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/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/rculist.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;
102 struct perf_counter_context;
105 * List of flags we want to share for kernel threads,
106 * if only because they are not used by them anyway.
108 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
111 * These are the constant used to fake the fixed-point load-average
112 * counting. Some notes:
113 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
114 * a load-average precision of 10 bits integer + 11 bits fractional
115 * - if you want to count load-averages more often, you need more
116 * precision, or rounding will get you. With 2-second counting freq,
117 * the EXP_n values would be 1981, 2034 and 2043 if still using only
120 extern unsigned long avenrun[]; /* Load averages */
121 extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
123 #define FSHIFT 11 /* nr of bits of precision */
124 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
125 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
126 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
127 #define EXP_5 2014 /* 1/exp(5sec/5min) */
128 #define EXP_15 2037 /* 1/exp(5sec/15min) */
130 #define CALC_LOAD(load,exp,n) \
132 load += n*(FIXED_1-exp); \
135 extern unsigned long total_forks;
136 extern int nr_threads;
137 DECLARE_PER_CPU(unsigned long, process_counts);
138 extern int nr_processes(void);
139 extern unsigned long nr_running(void);
140 extern unsigned long nr_uninterruptible(void);
141 extern unsigned long nr_iowait(void);
142 extern void calc_global_load(void);
143 extern u64 cpu_nr_migrations(int cpu);
145 extern unsigned long get_parent_ip(unsigned long addr);
150 #ifdef CONFIG_SCHED_DEBUG
151 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
152 extern void proc_sched_set_task(struct task_struct *p);
154 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
157 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
160 static inline void proc_sched_set_task(struct task_struct *p)
164 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
169 extern unsigned long long time_sync_thresh;
172 * Task state bitmask. NOTE! These bits are also
173 * encoded in fs/proc/array.c: get_task_state().
175 * We have two separate sets of flags: task->state
176 * is about runnability, while task->exit_state are
177 * about the task exiting. Confusing, but this way
178 * modifying one set can't modify the other one by
181 #define TASK_RUNNING 0
182 #define TASK_INTERRUPTIBLE 1
183 #define TASK_UNINTERRUPTIBLE 2
184 #define __TASK_STOPPED 4
185 #define __TASK_TRACED 8
186 /* in tsk->exit_state */
187 #define EXIT_ZOMBIE 16
189 /* in tsk->state again */
191 #define TASK_WAKEKILL 128
193 /* Convenience macros for the sake of set_task_state */
194 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
195 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
196 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
198 /* Convenience macros for the sake of wake_up */
199 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
200 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
202 /* get_task_state() */
203 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
204 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
207 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
208 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
209 #define task_is_stopped_or_traced(task) \
210 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
211 #define task_contributes_to_load(task) \
212 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
213 (task->flags & PF_FROZEN) == 0)
215 #define __set_task_state(tsk, state_value) \
216 do { (tsk)->state = (state_value); } while (0)
217 #define set_task_state(tsk, state_value) \
218 set_mb((tsk)->state, (state_value))
221 * set_current_state() includes a barrier so that the write of current->state
222 * is correctly serialised wrt the caller's subsequent test of whether to
225 * set_current_state(TASK_UNINTERRUPTIBLE);
226 * if (do_i_need_to_sleep())
229 * If the caller does not need such serialisation then use __set_current_state()
231 #define __set_current_state(state_value) \
232 do { current->state = (state_value); } while (0)
233 #define set_current_state(state_value) \
234 set_mb(current->state, (state_value))
236 /* Task command name length */
237 #define TASK_COMM_LEN 16
239 #include <linux/spinlock.h>
242 * This serializes "schedule()" and also protects
243 * the run-queue from deletions/modifications (but
244 * _adding_ to the beginning of the run-queue has
247 extern rwlock_t tasklist_lock;
248 extern spinlock_t mmlist_lock;
252 extern void sched_init(void);
253 extern void sched_init_smp(void);
254 extern asmlinkage void schedule_tail(struct task_struct *prev);
255 extern void init_idle(struct task_struct *idle, int cpu);
256 extern void init_idle_bootup_task(struct task_struct *idle);
258 extern int runqueue_is_locked(void);
259 extern void task_rq_unlock_wait(struct task_struct *p);
261 extern cpumask_var_t nohz_cpu_mask;
262 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
263 extern int select_nohz_load_balancer(int cpu);
264 extern int get_nohz_load_balancer(void);
266 static inline int select_nohz_load_balancer(int cpu)
273 * Only dump TASK_* tasks. (0 for all tasks)
275 extern void show_state_filter(unsigned long state_filter);
277 static inline void show_state(void)
279 show_state_filter(0);
282 extern void show_regs(struct pt_regs *);
285 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
286 * task), SP is the stack pointer of the first frame that should be shown in the back
287 * trace (or NULL if the entire call-chain of the task should be shown).
289 extern void show_stack(struct task_struct *task, unsigned long *sp);
291 void io_schedule(void);
292 long io_schedule_timeout(long timeout);
294 extern void cpu_init (void);
295 extern void trap_init(void);
296 extern void update_process_times(int user);
297 extern void scheduler_tick(void);
299 extern void sched_show_task(struct task_struct *p);
301 #ifdef CONFIG_DETECT_SOFTLOCKUP
302 extern void softlockup_tick(void);
303 extern void touch_softlockup_watchdog(void);
304 extern void touch_all_softlockup_watchdogs(void);
305 extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
306 struct file *filp, void __user *buffer,
307 size_t *lenp, loff_t *ppos);
308 extern unsigned int softlockup_panic;
309 extern int softlockup_thresh;
311 static inline void softlockup_tick(void)
314 static inline void touch_softlockup_watchdog(void)
317 static inline void touch_all_softlockup_watchdogs(void)
322 #ifdef CONFIG_DETECT_HUNG_TASK
323 extern unsigned int sysctl_hung_task_panic;
324 extern unsigned long sysctl_hung_task_check_count;
325 extern unsigned long sysctl_hung_task_timeout_secs;
326 extern unsigned long sysctl_hung_task_warnings;
327 extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
328 struct file *filp, void __user *buffer,
329 size_t *lenp, loff_t *ppos);
332 /* Attach to any functions which should be ignored in wchan output. */
333 #define __sched __attribute__((__section__(".sched.text")))
335 /* Linker adds these: start and end of __sched functions */
336 extern char __sched_text_start[], __sched_text_end[];
338 /* Is this address in the __sched functions? */
339 extern int in_sched_functions(unsigned long addr);
341 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
342 extern signed long schedule_timeout(signed long timeout);
343 extern signed long schedule_timeout_interruptible(signed long timeout);
344 extern signed long schedule_timeout_killable(signed long timeout);
345 extern signed long schedule_timeout_uninterruptible(signed long timeout);
346 asmlinkage void __schedule(void);
347 asmlinkage void schedule(void);
348 extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
351 struct user_namespace;
353 /* Maximum number of active map areas.. This is a random (large) number */
354 #define DEFAULT_MAX_MAP_COUNT 65536
356 extern int sysctl_max_map_count;
358 #include <linux/aio.h>
361 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
362 unsigned long, unsigned long);
364 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
365 unsigned long len, unsigned long pgoff,
366 unsigned long flags);
367 extern void arch_unmap_area(struct mm_struct *, unsigned long);
368 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
370 #if USE_SPLIT_PTLOCKS
372 * The mm counters are not protected by its page_table_lock,
373 * so must be incremented atomically.
375 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
376 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
377 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
378 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
379 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
381 #else /* !USE_SPLIT_PTLOCKS */
383 * The mm counters are protected by its page_table_lock,
384 * so can be incremented directly.
386 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
387 #define get_mm_counter(mm, member) ((mm)->_##member)
388 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
389 #define inc_mm_counter(mm, member) (mm)->_##member++
390 #define dec_mm_counter(mm, member) (mm)->_##member--
392 #endif /* !USE_SPLIT_PTLOCKS */
394 #define get_mm_rss(mm) \
395 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
396 #define update_hiwater_rss(mm) do { \
397 unsigned long _rss = get_mm_rss(mm); \
398 if ((mm)->hiwater_rss < _rss) \
399 (mm)->hiwater_rss = _rss; \
401 #define update_hiwater_vm(mm) do { \
402 if ((mm)->hiwater_vm < (mm)->total_vm) \
403 (mm)->hiwater_vm = (mm)->total_vm; \
406 static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
408 return max(mm->hiwater_rss, get_mm_rss(mm));
411 static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
413 return max(mm->hiwater_vm, mm->total_vm);
416 extern void set_dumpable(struct mm_struct *mm, int value);
417 extern int get_dumpable(struct mm_struct *mm);
421 #define MMF_DUMPABLE 0 /* core dump is permitted */
422 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
423 #define MMF_DUMPABLE_BITS 2
425 /* coredump filter bits */
426 #define MMF_DUMP_ANON_PRIVATE 2
427 #define MMF_DUMP_ANON_SHARED 3
428 #define MMF_DUMP_MAPPED_PRIVATE 4
429 #define MMF_DUMP_MAPPED_SHARED 5
430 #define MMF_DUMP_ELF_HEADERS 6
431 #define MMF_DUMP_HUGETLB_PRIVATE 7
432 #define MMF_DUMP_HUGETLB_SHARED 8
433 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
434 #define MMF_DUMP_FILTER_BITS 7
435 #define MMF_DUMP_FILTER_MASK \
436 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
437 #define MMF_DUMP_FILTER_DEFAULT \
438 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
439 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
441 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
442 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
444 # define MMF_DUMP_MASK_DEFAULT_ELF 0
447 struct sighand_struct {
449 struct k_sigaction action[_NSIG];
451 wait_queue_head_t signalfd_wqh;
454 struct pacct_struct {
457 unsigned long ac_mem;
458 cputime_t ac_utime, ac_stime;
459 unsigned long ac_minflt, ac_majflt;
463 * struct task_cputime - collected CPU time counts
464 * @utime: time spent in user mode, in &cputime_t units
465 * @stime: time spent in kernel mode, in &cputime_t units
466 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
468 * This structure groups together three kinds of CPU time that are
469 * tracked for threads and thread groups. Most things considering
470 * CPU time want to group these counts together and treat all three
471 * of them in parallel.
