Implement the "staircase" hybrid foreground-background single priority array cpu scheduler policy. Signed-off-by: Con Kolivas --- fs/proc/array.c | 4 include/linux/sched.h | 20 kernel/exit.c | 1 kernel/sched.c | 1107 ++++++++++++++++++-------------------------------- 4 files changed, 418 insertions(+), 714 deletions(-) Index: linux-2.6.21-rc7-ck1/fs/proc/array.c =================================================================== --- linux-2.6.21-rc7-ck1.orig/fs/proc/array.c 2007-04-16 20:35:14.000000000 +1000 +++ linux-2.6.21-rc7-ck1/fs/proc/array.c 2007-04-16 20:35:18.000000000 +1000 @@ -165,7 +165,7 @@ static inline char * task_state(struct t rcu_read_lock(); buffer += sprintf(buffer, "State:\t%s\n" - "SleepAVG:\t%lu%%\n" + "Bonus:\t%d\n" "Tgid:\t%d\n" "Pid:\t%d\n" "PPid:\t%d\n" @@ -173,7 +173,7 @@ static inline char * task_state(struct t "Uid:\t%d\t%d\t%d\t%d\n" "Gid:\t%d\t%d\t%d\t%d\n", get_task_state(p), - (p->sleep_avg/1024)*100/(1020000000/1024), + p->bonus, p->tgid, p->pid, pid_alive(p) ? rcu_dereference(p->real_parent)->tgid : 0, pid_alive(p) && p->ptrace ? rcu_dereference(p->parent)->pid : 0, Index: linux-2.6.21-rc7-ck1/kernel/exit.c =================================================================== --- linux-2.6.21-rc7-ck1.orig/kernel/exit.c 2007-04-16 20:35:15.000000000 +1000 +++ linux-2.6.21-rc7-ck1/kernel/exit.c 2007-04-16 20:35:18.000000000 +1000 @@ -170,7 +170,6 @@ repeat: zap_leader = (leader->exit_signal == -1); } - sched_exit(p); write_unlock_irq(&tasklist_lock); proc_flush_task(p); release_thread(p); Index: linux-2.6.21-rc7-ck1/include/linux/sched.h =================================================================== --- linux-2.6.21-rc7-ck1.orig/include/linux/sched.h 2007-04-16 20:35:15.000000000 +1000 +++ linux-2.6.21-rc7-ck1/include/linux/sched.h 2007-04-16 20:35:18.000000000 +1000 @@ -524,6 +524,7 @@ struct signal_struct { #define MAX_RT_PRIO MAX_USER_RT_PRIO #define MAX_PRIO (MAX_RT_PRIO + 40) +#define MIN_USER_PRIO (MAX_PRIO - 1) #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO) #define rt_task(p) rt_prio((p)->prio) @@ -788,15 +789,6 @@ struct mempolicy; struct pipe_inode_info; struct uts_namespace; -enum sleep_type { - SLEEP_NORMAL, - SLEEP_NONINTERACTIVE, - SLEEP_INTERACTIVE, - SLEEP_INTERRUPTED, -}; - -struct prio_array; - struct task_struct { volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */ struct thread_info *thread_info; @@ -814,20 +806,19 @@ struct task_struct { int load_weight; /* for niceness load balancing purposes */ int prio, static_prio, normal_prio; struct list_head run_list; - struct prio_array *array; unsigned short ioprio; #ifdef CONFIG_BLK_DEV_IO_TRACE unsigned int btrace_seq; #endif - unsigned long sleep_avg; unsigned long long timestamp, last_ran; + unsigned long runtime, totalrun, systime; + unsigned int bonus; + unsigned int slice, time_slice; unsigned long long sched_time; /* sched_clock time spent running */ - enum sleep_type sleep_type; unsigned long policy; cpumask_t cpus_allowed; - unsigned int time_slice, first_time_slice; #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) struct sched_info sched_info; @@ -1158,6 +1149,8 @@ static inline void put_task_struct(struc #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */ #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */ #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */ +#define PF_NONSLEEP 0x40000000 /* Waiting on in kernel activity */ +#define PF_FORKED 0x80000000 /* Task just forked another process */ /* * Only the _current_ task can read/write to tsk->flags, but other @@ -1292,7 +1285,6 @@ extern void FASTCALL(wake_up_new_task(st static inline void kick_process(struct task_struct *tsk) { } #endif extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags)); -extern void FASTCALL(sched_exit(struct task_struct * p)); extern int in_group_p(gid_t); extern int in_egroup_p(gid_t); Index: linux-2.6.21-rc7-ck1/kernel/sched.c =================================================================== --- linux-2.6.21-rc7-ck1.orig/kernel/sched.c 2007-04-16 20:35:15.000000000 +1000 +++ linux-2.6.21-rc7-ck1/kernel/sched.c 2007-04-16 20:35:18.000000000 +1000 @@ -16,6 +16,10 @@ * by Davide Libenzi, preemptible kernel bits by Robert Love. * 2003-09-03 Interactivity tuning by Con Kolivas. * 2004-04-02 Scheduler domains code by Nick Piggin + * 2007-04-16 Staircase scheduling policy by Con Kolivas with help + * from William Lee Irwin III, Zwane Mwaikambo, Peter Williams + * and Andreas Mohr. + * Staircase v17.1 */ #include @@ -87,123 +91,19 @@ unsigned long long __attribute__((weak)) /* * Some helpers for converting nanosecond timing to jiffy resolution */ -#define NS_TO_JIFFIES(TIME) ((TIME) / (1000000000 / HZ)) -#define JIFFIES_TO_NS(TIME) ((TIME) * (1000000000 / HZ)) - -/* - * These are the 'tuning knobs' of the scheduler: - * - * 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 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 EXIT_WEIGHT 3 -#define PRIO_BONUS_RATIO 25 -#define MAX_BONUS (MAX_USER_PRIO * PRIO_BONUS_RATIO / 100) -#define INTERACTIVE_DELTA 2 -#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)) - -/* - * If a task is 'interactive' then we reinsert it in the active - * array after it has expired its current timeslice. (it will not - * continue to run immediately, it will still roundrobin with - * other interactive tasks.) - * - * This part scales the interactivity limit depending on niceness. - * - * We scale it linearly, offset by the INTERACTIVE_DELTA delta. - * Here are a few examples of different nice levels: - * - * TASK_INTERACTIVE(-20): [1,1,1,1,1,1,1,1,1,0,0] - * TASK_INTERACTIVE(-10): [1,1,1,1,1,1,1,0,0,0,0] - * TASK_INTERACTIVE( 0): [1,1,1,1,0,0,0,0,0,0,0] - * TASK_INTERACTIVE( 10): [1,1,0,0,0,0,0,0,0,0,0] - * TASK_INTERACTIVE( 19): [0,0,0,0,0,0,0,0,0,0,0] - * - * (the X axis represents the possible -5 ... 