This patch introduces two new structures:

struct sched_entity
        stores essential attributes/execution-history used by CFS core
        to drive fairness between 'schedulable entities' (tasks, users etc)

struct lrq
        runqueue used to hold ready-to-run entities

These new structures are formed by grouping together existing fields in
existing structures (task_struct and rq) and hence represents rework
with zero functionality change.

Signed-off-by : Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>


---
 fs/proc/array.c           |    4 
 include/linux/sched.h     |   43 +++++----
 kernel/exit.c             |    2 
 kernel/posix-cpu-timers.c |   16 +--
 kernel/sched.c            |  186 +++++++++++++++++++++-------------------
 kernel/sched_debug.c      |   86 +++++++++---------
 kernel/sched_fair.c       |  211 +++++++++++++++++++++++-----------------------
 kernel/sched_rt.c         |   14 +--
 8 files changed, 289 insertions(+), 273 deletions(-)

Index: current/include/linux/sched.h
===================================================================
--- current.orig/include/linux/sched.h	2007-06-09 15:01:39.000000000 +0530
+++ current/include/linux/sched.h	2007-06-09 15:04:54.000000000 +0530
@@ -872,6 +872,29 @@
 	void (*task_new) (struct rq *rq, struct task_struct *p);
 };
 
+/* CFS stats for a schedulable entity (task, task-group etc) */
+struct sched_entity {
+	int load_weight;	/* for niceness load balancing purposes */
+	int on_rq;
+	struct rb_node run_node;
+	u64 wait_start_fair;
+	u64 wait_start;
+	u64 exec_start;
+	u64 sleep_start, sleep_start_fair;
+	u64 block_start;
+	u64 sleep_max;
+	u64 block_max;
+	u64 exec_max;
+	u64 wait_max;
+	u64 last_ran;
+
+	s64 wait_runtime;
+	u64 sum_exec_runtime;
+	s64 fair_key;
+	s64 sum_wait_runtime, sum_sleep_runtime;
+	unsigned long wait_runtime_overruns, wait_runtime_underruns;
+};
+
 struct task_struct {
 	volatile long state;	/* -1 unrunnable, 0 runnable, >0 stopped */
 	void *stack;
@@ -886,33 +909,15 @@
 	int oncpu;
 #endif
 #endif
-	int load_weight;	/* for niceness load balancing purposes */
 
 	int prio, static_prio, normal_prio;
-	int on_rq;
 	struct list_head run_list;
-	struct rb_node run_node;
+	struct sched_entity se;
 
 	unsigned short ioprio;
 #ifdef CONFIG_BLK_DEV_IO_TRACE
 	unsigned int btrace_seq;
 #endif
-	/* CFS scheduling class statistics fields: */
-	u64 wait_start_fair;
-	u64 wait_start;
-	u64 exec_start;
-	u64 sleep_start, sleep_start_fair;
-	u64 block_start;
-	u64 sleep_max;
-	u64 block_max;
-	u64 exec_max;
-	u64 wait_max;
-
-	s64 wait_runtime;
-	u64 sum_exec_runtime;
-	s64 fair_key;
-	s64 sum_wait_runtime, sum_sleep_runtime;
-	unsigned long wait_runtime_overruns, wait_runtime_underruns;
 
 	unsigned int policy;
 	cpumask_t cpus_allowed;
Index: current/fs/proc/array.c
===================================================================
--- current.orig/fs/proc/array.c	2007-06-09 15:01:39.000000000 +0530
+++ current/fs/proc/array.c	2007-06-09 15:04:54.000000000 +0530
@@ -329,7 +329,7 @@
 	 * Use CFS's precise accounting, if available:
 	 */
 	if (!(sysctl_sched_features & 128)) {
-		u64 temp = (u64)nsec_to_clock_t(p->sum_exec_runtime);
+		u64 temp = (u64)nsec_to_clock_t(p->se.sum_exec_runtime);
 
 		if (total) {
 			temp *= utime;
@@ -351,7 +351,7 @@
 	 * by userspace grows monotonically - apps rely on that):
 	 */
 	if (!(sysctl_sched_features & 128))
-		stime = nsec_to_clock_t(p->sum_exec_runtime) - task_utime(p);
+		stime = nsec_to_clock_t(p->se.sum_exec_runtime) - task_utime(p);
 
