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_NEWPID 0x20000000 /* New pid namespace */
29 #define CLONE_NEWNET 0x40000000 /* New network namespace */
30 #define CLONE_IO 0x80000000 /* Clone io context */
35 #define SCHED_NORMAL 0
39 /* SCHED_ISO: reserved but not implemented yet */
48 #include <asm/param.h> /* for HZ */
50 #include <linux/capability.h>
51 #include <linux/threads.h>
52 #include <linux/kernel.h>
53 #include <linux/types.h>
54 #include <linux/timex.h>
55 #include <linux/jiffies.h>
56 #include <linux/rbtree.h>
57 #include <linux/thread_info.h>
58 #include <linux/cpumask.h>
59 #include <linux/errno.h>
60 #include <linux/nodemask.h>
61 #include <linux/mm_types.h>
63 #include <asm/system.h>
65 #include <asm/ptrace.h>
66 #include <asm/cputime.h>
68 #include <linux/smp.h>
69 #include <linux/sem.h>
70 #include <linux/signal.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/rtmutex.h>
82 #include <linux/time.h>
83 #include <linux/param.h>
84 #include <linux/resource.h>
85 #include <linux/timer.h>
86 #include <linux/hrtimer.h>
87 #include <linux/task_io_accounting.h>
88 #include <linux/kobject.h>
89 #include <linux/latencytop.h>
90 #include <linux/cred.h>
92 #include <asm/processor.h>
96 struct futex_pi_state;
97 struct robust_list_head;
102 * List of flags we want to share for kernel threads,
103 * if only because they are not used by them anyway.
105 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
108 * These are the constant used to fake the fixed-point load-average
109 * counting. Some notes:
110 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
111 * a load-average precision of 10 bits integer + 11 bits fractional
112 * - if you want to count load-averages more often, you need more
113 * precision, or rounding will get you. With 2-second counting freq,
114 * the EXP_n values would be 1981, 2034 and 2043 if still using only
117 extern unsigned long avenrun[]; /* Load averages */
119 #define FSHIFT 11 /* nr of bits of precision */
120 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
121 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
122 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
123 #define EXP_5 2014 /* 1/exp(5sec/5min) */
124 #define EXP_15 2037 /* 1/exp(5sec/15min) */
126 #define CALC_LOAD(load,exp,n) \
128 load += n*(FIXED_1-exp); \
131 extern unsigned long total_forks;
132 extern int nr_threads;
133 DECLARE_PER_CPU(unsigned long, process_counts);
134 extern int nr_processes(void);
135 extern unsigned long nr_running(void);
136 extern unsigned long nr_uninterruptible(void);
137 extern unsigned long nr_active(void);
138 extern unsigned long nr_iowait(void);
140 extern unsigned long get_parent_ip(unsigned long addr);
145 #ifdef CONFIG_SCHED_DEBUG
146 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
147 extern void proc_sched_set_task(struct task_struct *p);
149 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
152 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
155 static inline void proc_sched_set_task(struct task_struct *p)
159 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
164 extern unsigned long long time_sync_thresh;
167 * Task state bitmask. NOTE! These bits are also
168 * encoded in fs/proc/array.c: get_task_state().
170 * We have two separate sets of flags: task->state
171 * is about runnability, while task->exit_state are
172 * about the task exiting. Confusing, but this way
173 * modifying one set can't modify the other one by
176 #define TASK_RUNNING 0
177 #define TASK_INTERRUPTIBLE 1
178 #define TASK_UNINTERRUPTIBLE 2
179 #define __TASK_STOPPED 4
180 #define __TASK_TRACED 8
181 /* in tsk->exit_state */
182 #define EXIT_ZOMBIE 16
184 /* in tsk->state again */
186 #define TASK_WAKEKILL 128
188 /* Convenience macros for the sake of set_task_state */
189 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
190 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
191 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
193 /* Convenience macros for the sake of wake_up */
194 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
195 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
197 /* get_task_state() */
198 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
199 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
202 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
203 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
204 #define task_is_stopped_or_traced(task) \
205 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
206 #define task_contributes_to_load(task) \
207 ((task->state & TASK_UNINTERRUPTIBLE) != 0)
209 #define __set_task_state(tsk, state_value) \
210 do { (tsk)->state = (state_value); } while (0)
211 #define set_task_state(tsk, state_value) \
212 set_mb((tsk)->state, (state_value))
215 * set_current_state() includes a barrier so that the write of current->state
216 * is correctly serialised wrt the caller's subsequent test of whether to
219 * set_current_state(TASK_UNINTERRUPTIBLE);
220 * if (do_i_need_to_sleep())
223 * If the caller does not need such serialisation then use __set_current_state()
225 #define __set_current_state(state_value) \
226 do { current->state = (state_value); } while (0)
227 #define set_current_state(state_value) \
228 set_mb(current->state, (state_value))
230 /* Task command name length */
231 #define TASK_COMM_LEN 16
233 #include <linux/spinlock.h>
236 * This serializes "schedule()" and also protects
237 * the run-queue from deletions/modifications (but
238 * _adding_ to the beginning of the run-queue has
241 extern rwlock_t tasklist_lock;
242 extern spinlock_t mmlist_lock;
246 extern void sched_init(void);
247 extern void sched_init_smp(void);
248 extern asmlinkage void schedule_tail(struct task_struct *prev);
249 extern void init_idle(struct task_struct *idle, int cpu);
250 extern void init_idle_bootup_task(struct task_struct *idle);
252 extern int runqueue_is_locked(void);
253 extern void task_rq_unlock_wait(struct task_struct *p);
255 extern cpumask_var_t nohz_cpu_mask;
256 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
257 extern int select_nohz_load_balancer(int cpu);
259 static inline int select_nohz_load_balancer(int cpu)
266 * Only dump TASK_* tasks. (0 for all tasks)
268 extern void show_state_filter(unsigned long state_filter);
270 static inline void show_state(void)
272 show_state_filter(0);
275 extern void show_regs(struct pt_regs *);
278 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
279 * task), SP is the stack pointer of the first frame that should be shown in the back
280 * trace (or NULL if the entire call-chain of the task should be shown).
282 extern void show_stack(struct task_struct *task, unsigned long *sp);
284 void io_schedule(void);
285 long io_schedule_timeout(long timeout);
287 extern void cpu_init (void);
288 extern void trap_init(void);
289 extern void update_process_times(int user);
290 extern void scheduler_tick(void);
292 extern void sched_show_task(struct task_struct *p);
294 #ifdef CONFIG_DETECT_SOFTLOCKUP
295 extern void softlockup_tick(void);
296 extern void touch_softlockup_watchdog(void);
297 extern void touch_all_softlockup_watchdogs(void);
298 extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
299 struct file *filp, void __user *buffer,
300 size_t *lenp, loff_t *ppos);
301 extern unsigned int softlockup_panic;
302 extern unsigned long sysctl_hung_task_check_count;
303 extern unsigned long sysctl_hung_task_timeout_secs;
304 extern unsigned long sysctl_hung_task_warnings;
305 extern int softlockup_thresh;
307 static inline void softlockup_tick(void)
310 static inline void spawn_softlockup_task(void)
313 static inline void touch_softlockup_watchdog(void)
316 static inline void touch_all_softlockup_watchdogs(void)
322 /* Attach to any functions which should be ignored in wchan output. */
323 #define __sched __attribute__((__section__(".sched.text")))
325 /* Linker adds these: start and end of __sched functions */
326 extern char __sched_text_start[], __sched_text_end[];
328 /* Is this address in the __sched functions? */
329 extern int in_sched_functions(unsigned long addr);
331 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
332 extern signed long schedule_timeout(signed long timeout);
333 extern signed long schedule_timeout_interruptible(signed long timeout);
334 extern signed long schedule_timeout_killable(signed long timeout);
335 extern signed long schedule_timeout_uninterruptible(signed long timeout);
336 asmlinkage void __schedule(void);
337 asmlinkage void schedule(void);
338 extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
341 struct user_namespace;
343 /* Maximum number of active map areas.. This is a random (large) number */
344 #define DEFAULT_MAX_MAP_COUNT 65536
346 extern int sysctl_max_map_count;
348 #include <linux/aio.h>
351 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
352 unsigned long, unsigned long);
354 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
355 unsigned long len, unsigned long pgoff,
356 unsigned long flags);
357 extern void arch_unmap_area(struct mm_struct *, unsigned long);
358 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
360 #if USE_SPLIT_PTLOCKS
362 * The mm counters are not protected by its page_table_lock,
363 * so must be incremented atomically.
