sched_mc>0 Vaidyanathan Srinivasan
2008-12-18 17:56 ` [PATCH v7 8/8] sched: idle_balance() does not call load_balance_newidle() Vaidyanathan Srinivasan
2008-12-18 18:12   ` Balbir Singh
2008-12-18 20:17     ` Ingo Molnar
2008-12-18 20:19 ` [PATCH v7 0/8] Tunable sched_mc_power_savings=n Ingo Molnar
2008-12-18 20:31   ` Ingo Molnar
2008-12-19  8:29     ` Vaidyanathan Srinivasan
2008-12-19  8:24   ` Vaidyanathan Srinivasan
2008-12-19 13:34   ` Vaidyanathan Srinivasan
2008-12-29 23:43 ` MinChan Kim
2008-12-30  2:48   ` Balbir Singh
2008-12-30  6:21     ` Ingo Molnar
2008-12-30  6:44       ` Balbir Singh
2008-12-30  7:20         ` Ingo Molnar
2008-12-30 18:07       ` Vaidyanathan Srinivasan
2009-01-02  7:26         ` Vaidyanathan Srinivasan
2009-01-02 22:16           ` Ingo Molnar
2009-01-03  7:29             ` Mike Galbraith
2009-01-03 10:16               ` Vaidyanathan Srinivasan
2009-01-03 11:22                 ` Mike Galbraith
2009-01-04 15:00                   ` Mike Galbraith
2009-01-04 18:19                     ` Vaidyanathan Srinivasan
2009-01-04 19:52                       ` Mike Galbraith
2009-01-05  3:20                         ` Vaidyanathan Srinivasan
2009-01-05  4:40                           ` Mike Galbraith
2009-01-05  6:36                             ` Mike Galbraith
2009-01-05 15:19                               ` Mike Galbraith
2009-01-06  9:31                                 ` Mike Galbraith
2009-01-06 15:07                                   ` Vaidyanathan Srinivasan
2009-01-06 17:48                                     ` Mike Galbraith
2009-01-06 18:45                                       ` Balbir Singh
2009-01-07  8:59                                         ` Mike Galbraith
2009-01-07 11:26                                           ` Vaidyanathan Srinivasan
2009-01-07 14:36                                             ` Mike Galbraith
2009-01-07 15:35                                               ` Vaidyanathan Srinivasan
2009-01-08  8:06                                                 ` Mike Galbraith
2009-01-08 17:46                                                   ` Vaidyanathan Srinivasan
2009-01-09  6:00                                                     ` Mike Galbraith
2009-01-06 14:54                             ` Vaidyanathan Srinivasan
2008-12-30 17:31     ` Vaidyanathan Srinivasan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20081218175635.29812.82242.stgit@drishya.in.ibm.com \
    --to=svaidy@linux.vnet.ibm.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=akpm@linux-foundation.org \
    --cc=andi@firstfloor.org \
    --cc=balbir@linux.vnet.ibm.com \
    --cc=davecb@sun.com \
    --cc=dipankar@in.ibm.com \
    --cc=ego@in.ibm.com \
    --cc=gregory.haskins@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maxk@qualcomm.com \
    --cc=mingo@elte.hu \
    --cc=pavel@suse.cz \
    --cc=suresh.b.siddha@intel.com \
    --cc=tconnors@astro.swin.edu.au \
    --cc=vatsa@linux.vnet.ibm.com \
    --cc=venkatesh.pallipadi@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome