7 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
9 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15 #define CLONE_THREAD 0x00010000 /* Same thread group? */
16 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
17 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
24 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
25 #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
26 #define CLONE_NEWIPC 0x08000000 /* New ipcs */
27 #define CLONE_NEWUSER 0x10000000 /* New user namespace */
28 #define CLONE_NEWNET 0x40000000 /* New network namespace */
33 #define SCHED_NORMAL 0
37 /* SCHED_ISO: reserved but not implemented yet */
46 #include <asm/param.h> /* for HZ */
48 #include <linux/capability.h>
49 #include <linux/threads.h>
50 #include <linux/kernel.h>
51 #include <linux/types.h>
52 #include <linux/timex.h>
53 #include <linux/jiffies.h>
54 #include <linux/rbtree.h>
55 #include <linux/thread_info.h>
56 #include <linux/cpumask.h>
57 #include <linux/errno.h>
58 #include <linux/nodemask.h>
59 #include <linux/mm_types.h>
61 #include <asm/system.h>
62 #include <asm/semaphore.h>
64 #include <asm/ptrace.h>
65 #include <asm/cputime.h>
67 #include <linux/smp.h>
68 #include <linux/sem.h>
69 #include <linux/signal.h>
70 #include <linux/securebits.h>
71 #include <linux/fs_struct.h>
72 #include <linux/compiler.h>
73 #include <linux/completion.h>
74 #include <linux/pid.h>
75 #include <linux/percpu.h>
76 #include <linux/topology.h>
77 #include <linux/proportions.h>
78 #include <linux/seccomp.h>
79 #include <linux/rcupdate.h>
80 #include <linux/futex.h>
81 #include <linux/rtmutex.h>
83 #include <linux/time.h>
84 #include <linux/param.h>
85 #include <linux/resource.h>
86 #include <linux/timer.h>
87 #include <linux/hrtimer.h>
88 #include <linux/task_io_accounting.h>
89 #include <linux/kobject.h>
91 #include <asm/processor.h>
94 struct futex_pi_state
;
98 * List of flags we want to share for kernel threads,
99 * if only because they are not used by them anyway.
101 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
104 * These are the constant used to fake the fixed-point load-average
105 * counting. Some notes:
106 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
107 * a load-average precision of 10 bits integer + 11 bits fractional
108 * - if you want to count load-averages more often, you need more
109 * precision, or rounding will get you. With 2-second counting freq,
110 * the EXP_n values would be 1981, 2034 and 2043 if still using only
113 extern unsigned long avenrun
[]; /* Load averages */
115 #define FSHIFT 11 /* nr of bits of precision */
116 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
117 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
118 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
119 #define EXP_5 2014 /* 1/exp(5sec/5min) */
120 #define EXP_15 2037 /* 1/exp(5sec/15min) */
122 #define CALC_LOAD(load,exp,n) \
124 load += n*(FIXED_1-exp); \
127 extern unsigned long total_forks
;
128 extern int nr_threads
;
129 DECLARE_PER_CPU(unsigned long, process_counts
);
130 extern int nr_processes(void);
131 extern unsigned long nr_running(void);
132 extern unsigned long nr_uninterruptible(void);
133 extern unsigned long nr_active(void);
134 extern unsigned long nr_iowait(void);
135 extern unsigned long weighted_cpuload(const int cpu
);
140 #ifdef CONFIG_SCHED_DEBUG
141 extern void proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
);
142 extern void proc_sched_set_task(struct task_struct
*p
);
144 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
);
147 proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
)
150 static inline void proc_sched_set_task(struct task_struct
*p
)
154 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
)
160 * Task state bitmask. NOTE! These bits are also
161 * encoded in fs/proc/array.c: get_task_state().
163 * We have two separate sets of flags: task->state
164 * is about runnability, while task->exit_state are
165 * about the task exiting. Confusing, but this way
166 * modifying one set can't modify the other one by
169 #define TASK_RUNNING 0
170 #define TASK_INTERRUPTIBLE 1
171 #define TASK_UNINTERRUPTIBLE 2
172 #define TASK_STOPPED 4
173 #define TASK_TRACED 8
174 /* in tsk->exit_state */
175 #define EXIT_ZOMBIE 16
177 /* in tsk->state again */
180 #define __set_task_state(tsk, state_value) \
181 do { (tsk)->state = (state_value); } while (0)
182 #define set_task_state(tsk, state_value) \
183 set_mb((tsk)->state, (state_value))
186 * set_current_state() includes a barrier so that the write of current->state
187 * is correctly serialised wrt the caller's subsequent test of whether to
190 * set_current_state(TASK_UNINTERRUPTIBLE);
191 * if (do_i_need_to_sleep())
194 * If the caller does not need such serialisation then use __set_current_state()
196 #define __set_current_state(state_value) \
197 do { current->state = (state_value); } while (0)
198 #define set_current_state(state_value) \
199 set_mb(current->state, (state_value))
201 /* Task command name length */
202 #define TASK_COMM_LEN 16
204 #include <linux/spinlock.h>
207 * This serializes "schedule()" and also protects
208 * the run-queue from deletions/modifications (but
209 * _adding_ to the beginning of the run-queue has
212 extern rwlock_t tasklist_lock
;
213 extern spinlock_t mmlist_lock
;
217 extern void sched_init(void);
218 extern void sched_init_smp(void);
219 extern void init_idle(struct task_struct
*idle
, int cpu
);
220 extern void init_idle_bootup_task(struct task_struct
*idle
);
222 extern cpumask_t nohz_cpu_mask
;
223 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
224 extern int select_nohz_load_balancer(int cpu
);
226 static inline int select_nohz_load_balancer(int cpu
)
233 * Only dump TASK_* tasks. (0 for all tasks)
235 extern void show_state_filter(unsigned long state_filter
);
237 static inline void show_state(void)
239 show_state_filter(0);
242 extern void show_regs(struct pt_regs
*);
245 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
246 * task), SP is the stack pointer of the first frame that should be shown in the back
247 * trace (or NULL if the entire call-chain of the task should be shown).
249 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
251 void io_schedule(void);
252 long io_schedule_timeout(long timeout
);
254 extern void cpu_init (void);
255 extern void trap_init(void);
256 extern void update_process_times(int user
);
257 extern void scheduler_tick(void);
259 #ifdef CONFIG_DETECT_SOFTLOCKUP
260 extern void softlockup_tick(void);
261 extern void spawn_softlockup_task(void);
262 extern void touch_softlockup_watchdog(void);
263 extern void touch_all_softlockup_watchdogs(void);
265 static inline void softlockup_tick(void)
268 static inline void spawn_softlockup_task(void)
271 static inline void touch_softlockup_watchdog(void)
274 static inline void touch_all_softlockup_watchdogs(void)
280 /* Attach to any functions which should be ignored in wchan output. */
281 #define __sched __attribute__((__section__(".sched.text")))
282 /* Is this address in the __sched functions? */
283 extern int in_sched_functions(unsigned long addr
);
285 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
286 extern signed long FASTCALL(schedule_timeout(signed long timeout
));
287 extern signed long schedule_timeout_interruptible(signed long timeout
);
288 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
289 asmlinkage
void schedule(void);
292 struct user_namespace
;
294 /* Maximum number of active map areas.. This is a random (large) number */
295 #define DEFAULT_MAX_MAP_COUNT 65536
297 extern int sysctl_max_map_count
;
299 #include <linux/aio.h>
302 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
303 unsigned long, unsigned long);
305 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
306 unsigned long len
, unsigned long pgoff
,
307 unsigned long flags
);
308 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
309 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
311 #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
313 * The mm counters are not protected by its page_table_lock,
314 * so must be incremented atomically.
316 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
317 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
318 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
319 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
320 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
322 #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
324 * The mm counters are protected by its page_table_lock,
325 * so can be incremented directly.
327 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
328 #define get_mm_counter(mm, member) ((mm)->_##member)
329 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
330 #define inc_mm_counter(mm, member) (mm)->_##member++
331 #define dec_mm_counter(mm, member) (mm)->_##member--
333 #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
335 #define get_mm_rss(mm) \
336 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
337 #define update_hiwater_rss(mm) do { \
338 unsigned long _rss = get_mm_rss(mm); \
339 if ((mm)->hiwater_rss < _rss) \
340 (mm)->hiwater_rss = _rss; \
342 #define update_hiwater_vm(mm) do { \
343 if ((mm)->hiwater_vm < (mm)->total_vm) \
344 (mm)->hiwater_vm = (mm)->total_vm; \
347 extern void set_dumpable(struct mm_struct
*mm
, int value
);
348 extern int get_dumpable(struct mm_struct
*mm
);
352 #define MMF_DUMPABLE 0 /* core dump is permitted */
353 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
354 #define MMF_DUMPABLE_BITS 2
356 /* coredump filter bits */
357 #define MMF_DUMP_ANON_PRIVATE 2
358 #define MMF_DUMP_ANON_SHARED 3
359 #define MMF_DUMP_MAPPED_PRIVATE 4
360 #define MMF_DUMP_MAPPED_SHARED 5
361 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
362 #define MMF_DUMP_FILTER_BITS 4
363 #define MMF_DUMP_FILTER_MASK \
364 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
365 #define MMF_DUMP_FILTER_DEFAULT \
366 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED))
368 struct sighand_struct
{
370 struct k_sigaction action
[_NSIG
];
372 wait_queue_head_t signalfd_wqh
;
375 struct pacct_struct
{
378 unsigned long ac_mem
;
379 cputime_t ac_utime
, ac_stime
;
380 unsigned long ac_minflt
, ac_majflt
;
384 * NOTE! "signal_struct" does not have it's own
385 * locking, because a shared signal_struct always
386 * implies a shared sighand_struct, so locking
387 * sighand_struct is always a proper superset of
388 * the locking of signal_struct.
390 struct signal_struct
{
394 wait_queue_head_t wait_chldexit
; /* for wait4() */
396 /* current thread group signal load-balancing target: */
397 struct task_struct
*curr_target
;
399 /* shared signal handling: */
400 struct sigpending shared_pending
;
402 /* thread group exit support */
405 * - notify group_exit_task when ->count is equal to notify_count
406 * - everyone except group_exit_task is stopped during signal delivery
407 * of fatal signals, group_exit_task processes the signal.
409 struct task_struct
*group_exit_task
;
412 /* thread group stop support, overloads group_exit_code too */
413 int group_stop_count
;
414 unsigned int flags
; /* see SIGNAL_* flags below */
416 /* POSIX.1b Interval Timers */
417 struct list_head posix_timers
;
419 /* ITIMER_REAL timer for the process */
420 struct hrtimer real_timer
;
421 struct task_struct
*tsk
;
422 ktime_t it_real_incr
;
424 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
425 cputime_t it_prof_expires
, it_virt_expires
;
426 cputime_t it_prof_incr
, it_virt_incr
;
428 /* job control IDs */
430 struct pid
*tty_old_pgrp
;
433 pid_t session __deprecated
;
437 /* boolean value for session group leader */
440 struct tty_struct
*tty
; /* NULL if no tty */
443 * Cumulative resource counters for dead threads in the group,
444 * and for reaped dead child processes forked by this group.
445 * Live threads maintain their own counters and add to these
446 * in __exit_signal, except for the group leader.
448 cputime_t utime
, stime
, cutime
, cstime
;
451 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
452 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
453 unsigned long inblock
, oublock
, cinblock
, coublock
;
456 * Cumulative ns of scheduled CPU time for dead threads in the
457 * group, not including a zombie group leader. (This only differs
458 * from jiffies_to_ns(utime + stime) if sched_clock uses something
459 * other than jiffies.)
461 unsigned long long sum_sched_runtime
;
464 * We don't bother to synchronize most readers of this at all,
465 * because there is no reader checking a limit that actually needs
466 * to get both rlim_cur and rlim_max atomically, and either one
467 * alone is a single word that can safely be read normally.
468 * getrlimit/setrlimit use task_lock(current->group_leader) to
469 * protect this instead of the siglock, because they really
470 * have no need to disable irqs.
472 struct rlimit rlim
[RLIM_NLIMITS
];
474 struct list_head cpu_timers
[3];
476 /* keep the process-shared keyrings here so that they do the right
477 * thing in threads created with CLONE_THREAD */
479 struct key
*session_keyring
; /* keyring inherited over fork */
480 struct key
*process_keyring
; /* keyring private to this process */
482 #ifdef CONFIG_BSD_PROCESS_ACCT
483 struct pacct_struct pacct
; /* per-process accounting information */
485 #ifdef CONFIG_TASKSTATS
486 struct taskstats
*stats
;
490 struct tty_audit_buf
*tty_audit_buf
;
494 /* Context switch must be unlocked if interrupts are to be enabled */
495 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
496 # define __ARCH_WANT_UNLOCKED_CTXSW
500 * Bits in flags field of signal_struct.
502 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
503 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
504 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
505 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
508 * Some day this will be a full-fledged user tracking system..
511 atomic_t __count
; /* reference count */
512 atomic_t processes
; /* How many processes does this user have? */
513 atomic_t files
; /* How many open files does this user have? */
514 atomic_t sigpending
; /* How many pending signals does this user have? */
515 #ifdef CONFIG_INOTIFY_USER
516 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
517 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
519 /* protected by mq_lock */
520 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
521 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
524 struct key
*uid_keyring
; /* UID specific keyring */
525 struct key
*session_keyring
; /* UID's default session keyring */
528 /* Hash table maintenance information */
529 struct hlist_node uidhash_node
;
532 #ifdef CONFIG_FAIR_USER_SCHED
533 struct task_group
*tg
;
535 struct subsys_attribute user_attr
;
536 struct work_struct work
;
540 #ifdef CONFIG_FAIR_USER_SCHED
541 extern int uids_kobject_init(void);
543 static inline int uids_kobject_init(void) { return 0; }
546 extern struct user_struct
*find_user(uid_t
);
548 extern struct user_struct root_user
;
549 #define INIT_USER (&root_user)
551 struct backing_dev_info
;
552 struct reclaim_state
;
554 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
556 /* cumulative counters */
557 unsigned long pcount
; /* # of times run on this cpu */
558 unsigned long long cpu_time
, /* time spent on the cpu */
559 run_delay
; /* time spent waiting on a runqueue */
562 unsigned long long last_arrival
,/* when we last ran on a cpu */
563 last_queued
; /* when we were last queued to run */
564 #ifdef CONFIG_SCHEDSTATS
566 unsigned long bkl_count
;
569 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
571 #ifdef CONFIG_SCHEDSTATS
572 extern const struct file_operations proc_schedstat_operations
;
573 #endif /* CONFIG_SCHEDSTATS */
575 #ifdef CONFIG_TASK_DELAY_ACCT
576 struct task_delay_info
{
578 unsigned int flags
; /* Private per-task flags */
580 /* For each stat XXX, add following, aligned appropriately
582 * struct timespec XXX_start, XXX_end;
586 * Atomicity of updates to XXX_delay, XXX_count protected by
587 * single lock above (split into XXX_lock if contention is an issue).
591 * XXX_count is incremented on every XXX operation, the delay
592 * associated with the operation is added to XXX_delay.
593 * XXX_delay contains the accumulated delay time in nanoseconds.
595 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
596 u64 blkio_delay
; /* wait for sync block io completion */
597 u64 swapin_delay
; /* wait for swapin block io completion */
598 u32 blkio_count
; /* total count of the number of sync block */
599 /* io operations performed */
600 u32 swapin_count
; /* total count of the number of swapin block */
601 /* io operations performed */
603 #endif /* CONFIG_TASK_DELAY_ACCT */
605 static inline int sched_info_on(void)
607 #ifdef CONFIG_SCHEDSTATS
609 #elif defined(CONFIG_TASK_DELAY_ACCT)
610 extern int delayacct_on
;
625 * sched-domains (multiprocessor balancing) declarations:
629 * Increase resolution of nice-level calculations:
631 #define SCHED_LOAD_SHIFT 10
632 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
634 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
637 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
638 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
639 #define SD_BALANCE_EXEC 4 /* Balance on exec */
640 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
641 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
642 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
643 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
644 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
645 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
646 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
647 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
649 #define BALANCE_FOR_MC_POWER \
650 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
652 #define BALANCE_FOR_PKG_POWER \
653 ((sched_mc_power_savings || sched_smt_power_savings) ? \
654 SD_POWERSAVINGS_BALANCE : 0)
656 #define test_sd_parent(sd, flag) ((sd->parent && \
657 (sd->parent->flags & flag)) ? 1 : 0)
661 struct sched_group
*next
; /* Must be a circular list */
665 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
666 * single CPU. This is read only (except for setup, hotplug CPU).
667 * Note : Never change cpu_power without recompute its reciprocal
669 unsigned int __cpu_power
;
671 * reciprocal value of cpu_power to avoid expensive divides
672 * (see include/linux/reciprocal_div.h)
674 u32 reciprocal_cpu_power
;
677 struct sched_domain
{
678 /* These fields must be setup */
679 struct sched_domain
*parent
; /* top domain must be null terminated */
680 struct sched_domain
*child
; /* bottom domain must be null terminated */
681 struct sched_group
*groups
; /* the balancing groups of the domain */
682 cpumask_t span
; /* span of all CPUs in this domain */
683 unsigned long min_interval
; /* Minimum balance interval ms */
684 unsigned long max_interval
; /* Maximum balance interval ms */
685 unsigned int busy_factor
; /* less balancing by factor if busy */
686 unsigned int imbalance_pct
; /* No balance until over watermark */
687 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
688 unsigned int busy_idx
;
689 unsigned int idle_idx
;
690 unsigned int newidle_idx
;
691 unsigned int wake_idx
;
692 unsigned int forkexec_idx
;
693 int flags
; /* See SD_* */
695 /* Runtime fields. */
696 unsigned long last_balance
; /* init to jiffies. units in jiffies */
697 unsigned int balance_interval
; /* initialise to 1. units in ms. */
698 unsigned int nr_balance_failed
; /* initialise to 0 */
700 #ifdef CONFIG_SCHEDSTATS
701 /* load_balance() stats */
702 unsigned long lb_count
[CPU_MAX_IDLE_TYPES
];
703 unsigned long lb_failed
[CPU_MAX_IDLE_TYPES
];
704 unsigned long lb_balanced
[CPU_MAX_IDLE_TYPES
];
705 unsigned long lb_imbalance
[CPU_MAX_IDLE_TYPES
];
706 unsigned long lb_gained
[CPU_MAX_IDLE_TYPES
];
707 unsigned long lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
708 unsigned long lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
709 unsigned long lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
711 /* Active load balancing */
712 unsigned long alb_count
;
713 unsigned long alb_failed
;
714 unsigned long alb_pushed
;
716 /* SD_BALANCE_EXEC stats */
717 unsigned long sbe_count
;
718 unsigned long sbe_balanced
;
719 unsigned long sbe_pushed
;
721 /* SD_BALANCE_FORK stats */
722 unsigned long sbf_count
;
723 unsigned long sbf_balanced
;
724 unsigned long sbf_pushed
;
726 /* try_to_wake_up() stats */
727 unsigned long ttwu_wake_remote
;
728 unsigned long ttwu_move_affine
;
729 unsigned long ttwu_move_balance
;
733 #endif /* CONFIG_SMP */
736 * A runqueue laden with a single nice 0 task scores a weighted_cpuload of
737 * SCHED_LOAD_SCALE. This function returns 1 if any cpu is laden with a
738 * task of nice 0 or enough lower priority tasks to bring up the
741 static inline int above_background_load(void)
745 for_each_online_cpu(cpu
) {
746 if (weighted_cpuload(cpu
) >= SCHED_LOAD_SCALE
)
752 struct io_context
; /* See blkdev.h */
755 #define NGROUPS_SMALL 32
756 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
760 gid_t small_block
[NGROUPS_SMALL
];
766 * get_group_info() must be called with the owning task locked (via task_lock())
767 * when task != current. The reason being that the vast majority of callers are
768 * looking at current->group_info, which can not be changed except by the
769 * current task. Changing current->group_info requires the task lock, too.
771 #define get_group_info(group_info) do { \
772 atomic_inc(&(group_info)->usage); \
775 #define put_group_info(group_info) do { \
776 if (atomic_dec_and_test(&(group_info)->usage)) \
777 groups_free(group_info); \
780 extern struct group_info
*groups_alloc(int gidsetsize
);
781 extern void groups_free(struct group_info
*group_info
);
782 extern int set_current_groups(struct group_info
*group_info
);
783 extern int groups_search(struct group_info
*group_info
, gid_t grp
);
784 /* access the groups "array" with this macro */
785 #define GROUP_AT(gi, i) \
786 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
788 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
789 extern void prefetch_stack(struct task_struct
*t
);
791 static inline void prefetch_stack(struct task_struct
*t
) { }
794 struct audit_context
; /* See audit.c */
796 struct pipe_inode_info
;
797 struct uts_namespace
;
803 const struct sched_class
*next
;
805 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
, int wakeup
);
806 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
, int sleep
);
807 void (*yield_task
) (struct rq
*rq
);
809 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
);
811 struct task_struct
* (*pick_next_task
) (struct rq
*rq
);
812 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
);
814 unsigned long (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
816 unsigned long max_nr_move
, unsigned long max_load_move
,
817 struct sched_domain
*sd
, enum cpu_idle_type idle
,
818 int *all_pinned
, int *this_best_prio
);
820 void (*set_curr_task
) (struct rq
*rq
);
821 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
);
822 void (*task_new
) (struct rq
*rq
, struct task_struct
*p
);
826 unsigned long weight
, inv_weight
;
830 * CFS stats for a schedulable entity (task, task-group etc)
832 * Current field usage histogram:
839 struct sched_entity
{
840 struct load_weight load
; /* for load-balancing */
841 struct rb_node run_node
;
846 u64 sum_exec_runtime
;
848 u64 prev_sum_exec_runtime
;
850 #ifdef CONFIG_SCHEDSTATS
856 s64 sum_sleep_runtime
;
864 u64 nr_migrations_cold
;
865 u64 nr_failed_migrations_affine
;
866 u64 nr_failed_migrations_running
;
867 u64 nr_failed_migrations_hot
;
868 u64 nr_forced_migrations
;
869 u64 nr_forced2_migrations
;
873 u64 nr_wakeups_migrate
;
874 u64 nr_wakeups_local
;
875 u64 nr_wakeups_remote
;
876 u64 nr_wakeups_affine
;
877 u64 nr_wakeups_affine_attempts
;
878 u64 nr_wakeups_passive
;
882 #ifdef CONFIG_FAIR_GROUP_SCHED
883 struct sched_entity
*parent
;
884 /* rq on which this entity is (to be) queued: */
885 struct cfs_rq
*cfs_rq
;
886 /* rq "owned" by this entity/group: */
892 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
895 unsigned int flags
; /* per process flags, defined below */
898 int lock_depth
; /* BKL lock depth */
901 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
906 int prio
, static_prio
, normal_prio
;
907 struct list_head run_list
;
908 const struct sched_class
*sched_class
;
909 struct sched_entity se
;
911 #ifdef CONFIG_PREEMPT_NOTIFIERS
912 /* list of struct preempt_notifier: */
913 struct hlist_head preempt_notifiers
;
916 unsigned short ioprio
;
917 #ifdef CONFIG_BLK_DEV_IO_TRACE
918 unsigned int btrace_seq
;
922 cpumask_t cpus_allowed
;
923 unsigned int time_slice
;
925 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
926 struct sched_info sched_info
;
929 struct list_head tasks
;
931 * ptrace_list/ptrace_children forms the list of my children
932 * that were stolen by a ptracer.