0 ... +5 dynamic - * priority range a task can explore, a value of '1' means the - * task is rated interactive.) - * - * Ie. nice +19 tasks can never get 'interactive' enough to be - * reinserted into the active array. And only heavily CPU-hog nice -20 - * tasks will be expired. Default nice 0 tasks are somewhere between, - * it takes some effort for them to get interactive, but it's not - * too hard. - */ - -#define CURRENT_BONUS(p) \ - (NS_TO_JIFFIES((p)->sleep_avg) * MAX_BONUS / \ - MAX_SLEEP_AVG) - -#define GRANULARITY (10 * HZ / 1000 ? : 1) - -#ifdef CONFIG_SMP -#define TIMESLICE_GRANULARITY(p) (GRANULARITY * \ - (1 << (((MAX_BONUS - CURRENT_BONUS(p)) ? : 1) - 1)) * \ - num_online_cpus()) -#else -#define TIMESLICE_GRANULARITY(p) (GRANULARITY * \ - (1 << (((MAX_BONUS - CURRENT_BONUS(p)) ? : 1) - 1))) -#endif - -#define SCALE(v1,v1_max,v2_max) \ - (v1) * (v2_max) / (v1_max) - -#define DELTA(p) \ - (SCALE(TASK_NICE(p) + 20, 40, MAX_BONUS) - 20 * MAX_BONUS / 40 + \ - INTERACTIVE_DELTA) - -#define TASK_INTERACTIVE(p) \ - ((p)->prio <= (p)->static_prio - DELTA(p)) - -#define INTERACTIVE_SLEEP(p) \ - (JIFFIES_TO_NS(MAX_SLEEP_AVG * \ - (MAX_BONUS / 2 + DELTA((p)) + 1) / MAX_BONUS - 1)) +#define NSJIFFY (1000000000 / HZ) /* One jiffy in ns */ +#define NS_TO_JIFFIES(TIME) ((TIME) / NSJIFFY) +#define JIFFIES_TO_NS(TIME) ((TIME) * NSJIFFY) #define TASK_PREEMPTS_CURR(p, rq) \ ((p)->prio < (rq)->curr->prio) -#define SCALE_PRIO(x, prio) \ - max(x * (MAX_PRIO - prio) / (MAX_USER_PRIO / 2), MIN_TIMESLICE) - -static unsigned int static_prio_timeslice(int static_prio) -{ - if (static_prio < NICE_TO_PRIO(0)) - return SCALE_PRIO(DEF_TIMESLICE * 4, static_prio); - else - return SCALE_PRIO(DEF_TIMESLICE, static_prio); -} - /* - * 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. + * This is the time all tasks within the same priority round robin. + * Set to a minimum of 6ms. */ - -static inline unsigned int task_timeslice(struct task_struct *p) -{ - return static_prio_timeslice(p->static_prio); -} - -/* - * These are the runqueue data structures: - */ - -struct prio_array { - unsigned int nr_active; - DECLARE_BITMAP(bitmap, MAX_PRIO+1); /* include 1 bit for delimiter */ - struct list_head queue[MAX_PRIO]; -}; +#define RR_INTERVAL ((6 * HZ / 1001) + 1) +#define DEF_TIMESLICE (RR_INTERVAL * 19) /* * This is the main, per-CPU runqueue data structure. @@ -234,14 +134,13 @@ struct rq { */ unsigned long nr_uninterruptible; - unsigned long expired_timestamp; /* Cached timestamp set by update_cpu_clock() */ unsigned long long most_recent_timestamp; struct task_struct *curr, *idle; unsigned long next_balance; struct mm_struct *prev_mm; - struct prio_array *active, *expired, arrays[2]; - int best_expired_prio; + unsigned long bitmap[BITS_TO_LONGS(MAX_PRIO + 1)]; + struct list_head queue[MAX_PRIO]; atomic_t nr_iowait; #ifdef CONFIG_SMP @@ -578,13 +477,7 @@ static inline struct rq *this_rq_lock(vo #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) /* - * 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()). + * Called when a process is dequeued and given the cpu. * * This function is only called from sched_info_arrive(), rather than * dequeue_task(). Even though a task may be queued and dequeued multiple @@ -617,13 +510,11 @@ static void sched_info_arrive(struct tas } /* - * 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(), + * Called when a process is queued. + * The time is noted and later used to determine how long we had to wait for + * us to reach the cpu. + * It is unusual but not impossible for tasks to be dequeued and immediately + * requeued: this can happen in sched_yield(), * set_user_nice(), and even load_balance() as it moves tasks from runqueue * to runqueue. * @@ -682,73 +573,75 @@ sched_info_switch(struct task_struct *pr #define sched_info_switch(t, next) do { } while (0) #endif /* CONFIG_SCHEDSTATS || CONFIG_TASK_DELAY_ACCT */ -/* - * Adding/removing a task to/from a priority array: - */ -static void dequeue_task(struct task_struct *p, struct prio_array *array) +#if BITS_PER_LONG < 64 +static inline void longlimit(unsigned long long *longlong) +{ + if (*longlong > (1 << 31)) + *longlong = 1 << 31; +} +#else +static inline void longlimit(unsigned long long *__unused) +{ +} +#endif + +/* Get nanosecond clock difference without overflowing unsigned long. */ +static unsigned long ns_diff(unsigned long long v1, unsigned long long v2) { - array->nr_active--; - list_del(&p->run_list); - if (list_empty(array->queue + p->prio)) - __clear_bit(p->prio, array->bitmap); + unsigned long long vdiff = 0; + + /* Rarely the clock appears to go backwards. Ignore that */ + if (likely(v1 > v2)) { + vdiff = v1 - v2; + longlimit(&vdiff); + } + return (unsigned long)vdiff; } -static void enqueue_task(struct task_struct *p, struct prio_array *array) +static int task_queued(struct task_struct *task) { - 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; + return !list_empty(&task->run_list); } /* - * Put task to the end of the run list without the overhead of dequeue - * followed by enqueue. + * Adding/removing a task to/from a runqueue: */ -static void requeue_task(struct task_struct *p, struct prio_array *array) +static void dequeue_task(struct task_struct *p, struct rq *rq) { - list_move_tail(&p->run_list, array->queue + p->prio); + list_del_init(&p->run_list); + if (list_empty(rq->queue + p->prio)) + __clear_bit(p->prio, rq->bitmap); } -static inline void -enqueue_task_head(struct task_struct *p, struct prio_array *array) +static void enqueue_task(struct task_struct *p, struct rq *rq) { - list_add(&p->run_list, array->queue + p->prio); - __set_bit(p->prio, array->bitmap); - array->nr_active++; - p->array = array; + list_add_tail(&p->run_list, rq->queue + p->prio); + __set_bit(p->prio, rq->bitmap); } /* - * __normal_prio - return the priority that is based on the static - * priority but is modified by bonuses/penalties. - * - * We scale the actual sleep average [0 .... MAX_SLEEP_AVG] - * into the -5 ... 