473 struct task_cputime {
476 unsigned long long sum_exec_runtime;
478 /* Alternate field names when used to cache expirations. */
479 #define prof_exp stime
480 #define virt_exp utime
481 #define sched_exp sum_exec_runtime
483 #define INIT_CPUTIME \
484 (struct task_cputime) { \
485 .utime = cputime_zero, \
486 .stime = cputime_zero, \
487 .sum_exec_runtime = 0, \
491 * struct thread_group_cputimer - thread group interval timer counts
492 * @cputime: thread group interval timers.
493 * @running: non-zero when there are timers running and
494 * @cputime receives updates.
495 * @lock: lock for fields in this struct.
497 * This structure contains the version of task_cputime, above, that is
498 * used for thread group CPU timer calculations.
500 struct thread_group_cputimer {
501 struct task_cputime cputime;
507 * NOTE! "signal_struct" does not have it's own
508 * locking, because a shared signal_struct always
509 * implies a shared sighand_struct, so locking
510 * sighand_struct is always a proper superset of
511 * the locking of signal_struct.
513 struct signal_struct {
517 wait_queue_head_t wait_chldexit; /* for wait4() */
519 /* current thread group signal load-balancing target: */
520 struct task_struct *curr_target;
522 /* shared signal handling: */
523 struct sigpending shared_pending;
525 /* thread group exit support */
528 * - notify group_exit_task when ->count is equal to notify_count
529 * - everyone except group_exit_task is stopped during signal delivery
530 * of fatal signals, group_exit_task processes the signal.
533 struct task_struct *group_exit_task;
535 /* thread group stop support, overloads group_exit_code too */
536 int group_stop_count;
537 unsigned int flags; /* see SIGNAL_* flags below */
539 /* POSIX.1b Interval Timers */
540 struct list_head posix_timers;
542 /* ITIMER_REAL timer for the process */
543 struct hrtimer real_timer;
544 struct pid *leader_pid;
545 ktime_t it_real_incr;
547 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
548 cputime_t it_prof_expires, it_virt_expires;
549 cputime_t it_prof_incr, it_virt_incr;
552 * Thread group totals for process CPU timers.
553 * See thread_group_cputimer(), et al, for details.
555 struct thread_group_cputimer cputimer;
557 /* Earliest-expiration cache. */
558 struct task_cputime cputime_expires;
560 struct list_head cpu_timers[3];
562 struct pid *tty_old_pgrp;
564 /* boolean value for session group leader */
567 struct tty_struct *tty; /* NULL if no tty */
570 * Cumulative resource counters for dead threads in the group,
571 * and for reaped dead child processes forked by this group.
572 * Live threads maintain their own counters and add to these
573 * in __exit_signal, except for the group leader.
575 cputime_t utime, stime, cutime, cstime;
578 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
579 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
580 unsigned long inblock, oublock, cinblock, coublock;
581 struct task_io_accounting ioac;
584 * Cumulative ns of schedule CPU time fo dead threads in the
585 * group, not including a zombie group leader, (This only differs
586 * from jiffies_to_ns(utime + stime) if sched_clock uses something
587 * other than jiffies.)
589 unsigned long long sum_sched_runtime;
592 * We don't bother to synchronize most readers of this at all,
593 * because there is no reader checking a limit that actually needs
594 * to get both rlim_cur and rlim_max atomically, and either one
595 * alone is a single word that can safely be read normally.
596 * getrlimit/setrlimit use task_lock(current->group_leader) to
597 * protect this instead of the siglock, because they really
598 * have no need to disable irqs.
600 struct rlimit rlim[RLIM_NLIMITS];
602 #ifdef CONFIG_BSD_PROCESS_ACCT
603 struct pacct_struct pacct; /* per-process accounting information */
605 #ifdef CONFIG_TASKSTATS
606 struct taskstats *stats;
610 struct tty_audit_buf *tty_audit_buf;
614 /* Context switch must be unlocked if interrupts are to be enabled */
615 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
616 # define __ARCH_WANT_UNLOCKED_CTXSW
620 * Bits in flags field of signal_struct.
622 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
623 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
624 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
625 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
627 * Pending notifications to parent.
629 #define SIGNAL_CLD_STOPPED 0x00000010
630 #define SIGNAL_CLD_CONTINUED 0x00000020
631 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
633 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
635 /* If true, all threads except ->group_exit_task have pending SIGKILL */
636 static inline int signal_group_exit(const struct signal_struct *sig)
638 return (sig->flags & SIGNAL_GROUP_EXIT) ||
639 (sig->group_exit_task != NULL);
643 * Some day this will be a full-fledged user tracking system..