934 struct list_head ptrace_children
;
935 struct list_head ptrace_list
;
937 struct mm_struct
*mm
, *active_mm
;
940 struct linux_binfmt
*binfmt
;
942 int exit_code
, exit_signal
;
943 int pdeath_signal
; /* The signal sent when the parent dies */
945 unsigned int personality
;
950 #ifdef CONFIG_CC_STACKPROTECTOR
951 /* Canary value for the -fstack-protector gcc feature */
952 unsigned long stack_canary
;
955 * pointers to (original) parent process, youngest child, younger sibling,
956 * older sibling, respectively. (p->father can be replaced with
959 struct task_struct
*real_parent
; /* real parent process (when being debugged) */
960 struct task_struct
*parent
; /* parent process */
962 * children/sibling forms the list of my children plus the
963 * tasks I'm ptracing.
965 struct list_head children
; /* list of my children */
966 struct list_head sibling
; /* linkage in my parent's children list */
967 struct task_struct
*group_leader
; /* threadgroup leader */
969 /* PID/PID hash table linkage. */
970 struct pid_link pids
[PIDTYPE_MAX
];
971 struct list_head thread_group
;
973 struct completion
*vfork_done
; /* for vfork() */
974 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
975 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
977 unsigned int rt_priority
;
978 cputime_t utime
, stime
;
980 unsigned long nvcsw
, nivcsw
; /* context switch counts */
981 struct timespec start_time
; /* monotonic time */
982 struct timespec real_start_time
; /* boot based time */
983 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
984 unsigned long min_flt
, maj_flt
;
986 cputime_t it_prof_expires
, it_virt_expires
;
987 unsigned long long it_sched_expires
;
988 struct list_head cpu_timers
[3];
990 /* process credentials */
991 uid_t uid
,euid
,suid
,fsuid
;
992 gid_t gid
,egid
,sgid
,fsgid
;
993 struct group_info
*group_info
;
994 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
;
995 unsigned keep_capabilities
:1;
996 struct user_struct
*user
;
998 struct key
*request_key_auth
; /* assumed request_key authority */
999 struct key
*thread_keyring
; /* keyring private to this thread */
1000 unsigned char jit_keyring
; /* default keyring to attach requested keys to */
1003 * fpu_counter contains the number of consecutive context switches
1004 * that the FPU is used. If this is over a threshold, the lazy fpu
1005 * saving becomes unlazy to save the trap. This is an unsigned char
1006 * so that after 256 times the counter wraps and the behavior turns
1007 * lazy again; this to deal with bursty apps that only use FPU for
1010 unsigned char fpu_counter
;
1011 int oomkilladj
; /* OOM kill score adjustment (bit shift). */
1012 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1013 - access with [gs]et_task_comm (which lock
1014 it with task_lock())
1015 - initialized normally by flush_old_exec */
1016 /* file system info */
1017 int link_count
, total_link_count
;
1018 #ifdef CONFIG_SYSVIPC
1020 struct sysv_sem sysvsem
;
1022 /* CPU-specific state of this task */
1023 struct thread_struct thread
;
1024 /* filesystem information */
1025 struct fs_struct
*fs
;
1026 /* open file information */
1027 struct files_struct
*files
;
1029 struct nsproxy
*nsproxy
;
1030 /* signal handlers */
1031 struct signal_struct
*signal
;
1032 struct sighand_struct
*sighand
;
1034 sigset_t blocked
, real_blocked
;
1035 sigset_t saved_sigmask
; /* To be restored with TIF_RESTORE_SIGMASK */
1036 struct sigpending pending
;
1038 unsigned long sas_ss_sp
;
1040 int (*notifier
)(void *priv
);
1041 void *notifier_data
;
1042 sigset_t
*notifier_mask
;
1045 struct audit_context
*audit_context
;
1048 /* Thread group tracking */
1051 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1052 spinlock_t alloc_lock
;
1054 /* Protection of the PI data structures: */
1057 #ifdef CONFIG_RT_MUTEXES
1058 /* PI waiters blocked on a rt_mutex held by this task */
1059 struct plist_head pi_waiters
;
1060 /* Deadlock detection and priority inheritance handling */
1061 struct rt_mutex_waiter
*pi_blocked_on
;
1064 #ifdef CONFIG_DEBUG_MUTEXES
1065 /* mutex deadlock detection */
1066 struct mutex_waiter
*blocked_on
;
1068 #ifdef CONFIG_TRACE_IRQFLAGS
1069 unsigned int irq_events
;
1070 int hardirqs_enabled
;
1071 unsigned long hardirq_enable_ip
;
1072 unsigned int hardirq_enable_event
;
1073 unsigned long hardirq_disable_ip
;
1074 unsigned int hardirq_disable_event
;
1075 int softirqs_enabled
;
1076 unsigned long softirq_disable_ip
;
1077 unsigned int softirq_disable_event
;
1078 unsigned long softirq_enable_ip
;
1079 unsigned int softirq_enable_event
;
1080 int hardirq_context
;
1081 int softirq_context
;
1083 #ifdef CONFIG_LOCKDEP
1084 # define MAX_LOCK_DEPTH 30UL
1087 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1088 unsigned int lockdep_recursion
;
1091 /* journalling filesystem info */
1094 /* stacked block device info */
1095 struct bio
*bio_list
, **bio_tail
;
1098 struct reclaim_state
*reclaim_state
;
1100 struct backing_dev_info
*backing_dev_info
;
1102 struct io_context
*io_context
;
1104 unsigned long ptrace_message
;
1105 siginfo_t
*last_siginfo
; /* For ptrace use. */
1107 * current io wait handle: wait queue entry to use for io waits
1108 * If this thread is processing aio, this points at the waitqueue
1109 * inside the currently handled kiocb. It may be NULL (i.e. default
1110 * to a stack based synchronous wait) if its doing sync IO.
1112 wait_queue_t
*io_wait
;
1113 #ifdef CONFIG_TASK_XACCT
1114 /* i/o counters(bytes read/written, #syscalls */
1115 u64 rchar
, wchar
, syscr
, syscw
;
1117 struct task_io_accounting ioac
;
1118 #if defined(CONFIG_TASK_XACCT)
1119 u64 acct_rss_mem1
; /* accumulated rss usage */
1120 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1121 cputime_t acct_stimexpd
;/* stime since last update */
1124 struct mempolicy
*mempolicy
;
1127 #ifdef CONFIG_CPUSETS
1128 struct cpuset
*cpuset
;
1129 nodemask_t mems_allowed
;
1130 int cpuset_mems_generation
;
1131 int cpuset_mem_spread_rotor
;
1133 struct robust_list_head __user
*robust_list
;
1134 #ifdef CONFIG_COMPAT
1135 struct compat_robust_list_head __user
*compat_robust_list
;
1137 struct list_head pi_state_list
;
1138 struct futex_pi_state
*pi_state_cache
;
1140 atomic_t fs_excl
; /* holding fs exclusive resources */
1141 struct rcu_head rcu
;
1144 * cache last used pipe for splice
1146 struct pipe_inode_info
*splice_pipe
;
1147 #ifdef CONFIG_TASK_DELAY_ACCT
1148 struct task_delay_info
*delays
;
1150 #ifdef CONFIG_FAULT_INJECTION
1153 struct prop_local_single dirties
;
1157 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1158 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1159 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1160 * values are inverted: lower p->prio value means higher priority.
1162 * The MAX_USER_RT_PRIO value allows the actual maximum
1163 * RT priority to be separate from the value exported to
1164 * user-space. This allows kernel threads to set their
1165 * priority to a value higher than any user task. Note:
1166 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1169 #define MAX_USER_RT_PRIO 100
1170 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1172 #define MAX_PRIO (MAX_RT_PRIO + 40)
1173 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1175 static inline int rt_prio(int prio
)
1177 if (unlikely(prio
< MAX_RT_PRIO
))
1182 static inline int rt_task(struct task_struct
*p
)
1184 return rt_prio(p
->prio
);
1187 static inline pid_t
process_group(struct task_struct
*tsk
)
1189 return tsk
->signal
->pgrp
;
1192 static inline pid_t
signal_session(struct signal_struct
*sig
)
1194 return sig
->__session
;
1197 static inline pid_t
process_session(struct task_struct
*tsk
)
1199 return signal_session(tsk
->signal
);
1202 static inline void set_signal_session(struct signal_struct
*sig
, pid_t session
)
1204 sig
->__session
= session
;
1207 static inline struct pid
*task_pid(struct task_struct
*task
)
1209 return task
->pids
[PIDTYPE_PID
].pid
;
1212 static inline struct pid
*task_tgid(struct task_struct
*task
)
1214 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1217 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1219 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1222 static inline struct pid
*task_session(struct task_struct
*task
)
1224 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1228 * pid_alive - check that a task structure is not stale
1229 * @p: Task structure to be checked.
1231 * Test if a process is not yet dead (at most zombie state)
1232 * If pid_alive fails, then pointers within the task structure
1233 * can be stale and must not be dereferenced.
1235 static inline int pid_alive(struct task_struct
*p
)
1237 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1241 * is_init - check if a task structure is init
1242 * @tsk: Task structure to be checked.
1244 * Check if a task structure is the first user space task the kernel created.
1246 static inline int is_init(struct task_struct
*tsk
)
1248 return tsk
->pid
== 1;
1251 extern struct pid
*cad_pid
;
1253 extern void free_task(struct task_struct
*tsk
);
1254 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1256 extern void __put_task_struct(struct task_struct
*t
);
1258 static inline void put_task_struct(struct task_struct
*t
)
1260 if (atomic_dec_and_test(&t
->usage
))
1261 __put_task_struct(t
);
1267 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1268 /* Not implemented yet, only for 486*/
1269 #define PF_STARTING 0x00000002 /* being created */
1270 #define PF_EXITING 0x00000004 /* getting shut down */
1271 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1272 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1273 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1274 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1275 #define PF_DUMPCORE 0x00000200 /* dumped core */
1276 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1277 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1278 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1279 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1280 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1281 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1282 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1283 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1284 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1285 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1286 #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1287 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1288 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1289 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1290 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1291 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1292 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1293 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1296 * Only the _current_ task can read/write to tsk->flags, but other
1297 * tasks can access tsk->flags in readonly mode for example
1298 * with tsk_used_math (like during threaded core dumping).
1299 * There is however an exception to this rule during ptrace
1300 * or during fork: the ptracer task is allowed to write to the
1301 * child->flags of its traced child (same goes for fork, the parent
1302 * can write to the child->flags), because we're guaranteed the
1303 * child is not running and in turn not changing child->flags
1304 * at the same time the parent does it.
1306 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1307 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1308 #define clear_used_math() clear_stopped_child_used_math(current)
1309 #define set_used_math() set_stopped_child_used_math(current)
1310 #define conditional_stopped_child_used_math(condition, child) \
1311 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1312 #define conditional_used_math(condition) \
1313 conditional_stopped_child_used_math(condition, current)
1314 #define copy_to_stopped_child_used_math(child) \
1315 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1316 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1317 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1318 #define used_math() tsk_used_math(current)
1321 extern int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
);
1323 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1325 if (!cpu_isset(0, new_mask
))
1331 extern unsigned long long sched_clock(void);
1334 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1335 * clock constructed from sched_clock():
1337 extern unsigned long long cpu_clock(int cpu
);
1339 extern unsigned long long
1340 task_sched_runtime(struct task_struct
*task
);
1342 /* sched_exec is called by processes performing an exec */
1344 extern void sched_exec(void);
1346 #define sched_exec() {}
1349 extern void sched_clock_idle_sleep_event(void);
1350 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1352 #ifdef CONFIG_HOTPLUG_CPU
1353 extern void idle_task_exit(void);
1355 static inline void idle_task_exit(void) {}
1358 extern void sched_idle_next(void);
1360 #ifdef CONFIG_SCHED_DEBUG
1361 extern unsigned int sysctl_sched_latency
;
1362 extern unsigned int sysctl_sched_nr_latency
;
1363 extern unsigned int sysctl_sched_wakeup_granularity
;
1364 extern unsigned int sysctl_sched_batch_wakeup_granularity
;
1365 extern unsigned int sysctl_sched_child_runs_first
;
1366 extern unsigned int sysctl_sched_features
;
1367 extern unsigned int sysctl_sched_migration_cost
;
1370 extern unsigned int sysctl_sched_compat_yield
;
1372 #ifdef CONFIG_RT_MUTEXES
1373 extern int rt_mutex_getprio(struct task_struct
*p
);
1374 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1375 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1377 static inline int rt_mutex_getprio(struct task_struct
*p
)
1379 return p
->normal_prio
;
1381 # define rt_mutex_adjust_pi(p) do { } while (0)
1384 extern void set_user_nice(struct task_struct
*p
, long nice
);
1385 extern int task_prio(const struct task_struct
*p
);
1386 extern int task_nice(const struct task_struct
*p
);
1387 extern int can_nice(const struct task_struct
*p
, const int nice
);
1388 extern int task_curr(const struct task_struct
*p
);
1389 extern int idle_cpu(int cpu
);
1390 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1391 extern struct task_struct
*idle_task(int cpu
);
1392 extern struct task_struct
*curr_task(int cpu
);
1393 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1398 * The default (Linux) execution domain.
1400 extern struct exec_domain default_exec_domain
;
1402 union thread_union
{
1403 struct thread_info thread_info
;
1404 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1407 #ifndef __HAVE_ARCH_KSTACK_END
1408 static inline int kstack_end(void *addr
)
1410 /* Reliable end of stack detection:
1411 * Some APM bios versions misalign the stack
1413 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1417 extern union thread_union init_thread_union
;
1418 extern struct task_struct init_task
;
1420 extern struct mm_struct init_mm
;
1422 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1423 extern struct task_struct
*find_task_by_pid_type(int type
, int pid
);
1424 extern void __set_special_pids(pid_t session
, pid_t pgrp
);
1426 /* per-UID process charging. */
1427 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
1428 static inline struct user_struct
*get_uid(struct user_struct
*u
)
1430 atomic_inc(&u
->__count
);
1433 extern void free_uid(struct user_struct
*);
1434 extern void switch_uid(struct user_struct
*);
1435 extern void release_uids(struct user_namespace
*ns
);
1437 #include <asm/current.h>
1439 extern void do_timer(unsigned long ticks
);
1441 extern int FASTCALL(wake_up_state(struct task_struct
* tsk
, unsigned int state
));
1442 extern int FASTCALL(wake_up_process(struct task_struct
* tsk
));
1443 extern void FASTCALL(wake_up_new_task(struct task_struct
* tsk
,
1444 unsigned long clone_flags
));
1446 extern void kick_process(struct task_struct
*tsk
);
1448 static inline void kick_process(struct task_struct
*tsk
) { }
1450 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
1451 extern void sched_dead(struct task_struct
*p
);
1453 extern int in_group_p(gid_t
);
1454 extern int in_egroup_p(gid_t
);
1456 extern void proc_caches_init(void);
1457 extern void flush_signals(struct task_struct
*);
1458 extern void ignore_signals(struct task_struct
*);
1459 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
1460 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
1462 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
1464 unsigned long flags
;
1467 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
1468 ret
= dequeue_signal(tsk
, mask
, info
);
1469 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
1474 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
1476 extern void unblock_all_signals(void);
1477 extern void release_task(struct task_struct
* p
);
1478 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
1479 extern int send_group_sig_info(int, struct siginfo
*, struct task_struct
*);
1480 extern int force_sigsegv(int, struct task_struct
*);
1481 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
1482 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1483 extern int kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1484 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
1485 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
1486 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
1487 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
1488 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
1489 extern void do_notify_parent(struct task_struct
*, int);
1490 extern void force_sig(int, struct task_struct
*);
1491 extern void force_sig_specific(int, struct task_struct
*);
1492 extern int send_sig(int, struct task_struct
*, int);
1493 extern void zap_other_threads(struct task_struct
*p
);
1494 extern int kill_proc(pid_t
, int, int);
1495 extern struct sigqueue
*sigqueue_alloc(void);
1496 extern void sigqueue_free(struct sigqueue
*);
1497 extern int send_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1498 extern int send_group_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1499 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
1500 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
1502 static inline int kill_cad_pid(int sig
, int priv
)
1504 return kill_pid(cad_pid
, sig
, priv
);
1507 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1508 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1509 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1510 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1512 static inline int is_si_special(const struct siginfo
*info
)
1514 return info
<= SEND_SIG_FORCED
;
1517 /* True if we are on the alternate signal stack. */
1519 static inline int on_sig_stack(unsigned long sp
)
1521 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
1524 static inline int sas_ss_flags(unsigned long sp
)
1526 return (current
->sas_ss_size
== 0 ? SS_DISABLE
1527 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
1531 * Routines for handling mm_structs
1533 extern struct mm_struct
* mm_alloc(void);
1535 /* mmdrop drops the mm and the page tables */
1536 extern void FASTCALL(__mmdrop(struct mm_struct
*));
1537 static inline void mmdrop(struct mm_struct
* mm
)
1539 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
1543 /* mmput gets rid of the mappings and all user-space */
1544 extern void mmput(struct mm_struct
*);
1545 /* Grab a reference to a task's mm, if it is not already going away */
1546 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1547 /* Remove the current tasks stale references to the old mm_struct */
1548 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1550 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1551 extern void flush_thread(void);
1552 extern void exit_thread(void);
1554 extern void exit_files(struct task_struct
*);
1555 extern void __cleanup_signal(struct signal_struct
*);
1556 extern void __cleanup_sighand(struct sighand_struct
*);
1557 extern void exit_itimers(struct signal_struct
*);
1559 extern NORET_TYPE
void do_group_exit(int);
1561 extern void daemonize(const char *, ...);
1562 extern int allow_signal(int);
1563 extern int disallow_signal(int);
1565 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1566 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1567 struct task_struct
*fork_idle(int);
1569 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1570 extern void get_task_comm(char *to
, struct task_struct
*tsk
);
1573 extern void wait_task_inactive(struct task_struct
* p
);
1575 #define wait_task_inactive(p) do { } while (0)
1578 #define remove_parent(p) list_del_init(&(p)->sibling)
1579 #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1581 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1583 #define for_each_process(p) \
1584 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1587 * Careful: do_each_thread/while_each_thread is a double loop so
1588 * 'break' will not work as expected - use goto instead.