0 ... +5 bonus/penalty range. - * - * We use 25% of the full 0...39 priority range so that: - * - * 1) nice +19 interactive tasks do not preempt nice 0 CPU hogs. - * 2) nice -20 CPU hogs do not get preempted by nice 0 tasks. - * - * Both properties are important to certain workloads. + * Put task to the end of the run list without the overhead of dequeue + * followed by enqueue. */ - -static inline int __normal_prio(struct task_struct *p) +static void requeue_task(struct task_struct *p, struct rq *rq, int prio) { - int bonus, prio; - - bonus = CURRENT_BONUS(p) - MAX_BONUS / 2; + list_move_tail(&p->run_list, rq->queue + prio); + if (p->prio != prio) { + if (list_empty(rq->queue + p->prio)) + __clear_bit(p->prio, rq->bitmap); + p->prio = prio; + __set_bit(prio, rq->bitmap); + } +} - prio = p->static_prio - bonus; - if (prio < MAX_RT_PRIO) - prio = MAX_RT_PRIO; - if (prio > MAX_PRIO-1) - prio = MAX_PRIO-1; - return prio; +static void enqueue_task_head(struct task_struct *p, struct rq *rq) +{ + list_add(&p->run_list, rq->queue + p->prio); + __set_bit(p->prio, rq->bitmap); } +static unsigned int slice(const struct task_struct *p); + /* * To aid in avoiding the subversion of "niceness" due to uneven distribution * of tasks with abnormal "nice" values across CPUs the contribution that @@ -766,10 +659,9 @@ static inline int __normal_prio(struct t #define TIME_SLICE_NICE_ZERO DEF_TIMESLICE #define LOAD_WEIGHT(lp) \ (((lp) * SCHED_LOAD_SCALE) / TIME_SLICE_NICE_ZERO) -#define PRIO_TO_LOAD_WEIGHT(prio) \ - LOAD_WEIGHT(static_prio_timeslice(prio)) -#define RTPRIO_TO_LOAD_WEIGHT(rp) \ - (PRIO_TO_LOAD_WEIGHT(MAX_RT_PRIO) + LOAD_WEIGHT(rp)) +#define TASK_LOAD_WEIGHT(p) LOAD_WEIGHT(slice(p)) +#define RTPRIO_TO_LOAD_WEIGHT(rp) \ + (LOAD_WEIGHT((RR_INTERVAL + 20 + (rp)))) static void set_load_weight(struct task_struct *p) { @@ -786,7 +678,7 @@ static void set_load_weight(struct task_ #endif p->load_weight = RTPRIO_TO_LOAD_WEIGHT(p->rt_priority); } else - p->load_weight = PRIO_TO_LOAD_WEIGHT(p->static_prio); + p->load_weight = TASK_LOAD_WEIGHT(p); } static inline void @@ -814,13 +706,189 @@ static inline void dec_nr_running(struct } /* + * __activate_task - move a task to the runqueue. + */ +static void __activate_task(struct task_struct *p, struct rq *rq) +{ + enqueue_task(p, rq); + inc_nr_running(p, rq); +} + +/* + * __activate_idle_task - move idle task to the _front_ of runqueue. + */ +static void __activate_idle_task(struct task_struct *p, struct rq *rq) +{ + enqueue_task_head(p, rq); + inc_nr_running(p, rq); +} + +/* + * Bonus - How much higher than its base priority an interactive task can run. + */ +static unsigned int bonus(const struct task_struct *p) +{ + return TASK_USER_PRIO(p); +} + +static unsigned int rr_interval(const struct task_struct *p) +{ + int nice = TASK_NICE(p); + + if (nice < 0 && !rt_task(p)) + return RR_INTERVAL * (20 - nice) / 20; + return RR_INTERVAL; +} + +/* + * slice - the duration a task runs before getting requeued at its best + * priority and has its bonus decremented. + */ +static unsigned int slice(const struct task_struct *p) +{ + unsigned int slice, rr; + + slice = rr = rr_interval(p); + if (likely(!rt_task(p))) + slice += (39 - TASK_USER_PRIO(p)) * rr; + return slice; +} + +/* + * We increase our bonus by sleeping more than the time we ran. + * The ratio of sleep to run gives us the cpu% that we last ran and determines + * the maximum bonus we can acquire. + */ +static void inc_bonus(struct task_struct *p, unsigned long totalrun, unsigned long sleep) +{ + unsigned int best_bonus = sleep / (totalrun + 1); + + if (p->bonus >= best_bonus) + return; + best_bonus = bonus(p); + if (p->bonus < best_bonus) + p->bonus++; +} + +static void dec_bonus(struct task_struct *p) +{ + if (p->bonus) + p->bonus--; +} + +static void slice_overrun(struct task_struct *p) +{ + unsigned long ns_slice = JIFFIES_TO_NS(p->slice); + + do { + p->totalrun -= ns_slice; + dec_bonus(p); + } while (unlikely(p->totalrun > ns_slice)); +} + +static void continue_slice(struct task_struct *p) +{ + unsigned long total_run = NS_TO_JIFFIES(p->totalrun); + + if (unlikely(total_run >= p->slice)) + slice_overrun(p); + else { + unsigned long remainder; + + p->slice -= total_run; + remainder = p->slice % rr_interval(p); + if (remainder) + p->time_slice = remainder; + } +} + +/* + * recalc_task_prio - this checks for tasks that have run less than a full + * slice and have woken up again soon after, or have just forked a + * thread/process and make them continue their old slice instead of starting + * a new one at high priority. + */ +static void recalc_task_prio(struct task_struct *p, unsigned long long now) +{ + unsigned long sleep_time; + + /* + * If this task has managed to run to its lowest priority then + * decrease its bonus and requeue it now at best priority instead + * of possibly flagging around lowest priority. Save up any systime + * that may affect priority on the next reschedule. + */ + if (p->slice > p->time_slice && + p->slice - NS_TO_JIFFIES(p->totalrun) < p->time_slice) { + dec_bonus(p); + p->totalrun = 0; + return; + } + + /* + * Add the total for this last scheduled run (p->runtime) and system + * time (p->systime) done on behalf of p to the running total so far + * used (p->totalrun). + */ + p->totalrun += p->runtime + p->systime; + sleep_time = ns_diff(now, p->timestamp); + + if (p->systime > sleep_time || p->flags & PF_FORKED) + sleep_time = 0; + else { + sleep_time -= p->systime; + /* + * We elevate priority by the amount of time we slept. If we + * sleep longer than our running total and have not set the + * PF_NONSLEEP flag we gain a bonus. + */ + if (sleep_time >= p->totalrun) { + if (!(p->flags & PF_NONSLEEP)) + inc_bonus(p, p->totalrun, sleep_time); + p->totalrun = 0; + return; + } + p->totalrun -= sleep_time; + } + continue_slice(p); +} + +/* + * __normal_prio - dynamic priority dependent on bonus. + * The priority normally decreases by one each RR_INTERVAL. + * As the bonus increases the initial priority starts at a higher "stair" or + * priority for longer. + */ +static int __normal_prio(struct task_struct *p) +{ + int prio; + unsigned int full_slice, used_slice = 0; + unsigned int best_bonus, rr; + + full_slice = slice(p); + if (full_slice > p->slice) + used_slice = full_slice - p->slice; + + best_bonus = bonus(p); + prio = MAX_RT_PRIO + best_bonus; + if (!batch_task(p)) + prio -= p->bonus; + + rr = rr_interval(p); + prio += used_slice / rr; + if (prio > MIN_USER_PRIO) + prio = MIN_USER_PRIO; + return prio; +} + +/* * Calculate the expected normal priority: i.e. priority * without taking RT-inheritance into account. Might be * boosted by interactivity modifiers. Changes upon fork, * setprio syscalls, and whenever the interactivity * estimator recalculates. */ -static inline int normal_prio(struct task_struct *p) +static int normal_prio(struct task_struct *p) { int prio; @@ -852,111 +920,14 @@ static int effective_prio(struct task_st } /* - * __activate_task - move a task to the runqueue. - */ -static void __activate_task(struct task_struct *p, struct rq *rq) -{ - struct prio_array *target = rq->active; - - if (batch_task(p)) - target = rq->expired; - enqueue_task(p, target); - inc_nr_running(p, rq); -} - -/* - * __activate_idle_task - move idle task to the _front_ of runqueue. - */ -static inline void __activate_idle_task(struct task_struct *p, struct rq *rq) -{ - enqueue_task_head(p, rq->active); - inc_nr_running(p, rq); -} - -/* - * Recalculate p->normal_prio and p->prio after having slept, - * updating the sleep-average too: - */ -static int recalc_task_prio(struct task_struct *p, unsigned long long now) -{ - /* Caller must always ensure 'now >= p->timestamp' */ - unsigned long sleep_time = now - p->timestamp; - - if (batch_task(p)) - sleep_time = 0; - - if (likely(sleep_time > 0)) { - /* - * This ceiling is set to the lowest priority that would allow - * a task to be reinserted into the active array on timeslice - * completion. - */ - unsigned long ceiling = INTERACTIVE_SLEEP(p); - - if (p->mm && sleep_time > ceiling && p->sleep_avg < ceiling) { - /* - * Prevents user tasks from achieving best priority - * with one single large enough sleep. - */ - p->sleep_avg = ceiling; - /* - * Using INTERACTIVE_SLEEP() as a ceiling places a - * nice(0) task 1ms sleep away from promotion, and - * gives it 700ms to round-robin with no chance of - * being demoted. This is more than generous, so - * mark this sleep as non-interactive to prevent the - * on-runqueue bonus logic from intervening should - * this task not receive cpu immediately. - */ - p->sleep_type = SLEEP_NONINTERACTIVE; - } else { - /* - * Tasks waking from uninterruptible sleep are - * limited in their sleep_avg rise as they - * are likely to be waiting on I/O - */ - if (p->sleep_type == SLEEP_NONINTERACTIVE && p->mm) { - if (p->sleep_avg >= ceiling) - sleep_time = 0; - else if (p->sleep_avg + sleep_time >= - ceiling) { - p->sleep_avg = ceiling; - sleep_time = 0; - } - } - - /* - * This code gives a bonus to interactive tasks. - * - * The boost works by updating the 'average sleep time' - * value here, based on ->timestamp. The more time a - * task spends sleeping, the higher the average gets - - * and the higher the priority boost gets as well. - */ - p->sleep_avg += sleep_time; - - } - if (p->sleep_avg > NS_MAX_SLEEP_AVG) - p->sleep_avg = NS_MAX_SLEEP_AVG; - } - - return effective_prio(p); -} - -/* * activate_task - move a task to the runqueue and do priority recalculation * - * Update all the scheduling statistics stuff. (sleep average - * calculation, priority modifiers, etc.) */ static void activate_task(struct task_struct *p, struct rq *rq, int local) { - unsigned long long now; - - if (rt_task(p)) - goto out; + unsigned long long now = sched_clock(); + unsigned long rr = rr_interval(p); - now = sched_clock(); #ifdef CONFIG_SMP if (!local) { /* Compensate for drifting sched_clock */ @@ -977,32 +948,15 @@ static void activate_task(struct task_st (now - p->timestamp) >> 20); } - p->prio = recalc_task_prio(p, now); - - /* - * This checks to make sure it's not an uninterruptible task - * that is now waking up. - */ - if (p->sleep_type == SLEEP_NORMAL) { - /* - * Tasks which were woken up by interrupts (ie. hw events) - * are most likely of interactive nature. So we give them - * the credit of extending their sleep time to the period - * of time they spend on the runqueue, waiting for execution - * on a CPU, first time around: - */ - if (in_interrupt()) - p->sleep_type = SLEEP_INTERRUPTED; - else { - /* - * Normal first-time wakeups get a credit too for - * on-runqueue time, but it will be weighted down: - */ - p->sleep_type = SLEEP_INTERACTIVE; - } + p->slice = slice(p); + p->time_slice = p->slice % rr ? : rr; + if (!rt_task(p)) { + recalc_task_prio(p, now); + p->prio = effective_prio(p); + p->systime = 0; + p->flags &= ~(PF_FORKED | PF_NONSLEEP); } p->timestamp = now; -out: __activate_task(p, rq); } @@ -1012,8 +966,7 @@ out: static void deactivate_task(struct task_struct *p, struct rq *rq) { dec_nr_running(p, rq); - dequeue_task(p, p->array); - p->array = NULL; + dequeue_task(p, rq); } /* @@ -1095,7 +1048,7 @@ migrate_task(struct task_struct *p, int * If the task is not on a runqueue (and not running), then * it is sufficient to simply update the task's cpu field. */ - if (!p->array && !task_running(rq, p)) { + if (!task_queued(p) && !task_running(rq, p)) { set_task_cpu(p, dest_cpu); return 0; } @@ -1126,7 +1079,7 @@ void wait_task_inactive(struct task_stru repeat: rq = task_rq_lock(p, &flags); /* Must be off runqueue entirely, not preempted. */ - if (unlikely(p->array || task_running(rq, p))) { + if (unlikely(task_queued(p) || task_running(rq, p))) { /* If it's preempted, we yield. It could be a while. */ preempted = !task_running(rq, p); task_rq_unlock(rq, &flags); @@ -1391,6 +1344,16 @@ static inline int wake_idle(int cpu, str } #endif +/* + * Check to see if p preempts rq->curr and resched if it does. + */ +static void try_preempt(struct task_struct *p, struct rq *rq) +{ + if (TASK_PREEMPTS_CURR(p, rq)) + resched_task(rq->curr); +} + + /*** * try_to_wake_up - wake up a thread * @p: the to-be-woken-up thread @@ -1422,7 +1385,7 @@ static int try_to_wake_up(struct task_st if (!(old_state & state)) goto out; - if (p->array) + if (task_queued(p)) goto out_running; cpu = task_cpu(p); @@ -1515,7 +1478,7 @@ out_set_cpu: old_state = p->state; if (!