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/path.h>
72 #include <linux/compiler.h>
73 #include <linux/completion.h>
74 #include <linux/perf_counter.h>
75 #include <linux/pid.h>
76 #include <linux/percpu.h>
77 #include <linux/topology.h>
78 #include <linux/proportions.h>
79 #include <linux/seccomp.h>
80 #include <linux/rcupdate.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>
90 #include <linux/latencytop.h>
91 #include <linux/cred.h>
93 #include <asm/processor.h>
97 struct futex_pi_state;
98 struct robust_list_head;
104 * List of flags we want to share for kernel threads,
105 * if only because they are not used by them anyway.
107 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
110 * These are the constant used to fake the fixed-point load-average
111 * counting. Some notes:
112 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
113 * a load-average precision of 10 bits integer + 11 bits fractional
114 * - if you want to count load-averages more often, you need more
115 * precision, or rounding will get you. With 2-second counting freq,
116 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 extern unsigned long avenrun[]; /* Load averages */
121 #define FSHIFT 11 /* nr of bits of precision */
122 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
123 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
124 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
125 #define EXP_5 2014 /* 1/exp(5sec/5min) */
126 #define EXP_15 2037 /* 1/exp(5sec/15min) */
128 #define CALC_LOAD(load,exp,n) \
130 load += n*(FIXED_1-exp); \
133 extern unsigned long total_forks;
134 extern int nr_threads;
135 DECLARE_PER_CPU(unsigned long, process_counts);
136 extern int nr_processes(void);
137 extern unsigned long nr_running(void);
138 extern unsigned long nr_uninterruptible(void);
139 extern unsigned long nr_active(void);
140 extern unsigned long nr_iowait(void);
141 extern u64 cpu_nr_switches(int cpu);
142 extern u64 cpu_nr_migrations(int cpu);
144 extern unsigned long get_parent_ip(unsigned long addr);
149 #ifdef CONFIG_SCHED_DEBUG
150 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
151 extern void proc_sched_set_task(struct task_struct *p);
153 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
156 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
159 static inline void proc_sched_set_task(struct task_struct *p)
163 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
168 extern unsigned long long time_sync_thresh;
171 * Task state bitmask. NOTE! These bits are also
172 * encoded in fs/proc/array.c: get_task_state().
174 * We have two separate sets of flags: task->state
175 * is about runnability, while task->exit_state are
176 * about the task exiting. Confusing, but this way
177 * modifying one set can't modify the other one by
180 #define TASK_RUNNING 0
181 #define TASK_INTERRUPTIBLE 1
182 #define TASK_UNINTERRUPTIBLE 2
183 #define __TASK_STOPPED 4
184 #define __TASK_TRACED 8
185 /* in tsk->exit_state */
186 #define EXIT_ZOMBIE 16
188 /* in tsk->state again */
190 #define TASK_WAKEKILL 128
192 /* Convenience macros for the sake of set_task_state */
193 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
194 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
195 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
197 /* Convenience macros for the sake of wake_up */
198 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
199 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
201 /* get_task_state() */
202 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
203 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
206 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
207 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
208 #define task_is_stopped_or_traced(task) \
209 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
210 #define task_contributes_to_load(task) \
211 ((task->state & TASK_UNINTERRUPTIBLE) != 0)
213 #define __set_task_state(tsk, state_value) \
214 do { (tsk)->state = (state_value); } while (0)
215 #define set_task_state(tsk, state_value) \
216 set_mb((tsk)->state, (state_value))
219 * set_current_state() includes a barrier so that the write of current->state
220 * is correctly serialised wrt the caller's subsequent test of whether to
223 * set_current_state(TASK_UNINTERRUPTIBLE);
224 * if (do_i_need_to_sleep())
227 * If the caller does not need such serialisation then use __set_current_state()
229 #define __set_current_state(state_value) \
230 do { current->state = (state_value); } while (0)
231 #define set_current_state(state_value) \
232 set_mb(current->state, (state_value))
234 /* Task command name length */
235 #define TASK_COMM_LEN 16
237 #include <linux/spinlock.h>
240 * This serializes "schedule()" and also protects
241 * the run-queue from deletions/modifications (but
242 * _adding_ to the beginning of the run-queue has
245 extern rwlock_t tasklist_lock;
246 extern spinlock_t mmlist_lock;
250 extern void sched_init(void);
251 extern void sched_init_smp(void);
252 extern asmlinkage void schedule_tail(struct task_struct *prev);
253 extern void init_idle(struct task_struct *idle, int cpu);
254 extern void init_idle_bootup_task(struct task_struct *idle);
256 extern int runqueue_is_locked(void);
257 extern void task_rq_unlock_wait(struct task_struct *p);
259 extern cpumask_var_t nohz_cpu_mask;
260 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
261 extern int select_nohz_load_balancer(int cpu);
263 static inline int select_nohz_load_balancer(int cpu)
270 * Only dump TASK_* tasks. (0 for all tasks)
272 extern void show_state_filter(unsigned long state_filter);
274 static inline void show_state(void)
276 show_state_filter(0);
279 extern void show_regs(struct pt_regs *);
282 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
283 * task), SP is the stack pointer of the first frame that should be shown in the back
284 * trace (or NULL if the entire call-chain of the task should be shown).
286 extern void show_stack(struct task_struct *task, unsigned long *sp);
288 void io_schedule(void);
289 long io_schedule_timeout(long timeout);
291 extern void cpu_init (void);
292 extern void trap_init(void);
293 extern void update_process_times(int user);
294 extern void scheduler_tick(void);
296 extern void sched_show_task(struct task_struct *p);
298 #ifdef CONFIG_DETECT_SOFTLOCKUP
299 extern void softlockup_tick(void);
300 extern void touch_softlockup_watchdog(void);
301 extern void touch_all_softlockup_watchdogs(void);
302 extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
303 struct file *filp, void __user *buffer,
304 size_t *lenp, loff_t *ppos);
305 extern unsigned int softlockup_panic;
306 extern unsigned long sysctl_hung_task_check_count;
307 extern unsigned long sysctl_hung_task_timeout_secs;
308 extern unsigned long sysctl_hung_task_warnings;
309 extern int softlockup_thresh;
311 static inline void softlockup_tick(void)
314 static inline void spawn_softlockup_task(void)
317 static inline void touch_softlockup_watchdog(void)
320 static inline void touch_all_softlockup_watchdogs(void)
326 /* Attach to any functions which should be ignored in wchan output. */
327 #define __sched __attribute__((__section__(".sched.text")))
329 /* Linker adds these: start and end of __sched functions */
330 extern char __sched_text_start[], __sched_text_end[];
332 /* Is this address in the __sched functions? */
333 extern int in_sched_functions(unsigned long addr);
335 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
336 extern signed long schedule_timeout(signed long timeout);
337 extern signed long schedule_timeout_interruptible(signed long timeout);
338 extern signed long schedule_timeout_killable(signed long timeout);
339 extern signed long schedule_timeout_uninterruptible(signed long timeout);
340 asmlinkage void __schedule(void);
341 asmlinkage void schedule(void);
342 extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
345 struct user_namespace;
347 /* Maximum number of active map areas.. This is a random (large) number */
348 #define DEFAULT_MAX_MAP_COUNT 65536
350 extern int sysctl_max_map_count;
352 #include <linux/aio.h>
355 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
356 unsigned long, unsigned long);
358 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
359 unsigned long len, unsigned long pgoff,
360 unsigned long flags);
361 extern void arch_unmap_area(struct mm_struct *, unsigned long);
362 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
364 #if USE_SPLIT_PTLOCKS
366 * The mm counters are not protected by its page_table_lock,
367 * so must be incremented atomically.
