From: Gautham R Shenoy <ego@in.ibm.com>
To: linux-tip-commits@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, ego@in.ibm.com, hpa@zytor.com,
mingo@redhat.com, a.p.zijlstra@chello.nl,
dhaval@linux.vnet.ibm.com, balbir@in.ibm.com,
bharata@linux.vnet.ibm.com, suresh.b.siddha@intel.com,
tglx@linutronix.de, mingo@elte.hu, nickpiggin@yahoo.com.au
Subject: [tip:sched/balancing] sched: Create helper to calculate small_imbalance in fbg()
Date: Wed, 25 Mar 2009 09:46:46 GMT [thread overview]
Message-ID: <tip-2e6f44aeda426054fc58464df1ad571aecca0c92@git.kernel.org> (raw)
In-Reply-To: <20090325091406.13992.54316.stgit@sofia.in.ibm.com>
Commit-ID: 2e6f44aeda426054fc58464df1ad571aecca0c92
Gitweb: http://git.kernel.org/tip/2e6f44aeda426054fc58464df1ad571aecca0c92
Author: Gautham R Shenoy <ego@in.ibm.com>
AuthorDate: Wed, 25 Mar 2009 14:44:06 +0530
Committer: Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 25 Mar 2009 10:30:47 +0100
sched: Create helper to calculate small_imbalance in fbg()
Impact: cleanup
We have two places in find_busiest_group() where we need to calculate
the minor imbalance before returning the busiest group. Encapsulate
this functionality into a seperate helper function.
Credit: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
Signed-off-by: Gautham R Shenoy <ego@in.ibm.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Balbir Singh" <balbir@in.ibm.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: "Dhaval Giani" <dhaval@linux.vnet.ibm.com>
Cc: Bharata B Rao <bharata@linux.vnet.ibm.com>
LKML-Reference: <20090325091406.13992.54316.stgit@sofia.in.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
kernel/sched.c | 131 ++++++++++++++++++++++++++++++--------------------------
1 files changed, 70 insertions(+), 61 deletions(-)
diff --git a/kernel/sched.c b/kernel/sched.c
index ec715f9..540147e 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -3484,6 +3484,71 @@ group_next:
} while (group != sd->groups);
}
+
+/**
+ * fix_small_imbalance - Calculate the minor imbalance that exists
+ * amongst the groups of a sched_domain, during
+ * load balancing.
+ * @sds: Statistics of the sched_domain whose imbalance is to be calculated.
+ * @this_cpu: The cpu at whose sched_domain we're performing load-balance.
+ * @imbalance: Variable to store the imbalance.
+ */
+static inline void fix_small_imbalance(struct sd_lb_stats *sds,
+ int this_cpu, unsigned long *imbalance)
+{
+ unsigned long tmp, pwr_now = 0, pwr_move = 0;
+ unsigned int imbn = 2;
+
+ if (sds->this_nr_running) {
+ sds->this_load_per_task /= sds->this_nr_running;
+ if (sds->busiest_load_per_task >
+ sds->this_load_per_task)
+ imbn = 1;
+ } else
+ sds->this_load_per_task =
+ cpu_avg_load_per_task(this_cpu);
+
+ if (sds->max_load - sds->this_load + sds->busiest_load_per_task >=
+ sds->busiest_load_per_task * imbn) {
+ *imbalance = sds->busiest_load_per_task;
+ return;
+ }
+
+ /*
+ * OK, we don't have enough imbalance to justify moving tasks,
+ * however we may be able to increase total CPU power used by
+ * moving them.
+ */
+
+ pwr_now += sds->busiest->__cpu_power *
+ min(sds->busiest_load_per_task, sds->max_load);
+ pwr_now += sds->this->__cpu_power *
+ min(sds->this_load_per_task, sds->this_load);
+ pwr_now /= SCHED_LOAD_SCALE;
+
+ /* Amount of load we'd subtract */
+ tmp = sg_div_cpu_power(sds->busiest,
+ sds->busiest_load_per_task * SCHED_LOAD_SCALE);
+ if (sds->max_load > tmp)
+ pwr_move += sds->busiest->__cpu_power *
+ min(sds->busiest_load_per_task, sds->max_load - tmp);
+
+ /* Amount of load we'd add */
+ if (sds->max_load * sds->busiest->__cpu_power <
+ sds->busiest_load_per_task * SCHED_LOAD_SCALE)
+ tmp = sg_div_cpu_power(sds->this,
+ sds->max_load * sds->busiest->__cpu_power);
+ else
+ tmp = sg_div_cpu_power(sds->this,
+ sds->busiest_load_per_task * SCHED_LOAD_SCALE);
+ pwr_move += sds->this->__cpu_power *
+ min(sds->this_load_per_task, sds->this_load + tmp);
+ pwr_move /= SCHED_LOAD_SCALE;
+
+ /* Move if we gain throughput */
+ if (pwr_move > pwr_now)
+ *imbalance = sds->busiest_load_per_task;
+}
/******* find_busiest_group() helpers end here *********************/
/*
@@ -3547,7 +3612,8 @@ find_busiest_group(struct sched_domain *sd, int this_cpu,
*/
if (sds.max_load < sds.avg_load) {
*imbalance = 0;
- goto small_imbalance;
+ fix_small_imbalance(&sds, this_cpu, imbalance);
+ goto ret_busiest;
}
/* Don't want to pull so many tasks that a group would go idle */
@@ -3565,67 +3631,10 @@ find_busiest_group(struct sched_domain *sd, int this_cpu,
* a think about bumping its value to force at least one task to be
* moved
*/
- if (*imbalance < sds.busiest_load_per_task) {
- unsigned long tmp, pwr_now, pwr_move;
- unsigned int imbn;
-
-small_imbalance:
- pwr_move = pwr_now = 0;
- imbn = 2;
- if (sds.this_nr_running) {
- sds.this_load_per_task /= sds.this_nr_running;
- if (sds.busiest_load_per_task >
- sds.this_load_per_task)
- imbn = 1;
- } else
- sds.this_load_per_task =
- cpu_avg_load_per_task(this_cpu);
-
- if (sds.max_load - sds.this_load +
- sds.busiest_load_per_task >=
- sds.busiest_load_per_task * imbn) {
- *imbalance = sds.busiest_load_per_task;
- return sds.busiest;
- }
-
- /*
- * OK, we don't have enough imbalance to justify moving tasks,
- * however we may be able to increase total CPU power used by
- * moving them.