(old_state & state)) goto out; - if (p->array) + if (task_queued(p)) goto out_running; this_cpu = smp_processor_id(); @@ -1524,26 +1487,10 @@ out_set_cpu: out_activate: #endif /* CONFIG_SMP */ - if (old_state == TASK_UNINTERRUPTIBLE) { + if (old_state == TASK_UNINTERRUPTIBLE) rq->nr_uninterruptible--; - /* - * Tasks on involuntary sleep don't earn - * sleep_avg beyond just interactive state. - */ - p->sleep_type = SLEEP_NONINTERACTIVE; - } else /* - * Tasks that have marked their sleep as noninteractive get - * woken up with their sleep average not weighted in an - * interactive way. - */ - if (old_state & TASK_NONINTERACTIVE) - p->sleep_type = SLEEP_NONINTERACTIVE; - - - activate_task(p, rq, cpu == this_cpu); - /* * Sync wakeups (i.e. those types of wakeups where the waker * has indicated that it will leave the CPU in short order) * don't trigger a preemption, if the woken up task will run on @@ -1551,10 +1498,9 @@ out_activate: * the waker guarantees that the freshly woken up task is going * to be considered on this CPU.) */ - if (!sync || cpu != this_cpu) { - if (TASK_PREEMPTS_CURR(p, rq)) - resched_task(rq->curr); - } + activate_task(p, rq, cpu == this_cpu); + if (!sync || cpu != this_cpu) + try_preempt(p, rq); success = 1; out_running: @@ -1577,7 +1523,6 @@ int fastcall wake_up_state(struct task_s return try_to_wake_up(p, state, 0); } -static void task_running_tick(struct rq *rq, struct task_struct *p); /* * Perform scheduler related setup for a newly forked process p. * p is forked by current. @@ -1605,7 +1550,6 @@ void fastcall sched_fork(struct task_str p->prio = current->normal_prio; INIT_LIST_HEAD(&p->run_list); - p->array = NULL; #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) if (unlikely(sched_info_on())) memset(&p->sched_info, 0, sizeof(p->sched_info)); @@ -1617,30 +1561,6 @@ void fastcall sched_fork(struct task_str /* Want to start with kernel preemption disabled. */ task_thread_info(p)->preempt_count = 1; #endif - /* - * Share the timeslice between parent and child, thus the - * total amount of pending timeslices in the system doesn't change, - * resulting in more scheduling fairness. - */ - local_irq_disable(); - p->time_slice = (current->time_slice + 1) >> 1; - /* - * The remainder of the first timeslice might be recovered by - * the parent if the child exits early enough. - */ - p->first_time_slice = 1; - current->time_slice >>= 1; - p->timestamp = sched_clock(); - 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 - * runqueue lock is not a problem. - */ - current->time_slice = 1; - task_running_tick(cpu_rq(cpu), current); - } - local_irq_enable(); put_cpu(); } @@ -1662,38 +1582,20 @@ void fastcall wake_up_new_task(struct ta this_cpu = smp_processor_id(); cpu = task_cpu(p); - /* - * We decrease the sleep average of forking parents - * and children as well, to keep max-interactive tasks - * from forking tasks that are max-interactive. The parent - * (current) is done further down, under its lock. - */ - p->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(p) * - CHILD_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS); - - p->prio = effective_prio(p); + /* Forked process gets no bonus to prevent fork bombs. */ + p->bonus = 0; + current->flags |= PF_FORKED; if (likely(cpu == this_cpu)) { + activate_task(p, rq, 1); 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; - p->normal_prio = current->normal_prio; - list_add_tail(&p->run_list, ¤t->run_list); - p->array = current->array; - p->array->nr_active++; - inc_nr_running(p, rq); - } set_need_resched(); - } else - /* Run child last */ - __activate_task(p, rq); + } /* * We skip the following code due to cpu == this_cpu * @@ -1710,53 +1612,19 @@ void fastcall wake_up_new_task(struct ta */ p->timestamp = (p->timestamp - this_rq->most_recent_timestamp) + rq->most_recent_timestamp; - __activate_task(p, rq); - if (TASK_PREEMPTS_CURR(p, rq)) - resched_task(rq->curr); + activate_task(p, rq, 0); + try_preempt(p, rq); /* * Parent and child are on different CPUs, now get the - * parent runqueue to update the parent's ->sleep_avg: + * parent runqueue to update the parent's ->flags: */ task_rq_unlock(rq, &flags); this_rq = task_rq_lock(current, &flags); } - current->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(current) * - PARENT_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS); task_rq_unlock(this_rq, &flags); } -/* - * Potentially available exiting-child timeslices are - * retrieved here - this way the parent does not get - * penalized for creating too many threads. - * - * (this cannot be used to 'generate' timeslices - * artificially, because any timeslice recovered here - * was given away by the parent in the first place.) - */ -void fastcall sched_exit(struct task_struct *p) -{ - unsigned long flags; - struct rq *rq; - - /* - * 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 && task_cpu(p) == task_cpu(p->parent)) { - 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 / - (EXIT_WEIGHT + 1); - task_rq_unlock(rq, &flags); -} - /** * prepare_task_switch - prepare to switch tasks * @rq: the runqueue preparing to switch @@ -2085,23 +1953,21 @@ void sched_exec(void) * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. */ -static void pull_task(struct rq *src_rq, struct prio_array *src_array, - struct task_struct *p, struct rq *this_rq, - struct prio_array *this_array, int this_cpu) +static void pull_task(struct rq *src_rq, struct task_struct *p, + struct rq *this_rq, int this_cpu) { - dequeue_task(p, src_array); + dequeue_task(p, src_rq); dec_nr_running(p, src_rq); set_task_cpu(p, this_cpu); inc_nr_running(p, this_rq); - enqueue_task(p, this_array); + enqueue_task(p, this_rq); p->timestamp = (p->timestamp - src_rq->most_recent_timestamp) + this_rq->most_recent_timestamp; /* * Note that idle threads have a prio of MAX_PRIO, for this test * to be always true for them. */ - if (TASK_PREEMPTS_CURR(p, this_rq)) - resched_task(this_rq->curr); + try_preempt(p, this_rq); } /* @@ -2144,8 +2010,6 @@ int can_migrate_task(struct task_struct return 1; } -#define rq_best_prio(rq) min((rq)->curr->prio, (rq)->best_expired_prio) - /* * move_tasks tries to move up to max_nr_move tasks and max_load_move weighted * load from busiest to