365 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
366 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
367 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
368 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
369 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
371 #else /* !USE_SPLIT_PTLOCKS */
373 * The mm counters are protected by its page_table_lock,
374 * so can be incremented directly.
376 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
377 #define get_mm_counter(mm, member) ((mm)->_##member)
378 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
379 #define inc_mm_counter(mm, member) (mm)->_##member++
380 #define dec_mm_counter(mm, member) (mm)->_##member--
382 #endif /* !USE_SPLIT_PTLOCKS */
384 #define get_mm_rss(mm) \
385 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
386 #define update_hiwater_rss(mm) do { \
387 unsigned long _rss = get_mm_rss(mm); \
388 if ((mm)->hiwater_rss < _rss) \
389 (mm)->hiwater_rss = _rss; \
391 #define update_hiwater_vm(mm) do { \
392 if ((mm)->hiwater_vm < (mm)->total_vm) \
393 (mm)->hiwater_vm = (mm)->total_vm; \
396 static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
398 return max(mm->hiwater_rss, get_mm_rss(mm));
401 static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
403 return max(mm->hiwater_vm, mm->total_vm);
406 extern void set_dumpable(struct mm_struct *mm, int value);
407 extern int get_dumpable(struct mm_struct *mm);
411 #define MMF_DUMPABLE 0 /* core dump is permitted */
412 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
413 #define MMF_DUMPABLE_BITS 2
415 /* coredump filter bits */
416 #define MMF_DUMP_ANON_PRIVATE 2
417 #define MMF_DUMP_ANON_SHARED 3
418 #define MMF_DUMP_MAPPED_PRIVATE 4
419 #define MMF_DUMP_MAPPED_SHARED 5
420 #define MMF_DUMP_ELF_HEADERS 6
421 #define MMF_DUMP_HUGETLB_PRIVATE 7
422 #define MMF_DUMP_HUGETLB_SHARED 8
423 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
424 #define MMF_DUMP_FILTER_BITS 7
425 #define MMF_DUMP_FILTER_MASK \
426 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
427 #define MMF_DUMP_FILTER_DEFAULT \
428 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
429 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
431 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
432 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
434 # define MMF_DUMP_MASK_DEFAULT_ELF 0
437 struct sighand_struct {
439 struct k_sigaction action[_NSIG];
441 wait_queue_head_t signalfd_wqh;
444 struct pacct_struct {
447 unsigned long ac_mem;
448 cputime_t ac_utime, ac_stime;
449 unsigned long ac_minflt, ac_majflt;
453 * struct task_cputime - collected CPU time counts
454 * @utime: time spent in user mode, in &cputime_t units
455 * @stime: time spent in kernel mode, in &cputime_t units
456 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
458 * This structure groups together three kinds of CPU time that are
459 * tracked for threads and thread groups. Most things considering
460 * CPU time want to group these counts together and treat all three
461 * of them in parallel.
463 struct task_cputime {
466 unsigned long long sum_exec_runtime;
468 /* Alternate field names when used to cache expirations. */
469 #define prof_exp stime
470 #define virt_exp utime
471 #define sched_exp sum_exec_runtime
473 #define INIT_CPUTIME \
474 (struct task_cputime) { \
475 .utime = cputime_zero, \
476 .stime = cputime_zero, \
477 .sum_exec_runtime = 0, \
481 * struct thread_group_cputimer - thread group interval timer counts
482 * @cputime: thread group interval timers.
483 * @running: non-zero when there are timers running and
484 * @cputime receives updates.
485 * @lock: lock for fields in this struct.
487 * This structure contains the version of task_cputime, above, that is
488 * used for thread group CPU timer calculations.
490 struct thread_group_cputimer {
491 struct task_cputime cputime;
497 * NOTE! "signal_struct" does not have it's own
498 * locking, because a shared signal_struct always
499 * implies a shared sighand_struct, so locking
500 * sighand_struct is always a proper superset of
501 * the locking of signal_struct.
503 struct signal_struct {
507 wait_queue_head_t wait_chldexit; /* for wait4() */
509 /* current thread group signal load-balancing target: */
510 struct task_struct *curr_target;
512 /* shared signal handling: */
513 struct sigpending shared_pending;
515 /* thread group exit support */
518 * - notify group_exit_task when ->count is equal to notify_count
519 * - everyone except group_exit_task is stopped during signal delivery
520 * of fatal signals, group_exit_task processes the signal.
523 struct task_struct *group_exit_task;
525 /* thread group stop support, overloads group_exit_code too */
526 int group_stop_count;
527 unsigned int flags; /* see SIGNAL_* flags below */
529 /* POSIX.1b Interval Timers */
530 struct list_head posix_timers;
532 /* ITIMER_REAL timer for the process */
533 struct hrtimer real_timer;
534 struct pid *leader_pid;
535 ktime_t it_real_incr;
537 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
538 cputime_t it_prof_expires, it_virt_expires;
539 cputime_t it_prof_incr, it_virt_incr;
542 * Thread group totals for process CPU timers.
543 * See thread_group_cputimer(), et al, for details.
545 struct thread_group_cputimer cputimer;
547 /* Earliest-expiration cache. */
548 struct task_cputime cputime_expires;
550 struct list_head cpu_timers[3];
552 /* job control IDs */
555 * pgrp and session fields are deprecated.
556 * use the task_session_Xnr and task_pgrp_Xnr routines below
560 pid_t pgrp __deprecated;
564 struct pid *tty_old_pgrp;
567 pid_t session __deprecated;
571 /* boolean value for session group leader */
574 struct tty_struct *tty; /* NULL if no tty */
577 * Cumulative resource counters for dead threads in the group,
578 * and for reaped dead child processes forked by this group.
579 * Live threads maintain their own counters and add to these
580 * in __exit_signal, except for the group leader.
582 cputime_t utime, stime, cutime, cstime;
585 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
586 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
587 unsigned long inblock, oublock, cinblock, coublock;
588 struct task_io_accounting ioac;
591 * Cumulative ns of schedule CPU time fo dead threads in the
592 * group, not including a zombie group leader, (This only differs
593 * from jiffies_to_ns(utime + stime) if sched_clock uses something
594 * other than jiffies.)