646 atomic_t __count; /* reference count */
647 atomic_t processes; /* How many processes does this user have? */
648 atomic_t files; /* How many open files does this user have? */
649 atomic_t sigpending; /* How many pending signals does this user have? */
650 #ifdef CONFIG_INOTIFY_USER
651 atomic_t inotify_watches; /* How many inotify watches does this user have? */
652 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
655 atomic_t epoll_watches; /* The number of file descriptors currently watched */
657 #ifdef CONFIG_POSIX_MQUEUE
658 /* protected by mq_lock */
659 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
661 unsigned long locked_shm; /* How many pages of mlocked shm ? */
664 struct key *uid_keyring; /* UID specific keyring */
665 struct key *session_keyring; /* UID's default session keyring */
668 /* Hash table maintenance information */
669 struct hlist_node uidhash_node;
671 struct user_namespace *user_ns;
673 #ifdef CONFIG_USER_SCHED
674 struct task_group *tg;
677 struct work_struct work;
681 #ifdef CONFIG_PERF_COUNTERS
682 atomic_long_t locked_vm;
686 extern int uids_sysfs_init(void);
688 extern struct user_struct *find_user(uid_t);
690 extern struct user_struct root_user;
691 #define INIT_USER (&root_user)
694 struct backing_dev_info;
695 struct reclaim_state;
697 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
699 /* cumulative counters */
700 unsigned long pcount; /* # of times run on this cpu */
701 unsigned long long run_delay; /* time spent waiting on a runqueue */
704 unsigned long long last_arrival,/* when we last ran on a cpu */
705 last_queued; /* when we were last queued to run */
706 #ifdef CONFIG_SCHEDSTATS
708 unsigned int bkl_count;
711 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
713 #ifdef CONFIG_TASK_DELAY_ACCT
714 struct task_delay_info {
716 unsigned int flags; /* Private per-task flags */
718 /* For each stat XXX, add following, aligned appropriately
720 * struct timespec XXX_start, XXX_end;
724 * Atomicity of updates to XXX_delay, XXX_count protected by
725 * single lock above (split into XXX_lock if contention is an issue).
729 * XXX_count is incremented on every XXX operation, the delay
730 * associated with the operation is added to XXX_delay.
731 * XXX_delay contains the accumulated delay time in nanoseconds.
733 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
734 u64 blkio_delay; /* wait for sync block io completion */
735 u64 swapin_delay; /* wait for swapin block io completion */
736 u32 blkio_count; /* total count of the number of sync block */
737 /* io operations performed */
738 u32 swapin_count; /* total count of the number of swapin block */
739 /* io operations performed */
741 struct timespec freepages_start, freepages_end;
742 u64 freepages_delay; /* wait for memory reclaim */
743 u32 freepages_count; /* total count of memory reclaim */
745 #endif /* CONFIG_TASK_DELAY_ACCT */
747 static inline int sched_info_on(void)
749 #ifdef CONFIG_SCHEDSTATS
751 #elif defined(CONFIG_TASK_DELAY_ACCT)
752 extern int delayacct_on;
767 * sched-domains (multiprocessor balancing) declarations:
771 * Increase resolution of nice-level calculations:
773 #define SCHED_LOAD_SHIFT 10
774 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
776 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
779 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
780 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
781 #define SD_BALANCE_EXEC 4 /* Balance on exec */
782 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
783 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
784 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
785 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
786 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
787 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
788 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
789 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
790 #define SD_WAKE_IDLE_FAR 2048 /* Gain latency sacrificing cache hit */
792 enum powersavings_balance_level {
793 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
794 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
795 * first for long running threads
797 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
798 * cpu package for power savings
800 MAX_POWERSAVINGS_BALANCE_LEVELS
803 extern int sched_mc_power_savings, sched_smt_power_savings;
805 static inline int sd_balance_for_mc_power(void)
807 if (sched_smt_power_savings)
808 return SD_POWERSAVINGS_BALANCE;
813 static inline int sd_balance_for_package_power(void)
815 if (sched_mc_power_savings | sched_smt_power_savings)
816 return SD_POWERSAVINGS_BALANCE;
822 * Optimise SD flags for power savings:
823 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
824 * Keep default SD flags if sched_{smt,mc}_power_saving=0
827 static inline int sd_power_saving_flags(void)
829 if (sched_mc_power_savings | sched_smt_power_savings)
830 return SD_BALANCE_NEWIDLE;
836 struct sched_group *next; /* Must be a circular list */
839 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
840 * single CPU. This is read only (except for setup, hotplug CPU).
841 * Note : Never change cpu_power without recompute its reciprocal
843 unsigned int __cpu_power;
845 * reciprocal value of cpu_power to avoid expensive divides
846 * (see include/linux/reciprocal_div.h)
848 u32 reciprocal_cpu_power;
851 * The CPUs this group covers.
