rpms/kernel/F-9 linux-2.6-sched-fine-tune-SD_MC_INIT.patch, NONE, 1.1 linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch, NONE, 1.1 linux-2.6-sched-wakeup-preempt-when-small-overlap.patch, NONE, 1.1 kernel.spec, 1.872, 1.873

[Chuck Ebbert]

Author: cebbert

Update of /cvs/pkgs/rpms/kernel/F-9
In directory cvs1.fedora.phx.redhat.com:/tmp/cvs-serv19506

Modified Files:
	kernel.spec 
Added Files:
	linux-2.6-sched-fine-tune-SD_MC_INIT.patch 
	linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch 
	linux-2.6-sched-wakeup-preempt-when-small-overlap.patch 
Log Message:
Scheduler fixes from 2.6.28

linux-2.6-sched-fine-tune-SD_MC_INIT.patch:

--- NEW FILE linux-2.6-sched-fine-tune-SD_MC_INIT.patch ---
From: Mike Galbraith <efault at gmx.de>
Date: Fri, 7 Nov 2008 14:26:50 +0000 (+0100)
Subject: sched: fine-tune SD_MC_INIT
X-Git-Url: http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Ftorvalds%2Flinux-2.6.git;a=commitdiff_plain;h=14800984706bf6936bbec5187f736e928be5c218

sched: fine-tune SD_MC_INIT

Tune SD_MC_INIT the same way as SD_CPU_INIT:
unset SD_BALANCE_NEWIDLE, and set SD_WAKE_BALANCE.

This improves vmark by 5%:

vmark         132102 125968 125497 messages/sec    avg 127855.66    .984
vmark         139404 131719 131272 messages/sec    avg 134131.66   1.033

Signed-off-by: Mike Galbraith <efault at gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra at chello.nl>
Signed-off-by: Ingo Molnar <mingo at elte.hu>

 # *DOCUMENTATION*
---

diff --git a/include/linux/topology.h b/include/linux/topology.h
index 34a7ee0..a8d8405 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -120,10 +120,10 @@ void arch_update_cpu_topology(void);
 	.wake_idx		= 1,			\
 	.forkexec_idx		= 1,			\
 	.flags			= SD_LOAD_BALANCE	\
-				| SD_BALANCE_NEWIDLE	\
 				| SD_BALANCE_FORK	\
 				| SD_BALANCE_EXEC	\
 				| SD_WAKE_AFFINE	\
+				| SD_WAKE_BALANCE	\
 				| SD_SHARE_PKG_RESOURCES\
 				| BALANCE_FOR_MC_POWER,	\
 	.last_balance		= jiffies,		\

linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch:

--- NEW FILE linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch ---
From: Ingo Molnar <mingo at elte.hu>
Date: Fri, 7 Nov 2008 15:09:23 +0000 (+0100)
Subject: sched: fine-tune SD_SIBLING_INIT
X-Git-Url: http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Ftorvalds%2Flinux-2.6.git;a=commitdiff_plain;h=52c642f33b14bfa1b00ef2b68296effb34a573f3

sched: fine-tune SD_SIBLING_INIT

fine-tune the HT sched-domains parameters as well.

On a HT capable box, this increases lat_ctx performance from 23.87
usecs to 1.49 usecs:

 # before

 $ ./lat_ctx -s 0 2

   "size=0k ovr=1.89
    2 23.87

 # after

 $ ./lat_ctx -s 0 2

   "size=0k ovr=1.84
     2 1.49

Signed-off-by: Ingo Molnar <mingo at elte.hu>
---

diff --git a/include/linux/topology.h b/include/linux/topology.h
index a8d8405..117f1b7 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -99,7 +99,7 @@ void arch_update_cpu_topology(void);
 				| SD_BALANCE_FORK	\
 				| SD_BALANCE_EXEC	\
 				| SD_WAKE_AFFINE	\
-				| SD_WAKE_IDLE		\
+				| SD_WAKE_BALANCE	\
 				| SD_SHARE_CPUPOWER,	\
 	.last_balance		= jiffies,		\
 	.balance_interval	= 1,			\

linux-2.6-sched-wakeup-preempt-when-small-overlap.patch:

--- NEW FILE linux-2.6-sched-wakeup-preempt-when-small-overlap.patch ---
From: Peter Zijlstra <a.p.zijlstra at chello.nl>
Date: Sat, 20 Sep 2008 21:38:02 +0000 (+0200)
Subject: sched: wakeup preempt when small overlap
X-Git-Tag: v2.6.28-rc1~725^2~2^2~15
X-Git-Url: http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Ftorvalds%2Flinux-2.6.git;a=commitdiff_plain;h=15afe09bf496ae10c989e1a375a6b5da7bd3e16e

sched: wakeup preempt when small overlap

Lin Ming reported a 10% OLTP regression against 2.6.27-rc4.