596 unsigned long long sum_sched_runtime;
599 * We don't bother to synchronize most readers of this at all,
600 * because there is no reader checking a limit that actually needs
601 * to get both rlim_cur and rlim_max atomically, and either one
602 * alone is a single word that can safely be read normally.
603 * getrlimit/setrlimit use task_lock(current->group_leader) to
604 * protect this instead of the siglock, because they really
605 * have no need to disable irqs.
607 struct rlimit rlim[RLIM_NLIMITS];
609 #ifdef CONFIG_BSD_PROCESS_ACCT
610 struct pacct_struct pacct; /* per-process accounting information */
612 #ifdef CONFIG_TASKSTATS
613 struct taskstats *stats;
617 struct tty_audit_buf *tty_audit_buf;
621 /* Context switch must be unlocked if interrupts are to be enabled */
622 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
623 # define __ARCH_WANT_UNLOCKED_CTXSW
627 * Bits in flags field of signal_struct.
629 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
630 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
631 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
632 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
634 * Pending notifications to parent.
636 #define SIGNAL_CLD_STOPPED 0x00000010
637 #define SIGNAL_CLD_CONTINUED 0x00000020
638 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
640 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
642 /* If true, all threads except ->group_exit_task have pending SIGKILL */
643 static inline int signal_group_exit(const struct signal_struct *sig)
645 return (sig->flags & SIGNAL_GROUP_EXIT) ||
646 (sig->group_exit_task != NULL);
650 * Some day this will be a full-fledged user tracking system..
653 atomic_t __count; /* reference count */
654 atomic_t processes; /* How many processes does this user have? */
655 atomic_t files; /* How many open files does this user have? */
656 atomic_t sigpending; /* How many pending signals does this user have? */
657 #ifdef CONFIG_INOTIFY_USER
658 atomic_t inotify_watches; /* How many inotify watches does this user have? */
659 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
662 atomic_t epoll_watches; /* The number of file descriptors currently watched */
664 #ifdef CONFIG_POSIX_MQUEUE
665 /* protected by mq_lock */
666 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
668 unsigned long locked_shm; /* How many pages of mlocked shm ? */
671 struct key *uid_keyring; /* UID specific keyring */
672 struct key *session_keyring; /* UID's default session keyring */
675 /* Hash table maintenance information */
676 struct hlist_node uidhash_node;
678 struct user_namespace *user_ns;
680 #ifdef CONFIG_USER_SCHED
681 struct task_group *tg;
684 struct work_struct work;
689 extern int uids_sysfs_init(void);
691 extern struct user_struct *find_user(uid_t);
693 extern struct user_struct root_user;
694 #define INIT_USER (&root_user)
697 struct backing_dev_info;
698 struct reclaim_state;
700 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
702 /* cumulative counters */
703 unsigned long pcount; /* # of times run on this cpu */
704 unsigned long long run_delay; /* time spent waiting on a runqueue */
707 unsigned long long last_arrival,/* when we last ran on a cpu */
708 last_queued; /* when we were last queued to run */
709 #ifdef CONFIG_SCHEDSTATS
711 unsigned int bkl_count;
714 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
716 #ifdef CONFIG_TASK_DELAY_ACCT
717 struct task_delay_info {
719 unsigned int flags; /* Private per-task flags */
721 /* For each stat XXX, add following, aligned appropriately
723 * struct timespec XXX_start, XXX_end;
727 * Atomicity of updates to XXX_delay, XXX_count protected by
728 * single lock above (split into XXX_lock if contention is an issue).
732 * XXX_count is incremented on every XXX operation, the delay
733 * associated with the operation is added to XXX_delay.
734 * XXX_delay contains the accumulated delay time in nanoseconds.
736 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
737 u64 blkio_delay; /* wait for sync block io completion */
738 u64 swapin_delay; /* wait for swapin block io completion */
739 u32 blkio_count; /* total count of the number of sync block */
740 /* io operations performed */
741 u32 swapin_count; /* total count of the number of swapin block */
742 /* io operations performed */
744 struct timespec freepages_start, freepages_end;
745 u64 freepages_delay; /* wait for memory reclaim */
746 u32 freepages_count; /* total count of memory reclaim */
748 #endif /* CONFIG_TASK_DELAY_ACCT */
750 static inline int sched_info_on(void)
752 #ifdef CONFIG_SCHEDSTATS
754 #elif defined(CONFIG_TASK_DELAY_ACCT)
755 extern int delayacct_on;
770 * sched-domains (multiprocessor balancing) declarations:
774 * Increase resolution of nice-level calculations:
776 #define SCHED_LOAD_SHIFT 10
777 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
779 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
782 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
783 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
784 #define SD_BALANCE_EXEC 4 /* Balance on exec */
785 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
786 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
787 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
788 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
789 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
790 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
791 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
792 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
793 #define SD_WAKE_IDLE_FAR 2048 /* Gain latency sacrificing cache hit */
795 enum powersavings_balance_level {
796 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
797 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
798 * first for long running threads
800 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
801 * cpu package for power savings
803 MAX_POWERSAVINGS_BALANCE_LEVELS
806 extern int sched_mc_power_savings, sched_smt_power_savings;
808 static inline int sd_balance_for_mc_power(void)
810 if (sched_smt_power_savings)
811 return SD_POWERSAVINGS_BALANCE;
816 static inline int sd_balance_for_package_power(void)
818 if (sched_mc_power_savings | sched_smt_power_savings)
819 return SD_POWERSAVINGS_BALANCE;
825 * Optimise SD flags for power savings:
826 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
827 * Keep default SD flags if sched_{smt,mc}_power_saving=0
830 static inline int sd_power_saving_flags(void)
832 if (sched_mc_power_savings | sched_smt_power_savings)
833 return SD_BALANCE_NEWIDLE;
839 struct sched_group *next; /* Must be a circular list */
842 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
843 * single CPU. This is read only (except for setup, hotplug CPU).