853 * NOTE: this field is variable length. (Allocated dynamically
854 * by attaching extra space to the end of the structure,
855 * depending on how many CPUs the kernel has booted up with)
857 * It is also be embedded into static data structures at build
858 * time. (See 'struct static_sched_group' in kernel/sched.c)
860 unsigned long cpumask[0];
863 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
865 return to_cpumask(sg->cpumask);
868 enum sched_domain_level {
878 struct sched_domain_attr {
879 int relax_domain_level;
882 #define SD_ATTR_INIT (struct sched_domain_attr) { \
883 .relax_domain_level = -1, \
886 struct sched_domain {
887 /* These fields must be setup */
888 struct sched_domain *parent; /* top domain must be null terminated */
889 struct sched_domain *child; /* bottom domain must be null terminated */
890 struct sched_group *groups; /* the balancing groups of the domain */
891 unsigned long min_interval; /* Minimum balance interval ms */
892 unsigned long max_interval; /* Maximum balance interval ms */
893 unsigned int busy_factor; /* less balancing by factor if busy */
894 unsigned int imbalance_pct; /* No balance until over watermark */
895 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
896 unsigned int busy_idx;
897 unsigned int idle_idx;
898 unsigned int newidle_idx;
899 unsigned int wake_idx;
900 unsigned int forkexec_idx;
901 int flags; /* See SD_* */
902 enum sched_domain_level level;
904 /* Runtime fields. */
905 unsigned long last_balance; /* init to jiffies. units in jiffies */
906 unsigned int balance_interval; /* initialise to 1. units in ms. */
907 unsigned int nr_balance_failed; /* initialise to 0 */
911 #ifdef CONFIG_SCHEDSTATS
912 /* load_balance() stats */
913 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
914 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
915 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
916 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
917 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
918 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
919 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
920 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
922 /* Active load balancing */
923 unsigned int alb_count;
924 unsigned int alb_failed;
925 unsigned int alb_pushed;
927 /* SD_BALANCE_EXEC stats */
928 unsigned int sbe_count;
929 unsigned int sbe_balanced;
930 unsigned int sbe_pushed;
932 /* SD_BALANCE_FORK stats */
933 unsigned int sbf_count;
934 unsigned int sbf_balanced;
935 unsigned int sbf_pushed;
937 /* try_to_wake_up() stats */
938 unsigned int ttwu_wake_remote;
939 unsigned int ttwu_move_affine;
940 unsigned int ttwu_move_balance;
942 #ifdef CONFIG_SCHED_DEBUG
947 * Span of all CPUs in this domain.
949 * NOTE: this field is variable length. (Allocated dynamically
950 * by attaching extra space to the end of the structure,
951 * depending on how many CPUs the kernel has booted up with)
953 * It is also be embedded into static data structures at build
954 * time. (See 'struct static_sched_domain' in kernel/sched.c)
956 unsigned long span[0];
959 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
961 return to_cpumask(sd->span);
964 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
965 struct sched_domain_attr *dattr_new);
967 /* Test a flag in parent sched domain */
968 static inline int test_sd_parent(struct sched_domain *sd, int flag)
970 if (sd->parent && (sd->parent->flags & flag))
976 #else /* CONFIG_SMP */
978 struct sched_domain_attr;
981 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
982 struct sched_domain_attr *dattr_new)
985 #endif /* !CONFIG_SMP */
987 struct io_context; /* See blkdev.h */
990 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
991 extern void prefetch_stack(struct task_struct *t);
993 static inline void prefetch_stack(struct task_struct *t) { }
996 struct audit_context; /* See audit.c */
998 struct pipe_inode_info;
999 struct uts_namespace;
1002 struct sched_domain;
1004 struct sched_class {
1005 const struct sched_class *next;
1007 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
1008 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
1009 void (*yield_task) (struct rq *rq);
1011 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
1013 struct task_struct * (*pick_next_task) (struct rq *rq);
1014 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
1017 int (*select_task_rq)(struct task_struct *p, int sync);
1019 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
1020 struct rq *busiest, unsigned long max_load_move,
1021 struct sched_domain *sd, enum cpu_idle_type idle,
1022 int *all_pinned, int *this_best_prio);
1024 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1025 struct rq *busiest, struct sched_domain *sd,
1026 enum cpu_idle_type idle);
1027 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1028 int (*needs_post_schedule) (struct rq *this_rq);