The difference seems to come from different preemption agressiveness,
which affects the cache footprint of the workload and its effective
cache trashing.

Aggresively preempt a task if its avg overlap is very small, this should
avoid the task going to sleep and find it still running when we schedule
back to it - saving a wakeup.

Reported-by: Lin Ming <ming.m.lin at intel.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra at chello.nl>
Signed-off-by: Ingo Molnar <mingo at elte.hu>
---

diff --git a/include/linux/sched.h b/include/linux/sched.h
index b3b7a8f..d8e699b 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -897,7 +897,7 @@ struct sched_class {
 	void (*yield_task) (struct rq *rq);
 	int  (*select_task_rq)(struct task_struct *p, int sync);
 
-	void (*check_preempt_curr) (struct rq *rq, struct task_struct *p);
+	void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
 
 	struct task_struct * (*pick_next_task) (struct rq *rq);
 	void (*put_prev_task) (struct rq *rq, struct task_struct *p);
diff --git a/kernel/sched.c b/kernel/sched.c
index 0d8905a..ad9d39b 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -604,9 +604,9 @@ struct rq {
 
 static DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
 
-static inline void check_preempt_curr(struct rq *rq, struct task_struct *p)
+static inline void check_preempt_curr(struct rq *rq, struct task_struct *p, int sync)
 {
-	rq->curr->sched_class->check_preempt_curr(rq, p);
+	rq->curr->sched_class->check_preempt_curr(rq, p, sync);
 }
 
 static inline int cpu_of(struct rq *rq)
@@ -2282,7 +2282,7 @@ out_running:
 	trace_mark(kernel_sched_wakeup,
 		"pid %d state %ld ## rq %p task %p rq->curr %p",
 		p->pid, p->state, rq, p, rq->curr);
-	check_preempt_curr(rq, p);
+	check_preempt_curr(rq, p, sync);
 
 	p->state = TASK_RUNNING;
 #ifdef CONFIG_SMP
@@ -2417,7 +2417,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags)
 	trace_mark(kernel_sched_wakeup_new,
 		"pid %d state %ld ## rq %p task %p rq->curr %p",
 		p->pid, p->state, rq, p, rq->curr);
-	check_preempt_curr(rq, p);
+	check_preempt_curr(rq, p, 0);
 #ifdef CONFIG_SMP
 	if (p->sched_class->task_wake_up)
 		p->sched_class->task_wake_up(rq, p);
@@ -2877,7 +2877,7 @@ static void pull_task(struct rq *src_rq, struct task_struct *p,
 	 * Note that idle threads have a prio of MAX_PRIO, for this test
 	 * to be always true for them.
 	 */
-	check_preempt_curr(this_rq, p);
+	check_preempt_curr(this_rq, p, 0);
 }
 
 /*
@@ -6007,7 +6007,7 @@ static int __migrate_task(struct task_struct *p, int src_cpu, int dest_cpu)
 	set_task_cpu(p, dest_cpu);
 	if (on_rq) {
 		activate_task(rq_dest, p, 0);
-		check_preempt_curr(rq_dest, p);
+		check_preempt_curr(rq_dest, p, 0);
 	}
 done:
 	ret = 1;
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index a10ac0b..7328383 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -1331,7 +1331,7 @@ static inline int depth_se(struct sched_entity *se)
 /*
  * Preempt the current task with a newly woken task if needed:
  */
-static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
+static void check_preempt_wakeup(struct rq *rq, struct task_struct *p, int sync)
 {
 	struct task_struct *curr = rq->curr;
 	struct cfs_rq *cfs_rq = task_cfs_rq(curr);
@@ -1367,6 +1367,13 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
 	if (!sched_feat(WAKEUP_PREEMPT))
 		return;
 
+	if (sched_feat(WAKEUP_OVERLAP) && sync &&
+			se->avg_overlap < sysctl_sched_migration_cost &&
+			pse->avg_overlap < sysctl_sched_migration_cost) {
+		resched_task(curr);
+		return;
+	}
+
 	/*
 	 * preemption test can be made between sibling entities who are in the
 	 * same cfs_rq i.e who have a common parent. Walk up the hierarchy of
@@ -1649,7 +1656,7 @@ static void prio_changed_fair(struct rq *rq, struct task_struct *p,
 		if (p->prio > oldprio)
 			resched_task(rq->curr);
 	} else
-		check_preempt_curr(rq, p);
+		check_preempt_curr(rq, p, 0);
 }
 