369 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
370 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
371 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
372 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
373 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
375 #else /* !USE_SPLIT_PTLOCKS */
377 * The mm counters are protected by its page_table_lock,
378 * so can be incremented directly.
380 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
381 #define get_mm_counter(mm, member) ((mm)->_##member)
382 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
383 #define inc_mm_counter(mm, member) (mm)->_##member++
384 #define dec_mm_counter(mm, member) (mm)->_##member--
386 #endif /* !USE_SPLIT_PTLOCKS */
388 #define get_mm_rss(mm) \
389 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
390 #define update_hiwater_rss(mm) do { \
391 unsigned long _rss = get_mm_rss(mm); \
392 if ((mm)->hiwater_rss < _rss) \
393 (mm)->hiwater_rss = _rss; \
395 #define update_hiwater_vm(mm) do { \
396 if ((mm)->hiwater_vm < (mm)->total_vm) \
397 (mm)->hiwater_vm = (mm)->total_vm; \
400 static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
402 return max(mm->hiwater_rss, get_mm_rss(mm));
405 static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
407 return max(mm->hiwater_vm, mm->total_vm);
410 extern void set_dumpable(struct mm_struct *mm, int value);
411 extern int get_dumpable(struct mm_struct *mm);
415 #define MMF_DUMPABLE 0 /* core dump is permitted */
416 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
417 #define MMF_DUMPABLE_BITS 2
419 /* coredump filter bits */
420 #define MMF_DUMP_ANON_PRIVATE 2
421 #define MMF_DUMP_ANON_SHARED 3
422 #define MMF_DUMP_MAPPED_PRIVATE 4
423 #define MMF_DUMP_MAPPED_SHARED 5
424 #define MMF_DUMP_ELF_HEADERS 6
425 #define MMF_DUMP_HUGETLB_PRIVATE 7
426 #define MMF_DUMP_HUGETLB_SHARED 8
427 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
428 #define MMF_DUMP_FILTER_BITS 7
429 #define MMF_DUMP_FILTER_MASK \
430 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
431 #define MMF_DUMP_FILTER_DEFAULT \
432 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
433 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
435 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
436 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
438 # define MMF_DUMP_MASK_DEFAULT_ELF 0
441 struct sighand_struct {
443 struct k_sigaction action[_NSIG];
445 wait_queue_head_t signalfd_wqh;
448 struct pacct_struct {
451 unsigned long ac_mem;
452 cputime_t ac_utime, ac_stime;
453 unsigned long ac_minflt, ac_majflt;
457 * struct task_cputime - collected CPU time counts
458 * @utime: time spent in user mode, in &cputime_t units
459 * @stime: time spent in kernel mode, in &cputime_t units
460 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
462 * This structure groups together three kinds of CPU time that are
463 * tracked for threads and thread groups. Most things considering
464 * CPU time want to group these counts together and treat all three
465 * of them in parallel.
467 struct task_cputime {
470 unsigned long long sum_exec_runtime;
472 /* Alternate field names when used to cache expirations. */
473 #define prof_exp stime
474 #define virt_exp utime
475 #define sched_exp sum_exec_runtime
477 #define INIT_CPUTIME \
478 (struct task_cputime) { \
479 .utime = cputime_zero, \
480 .stime = cputime_zero, \
481 .sum_exec_runtime = 0, \
485 * struct thread_group_cputimer - thread group interval timer counts
486 * @cputime: thread group interval timers.
487 * @running: non-zero when there are timers running and
488 * @cputime receives updates.
489 * @lock: lock for fields in this struct.
491 * This structure contains the version of task_cputime, above, that is
492 * used for thread group CPU timer calculations.
494 struct thread_group_cputimer {
495 struct task_cputime cputime;
501 * NOTE! "signal_struct" does not have it's own
502 * locking, because a shared signal_struct always
503 * implies a shared sighand_struct, so locking
504 * sighand_struct is always a proper superset of
505 * the locking of signal_struct.
507 struct signal_struct {
511 wait_queue_head_t wait_chldexit; /* for wait4() */
513 /* current thread group signal load-balancing target: */
514 struct task_struct *curr_target;
516 /* shared signal handling: */
517 struct sigpending shared_pending;
519 /* thread group exit support */
522 * - notify group_exit_task when ->count is equal to notify_count
523 * - everyone except group_exit_task is stopped during signal delivery
524 * of fatal signals, group_exit_task processes the signal.
527 struct task_struct *group_exit_task;
529 /* thread group stop support, overloads group_exit_code too */
530 int group_stop_count;
531 unsigned int flags; /* see SIGNAL_* flags below */
533 /* POSIX.1b Interval Timers */
534 struct list_head posix_timers;
536 /* ITIMER_REAL timer for the process */
537 struct hrtimer real_timer;
538 struct pid *leader_pid;
539 ktime_t it_real_incr;
541 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
542 cputime_t it_prof_expires, it_virt_expires;
543 cputime_t it_prof_incr, it_virt_incr;
546 * Thread group totals for process CPU timers.
547 * See thread_group_cputimer(), et al, for details.
549 struct thread_group_cputimer cputimer;
551 /* Earliest-expiration cache. */
552 struct task_cputime cputime_expires;
554 struct list_head cpu_timers[3];
556 struct pid *tty_old_pgrp;
558 /* boolean value for session group leader */
561 struct tty_struct *tty; /* NULL if no tty */
564 * Cumulative resource counters for dead threads in the group,
565 * and for reaped dead child processes forked by this group.
566 * Live threads maintain their own counters and add to these
567 * in __exit_signal, except for the group leader.
569 cputime_t utime, stime, cutime, cstime;
572 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
573 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
574 unsigned long inblock, oublock, cinblock, coublock;
575 struct task_io_accounting ioac;
578 * Cumulative ns of schedule CPU time fo dead threads in the
579 * group, not including a zombie group leader, (This only differs
580 * from jiffies_to_ns(utime + stime) if sched_clock uses something
581 * other than jiffies.)
583 unsigned long long sum_sched_runtime;
586 * We don't bother to synchronize most readers of this at all,
587 * because there is no reader checking a limit that actually needs
588 * to get both rlim_cur and rlim_max atomically, and either one
589 * alone is a single word that can safely be read normally.
590 * getrlimit/setrlimit use task_lock(current->group_leader) to
591 * protect this instead of the siglock, because they really
592 * have no need to disable irqs.
594 struct rlimit rlim[RLIM_NLIMITS];
596 #ifdef CONFIG_BSD_PROCESS_ACCT
597 struct pacct_struct pacct; /* per-process accounting information */
599 #ifdef CONFIG_TASKSTATS
600 struct taskstats *stats;
604 struct tty_audit_buf *tty_audit_buf;
608 /* Context switch must be unlocked if interrupts are to be enabled */
609 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
610 # define __ARCH_WANT_UNLOCKED_CTXSW
614 * Bits in flags field of signal_struct.
616 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
617 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
618 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
619 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
621 * Pending notifications to parent.