1029 void (*post_schedule) (struct rq *this_rq);
1030 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
1032 void (*set_cpus_allowed)(struct task_struct *p,
1033 const struct cpumask *newmask);
1035 void (*rq_online)(struct rq *rq);
1036 void (*rq_offline)(struct rq *rq);
1039 void (*set_curr_task) (struct rq *rq);
1040 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1041 void (*task_new) (struct rq *rq, struct task_struct *p);
1043 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1045 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1047 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1048 int oldprio, int running);
1050 #ifdef CONFIG_FAIR_GROUP_SCHED
1051 void (*moved_group) (struct task_struct *p);
1055 struct load_weight {
1056 unsigned long weight, inv_weight;
1060 * CFS stats for a schedulable entity (task, task-group etc)
1062 * Current field usage histogram:
1069 struct sched_entity {
1070 struct load_weight load; /* for load-balancing */
1071 struct rb_node run_node;
1072 struct list_head group_node;
1076 u64 sum_exec_runtime;
1078 u64 prev_sum_exec_runtime;
1088 #ifdef CONFIG_SCHEDSTATS
1096 s64 sum_sleep_runtime;
1103 u64 nr_migrations_cold;
1104 u64 nr_failed_migrations_affine;
1105 u64 nr_failed_migrations_running;
1106 u64 nr_failed_migrations_hot;
1107 u64 nr_forced_migrations;
1108 u64 nr_forced2_migrations;
1111 u64 nr_wakeups_sync;
1112 u64 nr_wakeups_migrate;
1113 u64 nr_wakeups_local;
1114 u64 nr_wakeups_remote;
1115 u64 nr_wakeups_affine;
1116 u64 nr_wakeups_affine_attempts;
1117 u64 nr_wakeups_passive;
1118 u64 nr_wakeups_idle;
1121 #ifdef CONFIG_FAIR_GROUP_SCHED
1122 struct sched_entity *parent;
1123 /* rq on which this entity is (to be) queued: */
1124 struct cfs_rq *cfs_rq;
1125 /* rq "owned" by this entity/group: */
1126 struct cfs_rq *my_q;
1130 struct sched_rt_entity {
1131 struct list_head run_list;
1132 unsigned long timeout;
1133 unsigned int time_slice;
1134 int nr_cpus_allowed;
1136 struct sched_rt_entity *back;
1137 #ifdef CONFIG_RT_GROUP_SCHED
1138 struct sched_rt_entity *parent;
1139 /* rq on which this entity is (to be) queued: */
1140 struct rt_rq *rt_rq;
1141 /* rq "owned" by this entity/group: */
1146 struct task_struct {
1147 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
1150 unsigned int flags; /* per process flags, defined below */
1151 unsigned int ptrace;
1153 int lock_depth; /* BKL lock depth */
1156 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1161 int prio, static_prio, normal_prio;
1162 unsigned int rt_priority;
1163 const struct sched_class *sched_class;
1164 struct sched_entity se;
1165 struct sched_rt_entity rt;
1167 #ifdef CONFIG_PREEMPT_NOTIFIERS
1168 /* list of struct preempt_notifier: */
1169 struct hlist_head preempt_notifiers;
1173 * fpu_counter contains the number of consecutive context switches
1174 * that the FPU is used. If this is over a threshold, the lazy fpu
1175 * saving becomes unlazy to save the trap. This is an unsigned char
1176 * so that after 256 times the counter wraps and the behavior turns
1177 * lazy again; this to deal with bursty apps that only use FPU for
1180 unsigned char fpu_counter;
1181 s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1182 #ifdef CONFIG_BLK_DEV_IO_TRACE
1183 unsigned int btrace_seq;
1186 unsigned int policy;
1187 cpumask_t cpus_allowed;
1189 #ifdef CONFIG_PREEMPT_RCU
1190 int rcu_read_lock_nesting;
1191 int rcu_flipctr_idx;
1192 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1194 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1195 struct sched_info sched_info;
1198 struct list_head tasks;
1199 struct plist_node pushable_tasks;
1201 struct mm_struct *mm, *active_mm;
1204 struct linux_binfmt *binfmt;
1206 int exit_code, exit_signal;
1207 int pdeath_signal; /* The signal sent when the parent dies */
1209 unsigned int personality;
1210 unsigned did_exec:1;
1211 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1216 /* Canary value for the -fstack-protector gcc feature */
1217 unsigned long stack_canary;
1220 * pointers to (original) parent process, youngest child, younger sibling,
1221 * older sibling, respectively. (p->father can be replaced with
1222 * p->real_parent->pid)
1224 struct task_struct *real_parent; /* real parent process */
1225 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1227 * children/sibling forms the list of my natural children
1229 struct list_head children; /* list of my children */
1230 struct list_head sibling; /* linkage in my parent's children list */
1231 struct task_struct *group_leader; /* threadgroup leader */
1234 * ptraced is the list of tasks this task is using ptrace on.
1235 * This includes both natural children and PTRACE_ATTACH targets.
1236 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1238 struct list_head ptraced;
1239 struct list_head ptrace_entry;
1242 * This is the tracer handle for the ptrace BTS extension.
1243 * This field actually belongs to the ptracer task.