844 * Note : Never change cpu_power without recompute its reciprocal
846 unsigned int __cpu_power;
848 * reciprocal value of cpu_power to avoid expensive divides
849 * (see include/linux/reciprocal_div.h)
851 u32 reciprocal_cpu_power;
853 unsigned long cpumask[];
856 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
858 return to_cpumask(sg->cpumask);
861 enum sched_domain_level {
871 struct sched_domain_attr {
872 int relax_domain_level;
875 #define SD_ATTR_INIT (struct sched_domain_attr) { \
876 .relax_domain_level = -1, \
879 struct sched_domain {
880 /* These fields must be setup */
881 struct sched_domain *parent; /* top domain must be null terminated */
882 struct sched_domain *child; /* bottom domain must be null terminated */
883 struct sched_group *groups; /* the balancing groups of the domain */
884 unsigned long min_interval; /* Minimum balance interval ms */
885 unsigned long max_interval; /* Maximum balance interval ms */
886 unsigned int busy_factor; /* less balancing by factor if busy */
887 unsigned int imbalance_pct; /* No balance until over watermark */
888 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
889 unsigned int busy_idx;
890 unsigned int idle_idx;
891 unsigned int newidle_idx;
892 unsigned int wake_idx;
893 unsigned int forkexec_idx;
894 int flags; /* See SD_* */
895 enum sched_domain_level level;
897 /* Runtime fields. */
898 unsigned long last_balance; /* init to jiffies. units in jiffies */
899 unsigned int balance_interval; /* initialise to 1. units in ms. */
900 unsigned int nr_balance_failed; /* initialise to 0 */
904 #ifdef CONFIG_SCHEDSTATS
905 /* load_balance() stats */
906 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
907 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
908 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
909 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
910 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
911 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
912 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
913 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
915 /* Active load balancing */
916 unsigned int alb_count;
917 unsigned int alb_failed;
918 unsigned int alb_pushed;
920 /* SD_BALANCE_EXEC stats */
921 unsigned int sbe_count;
922 unsigned int sbe_balanced;
923 unsigned int sbe_pushed;
925 /* SD_BALANCE_FORK stats */
926 unsigned int sbf_count;
927 unsigned int sbf_balanced;
928 unsigned int sbf_pushed;
930 /* try_to_wake_up() stats */
931 unsigned int ttwu_wake_remote;
932 unsigned int ttwu_move_affine;
933 unsigned int ttwu_move_balance;
935 #ifdef CONFIG_SCHED_DEBUG
939 /* span of all CPUs in this domain */
940 unsigned long span[];
943 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
945 return to_cpumask(sd->span);
948 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
949 struct sched_domain_attr *dattr_new);
951 /* Test a flag in parent sched domain */
952 static inline int test_sd_parent(struct sched_domain *sd, int flag)
954 if (sd->parent && (sd->parent->flags & flag))
960 #else /* CONFIG_SMP */
962 struct sched_domain_attr;
965 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
966 struct sched_domain_attr *dattr_new)
969 #endif /* !CONFIG_SMP */
971 struct io_context; /* See blkdev.h */
974 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
975 extern void prefetch_stack(struct task_struct *t);
977 static inline void prefetch_stack(struct task_struct *t) { }
980 struct audit_context; /* See audit.c */
982 struct pipe_inode_info;
983 struct uts_namespace;
989 const struct sched_class *next;
991 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
992 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
993 void (*yield_task) (struct rq *rq);
995 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
997 struct task_struct * (*pick_next_task) (struct rq *rq);
998 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
1001 int (*select_task_rq)(struct task_struct *p, int sync);
1003 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
1004 struct rq *busiest, unsigned long max_load_move,
1005 struct sched_domain *sd, enum cpu_idle_type idle,
1006 int *all_pinned, int *this_best_prio);
1008 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1009 struct rq *busiest, struct sched_domain *sd,
1010 enum cpu_idle_type idle);
1011 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1012 int (*needs_post_schedule) (struct rq *this_rq);
1013 void (*post_schedule) (struct rq *this_rq);
1014 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
1016 void (*set_cpus_allowed)(struct task_struct *p,
1017 const struct cpumask *newmask);
1019 void (*rq_online)(struct rq *rq);
1020 void (*rq_offline)(struct rq *rq);
1023 void (*set_curr_task) (struct rq *rq);
1024 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1025 void (*task_new) (struct rq *rq, struct task_struct *p);
1027 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1029 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1031 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1032 int oldprio, int running);
1034 #ifdef CONFIG_FAIR_GROUP_SCHED
1035 void (*moved_group) (struct task_struct *p);
1039 struct load_weight {
1040 unsigned long weight, inv_weight;
1044 * CFS stats for a schedulable entity (task, task-group etc)
1046 * Current field usage histogram:
1053 struct sched_entity {
1054 struct load_weight load; /* for load-balancing */
1055 struct rb_node run_node;
1056 struct list_head group_node;
1060 u64 sum_exec_runtime;
1062 u64 prev_sum_exec_runtime;
1071 #ifdef CONFIG_SCHEDSTATS
1079 s64 sum_sleep_runtime;
1086 u64 nr_migrations_cold;
1087 u64 nr_failed_migrations_affine;
1088 u64 nr_failed_migrations_running;
1089 u64 nr_failed_migrations_hot;
1090 u64 nr_forced_migrations;
1091 u64 nr_forced2_migrations;
1094 u64 nr_wakeups_sync;
1095 u64 nr_wakeups_migrate;
1096 u64 nr_wakeups_local;
1097 u64 nr_wakeups_remote;
1098 u64 nr_wakeups_affine;
1099 u64 nr_wakeups_affine_attempts;
1100 u64 nr_wakeups_passive;
1101 u64 nr_wakeups_idle;
1104 #ifdef CONFIG_FAIR_GROUP_SCHED
1105 struct sched_entity *parent;
1106 /* rq on which this entity is (to be) queued: */
1107 struct cfs_rq *cfs_rq;
1108 /* rq "owned" by this entity/group: */
1109 struct cfs_rq *my_q;
1113 struct sched_rt_entity {
1114 struct list_head run_list;
1115 unsigned long timeout;
1116 unsigned int time_slice;
1117 int nr_cpus_allowed;
1119 struct sched_rt_entity *back;
1120 #ifdef CONFIG_RT_GROUP_SCHED
1121 struct sched_rt_entity *parent;
1122 /* rq on which this entity is (to be) queued: */
1123 struct rt_rq *rt_rq;
1124 /* rq "owned" by this entity/group: */
1129 struct task_struct {
1130 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
1133 unsigned int flags; /* per process flags, defined below */
1134 unsigned int ptrace;
1136 int lock_depth; /* BKL lock depth */
1139 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1144 int prio, static_prio, normal_prio;
1145 unsigned int rt_priority;
1146 const struct sched_class *sched_class;
1147 struct sched_entity se;
1148 struct sched_rt_entity rt;
1150 #ifdef CONFIG_PREEMPT_NOTIFIERS
1151 /* list of struct preempt_notifier: */
1152 struct hlist_head preempt_notifiers;
1156 * fpu_counter contains the number of consecutive context switches
1157 * that the FPU is used. If this is over a threshold, the lazy fpu
1158 * saving becomes unlazy to save the trap. This is an unsigned char
1159 * so that after 256 times the counter wraps and the behavior turns
1160 * lazy again; this to deal with bursty apps that only use FPU for
1163 unsigned char fpu_counter;
1164 s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1165 #ifdef CONFIG_BLK_DEV_IO_TRACE
1166 unsigned int btrace_seq;
1169 unsigned int policy;
1170 cpumask_t cpus_allowed;
1172 #ifdef CONFIG_PREEMPT_RCU
1173 int rcu_read_lock_nesting;
1174 int rcu_flipctr_idx;
1175 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1177 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1178 struct sched_info sched_info;
1181 struct list_head tasks;
1182 struct plist_node pushable_tasks;
1184 struct mm_struct *mm, *active_mm;
1187 struct linux_binfmt *binfmt;
1189 int exit_code, exit_signal;
1190 int pdeath_signal; /* The signal sent when the parent dies */
1192 unsigned int personality;
1193 unsigned did_exec:1;
1194 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1199 /* Canary value for the -fstack-protector gcc feature */
1200 unsigned long stack_canary;
1203 * pointers to (original) parent process, youngest child, younger sibling,
1204 * older sibling, respectively. (p->father can be replaced with
1205 * p->real_parent->pid)
1207 struct task_struct *real_parent; /* real parent process */
1208 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1210 * children/sibling forms the list of my natural children
1212 struct list_head children; /* list of my children */
1213 struct list_head sibling; /* linkage in my parent's children list */
1214 struct task_struct *group_leader; /* threadgroup leader */
1217 * ptraced is the list of tasks this task is using ptrace on.
1218 * This includes both natural children and PTRACE_ATTACH targets.