623 #define SIGNAL_CLD_STOPPED 0x00000010
624 #define SIGNAL_CLD_CONTINUED 0x00000020
625 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
627 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
629 /* If true, all threads except ->group_exit_task have pending SIGKILL */
630 static inline int signal_group_exit(const struct signal_struct *sig)
632 return (sig->flags & SIGNAL_GROUP_EXIT) ||
633 (sig->group_exit_task != NULL);
637 * Some day this will be a full-fledged user tracking system..
640 atomic_t __count; /* reference count */
641 atomic_t processes; /* How many processes does this user have? */
642 atomic_t files; /* How many open files does this user have? */
643 atomic_t sigpending; /* How many pending signals does this user have? */
644 #ifdef CONFIG_INOTIFY_USER
645 atomic_t inotify_watches; /* How many inotify watches does this user have? */
646 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
649 atomic_t epoll_watches; /* The number of file descriptors currently watched */
651 #ifdef CONFIG_POSIX_MQUEUE
652 /* protected by mq_lock */
653 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
655 unsigned long locked_shm; /* How many pages of mlocked shm ? */
658 struct key *uid_keyring; /* UID specific keyring */
659 struct key *session_keyring; /* UID's default session keyring */
662 /* Hash table maintenance information */
663 struct hlist_node uidhash_node;
665 struct user_namespace *user_ns;
667 #ifdef CONFIG_USER_SCHED
668 struct task_group *tg;
671 struct work_struct work;
676 extern int uids_sysfs_init(void);
678 extern struct user_struct *find_user(uid_t);
680 extern struct user_struct root_user;
681 #define INIT_USER (&root_user)
684 struct backing_dev_info;
685 struct reclaim_state;
687 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
689 /* cumulative counters */
690 unsigned long pcount; /* # of times run on this cpu */
691 unsigned long long run_delay; /* time spent waiting on a runqueue */
694 unsigned long long last_arrival,/* when we last ran on a cpu */
695 last_queued; /* when we were last queued to run */
696 #ifdef CONFIG_SCHEDSTATS
698 unsigned int bkl_count;
701 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
703 #ifdef CONFIG_TASK_DELAY_ACCT
704 struct task_delay_info {
706 unsigned int flags; /* Private per-task flags */
708 /* For each stat XXX, add following, aligned appropriately
710 * struct timespec XXX_start, XXX_end;
714 * Atomicity of updates to XXX_delay, XXX_count protected by
715 * single lock above (split into XXX_lock if contention is an issue).
719 * XXX_count is incremented on every XXX operation, the delay
720 * associated with the operation is added to XXX_delay.
721 * XXX_delay contains the accumulated delay time in nanoseconds.
723 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
724 u64 blkio_delay; /* wait for sync block io completion */
725 u64 swapin_delay; /* wait for swapin block io completion */
726 u32 blkio_count; /* total count of the number of sync block */
727 /* io operations performed */
728 u32 swapin_count; /* total count of the number of swapin block */
729 /* io operations performed */
731 struct timespec freepages_start, freepages_end;
732 u64 freepages_delay; /* wait for memory reclaim */
733 u32 freepages_count; /* total count of memory reclaim */
735 #endif /* CONFIG_TASK_DELAY_ACCT */
737 static inline int sched_info_on(void)
739 #ifdef CONFIG_SCHEDSTATS
741 #elif defined(CONFIG_TASK_DELAY_ACCT)
742 extern int delayacct_on;
757 * sched-domains (multiprocessor balancing) declarations:
761 * Increase resolution of nice-level calculations:
763 #define SCHED_LOAD_SHIFT 10
764 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
766 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
769 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
770 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
771 #define SD_BALANCE_EXEC 4 /* Balance on exec */
772 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
773 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
774 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
775 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
776 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
777 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
778 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
779 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
780 #define SD_WAKE_IDLE_FAR 2048 /* Gain latency sacrificing cache hit */
782 enum powersavings_balance_level {
783 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
784 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
785 * first for long running threads
787 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
788 * cpu package for power savings
790 MAX_POWERSAVINGS_BALANCE_LEVELS
793 extern int sched_mc_power_savings, sched_smt_power_savings;
795 static inline int sd_balance_for_mc_power(void)
797 if (sched_smt_power_savings)
798 return SD_POWERSAVINGS_BALANCE;
803 static inline int sd_balance_for_package_power(void)
805 if (sched_mc_power_savings | sched_smt_power_savings)
806 return SD_POWERSAVINGS_BALANCE;
812 * Optimise SD flags for power savings:
813 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
814 * Keep default SD flags if sched_{smt,mc}_power_saving=0
817 static inline int sd_power_saving_flags(void)
819 if (sched_mc_power_savings | sched_smt_power_savings)
820 return SD_BALANCE_NEWIDLE;
826 struct sched_group *next; /* Must be a circular list */
829 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
830 * single CPU. This is read only (except for setup, hotplug CPU).
831 * Note : Never change cpu_power without recompute its reciprocal
833 unsigned int __cpu_power;
835 * reciprocal value of cpu_power to avoid expensive divides
836 * (see include/linux/reciprocal_div.h)
838 u32 reciprocal_cpu_power;
840 unsigned long cpumask[];
843 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
845 return to_cpumask(sg->cpumask);
848 enum sched_domain_level {
858 struct sched_domain_attr {
859 int relax_domain_level;
862 #define SD_ATTR_INIT (struct sched_domain_attr) { \
863 .relax_domain_level = -1, \
866 struct sched_domain {
867 /* These fields must be setup */
868 struct sched_domain *parent; /* top domain must be null terminated */
869 struct sched_domain *child; /* bottom domain must be null terminated */
870 struct sched_group *groups; /* the balancing groups of the domain */
871 unsigned long min_interval; /* Minimum balance interval ms */
872 unsigned long max_interval; /* Maximum balance interval ms */
873 unsigned int busy_factor; /* less balancing by factor if busy */
874 unsigned int imbalance_pct; /* No balance until over watermark */
875 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
876 unsigned int busy_idx;
877 unsigned int idle_idx;
878 unsigned int newidle_idx;
879 unsigned int wake_idx;
880 unsigned int forkexec_idx;
881 int flags; /* See SD_* */
882 enum sched_domain_level level;
884 /* Runtime fields. */
885 unsigned long last_balance; /* init to jiffies. units in jiffies */
886 unsigned int balance_interval; /* initialise to 1. units in ms. */
887 unsigned int nr_balance_failed; /* initialise to 0 */
891 #ifdef CONFIG_SCHEDSTATS
892 /* load_balance() stats */
893 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
894 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
895 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
896 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
897 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