 	return stime;
 }
Index: current/kernel/exit.c
===================================================================
--- current.orig/kernel/exit.c	2007-06-09 14:56:50.000000000 +0530
+++ current/kernel/exit.c	2007-06-09 15:04:54.000000000 +0530
@@ -126,7 +126,7 @@
 		sig->nivcsw += tsk->nivcsw;
 		sig->inblock += task_io_get_inblock(tsk);
 		sig->oublock += task_io_get_oublock(tsk);
-		sig->sum_sched_runtime += tsk->sum_exec_runtime;
+		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
 		sig = NULL; /* Marker for below. */
 	}
 
Index: current/kernel/posix-cpu-timers.c
===================================================================
--- current.orig/kernel/posix-cpu-timers.c	2007-06-09 15:01:39.000000000 +0530
+++ current/kernel/posix-cpu-timers.c	2007-06-09 15:04:54.000000000 +0530
@@ -249,7 +249,7 @@
 		cpu->sched = p->signal->sum_sched_runtime;
 		/* Add in each other live thread.  */
 		while ((t = next_thread(t)) != p) {
-			cpu->sched += t->sum_exec_runtime;
+			cpu->sched += t->se.sum_exec_runtime;
 		}
 		cpu->sched += sched_ns(p);
 		break;
@@ -467,7 +467,7 @@
 void posix_cpu_timers_exit(struct task_struct *tsk)
 {
 	cleanup_timers(tsk->cpu_timers,
-		       tsk->utime, tsk->stime, tsk->sum_exec_runtime);
+		       tsk->utime, tsk->stime, tsk->se.sum_exec_runtime);
 
 }
 void posix_cpu_timers_exit_group(struct task_struct *tsk)
@@ -475,7 +475,7 @@
 	cleanup_timers(tsk->signal->cpu_timers,
 		       cputime_add(tsk->utime, tsk->signal->utime),
 		       cputime_add(tsk->stime, tsk->signal->stime),
-		       tsk->sum_exec_runtime + tsk->signal->sum_sched_runtime);
+		     tsk->se.sum_exec_runtime + tsk->signal->sum_sched_runtime);
 }
 
 
@@ -536,7 +536,7 @@
 		nsleft = max_t(unsigned long long, nsleft, 1);
 		do {
 			if (likely(!(t->flags & PF_EXITING))) {
-				ns = t->sum_exec_runtime + nsleft;
+				ns = t->se.sum_exec_runtime + nsleft;
 				if (t->it_sched_expires == 0 ||
 				    t->it_sched_expires > ns) {
 					t->it_sched_expires = ns;
@@ -1004,7 +1004,7 @@
 		struct cpu_timer_list *t = list_first_entry(timers,
 						      struct cpu_timer_list,
 						      entry);
-		if (!--maxfire || tsk->sum_exec_runtime < t->expires.sched) {
+		if (!--maxfire || tsk->se.sum_exec_runtime < t->expires.sched) {
 			tsk->it_sched_expires = t->expires.sched;
 			break;
 		}
@@ -1049,7 +1049,7 @@
 	do {
 		utime = cputime_add(utime, t->utime);
 		stime = cputime_add(stime, t->stime);
-		sum_sched_runtime += t->sum_exec_runtime;
+		sum_sched_runtime += t->se.sum_exec_runtime;
 		t = next_thread(t);
 	} while (t != tsk);
 	ptime = cputime_add(utime, stime);
@@ -1208,7 +1208,7 @@
 				t->it_virt_expires = ticks;
 			}
 
-			sched = t->sum_exec_runtime + sched_left;
+			sched = t->se.sum_exec_runtime + sched_left;
 			if (sched_expires && (t->it_sched_expires == 0 ||
 					      t->it_sched_expires > sched)) {
 				t->it_sched_expires = sched;
@@ -1300,7 +1300,7 @@
 
 	if (UNEXPIRED(prof) && UNEXPIRED(virt) &&
 	    (tsk->it_sched_expires == 0 ||
-	     tsk->sum_exec_runtime < tsk->it_sched_expires))
+	     tsk->se.sum_exec_runtime < tsk->it_sched_expires))
 		return;
 