1219 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1221 struct list_head ptraced;
1222 struct list_head ptrace_entry;
1224 #ifdef CONFIG_X86_PTRACE_BTS
1226 * This is the tracer handle for the ptrace BTS extension.
1227 * This field actually belongs to the ptracer task.
1229 struct bts_tracer *bts;
1231 * The buffer to hold the BTS data.
1235 #endif /* CONFIG_X86_PTRACE_BTS */
1237 /* PID/PID hash table linkage. */
1238 struct pid_link pids[PIDTYPE_MAX];
1239 struct list_head thread_group;
1241 struct completion *vfork_done; /* for vfork() */
1242 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1243 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1245 cputime_t utime, stime, utimescaled, stimescaled;
1247 cputime_t prev_utime, prev_stime;
1248 unsigned long nvcsw, nivcsw; /* context switch counts */
1249 struct timespec start_time; /* monotonic time */
1250 struct timespec real_start_time; /* boot based time */
1251 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1252 unsigned long min_flt, maj_flt;
1254 struct task_cputime cputime_expires;
1255 struct list_head cpu_timers[3];
1257 /* process credentials */
1258 const struct cred *real_cred; /* objective and real subjective task
1259 * credentials (COW) */
1260 const struct cred *cred; /* effective (overridable) subjective task
1261 * credentials (COW) */
1262 struct mutex cred_exec_mutex; /* execve vs ptrace cred calculation mutex */
1264 char comm[TASK_COMM_LEN]; /* executable name excluding path
1265 - access with [gs]et_task_comm (which lock
1266 it with task_lock())
1267 - initialized normally by flush_old_exec */
1268 /* file system info */
1269 int link_count, total_link_count;
1270 #ifdef CONFIG_SYSVIPC
1272 struct sysv_sem sysvsem;
1274 #ifdef CONFIG_DETECT_SOFTLOCKUP
1275 /* hung task detection */
1276 unsigned long last_switch_timestamp;
1277 unsigned long last_switch_count;
1279 /* CPU-specific state of this task */
1280 struct thread_struct thread;
1281 /* filesystem information */
1282 struct fs_struct *fs;
1283 /* open file information */
1284 struct files_struct *files;
1286 struct nsproxy *nsproxy;
1287 /* signal handlers */
1288 struct signal_struct *signal;
1289 struct sighand_struct *sighand;
1291 sigset_t blocked, real_blocked;
1292 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1293 struct sigpending pending;
1295 unsigned long sas_ss_sp;
1297 int (*notifier)(void *priv);
1298 void *notifier_data;
1299 sigset_t *notifier_mask;
1300 struct audit_context *audit_context;
1301 #ifdef CONFIG_AUDITSYSCALL
1303 unsigned int sessionid;
1307 /* Thread group tracking */
1310 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1311 spinlock_t alloc_lock;
1313 /* Protection of the PI data structures: */
1316 #ifdef CONFIG_RT_MUTEXES
1317 /* PI waiters blocked on a rt_mutex held by this task */
1318 struct plist_head pi_waiters;
1319 /* Deadlock detection and priority inheritance handling */
1320 struct rt_mutex_waiter *pi_blocked_on;
1323 #ifdef CONFIG_DEBUG_MUTEXES
1324 /* mutex deadlock detection */
1325 struct mutex_waiter *blocked_on;
1327 #ifdef CONFIG_TRACE_IRQFLAGS
1328 unsigned int irq_events;
1329 int hardirqs_enabled;
1330 unsigned long hardirq_enable_ip;
1331 unsigned int hardirq_enable_event;
1332 unsigned long hardirq_disable_ip;
1333 unsigned int hardirq_disable_event;
1334 int softirqs_enabled;
1335 unsigned long softirq_disable_ip;
1336 unsigned int softirq_disable_event;
1337 unsigned long softirq_enable_ip;
1338 unsigned int softirq_enable_event;
1339 int hardirq_context;
1340 int softirq_context;
1342 #ifdef CONFIG_LOCKDEP
1343 # define MAX_LOCK_DEPTH 48UL
1346 unsigned int lockdep_recursion;
1347 struct held_lock held_locks[MAX_LOCK_DEPTH];
1348 gfp_t lockdep_reclaim_gfp;
1351 /* journalling filesystem info */
1354 /* stacked block device info */
1355 struct bio *bio_list, **bio_tail;
1358 struct reclaim_state *reclaim_state;
1360 struct backing_dev_info *backing_dev_info;
1362 struct io_context *io_context;
1364 unsigned long ptrace_message;
1365 siginfo_t *last_siginfo; /* For ptrace use. */
1366 struct task_io_accounting ioac;
1367 #if defined(CONFIG_TASK_XACCT)
1368 u64 acct_rss_mem1; /* accumulated rss usage */
1369 u64 acct_vm_mem1; /* accumulated virtual memory usage */
1370 cputime_t acct_timexpd; /* stime + utime since last update */
1372 #ifdef CONFIG_CPUSETS
1373 nodemask_t mems_allowed;
1374 int cpuset_mems_generation;
1375 int cpuset_mem_spread_rotor;
1377 #ifdef CONFIG_CGROUPS
1378 /* Control Group info protected by css_set_lock */
1379 struct css_set *cgroups;
1380 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1381 struct list_head cg_list;
1384 struct robust_list_head __user *robust_list;
1385 #ifdef CONFIG_COMPAT
1386 struct compat_robust_list_head __user *compat_robust_list;
1388 struct list_head pi_state_list;
1389 struct futex_pi_state *pi_state_cache;
1392 struct mempolicy *mempolicy;
1395 atomic_t fs_excl; /* holding fs exclusive resources */
1396 struct rcu_head rcu;
1399 * cache last used pipe for splice
1401 struct pipe_inode_info *splice_pipe;
1402 #ifdef CONFIG_TASK_DELAY_ACCT
1403 struct task_delay_info *delays;
1405 #ifdef CONFIG_FAULT_INJECTION
1408 struct prop_local_single dirties;
1409 #ifdef CONFIG_LATENCYTOP
1410 int latency_record_count;
1411 struct latency_record latency_record[LT_SAVECOUNT];
1414 * time slack values; these are used to round up poll() and
1415 * select() etc timeout values. These are in nanoseconds.
1417 unsigned long timer_slack_ns;
1418 unsigned long default_timer_slack_ns;
1420 struct list_head *scm_work_list;
1421 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1422 /* Index of current stored adress in ret_stack */
1424 /* Stack of return addresses for return function tracing */
1425 struct ftrace_ret_stack *ret_stack;
1426 /* time stamp for last schedule */
1427 unsigned long long ftrace_timestamp;
1429 * Number of functions that haven't been traced
1430 * because of depth overrun.
1432 atomic_t trace_overrun;
1433 /* Pause for the tracing */
1434 atomic_t tracing_graph_pause;
1436 #ifdef CONFIG_TRACING
1437 /* state flags for use by tracers */
1438 unsigned long trace;
1442 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1443 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1446 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1447 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1448 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1449 * values are inverted: lower p->prio value means higher priority.