898 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
899 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
900 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
902 /* Active load balancing */
903 unsigned int alb_count;
904 unsigned int alb_failed;
905 unsigned int alb_pushed;
907 /* SD_BALANCE_EXEC stats */
908 unsigned int sbe_count;
909 unsigned int sbe_balanced;
910 unsigned int sbe_pushed;
912 /* SD_BALANCE_FORK stats */
913 unsigned int sbf_count;
914 unsigned int sbf_balanced;
915 unsigned int sbf_pushed;
917 /* try_to_wake_up() stats */
918 unsigned int ttwu_wake_remote;
919 unsigned int ttwu_move_affine;
920 unsigned int ttwu_move_balance;
922 #ifdef CONFIG_SCHED_DEBUG
926 /* span of all CPUs in this domain */
927 unsigned long span[];
930 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
932 return to_cpumask(sd->span);
935 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
936 struct sched_domain_attr *dattr_new);
938 /* Test a flag in parent sched domain */
939 static inline int test_sd_parent(struct sched_domain *sd, int flag)
941 if (sd->parent && (sd->parent->flags & flag))
947 #else /* CONFIG_SMP */
949 struct sched_domain_attr;
952 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
953 struct sched_domain_attr *dattr_new)
956 #endif /* !CONFIG_SMP */
958 struct io_context; /* See blkdev.h */
961 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
962 extern void prefetch_stack(struct task_struct *t);
964 static inline void prefetch_stack(struct task_struct *t) { }
967 struct audit_context; /* See audit.c */
969 struct pipe_inode_info;
970 struct uts_namespace;
976 const struct sched_class *next;
978 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
979 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
980 void (*yield_task) (struct rq *rq);
982 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
984 struct task_struct * (*pick_next_task) (struct rq *rq);
985 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
988 int (*select_task_rq)(struct task_struct *p, int sync);
990 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
991 struct rq *busiest, unsigned long max_load_move,
992 struct sched_domain *sd, enum cpu_idle_type idle,
993 int *all_pinned, int *this_best_prio);
995 int (*move_one_task) (struct rq *this_rq, int this_cpu,
996 struct rq *busiest, struct sched_domain *sd,
997 enum cpu_idle_type idle);
998 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
999 int (*needs_post_schedule) (struct rq *this_rq);
1000 void (*post_schedule) (struct rq *this_rq);
1001 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
1003 void (*set_cpus_allowed)(struct task_struct *p,
1004 const struct cpumask *newmask);
1006 void (*rq_online)(struct rq *rq);
1007 void (*rq_offline)(struct rq *rq);
1010 void (*set_curr_task) (struct rq *rq);
1011 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1012 void (*task_new) (struct rq *rq, struct task_struct *p);
1014 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1016 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1018 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1019 int oldprio, int running);
1021 #ifdef CONFIG_FAIR_GROUP_SCHED
1022 void (*moved_group) (struct task_struct *p);
1026 struct load_weight {
1027 unsigned long weight, inv_weight;
1031 * CFS stats for a schedulable entity (task, task-group etc)
1033 * Current field usage histogram:
1040 struct sched_entity {
1041 struct load_weight load; /* for load-balancing */
1042 struct rb_node run_node;
1043 struct list_head group_node;
1047 u64 sum_exec_runtime;
1049 u64 prev_sum_exec_runtime;
1059 #ifdef CONFIG_SCHEDSTATS
1067 s64 sum_sleep_runtime;
1074 u64 nr_migrations_cold;
1075 u64 nr_failed_migrations_affine;
1076 u64 nr_failed_migrations_running;
1077 u64 nr_failed_migrations_hot;
1078 u64 nr_forced_migrations;
1079 u64 nr_forced2_migrations;
1082 u64 nr_wakeups_sync;
1083 u64 nr_wakeups_migrate;
1084 u64 nr_wakeups_local;
1085 u64 nr_wakeups_remote;
1086 u64 nr_wakeups_affine;
1087 u64 nr_wakeups_affine_attempts;
1088 u64 nr_wakeups_passive;
1089 u64 nr_wakeups_idle;
1092 #ifdef CONFIG_FAIR_GROUP_SCHED
1093 struct sched_entity *parent;
1094 /* rq on which this entity is (to be) queued: */
1095 struct cfs_rq *cfs_rq;
1096 /* rq "owned" by this entity/group: */
1097 struct cfs_rq *my_q;
1101 struct sched_rt_entity {
1102 struct list_head run_list;
1103 unsigned long timeout;
1104 unsigned int time_slice;
1105 int nr_cpus_allowed;
1107 struct sched_rt_entity *back;
1108 #ifdef CONFIG_RT_GROUP_SCHED
1109 struct sched_rt_entity *parent;
1110 /* rq on which this entity is (to be) queued: */
1111 struct rt_rq *rt_rq;
1112 /* rq "owned" by this entity/group: */
1117 struct task_struct {
1118 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
1121 unsigned int flags; /* per process flags, defined below */
1122 unsigned int ptrace;
1124 int lock_depth; /* BKL lock depth */
1127 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1132 int prio, static_prio, normal_prio;
1133 unsigned int rt_priority;
1134 const struct sched_class *sched_class;
1135 struct sched_entity se;
1136 struct sched_rt_entity rt;
1138 #ifdef CONFIG_PREEMPT_NOTIFIERS
1139 /* list of struct preempt_notifier: */
1140 struct hlist_head preempt_notifiers;
1144 * fpu_counter contains the number of consecutive context switches
1145 * that the FPU is used. If this is over a threshold, the lazy fpu
1146 * saving becomes unlazy to save the trap. This is an unsigned char
1147 * so that after 256 times the counter wraps and the behavior turns
1148 * lazy again; this to deal with bursty apps that only use FPU for
1151 unsigned char fpu_counter;
1152 s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1153 #ifdef CONFIG_BLK_DEV_IO_TRACE
1154 unsigned int btrace_seq;
1157 unsigned int policy;
1158 cpumask_t cpus_allowed;
1160 #ifdef CONFIG_PREEMPT_RCU
1161 int rcu_read_lock_nesting;
1162 int rcu_flipctr_idx;
1163 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1165 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1166 struct sched_info sched_info;
1169 struct list_head tasks;
1170 struct plist_node pushable_tasks;
1172 struct mm_struct *mm, *active_mm;
1175 struct linux_binfmt *binfmt;
1177 int exit_code, exit_signal;
1178 int pdeath_signal; /* The signal sent when the parent dies */
1180 unsigned int personality;
1181 unsigned did_exec:1;
1182 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1187 /* Canary value for the -fstack-protector gcc feature */
1188 unsigned long stack_canary;
1191 * pointers to (original) parent process, youngest child, younger sibling,
1192 * older sibling, respectively. (p->father can be replaced with
1193 * p->real_parent->pid)
1195 struct task_struct *real_parent; /* real parent process */
1196 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1198 * children/sibling forms the list of my natural children
1200 struct list_head children; /* list of my children */
1201 struct list_head sibling; /* linkage in my parent's children list */
1202 struct task_struct *group_leader; /* threadgroup leader */
1205 * ptraced is the list of tasks this task is using ptrace on.
1206 * This includes both natural children and PTRACE_ATTACH targets.
1207 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1209 struct list_head ptraced;
1210 struct list_head ptrace_entry;
1212 #ifdef CONFIG_X86_PTRACE_BTS
1214 * This is the tracer handle for the ptrace BTS extension.
1215 * This field actually belongs to the ptracer task.
1217 struct bts_tracer *bts;
1219 * The buffer to hold the BTS data.