 #undef	UNEXPIRED
Index: current/kernel/sched.c
===================================================================
--- current.orig/kernel/sched.c	2007-06-09 15:01:39.000000000 +0530
+++ current/kernel/sched.c	2007-06-09 15:07:17.000000000 +0530
@@ -116,6 +116,23 @@
 	struct list_head queue[MAX_RT_PRIO];
 };
 
+/* CFS-related fields in a runqueue */
+struct lrq {
+	unsigned long raw_weighted_load;
+	#define CPU_LOAD_IDX_MAX 5
+	unsigned long cpu_load[CPU_LOAD_IDX_MAX];
+	unsigned long nr_load_updates;
+
+	u64 fair_clock, delta_fair_clock;
+	u64 exec_clock, delta_exec_clock;
+	s64 wait_runtime;
+	unsigned long wait_runtime_overruns, wait_runtime_underruns;
+
+	struct rb_root tasks_timeline;
+	struct rb_node *rb_leftmost;
+	struct rb_node *rb_load_balance_curr;
+};
+
 /*
  * This is the main, per-CPU runqueue data structure.
  *
@@ -131,16 +148,13 @@
 	 * remote CPUs use both these fields when doing load calculation.
 	 */
 	long nr_running;
-	unsigned long raw_weighted_load;
-	#define CPU_LOAD_IDX_MAX 5
-	unsigned long cpu_load[CPU_LOAD_IDX_MAX];
+	struct lrq lrq;
 
 	unsigned char idle_at_tick;
 #ifdef CONFIG_NO_HZ
 	unsigned char in_nohz_recently;
 #endif
 	u64 nr_switches;
-	unsigned long nr_load_updates;
 
 	/*
 	 * This is part of a global counter where only the total sum
@@ -156,10 +170,6 @@
 
 	u64 clock, prev_clock_raw;
 	s64 clock_max_delta;
-	u64 fair_clock, delta_fair_clock;
-	u64 exec_clock, delta_exec_clock;
-	s64 wait_runtime;
-	unsigned long wait_runtime_overruns, wait_runtime_underruns;
 
 	unsigned int clock_warps, clock_overflows;
 	unsigned int clock_unstable_events;
@@ -170,10 +180,6 @@
 	int rt_load_balance_idx;
 	struct list_head *rt_load_balance_head, *rt_load_balance_curr;
 
-	struct rb_root tasks_timeline;
-	struct rb_node *rb_leftmost;
-	struct rb_node *rb_load_balance_curr;
-
 	atomic_t nr_iowait;
 
 #ifdef CONFIG_SMP
@@ -583,13 +589,13 @@
 static inline void
 inc_raw_weighted_load(struct rq *rq, const struct task_struct *p)
 {
-	rq->raw_weighted_load += p->load_weight;
+	rq->lrq.raw_weighted_load += p->se.load_weight;
 }
 
 static inline void
 dec_raw_weighted_load(struct rq *rq, const struct task_struct *p)
 {
-	rq->raw_weighted_load -= p->load_weight;
+	rq->lrq.raw_weighted_load -= p->se.load_weight;
 }
 
 static inline void inc_nr_running(struct task_struct *p, struct rq *rq)
@@ -615,22 +621,22 @@
 
 static void set_load_weight(struct task_struct *p)
 {
-	task_rq(p)->wait_runtime -= p->wait_runtime;
-	p->wait_runtime = 0;
+	task_rq(p)->lrq.wait_runtime -= p->se.wait_runtime;
+	p->se.wait_runtime = 0;
 
 	if (has_rt_policy(p)) {
-		p->load_weight = prio_to_weight[0] * 2;
+		p->se.load_weight = prio_to_weight[0] * 2;
 		return;
 	}
 	/*
 	 * SCHED_IDLEPRIO tasks get minimal weight:
 	 */
 	if (p->policy == SCHED_IDLEPRIO) {
-		p->load_weight = 1;
+		p->se.load_weight = 1;
 		return;
 	}
 
-	p->load_weight = prio_to_weight[p->static_prio - MAX_RT_PRIO];
+	p->se.load_weight = prio_to_weight[p->static_prio - MAX_RT_PRIO];
 }
 
 static v