1451 * The MAX_USER_RT_PRIO value allows the actual maximum
1452 * RT priority to be separate from the value exported to
1453 * user-space. This allows kernel threads to set their
1454 * priority to a value higher than any user task. Note:
1455 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1458 #define MAX_USER_RT_PRIO 100
1459 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1461 #define MAX_PRIO (MAX_RT_PRIO + 40)
1462 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1464 static inline int rt_prio(int prio)
1466 if (unlikely(prio < MAX_RT_PRIO))
1471 static inline int rt_task(struct task_struct *p)
1473 return rt_prio(p->prio);
1476 static inline void set_task_session(struct task_struct *tsk, pid_t session)
1478 tsk->signal->__session = session;
1481 static inline void set_task_pgrp(struct task_struct *tsk, pid_t pgrp)
1483 tsk->signal->__pgrp = pgrp;
1486 static inline struct pid *task_pid(struct task_struct *task)
1488 return task->pids[PIDTYPE_PID].pid;
1491 static inline struct pid *task_tgid(struct task_struct *task)
1493 return task->group_leader->pids[PIDTYPE_PID].pid;
1496 static inline struct pid *task_pgrp(struct task_struct *task)
1498 return task->group_leader->pids[PIDTYPE_PGID].pid;
1501 static inline struct pid *task_session(struct task_struct *task)
1503 return task->group_leader->pids[PIDTYPE_SID].pid;
1506 struct pid_namespace;
1509 * the helpers to get the task's different pids as they are seen
1510 * from various namespaces
1512 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1513 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1515 * task_xid_nr_ns() : id seen from the ns specified;
1517 * set_task_vxid() : assigns a virtual id to a task;
1519 * see also pid_nr() etc in include/linux/pid.h
1522 static inline pid_t task_pid_nr(struct task_struct *tsk)
1527 pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1529 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1531 return pid_vnr(task_pid(tsk));
1535 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1540 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1542 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1544 return pid_vnr(task_tgid(tsk));
1548 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1550 return tsk->signal->__pgrp;
1553 pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1555 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1557 return pid_vnr(task_pgrp(tsk));
1561 static inline pid_t task_session_nr(struct task_struct *tsk)
1563 return tsk->signal->__session;
1566 pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1568 static inline pid_t task_session_vnr(struct task_struct *tsk)
1570 return pid_vnr(task_session(tsk));
1575 * pid_alive - check that a task structure is not stale
1576 * @p: Task structure to be checked.
1578 * Test if a process is not yet dead (at most zombie state)
1579 * If pid_alive fails, then pointers within the task structure
1580 * can be stale and must not be dereferenced.
1582 static inline int pid_alive(struct task_struct *p)
1584 return p->pids[PIDTYPE_PID].pid != NULL;
1588 * is_global_init - check if a task structure is init
1589 * @tsk: Task structure to be checked.
1591 * Check if a task structure is the first user space task the kernel created.
1593 static inline int is_global_init(struct task_struct *tsk)
1595 return tsk->pid == 1;
1599 * is_container_init:
1600 * check whether in the task is init in its own pid namespace.
1602 extern int is_container_init(struct task_struct *tsk);
1604 extern struct pid *cad_pid;
1606 extern void free_task(struct task_struct *tsk);
1607 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1609 extern void __put_task_struct(struct task_struct *t);
1611 static inline void put_task_struct(struct task_struct *t)
1613 if (atomic_dec_and_test(&t->usage))
1614 __put_task_struct(t);
1617 extern cputime_t task_utime(struct task_struct *p);
1618 extern cputime_t task_stime(struct task_struct *p);
1619 extern cputime_t task_gtime(struct task_struct *p);
1624 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1625 /* Not implemented yet, only for 486*/
1626 #define PF_STARTING 0x00000002 /* being created */
1627 #define PF_EXITING 0x00000004 /* getting shut down */
1628 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1629 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1630 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1631 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1632 #define PF_DUMPCORE 0x00000200 /* dumped core */
1633 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1634 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1635 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1636 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1637 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1638 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1639 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1640 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1641 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1642 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1643 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1644 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1645 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1646 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1647 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1648 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1649 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1650 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1651 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1652 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1655 * Only the _current_ task can read/write to tsk->flags, but other
1656 * tasks can access tsk->flags in readonly mode for example
1657 * with tsk_used_math (like during threaded core dumping).
1658 * There is however an exception to this rule during ptrace
1659 * or during fork: the ptracer task is allowed to write to the
1660 * child->flags of its traced child (same goes for fork, the parent
1661 * can write to the child->flags), because we're guaranteed the
1662 * child is not running and in turn not changing child->flags
1663 * at the same time the parent does it.
1665 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1666 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1667 #define clear_used_math() clear_stopped_child_used_math(current)
1668 #define set_used_math() set_stopped_child_used_math(current)
1669 #define conditional_stopped_child_used_math(condition, child) \
1670 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1671 #define conditional_used_math(condition) \
1672 conditional_stopped_child_used_math(condition, current)
1673 #define copy_to_stopped_child_used_math(child) \
1674 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1675 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1676 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1677 #define used_math() tsk_used_math(current)
1680 extern int set_cpus_allowed_ptr(struct task_struct *p,
1681 const struct cpumask *new_mask);
1683 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1684 const struct cpumask *new_mask)
1686 if (!cpumask_test_cpu(0, new_mask))
1691 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1693 return set_cpus_allowed_ptr(p, &new_mask);
1697 * Architectures can set this to 1 if they have specified
1698 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1699 * but then during bootup it turns out that sched_clock()
1700 * is reliable after all:
1702 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1703 extern int sched_clock_stable;
1706 extern unsigned long long sched_clock(void);
1708 extern void sched_clock_init(void);
1709 extern u64 sched_clock_cpu(int cpu);
1711 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1712 static inline void sched_clock_tick(void)
1716 static inline void sched_clock_idle_sleep_event(void)
1720 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1724 extern void sched_clock_tick(void);
1725 extern void sched_clock_idle_sleep_event(void);
1726 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1730 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1731 * clock constructed from sched_clock():
1733 extern unsigned long long cpu_clock(int cpu);
1735 extern unsigned long long
1736 task_sched_runtime(struct task_struct *task);
1737 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1739 /* sched_exec is called by processes performing an exec */
1741 extern void sched_exec(void);
1743 #define sched_exec() {}
1746 extern void sched_clock_idle_sleep_event(void);
1747 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1749 #ifdef CONFIG_HOTPLUG_CPU
1750 extern void idle_task_exit(void);
1752 static inline void idle_task_exit(void) {}
1755 extern void sched_idle_next(void);
1757 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1758 extern void wake_up_idle_cpu(int cpu);
1760 static inline void wake_up_idle_cpu(int cpu) { }
1763 extern unsigned int sysctl_sched_latency;
1764 extern unsigned int sysctl_sched_min_granularity;
1765 extern unsigned int sysctl_sched_wakeup_granularity;
1766 extern unsigned int sysctl_sched_shares_ratelimit;
1767 extern unsigned int sysctl_sched_shares_thresh;
1768 #ifdef CONFIG_SCHED_DEBUG
1769 extern unsigned int sysctl_sched_child_runs_first;
1770 extern unsigned int sysctl_sched_features;
1771 extern unsigned int sysctl_sched_migration_cost;
1772 extern unsigned int sysctl_sched_nr_migrate;
1774 int sched_nr_latency_handler(struct ctl_table *table, int write,
1775 struct file *file, void __user *buffer, size_t *length,
1778 extern unsigned int sysctl_sched_rt_period;
1779 extern int sysctl_sched_rt_runtime;
1781 int sched_rt_handler(struct ctl_table *table, int write,
1782 struct file *filp, void __user *buffer, size_t *lenp,
1785 extern unsigned int sysctl_sched_compat_yield;
1787 #ifdef CONFIG_RT_MUTEXES
1788 extern int rt_mutex_getprio(struct task_struct *p);
1789 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1790 extern void rt_mutex_adjust_pi(struct task_struct *p);
1792 static inline int rt_mutex_getprio(struct task_struct *p)
1794 return p->normal_prio;
1796 # define rt_mutex_adjust_pi(p) do { } while (0)
1799 extern void set_user_nice(struct task_struct *p, long nice);
1800 extern int task_prio(const struct task_struct *p);
1801 extern int task_nice(const struct task_struct *p);
1802 extern int can_nice(const struct task_struct *p, const int nice);
1803 extern int task_curr(const struct task_struct *p);
1804 extern int idle_cpu(int cpu);
1805 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1806 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1807 struct sched_param *);
1808 extern struct task_struct *idle_task(int cpu);
1809 extern struct task_struct *curr_task(int cpu);
1810 extern void set_curr_task(int cpu, struct task_struct *p);
1815 * The default (Linux) execution domain.