1245 struct bts_context *bts;
1247 /* PID/PID hash table linkage. */
1248 struct pid_link pids[PIDTYPE_MAX];
1249 struct list_head thread_group;
1251 struct completion *vfork_done; /* for vfork() */
1252 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1253 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1255 cputime_t utime, stime, utimescaled, stimescaled;
1257 cputime_t prev_utime, prev_stime;
1258 unsigned long nvcsw, nivcsw; /* context switch counts */
1259 struct timespec start_time; /* monotonic time */
1260 struct timespec real_start_time; /* boot based time */
1261 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1262 unsigned long min_flt, maj_flt;
1264 struct task_cputime cputime_expires;
1265 struct list_head cpu_timers[3];
1267 /* process credentials */
1268 const struct cred *real_cred; /* objective and real subjective task
1269 * credentials (COW) */
1270 const struct cred *cred; /* effective (overridable) subjective task
1271 * credentials (COW) */
1272 struct mutex cred_guard_mutex; /* guard against foreign influences on
1273 * credential calculations
1274 * (notably. ptrace) */
1276 char comm[TASK_COMM_LEN]; /* executable name excluding path
1277 - access with [gs]et_task_comm (which lock
1278 it with task_lock())
1279 - initialized normally by flush_old_exec */
1280 /* file system info */
1281 int link_count, total_link_count;
1282 #ifdef CONFIG_SYSVIPC
1284 struct sysv_sem sysvsem;
1286 #ifdef CONFIG_DETECT_HUNG_TASK
1287 /* hung task detection */
1288 unsigned long last_switch_count;
1290 /* CPU-specific state of this task */
1291 struct thread_struct thread;
1292 /* filesystem information */
1293 struct fs_struct *fs;
1294 /* open file information */
1295 struct files_struct *files;
1297 struct nsproxy *nsproxy;
1298 /* signal handlers */
1299 struct signal_struct *signal;
1300 struct sighand_struct *sighand;
1302 sigset_t blocked, real_blocked;
1303 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1304 struct sigpending pending;
1306 unsigned long sas_ss_sp;
1308 int (*notifier)(void *priv);
1309 void *notifier_data;
1310 sigset_t *notifier_mask;
1311 struct audit_context *audit_context;
1312 #ifdef CONFIG_AUDITSYSCALL
1314 unsigned int sessionid;
1318 /* Thread group tracking */
1321 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1323 spinlock_t alloc_lock;
1325 #ifdef CONFIG_GENERIC_HARDIRQS
1326 /* IRQ handler threads */
1327 struct irqaction *irqaction;
1330 /* Protection of the PI data structures: */
1333 #ifdef CONFIG_RT_MUTEXES
1334 /* PI waiters blocked on a rt_mutex held by this task */
1335 struct plist_head pi_waiters;
1336 /* Deadlock detection and priority inheritance handling */
1337 struct rt_mutex_waiter *pi_blocked_on;
1340 #ifdef CONFIG_DEBUG_MUTEXES
1341 /* mutex deadlock detection */
1342 struct mutex_waiter *blocked_on;
1344 #ifdef CONFIG_TRACE_IRQFLAGS
1345 unsigned int irq_events;
1346 int hardirqs_enabled;
1347 unsigned long hardirq_enable_ip;
1348 unsigned int hardirq_enable_event;
1349 unsigned long hardirq_disable_ip;
1350 unsigned int hardirq_disable_event;
1351 int softirqs_enabled;
1352 unsigned long softirq_disable_ip;
1353 unsigned int softirq_disable_event;
1354 unsigned long softirq_enable_ip;
1355 unsigned int softirq_enable_event;
1356 int hardirq_context;
1357 int softirq_context;
1359 #ifdef CONFIG_LOCKDEP
1360 # define MAX_LOCK_DEPTH 48UL
1363 unsigned int lockdep_recursion;
1364 struct held_lock held_locks[MAX_LOCK_DEPTH];
1365 gfp_t lockdep_reclaim_gfp;
1368 /* journalling filesystem info */
1371 /* stacked block device info */
1372 struct bio *bio_list, **bio_tail;
1375 struct reclaim_state *reclaim_state;
1377 struct backing_dev_info *backing_dev_info;
1379 struct io_context *io_context;
1381 unsigned long ptrace_message;
1382 siginfo_t *last_siginfo; /* For ptrace use. */
1383 struct task_io_accounting ioac;
1384 #if defined(CONFIG_TASK_XACCT)
1385 u64 acct_rss_mem1; /* accumulated rss usage */
1386 u64 acct_vm_mem1; /* accumulated virtual memory usage */
1387 cputime_t acct_timexpd; /* stime + utime since last update */
1389 #ifdef CONFIG_CPUSETS
1390 nodemask_t mems_allowed; /* Protected by alloc_lock */
1391 int cpuset_mem_spread_rotor;
1393 #ifdef CONFIG_CGROUPS
1394 /* Control Group info protected by css_set_lock */
1395 struct css_set *cgroups;
1396 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1397 struct list_head cg_list;
1400 struct robust_list_head __user *robust_list;
1401 #ifdef CONFIG_COMPAT
1402 struct compat_robust_list_head __user *compat_robust_list;
1404 struct list_head pi_state_list;
1405 struct futex_pi_state *pi_state_cache;
1407 #ifdef CONFIG_PERF_COUNTERS
1408 struct perf_counter_context *perf_counter_ctxp;
1409 struct mutex perf_counter_mutex;
1410 struct list_head perf_counter_list;
1413 struct mempolicy *mempolicy; /* Protected by alloc_lock */
1416 atomic_t fs_excl; /* holding fs exclusive resources */
1417 struct rcu_head rcu;
1420 * cache last used pipe for splice
1422 struct pipe_inode_info *splice_pipe;
1423 #ifdef CONFIG_TASK_DELAY_ACCT
1424 struct task_delay_info *delays;
1426 #ifdef CONFIG_FAULT_INJECTION
1429 struct prop_local_single dirties;
1430 #ifdef CONFIG_LATENCYTOP
1431 int latency_record_count;
1432 struct latency_record latency_record[LT_SAVECOUNT];
1435 * time slack values; these are used to round up poll() and
1436 * select() etc timeout values. These are in nanoseconds.
1438 unsigned long timer_slack_ns;
1439 unsigned long default_timer_slack_ns;
1441 struct list_head *scm_work_list;
1442 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1443 /* Index of current stored adress in ret_stack */
1445 /* Stack of return addresses for return function tracing */
1446 struct ftrace_ret_stack *ret_stack;
1447 /* time stamp for last schedule */
1448 unsigned long long ftrace_timestamp;
1450 * Number of functions that haven't been traced
1451 * because of depth overrun.
1453 atomic_t trace_overrun;
1454 /* Pause for the tracing */
1455 atomic_t tracing_graph_pause;
1457 #ifdef CONFIG_TRACING
1458 /* state flags for use by tracers */
1459 unsigned long trace;
1460 /* bitmask of trace recursion */
1461 unsigned long trace_recursion;
1462 #endif /* CONFIG_TRACING */
1465 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1466 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1469 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1470 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1471 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1472 * values are inverted: lower p->prio value means higher priority.
1474 * The MAX_USER_RT_PRIO value allows the actual maximum
1475 * RT priority to be separate from the value exported to
1476 * user-space. This allows kernel threads to set their
1477 * priority to a value higher than any user task. Note:
1478 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1481 #define MAX_USER_RT_PRIO 100
1482 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1484 #define MAX_PRIO (MAX_RT_PRIO + 40)
1485 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1487 static inline int rt_prio(int prio)
1489 if (unlikely(prio < MAX_RT_PRIO))
1494 static inline int rt_task(struct task_struct *p)
1496 return rt_prio(p->prio);
1499 static inline struct pid *task_pid(struct task_struct *task)
1501 return task->pids[PIDTYPE_PID].pid;
1504 static inline struct pid *task_tgid(struct task_struct *task)
1506 return task->group_leader->pids[PIDTYPE_PID].pid;
1510 * Without tasklist or rcu lock it is not safe to dereference
1511 * the result of task_pgrp/task_session even if task == current,
1512 * we can race with another thread doing sys_setsid/sys_setpgid.
1514 static inline struct pid *task_pgrp(struct task_struct *task)
1516 return task->group_leader->pids[PIDTYPE_PGID].pid;
1519 static inline struct pid *task_session(struct task_struct *task)
1521 return task->group_leader->pids[PIDTYPE_SID].pid;
1524 struct pid_namespace;
1527 * the helpers to get the task's different pids as they are seen
1528 * from various namespaces
1530 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1531 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1533 * task_xid_nr_ns() : id seen from the ns specified;
1535 * set_task_vxid() : assigns a virtual id to a task;
1537 * see also pid_nr() etc in include/linux/pid.h
1539 pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1540 struct pid_namespace *ns);
1542 static inline pid_t task_pid_nr(struct task_struct *tsk)
1547 static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1548 struct pid_namespace *ns)
1550 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1553 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1555 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
1559 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1564 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1566 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1568 return pid_vnr(task_tgid(tsk));
1572 static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1573 struct pid_namespace *ns)
1575 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
1578 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1580 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
1584 static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1585 struct pid_namespace *ns)
1587 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
1590 static inline pid_t task_session_vnr(struct task_struct *tsk)
1592 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
1595 /* obsolete, do not use */
1596 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1598 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1602 * pid_alive - check that a task structure is not stale
1603 * @p: Task structure to be checked.
1605 * Test if a process is not yet dead (at most zombie state)
1606 * If pid_alive fails, then pointers within the task structure
1607 * can be stale and must not be dereferenced.
1609 static inline int pid_alive(struct task_struct *p)
1611 return p->pids[PIDTYPE_PID].pid != NULL;
1615 * is_global_init - check if a task structure is init
1616 * @tsk: Task structure to be checked.
1618 * Check if a task structure is the first user space task the kernel created.
1620 static inline int is_global_init(struct task_struct *tsk)
1622 return tsk->pid == 1;
1626 * is_container_init:
1627 * check whether in the task is init in its own pid namespace.
1629 extern int is_container_init(struct task_struct *tsk);
1631 extern struct pid *cad_pid;
1633 extern void free_task(struct task_struct *tsk);
1634 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1636 extern void __put_task_struct(struct task_struct *t);
1638 static inline void put_task_struct(struct task_struct *t)
1640 if (atomic_dec_and_test(&t->usage))
1641 __put_task_struct(t);
1644 extern cputime_t task_utime(struct task_struct *p);
1645 extern cputime_t task_stime(struct task_struct *p);
1646 extern cputime_t task_gtime(struct task_struct *p);
1651 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1652 /* