* 2003-09-03 Interactivity tuning by Con Kolivas.
* 2004-04-02 Scheduler domains code by Nick Piggin
*/
-
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/nmi.h>
#include <linux/kernel_stat.h>
#include <linux/security.h>
#include <linux/notifier.h>
+#include <linux/profile.h>
#include <linux/suspend.h>
#include <linux/blkdev.h>
#include <linux/delay.h>
#include <linux/cpu.h>
#include <linux/percpu.h>
#include <linux/kthread.h>
+#include <linux/seq_file.h>
+#include <linux/times.h>
+#include <linux/vserver/sched.h>
+#include <linux/vs_base.h>
+#include <linux/vs_context.h>
+#include <linux/vs_cvirt.h>
+#include <asm/tlb.h>
#include <asm/unistd.h>
#define cpu_to_node_mask(cpu) (cpu_online_map)
#endif
+/* used to soft spin in sched while dump is in progress */
+unsigned long dump_oncpu;
+EXPORT_SYMBOL(dump_oncpu);
+
/*
* Convert user-nice values [ -20 ... 0 ... 19 ]
* to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ],
#define USER_PRIO(p) ((p)-MAX_RT_PRIO)
#define TASK_USER_PRIO(p) USER_PRIO((p)->static_prio)
#define MAX_USER_PRIO (USER_PRIO(MAX_PRIO))
-#define AVG_TIMESLICE (MIN_TIMESLICE + ((MAX_TIMESLICE - MIN_TIMESLICE) *\
- (MAX_PRIO-1-NICE_TO_PRIO(0))/(MAX_USER_PRIO - 1)))
/*
* Some helpers for converting nanosecond timing to jiffy resolution
/*
* These are the 'tuning knobs' of the scheduler:
*
- * Minimum timeslice is 10 msecs, default timeslice is 100 msecs,
- * maximum timeslice is 200 msecs. Timeslices get refilled after
- * they expire.
+ * Minimum timeslice is 5 msecs (or 1 jiffy, whichever is larger),
+ * default timeslice is 100 msecs, maximum timeslice is 800 msecs.
+ * Timeslices get refilled after they expire.
*/
-#define MIN_TIMESLICE ( 10 * HZ / 1000)
-#define MAX_TIMESLICE (200 * HZ / 1000)
+#define MIN_TIMESLICE max(5 * HZ / 1000, 1)
+#define DEF_TIMESLICE (100 * HZ / 1000)
#define ON_RUNQUEUE_WEIGHT 30
#define CHILD_PENALTY 95
#define PARENT_PENALTY 100
#define PRIO_BONUS_RATIO 25
#define MAX_BONUS (MAX_USER_PRIO * PRIO_BONUS_RATIO / 100)
#define INTERACTIVE_DELTA 2
-#define MAX_SLEEP_AVG (AVG_TIMESLICE * MAX_BONUS)
+#define MAX_SLEEP_AVG (DEF_TIMESLICE * MAX_BONUS)
#define STARVATION_LIMIT (MAX_SLEEP_AVG)
#define NS_MAX_SLEEP_AVG (JIFFIES_TO_NS(MAX_SLEEP_AVG))
#define CREDIT_LIMIT 100
#define LOW_CREDIT(p) \
((p)->interactive_credit < -CREDIT_LIMIT)
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+/*
+ * if belong to different class, compare class priority
+ * otherwise compare task priority
+ */
+#define TASK_PREEMPTS_CURR(p, rq) \
+ ( ((p)->cpu_class != (rq)->curr->cpu_class) \
+ && ((rq)->curr != (rq)->idle) && ((p) != (rq)->idle )) \
+ ? class_preempts_curr((p),(rq)->curr) \
+ : ((p)->prio < (rq)->curr->prio)
+#else
#define TASK_PREEMPTS_CURR(p, rq) \
((p)->prio < (rq)->curr->prio)
+#endif
/*
- * BASE_TIMESLICE scales user-nice values [ -20 ... 19 ]
- * to time slice values.
+ * task_timeslice() scales user-nice values [ -20 ... 0 ... 19 ]
+ * to time slice values: [800ms ... 100ms ... 5ms]
*
* The higher a thread's priority, the bigger timeslices
* it gets during one round of execution. But even the lowest
* priority thread gets MIN_TIMESLICE worth of execution time.
- *
- * task_timeslice() is the interface that is used by the scheduler.
*/
-#define BASE_TIMESLICE(p) (MIN_TIMESLICE + \
- ((MAX_TIMESLICE - MIN_TIMESLICE) * \
- (MAX_PRIO-1 - (p)->static_prio) / (MAX_USER_PRIO-1)))
+#define SCALE_PRIO(x, prio) \
+ max(x * (MAX_PRIO - prio) / (MAX_USER_PRIO/2), MIN_TIMESLICE)
-static unsigned int task_timeslice(task_t *p)
+unsigned int task_timeslice(task_t *p)
{
- return BASE_TIMESLICE(p);
+ if (p->static_prio < NICE_TO_PRIO(0))
+ return SCALE_PRIO(DEF_TIMESLICE*4, p->static_prio);
+ else
+ return SCALE_PRIO(DEF_TIMESLICE, p->static_prio);
}
+#define task_hot(p, now, sd) ((long long) ((now) - (p)->last_ran) \
+ < (long long) (sd)->cache_hot_time)
+
+enum idle_type
+{
+ IDLE,
+ NOT_IDLE,
+ NEWLY_IDLE,
+ MAX_IDLE_TYPES
+};
-#define task_hot(p, now, sd) ((now) - (p)->timestamp < (sd)->cache_hot_time)
+struct sched_domain;
/*
* These are the runqueue data structures:
*/
-#define BITMAP_SIZE ((((MAX_PRIO+1+7)/8)+sizeof(long)-1)/sizeof(long))
-
typedef struct runqueue runqueue_t;
-
-struct prio_array {
- unsigned int nr_active;
- unsigned long bitmap[BITMAP_SIZE];
- struct list_head queue[MAX_PRIO];
-};
+#include <linux/ckrm_classqueue.h>
+#include <linux/ckrm_sched.h>
/*
* This is the main, per-CPU runqueue data structure.
unsigned long long timestamp_last_tick;
task_t *curr, *idle;
struct mm_struct *prev_mm;
- prio_array_t *active, *expired, arrays[2];
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ struct classqueue_struct classqueue;
+ ckrm_load_t ckrm_load;
+#else
+ prio_array_t *active, *expired, arrays[2];
+#endif
int best_expired_prio;
atomic_t nr_iowait;
task_t *migration_thread;
struct list_head migration_queue;
#endif
+
+#ifdef CONFIG_VSERVER_HARDCPU
+ struct list_head hold_queue;
+ int idle_tokens;
+#endif
+
+#ifdef CONFIG_SCHEDSTATS
+ /* latency stats */
+ struct sched_info rq_sched_info;
+
+ /* sys_sched_yield() stats */
+ unsigned long yld_exp_empty;
+ unsigned long yld_act_empty;
+ unsigned long yld_both_empty;
+ unsigned long yld_cnt;
+
+ /* schedule() stats */
+ unsigned long sched_noswitch;
+ unsigned long sched_switch;
+ unsigned long sched_cnt;
+ unsigned long sched_goidle;
+
+ /* pull_task() stats */
+ unsigned long pt_gained[MAX_IDLE_TYPES];
+ unsigned long pt_lost[MAX_IDLE_TYPES];
+
+ /* active_load_balance() stats */
+ unsigned long alb_cnt;
+ unsigned long alb_lost;
+ unsigned long alb_gained;
+ unsigned long alb_failed;
+
+ /* try_to_wake_up() stats */
+ unsigned long ttwu_cnt;
+ unsigned long ttwu_attempts;
+ unsigned long ttwu_moved;
+
+ /* wake_up_new_task() stats */
+ unsigned long wunt_cnt;
+ unsigned long wunt_moved;
+
+ /* sched_migrate_task() stats */
+ unsigned long smt_cnt;
+
+ /* sched_balance_exec() stats */
+ unsigned long sbe_cnt;
+#endif
};
static DEFINE_PER_CPU(struct runqueue, runqueues);
+/*
+ * sched-domains (multiprocessor balancing) declarations:
+ */
+#ifdef CONFIG_SMP
+#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
+
+#define SD_BALANCE_NEWIDLE 1 /* Balance when about to become idle */
+#define SD_BALANCE_EXEC 2 /* Balance on exec */
+#define SD_WAKE_IDLE 4 /* Wake to idle CPU on task wakeup */
+#define SD_WAKE_AFFINE 8 /* Wake task to waking CPU */
+#define SD_WAKE_BALANCE 16 /* Perform balancing at task wakeup */
+#define SD_SHARE_CPUPOWER 32 /* Domain members share cpu power */
+
+struct sched_group {
+ struct sched_group *next; /* Must be a circular list */
+ cpumask_t cpumask;
+
+ /*
+ * CPU power of this group, SCHED_LOAD_SCALE being max power for a
+ * single CPU. This is read only (except for setup, hotplug CPU).
+ */
+ unsigned long cpu_power;
+};
+
+struct sched_domain {
+ /* These fields must be setup */
+ struct sched_domain *parent; /* top domain must be null terminated */
+ struct sched_group *groups; /* the balancing groups of the domain */
+ cpumask_t span; /* span of all CPUs in this domain */
+ unsigned long min_interval; /* Minimum balance interval ms */
+ unsigned long max_interval; /* Maximum balance interval ms */
+ unsigned int busy_factor; /* less balancing by factor if busy */
+ unsigned int imbalance_pct; /* No balance until over watermark */
+ unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
+ unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
+ unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
+ int flags; /* See SD_* */
+
+ /* Runtime fields. */
+ unsigned long last_balance; /* init to jiffies. units in jiffies */
+ unsigned int balance_interval; /* initialise to 1. units in ms. */
+ unsigned int nr_balance_failed; /* initialise to 0 */
+
+#ifdef CONFIG_SCHEDSTATS
+ /* load_balance() stats */
+ unsigned long lb_cnt[MAX_IDLE_TYPES];
+ unsigned long lb_failed[MAX_IDLE_TYPES];
+ unsigned long lb_imbalance[MAX_IDLE_TYPES];
+ unsigned long lb_nobusyg[MAX_IDLE_TYPES];
+ unsigned long lb_nobusyq[MAX_IDLE_TYPES];
+
+ /* sched_balance_exec() stats */
+ unsigned long sbe_attempts;
+ unsigned long sbe_pushed;
+
+ /* try_to_wake_up() stats */
+ unsigned long ttwu_wake_affine;
+ unsigned long ttwu_wake_balance;
+#endif
+};
+
+#ifndef ARCH_HAS_SCHED_TUNE
+#ifdef CONFIG_SCHED_SMT
+#define ARCH_HAS_SCHED_WAKE_IDLE
+/* Common values for SMT siblings */
+#define SD_SIBLING_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 1, \
+ .max_interval = 2, \
+ .busy_factor = 8, \
+ .imbalance_pct = 110, \
+ .cache_hot_time = 0, \
+ .cache_nice_tries = 0, \
+ .per_cpu_gain = 25, \
+ .flags = SD_BALANCE_NEWIDLE \
+ | SD_BALANCE_EXEC \
+ | SD_WAKE_AFFINE \
+ | SD_WAKE_IDLE \
+ | SD_SHARE_CPUPOWER, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+#endif
+
+/* Common values for CPUs */
+#define SD_CPU_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 1, \
+ .max_interval = 4, \
+ .busy_factor = 64, \
+ .imbalance_pct = 125, \
+ .cache_hot_time = cache_decay_ticks*1000000 ? : (5*1000000/2),\
+ .cache_nice_tries = 1, \
+ .per_cpu_gain = 100, \
+ .flags = SD_BALANCE_NEWIDLE \
+ | SD_BALANCE_EXEC \
+ | SD_WAKE_AFFINE \
+ | SD_WAKE_BALANCE, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+
+/* Arch can override this macro in processor.h */
+#if defined(CONFIG_NUMA) && !defined(SD_NODE_INIT)
+#define SD_NODE_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 8, \
+ .max_interval = 32, \
+ .busy_factor = 32, \
+ .imbalance_pct = 125, \
+ .cache_hot_time = (10*1000000), \
+ .cache_nice_tries = 1, \
+ .per_cpu_gain = 100, \
+ .flags = SD_BALANCE_EXEC \
+ | SD_WAKE_BALANCE, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+#endif
+#endif /* ARCH_HAS_SCHED_TUNE */
+#endif
+
+
#define for_each_domain(cpu, domain) \
for (domain = cpu_rq(cpu)->sd; domain; domain = domain->parent)
spin_unlock_irqrestore(&rq->lock, *flags);
}
+#ifdef CONFIG_SCHEDSTATS
+/*
+ * bump this up when changing the output format or the meaning of an existing
+ * format, so that tools can adapt (or abort)
+ */
+#define SCHEDSTAT_VERSION 10
+
+static int show_schedstat(struct seq_file *seq, void *v)
+{
+ int cpu;
+ enum idle_type itype;
+
+ seq_printf(seq, "version %d\n", SCHEDSTAT_VERSION);
+ seq_printf(seq, "timestamp %lu\n", jiffies);
+ for_each_online_cpu(cpu) {
+ runqueue_t *rq = cpu_rq(cpu);
+#ifdef CONFIG_SMP
+ struct sched_domain *sd;
+ int dcnt = 0;
+#endif
+
+ /* runqueue-specific stats */
+ seq_printf(seq,
+ "cpu%d %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu "
+ "%lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+ cpu, rq->yld_both_empty,
+ rq->yld_act_empty, rq->yld_exp_empty,
+ rq->yld_cnt, rq->sched_noswitch,
+ rq->sched_switch, rq->sched_cnt, rq->sched_goidle,
+ rq->alb_cnt, rq->alb_gained, rq->alb_lost,
+ rq->alb_failed,
+ rq->ttwu_cnt, rq->ttwu_moved, rq->ttwu_attempts,
+ rq->wunt_cnt, rq->wunt_moved,
+ rq->smt_cnt, rq->sbe_cnt, rq->rq_sched_info.cpu_time,
+ rq->rq_sched_info.run_delay, rq->rq_sched_info.pcnt);
+
+ for (itype = IDLE; itype < MAX_IDLE_TYPES; itype++)
+ seq_printf(seq, " %lu %lu", rq->pt_gained[itype],
+ rq->pt_lost[itype]);
+ seq_printf(seq, "\n");
+
+#ifdef CONFIG_SMP
+ /* domain-specific stats */
+ for_each_domain(cpu, sd) {
+ char mask_str[NR_CPUS];
+
+ cpumask_scnprintf(mask_str, NR_CPUS, sd->span);
+ seq_printf(seq, "domain%d %s", dcnt++, mask_str);
+ for (itype = IDLE; itype < MAX_IDLE_TYPES; itype++) {
+ seq_printf(seq, " %lu %lu %lu %lu %lu",
+ sd->lb_cnt[itype],
+ sd->lb_failed[itype],
+ sd->lb_imbalance[itype],
+ sd->lb_nobusyq[itype],
+ sd->lb_nobusyg[itype]);
+ }
+ seq_printf(seq, " %lu %lu %lu %lu\n",
+ sd->sbe_pushed, sd->sbe_attempts,
+ sd->ttwu_wake_affine, sd->ttwu_wake_balance);
+ }
+#endif
+ }
+ return 0;
+}
+
+static int schedstat_open(struct inode *inode, struct file *file)
+{
+ unsigned int size = PAGE_SIZE * (1 + num_online_cpus() / 32);
+ char *buf = kmalloc(size, GFP_KERNEL);
+ struct seq_file *m;
+ int res;
+
+ if (!buf)
+ return -ENOMEM;
+ res = single_open(file, show_schedstat, NULL);
+ if (!res) {
+ m = file->private_data;
+ m->buf = buf;
+ m->size = size;
+ } else
+ kfree(buf);
+ return res;
+}
+
+struct file_operations proc_schedstat_operations = {
+ .open = schedstat_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+# define schedstat_inc(rq, field) rq->field++;
+# define schedstat_add(rq, field, amt) rq->field += amt;
+#else /* !CONFIG_SCHEDSTATS */
+# define schedstat_inc(rq, field) do { } while (0);
+# define schedstat_add(rq, field, amt) do { } while (0);
+#endif
+
/*
* rq_lock - lock a given runqueue and disable interrupts.
*/
spin_unlock_irq(&rq->lock);
}
+#ifdef CONFIG_SCHEDSTATS
+/*
+ * Called when a process is dequeued from the active array and given
+ * the cpu. We should note that with the exception of interactive
+ * tasks, the expired queue will become the active queue after the active
+ * queue is empty, without explicitly dequeuing and requeuing tasks in the
+ * expired queue. (Interactive tasks may be requeued directly to the
+ * active queue, thus delaying tasks in the expired queue from running;
+ * see scheduler_tick()).
+ *
+ * This function is only called from sched_info_arrive(), rather than
+ * dequeue_task(). Even though a task may be queued and dequeued multiple
+ * times as it is shuffled about, we're really interested in knowing how
+ * long it was from the *first* time it was queued to the time that it
+ * finally hit a cpu.
+ */
+static inline void sched_info_dequeued(task_t *t)
+{
+ t->sched_info.last_queued = 0;
+}
+
+/*
+ * Called when a task finally hits the cpu. We can now calculate how
+ * long it was waiting to run. We also note when it began so that we
+ * can keep stats on how long its timeslice is.
+ */
+static inline void sched_info_arrive(task_t *t)
+{
+ unsigned long now = jiffies, diff = 0;
+ struct runqueue *rq = task_rq(t);
+
+ if (t->sched_info.last_queued)
+ diff = now - t->sched_info.last_queued;
+ sched_info_dequeued(t);
+ t->sched_info.run_delay += diff;
+ t->sched_info.last_arrival = now;
+ t->sched_info.pcnt++;
+
+ if (!rq)
+ return;
+
+ rq->rq_sched_info.run_delay += diff;
+ rq->rq_sched_info.pcnt++;
+}
+
+/*
+ * Called when a process is queued into either the active or expired
+ * array. The time is noted and later used to determine how long we
+ * had to wait for us to reach the cpu. Since the expired queue will
+ * become the active queue after active queue is empty, without dequeuing
+ * and requeuing any tasks, we are interested in queuing to either. It
+ * is unusual but not impossible for tasks to be dequeued and immediately
+ * requeued in the same or another array: this can happen in sched_yield(),
+ * set_user_nice(), and even load_balance() as it moves tasks from runqueue
+ * to runqueue.
+ *
+ * This function is only called from enqueue_task(), but also only updates
+ * the timestamp if it is already not set. It's assumed that
+ * sched_info_dequeued() will clear that stamp when appropriate.
+ */
+static inline void sched_info_queued(task_t *t)
+{
+ if (!t->sched_info.last_queued)
+ t->sched_info.last_queued = jiffies;
+}
+
+/*
+ * Called when a process ceases being the active-running process, either
+ * voluntarily or involuntarily. Now we can calculate how long we ran.
+ */
+static inline void sched_info_depart(task_t *t)
+{
+ struct runqueue *rq = task_rq(t);
+ unsigned long diff = jiffies - t->sched_info.last_arrival;
+
+ t->sched_info.cpu_time += diff;
+
+ if (rq)
+ rq->rq_sched_info.cpu_time += diff;
+}
+
+/*
+ * Called when tasks are switched involuntarily due, typically, to expiring
+ * their time slice. (This may also be called when switching to or from
+ * the idle task.) We are only called when prev != next.
+ */
+static inline void sched_info_switch(task_t *prev, task_t *next)
+{
+ struct runqueue *rq = task_rq(prev);
+
+ /*
+ * prev now departs the cpu. It's not interesting to record
+ * stats about how efficient we were at scheduling the idle
+ * process, however.
+ */
+ if (prev != rq->idle)
+ sched_info_depart(prev);
+
+ if (next != rq->idle)
+ sched_info_arrive(next);
+}
+#else
+#define sched_info_queued(t) do { } while (0)
+#define sched_info_switch(t, next) do { } while (0)
+#endif /* CONFIG_SCHEDSTATS */
+
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+static inline ckrm_lrq_t *rq_get_next_class(struct runqueue *rq)
+{
+ cq_node_t *node = classqueue_get_head(&rq->classqueue);
+ return ((node) ? class_list_entry(node) : NULL);
+}
+
+/*
+ * return the cvt of the current running class
+ * if no current running class, return 0
+ * assume cpu is valid (cpu_online(cpu) == 1)
+ */
+CVT_t get_local_cur_cvt(int cpu)
+{
+ ckrm_lrq_t * lrq = rq_get_next_class(cpu_rq(cpu));
+
+ if (lrq)
+ return lrq->local_cvt;
+ else
+ return 0;
+}
+
+static inline struct task_struct * rq_get_next_task(struct runqueue* rq)
+{
+ prio_array_t *array;
+ struct task_struct *next;
+ ckrm_lrq_t *queue;
+ int idx;
+ int cpu = smp_processor_id();
+
+ // it is guaranteed be the ( rq->nr_running > 0 ) check in
+ // schedule that a task will be found.
+
+ retry_next_class:
+ queue = rq_get_next_class(rq);
+ // BUG_ON( !queue );
+
+ array = queue->active;
+ if (unlikely(!array->nr_active)) {
+ queue->active = queue->expired;
+ queue->expired = array;
+ queue->expired_timestamp = 0;
+
+ schedstat_inc(rq, sched_switch);
+ if (queue->active->nr_active)
+ set_top_priority(queue,
+ find_first_bit(queue->active->bitmap, MAX_PRIO));
+ else {
+ classqueue_dequeue(queue->classqueue,
+ &queue->classqueue_linkobj);
+ cpu_demand_event(get_rq_local_stat(queue,cpu),CPU_DEMAND_DEQUEUE,0);
+ }
+ goto retry_next_class;
+ } else
+ schedstat_inc(rq, sched_noswitch);
+ // BUG_ON(!array->nr_active);
+
+ idx = queue->top_priority;
+ // BUG_ON (idx == MAX_PRIO);
+ next = task_list_entry(array->queue[idx].next);
+ return next;
+}
+#else /*! CONFIG_CKRM_CPU_SCHEDULE*/
+static inline struct task_struct * rq_get_next_task(struct runqueue* rq)
+{
+ prio_array_t *array;
+ struct list_head *queue;
+ int idx;
+
+ array = rq->active;
+ if (unlikely(!array->nr_active)) {
+ /*
+ * Switch the active and expired arrays.
+ */
+ schedstat_inc(rq, sched_switch);
+ rq->active = rq->expired;
+ rq->expired = array;
+ array = rq->active;
+ rq->expired_timestamp = 0;
+ rq->best_expired_prio = MAX_PRIO;
+ } else
+ schedstat_inc(rq, sched_noswitch);
+
+ idx = sched_find_first_bit(array->bitmap);
+ queue = array->queue + idx;
+ return list_entry(queue->next, task_t, run_list);
+}
+
+static inline void class_enqueue_task(struct task_struct* p, prio_array_t *array) { }
+static inline void class_dequeue_task(struct task_struct* p, prio_array_t *array) { }
+static inline void init_cpu_classes(void) { }
+#define rq_ckrm_load(rq) NULL
+static inline void ckrm_sched_tick(int j,int this_cpu,void* name) {}
+#endif /* CONFIG_CKRM_CPU_SCHEDULE */
+
/*
* Adding/removing a task to/from a priority array:
*/
list_del(&p->run_list);
if (list_empty(array->queue + p->prio))
__clear_bit(p->prio, array->bitmap);
+ class_dequeue_task(p,array);
}
static void enqueue_task(struct task_struct *p, prio_array_t *array)
{
+ sched_info_queued(p);
list_add_tail(&p->run_list, array->queue + p->prio);
__set_bit(p->prio, array->bitmap);
array->nr_active++;
p->array = array;
+ class_enqueue_task(p,array);
}
/*
__set_bit(p->prio, array->bitmap);
array->nr_active++;
p->array = array;
+ class_enqueue_task(p,array);
}
/*
bonus = CURRENT_BONUS(p) - MAX_BONUS / 2;
prio = p->static_prio - bonus;
+
+#ifdef CONFIG_VSERVER_HARDCPU
+ if (task_vx_flags(p, VXF_SCHED_PRIO, 0))
+ prio += effective_vavavoom(p, MAX_USER_PRIO);
+#endif
+
if (prio < MAX_RT_PRIO)
prio = MAX_RT_PRIO;
if (prio > MAX_PRIO-1)
*/
static inline void __activate_task(task_t *p, runqueue_t *rq)
{
- enqueue_task(p, rq->active);
+ enqueue_task(p, rq_active(p,rq));
rq->nr_running++;
}
*/
static inline void __activate_idle_task(task_t *p, runqueue_t *rq)
{
- enqueue_task_head(p, rq->active);
+ enqueue_task_head(p, rq_active(p,rq));
rq->nr_running++;
}
if (p->mm && p->activated != -1 &&
sleep_time > INTERACTIVE_SLEEP(p)) {
p->sleep_avg = JIFFIES_TO_NS(MAX_SLEEP_AVG -
- AVG_TIMESLICE);
+ DEF_TIMESLICE);
if (!HIGH_CREDIT(p))
p->interactive_credit++;
} else {
}
p->timestamp = now;
+ vx_activate_task(p);
__activate_task(p, rq);
}
/*
* deactivate_task - remove a task from the runqueue.
*/
-static void deactivate_task(struct task_struct *p, runqueue_t *rq)
+static void __deactivate_task(struct task_struct *p, runqueue_t *rq)
{
rq->nr_running--;
if (p->state == TASK_UNINTERRUPTIBLE)
rq->nr_uninterruptible++;
dequeue_task(p, p->array);
+
p->array = NULL;
}
+static void deactivate_task(struct task_struct *p, runqueue_t *rq)
+{
+ __deactivate_task(p, rq);
+ vx_deactivate_task(p);
+}
+
/*
* resched_task - mark a task 'to be rescheduled now'.
*
{
int need_resched, nrpolling;
- preempt_disable();
+ BUG_ON(!spin_is_locked(&task_rq(p)->lock));
+
/* minimise the chance of sending an interrupt to poll_idle() */
nrpolling = test_tsk_thread_flag(p,TIF_POLLING_NRFLAG);
need_resched = test_and_set_tsk_thread_flag(p,TIF_NEED_RESCHED);
if (!need_resched && !nrpolling && (task_cpu(p) != smp_processor_id()))
smp_send_reschedule(task_cpu(p));
- preempt_enable();
}
#else
static inline void resched_task(task_t *p)
* task_curr - is this task currently executing on a CPU?
* @p: the task in question.
*/
-inline int task_curr(task_t *p)
+inline int task_curr(const task_t *p)
{
return cpu_curr(task_cpu(p)) == p;
}
preempt_enable();
}
-EXPORT_SYMBOL_GPL(kick_process);
-
/*
* Return a low guess at the load of a migration-source cpu.
*
if (!(sd->flags & SD_WAKE_IDLE))
return cpu;
- cpus_and(tmp, sd->span, cpu_online_map);
- for_each_cpu_mask(i, tmp) {
- if (!cpu_isset(i, p->cpus_allowed))
- continue;
+ cpus_and(tmp, sd->span, p->cpus_allowed);
+ for_each_cpu_mask(i, tmp) {
if (idle_cpu(i))
return i;
}
#endif
rq = task_rq_lock(p, &flags);
+ schedstat_inc(rq, ttwu_cnt);
old_state = p->state;
if (!(old_state & state))
goto out;
*/
imbalance = sd->imbalance_pct + (sd->imbalance_pct - 100) / 2;
- if ( ((sd->flags & SD_WAKE_AFFINE) &&
- !task_hot(p, rq->timestamp_last_tick, sd))
- || ((sd->flags & SD_WAKE_BALANCE) &&
- imbalance*this_load <= 100*load) ) {
+ if ((sd->flags & SD_WAKE_AFFINE) &&
+ !task_hot(p, rq->timestamp_last_tick, sd)) {
+ /*
+ * This domain has SD_WAKE_AFFINE and p is cache cold
+ * in this domain.
+ */
+ if (cpu_isset(cpu, sd->span)) {
+ schedstat_inc(sd, ttwu_wake_affine);
+ goto out_set_cpu;
+ }
+ } else if ((sd->flags & SD_WAKE_BALANCE) &&
+ imbalance*this_load <= 100*load) {
/*
- * Now sd has SD_WAKE_AFFINE and p is cache cold in sd
- * or sd has SD_WAKE_BALANCE and there is an imbalance
+ * This domain has SD_WAKE_BALANCE and there is
+ * an imbalance.
*/
- if (cpu_isset(cpu, sd->span))
+ if (cpu_isset(cpu, sd->span)) {
+ schedstat_inc(sd, ttwu_wake_balance);
goto out_set_cpu;
+ }
}
}
new_cpu = cpu; /* Could not wake to this_cpu. Wake to cpu instead */
out_set_cpu:
+ schedstat_inc(rq, ttwu_attempts);
new_cpu = wake_idle(new_cpu, p);
if (new_cpu != cpu && cpu_isset(new_cpu, p->cpus_allowed)) {
+ schedstat_inc(rq, ttwu_moved);
set_task_cpu(p, new_cpu);
task_rq_unlock(rq, &flags);
/* might preempt at this point */
int fastcall wake_up_process(task_t * p)
{
- return try_to_wake_up(p, TASK_STOPPED |
+ return try_to_wake_up(p, TASK_STOPPED | TASK_TRACED |
TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE, 0);
}
return try_to_wake_up(p, state, 0);
}
+#ifdef CONFIG_SMP
+static int find_idlest_cpu(struct task_struct *p, int this_cpu,
+ struct sched_domain *sd);
+#endif
+
/*
* Perform scheduler related setup for a newly forked process p.
* p is forked by current.
INIT_LIST_HEAD(&p->run_list);
p->array = NULL;
spin_lock_init(&p->switch_lock);
+#ifdef CONFIG_SCHEDSTATS
+ memset(&p->sched_info, 0, sizeof(p->sched_info));
+#endif
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ cpu_demand_event(&p->demand_stat,CPU_DEMAND_INIT,0);
+#endif
#ifdef CONFIG_PREEMPT
/*
* During context-switch we hold precisely one spinlock, which
p->first_time_slice = 1;
current->time_slice >>= 1;
p->timestamp = sched_clock();
- if (!current->time_slice) {
+ if (unlikely(!current->time_slice)) {
/*
* This case is rare, it happens when the parent has only
* a single jiffy left from its timeslice. Taking the
}
/*
- * wake_up_forked_process - wake up a freshly forked process.
+ * wake_up_new_task - wake up a newly created task for the first time.
*
* This function will do some initial scheduler statistics housekeeping
- * that must be done for every newly created process.
+ * that must be done for every newly created context, then puts the task
+ * on the runqueue and wakes it.
*/
-void fastcall wake_up_forked_process(task_t * p)
+void fastcall wake_up_new_task(task_t * p, unsigned long clone_flags)
{
unsigned long flags;
- runqueue_t *rq = task_rq_lock(current, &flags);
+ int this_cpu, cpu;
+ runqueue_t *rq, *this_rq;
+
+ rq = task_rq_lock(p, &flags);
+ cpu = task_cpu(p);
+ this_cpu = smp_processor_id();
BUG_ON(p->state != TASK_RUNNING);
+ schedstat_inc(rq, wunt_cnt);
/*
* We decrease the sleep average of forking parents
* and children as well, to keep max-interactive tasks
- * from forking tasks that are max-interactive.
+ * from forking tasks that are max-interactive. The parent
+ * (current) is done further down, under its lock.
*/
- current->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(current) *
- PARENT_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
-
p->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(p) *
CHILD_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
p->interactive_credit = 0;
p->prio = effective_prio(p);
- set_task_cpu(p, smp_processor_id());
- if (unlikely(!current->array))
+ vx_activate_task(p);
+ if (likely(cpu == this_cpu)) {
+ if (!(clone_flags & CLONE_VM)) {
+ /*
+ * The VM isn't cloned, so we're in a good position to
+ * do child-runs-first in anticipation of an exec. This
+ * usually avoids a lot of COW overhead.
+ */
+ if (unlikely(!current->array))
+ __activate_task(p, rq);
+ else {
+ p->prio = current->prio;
+ list_add_tail(&p->run_list, ¤t->run_list);
+ p->array = current->array;
+ p->array->nr_active++;
+ rq->nr_running++;
+ class_enqueue_task(p,p->array);
+ }
+ set_need_resched();
+ } else
+ /* Run child last */
+ __activate_task(p, rq);
+ /*
+ * We skip the following code due to cpu == this_cpu
+ *
+ * task_rq_unlock(rq, &flags);
+ * this_rq = task_rq_lock(current, &flags);
+ */
+ this_rq = rq;
+ } else {
+ this_rq = cpu_rq(this_cpu);
+
+ /*
+ * Not the local CPU - must adjust timestamp. This should
+ * get optimised away in the !CONFIG_SMP case.
+ */
+ p->timestamp = (p->timestamp - this_rq->timestamp_last_tick)
+ + rq->timestamp_last_tick;
__activate_task(p, rq);
- else {
- p->prio = current->prio;
- list_add_tail(&p->run_list, ¤t->run_list);
- p->array = current->array;
- p->array->nr_active++;
- rq->nr_running++;
+ if (TASK_PREEMPTS_CURR(p, rq))
+ resched_task(rq->curr);
+
+ schedstat_inc(rq, wunt_moved);
+ /*
+ * Parent and child are on different CPUs, now get the
+ * parent runqueue to update the parent's ->sleep_avg:
+ */
+ task_rq_unlock(rq, &flags);
+ this_rq = task_rq_lock(current, &flags);
}
- task_rq_unlock(rq, &flags);
+ current->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(current) *
+ PARENT_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
+ task_rq_unlock(this_rq, &flags);
}
/*
unsigned long flags;
runqueue_t *rq;
- local_irq_save(flags);
- if (p->first_time_slice) {
- p->parent->time_slice += p->time_slice;
- if (unlikely(p->parent->time_slice > MAX_TIMESLICE))
- p->parent->time_slice = MAX_TIMESLICE;
- }
- local_irq_restore(flags);
/*
* If the child was a (relative-) CPU hog then decrease
* the sleep_avg of the parent as well.
*/
rq = task_rq_lock(p->parent, &flags);
+ if (p->first_time_slice) {
+ p->parent->time_slice += p->time_slice;
+ if (unlikely(p->parent->time_slice > task_timeslice(p)))
+ p->parent->time_slice = task_timeslice(p);
+ }
if (p->sleep_avg < p->parent->sleep_avg)
p->parent->sleep_avg = p->parent->sleep_avg /
(EXIT_WEIGHT + 1) * EXIT_WEIGHT + p->sleep_avg /
/*
* A task struct has one reference for the use as "current".
- * If a task dies, then it sets TASK_ZOMBIE in tsk->state and calls
- * schedule one last time. The schedule call will never return,
+ * If a task dies, then it sets EXIT_ZOMBIE in tsk->exit_state and
+ * calls schedule one last time. The schedule call will never return,
* and the scheduled task must drop that reference.
- * The test for TASK_ZOMBIE must occur while the runqueue locks are
+ * The test for EXIT_ZOMBIE must occur while the runqueue locks are
* still held, otherwise prev could be scheduled on another cpu, die
* there before we look at prev->state, and then the reference would
* be dropped twice.
{
unsigned long i, sum = 0;
- for_each_cpu(i)
+ for_each_online_cpu(i)
sum += cpu_rq(i)->nr_running;
return sum;
{
unsigned long i, sum = 0;
- for_each_online_cpu(i)
+ for_each_cpu(i)
sum += cpu_rq(i)->nr_uninterruptible;
return sum;
{
unsigned long long i, sum = 0;
- for_each_online_cpu(i)
+ for_each_cpu(i)
sum += cpu_rq(i)->nr_switches;
return sum;
{
unsigned long i, sum = 0;
- for_each_online_cpu(i)
+ for_each_cpu(i)
sum += atomic_read(&cpu_rq(i)->nr_iowait);
return sum;
}
+#ifdef CONFIG_SMP
+
/*
* double_rq_lock - safely lock two runqueues
*
spin_unlock(&rq2->lock);
}
-enum idle_type
+/*
+ * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
+ */
+static void double_lock_balance(runqueue_t *this_rq, runqueue_t *busiest)
{
- IDLE,
- NOT_IDLE,
- NEWLY_IDLE,
-};
-
-#ifdef CONFIG_SMP
+ if (unlikely(!spin_trylock(&busiest->lock))) {
+ if (busiest < this_rq) {
+ spin_unlock(&this_rq->lock);
+ spin_lock(&busiest->lock);
+ spin_lock(&this_rq->lock);
+ } else
+ spin_lock(&busiest->lock);
+ }
+}
/*
* find_idlest_cpu - find the least busy runqueue.
min_cpu = UINT_MAX;
min_load = ULONG_MAX;
- cpus_and(mask, sd->span, cpu_online_map);
- cpus_and(mask, mask, p->cpus_allowed);
+ cpus_and(mask, sd->span, p->cpus_allowed);
for_each_cpu_mask(i, mask) {
load = target_load(i);
}
/*
- * wake_up_forked_thread - wake up a freshly forked thread.
- *
- * This function will do some initial scheduler statistics housekeeping
- * that must be done for every newly created context, and it also does
- * runqueue balancing.
+ * If dest_cpu is allowed for this process, migrate the task to it.
+ * This is accomplished by forcing the cpu_allowed mask to only
+ * allow dest_cpu, which will force the cpu onto dest_cpu. Then
+ * the cpu_allowed mask is restored.
*/
-void fastcall wake_up_forked_thread(task_t * p)
+static void sched_migrate_task(task_t *p, int dest_cpu)
{
- unsigned long flags;
- int this_cpu = get_cpu(), cpu;
- struct sched_domain *tmp, *sd = NULL;
- runqueue_t *this_rq = cpu_rq(this_cpu), *rq;
-
- /*
- * Find the largest domain that this CPU is part of that
- * is willing to balance on clone:
- */
- for_each_domain(this_cpu, tmp)
- if (tmp->flags & SD_BALANCE_CLONE)
- sd = tmp;
- if (sd)
- cpu = find_idlest_cpu(p, this_cpu, sd);
- else
- cpu = this_cpu;
-
- local_irq_save(flags);
-lock_again:
- rq = cpu_rq(cpu);
- double_rq_lock(this_rq, rq);
-
- BUG_ON(p->state != TASK_RUNNING);
-
- /*
- * We did find_idlest_cpu() unlocked, so in theory
- * the mask could have changed - just dont migrate
- * in this case:
- */
- if (unlikely(!cpu_isset(cpu, p->cpus_allowed))) {
- cpu = this_cpu;
- double_rq_unlock(this_rq, rq);
- goto lock_again;
- }
- /*
- * We decrease the sleep average of forking parents
- * and children as well, to keep max-interactive tasks
- * from forking tasks that are max-interactive.
- */
- current->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(current) *
- PARENT_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
-
- p->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(p) *
- CHILD_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
-
- p->interactive_credit = 0;
-
- p->prio = effective_prio(p);
- set_task_cpu(p, cpu);
-
- if (cpu == this_cpu) {
- if (unlikely(!current->array))
- __activate_task(p, rq);
- else {
- p->prio = current->prio;
- list_add_tail(&p->run_list, ¤t->run_list);
- p->array = current->array;
- p->array->nr_active++;
- rq->nr_running++;
- }
- } else {
- /* Not the local CPU - must adjust timestamp */
- p->timestamp = (p->timestamp - this_rq->timestamp_last_tick)
- + rq->timestamp_last_tick;
- __activate_task(p, rq);
- if (TASK_PREEMPTS_CURR(p, rq))
- resched_task(rq->curr);
- }
-
- double_rq_unlock(this_rq, rq);
- local_irq_restore(flags);
- put_cpu();
-}
-
-/*
- * If dest_cpu is allowed for this process, migrate the task to it.
- * This is accomplished by forcing the cpu_allowed mask to only
- * allow dest_cpu, which will force the cpu onto dest_cpu. Then
- * the cpu_allowed mask is restored.
- */
-static void sched_migrate_task(task_t *p, int dest_cpu)
-{
- migration_req_t req;
- runqueue_t *rq;
+ migration_req_t req;
+ runqueue_t *rq;
unsigned long flags;
rq = task_rq_lock(p, &flags);
|| unlikely(cpu_is_offline(dest_cpu)))
goto out;
+ schedstat_inc(rq, smt_cnt);
/* force the process onto the specified CPU */
if (migrate_task(p, dest_cpu, &req)) {
/* Need to wait for migration thread (might exit: take ref). */
}
/*
- * sched_balance_exec(): find the highest-level, exec-balance-capable
+ * sched_exec(): find the highest-level, exec-balance-capable
* domain and try to migrate the task to the least loaded CPU.
*
* execve() is a valuable balancing opportunity, because at this point
* the task has the smallest effective memory and cache footprint.
*/
-void sched_balance_exec(void)
+void sched_exec(void)
{
struct sched_domain *tmp, *sd = NULL;
int new_cpu, this_cpu = get_cpu();
+ schedstat_inc(this_rq(), sbe_cnt);
/* Prefer the current CPU if there's only this task running */
if (this_rq()->nr_running <= 1)
goto out;
sd = tmp;
if (sd) {
+ schedstat_inc(sd, sbe_attempts);
new_cpu = find_idlest_cpu(current, this_cpu, sd);
if (new_cpu != this_cpu) {
+ schedstat_inc(sd, sbe_pushed);
put_cpu();
sched_migrate_task(current, new_cpu);
return;
put_cpu();
}
-/*
- * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
- */
-static void double_lock_balance(runqueue_t *this_rq, runqueue_t *busiest)
-{
- if (unlikely(!spin_trylock(&busiest->lock))) {
- if (busiest < this_rq) {
- spin_unlock(&this_rq->lock);
- spin_lock(&busiest->lock);
- spin_lock(&this_rq->lock);
- } else
- spin_lock(&busiest->lock);
- }
-}
-
/*
* pull_task - move a task from a remote runqueue to the local runqueue.
* Both runqueues must be locked.
return 1;
}
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+static inline int ckrm_preferred_task(task_t *tmp,long min, long max,
+ int phase, enum idle_type idle)
+{
+ long pressure = task_load(tmp);
+
+ if (pressure > max)
+ return 0;
+
+ if ((idle == NOT_IDLE) && ! phase && (pressure <= min))
+ return 0;
+ return 1;
+}
+
+/*
+ * move tasks for a specic local class
+ * return number of tasks pulled
+ */
+static inline int ckrm_cls_move_tasks(ckrm_lrq_t* src_lrq,ckrm_lrq_t*dst_lrq,
+ runqueue_t *this_rq,
+ runqueue_t *busiest,
+ struct sched_domain *sd,
+ int this_cpu,
+ enum idle_type idle,
+ long* pressure_imbalance)
+{
+ prio_array_t *array, *dst_array;
+ struct list_head *head, *curr;
+ task_t *tmp;
+ int idx;
+ int pulled = 0;
+ int phase = -1;
+ long pressure_min, pressure_max;
+ /*hzheng: magic : 90% balance is enough*/
+ long balance_min = *pressure_imbalance / 10;
+/*
+ * we don't want to migrate tasks that will reverse the balance
+ * or the tasks that make too small difference
+ */
+#define CKRM_BALANCE_MAX_RATIO 100
+#define CKRM_BALANCE_MIN_RATIO 1
+ start:
+ phase ++;
+ /*
+ * We first consider expired tasks. Those will likely not be
+ * executed in the near future, and they are most likely to
+ * be cache-cold, thus switching CPUs has the least effect
+ * on them.
+ */
+ if (src_lrq->expired->nr_active) {
+ array = src_lrq->expired;
+ dst_array = dst_lrq->expired;
+ } else {
+ array = src_lrq->active;
+ dst_array = dst_lrq->active;
+ }
+
+ new_array:
+ /* Start searching at priority 0: */
+ idx = 0;
+ skip_bitmap:
+ if (!idx)
+ idx = sched_find_first_bit(array->bitmap);
+ else
+ idx = find_next_bit(array->bitmap, MAX_PRIO, idx);
+ if (idx >= MAX_PRIO) {
+ if (array == src_lrq->expired && src_lrq->active->nr_active) {
+ array = src_lrq->active;
+ dst_array = dst_lrq->active;
+ goto new_array;
+ }
+ if ((! phase) && (! pulled) && (idle != IDLE))
+ goto start; //try again
+ else
+ goto out; //finished search for this lrq
+ }
+
+ head = array->queue + idx;
+ curr = head->prev;
+ skip_queue:
+ tmp = list_entry(curr, task_t, run_list);
+
+ curr = curr->prev;
+
+ if (!can_migrate_task(tmp, busiest, this_cpu, sd, idle)) {
+ if (curr != head)
+ goto skip_queue;
+ idx++;
+ goto skip_bitmap;
+ }
+
+ pressure_min = *pressure_imbalance * CKRM_BALANCE_MIN_RATIO/100;
+ pressure_max = *pressure_imbalance * CKRM_BALANCE_MAX_RATIO/100;
+ /*
+ * skip the tasks that will reverse the balance too much
+ */
+ if (ckrm_preferred_task(tmp,pressure_min,pressure_max,phase,idle)) {
+ *pressure_imbalance -= task_load(tmp);
+ pull_task(busiest, array, tmp,
+ this_rq, dst_array, this_cpu);
+ pulled++;
+
+ if (*pressure_imbalance <= balance_min)
+ goto out;
+ }
+
+ if (curr != head)
+ goto skip_queue;
+ idx++;
+ goto skip_bitmap;
+ out:
+ return pulled;
+}
+
+static inline long ckrm_rq_imbalance(runqueue_t *this_rq,runqueue_t *dst_rq)
+{
+ long imbalance;
+ /*
+ * make sure after balance, imbalance' > - imbalance/2
+ * we don't want the imbalance be reversed too much
+ */
+ imbalance = pid_get_pressure(rq_ckrm_load(dst_rq),0)
+ - pid_get_pressure(rq_ckrm_load(this_rq),1);
+ imbalance /= 2;
+ return imbalance;
+}
+
+/*
+ * try to balance the two runqueues
+ *
+ * Called with both runqueues locked.