1817 extern struct exec_domain default_exec_domain;
1819 union thread_union {
1820 struct thread_info thread_info;
1821 unsigned long stack[THREAD_SIZE/sizeof(long)];
1824 #ifndef __HAVE_ARCH_KSTACK_END
1825 static inline int kstack_end(void *addr)
1827 /* Reliable end of stack detection:
1828 * Some APM bios versions misalign the stack
1830 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1834 extern union thread_union init_thread_union;
1835 extern struct task_struct init_task;
1837 extern struct mm_struct init_mm;
1839 extern struct pid_namespace init_pid_ns;
1842 * find a task by one of its numerical ids
1844 * find_task_by_pid_type_ns():
1845 * it is the most generic call - it finds a task by all id,
1846 * type and namespace specified
1847 * find_task_by_pid_ns():
1848 * finds a task by its pid in the specified namespace
1849 * find_task_by_vpid():
1850 * finds a task by its virtual pid
1852 * see also find_vpid() etc in include/linux/pid.h
1855 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1856 struct pid_namespace *ns);
1858 extern struct task_struct *find_task_by_vpid(pid_t nr);
1859 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1860 struct pid_namespace *ns);
1862 extern void __set_special_pids(struct pid *pid);
1864 /* per-UID process charging. */
1865 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1866 static inline struct user_struct *get_uid(struct user_struct *u)
1868 atomic_inc(&u->__count);
1871 extern void free_uid(struct user_struct *);
1872 extern void release_uids(struct user_namespace *ns);
1874 #include <asm/current.h>
1876 extern void do_timer(unsigned long ticks);
1878 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1879 extern int wake_up_process(struct task_struct *tsk);
1880 extern void wake_up_new_task(struct task_struct *tsk,
1881 unsigned long clone_flags);
1883 extern void kick_process(struct task_struct *tsk);
1885 static inline void kick_process(struct task_struct *tsk) { }
1887 extern void sched_fork(struct task_struct *p, int clone_flags);
1888 extern void sched_dead(struct task_struct *p);
1890 extern void proc_caches_init(void);
1891 extern void flush_signals(struct task_struct *);
1892 extern void ignore_signals(struct task_struct *);
1893 extern void flush_signal_handlers(struct task_struct *, int force_default);
1894 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1896 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1898 unsigned long flags;
1901 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1902 ret = dequeue_signal(tsk, mask, info);
1903 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1908 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1910 extern void unblock_all_signals(void);
1911 extern void release_task(struct task_struct * p);
1912 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1913 extern int force_sigsegv(int, struct task_struct *);
1914 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1915 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1916 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1917 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1918 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1919 extern int kill_pid(struct pid *pid, int sig, int priv);
1920 extern int kill_proc_info(int, struct siginfo *, pid_t);
1921 extern int do_notify_parent(struct task_struct *, int);
1922 extern void force_sig(int, struct task_struct *);
1923 extern void force_sig_specific(int, struct task_struct *);
1924 extern int send_sig(int, struct task_struct *, int);
1925 extern void zap_other_threads(struct task_struct *p);
1926 extern struct sigqueue *sigqueue_alloc(void);
1927 extern void sigqueue_free(struct sigqueue *);
1928 extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
1929 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1930 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1932 static inline int kill_cad_pid(int sig, int priv)
1934 return kill_pid(cad_pid, sig, priv);
1937 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1938 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1939 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1940 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1942 static inline int is_si_special(const struct siginfo *info)
1944 return info <= SEND_SIG_FORCED;
1947 /* True if we are on the alternate signal stack. */
1949 static inline int on_sig_stack(unsigned long sp)
1951 return (sp - current->sas_ss_sp < current->sas_ss_size);
1954 static inline int sas_ss_flags(unsigned long sp)
1956 return (current->sas_ss_size == 0 ? SS_DISABLE
1957 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1961 * Routines for handling mm_structs
1963 extern struct mm_struct * mm_alloc(void);
1965 /* mmdrop drops the mm and the page tables */
1966 extern void __mmdrop(struct mm_struct *);
1967 static inline void mmdrop(struct mm_struct * mm)
1969 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1973 /* mmput gets rid of the mappings and all user-space */
1974 extern void mmput(struct mm_struct *);
1975 /* Grab a reference to a task's mm, if it is not already going away */
1976 extern struct mm_struct *get_task_mm(struct task_struct *task);
1977 /* Remove the current tasks stale references to the old mm_struct */
1978 extern void mm_release(struct task_struct *, struct mm_struct *);
1979 /* Allocate a new mm structure and copy contents from tsk->mm */
1980 extern struct mm_struct *dup_mm(struct task_struct *tsk);
1982 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1983 extern void flush_thread(void);
1984 extern void exit_thread(void);
1986 extern void exit_files(struct task_struct *);
1987 extern void __cleanup_signal(struct signal_struct *);
1988 extern void __cleanup_sighand(struct sighand_struct *);
1990 extern void exit_itimers(struct signal_struct *);
1991 extern void flush_itimer_signals(void);
1993 extern NORET_TYPE void do_group_exit(int);
1995 extern void daemonize(const char *, ...);
1996 extern int allow_signal(int);
1997 extern int disallow_signal(int);
1999 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2000 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
2001 struct task_struct *fork_idle(int);
2003 extern void set_task_comm(struct task_struct *tsk, char *from);
2004 extern char *get_task_comm(char *to, struct task_struct *tsk);
2007 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
2009 static inline unsigned long wait_task_inactive(struct task_struct *p,
2016 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
2018 #define for_each_process(p) \
2019 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2021 extern bool is_single_threaded(struct task_struct *);
2024 * Careful: do_each_thread/while_each_thread is a double loop so
2025 * 'break' will not work as expected - use goto instead.
2027 #define do_each_thread(g, t) \
2028 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2030 #define while_each_thread(g, t) \
2031 while ((t = next_thread(t)) != g)
2033 /* de_thread depends on thread_group_leader not being a pid based check */
2034 #define thread_group_leader(p) (p == p->group_leader)
2036 /* Do to the insanities of de_thread it is possible for a process
2037 * to have the pid of the thread group leader without actually being
2038 * the thread group leader. For iteration through the pids in proc
2039 * all we care about is that we have a task with the appropriate
2040 * pid, we don't actually care if we have the right task.
2042 static inline int has_group_leader_pid(struct task_struct *p)
2044 return p->pid == p->tgid;
2048 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2050 return p1->tgid == p2->tgid;
2053 static inline struct task_struct *next_thread(const struct task_struct *p)
2055 return list_entry(rcu_dereference(p->thread_group.next),
2056 struct task_struct, thread_group);
2059 static inline int thread_group_empty(struct task_struct *p)
2061 return list_empty(&p->thread_group);
2064 #define delay_group_leader(p) \
2065 (thread_group_leader(p) && !thread_group_empty(p))
2068 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2069 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2070 * pins the final release of task.io_context. Also protects ->cpuset and
2071 * ->cgroup.subsys[].
2073 * Nests both inside and outside of read_lock(&tasklist_lock).
2074 * It must not be nested with write_lock_irq(&tasklist_lock),
2075 * neither inside nor outside.
2077 static inline void task_lock(struct task_struct *p)
2079 spin_lock(&p->alloc_lock);
2082 static inline void task_unlock(struct task_struct *p)
2084 spin_unlock(&p->alloc_lock);
2087 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2088 unsigned long *flags);
2090 static inline void unlock_task_sighand(struct task_struct *tsk,
2091 unsigned long *flags)
2093 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2096 #ifndef __HAVE_THREAD_FUNCTIONS
2098 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2099 #define task_stack_page(task) ((task)->stack)
2101 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2103 *task_thread_info(p) = *task_thread_info(org);
2104 task_thread_info(p)->task = p;
2107 static inline unsigned long *end_of_stack(struct task_struct *p)
2109 return (unsigned long *)(task_thread_info(p) + 1);
2114 static inline int object_is_on_stack(void *obj)
2116 void *stack = task_stack_page(current);
2118 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2121 extern void thread_info_cache_init(void);
2123 #ifdef CONFIG_DEBUG_STACK_USAGE
2124 static inline unsigned long stack_not_used(struct task_struct *p)
2126 unsigned long *n = end_of_stack(p);
2128 do { /* Skip over canary */
2132 return (unsigned long)n - (unsigned long)end_of_stack(p);
2136 /* set thread flags in other task's structures
2137 * - see asm/thread_info.h for TIF_xxxx flags available
2139 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2141 set_ti_thread_flag(task_thread_info(tsk), flag);
2144 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2146 clear_ti_thread_flag(task_thread_info(tsk), flag);
2149 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2151 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2154 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2156 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2159 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2161 return test_ti_thread_flag(task_thread_info(tsk), flag);
2164 static inline void set_tsk_need_resched(struct task_struct *tsk)
2166 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2169 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2171 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2174 static inline int test_tsk_need_resched(struct task_struct *tsk)
2176 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2179 static inline int signal_pending(struct task_struct *p)
2181 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2184 extern int __fatal_signal_pending(struct task_struct *p);
2186 static inline int fatal_signal_pending(struct task_struct *p)
2188 return signal_pending(p) && __fatal_signal_pending(p);
2191 static inline int signal_pending_state(long state, struct task_struct *p)
2193 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2195 if (!signal_pending(p))
2198 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2201 static inline int need_resched(void)
2203 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2207 * cond_resched() and cond_resched_lock(): latency reduction via
2208 * explicit rescheduling in places that are safe. The return
2209 * value indicates whether a reschedule was done in fact.
2210 * cond_resched_lock() will drop the spinlock before scheduling,
2211 * cond_resched_softirq() will enable bhs before scheduling.
2213 extern int _cond_resched(void);
2214 #ifdef CONFIG_PREEMPT_BKL
2215 static inline int cond_resched(void)
2220 static inline int cond_resched(void)
2222 return _cond_resched();
2225 extern int cond_resched_lock(spinlock_t * lock);
2226 extern int cond_resched_softirq(void);
2227 static inline int cond_resched_bkl(void)
2229 return _cond_resched();
2233 * Does a critical section need to be broken due to another
2234 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2235 * but a general need for low latency)
2237 static inline int spin_needbreak(spinlock_t *lock)
2239 #ifdef CONFIG_PREEMPT
2240 return spin_is_contended(lock);
2247 * Thread group CPU time accounting.
2249 void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2250 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2252 static inline void thread_group_cputime_init(struct signal_struct *sig)
2254 sig->cputimer.cputime = INIT_CPUTIME;
2255 spin_lock_init(&sig->cputimer.lock);
2256 sig->cputimer.running = 0;
2259 static inline void thread_group_cputime_free(struct signal_struct *sig)
2264 * Reevaluate whether the task has signals pending delivery.
2265 * Wake the task if so.
2266 * This is required every time the blocked sigset_t changes.
2267 * callers must hold sighand->siglock.
2269 extern void recalc_sigpending_and_wake(struct task_struct *t);
2270 extern void recalc_sigpending(void);
2272 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2275 * Wrappers for p->thread_info->cpu access. No-op on UP.
2279 static inline unsigned int task_cpu(const struct task_struct *p)
2281 return task_thread_info(p)->cpu;
2284 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2288 static inline unsigned int task_cpu(const struct task_struct *p)
2293 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2297 #endif /* CONFIG_SMP */
2299 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2301 #ifdef CONFIG_TRACING
2303 __trace_special(void *__tr, void *__data,
2304 unsigned long arg1, unsigned long arg2, unsigned long arg3);
2307 __trace_special(void *__tr, void *__data,
2308 unsigned long arg1, unsigned long arg2, unsigned long arg3)
2313 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2314 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2316 extern void normalize_rt_tasks(void);
2318 #ifdef CONFIG_GROUP_SCHED
2320 extern struct task_group init_task_group;
2321 #ifdef CONFIG_USER_SCHED
2322 extern struct task_group root_task_group;
2323 extern void set_tg_uid(struct user_struct *user);
2326 extern struct task_group *sched_create_group(struct task_group *parent);
2327 extern void sched_destroy_group(struct task_group *tg);
2328 extern void sched_move_task(struct task_struct *tsk);
2329 #ifdef CONFIG_FAIR_GROUP_SCHED
2330 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2331 extern unsigned long sched_group_shares(struct task_group *tg);
2333 #ifdef CONFIG_RT_GROUP_SCHED
2334 extern int sched_group_set_rt_runtime(struct task_group *tg,
2335 long rt_runtime_us);
2336 extern long sched_group_rt_runtime(struct task_group *tg);
2337 extern int sched_group_set_rt_period(struct task_group *tg,
2339 extern long sched_group_rt_period(struct task_group *tg);
2340 extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
2344 extern int task_can_switch_user(struct user_struct *up,
2345 struct task_struct *tsk);
2347 #ifdef CONFIG_TASK_XACCT
2348 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2350 tsk->ioac.rchar += amt;
2353 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2355 tsk->ioac.wchar += amt;
2358 static inline void inc_syscr(struct task_struct *tsk)
2363 static inline void inc_syscw(struct task_struct *tsk)
2368 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2372 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2376 static inline void inc_syscr(struct task_struct *tsk)
2380 static inline void inc_syscw(struct task_struct *tsk)
2385 #ifndef TASK_SIZE_OF
2386 #define TASK_SIZE_OF(tsk) TASK_SIZE
2389 #ifdef CONFIG_MM_OWNER
2390 extern void mm_update_next_owner(struct mm_struct *mm);
2391 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2393 static inline void mm_update_next_owner(struct mm_struct *mm)
2397 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2400 #endif /* CONFIG_MM_OWNER */
2402 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2404 #endif /* __KERNEL__ */