1590 #define do_each_thread(g, t) \
1591 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1593 #define while_each_thread(g, t) \
1594 while ((t = next_thread(t)) != g)
1596 /* de_thread depends on thread_group_leader not being a pid based check */
1597 #define thread_group_leader(p) (p == p->group_leader)
1599 /* Do to the insanities of de_thread it is possible for a process
1600 * to have the pid of the thread group leader without actually being
1601 * the thread group leader. For iteration through the pids in proc
1602 * all we care about is that we have a task with the appropriate
1603 * pid, we don't actually care if we have the right task.
1605 static inline int has_group_leader_pid(struct task_struct
*p
)
1607 return p
->pid
== p
->tgid
;
1610 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
1612 return list_entry(rcu_dereference(p
->thread_group
.next
),
1613 struct task_struct
, thread_group
);
1616 static inline int thread_group_empty(struct task_struct
*p
)
1618 return list_empty(&p
->thread_group
);
1621 #define delay_group_leader(p) \
1622 (thread_group_leader(p) && !thread_group_empty(p))
1625 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
1626 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1627 * pins the final release of task.io_context. Also protects ->cpuset.
1629 * Nests both inside and outside of read_lock(&tasklist_lock).
1630 * It must not be nested with write_lock_irq(&tasklist_lock),
1631 * neither inside nor outside.
1633 static inline void task_lock(struct task_struct
*p
)
1635 spin_lock(&p
->alloc_lock
);
1638 static inline void task_unlock(struct task_struct
*p
)
1640 spin_unlock(&p
->alloc_lock
);
1643 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
1644 unsigned long *flags
);
1646 static inline void unlock_task_sighand(struct task_struct
*tsk
,
1647 unsigned long *flags
)
1649 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
1652 #ifndef __HAVE_THREAD_FUNCTIONS
1654 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
1655 #define task_stack_page(task) ((task)->stack)
1657 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
1659 *task_thread_info(p
) = *task_thread_info(org
);
1660 task_thread_info(p
)->task
= p
;
1663 static inline unsigned long *end_of_stack(struct task_struct
*p
)
1665 return (unsigned long *)(task_thread_info(p
) + 1);
1670 /* set thread flags in other task's structures
1671 * - see asm/thread_info.h for TIF_xxxx flags available
1673 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1675 set_ti_thread_flag(task_thread_info(tsk
), flag
);
1678 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1680 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1683 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1685 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
1688 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1690 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1693 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1695 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
1698 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
1700 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1703 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
1705 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1708 static inline int signal_pending(struct task_struct
*p
)
1710 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
1713 static inline int need_resched(void)
1715 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
1719 * cond_resched() and cond_resched_lock(): latency reduction via
1720 * explicit rescheduling in places that are safe. The return
1721 * value indicates whether a reschedule was done in fact.
1722 * cond_resched_lock() will drop the spinlock before scheduling,
1723 * cond_resched_softirq() will enable bhs before scheduling.
1725 extern int cond_resched(void);
1726 extern int cond_resched_lock(spinlock_t
* lock
);
1727 extern int cond_resched_softirq(void);
1730 * Does a critical section need to be broken due to another
1733 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1734 # define need_lockbreak(lock) ((lock)->break_lock)
1736 # define need_lockbreak(lock) 0
1740 * Does a critical section need to be broken due to another
1741 * task waiting or preemption being signalled:
1743 static inline int lock_need_resched(spinlock_t
*lock
)
1745 if (need_lockbreak(lock
) || need_resched())
1751 * Reevaluate whether the task has signals pending delivery.
1752 * Wake the task if so.
1753 * This is required every time the blocked sigset_t changes.
1754 * callers must hold sighand->siglock.
1756 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
1757 extern void recalc_sigpending(void);
1759 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
1762 * Wrappers for p->thread_info->cpu access. No-op on UP.
1766 static inline unsigned int task_cpu(const struct task_struct
*p
)
1768 return task_thread_info(p
)->cpu
;
1771 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
1775 static inline unsigned int task_cpu(const struct task_struct
*p
)
1780 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1784 #endif /* CONFIG_SMP */
1786 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1787 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
1789 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
)
1791 mm
->mmap_base
= TASK_UNMAPPED_BASE
;
1792 mm
->get_unmapped_area
= arch_get_unmapped_area
;
1793 mm
->unmap_area
= arch_unmap_area
;
1797 extern long sched_setaffinity(pid_t pid
, cpumask_t new_mask
);
1798 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
1800 extern int sched_mc_power_savings
, sched_smt_power_savings
;
1802 extern void normalize_rt_tasks(void);
1804 #ifdef CONFIG_FAIR_GROUP_SCHED
1806 extern struct task_group init_task_group
;
1808 extern struct task_group
*sched_create_group(void);
1809 extern void sched_destroy_group(struct task_group
*tg
);
1810 extern void sched_move_task(struct task_struct
*tsk
);
1811 extern int sched_group_set_shares(struct task_group
*tg
, unsigned long shares
);
1812 extern unsigned long sched_group_shares(struct task_group
*tg
);
1816 #ifdef CONFIG_TASK_XACCT
1817 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1822 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1827 static inline void inc_syscr(struct task_struct
*tsk
)
1832 static inline void inc_syscw(struct task_struct
*tsk
)
1837 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1841 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1845 static inline void inc_syscr(struct task_struct
*tsk
)
1849 static inline void inc_syscw(struct task_struct
*tsk
)
1854 #endif /* __KERNEL__ */