1223 #endif /* CONFIG_X86_PTRACE_BTS */
1225 /* PID/PID hash table linkage. */
1226 struct pid_link pids[PIDTYPE_MAX];
1227 struct list_head thread_group;
1229 struct completion *vfork_done; /* for vfork() */
1230 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1231 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1233 cputime_t utime, stime, utimescaled, stimescaled;
1235 cputime_t prev_utime, prev_stime;
1236 unsigned long nvcsw, nivcsw; /* context switch counts */
1237 struct timespec start_time; /* monotonic time */
1238 struct timespec real_start_time; /* boot based time */
1239 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1240 unsigned long min_flt, maj_flt;
1242 struct task_cputime cputime_expires;
1243 struct list_head cpu_timers[3];
1245 /* process credentials */
1246 const struct cred *real_cred; /* objective and real subjective task
1247 * credentials (COW) */
1248 const struct cred *cred; /* effective (overridable) subjective task
1249 * credentials (COW) */
1250 struct mutex cred_exec_mutex; /* execve vs ptrace cred calculation mutex */
1252 char comm[TASK_COMM_LEN]; /* executable name excluding path
1253 - access with [gs]et_task_comm (which lock
1254 it with task_lock())
1255 - initialized normally by flush_old_exec */
1256 /* file system info */
1257 int link_count, total_link_count;
1258 #ifdef CONFIG_SYSVIPC
1260 struct sysv_sem sysvsem;
1262 #ifdef CONFIG_DETECT_SOFTLOCKUP
1263 /* hung task detection */
1264 unsigned long last_switch_timestamp;
1265 unsigned long last_switch_count;
1267 /* CPU-specific state of this task */
1268 struct thread_struct thread;
1269 /* filesystem information */
1270 struct fs_struct *fs;
1271 /* open file information */
1272 struct files_struct *files;
1274 struct nsproxy *nsproxy;
1275 /* signal handlers */
1276 struct signal_struct *signal;
1277 struct sighand_struct *sighand;
1279 sigset_t blocked, real_blocked;
1280 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1281 struct sigpending pending;
1283 unsigned long sas_ss_sp;
1285 int (*notifier)(void *priv);
1286 void *notifier_data;
1287 sigset_t *notifier_mask;
1288 struct audit_context *audit_context;
1289 #ifdef CONFIG_AUDITSYSCALL
1291 unsigned int sessionid;
1295 /* Thread group tracking */
1298 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1299 spinlock_t alloc_lock;
1301 /* Protection of the PI data structures: */
1304 #ifdef CONFIG_RT_MUTEXES
1305 /* PI waiters blocked on a rt_mutex held by this task */
1306 struct plist_head pi_waiters;
1307 /* Deadlock detection and priority inheritance handling */
1308 struct rt_mutex_waiter *pi_blocked_on;
1311 #ifdef CONFIG_DEBUG_MUTEXES
1312 /* mutex deadlock detection */
1313 struct mutex_waiter *blocked_on;
1315 #ifdef CONFIG_TRACE_IRQFLAGS
1316 unsigned int irq_events;
1317 int hardirqs_enabled;
1318 unsigned long hardirq_enable_ip;
1319 unsigned int hardirq_enable_event;
1320 unsigned long hardirq_disable_ip;
1321 unsigned int hardirq_disable_event;
1322 int softirqs_enabled;
1323 unsigned long softirq_disable_ip;
1324 unsigned int softirq_disable_event;
1325 unsigned long softirq_enable_ip;
1326 unsigned int softirq_enable_event;
1327 int hardirq_context;
1328 int softirq_context;
1330 #ifdef CONFIG_LOCKDEP
1331 # define MAX_LOCK_DEPTH 48UL
1334 unsigned int lockdep_recursion;
1335 struct held_lock held_locks[MAX_LOCK_DEPTH];
1336 gfp_t lockdep_reclaim_gfp;
1339 /* journalling filesystem info */
1342 /* stacked block device info */
1343 struct bio *bio_list, **bio_tail;
1346 struct reclaim_state *reclaim_state;
1348 struct backing_dev_info *backing_dev_info;
1350 struct io_context *io_context;
1352 unsigned long ptrace_message;
1353 siginfo_t *last_siginfo; /* For ptrace use. */
1354 struct task_io_accounting ioac;
1355 #if defined(CONFIG_TASK_XACCT)
1356 u64 acct_rss_mem1; /* accumulated rss usage */
1357 u64 acct_vm_mem1; /* accumulated virtual memory usage */
1358 cputime_t acct_timexpd; /* stime + utime since last update */
1360 #ifdef CONFIG_CPUSETS
1361 nodemask_t mems_allowed;
1362 int cpuset_mems_generation;
1363 int cpuset_mem_spread_rotor;
1365 #ifdef CONFIG_CGROUPS
1366 /* Control Group info protected by css_set_lock */
1367 struct css_set *cgroups;
1368 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1369 struct list_head cg_list;
1372 struct robust_list_head __user *robust_list;
1373 #ifdef CONFIG_COMPAT
1374 struct compat_robust_list_head __user *compat_robust_list;
1376 struct list_head pi_state_list;
1377 struct futex_pi_state *pi_state_cache;
1379 struct perf_counter_context perf_counter_ctx;
1381 struct mempolicy *mempolicy;
1384 atomic_t fs_excl; /* holding fs exclusive resources */
1385 struct rcu_head rcu;
1388 * cache last used pipe for splice
1390 struct pipe_inode_info *splice_pipe;
1391 #ifdef CONFIG_TASK_DELAY_ACCT
1392 struct task_delay_info *delays;
1394 #ifdef CONFIG_FAULT_INJECTION
1397 struct prop_local_single dirties;
1398 #ifdef CONFIG_LATENCYTOP
1399 int latency_record_count;
1400 struct latency_record latency_record[LT_SAVECOUNT];
1403 * time slack values; these are used to round up poll() and
1404 * select() etc timeout values. These are in nanoseconds.
1406 unsigned long timer_slack_ns;
1407 unsigned long default_timer_slack_ns;
1409 struct list_head *scm_work_list;
1410 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1411 /* Index of current stored adress in ret_stack */
1413 /* Stack of return addresses for return function tracing */
1414 struct ftrace_ret_stack *ret_stack;
1415 /* time stamp for last schedule */
1416 unsigned long long ftrace_timestamp;
1418 * Number of functions that haven't been traced
1419 * because of depth overrun.
1421 atomic_t trace_overrun;
1422 /* Pause for the tracing */
1423 atomic_t tracing_graph_pause;
1425 #ifdef CONFIG_TRACING
1426 /* state flags for use by tracers */
1427 unsigned long trace;
1431 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1432 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1435 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1436 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1437 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1438 * values are inverted: lower p->prio value means higher priority.
1440 * The MAX_USER_RT_PRIO value allows the actual maximum
1441 * RT priority to be separate from the value exported to
1442 * user-space. This allows kernel threads to set their
1443 * priority to a value higher than any user task. Note:
1444 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1447 #define MAX_USER_RT_PRIO 100
1448 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1450 #define MAX_PRIO (MAX_RT_PRIO + 40)
1451 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1453 static inline int rt_prio(int prio)
1455 if (unlikely(prio < MAX_RT_PRIO))
1460 static inline int rt_task(struct task_struct *p)
1462 return rt_prio(p->prio);
1465 static inline struct pid *task_pid(struct task_struct *task)
1467 return task->pids[PIDTYPE_PID].pid;
1470 static inline struct pid *task_tgid(struct task_struct *task)
1472 return task->group_leader->pids[PIDTYPE_PID].pid;
1476 * Without tasklist or rcu lock it is not safe to dereference
1477 * the result of task_pgrp/task_session even if task == current,
1478 * we can race with another thread doing sys_setsid/sys_setpgid.