this_rq, as part of a balancing operation within @@ -2160,7 +2024,6 @@ static int move_tasks(struct rq *this_rq { int idx, pulled = 0, pinned = 0, this_best_prio, best_prio, best_prio_seen, skip_for_load; - struct prio_array *array, *dst_array; struct list_head *head, *curr; struct task_struct *tmp; long rem_load_move; @@ -2170,8 +2033,8 @@ static int move_tasks(struct rq *this_rq rem_load_move = max_load_move; pinned = 1; - this_best_prio = rq_best_prio(this_rq); - best_prio = rq_best_prio(busiest); + this_best_prio = this_rq->curr->prio; + best_prio = busiest->curr->prio; /* * Enable handling of the case where there is more than one task * with the best priority. If the current running task is one @@ -2181,38 +2044,17 @@ static int move_tasks(struct rq *this_rq */ best_prio_seen = best_prio == busiest->curr->prio; - /* - * 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 (busiest->expired->nr_active) { - array = busiest->expired; - dst_array = this_rq->expired; - } else { - array = busiest->active; - dst_array = this_rq->active; - } - -new_array: /* Start searching at priority 0: */ idx = 0; skip_bitmap: if (!idx) - idx = sched_find_first_bit(array->bitmap); + idx = sched_find_first_bit(busiest->bitmap); else - idx = find_next_bit(array->bitmap, MAX_PRIO, idx); - if (idx >= MAX_PRIO) { - if (array == busiest->expired && busiest->active->nr_active) { - array = busiest->active; - dst_array = this_rq->active; - goto new_array; - } + idx = find_next_bit(busiest->bitmap, MAX_PRIO, idx); + if (idx >= MAX_PRIO) goto out; - } - head = array->queue + idx; + head = busiest->queue + idx; curr = head->prev; skip_queue: tmp = list_entry(curr, struct task_struct, run_list); @@ -2237,7 +2079,7 @@ skip_queue: goto skip_bitmap; } - pull_task(busiest, array, tmp, this_rq, dst_array, this_cpu); + pull_task(busiest, tmp, this_rq, this_cpu); pulled++; rem_load_move -= tmp->load_weight; @@ -3013,10 +2855,35 @@ EXPORT_PER_CPU_SYMBOL(kstat); /* * This is called on clock ticks and on context switches. * Bank in p->sched_time the ns elapsed since the last tick or switch. + * The value returned from sched_clock() occasionally gives bogus values so + * some sanity checking is required. */ -static inline void -update_cpu_clock(struct task_struct *p, struct rq *rq, unsigned long long now) +static void +update_cpu_clock(struct task_struct *p, struct rq *rq, unsigned long long now, + int tick) { + static const unsigned int min_diff = 1000; + cputime64_t time_diff = now - p->last_ran; + + if (tick) { + /* + * Called from scheduler_tick() there should be less than two + * jiffies worth, and not negative/overflow. + */ + if (time_diff > JIFFIES_TO_NS(2) || time_diff < min_diff) + time_diff = NSJIFFY; + } else { + /* + * Called from context_switch there should be less than one + * jiffy worth, and not negative/overflowed. In the case when + * sched_clock fails to return high resolution values this + * also ensures at least 1 min_diff gets banked. + */ + if (time_diff > NSJIFFY || time_diff < min_diff) + time_diff = min_diff; + } + if (p != rq->idle && p->policy != SCHED_FIFO) + p->runtime += time_diff; p->sched_time += now - p->last_ran; p->last_ran = rq->most_recent_timestamp = now; } @@ -3038,27 +2905,6 @@ unsigned long long current_sched_time(co } /* - * We place interactive tasks back into the active array, if possible. - * - * To guarantee that this does not starve expired tasks we ignore the - * interactivity of a task if the first expired task had to wait more - * than a 'reasonable' amount of time. This deadline timeout is - * load-dependent, as the frequency of array switched decreases with - * increasing number of running tasks. We also ignore the interactivity - * if a better static_prio task has expired: - */ -static inline int expired_starving(struct rq *rq) -{ - if (rq->curr->static_prio > rq->best_expired_prio) - return 1; - if (!STARVATION_LIMIT || !rq->expired_timestamp) - return 0; - if (jiffies - rq->expired_timestamp > STARVATION_LIMIT * rq->nr_running) - return 1; - return 0; -} - -/* * Account user cpu time to a process. * @p: the process that the cpu time gets accounted to * @hardirq_offset: the offset to subtract from hardirq_count() @@ -3106,6 +2952,7 @@ void account_system_time(struct task_str cpustat->iowait = cputime64_add(cpustat->iowait, tmp); else cpustat->idle = cputime64_add(cpustat->idle, tmp); + p->systime += NSJIFFY; /* Account for system time used */ acct_update_integrals(p); } @@ -3131,77 +2978,38 @@ void account_steal_time(struct task_stru cpustat->steal = cputime64_add(cpustat->steal, tmp); } -static void task_running_tick(struct rq *rq, struct task_struct *p) +static void time_slice_expired(struct task_struct *p, struct rq *rq) +{ + set_tsk_need_resched(p); + p->time_slice = rr_interval(p); + requeue_task(p, rq, effective_prio(p)); +} + +static void task_used_tick(struct rq *rq, struct task_struct *p) { - if (p->array != rq->active) { - /* Task has expired but was not scheduled yet */ + if (unlikely(!task_queued(p))) { + /* Task has expired but was not scheduled off yet */ set_tsk_need_resched(p); return; } spin_lock(&rq->lock); + p->runtime -= NSJIFFY; /* - * The task was running during this tick - update the - * time slice counter. Note: we do not update a thread's - * priority until it either goes to sleep or uses up its - * timeslice. This makes it possible for interactive tasks - * to use up their timeslices at their highest priority levels. + * Tasks lose bonus each time they use up a full slice(). */ - if (rt_task(p)) { - /* - * RR tasks need a special form of timeslice management. - * FIFO tasks have no timeslices. - */ - if ((p->policy == SCHED_RR) && !--p->time_slice) { - p->time_slice = task_timeslice(p); - p->first_time_slice = 0; - set_tsk_need_resched(p); - - /* put it at the end of the queue: */ - requeue_task(p, rq->active); - } + if (!--p->slice) { + dec_bonus(p); + p->totalrun = 0; + p->slice = slice(p); + time_slice_expired(p, rq); goto out_unlock; } - if (!