+ * if move_tasks is called, it will try to move at least one task over
+ */
+static int move_tasks(runqueue_t *this_rq, int this_cpu, runqueue_t *busiest,
+ unsigned long max_nr_move, struct sched_domain *sd,
+ enum idle_type idle)
+{
+ struct ckrm_cpu_class *clsptr,*vip_cls = NULL;
+ ckrm_lrq_t* src_lrq,*dst_lrq;
+ long pressure_imbalance, pressure_imbalance_old;
+ int src_cpu = task_cpu(busiest->curr);
+ struct list_head *list;
+ int pulled = 0;
+ long imbalance;
+
+ imbalance = ckrm_rq_imbalance(this_rq,busiest);
+
+ if ((idle == NOT_IDLE && imbalance <= 0) || busiest->nr_running <= 1)
+ goto out;
+
+ //try to find the vip class
+ list_for_each_entry(clsptr,&active_cpu_classes,links) {
+ src_lrq = get_ckrm_lrq(clsptr,src_cpu);
+
+ if (! lrq_nr_running(src_lrq))
+ continue;
+
+ if (! vip_cls || cpu_class_weight(vip_cls) < cpu_class_weight(clsptr) )
+ {
+ vip_cls = clsptr;
+ }
+ }
+
+ /*
+ * do search from the most significant class
+ * hopefully, less tasks will be migrated this way
+ */
+ clsptr = vip_cls;
+
+ move_class:
+ if (! clsptr)
+ goto out;
+
+
+ src_lrq = get_ckrm_lrq(clsptr,src_cpu);
+ if (! lrq_nr_running(src_lrq))
+ goto other_class;
+
+ dst_lrq = get_ckrm_lrq(clsptr,this_cpu);
+
+ //how much pressure for this class should be transferred
+ pressure_imbalance = src_lrq->lrq_load * imbalance/src_lrq->local_weight;
+ if (pulled && ! pressure_imbalance)
+ goto other_class;
+
+ pressure_imbalance_old = pressure_imbalance;
+
+ //move tasks
+ pulled +=
+ ckrm_cls_move_tasks(src_lrq,dst_lrq,
+ this_rq,
+ busiest,
+ sd,this_cpu,idle,
+ &pressure_imbalance);
+
+ /*
+ * hzheng: 2 is another magic number
+ * stop balancing if the imbalance is less than 25% of the orig
+ */
+ if (pressure_imbalance <= (pressure_imbalance_old >> 2))
+ goto out;
+
+ //update imbalance
+ imbalance *= pressure_imbalance / pressure_imbalance_old;
+ other_class:
+ //who is next?
+ list = clsptr->links.next;
+ if (list == &active_cpu_classes)
+ list = list->next;
+ clsptr = list_entry(list, typeof(*clsptr), links);
+ if (clsptr != vip_cls)
+ goto move_class;
+ out:
+ return pulled;
+}
+
+/**
+ * ckrm_check_balance - is load balancing necessary?
+ * return 0 if load balancing is not necessary
+ * otherwise return the average load of the system
+ * also, update nr_group
+ *
+ * heuristics:
+ * no load balancing if it's load is over average
+ * no load balancing if it's load is far more than the min
+ * task:
+ * read the status of all the runqueues
+ */
+static unsigned long ckrm_check_balance(struct sched_domain *sd, int this_cpu,
+ enum idle_type idle, int* nr_group)
+{
+ struct sched_group *group = sd->groups;
+ unsigned long min_load, max_load, avg_load;
+ unsigned long total_load, this_load, total_pwr;
+
+ max_load = this_load = total_load = total_pwr = 0;
+ min_load = 0xFFFFFFFF;
+ *nr_group = 0;
+
+ do {
+ cpumask_t tmp;
+ unsigned long load;
+ int local_group;
+ int i, nr_cpus = 0;
+
+ /* Tally up the load of all CPUs in the group */
+ cpus_and(tmp, group->cpumask, cpu_online_map);
+ if (unlikely(cpus_empty(tmp)))
+ goto nextgroup;
+
+ avg_load = 0;
+ local_group = cpu_isset(this_cpu, group->cpumask);
+
+ for_each_cpu_mask(i, tmp) {
+ load = pid_get_pressure(rq_ckrm_load(cpu_rq(i)),local_group);
+ nr_cpus++;
+ avg_load += load;
+ }
+
+ if (!nr_cpus)
+ goto nextgroup;
+
+ total_load += avg_load;
+ total_pwr += group->cpu_power;
+
+ /* Adjust by relative CPU power of the group */
+ avg_load = (avg_load * SCHED_LOAD_SCALE) / group->cpu_power;
+
+ if (local_group) {
+ this_load = avg_load;
+ goto nextgroup;
+ } else if (avg_load > max_load) {
+ max_load = avg_load;
+ }
+ if (avg_load < min_load) {
+ min_load = avg_load;
+ }
+nextgroup:
+ group = group->next;
+ *nr_group = *nr_group + 1;
+ } while (group != sd->groups);
+
+ if (!max_load || this_load >= max_load)
+ goto out_balanced;
+
+ avg_load = (SCHED_LOAD_SCALE * total_load) / total_pwr;
+
+ /* hzheng: debugging: 105 is a magic number
+ * 100*max_load <= sd->imbalance_pct*this_load)
+ * should use imbalance_pct instead
+ */
+ if (this_load > avg_load
+ || 100*max_load < 105*this_load
+ || 100*min_load < 70*this_load
+ )
+ goto out_balanced;
+
+ return avg_load;
+ out_balanced:
+ return 0;
+}
+
+/**
+ * any group that has above average load is considered busy
+ * find the busiest queue from any of busy group
+ */
+static runqueue_t *
+ckrm_find_busy_queue(struct sched_domain *sd, int this_cpu,
+ unsigned long avg_load, enum idle_type idle,
+ int nr_group)
+{
+ struct sched_group *group;
+ runqueue_t * busiest=NULL;
+ unsigned long rand;
+
+ group = sd->groups;
+ rand = get_ckrm_rand(nr_group);
+ nr_group = 0;
+
+ do {
+ unsigned long load,total_load,max_load;
+ cpumask_t tmp;
+ int i;
+ runqueue_t * grp_busiest;
+
+ cpus_and(tmp, group->cpumask, cpu_online_map);
+ if (unlikely(cpus_empty(tmp)))
+ goto find_nextgroup;
+
+ total_load = 0;
+ max_load = 0;
+ grp_busiest = NULL;
+ for_each_cpu_mask(i, tmp) {
+ load = pid_get_pressure(rq_ckrm_load(cpu_rq(i)),0);
+ total_load += load;
+ if (load > max_load) {
+ max_load = load;
+ grp_busiest = cpu_rq(i);
+ }
+ }
+
+ total_load = (total_load * SCHED_LOAD_SCALE) / group->cpu_power;
+ if (total_load > avg_load) {
+ busiest = grp_busiest;
+ if (nr_group >= rand)
+ break;
+ }
+ find_nextgroup:
+ group = group->next;
+ nr_group ++;
+ } while (group != sd->groups);
+
+ return busiest;
+}
+
+/**
+ * load_balance - pressure based load balancing algorithm used by ckrm
+ */
+static int ckrm_load_balance(int this_cpu, runqueue_t *this_rq,
+ struct sched_domain *sd, enum idle_type idle)
+{
+ runqueue_t *busiest;
+ unsigned long avg_load;
+ int nr_moved,nr_group;
+
+ avg_load = ckrm_check_balance(sd, this_cpu, idle, &nr_group);
+ if (! avg_load)
+ goto out_balanced;
+
+ busiest = ckrm_find_busy_queue(sd,this_cpu,avg_load,idle,nr_group);
+ if (! busiest)
+ goto out_balanced;
+ /*
+ * This should be "impossible", but since load
+ * balancing is inherently racy and statistical,
+ * it could happen in theory.
+ */
+ if (unlikely(busiest == this_rq)) {
+ WARN_ON(1);
+ goto out_balanced;
+ }
+
+ nr_moved = 0;
+ if (busiest->nr_running > 1) {
+ /*
+ * Attempt to move tasks. If find_busiest_group has found
+ * an imbalance but busiest->nr_running <= 1, the group is
+ * still unbalanced. nr_moved simply stays zero, so it is
+ * correctly treated as an imbalance.
+ */
+ double_lock_balance(this_rq, busiest);
+ nr_moved = move_tasks(this_rq, this_cpu, busiest,
+ 0,sd, idle);
+ spin_unlock(&busiest->lock);
+ if (nr_moved) {
+ adjust_local_weight();
+ }
+ }
+
+ if (!nr_moved)
+ sd->nr_balance_failed ++;
+ else
+ sd->nr_balance_failed = 0;
+
+ /* We were unbalanced, so reset the balancing interval */
+ sd->balance_interval = sd->min_interval;
+
+ return nr_moved;
+
+out_balanced:
+ /* tune up the balancing interval */
+ if (sd->balance_interval < sd->max_interval)
+ sd->balance_interval *= 2;
+
+ return 0;
+}
+
+/*
+ * this_rq->lock is already held
+ */
+static inline int load_balance_newidle(int this_cpu, runqueue_t *this_rq,
+ struct sched_domain *sd)
+{
+ int ret;
+ read_lock(&class_list_lock);
+ ret = ckrm_load_balance(this_cpu,this_rq,sd,NEWLY_IDLE);
+ read_unlock(&class_list_lock);
+ return ret;
+}
+
+static inline int load_balance(int this_cpu, runqueue_t *this_rq,
+ struct sched_domain *sd, enum idle_type idle)
+{
+ int ret;
+
+ spin_lock(&this_rq->lock);
+ read_lock(&class_list_lock);
+ ret= ckrm_load_balance(this_cpu,this_rq,sd,NEWLY_IDLE);
+ read_unlock(&class_list_lock);
+ spin_unlock(&this_rq->lock);
+ return ret;
+}
+#else /*! CONFIG_CKRM_CPU_SCHEDULE */
/*
* move_tasks tries to move up to max_nr_move tasks from busiest to this_rq,
* as part of a balancing operation within "domain". Returns the number of
idx++;
goto skip_bitmap;
}
+
+ /*
+ * Right now, this is the only place pull_task() is called,
+ * so we can safely collect pull_task() stats here rather than
+ * inside pull_task().
+ */
+ schedstat_inc(this_rq, pt_gained[idle]);
+ schedstat_inc(busiest, pt_lost[idle]);
+
pull_task(busiest, array, tmp, this_rq, dst_array, this_cpu);
pulled++;
max_load = this_load = total_load = total_pwr = 0;
do {
- cpumask_t tmp;
unsigned long load;
int local_group;
int i, nr_cpus = 0;
/* Tally up the load of all CPUs in the group */
avg_load = 0;
- cpus_and(tmp, group->cpumask, cpu_online_map);
- if (unlikely(cpus_empty(tmp)))
- goto nextgroup;
- for_each_cpu_mask(i, tmp) {
+ for_each_cpu_mask(i, group->cpumask) {
/* Bias balancing toward cpus of our domain */
if (local_group)
load = target_load(i);
*/
static runqueue_t *find_busiest_queue(struct sched_group *group)
{
- cpumask_t tmp;
unsigned long load, max_load = 0;
runqueue_t *busiest = NULL;
int i;
- cpus_and(tmp, group->cpumask, cpu_online_map);
- for_each_cpu_mask(i, tmp) {
+ for_each_cpu_mask(i, group->cpumask) {
load = source_load(i);
if (load > max_load) {
int nr_moved;
spin_lock(&this_rq->lock);
+ schedstat_inc(sd, lb_cnt[idle]);
group = find_busiest_group(sd, this_cpu, &imbalance, idle);
- if (!group)
+ if (!group) {
+ schedstat_inc(sd, lb_nobusyg[idle]);
goto out_balanced;
+ }
busiest = find_busiest_queue(group);
- if (!busiest)
+ if (!busiest) {
+ schedstat_inc(sd, lb_nobusyq[idle]);
goto out_balanced;
+ }
+
/*
* This should be "impossible", but since load
* balancing is inherently racy and statistical,
goto out_balanced;
}
+ schedstat_add(sd, lb_imbalance[idle], imbalance);
+
nr_moved = 0;
if (busiest->nr_running > 1) {
/*
spin_unlock(&this_rq->lock);
if (!nr_moved) {
+ schedstat_inc(sd, lb_failed[idle]);
sd->nr_balance_failed++;
if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) {
unsigned long imbalance;
int nr_moved = 0;
+ schedstat_inc(sd, lb_cnt[NEWLY_IDLE]);
group = find_busiest_group(sd, this_cpu, &imbalance, NEWLY_IDLE);
- if (!group)
+ if (!group) {
+ schedstat_inc(sd, lb_nobusyg[NEWLY_IDLE]);
goto out;
+ }
busiest = find_busiest_queue(group);
- if (!busiest || busiest == this_rq)
+ if (!busiest || busiest == this_rq) {
+ schedstat_inc(sd, lb_nobusyq[NEWLY_IDLE]);
goto out;
+ }
/* Attempt to move tasks */
double_lock_balance(this_rq, busiest);
+ schedstat_add(sd, lb_imbalance[NEWLY_IDLE], imbalance);
nr_moved = move_tasks(this_rq, this_cpu, busiest,
imbalance, sd, NEWLY_IDLE);
+ if (!nr_moved)
+ schedstat_inc(sd, lb_failed[NEWLY_IDLE]);
spin_unlock(&busiest->lock);
out:
return nr_moved;
}
+#endif /* CONFIG_CKRM_CPU_SCHEDULE*/
+
/*
* idle_balance is called by schedule() if this_cpu is about to become
struct sched_group *group, *busy_group;
int i;
+ schedstat_inc(busiest, alb_cnt);
if (busiest->nr_running <= 1)
return;
for_each_domain(busiest_cpu, sd)
if (cpu_isset(busiest->push_cpu, sd->span))
break;
- if (!sd) {
- WARN_ON(1);
+ if (!sd)
return;
- }
- group = sd->groups;
+ group = sd->groups;
while (!cpu_isset(busiest_cpu, group->cpumask))
- group = group->next;
- busy_group = group;
+ group = group->next;
+ busy_group = group;
- group = sd->groups;
- do {
- cpumask_t tmp;
+ group = sd->groups;
+ do {
runqueue_t *rq;
int push_cpu = 0;
- if (group == busy_group)
- goto next_group;
-
- cpus_and(tmp, group->cpumask, cpu_online_map);
- if (!cpus_weight(tmp))
+ if (group == busy_group)
goto next_group;
- for_each_cpu_mask(i, tmp) {
+ for_each_cpu_mask(i, group->cpumask) {
if (!idle_cpu(i))
goto next_group;
- push_cpu = i;
- }
+ push_cpu = i;
+ }
rq = cpu_rq(push_cpu);
if (unlikely(busiest == rq))
goto next_group;
double_lock_balance(busiest, rq);
- move_tasks(rq, push_cpu, busiest, 1, sd, IDLE);
+ if (move_tasks(rq, push_cpu, busiest, 1, sd, IDLE)) {
+ schedstat_inc(busiest, alb_lost);
+ schedstat_inc(rq, alb_gained);
+ } else {
+ schedstat_inc(busiest, alb_failed);
+ }
spin_unlock(&rq->lock);
next_group:
group = group->next;
}
}
}
-#else
+#else /* SMP*/
/*
* on UP we do not need to balance between CPUs:
*/
static inline int wake_priority_sleeper(runqueue_t *rq)
{
+ int ret = 0;
#ifdef CONFIG_SCHED_SMT
+ spin_lock(&rq->lock);
/*
* If an SMT sibling task has been put to sleep for priority
* reasons reschedule the idle task to see if it can now run.