 /*
@@ -1666,7 +1673,7 @@ static void switched_to_fair(struct rq *rq, struct task_struct *p,
 	if (running)
 		resched_task(rq->curr);
 	else
-		check_preempt_curr(rq, p);
+		check_preempt_curr(rq, p, 0);
 }
 
 /* Account for a task changing its policy or group.
diff --git a/kernel/sched_features.h b/kernel/sched_features.h
index 9353ca7..bf027a7 100644
--- a/kernel/sched_features.h
+++ b/kernel/sched_features.h
@@ -11,3 +11,4 @@ SCHED_FEAT(ASYM_GRAN, 1)
 SCHED_FEAT(LB_BIAS, 1)
 SCHED_FEAT(LB_WAKEUP_UPDATE, 1)
 SCHED_FEAT(ASYM_EFF_LOAD, 1)
+SCHED_FEAT(WAKEUP_OVERLAP, 1)
diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c
index 3a4f92d..dec4cca 100644
--- a/kernel/sched_idletask.c
+++ b/kernel/sched_idletask.c
@@ -14,7 +14,7 @@ static int select_task_rq_idle(struct task_struct *p, int sync)
 /*
  * Idle tasks are unconditionally rescheduled:
  */
-static void check_preempt_curr_idle(struct rq *rq, struct task_struct *p)
+static void check_preempt_curr_idle(struct rq *rq, struct task_struct *p, int sync)
 {
 	resched_task(rq->idle);
 }
@@ -76,7 +76,7 @@ static void switched_to_idle(struct rq *rq, struct task_struct *p,
 	if (running)
 		resched_task(rq->curr);
 	else
-		check_preempt_curr(rq, p);
+		check_preempt_curr(rq, p, 0);
 }
 
 static void prio_changed_idle(struct rq *rq, struct task_struct *p,
@@ -93,7 +93,7 @@ static void prio_changed_idle(struct rq *rq, struct task_struct *p,
 		if (p->prio > oldprio)
 			resched_task(rq->curr);
 	} else
-		check_preempt_curr(rq, p);
+		check_preempt_curr(rq, p, 0);
 }
 
 /*
diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c
index 5523107..6d2d0a5 100644
--- a/kernel/sched_rt.c
+++ b/kernel/sched_rt.c
@@ -783,7 +783,7 @@ static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
 /*
  * Preempt the current task with a newly woken task if needed:
  */
-static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p)
+static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p, int sync)
 {
 	if (p->prio < rq->curr->prio) {
 		resched_task(rq->curr);


Index: kernel.spec
===================================================================
RCS file: /cvs/pkgs/rpms/kernel/F-9/kernel.spec,v
retrieving revision 1.872
retrieving revision 1.873
diff -u -r1.872 -r1.873
--- kernel.spec	8 Dec 2008 00:19:01 -0000	1.872
+++ kernel.spec	8 Dec 2008 05:54:10 -0000	1.873
@@ -604,6 +604,10 @@
 
 Patch41: linux-2.6-sysrq-c.patch
 
+Patch60: linux-2.6-sched-fine-tune-SD_MC_INIT.patch
+Patch61: linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch
+Patch62: linux-2.6-sched-wakeup-preempt-when-small-overlap.patch
+
 Patch70: linux-2.6-x86-tune-generic.patch
 Patch75: linux-2.6-x86-debug-boot.patch
 Patch101: linux-2.6-x86-check-for-null-irq-context.patch
@@ -1085,6 +1089,9 @@
 ApplyPatch linux-2.6-sysrq-c.patch
 
 # scheduler patches
+ApplyPatch linux-2.6-sched-fine-tune-SD_MC_INIT.patch
+ApplyPatch linux-2.6-sched-fine-tune-SD_SIBLING_INIT.patch
+ApplyPatch linux-2.6-sched-wakeup-preempt-when-small-overlap.patch
 
 # Architecture patches
 # x86(-64)
@@ -1891,6 +1898,9 @@
 %kernel_variant_files -a /%{image_install_path}/xen*-%{KVERREL}.xen -e /etc/ld.so.conf.d/kernelcap-%{KVERREL}.xen.conf %{with_xen} xen
 
 %changelog
+* Wed Dec 03 2008 Chuck Ebbert <cebbert at redhat.com> 2.6.27.8-60
+- Scheduler fixes from 2.6.28
+
 * Wed Dec 03 2008 Chuck Ebbert <cebbert at redhat.com> 2.6.27.8-59
 - Fix PCI config space size on AMD Barcelona.