--p->time_slice) { - dequeue_task(p, rq->active); - set_tsk_need_resched(p); - p->prio = effective_prio(p); - p->time_slice = task_timeslice(p); - p->first_time_slice = 0; - - if (!rq->expired_timestamp) - 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; - } else - enqueue_task(p, rq->active); - } else { - /* - * Prevent a too long timeslice allowing a task to monopolize - * the CPU. We do this by splitting up the timeslice into - * smaller pieces. - * - * Note: this does not mean the task's timeslices expire or - * get lost in any way, they just might be preempted by - * another task of equal priority. (one with higher - * priority would have preempted this task already.) We - * requeue this task to the end of the list on this priority - * level, which is in essence a round-robin of tasks with - * equal priority. - * - * This only applies to tasks in the interactive - * delta range with at least TIMESLICE_GRANULARITY to requeue. - */ - if (TASK_INTERACTIVE(p) && !((task_timeslice(p) - - p->time_slice) % TIMESLICE_GRANULARITY(p)) && - (p->time_slice >= TIMESLICE_GRANULARITY(p)) && - (p->array == rq->active)) { - - requeue_task(p, rq->active); - set_tsk_need_resched(p); - } - } + /* + * Tasks that run out of time_slice but still have slice left get + * requeued with a lower priority && RR_INTERVAL time_slice. + */ + if (!--p->time_slice) + time_slice_expired(p, rq); out_unlock: spin_unlock(&rq->lock); } @@ -3209,9 +3017,6 @@ out_unlock: /* * This function gets called by the timer code, with HZ frequency. * We call it with interrupts disabled. - * - * It also gets called by the fork code, when changing the parent's - * timeslices. */ void scheduler_tick(void) { @@ -3220,10 +3025,10 @@ void scheduler_tick(void) int cpu = smp_processor_id(); struct rq *rq = cpu_rq(cpu); - update_cpu_clock(p, rq, now); + update_cpu_clock(p, rq, now, 1); + if (p != rq->idle && p->policy != SCHED_FIFO && p->runtime >= NSJIFFY) + task_used_tick(rq, p); - if (p != rq->idle) - task_running_tick(rq, p); #ifdef CONFIG_SMP update_load(rq); if (time_after_eq(jiffies, rq->next_balance)) @@ -3269,25 +3074,17 @@ EXPORT_SYMBOL(sub_preempt_count); #endif -static inline int interactive_sleep(enum sleep_type sleep_type) -{ - return (sleep_type == SLEEP_INTERACTIVE || - sleep_type == SLEEP_INTERRUPTED); -} - /* * schedule() is the main scheduler function. */ asmlinkage void __sched schedule(void) { struct task_struct *prev, *next; - struct prio_array *array; struct list_head *queue; unsigned long long now; - unsigned long run_time; - int cpu, idx, new_prio; long *switch_count; struct rq *rq; + int cpu, idx; /* * Test if we are atomic. Since do_exit() needs to call into @@ -3323,18 +3120,6 @@ need_resched_nonpreemptible: schedstat_inc(rq, sched_cnt); now = sched_clock(); - if (likely((long long)(now - prev->timestamp) < NS_MAX_SLEEP_AVG)) { - run_time = now - prev->timestamp; - if (unlikely((long long)(now - prev->timestamp) < 0)) - run_time = 0; - } else - run_time = NS_MAX_SLEEP_AVG; - - /* - * Tasks charged proportionately less run_time at high sleep_avg to - * delay them losing their interactive status - */ - run_time /= (CURRENT_BONUS(prev) ? : 1); spin_lock_irq(&rq->lock); @@ -3345,8 +3130,10 @@ need_resched_nonpreemptible: unlikely(signal_pending(prev)))) prev->state = TASK_RUNNING; else { - if (prev->state == TASK_UNINTERRUPTIBLE) + if (prev->state == TASK_UNINTERRUPTIBLE) { + prev->flags |= PF_NONSLEEP; rq->nr_uninterruptible++; + } deactivate_task(prev, rq); } } @@ -3356,59 +3143,25 @@ need_resched_nonpreemptible: idle_balance(cpu, rq); if (!rq->nr_running) { next = rq->idle; - rq->expired_timestamp = 0; goto switch_tasks; } } - 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; - } - - idx = sched_find_first_bit(array->bitmap); - queue = array->queue + idx; + idx = sched_find_first_bit(rq->bitmap); + queue = rq->queue + idx; next = list_entry(queue->next, struct task_struct, run_list); - if (!rt_task(next) && interactive_sleep(next->sleep_type)) { - unsigned long long delta = now - next->timestamp; - if (unlikely((long long)(now - next->timestamp) < 0)) - delta = 0; - - if (next->sleep_type == SLEEP_INTERACTIVE) - delta = delta * (ON_RUNQUEUE_WEIGHT * 128 / 100) / 128; - - array = next->array; - new_prio = recalc_task_prio(next, next->timestamp + delta); - - if (unlikely(next->prio != new_prio)) { - dequeue_task(next, array); - next->prio = new_prio; - enqueue_task(next, array); - } - } - next->sleep_type = SLEEP_NORMAL; switch_tasks: if (next == rq->idle) schedstat_inc(rq, sched_goidle); - prefetch(next); - prefetch_stack(next); + else { + prefetch(next); + prefetch_stack(next); + } clear_tsk_need_resched(prev); rcu_qsctr_inc(task_cpu(prev)); - update_cpu_clock(prev, rq, now); - - prev->sleep_avg -= run_time; - if ((long)prev->sleep_avg <= 0) - prev->sleep_avg = 0; + update_cpu_clock(prev, rq, now, 0); prev->timestamp = prev->last_ran = now; sched_info_switch(prev, next); @@ -3844,29 +3597,22 @@ EXPORT_SYMBOL(sleep_on_timeout); */ void rt_mutex_setprio(struct task_struct *p, int prio) { - struct prio_array *array; unsigned long flags; + int queued, oldprio; struct rq *rq; - int oldprio; BUG_ON(prio < 0 || prio > MAX_PRIO); rq = task_rq_lock(p, &flags); oldprio = p->prio; - array = p->array; - if (array) - dequeue_task(p, array); + queued = task_queued(p); + if (queued) + dequeue_task(p, rq); p->prio = prio; - if (array) { - /* - * If changing to an RT priority then queue it - * in the active array! - */ - if (rt_task(p)) - array = rq->active; - enqueue_task(p, array); + if (queued) { + enqueue_task(p, rq); /* * Reschedule if we are currently running on this runqueue and * our priority decreased, or if we are not currently running on @@ -3875,8 +3621,8 @@ void rt_mutex_setprio(struct task_struct if (task_running(rq, p)) { if (p->prio > oldprio) resched_task(rq->curr); - } else if (TASK_PREEMPTS_CURR(p, rq)) - resched_task(rq->curr); + } else + try_preempt(p, rq); } task_rq_unlock(rq, &flags); } @@ -3885,8 +3631,7 @@ void rt_mutex_setprio(struct task_struct void set_user_nice(struct task_struct *p, long nice) { - struct prio_array *array; - int old_prio, delta; + int queued, old_prio,delta; unsigned long flags; struct rq *rq; @@ -3907,20 +3652,22 @@ void set_user_nice(struct task_struct *p p->static_prio = NICE_TO_PRIO(nice); goto out_unlock; } - array = p->array; - if (array) { - dequeue_task(p, array); + queued = task_queued(p); + if (queued) { + dequeue_task(p, rq); dec_raw_weighted_load(rq, p); } p->static_prio = NICE_TO_PRIO(nice); set_load_weight(p); old_prio = p->prio; + if (p->bonus > bonus(p)) + p->bonus= bonus(p); p->prio = effective_prio(p); delta = p->prio - old_prio; - if (array) { - enqueue_task(p, array); + if (queued) { + enqueue_task(p, rq); inc_raw_weighted_load(rq, p); /* * If the task increased its priority or is running and @@ -4043,18 +3790,13 @@ static inline struct task_struct *find_p /* Actually do priority change: must hold rq lock. */ static void __setscheduler(struct task_struct *p, int policy, int prio) { - BUG_ON(p->array); + BUG_ON(task_queued(p)); p->policy = policy; p->rt_priority = prio; p->normal_prio = normal_prio(p); /* we are holding p->pi_lock already */ p->prio = rt_mutex_getprio(p); - /* - * SCHED_BATCH tasks are treated as perpetual CPU hogs: - */ - if (policy == SCHED_BATCH) - p->sleep_avg = 0; set_load_weight(p); } @@ -4069,8 +3811,7 @@ static void __setscheduler(struct task_s int sched_setscheduler(struct task_struct *p, int policy, struct sched_param *param) { - int retval, oldprio, oldpolicy = -1; - struct prio_array *array; + int queued, retval, oldprio, oldpolicy = -1; unsigned long flags; struct rq *rq; @@ -4144,12 +3885,12 @@ recheck: spin_unlock_irqrestore(&p->pi_lock, flags); goto recheck; } - array = p->array; - if (array) + queued = task_queued(p); + if (queued) deactivate_task(p, rq); oldprio = p->prio; __setscheduler(p, policy, param->sched_priority); - if (array) { + if (queued) { __activate_task(p, rq); /* * Reschedule if we are currently running on this runqueue and @@ -4159,8 +3900,8 @@ recheck: if (task_running(rq, p)) { if (p->prio > oldprio) resched_task(rq->curr); - } else if (TASK_PREEMPTS_CURR(p, rq)) - resched_task(rq->curr); + } else + try_preempt(p, rq); } __task_rq_unlock(rq); spin_unlock_irqrestore(&p->pi_lock, flags); @@ -4432,41 +4173,23 @@ asmlinkage long sys_sched_getaffinity(pi /** * sys_sched_yield - yield the current processor to other threads. * - * This function yields the current CPU by moving the calling thread - * to the expired array. If there are no other threads running on this - * CPU then this function will return. + * This function yields the current CPU by dropping the priority of current + * to the lowest priority. */ asmlinkage long sys_sched_yield(void) { struct rq *rq = this_rq_lock(); - struct prio_array *array = current->array, *target = rq->expired; + int newprio = current->prio; 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 (rt_task(current)) - target = rq->active; + newprio = current->prio; + schedstat_inc(rq, yld_cnt); + current->slice = slice(current); + current->time_slice = rr_interval(current); + if (likely(!rt_task(current))) + newprio = MIN_USER_PRIO; - if (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); - - if (array != target) { - dequeue_task(current, array); - enqueue_task(current, target); - } else - /* - * requeue_task is cheaper so perform that if possible. - */ - requeue_task(current, array); + requeue_task(current, rq, newprio); /* * Since we are going to call schedule() anyway, there's @@ -4677,7 +4400,7 @@ long sys_sched_rr_get_interval(pid_t pid goto out_unlock; jiffies_to_timespec(p->policy == SCHED_FIFO ? - 0 : task_timeslice(p), &t); + 0 : slice(p), &t); read_unlock(&tasklist_lock); retval = copy_to_user(interval, &t, sizeof(t)) ? -EFAULT : 0; out_nounlock: @@ -4772,8 +4495,6 @@ void __cpuinit init_idle(struct task_str unsigned long flags; idle->timestamp = sched_clock(); - idle->sleep_avg = 0; - idle->array = NULL; idle->prio = idle->normal_prio = MAX_PRIO; idle->state = TASK_RUNNING; idle->cpus_allowed = cpumask_of_cpu(cpu); @@ -4893,7 +4614,7 @@ static int __migrate_task(struct task_st goto out; set_task_cpu(p, dest_cpu); - if (p->array) { + if (task_queued(p)) { /* * Sync timestamp with rq_dest's before activating. * The same thing could be achieved by doing this step @@ -4904,8 +4625,7 @@ static int __migrate_task(struct task_st + rq_dest->most_recent_timestamp; deactivate_task(p, rq_src); __activate_task(p, rq_dest); - if (TASK_PREEMPTS_CURR(p, rq_dest)) - resched_task(rq_dest->curr); + try_preempt(p, rq_dest); } ret = 1; out: @@ -5134,7 +4854,7 @@ static void migrate_dead_tasks(unsigned for (arr = 0; arr < 2; arr++) { for (i = 0; i < MAX_PRIO; i++) { - struct list_head *list = &rq->arrays[arr].queue[i]; + struct list_head *list = &rq->queue[i]; while (!list_empty(list)) migrate_dead(dead_cpu, list_entry(list->next, @@ -6725,19 +6445,16 @@ int in_sched_functions(unsigned long add void __init sched_init(void) { - int i, j, k; + int i; for_each_possible_cpu(i) { - struct prio_array *array; struct rq *rq; + int j; rq = cpu_rq(i); spin_lock_init(&rq->lock); lockdep_set_class(&rq->lock, &rq->rq_lock_key); rq->nr_running = 0; - rq->active = rq->arrays; - rq->expired = rq->arrays + 1; - rq->best_expired_prio = MAX_PRIO; #ifdef CONFIG_SMP rq->sd = NULL; @@ -6751,15 +6468,11 @@ void __init sched_init(void) #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); - } + for (j = 0; j < MAX_PRIO; j++) + INIT_LIST_HEAD(&rq->queue[j]); + memset(rq->bitmap, 0, BITS_TO_LONGS(MAX_PRIO)*sizeof(long)); + /* delimiter for bitsearch */ + __set_bit(MAX_PRIO, rq->bitmap); } set_load_weight(&init_task); @@ -6815,10 +6528,10 @@ EXPORT_SYMBOL(__might_sleep); #ifdef CONFIG_MAGIC_SYSRQ void normalize_rt_tasks(void) { - struct prio_array *array; struct task_struct *p; unsigned long flags; struct rq *rq; + int queued; read_lock_irq(&tasklist_lock); for_each_process(p) { @@ -6828,11 +6541,11 @@ void normalize_rt_tasks(void) spin_lock_irqsave(&p->pi_lock, flags); rq = __task_rq_lock(p); - array = p->array; - if (array) + queued = task_queued(p); + if (queued) deactivate_task(p, task_rq(p)); __setscheduler(p, SCHED_NORMAL, 0); - if (array) { + if (queued) { __activate_task(p, task_rq(p)); resched_task(rq->curr); }