1480 static inline struct pid *task_pgrp(struct task_struct *task)
1482 return task->group_leader->pids[PIDTYPE_PGID].pid;
1485 static inline struct pid *task_session(struct task_struct *task)
1487 return task->group_leader->pids[PIDTYPE_SID].pid;
1490 struct pid_namespace;
1493 * the helpers to get the task's different pids as they are seen
1494 * from various namespaces
1496 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1497 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1499 * task_xid_nr_ns() : id seen from the ns specified;
1501 * set_task_vxid() : assigns a virtual id to a task;
1503 * see also pid_nr() etc in include/linux/pid.h
1505 pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1506 struct pid_namespace *ns);
1508 static inline pid_t task_pid_nr(struct task_struct *tsk)
1513 static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1514 struct pid_namespace *ns)
1516 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1519 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1521 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
1525 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1530 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1532 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1534 return pid_vnr(task_tgid(tsk));
1538 static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1539 struct pid_namespace *ns)
1541 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
1544 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1546 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
1550 static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1551 struct pid_namespace *ns)
1553 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
1556 static inline pid_t task_session_vnr(struct task_struct *tsk)
1558 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
1561 /* obsolete, do not use */
1562 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1564 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1568 * pid_alive - check that a task structure is not stale
1569 * @p: Task structure to be checked.
1571 * Test if a process is not yet dead (at most zombie state)
1572 * If pid_alive fails, then pointers within the task structure
1573 * can be stale and must not be dereferenced.
1575 static inline int pid_alive(struct task_struct *p)
1577 return p->pids[PIDTYPE_PID].pid != NULL;
1581 * is_global_init - check if a task structure is init
1582 * @tsk: Task structure to be checked.
1584 * Check if a task structure is the first user space task the kernel created.
1586 static inline int is_global_init(struct task_struct *tsk)
1588 return tsk->pid == 1;
1592 * is_container_init:
1593 * check whether in the task is init in its own pid namespace.
1595 extern int is_container_init(struct task_struct *tsk);
1597 extern struct pid *cad_pid;
1599 extern void free_task(struct task_struct *tsk);
1600 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1602 extern void __put_task_struct(struct task_struct *t);
1604 static inline void put_task_struct(struct task_struct *t)
1606 if (atomic_dec_and_test(&t->usage))
1607 __put_task_struct(t);
1610 extern cputime_t task_utime(struct task_struct *p);
1611 extern cputime_t task_stime(struct task_struct *p);
1612 extern cputime_t task_gtime(struct task_struct *p);
1617 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1618 /* Not implemented yet, only for 486*/
1619 #define PF_STARTING 0x00000002 /* being created */
1620 #define PF_EXITING 0x00000004 /* getting shut down */
1621 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1622 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1623 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1624 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1625 #define PF_DUMPCORE 0x00000200 /* dumped core */
1626 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1627 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1628 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1629 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1630 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1631 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1632 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1633 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1634 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1635 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1636 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1637 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1638 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1639 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1640 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1641 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1642 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1643 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1644 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1645 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1648 * Only the _current_ task can read/write to tsk->flags, but other
1649 * tasks can access tsk->flags in readonly mode for example
1650 * with tsk_used_math (like during threaded core dumping).
1651 * There is however an exception to this rule during ptrace
1652 * or during fork: the ptracer task is allowed to write to the
1653 * child->flags of its traced child (same goes for fork, the parent
1654 * can write to the child->flags), because we're guaranteed the
1655 * child is not running and in turn not changing child->flags
1656 * at the same time the parent does it.
1658 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1659 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1660 #define clear_used_math() clear_stopped_child_used_math(current)
1661 #define set_used_math() set_stopped_child_used_math(current)
1662 #define conditional_stopped_child_used_math(condition, child) \
1663 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1664 #define conditional_used_math(condition) \
1665 conditional_stopped_child_used_math(condition, current)
1666 #define copy_to_stopped_child_used_math(child) \
1667 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1668 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1669 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1670 #define used_math() tsk_used_math(current)
1673 extern int set_cpus_allowed_ptr(struct task_struct *p,
1674 const struct cpumask *new_mask);
1676 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1677 const struct cpumask *new_mask)
1679 if (!cpumask_test_cpu(0, new_mask))
1684 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1686 return set_cpus_allowed_ptr(p, &new_mask);
1690 * Architectures can set this to 1 if they have specified
1691 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1692 * but then during bootup it turns out that sched_clock()
1693 * is reliable after all:
1695 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1696 extern int sched_clock_stable;
1699 extern unsigned long long sched_clock(void);
1701 extern void sched_clock_init(void);
1702 extern u64 sched_clock_cpu(int cpu);
1704 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1705 static inline void sched_clock_tick(void)
1709 static inline void sched_clock_idle_sleep_event(void)
1713 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1717 extern void sched_clock_tick(void);
1718 extern void sched_clock_idle_sleep_event(void);
1719 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1723 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1724 * clock constructed from sched_clock():
1726 extern unsigned long long cpu_clock(int cpu);
1728 extern unsigned long long
1729 task_sched_runtime(struct task_struct *task);
1730 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1732 /* sched_exec is called by processes performing an exec */
1734 extern void sched_exec(void);
1736 #define sched_exec() {}
1739 extern void sched_clock_idle_sleep_event(void);
1740 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1742 #ifdef CONFIG_HOTPLUG_CPU
1743 extern void idle_task_exit(void);
1745 static inline void idle_task_exit(void) {}
1748 extern void sched_idle_next(void);
1750 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1751 extern void wake_up_idle_cpu(int cpu);
1753 static inline void wake_up_idle_cpu(int cpu) { }
1756 extern unsigned int sysctl_sched_latency;
1757 extern unsigned int sysctl_sched_min_granularity;
1758 extern unsigned int sysctl_sched_wakeup_granularity;
1759 extern unsigned int sysctl_sched_shares_ratelimit;
1760 extern unsigned int sysctl_sched_shares_thresh;
1761 #ifdef CONFIG_SCHED_DEBUG
1762 extern unsigned int sysctl_sched_child_runs_first;
1763 extern unsigned int sysctl_sched_features;
1764 extern unsigned int sysctl_sched_migration_cost;
1765 extern unsigned int sysctl_sched_nr_migrate;
1767 int sched_nr_latency_handler(struct ctl_table *table, int write,
1768 struct file *file, void __user *buffer, size_t *length,
1771 extern unsigned int sysctl_sched_rt_period;
1772 extern int sysctl_sched_rt_runtime;
1774 int sched_rt_handler(struct ctl_table *table, int write,
1775 struct file *filp, void __user *buffer, size_t *lenp,
1778 extern unsigned int sysctl_sched_compat_yield;
1780 #ifdef CONFIG_RT_MUTEXES
1781 extern int rt_mutex_getprio(struct task_struct *p);
1782 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1783 extern void rt_mutex_adjust_pi(struct task_struct *p);
1785 static inline int rt_mutex_getprio(struct task_struct *p)
1787 return p->normal_prio;
1789 # define rt_mutex_adjust_pi(p) do { } while (0)
1792 extern void set_user_nice(struct task_struct *p, long nice);
1793 extern int task_prio(const struct task_struct *p);
1794 extern int task_nice(const struct task_struct *p);
1795 extern int can_nice(const struct task_struct *p, const int nice);
1796 extern int task_curr(const struct task_struct *p);
1797 extern int idle_cpu(int cpu);
1798 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1799 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1800 struct sched_param *);
1801 extern struct task_struct *idle_task(int cpu);
1802 extern struct task_struct *curr_task(int cpu);
1803 extern void set_curr_task(int cpu, struct task_struct *p);
1808 * The default (Linux) execution domain.