*/
if (rq->nr_running) {
resched_task(rq->idle);
- return 1;
+ ret = 1;
}
+ spin_unlock(&rq->lock);
#endif
- return 0;
+ return ret;
}
DEFINE_PER_CPU(struct kernel_stat, kstat);
-
EXPORT_PER_CPU_SYMBOL(kstat);
/*
* increasing number of running tasks. We also ignore the interactivity
* if a better static_prio task has expired:
*/
+
+#ifndef CONFIG_CKRM_CPU_SCHEDULE
#define EXPIRED_STARVING(rq) \
((STARVATION_LIMIT && ((rq)->expired_timestamp && \
(jiffies - (rq)->expired_timestamp >= \
STARVATION_LIMIT * ((rq)->nr_running) + 1))) || \
((rq)->curr->static_prio > (rq)->best_expired_prio))
+#else
+#define EXPIRED_STARVING(rq) \
+ (STARVATION_LIMIT && ((rq)->expired_timestamp && \
+ (jiffies - (rq)->expired_timestamp >= \
+ STARVATION_LIMIT * (lrq_nr_running(rq)) + 1)))
+#endif
/*
* This function gets called by the timer code, with HZ frequency.
struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
runqueue_t *rq = this_rq();
task_t *p = current;
+ struct vx_info *vxi = p->vx_info;
rq->timestamp_last_tick = sched_clock();
if (rcu_pending(cpu))
rcu_check_callbacks(cpu, user_ticks);
+
+ if (vxi) {
+ vxi->sched.cpu[cpu].user_ticks += user_ticks;
+ vxi->sched.cpu[cpu].sys_ticks += sys_ticks;
+ }
+
/* note: this timer irq context must be accounted for as well */
if (hardirq_count() - HARDIRQ_OFFSET) {
cpustat->irq += sys_ticks;
if (p == rq->idle) {
if (atomic_read(&rq->nr_iowait) > 0)
cpustat->iowait += sys_ticks;
+ // vx_cpustat_acc(vxi, iowait, cpu, cpustat, sys_ticks);
else
cpustat->idle += sys_ticks;
+ // vx_cpustat_acc(vxi, idle, cpu, cpustat, sys_ticks);
+
if (wake_priority_sleeper(rq))
goto out;
+
+ ckrm_sched_tick(jiffies,cpu,rq_ckrm_load(rq));
+
+#ifdef CONFIG_VSERVER_HARDCPU_IDLE
+ if (!--rq->idle_tokens && !list_empty(&rq->hold_queue))
+ set_need_resched();
+#endif
rebalance_tick(cpu, rq, IDLE);
return;
}
cpustat->system += sys_ticks;
/* Task might have expired already, but not scheduled off yet */
- if (p->array != rq->active) {
+ if (p->array != rq_active(p,rq)) {
set_tsk_need_resched(p);
goto out;
}
* timeslice. This makes it possible for interactive tasks
* to use up their timeslices at their highest priority levels.
*/
- if (unlikely(rt_task(p))) {
+ if (rt_task(p)) {
/*
* RR tasks need a special form of timeslice management.
* FIFO tasks have no timeslices.
set_tsk_need_resched(p);
/* put it at the end of the queue: */
- dequeue_task(p, rq->active);
- enqueue_task(p, rq->active);
+ dequeue_task(p, rq_active(p,rq));
+ enqueue_task(p, rq_active(p,rq));
}
goto out_unlock;
}
- if (!--p->time_slice) {
+ if (vx_need_resched(p)) {
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ /* Hubertus ... we can abstract this out */
+ ckrm_lrq_t* rq = get_task_lrq(p);
+#endif
dequeue_task(p, rq->active);
set_tsk_need_resched(p);
p->prio = effective_prio(p);
rq->expired_timestamp = jiffies;
if (!TASK_INTERACTIVE(p) || EXPIRED_STARVING(rq)) {
enqueue_task(p, rq->expired);
- if (p->static_prio < rq->best_expired_prio)
- rq->best_expired_prio = p->static_prio;
+ if (p->static_prio < this_rq()->best_expired_prio)
+ this_rq()->best_expired_prio = p->static_prio;
} else
enqueue_task(p, rq->active);
} else {
if (TASK_INTERACTIVE(p) && !((task_timeslice(p) -
p->time_slice) % TIMESLICE_GRANULARITY(p)) &&
(p->time_slice >= TIMESLICE_GRANULARITY(p)) &&
- (p->array == rq->active)) {
+ (p->array == rq_active(p,rq))) {
- dequeue_task(p, rq->active);
+ dequeue_task(p, rq_active(p,rq));
set_tsk_need_resched(p);
p->prio = effective_prio(p);
- enqueue_task(p, rq->active);
+ enqueue_task(p, rq_active(p,rq));
}
}
out_unlock:
spin_unlock(&rq->lock);
out:
+ ckrm_sched_tick(jiffies,cpu,rq_ckrm_load(rq));
rebalance_tick(cpu, rq, NOT_IDLE);
}
#ifdef CONFIG_SCHED_SMT
-static inline void wake_sleeping_dependent(int cpu, runqueue_t *rq)
+static inline void wake_sleeping_dependent(int this_cpu, runqueue_t *this_rq)
{
- int i;
- struct sched_domain *sd = rq->sd;
+ struct sched_domain *sd = this_rq->sd;
cpumask_t sibling_map;
+ int i;
if (!(sd->flags & SD_SHARE_CPUPOWER))
return;
- cpus_and(sibling_map, sd->span, cpu_online_map);
- for_each_cpu_mask(i, sibling_map) {
- runqueue_t *smt_rq;
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ if (prev != rq->idle) {
+ unsigned long long run = now - prev->timestamp;
+ ckrm_lrq_t * lrq = get_task_lrq(prev);
- if (i == cpu)
- continue;
+ lrq->lrq_load -= task_load(prev);
+ cpu_demand_event(&prev->demand_stat,CPU_DEMAND_DESCHEDULE,run);
+ lrq->lrq_load += task_load(prev);
+
+ cpu_demand_event(get_task_lrq_stat(prev),CPU_DEMAND_DESCHEDULE,run);
+ update_local_cvt(prev, run);
+ }
+#endif
+ /*
+ * Unlock the current runqueue because we have to lock in
+ * CPU order to avoid deadlocks. Caller knows that we might
+ * unlock. We keep IRQs disabled.
+ */
+ spin_unlock(&this_rq->lock);
- smt_rq = cpu_rq(i);
+ sibling_map = sd->span;
+
+ for_each_cpu_mask(i, sibling_map)
+ spin_lock(&cpu_rq(i)->lock);
+ /*
+ * We clear this CPU from the mask. This both simplifies the
+ * inner loop and keps this_rq locked when we exit:
+ */
+ cpu_clear(this_cpu, sibling_map);
+
+ for_each_cpu_mask(i, sibling_map) {
+ runqueue_t *smt_rq = cpu_rq(i);
/*
* If an SMT sibling task is sleeping due to priority
if (smt_rq->curr == smt_rq->idle && smt_rq->nr_running)
resched_task(smt_rq->idle);
}
+
+ for_each_cpu_mask(i, sibling_map)
+ spin_unlock(&cpu_rq(i)->lock);
+ /*
+ * We exit with this_cpu's rq still held and IRQs
+ * still disabled:
+ */
}
-static inline int dependent_sleeper(int cpu, runqueue_t *rq, task_t *p)
+static inline int dependent_sleeper(int this_cpu, runqueue_t *this_rq)
{
- struct sched_domain *sd = rq->sd;
+ struct sched_domain *sd = this_rq->sd;
cpumask_t sibling_map;
+ prio_array_t *array;
int ret = 0, i;
+ task_t *p;
if (!(sd->flags & SD_SHARE_CPUPOWER))
return 0;
- cpus_and(sibling_map, sd->span, cpu_online_map);
- for_each_cpu_mask(i, sibling_map) {
- runqueue_t *smt_rq;
- task_t *smt_curr;
+ /*
+ * The same locking rules and details apply as for
+ * wake_sleeping_dependent():
+ */
+ spin_unlock(&this_rq->lock);
+ sibling_map = sd->span;
+ for_each_cpu_mask(i, sibling_map)
+ spin_lock(&cpu_rq(i)->lock);
+ cpu_clear(this_cpu, sibling_map);
- if (i == cpu)
- continue;
+ /*
+ * Establish next task to be run - it might have gone away because
+ * we released the runqueue lock above:
+ */
+ if (!this_rq->nr_running)
+ goto out_unlock;
+ array = this_rq->active;
+ if (!array->nr_active)
+ array = this_rq->expired;
+ BUG_ON(!array->nr_active);
- smt_rq = cpu_rq(i);
- smt_curr = smt_rq->curr;
+ p = list_entry(array->queue[sched_find_first_bit(array->bitmap)].next,
+ task_t, run_list);
+
+ for_each_cpu_mask(i, sibling_map) {
+ runqueue_t *smt_rq = cpu_rq(i);
+ task_t *smt_curr = smt_rq->curr;
/*
* If a user task with lower static priority than the
(smt_curr == smt_rq->idle && smt_rq->nr_running))
resched_task(smt_curr);
}
+out_unlock:
+ for_each_cpu_mask(i, sibling_map)
+ spin_unlock(&cpu_rq(i)->lock);
return ret;
}
#else
-static inline void wake_sleeping_dependent(int cpu, runqueue_t *rq)
+static inline void wake_sleeping_dependent(int this_cpu, runqueue_t *this_rq)
{
}
-static inline int dependent_sleeper(int cpu, runqueue_t *rq, task_t *p)
+static inline int dependent_sleeper(int this_cpu, runqueue_t *this_rq)
{
return 0;
}
task_t *prev, *next;
runqueue_t *rq;
prio_array_t *array;
- struct list_head *queue;
unsigned long long now;
unsigned long run_time;
- int cpu, idx;
+ int cpu;
+#ifdef CONFIG_VSERVER_HARDCPU
+ struct vx_info *vxi;
+ int maxidle = -HZ;
+#endif
+
+ /*
+ * If crash dump is in progress, this other cpu's
+ * need to wait until it completes.
+ * NB: this code is optimized away for kernels without
+ * dumping enabled.
+ */
+ if (unlikely(dump_oncpu))
+ goto dump_scheduling_disabled;
+
+ //WARN_ON(system_state == SYSTEM_BOOTING);
/*
* Test if we are atomic. Since do_exit() needs to call into
* schedule() atomically, we ignore that path for now.
* Otherwise, whine if we are scheduling when we should not be.
*/
- if (likely(!(current->state & (TASK_DEAD | TASK_ZOMBIE)))) {
+ if (likely(!(current->exit_state & (EXIT_DEAD | EXIT_ZOMBIE)))) {
if (unlikely(in_atomic())) {
printk(KERN_ERR "bad: scheduling while atomic!\n");
dump_stack();
prev = current;
rq = this_rq();
+ /*
+ * The idle thread is not allowed to schedule!
+ * Remove this check after it has been exercised a bit.
+ */
+ if (unlikely(current == rq->idle) && current->state != TASK_RUNNING) {
+ printk(KERN_ERR "bad: scheduling from the idle thread!\n");
+ dump_stack();
+ }
+
release_kernel_lock(prev);
+ schedstat_inc(rq, sched_cnt);
now = sched_clock();
if (likely(now - prev->timestamp < NS_MAX_SLEEP_AVG))
run_time = now - prev->timestamp;
spin_lock_irq(&rq->lock);
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ if (prev != rq->idle) {
+ unsigned long long run = now - prev->timestamp;
+ ckrm_lrq_t * lrq = get_task_lrq(prev);
+
+ lrq->lrq_load -= task_load(prev);
+ cpu_demand_event(&prev->demand_stat,CPU_DEMAND_DESCHEDULE,run);
+ lrq->lrq_load += task_load(prev);
+
+ cpu_demand_event(get_task_lrq_stat(prev),CPU_DEMAND_DESCHEDULE,run);
+ update_local_cvt(prev, run);
+ }
+#endif
+
+ if (unlikely(current->flags & PF_DEAD))
+ current->state = EXIT_DEAD;
/*
* if entering off of a kernel preemption go straight
* to picking the next task.
deactivate_task(prev, rq);
}
+#ifdef CONFIG_VSERVER_HARDCPU
+ if (!list_empty(&rq->hold_queue)) {
+ struct list_head *l, *n;
+ int ret;
+
+ vxi = NULL;
+ list_for_each_safe(l, n, &rq->hold_queue) {
+ next = list_entry(l, task_t, run_list);
+ if (vxi == next->vx_info)
+ continue;
+
+ vxi = next->vx_info;
+ ret = vx_tokens_recalc(vxi);
+ // tokens = vx_tokens_avail(next);
+
+ if (ret > 0) {
+ list_del(&next->run_list);
+ next->state &= ~TASK_ONHOLD;
+ // one less waiting
+ vx_onhold_dec(vxi);
+ array = rq->expired;
+ next->prio = MAX_PRIO-1;
+ enqueue_task(next, array);
+ rq->nr_running++;
+ if (next->static_prio < rq->best_expired_prio)
+ rq->best_expired_prio = next->static_prio;
+
+ // printk("··· %8lu unhold %p [%d]\n", jiffies, next, next->prio);
+ break;
+ }
+ if ((ret < 0) && (maxidle < ret))
+ maxidle = ret;
+ }
+ }
+ rq->idle_tokens = -maxidle;
+
+pick_next:
+#endif
+
cpu = smp_processor_id();
if (unlikely(!rq->nr_running)) {
+go_idle:
idle_balance(cpu, rq);
if (!rq->nr_running) {
next = rq->idle;
rq->expired_timestamp = 0;
wake_sleeping_dependent(cpu, rq);
+ /*
+ * wake_sleeping_dependent() might have released
+ * the runqueue, so break out if we got new
+ * tasks meanwhile:
+ */
+ if (!rq->nr_running)
+ goto switch_tasks;
+ }
+ } else {
+ if (dependent_sleeper(cpu, rq)) {
+ schedstat_inc(rq, sched_goidle);
+ next = rq->idle;
goto switch_tasks;
}
- }
-
- array = rq->active;
- if (unlikely(!array->nr_active)) {
/*
- * Switch the active and expired arrays.
+ * dependent_sleeper() releases and reacquires the runqueue
+ * lock, hence go into the idle loop if the rq went
+ * empty meanwhile:
*/
- rq->active = rq->expired;
- rq->expired = array;
- array = rq->active;
- rq->expired_timestamp = 0;
- rq->best_expired_prio = MAX_PRIO;
+ if (unlikely(!rq->nr_running))
+ goto go_idle;
}
- idx = sched_find_first_bit(array->bitmap);
- queue = array->queue + idx;
- next = list_entry(queue->next, task_t, run_list);
-
- if (dependent_sleeper(cpu, rq, next)) {
- next = rq->idle;
- goto switch_tasks;
+ /* MEF: CKRM refactored code into rq_get_next_task(); make
+ * sure that when upgrading changes are reflected into both
+ * versions of the code.