- */
-
- pwr_now += sds.busiest->__cpu_power *
- min(sds.busiest_load_per_task, sds.max_load);
- pwr_now += sds.this->__cpu_power *
- min(sds.this_load_per_task, sds.this_load);
- pwr_now /= SCHED_LOAD_SCALE;
-
- /* Amount of load we'd subtract */
- tmp = sg_div_cpu_power(sds.busiest,
- sds.busiest_load_per_task * SCHED_LOAD_SCALE);
- if (sds.max_load > tmp)
- pwr_move += sds.busiest->__cpu_power *
- min(sds.busiest_load_per_task,
- sds.max_load - tmp);
-
- /* Amount of load we'd add */
- if (sds.max_load * sds.busiest->__cpu_power <
- sds.busiest_load_per_task * SCHED_LOAD_SCALE)
- tmp = sg_div_cpu_power(sds.this,
- sds.max_load * sds.busiest->__cpu_power);
- else
- tmp = sg_div_cpu_power(sds.this,
- sds.busiest_load_per_task * SCHED_LOAD_SCALE);
- pwr_move += sds.this->__cpu_power *
- min(sds.this_load_per_task,
- sds.this_load + tmp);
- pwr_move /= SCHED_LOAD_SCALE;
-
- /* Move if we gain throughput */
- if (pwr_move > pwr_now)
- *imbalance = sds.busiest_load_per_task;
- }
+ if (*imbalance < sds.busiest_load_per_task)
+ fix_small_imbalance(&sds, this_cpu, imbalance);
+ret_busiest:
return sds.busiest;
out_balanced:
next prev parent reply other threads:[~2009-03-25 9:49 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-25 9:13 [RFC PATCH 00/11] sched: find_busiest_group() cleanup Gautham R Shenoy
2009-03-25 9:13 ` [RFC PATCH 01/11] sched: Simple helper functions for find_busiest_group() Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:13 ` [RFC PATCH 02/11] sched: Fix indentations in find_busiest_group using gotos Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] sched: Fix indentations in find_busiest_group() " Gautham R Shenoy
2009-03-25 9:13 ` [RFC PATCH 03/11] sched: Define structure to store the sched_group statistics for fbg() Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:13 ` [RFC PATCH 04/11] sched: Create a helper function to calculate sched_group stats " Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:13 ` [RFC PATCH 05/11] sched: Define structure to store the sched_domain statistics " Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:14 ` [RFC PATCH 06/11] sched: Create a helper function to calculate sched_domain stats " Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:14 ` [RFC PATCH 07/11] sched: Create helper to calculate small_imbalance in find_busiest_group Gautham R Shenoy
2009-03-25 9:46 ` Gautham R Shenoy [this message]
2009-03-25 9:14 ` [RFC PATCH 08/11] sched: Create a helper function to calculate imbalance Gautham R Shenoy
2009-03-25 9:46 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:14 ` [RFC PATCH 09/11] sched: Optimize the !power_savings_balance during find_busiest_group Gautham R Shenoy
2009-03-25 9:47 ` [tip:sched/balancing] sched: Optimize the !power_savings_balance during fbg() Gautham R Shenoy
2009-03-25 9:14 ` [RFC PATCH 10/11] sched: Refactor the power savings balance code Gautham R Shenoy
2009-03-25 9:47 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 9:14 ` [RFC PATCH 11/11] sched: Add comments to find_busiest_group() function Gautham R Shenoy
2009-03-25 9:47 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 11:43 ` [RFC PATCH 11/11] " Gautham R Shenoy
2009-03-25 12:29 ` Ingo Molnar
2009-03-25 13:07 ` Gautham R Shenoy
2009-03-25 13:10 ` Ingo Molnar
2009-03-25 12:30 ` [tip:sched/balancing] " Gautham R Shenoy
2009-03-25 16:04 ` Ray Lee
2009-03-25 16:17 ` Ingo Molnar
2009-03-25 19:17 ` Gautham R Shenoy
2009-03-25 9:30 ` [RFC PATCH 00/11] sched: find_busiest_group() cleanup Ingo Molnar
2009-03-25 9:42 ` Ingo Molnar
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=tip-2e6f44aeda426054fc58464df1ad571aecca0c92@git.kernel.org \
--to=ego@in.ibm.com \
--cc=a.p.zijlstra@chello.nl \
--cc=balbir@in.ibm.com \
--cc=bharata@linux.vnet.ibm.com \
--cc=dhaval@linux.vnet.ibm.com \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-tip-commits@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=mingo@redhat.com \
--cc=nickpiggin@yahoo.com.au \
--cc=suresh.b.siddha@intel.com \
--cc=tglx@linutronix.de \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.