1810 extern struct exec_domain default_exec_domain;
1812 union thread_union {
1813 struct thread_info thread_info;
1814 unsigned long stack[THREAD_SIZE/sizeof(long)];
1817 #ifndef __HAVE_ARCH_KSTACK_END
1818 static inline int kstack_end(void *addr)
1820 /* Reliable end of stack detection:
1821 * Some APM bios versions misalign the stack
1823 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1827 extern union thread_union init_thread_union;
1828 extern struct task_struct init_task;
1830 extern struct mm_struct init_mm;
1832 extern struct pid_namespace init_pid_ns;
1835 * find a task by one of its numerical ids
1837 * find_task_by_pid_type_ns():
1838 * it is the most generic call - it finds a task by all id,
1839 * type and namespace specified
1840 * find_task_by_pid_ns():
1841 * finds a task by its pid in the specified namespace
1842 * find_task_by_vpid():
1843 * finds a task by its virtual pid
1845 * see also find_vpid() etc in include/linux/pid.h
1848 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1849 struct pid_namespace *ns);
1851 extern struct task_struct *find_task_by_vpid(pid_t nr);
1852 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1853 struct pid_namespace *ns);
1855 extern void __set_special_pids(struct pid *pid);
1857 /* per-UID process charging. */
1858 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1859 static inline struct user_struct *get_uid(struct user_struct *u)
1861 atomic_inc(&u->__count);
1864 extern void free_uid(struct user_struct *);
1865 extern void release_uids(struct user_namespace *ns);
1867 #include <asm/current.h>
1869 extern void do_timer(unsigned long ticks);
1871 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1872 extern int wake_up_process(struct task_struct *tsk);
1873 extern void wake_up_new_task(struct task_struct *tsk,
1874 unsigned long clone_flags);
1876 extern void kick_process(struct task_struct *tsk);
1878 static inline void kick_process(struct task_struct *tsk) { }
1880 extern void sched_fork(struct task_struct *p, int clone_flags);
1881 extern void sched_dead(struct task_struct *p);
1883 extern void proc_caches_init(void);
1884 extern void flush_signals(struct task_struct *);
1885 extern void ignore_signals(struct task_struct *);
1886 extern void flush_signal_handlers(struct task_struct *, int force_default);
1887 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1889 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1891 unsigned long flags;
1894 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1895 ret = dequeue_signal(tsk, mask, info);
1896 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1901 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1903 extern void unblock_all_signals(void);
1904 extern void release_task(struct task_struct * p);
1905 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1906 extern int force_sigsegv(int, struct task_struct *);
1907 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1908 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1909 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1910 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1911 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1912 extern int kill_pid(struct pid *pid, int sig, int priv);
1913 extern int kill_proc_info(int, struct siginfo *, pid_t);
1914 extern int do_notify_parent(struct task_struct *, int);
1915 extern void force_sig(int, struct task_struct *);
1916 extern void force_sig_specific(int, struct task_struct *);
1917 extern int send_sig(int, struct task_struct *, int);
1918 extern void zap_other_threads(struct task_struct *p);
1919 extern struct sigqueue *sigqueue_alloc(void);
1920 extern void sigqueue_free(struct sigqueue *);
1921 extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
1922 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1923 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1925 static inline int kill_cad_pid(int sig, int priv)
1927 return kill_pid(cad_pid, sig, priv);
1930 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1931 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1932 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1933 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1935 static inline int is_si_special(const struct siginfo *info)
1937 return info <= SEND_SIG_FORCED;
1940 /* True if we are on the alternate signal stack. */
1942 static inline int on_sig_stack(unsigned long sp)
1944 return (sp - current->sas_ss_sp < current->sas_ss_size);
1947 static inline int sas_ss_flags(unsigned long sp)
1949 return (current->sas_ss_size == 0 ? SS_DISABLE
1950 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1954 * Routines for handling mm_structs
1956 extern struct mm_struct * mm_alloc(void);
1958 /* mmdrop drops the mm and the page tables */
1959 extern void __mmdrop(struct mm_struct *);
1960 static inline void mmdrop(struct mm_struct * mm)
1962 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1966 /* mmput gets rid of the mappings and all user-space */
1967 extern void mmput(struct mm_struct *);
1968 /* Grab a reference to a task's mm, if it is not already going away */
1969 extern struct mm_struct *get_task_mm(struct task_struct *task);
1970 /* Remove the current tasks stale references to the old mm_struct */
1971 extern void mm_release(struct task_struct *, struct mm_struct *);
1972 /* Allocate a new mm structure and copy contents from tsk->mm */
1973 extern struct mm_struct *dup_mm(struct task_struct *tsk);
1975 extern int copy_thread(unsigned long, unsigned long, unsigned long,
1976 struct task_struct *, struct pt_regs *);
1977 extern void flush_thread(void);
1978 extern void exit_thread(void);
1980 extern void exit_files(struct task_struct *);
1981 extern void __cleanup_signal(struct signal_struct *);
1982 extern void __cleanup_sighand(struct sighand_struct *);
1984 extern void exit_itimers(struct signal_struct *);
1985 extern void flush_itimer_signals(void);
1987 extern NORET_TYPE void do_group_exit(int);
1989 extern void daemonize(const char *, ...);
1990 extern int allow_signal(int);
1991 extern int disallow_signal(int);
1993 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1994 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1995 struct task_struct *fork_idle(int);
1997 extern void set_task_comm(struct task_struct *tsk, char *from);
1998 extern char *get_task_comm(char *to, struct task_struct *tsk);
2001 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
2003 static inline unsigned long wait_task_inactive(struct task_struct *p,
2010 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
2012 #define for_each_process(p) \
2013 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2015 extern bool is_single_threaded(struct task_struct *);
2018 * Careful: do_each_thread/while_each_thread is a double loop so
2019 * 'break' will not work as expected - use goto instead.
2021 #define do_each_thread(g, t) \
2022 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2024 #define while_each_thread(g, t) \
2025 while ((t = next_thread(t)) != g)
2027 /* de_thread depends on thread_group_leader not being a pid based check */
2028 #define thread_group_leader(p) (p == p->group_leader)
2030 /* Do to the insanities of de_thread it is possible for a process
2031 * to have the pid of the thread group leader without actually being
2032 * the thread group leader. For iteration through the pids in proc
2033 * all we care about is that we have a task with the appropriate
2034 * pid, we don't actually care if we have the right task.
2036 static inline int has_group_leader_pid(struct task_struct *p)
2038 return p->pid == p->tgid;
2042 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2044 return p1->tgid == p2->tgid;
2047 static inline struct task_struct *next_thread(const struct task_struct *p)
2049 return list_entry(rcu_dereference(p->thread_group.next),
2050 struct task_struct, thread_group);
2053 static inline int thread_group_empty(struct task_struct *p)
2055 return list_empty(&p->thread_group);
2058 #define delay_group_leader(p) \
2059 (thread_group_leader(p) && !thread_group_empty(p))
2061 static inline int task_detached(struct task_struct *p)
2063 return p->exit_signal == -1;
2067 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2068 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2069 * pins the final release of task.io_context. Also protects ->cpuset and
2070 * ->cgroup.subsys[].