+ */
+ next = rq_get_next_task(rq);
+
+#ifdef CONFIG_VSERVER_HARDCPU
+ vxi = next->vx_info;
+ if (vx_info_flags(vxi, VXF_SCHED_PAUSE|VXF_SCHED_HARD, 0)) {
+ int ret = vx_tokens_recalc(vxi);
+
+ if (unlikely(ret <= 0)) {
+ if (ret && (rq->idle_tokens > -ret))
+ rq->idle_tokens = -ret;
+ __deactivate_task(next, rq);
+ recalc_task_prio(next, now);
+ // a new one on hold
+ vx_onhold_inc(vxi);
+ next->state |= TASK_ONHOLD;
+ list_add_tail(&next->run_list, &rq->hold_queue);
+ //printk("··· %8lu hold %p [%d]\n", jiffies, next, next->prio);
+ goto pick_next;
+ }
}
+#endif
if (!rt_task(next) && next->activated > 0) {
unsigned long long delta = now - next->timestamp;
switch_tasks:
prefetch(next);
clear_tsk_need_resched(prev);
- RCU_qsctr(task_cpu(prev))++;
+ rcu_qsctr_inc(task_cpu(prev));
prev->sleep_avg -= run_time;
if ((long)prev->sleep_avg <= 0) {
if (!(HIGH_CREDIT(prev) || LOW_CREDIT(prev)))
prev->interactive_credit--;
}
- prev->timestamp = now;
+ prev->timestamp = prev->last_ran = now;
+ sched_info_switch(prev, next);
if (likely(prev != next)) {
next->timestamp = now;
rq->nr_switches++;
reacquire_kernel_lock(current);
preempt_enable_no_resched();
- if (test_thread_flag(TIF_NEED_RESCHED))
+ if (unlikely(test_thread_flag(TIF_NEED_RESCHED)))
goto need_resched;
+
+ return;
+
+ dump_scheduling_disabled:
+ /* allow scheduling only if this is the dumping cpu */
+ if (dump_oncpu != smp_processor_id()+1) {
+ while (dump_oncpu)
+ cpu_relax();
+ }
+ return;
}
EXPORT_SYMBOL(schedule);
-
#ifdef CONFIG_PREEMPT
/*
* this is is the entry point to schedule() from in-kernel preemption
__remove_wait_queue(q, &wait); \
spin_unlock_irqrestore(&q->lock, flags);
+#define SLEEP_ON_BKLCHECK \
+ if (unlikely(!kernel_locked()) && \
+ sleep_on_bkl_warnings < 10) { \
+ sleep_on_bkl_warnings++; \
+ WARN_ON(1); \
+ }
+
+static int sleep_on_bkl_warnings;
+
void fastcall __sched interruptible_sleep_on(wait_queue_head_t *q)
{
SLEEP_ON_VAR
+ SLEEP_ON_BKLCHECK
+
current->state = TASK_INTERRUPTIBLE;
SLEEP_ON_HEAD
{
SLEEP_ON_VAR
+ SLEEP_ON_BKLCHECK
+
current->state = TASK_INTERRUPTIBLE;
SLEEP_ON_HEAD
EXPORT_SYMBOL(interruptible_sleep_on_timeout);
-void fastcall __sched sleep_on(wait_queue_head_t *q)
-{
- SLEEP_ON_VAR
-
- current->state = TASK_UNINTERRUPTIBLE;
-
- SLEEP_ON_HEAD
- schedule();
- SLEEP_ON_TAIL
-}
-
-EXPORT_SYMBOL(sleep_on);
-
long fastcall __sched sleep_on_timeout(wait_queue_head_t *q, long timeout)
{
SLEEP_ON_VAR
+ SLEEP_ON_BKLCHECK
+
current->state = TASK_UNINTERRUPTIBLE;
SLEEP_ON_HEAD
* and we have a single winner.
*/
if (increment < 0) {
+ if (vx_flags(VXF_IGNEG_NICE, 0))
+ return 0;
if (!capable(CAP_SYS_NICE))
return -EPERM;
if (increment < -40)
* RT tasks are offset by -200. Normal tasks are centered
* around 0, value goes from -16 to +15.
*/
-int task_prio(task_t *p)
+int task_prio(const task_t *p)
{
return p->prio - MAX_RT_PRIO;
}
* task_nice - return the nice value of a given task.
* @p: the task in question.
*/
-int task_nice(task_t *p)
+int task_nice(const task_t *p)
{
return TASK_NICE(p);
}
-EXPORT_SYMBOL(task_nice);
-
/**
* idle_cpu - is a given cpu idle currently?
* @cpu: the processor in question.
{
struct sched_param lp;
int retval = -EINVAL;
- int oldprio;
+ int oldprio, oldpolicy = -1;
prio_array_t *array;
unsigned long flags;
runqueue_t *rq;
retval = -ESRCH;
if (!p)
- goto out_unlock_tasklist;
-
- /*
- * To be able to change p->policy safely, the apropriate
- * runqueue lock must be held.
- */
- rq = task_rq_lock(p, &flags);
-
+ goto out_unlock;
+recheck:
+ /* double check policy once rq lock held */
if (policy < 0)
- policy = p->policy;
+ policy = oldpolicy = p->policy;
else {
retval = -EINVAL;
if (policy != SCHED_FIFO && policy != SCHED_RR &&
policy != SCHED_NORMAL)
goto out_unlock;
}
+ profile_hit(SCHED_PROFILING, __builtin_return_address(0));
/*
* Valid priorities for SCHED_FIFO and SCHED_RR are
if (retval)
goto out_unlock;
+ /*
+ * To be able to change p->policy safely, the apropriate
+ * runqueue lock must be held.
+ */
+ rq = task_rq_lock(p, &flags);
+ /* recheck policy now with rq lock held */
+ if (unlikely(oldpolicy != -1 && oldpolicy != p->policy)) {
+ policy = oldpolicy = -1;
+ task_rq_unlock(rq, &flags);
+ goto recheck;
+ }
array = p->array;
if (array)
deactivate_task(p, task_rq(p));
oldprio = p->prio;
__setscheduler(p, policy, lp.sched_priority);
if (array) {
+ vx_activate_task(p);
__activate_task(p, task_rq(p));
/*
* Reschedule if we are currently running on this runqueue and
} else if (TASK_PREEMPTS_CURR(p, rq))
resched_task(rq->curr);
}
-
-out_unlock:
task_rq_unlock(rq, &flags);
-out_unlock_tasklist:
+out_unlock:
read_unlock_irq(&tasklist_lock);
-
out_nounlock:
return retval;
}
return retval;
}
-/**
- * sys_sched_setaffinity - set the cpu affinity of a process
- * @pid: pid of the process
- * @len: length in bytes of the bitmask pointed to by user_mask_ptr
- * @user_mask_ptr: user-space pointer to the new cpu mask
- */
-asmlinkage long sys_sched_setaffinity(pid_t pid, unsigned int len,
- unsigned long __user *user_mask_ptr)
+long sched_setaffinity(pid_t pid, cpumask_t new_mask)
{
- cpumask_t new_mask;
- int retval;
task_t *p;
-
- if (len < sizeof(new_mask))
- return -EINVAL;
-
- if (copy_from_user(&new_mask, user_mask_ptr, sizeof(new_mask)))
- return -EFAULT;
+ int retval;
lock_cpu_hotplug();
read_lock(&tasklist_lock);
return retval;
}
+static int get_user_cpu_mask(unsigned long __user *user_mask_ptr, unsigned len,
+ cpumask_t *new_mask)
+{
+ if (len < sizeof(cpumask_t)) {
+ memset(new_mask, 0, sizeof(cpumask_t));
+ } else if (len > sizeof(cpumask_t)) {
+ len = sizeof(cpumask_t);
+ }
+ return copy_from_user(new_mask, user_mask_ptr, len) ? -EFAULT : 0;
+}
+
/**
- * sys_sched_getaffinity - get the cpu affinity of a process
+ * sys_sched_setaffinity - set the cpu affinity of a process
* @pid: pid of the process
* @len: length in bytes of the bitmask pointed to by user_mask_ptr
- * @user_mask_ptr: user-space pointer to hold the current cpu mask
+ * @user_mask_ptr: user-space pointer to the new cpu mask
*/
-asmlinkage long sys_sched_getaffinity(pid_t pid, unsigned int len,
+asmlinkage long sys_sched_setaffinity(pid_t pid, unsigned int len,
unsigned long __user *user_mask_ptr)
{
- unsigned int real_len;
- cpumask_t mask;
+ cpumask_t new_mask;
int retval;
- task_t *p;
- real_len = sizeof(mask);
- if (len < real_len)
- return -EINVAL;
+ retval = get_user_cpu_mask(user_mask_ptr, len, &new_mask);
+ if (retval)
+ return retval;
+
+ return sched_setaffinity(pid, new_mask);
+}
+
+/*
+ * Represents all cpu's present in the system
+ * In systems capable of hotplug, this map could dynamically grow
+ * as new cpu's are detected in the system via any platform specific
+ * method, such as ACPI for e.g.
+ */
+
+cpumask_t cpu_present_map;
+EXPORT_SYMBOL(cpu_present_map);
+
+#ifndef CONFIG_SMP
+cpumask_t cpu_online_map = CPU_MASK_ALL;
+cpumask_t cpu_possible_map = CPU_MASK_ALL;
+#endif
+
+long sched_getaffinity(pid_t pid, cpumask_t *mask)
+{
+ int retval;
+ task_t *p;
lock_cpu_hotplug();
read_lock(&tasklist_lock);
goto out_unlock;
retval = 0;
- cpus_and(mask, p->cpus_allowed, cpu_possible_map);
+ cpus_and(*mask, p->cpus_allowed, cpu_possible_map);
out_unlock:
read_unlock(&tasklist_lock);
unlock_cpu_hotplug();
if (retval)
return retval;
- if (copy_to_user(user_mask_ptr, &mask, real_len))
+
+ return 0;
+}
+
+/**
+ * sys_sched_getaffinity - get the cpu affinity of a process
+ * @pid: pid of the process
+ * @len: length in bytes of the bitmask pointed to by user_mask_ptr
+ * @user_mask_ptr: user-space pointer to hold the current cpu mask
+ */
+asmlinkage long sys_sched_getaffinity(pid_t pid, unsigned int len,
+ unsigned long __user *user_mask_ptr)
+{
+ int ret;
+ cpumask_t mask;
+
+ if (len < sizeof(cpumask_t))
+ return -EINVAL;
+
+ ret = sched_getaffinity(pid, &mask);
+ if (ret < 0)
+ return ret;
+
+ if (copy_to_user(user_mask_ptr, &mask, sizeof(cpumask_t)))
return -EFAULT;
- return real_len;
+
+ return sizeof(cpumask_t);
}
/**
{
runqueue_t *rq = this_rq_lock();
prio_array_t *array = current->array;
- prio_array_t *target = rq->expired;
+ prio_array_t *target = rq_expired(current,rq);
+ schedstat_inc(rq, yld_cnt);
/*
* We implement yielding by moving the task into the expired
* queue.
* (special rule: RT tasks will just roundrobin in the active
* array.)
*/
- if (unlikely(rt_task(current)))
- target = rq->active;
+ if (rt_task(current))
+ target = rq_active(current,rq);
+
+#warning MEF need to fix up SCHEDSTATS code, but I hope this is fixed by the 2.6.10 CKRM patch
+#ifdef CONFIG_SCHEDSTATS
+ if (current->array->nr_active == 1) {
+ schedstat_inc(rq, yld_act_empty);
+ if (!rq->expired->nr_active)
+ schedstat_inc(rq, yld_both_empty);
+ } else if (!rq->expired->nr_active)
+ schedstat_inc(rq, yld_exp_empty);
+#endif
dequeue_task(current, array);
enqueue_task(current, target);
void __sched __cond_resched(void)
{
- set_current_state(TASK_RUNNING);
- schedule();
+#ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
+ __might_sleep(__FILE__, __LINE__, 0);
+#endif
+ /*
+ * The system_state check is somewhat ugly but we might be
+ * called during early boot when we are not yet ready to reschedule.
+ */
+ if (need_resched() && system_state >= SYSTEM_BOOTING_SCHEDULER_OK) {
+ set_current_state(TASK_RUNNING);
+ schedule();
+ }
+
}
EXPORT_SYMBOL(__cond_resched);
+void __sched __cond_resched_lock(spinlock_t * lock)
+{
+ if (need_resched()) {
+ _raw_spin_unlock(lock);
+ preempt_enable_no_resched();
+ set_current_state(TASK_RUNNING);
+ schedule();
+ spin_lock(lock);
+ }
+}
+
+EXPORT_SYMBOL(__cond_resched_lock);
+
+
/**
* yield - yield the current processor to other threads.
*
task_t *relative;
unsigned state;
unsigned long free = 0;
- static const char *stat_nam[] = { "R", "S", "D", "T", "Z", "W" };
+ static const char *stat_nam[] = { "R", "S", "D", "T", "t", "Z", "X" };
printk("%-13.13s ", p->comm);
state = p->state ? __ffs(p->state) + 1 : 0;
void __devinit init_idle(task_t *idle, int cpu)
{
- runqueue_t *idle_rq = cpu_rq(cpu), *rq = cpu_rq(task_cpu(idle));
+ runqueue_t *rq = cpu_rq(cpu);
unsigned long flags;
- local_irq_save(flags);
- double_rq_lock(idle_rq, rq);
-
- idle_rq->curr = idle_rq->idle = idle;
- deactivate_task(idle, rq);
+ idle->sleep_avg = 0;
+ idle->interactive_credit = 0;
idle->array = NULL;
idle->prio = MAX_PRIO;
idle->state = TASK_RUNNING;
set_task_cpu(idle, cpu);
- double_rq_unlock(idle_rq, rq);
+
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ cpu_demand_event(&(idle->demand_stat),CPU_DEMAND_INIT,0);
+ idle->cpu_class = get_default_cpu_class();
+ idle->array = NULL;
+#endif
+
+ spin_lock_irqsave(&rq->lock, flags);
+ rq->curr = rq->idle = idle;
set_tsk_need_resched(idle);
- local_irq_restore(flags);
+ spin_unlock_irqrestore(&rq->lock, flags);
/* Set the preempt count _outside_ the spinlocks! */
#ifdef CONFIG_PREEMPT
runqueue_t *rq;
rq = task_rq_lock(p, &flags);
- if (any_online_cpu(new_mask) == NR_CPUS) {
+ if (!cpus_intersects(new_mask, cpu_online_map)) {
ret = -EINVAL;
goto out;
}
task_rq_unlock(rq, &flags);
wake_up_process(rq->migration_thread);
wait_for_completion(&req.done);
+ tlb_migrate_finish(p->mm);
return 0;
}
out:
* Move (not current) task off this cpu, onto dest cpu. We're doing
* this because either it can't run here any more (set_cpus_allowed()
* away from this CPU, or CPU going down), or because we're
- * attempting to rebalance this task on exec (sched_balance_exec).
+ * attempting to rebalance this task on exec (sched_exec).
*
* So we race with normal scheduler movements, but that's OK, as long
* as the task is no longer on this CPU.
if (unlikely(cpu_is_offline(dest_cpu)))
return;
- rq_src = cpu_rq(src_cpu);
+ rq_src = cpu_rq(src_cpu);
rq_dest = cpu_rq(dest_cpu);
double_rq_lock(rq_src, rq_dest);
if (!cpu_isset(dest_cpu, p->cpus_allowed))
goto out;
- set_task_cpu(p, dest_cpu);
if (p->array) {
/*
* Sync timestamp with rq_dest's before activating.
p->timestamp = p->timestamp - rq_src->timestamp_last_tick
+ rq_dest->timestamp_last_tick;
deactivate_task(p, rq_src);
+ set_task_cpu(p, dest_cpu);
activate_task(p, rq_dest, 0);
if (TASK_PREEMPTS_CURR(p, rq_dest))
resched_task(rq_dest->curr);
- }
+ } else
+ set_task_cpu(p, dest_cpu);
out:
double_rq_unlock(rq_src, rq_dest);
return 0;
}
-#ifdef CONFIG_HOTPLUG_CPU
-/* migrate_all_tasks - function to migrate all tasks from the dead cpu. */
-static void migrate_all_tasks(int src_cpu)
+#ifdef CONFIG_HOTPLUG_CPU
+/* Figure out where task on dead CPU should go, use force if neccessary. */
+static void move_task_off_dead_cpu(int dead_cpu, struct task_struct *tsk)
+{
+ int dest_cpu;
+ cpumask_t mask;
+
+ /* On same node? */
+ mask = node_to_cpumask(cpu_to_node(dead_cpu));
+ cpus_and(mask, mask, tsk->cpus_allowed);
+ dest_cpu = any_online_cpu(mask);
+
+ /* On any allowed CPU? */
+ if (dest_cpu == NR_CPUS)
+ dest_cpu = any_online_cpu(tsk->cpus_allowed);
+
+ /* No more Mr. Nice Guy. */
+ if (dest_cpu == NR_CPUS) {
+ cpus_setall(tsk->cpus_allowed);
+ dest_cpu = any_online_cpu(tsk->cpus_allowed);
+
+ /*
+ * Don't tell them about moving exiting tasks or
+ * kernel threads (both mm NULL), since they never
+ * leave kernel.
+ */
+ if (tsk->mm && printk_ratelimit())
+ printk(KERN_INFO "process %d (%s) no "
+ "longer affine to cpu%d\n",
+ tsk->pid, tsk->comm, dead_cpu);
+ }
+ __migrate_task(tsk, dead_cpu, dest_cpu);
+}
+
+/* Run through task list and migrate tasks from the dead cpu. */
+static void migrate_live_tasks(int src_cpu)
{
struct task_struct *tsk, *t;
- int dest_cpu;
- unsigned int node;
write_lock_irq(&tasklist_lock);
- /* watch out for per node tasks, let's stay on this node */
- node = cpu_to_node(src_cpu);
-
do_each_thread(t, tsk) {
- cpumask_t mask;
if (tsk == current)
continue;
- if (task_cpu(tsk) != src_cpu)
- continue;
-
- /* Figure out where this task should go (attempting to
- * keep it on-node), and check if it can be migrated
- * as-is. NOTE that kernel threads bound to more than
- * one online cpu will be migrated. */
- mask = node_to_cpumask(node);
- cpus_and(mask, mask, tsk->cpus_allowed);
- dest_cpu = any_online_cpu(mask);
- if (dest_cpu == NR_CPUS)
- dest_cpu = any_online_cpu(tsk->cpus_allowed);
- if (dest_cpu == NR_CPUS) {
- cpus_clear(tsk->cpus_allowed);
- cpus_complement(tsk->cpus_allowed);
- dest_cpu = any_online_cpu(tsk->cpus_allowed);
-
- /* Don't tell them about moving exiting tasks
- or kernel threads (both mm NULL), since
- they never leave kernel. */
- if (tsk->mm && printk_ratelimit())
- printk(KERN_INFO "process %d (%s) no "
- "longer affine to cpu%d\n",
- tsk->pid, tsk->comm, src_cpu);
- }
-
- __migrate_task(tsk, src_cpu, dest_cpu);
+ if (task_cpu(tsk) == src_cpu)
+ move_task_off_dead_cpu(src_cpu, tsk);
} while_each_thread(t, tsk);
write_unlock_irq(&tasklist_lock);
spin_unlock_irqrestore(&rq->lock, flags);
}
+
+static void migrate_dead(unsigned int dead_cpu, task_t *tsk)
+{
+ struct runqueue *rq = cpu_rq(dead_cpu);
+
+ /* Must be exiting, otherwise would be on tasklist. */
+ BUG_ON(tsk->exit_state != EXIT_ZOMBIE && tsk->exit_state != EXIT_DEAD);
+
+ /* Cannot have done final schedule yet: would have vanished. */
+ BUG_ON(tsk->flags & PF_DEAD);
+
+ get_task_struct(tsk);
+
+ /*
+ * Drop lock around migration; if someone else moves it,
+ * that's OK. No task can be added to this CPU, so iteration is
+ * fine.
+ */
+ spin_unlock_irq(&rq->lock);
+ move_task_off_dead_cpu(dead_cpu, tsk);
+ spin_lock_irq(&rq->lock);
+
+ put_task_struct(tsk);
+}
+
+/* release_task() removes task from tasklist, so we won't find dead tasks. */
+static void migrate_dead_tasks(unsigned int dead_cpu)
+{
+ unsigned arr, i;
+ struct runqueue *rq = cpu_rq(dead_cpu);
+
+ for (arr = 0; arr < 2; arr++) {
+ for (i = 0; i < MAX_PRIO; i++) {
+ struct list_head *list = &rq->arrays[arr].queue[i];
+ while (!list_empty(list))
+ migrate_dead(dead_cpu,
+ list_entry(list->next, task_t,
+ run_list));
+ }
+ }
+}
#endif /* CONFIG_HOTPLUG_CPU */
/*
p = kthread_create(migration_thread, hcpu, "migration/%d",cpu);
if (IS_ERR(p))
return NOTIFY_BAD;
+ p->flags |= PF_NOFREEZE;
kthread_bind(p, cpu);
/* Must be high prio: stop_machine expects to yield to it. */
rq = task_rq_lock(p, &flags);
cpu_rq(cpu)->migration_thread = NULL;
break;
case CPU_DEAD:
- migrate_all_tasks(cpu);
+ migrate_live_tasks(cpu);
rq = cpu_rq(cpu);
kthread_stop(rq->migration_thread);
rq->migration_thread = NULL;
deactivate_task(rq->idle, rq);
rq->idle->static_prio = MAX_PRIO;
__setscheduler(rq->idle, SCHED_NORMAL, 0);
+ migrate_dead_tasks(cpu);
task_rq_unlock(rq, &flags);
- BUG_ON(rq->nr_running != 0);
+ BUG_ON(rq->nr_running != 0);
/* No need to migrate the tasks: it was best-effort if
* they didn't do lock_cpu_hotplug(). Just wake up
complete(&req->done);
}
spin_unlock_irq(&rq->lock);
- break;
+ break;
#endif
}
return NOTIFY_OK;
EXPORT_SYMBOL(kernel_flag);
#ifdef CONFIG_SMP
-/* Attach the domain 'sd' to 'cpu' as its base domain */
-void cpu_attach_domain(struct sched_domain *sd, int cpu)
+/*
+ * Attach the domain 'sd' to 'cpu' as its base domain. Callers must
+ * hold the hotplug lock.
+ */
+static void cpu_attach_domain(struct sched_domain *sd, int cpu)
{
migration_req_t req;
unsigned long flags;
runqueue_t *rq = cpu_rq(cpu);
int local = 1;
- lock_cpu_hotplug();
-
spin_lock_irqsave(&rq->lock, flags);
if (cpu == smp_processor_id() || !cpu_online(cpu)) {
wake_up_process(rq->migration_thread);
wait_for_completion(&req.done);
}
+}
- unlock_cpu_hotplug();
+/*
+ * To enable disjoint top-level NUMA domains, define SD_NODES_PER_DOMAIN
+ * in arch code. That defines the number of nearby nodes in a node's top
+ * level scheduling domain.
+ */
+#ifdef CONFIG_NUMA
+#ifdef SD_NODES_PER_DOMAIN
+/**
+ * find_next_best_node - find the next node to include in a sched_domain
+ * @node: node whose sched_domain we're building
+ * @used_nodes: nodes already in the sched_domain
+ *
+ * Find the next node to include in a given scheduling domain. Simply
+ * finds the closest node not already in the @used_nodes map.
+ *
+ * Should use nodemask_t.
+ */
+static int __devinit find_next_best_node(int node, unsigned long *used_nodes)
+{
+ int i, n, val, min_val, best_node = 0;
+
+ min_val = INT_MAX;
+
+ for (i = 0; i < numnodes; i++) {
+ /* Start at @node */
+ n = (node + i) % numnodes;
+
+ /* Skip already used nodes */
+ if (test_bit(n, used_nodes))
+ continue;
+
+ /* Simple min distance search */
+ val = node_distance(node, i);
+
+ if (val < min_val) {
+ min_val = val;
+ best_node = n;
+ }
+ }
+
+ set_bit(best_node, used_nodes);
+ return best_node;
}
-#ifdef ARCH_HAS_SCHED_DOMAIN
-extern void __init arch_init_sched_domains(void);
-#else
-static struct sched_group sched_group_cpus[NR_CPUS];
+/**
+ * sched_domain_node_span - get a cpumask for a node's sched_domain
+ * @node: node whose cpumask we're constructing
+ * @size: number of nodes to include in this span
+ *
+ * Given a node, construct a good cpumask for its sched_domain to span. It
+ * should be one that prevents unnecessary balancing, but also spreads tasks
+ * out optimally.
+ */
+static cpumask_t __devinit sched_domain_node_span(int node)
+{
+ int i;
+ cpumask_t span;
+ DECLARE_BITMAP(used_nodes, MAX_NUMNODES);
+
+ cpus_clear(span);
+ bitmap_zero(used_nodes, MAX_NUMNODES);
+
+ for (i = 0; i < SD_NODES_PER_DOMAIN; i++) {
+ int next_node = find_next_best_node(node, used_nodes);
+ cpumask_t nodemask;
+
+ nodemask = node_to_cpumask(next_node);
+ cpus_or(span, span, nodemask);
+ }
+
+ return span;
+}
+#else /* SD_NODES_PER_DOMAIN */
+static cpumask_t __devinit sched_domain_node_span(int node)
+{
+ return cpu_possible_map;
+}
+#endif /* SD_NODES_PER_DOMAIN */
+#endif /* CONFIG_NUMA */
+
+#ifdef CONFIG_SCHED_SMT
static DEFINE_PER_CPU(struct sched_domain, cpu_domains);
+static struct sched_group sched_group_cpus[NR_CPUS];
+static int __devinit cpu_to_cpu_group(int cpu)
+{
+ return cpu;
+}
+#endif
+
+static DEFINE_PER_CPU(struct sched_domain, phys_domains);
+static struct sched_group sched_group_phys[NR_CPUS];
+static int __devinit cpu_to_phys_group(int cpu)
+{
+#ifdef CONFIG_SCHED_SMT
+ return first_cpu(cpu_sibling_map[cpu]);
+#else
+ return cpu;
+#endif
+}
+
#ifdef CONFIG_NUMA
-static struct sched_group sched_group_nodes[MAX_NUMNODES];
+
static DEFINE_PER_CPU(struct sched_domain, node_domains);
-static void __init arch_init_sched_domains(void)
+static struct sched_group sched_group_nodes[MAX_NUMNODES];
+static int __devinit cpu_to_node_group(int cpu)
{
- int i;
- struct sched_group *first_node = NULL, *last_node = NULL;
+ return cpu_to_node(cpu);
+}
+#endif
- /* Set up domains */
- for_each_cpu(i) {
- int node = cpu_to_node(i);
- cpumask_t nodemask = node_to_cpumask(node);
- struct sched_domain *node_sd = &per_cpu(node_domains, i);
- struct sched_domain *cpu_sd = &per_cpu(cpu_domains, i);
+/* Groups for isolated scheduling domains */
+static struct sched_group sched_group_isolated[NR_CPUS];
- *node_sd = SD_NODE_INIT;
- node_sd->span = cpu_possible_map;
- node_sd->groups = &sched_group_nodes[cpu_to_node(i)];
+/* cpus with isolated domains */
+cpumask_t __devinitdata cpu_isolated_map = CPU_MASK_NONE;
- *cpu_sd = SD_CPU_INIT;
- cpus_and(cpu_sd->span, nodemask, cpu_possible_map);
- cpu_sd->groups = &sched_group_cpus[i];
- cpu_sd->parent = node_sd;
- }
+static int __devinit cpu_to_isolated_group(int cpu)
+{
+ return cpu;
+}
- /* Set up groups */
- for (i = 0; i < MAX_NUMNODES; i++) {
- cpumask_t tmp = node_to_cpumask(i);
- cpumask_t nodemask;
- struct sched_group *first_cpu = NULL, *last_cpu = NULL;
- struct sched_group *node = &sched_group_nodes[i];
- int j;
+/* Setup the mask of cpus configured for isolated domains */
+static int __init isolated_cpu_setup(char *str)
+{
+ int ints[NR_CPUS], i;
- cpus_and(nodemask, tmp, cpu_possible_map);
+ str = get_options(str, ARRAY_SIZE(ints), ints);
+ cpus_clear(cpu_isolated_map);
+ for (i = 1; i <= ints[0]; i++)
+ cpu_set(ints[i], cpu_isolated_map);
+ return 1;
+}
- if (cpus_empty(nodemask))
- continue;
+__setup ("isolcpus=", isolated_cpu_setup);
- node->cpumask = nodemask;
- node->cpu_power = SCHED_LOAD_SCALE * cpus_weight(node->cpumask);
+/*
+ * init_sched_build_groups takes an array of groups, the cpumask we wish
+ * to span, and a pointer to a function which identifies what group a CPU
+ * belongs to. The return value of group_fn must be a valid index into the
+ * groups[] array, and must be >= 0 and < NR_CPUS (due to the fact that we
+ * keep track of groups covered with a cpumask_t).
+ *
+ * init_sched_build_groups will build a circular linked list of the groups
+ * covered by the given span, and will set each group's ->cpumask correctly,
+ * and ->cpu_power to 0.
+ */
+static void __devinit init_sched_build_groups(struct sched_group groups[],
+ cpumask_t span, int (*group_fn)(int cpu))
+{
+ struct sched_group *first = NULL, *last = NULL;
+ cpumask_t covered = CPU_MASK_NONE;
+ int i;
- for_each_cpu_mask(j, node->cpumask) {
- struct sched_group *cpu = &sched_group_cpus[j];
+ for_each_cpu_mask(i, span) {
+ int group = group_fn(i);
+ struct sched_group *sg = &groups[group];
+ int j;
- cpus_clear(cpu->cpumask);
- cpu_set(j, cpu->cpumask);
- cpu->cpu_power = SCHED_LOAD_SCALE;
+ if (cpu_isset(i, covered))
+ continue;
- if (!first_cpu)
- first_cpu = cpu;
- if (last_cpu)
- last_cpu->next = cpu;
- last_cpu = cpu;
- }
- last_cpu->next = first_cpu;
+ sg->cpumask = CPU_MASK_NONE;
+ sg->cpu_power = 0;
- if (!first_node)
- first_node = node;
- if (last_node)
- last_node->next = node;
- last_node = node;
- }
- last_node->next = first_node;
+ for_each_cpu_mask(j, span) {
+ if (group_fn(j) != group)
+ continue;
- mb();
- for_each_cpu(i) {
- struct sched_domain *cpu_sd = &per_cpu(cpu_domains, i);
- cpu_attach_domain(cpu_sd, i);
+ cpu_set(j, covered);
+ cpu_set(j, sg->cpumask);
+ }
+ if (!first)
+ first = sg;
+ if (last)
+ last->next = sg;
+ last = sg;
}
+ last->next = first;
}
-#else /* !CONFIG_NUMA */
-static void __init arch_init_sched_domains(void)
+/*
+ * Set up scheduler domains and groups. Callers must hold the hotplug lock.
+ */
+static void __devinit arch_init_sched_domains(void)
{
int i;
- struct sched_group *first_cpu = NULL, *last_cpu = NULL;
+ cpumask_t cpu_default_map;
+ cpumask_t cpu_isolated_online_map;
+
+ cpus_and(cpu_isolated_online_map, cpu_isolated_map, cpu_online_map);
+
+ /*
+ * Setup mask for cpus without special case scheduling requirements.
+ * For now this just excludes isolated cpus, but could be used to
+ * exclude other special cases in the future.
+ */
+ cpus_complement(cpu_default_map, cpu_isolated_map);
+ cpus_and(cpu_default_map, cpu_default_map, cpu_online_map);
/* Set up domains */
- for_each_cpu(i) {
- struct sched_domain *cpu_sd = &per_cpu(cpu_domains, i);
+ for_each_online_cpu(i) {
+ int group;
+ struct sched_domain *sd = NULL, *p;
+ cpumask_t nodemask = node_to_cpumask(cpu_to_node(i));
+
+ cpus_and(nodemask, nodemask, cpu_default_map);
+
+ /*
+ * Set up isolated domains.
+ * Unlike those of other cpus, the domains and groups are
+ * single level, and span a single cpu.
+ */
+ if (cpu_isset(i, cpu_isolated_online_map)) {
+#ifdef CONFIG_SCHED_SMT
+ sd = &per_cpu(cpu_domains, i);
+#else
+ sd = &per_cpu(phys_domains, i);
+#endif
+ group = cpu_to_isolated_group(i);
+ *sd = SD_CPU_INIT;
+ cpu_set(i, sd->span);
+ sd->balance_interval = INT_MAX; /* Don't balance */
+ sd->flags = 0; /* Avoid WAKE_ */
+ sd->groups = &sched_group_isolated[group];
+ printk(KERN_INFO "Setting up cpu %d isolated.\n", i);
+ /* Single level, so continue with next cpu */
+ continue;
+ }
+
+#ifdef CONFIG_NUMA
+ sd = &per_cpu(node_domains, i);
+ group = cpu_to_node_group(i);
+ *sd = SD_NODE_INIT;
+ /* FIXME: should be multilevel, in arch code */
+ sd->span = sched_domain_node_span(i);
+ cpus_and(sd->span, sd->span, cpu_default_map);
+ sd->groups = &sched_group_nodes[group];
+#endif
+
+ p = sd;
+ sd = &per_cpu(phys_domains, i);
+ group = cpu_to_phys_group(i);
+ *sd = SD_CPU_INIT;
+ sd->span = nodemask;
+ sd->parent = p;
+ sd->groups = &sched_group_phys[group];
+
+#ifdef CONFIG_SCHED_SMT
+ p = sd;
+ sd = &per_cpu(cpu_domains, i);
+ group = cpu_to_cpu_group(i);
+ *sd = SD_SIBLING_INIT;
+ sd->span = cpu_sibling_map[i];
+ cpus_and(sd->span, sd->span, cpu_default_map);
+ sd->parent = p;
+ sd->groups = &sched_group_cpus[group];
+#endif
+ }
+
+#ifdef CONFIG_SCHED_SMT
+ /* Set up CPU (sibling) groups */
+ for_each_online_cpu(i) {
+ cpumask_t this_sibling_map = cpu_sibling_map[i];
+ cpus_and(this_sibling_map, this_sibling_map, cpu_default_map);
+ if (i != first_cpu(this_sibling_map))
+ continue;
+
+ init_sched_build_groups(sched_group_cpus, this_sibling_map,
+ &cpu_to_cpu_group);
+ }
+#endif
- *cpu_sd = SD_CPU_INIT;
- cpu_sd->span = cpu_possible_map;
- cpu_sd->groups = &sched_group_cpus[i];
+ /* Set up isolated groups */
+ for_each_cpu_mask(i, cpu_isolated_online_map) {
+ cpumask_t mask = cpumask_of_cpu(i);
+ init_sched_build_groups(sched_group_isolated, mask,
+ &cpu_to_isolated_group);
}
- /* Set up CPU groups */
- for_each_cpu_mask(i, cpu_possible_map) {
- struct sched_group *cpu = &sched_group_cpus[i];
+ /* Set up physical groups */
+ for (i = 0; i < MAX_NUMNODES; i++) {
+ cpumask_t nodemask = node_to_cpumask(i);
+
+ cpus_and(nodemask, nodemask, cpu_default_map);
+ if (cpus_empty(nodemask))
+ continue;
+
+ init_sched_build_groups(sched_group_phys, nodemask,
+ &cpu_to_phys_group);
+ }
+
+
+#ifdef CONFIG_NUMA
+ /* Set up node groups */
+ init_sched_build_groups(sched_group_nodes, cpu_default_map,
+ &cpu_to_node_group);
+#endif
+
+
+ /* Calculate CPU power for physical packages and nodes */
+ for_each_cpu_mask(i, cpu_default_map) {
+ int power;
+ struct sched_domain *sd;
+#ifdef CONFIG_SCHED_SMT
+ sd = &per_cpu(cpu_domains, i);
+ power = SCHED_LOAD_SCALE;
+ sd->groups->cpu_power = power;
+#endif
+
+ sd = &per_cpu(phys_domains, i);
+ power = SCHED_LOAD_SCALE + SCHED_LOAD_SCALE *
+ (cpus_weight(sd->groups->cpumask)-1) / 10;
+ sd->groups->cpu_power = power;
- cpus_clear(cpu->cpumask);
- cpu_set(i, cpu->cpumask);
- cpu->cpu_power = SCHED_LOAD_SCALE;
- if (!first_cpu)
- first_cpu = cpu;
- if (last_cpu)
- last_cpu->next = cpu;
- last_cpu = cpu;
+#ifdef CONFIG_NUMA
+ if (i == first_cpu(sd->groups->cpumask)) {
+ /* Only add "power" once for each physical package. */
+ sd = &per_cpu(node_domains, i);
+ sd->groups->cpu_power += power;
+ }
+#endif
}
- last_cpu->next = first_cpu;
- mb(); /* domains were modified outside the lock */
- for_each_cpu(i) {
- struct sched_domain *cpu_sd = &per_cpu(cpu_domains, i);
- cpu_attach_domain(cpu_sd, i);
+ /* Attach the domains */
+ for_each_online_cpu(i) {
+ struct sched_domain *sd;
+#ifdef CONFIG_SCHED_SMT
+ sd = &per_cpu(cpu_domains, i);
+#else
+ sd = &per_cpu(phys_domains, i);
+#endif
+ cpu_attach_domain(sd, i);
}
+ last->next = first;
}
-#endif /* CONFIG_NUMA */
-#endif /* ARCH_HAS_SCHED_DOMAIN */
+#ifdef CONFIG_HOTPLUG_CPU
+static void __devinit arch_destroy_sched_domains(void)
+{
+ /* Do nothing: everything is statically allocated. */
+}
+#endif
-#define SCHED_DOMAIN_DEBUG
+#undef SCHED_DOMAIN_DEBUG
#ifdef SCHED_DOMAIN_DEBUG
void sched_domain_debug(void)
{
int i;
- for_each_cpu(i) {
+ for_each_online_cpu(i) {
runqueue_t *rq = cpu_rq(i);
struct sched_domain *sd;
int level = 0;
sd = rq->sd;
- printk(KERN_DEBUG "CPU%d: %s\n",
- i, (cpu_online(i) ? " online" : "offline"));
+ printk(KERN_DEBUG "CPU%d:\n", i);
do {
int j;
char str[NR_CPUS];
struct sched_group *group = sd->groups;
- cpumask_t groupmask, tmp;
+ cpumask_t groupmask;
cpumask_scnprintf(str, NR_CPUS, sd->span);
cpus_clear(groupmask);
if (!cpus_weight(group->cpumask))
printk(" ERROR empty group:");
- cpus_and(tmp, groupmask, group->cpumask);
- if (cpus_weight(tmp) > 0)
+ if (cpus_intersects(groupmask, group->cpumask))
printk(" ERROR repeated CPUs:");
cpus_or(groupmask, groupmask, group->cpumask);
sd = sd->parent;
if (sd) {
- cpus_and(tmp, groupmask, sd->span);
- if (!cpus_equal(tmp, groupmask))
+ if (!cpus_subset(groupmask, sd->span))
printk(KERN_DEBUG "ERROR parent span is not a superset of domain->span\n");
}
#define sched_domain_debug() {}
#endif
+#ifdef CONFIG_SMP
+/* Initial dummy domain for early boot and for hotplug cpu */
+static __devinitdata struct sched_domain sched_domain_dummy;
+static __devinitdata struct sched_group sched_group_dummy;
+#endif
+
+#ifdef CONFIG_HOTPLUG_CPU
+/*
+ * Force a reinitialization of the sched domains hierarchy. The domains
+ * and groups cannot be updated in place without racing with the balancing
+ * code, so we temporarily attach all running cpus to a "dummy" domain
+ * which will prevent rebalancing while the sched domains are recalculated.
+ */
+static int update_sched_domains(struct notifier_block *nfb,
+ unsigned long action, void *hcpu)
+{
+ int i;
+
+ switch (action) {
+ case CPU_UP_PREPARE:
+ case CPU_DOWN_PREPARE:
+ for_each_online_cpu(i)
+ cpu_attach_domain(&sched_domain_dummy, i);
+ arch_destroy_sched_domains();
+ return NOTIFY_OK;
+
+ case CPU_UP_CANCELED:
+ case CPU_DOWN_FAILED:
+ case CPU_ONLINE:
+ case CPU_DEAD:
+ /*
+ * Fall through and re-initialise the domains.
+ */
+ break;
+ default:
+ return NOTIFY_DONE;
+ }
+
+ /* The hotplug lock is already held by cpu_up/cpu_down */
+ arch_init_sched_domains();
+
+ sched_domain_debug();
+
+ return NOTIFY_OK;
+}
+#endif
+
void __init sched_init_smp(void)
{
+ lock_cpu_hotplug();
arch_init_sched_domains();
sched_domain_debug();
+ unlock_cpu_hotplug();
+ /* XXX: Theoretical race here - CPU may be hotplugged now */
+ hotcpu_notifier(update_sched_domains, 0);
}
#else
void __init sched_init_smp(void)
{
/* Linker adds these: start and end of __sched functions */
extern char __sched_text_start[], __sched_text_end[];
- return addr >= (unsigned long)__sched_text_start
- && addr < (unsigned long)__sched_text_end;
+ return in_lock_functions(addr) ||
+ (addr >= (unsigned long)__sched_text_start
+ && addr < (unsigned long)__sched_text_end);
}
void __init sched_init(void)
{
runqueue_t *rq;
- int i, j, k;
+ int i;
#ifdef CONFIG_SMP
/* Set up an initial dummy domain for early boot */
- static struct sched_domain sched_domain_init;
- static struct sched_group sched_group_init;
- cpumask_t cpu_mask_all = CPU_MASK_ALL;
- memset(&sched_domain_init, 0, sizeof(struct sched_domain));
- sched_domain_init.span = cpu_mask_all;
- sched_domain_init.groups = &sched_group_init;
- sched_domain_init.last_balance = jiffies;
- sched_domain_init.balance_interval = INT_MAX; /* Don't balance */
-
- memset(&sched_group_init, 0, sizeof(struct sched_group));
- sched_group_init.cpumask = cpu_mask_all;
- sched_group_init.next = &sched_group_init;
- sched_group_init.cpu_power = SCHED_LOAD_SCALE;
+ memset(&sched_domain_dummy, 0, sizeof(struct sched_domain));
+ sched_domain_dummy.span = CPU_MASK_ALL;
+ sched_domain_dummy.groups = &sched_group_dummy;
+ sched_domain_dummy.last_balance = jiffies;
+ sched_domain_dummy.balance_interval = INT_MAX; /* Don't balance */
+ sched_domain_dummy.busy_factor = 1;
+
+ memset(&sched_group_dummy, 0, sizeof(struct sched_group));
+ sched_group_dummy.cpumask = CPU_MASK_ALL;
+ sched_group_dummy.next = &sched_group_dummy;
+ sched_group_dummy.cpu_power = SCHED_LOAD_SCALE;
#endif
+ init_cpu_classes();
+
for (i = 0; i < NR_CPUS; i++) {
+#ifndef CONFIG_CKRM_CPU_SCHEDULE
+ int j, k;
prio_array_t *array;
rq = cpu_rq(i);
spin_lock_init(&rq->lock);
+
+ for (j = 0; j < 2; j++) {
+ array = rq->arrays + j;
+ for (k = 0; k < MAX_PRIO; k++) {
+ INIT_LIST_HEAD(array->queue + k);
+ __clear_bit(k, array->bitmap);
+ }
+ // delimiter for bitsearch
+ __set_bit(MAX_PRIO, array->bitmap);
+ }
+
rq->active = rq->arrays;
rq->expired = rq->arrays + 1;
rq->best_expired_prio = MAX_PRIO;
+#else
+ rq = cpu_rq(i);
+ spin_lock_init(&rq->lock);
+#endif
+
#ifdef CONFIG_SMP
- rq->sd = &sched_domain_init;
+ rq->sd = &sched_domain_dummy;
rq->cpu_load = 0;
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+ ckrm_load_init(rq_ckrm_load(rq));
+#endif
rq->active_balance = 0;
rq->push_cpu = 0;
rq->migration_thread = NULL;
INIT_LIST_HEAD(&rq->migration_queue);
+#endif
+#ifdef CONFIG_VSERVER_HARDCPU
+ INIT_LIST_HEAD(&rq->hold_queue);
#endif
atomic_set(&rq->nr_iowait, 0);
- for (j = 0; j < 2; j++) {
- array = rq->arrays + j;
- for (k = 0; k < MAX_PRIO; k++) {
- INIT_LIST_HEAD(array->queue + k);
- __clear_bit(k, array->bitmap);
- }
- // delimiter for bitsearch
- __set_bit(MAX_PRIO, array->bitmap);
- }
}
- /*
- * We have to do a little magic to get the first
- * thread right in SMP mode.
- */
- rq = this_rq();
- rq->curr = current;
- rq->idle = current;
- set_task_cpu(current, smp_processor_id());
- wake_up_forked_process(current);
/*
* The boot idle thread does lazy MMU switching as well:
*/
atomic_inc(&init_mm.mm_count);
enter_lazy_tlb(&init_mm, current);
+
+ /*
+ * Make us the idle thread. Technically, schedule() should not be
+ * called from this thread, however somewhere below it might be,
+ * but because we are the idle thread, we just pick up running again
+ * when this runqueue becomes "idle".
+ */
+ init_idle(current, smp_processor_id());
}
#ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
-void __might_sleep(char *file, int line)
+void __might_sleep(char *file, int line, int atomic_depth)
{
#if defined(in_atomic)
static unsigned long prev_jiffy; /* ratelimiting */
- if ((in_atomic() || irqs_disabled()) &&
+#ifndef CONFIG_PREEMPT
+ atomic_depth = 0;
+#endif
+ if (((in_atomic() != atomic_depth) || irqs_disabled()) &&
system_state == SYSTEM_RUNNING) {
if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy)
return;
prev_jiffy = jiffies;
printk(KERN_ERR "Debug: sleeping function called from invalid"
" context at %s:%d\n", file, line);
- printk("in_atomic():%d, irqs_disabled():%d\n",
- in_atomic(), irqs_disabled());
+ printk("in_atomic():%d[expected: %d], irqs_disabled():%d\n",
+ in_atomic(), atomic_depth, irqs_disabled());
dump_stack();
}
#endif
EXPORT_SYMBOL(__might_sleep);
#endif
-
-#if defined(CONFIG_SMP) && defined(CONFIG_PREEMPT)
-/*
- * This could be a long-held lock. If another CPU holds it for a long time,
- * and that CPU is not asked to reschedule then *this* CPU will spin on the
- * lock for a long time, even if *this* CPU is asked to reschedule.
- *
- * So what we do here, in the slow (contended) path is to spin on the lock by
- * hand while permitting preemption.
- *
- * Called inside preempt_disable().
+#ifdef CONFIG_CKRM_CPU_SCHEDULE
+/**
+ * return the classqueue object of a certain processor
*/
-void __sched __preempt_spin_lock(spinlock_t *lock)
+struct classqueue_struct * get_cpu_classqueue(int cpu)
{
- if (preempt_count() > 1) {
- _raw_spin_lock(lock);
- return;
- }
- do {
- preempt_enable();
- while (spin_is_locked(lock))
- cpu_relax();
- preempt_disable();
- } while (!_raw_spin_trylock(lock));
+ return (& (cpu_rq(cpu)->classqueue) );
}
-EXPORT_SYMBOL(__preempt_spin_lock);
-
-void __sched __preempt_write_lock(rwlock_t *lock)
+/**
+ * _ckrm_cpu_change_class - change the class of a task
+ */
+void _ckrm_cpu_change_class(task_t *tsk, struct ckrm_cpu_class *newcls)
{
- if (preempt_count() > 1) {
- _raw_write_lock(lock);
- return;
- }
+ prio_array_t *array;
+ struct runqueue *rq;
+ unsigned long flags;
- do {
- preempt_enable();
- while (rwlock_is_locked(lock))
- cpu_relax();
- preempt_disable();
- } while (!_raw_write_trylock(lock));
-}
+ rq = task_rq_lock(tsk,&flags);
+ array = tsk->array;
+ if (array) {
+ dequeue_task(tsk,array);
+ tsk->cpu_class = newcls;
+ enqueue_task(tsk,rq_active(tsk,rq));
+ } else
+ tsk->cpu_class = newcls;
-EXPORT_SYMBOL(__preempt_write_lock);
-#endif /* defined(CONFIG_SMP) && defined(CONFIG_PREEMPT) */
+ task_rq_unlock(rq,&flags);
+}
+#endif