2072 * Nests both inside and outside of read_lock(&tasklist_lock).
2073 * It must not be nested with write_lock_irq(&tasklist_lock),
2074 * neither inside nor outside.
2076 static inline void task_lock(struct task_struct *p)
2078 spin_lock(&p->alloc_lock);
2081 static inline void task_unlock(struct task_struct *p)
2083 spin_unlock(&p->alloc_lock);
2086 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2087 unsigned long *flags);
2089 static inline void unlock_task_sighand(struct task_struct *tsk,
2090 unsigned long *flags)
2092 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2095 #ifndef __HAVE_THREAD_FUNCTIONS
2097 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2098 #define task_stack_page(task) ((task)->stack)
2100 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2102 *task_thread_info(p) = *task_thread_info(org);
2103 task_thread_info(p)->task = p;
2106 static inline unsigned long *end_of_stack(struct task_struct *p)
2108 return (unsigned long *)(task_thread_info(p) + 1);
2113 static inline int object_is_on_stack(void *obj)
2115 void *stack = task_stack_page(current);
2117 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2120 extern void thread_info_cache_init(void);
2122 #ifdef CONFIG_DEBUG_STACK_USAGE
2123 static inline unsigned long stack_not_used(struct task_struct *p)
2125 unsigned long *n = end_of_stack(p);
2127 do { /* Skip over canary */
2131 return (unsigned long)n - (unsigned long)end_of_stack(p);
2135 /* set thread flags in other task's structures
2136 * - see asm/thread_info.h for TIF_xxxx flags available
2138 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2140 set_ti_thread_flag(task_thread_info(tsk), flag);
2143 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2145 clear_ti_thread_flag(task_thread_info(tsk), flag);
2148 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2150 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2153 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2155 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2158 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2160 return test_ti_thread_flag(task_thread_info(tsk), flag);
2163 static inline void set_tsk_need_resched(struct task_struct *tsk)
2165 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2168 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2170 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2173 static inline int test_tsk_need_resched(struct task_struct *tsk)
2175 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2178 static inline int signal_pending(struct task_struct *p)
2180 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2183 extern int __fatal_signal_pending(struct task_struct *p);
2185 static inline int fatal_signal_pending(struct task_struct *p)
2187 return signal_pending(p) && __fatal_signal_pending(p);
2190 static inline int signal_pending_state(long state, struct task_struct *p)
2192 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2194 if (!signal_pending(p))
2197 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2200 static inline int need_resched(void)
2202 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2206 * cond_resched() and cond_resched_lock(): latency reduction via
2207 * explicit rescheduling in places that are safe. The return
2208 * value indicates whether a reschedule was done in fact.
2209 * cond_resched_lock() will drop the spinlock before scheduling,
2210 * cond_resched_softirq() will enable bhs before scheduling.
2212 extern int _cond_resched(void);
2213 #ifdef CONFIG_PREEMPT_BKL
2214 static inline int cond_resched(void)
2219 static inline int cond_resched(void)
2221 return _cond_resched();
2224 extern int cond_resched_lock(spinlock_t * lock);
2225 extern int cond_resched_softirq(void);
2226 static inline int cond_resched_bkl(void)
2228 return _cond_resched();
2232 * Does a critical section need to be broken due to another
2233 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2234 * but a general need for low latency)
2236 static inline int spin_needbreak(spinlock_t *lock)
2238 #ifdef CONFIG_PREEMPT
2239 return spin_is_contended(lock);
2246 * Thread group CPU time accounting.
2248 void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2249 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2251 static inline void thread_group_cputime_init(struct signal_struct *sig)
2253 sig->cputimer.cputime = INIT_CPUTIME;
2254 spin_lock_init(&sig->cputimer.lock);
2255 sig->cputimer.running = 0;
2258 static inline void thread_group_cputime_free(struct signal_struct *sig)
2263 * Reevaluate whether the task has signals pending delivery.
2264 * Wake the task if so.
2265 * This is required every time the blocked sigset_t changes.
2266 * callers must hold sighand->siglock.
2268 extern void recalc_sigpending_and_wake(struct task_struct *t);
2269 extern void recalc_sigpending(void);
2271 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2274 * Wrappers for p->thread_info->cpu access. No-op on UP.
2278 static inline unsigned int task_cpu(const struct task_struct *p)
2280 return task_thread_info(p)->cpu;
2283 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2287 static inline unsigned int task_cpu(const struct task_struct *p)
2292 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2296 #endif /* CONFIG_SMP */
2298 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2300 #ifdef CONFIG_TRACING
2302 __trace_special(void *__tr, void *__data,
2303 unsigned long arg1, unsigned long arg2, unsigned long arg3);
2306 __trace_special(void *__tr, void *__data,
2307 unsigned long arg1, unsigned long arg2, unsigned long arg3)
2312 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2313 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2315 extern void normalize_rt_tasks(void);
2317 #ifdef CONFIG_GROUP_SCHED
2319 extern struct task_group init_task_group;
2320 #ifdef CONFIG_USER_SCHED
2321 extern struct task_group root_task_group;
2322 extern void set_tg_uid(struct user_struct *user);
2325 extern struct task_group *sched_create_group(struct task_group *parent);
2326 extern void sched_destroy_group(struct task_group *tg);
2327 extern void sched_move_task(struct task_struct *tsk);
2328 #ifdef CONFIG_FAIR_GROUP_SCHED
2329 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2330 extern unsigned long sched_group_shares(struct task_group *tg);
2332 #ifdef CONFIG_RT_GROUP_SCHED
2333 extern int sched_group_set_rt_runtime(struct task_group *tg,
2334 long rt_runtime_us);
2335 extern long sched_group_rt_runtime(struct task_group *tg);
2336 extern int sched_group_set_rt_period(struct task_group *tg,
2338 extern long sched_group_rt_period(struct task_group *tg);
2339 extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
2343 extern int task_can_switch_user(struct user_struct *up,
2344 struct task_struct *tsk);
2346 #ifdef CONFIG_TASK_XACCT
2347 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2349 tsk->ioac.rchar += amt;
2352 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2354 tsk->ioac.wchar += amt;
2357 static inline void inc_syscr(struct task_struct *tsk)
2362 static inline void inc_syscw(struct task_struct *tsk)
2367 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2371 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2375 static inline void inc_syscr(struct task_struct *tsk)
2379 static inline void inc_syscw(struct task_struct *tsk)
2384 #ifndef TASK_SIZE_OF
2385 #define TASK_SIZE_OF(tsk) TASK_SIZE
2389 * Call the function if the target task is executing on a CPU right now:
2391 extern void task_oncpu_function_call(struct task_struct *p,
2392 void (*func) (void *info), void *info);
2395 #ifdef CONFIG_MM_OWNER
2396 extern void mm_update_next_owner(struct mm_struct *mm);
2397 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2399 static inline void mm_update_next_owner(struct mm_struct *mm)
2403 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2406 #endif /* CONFIG_MM_OWNER */
2408 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2410 #endif /* __KERNEL__ */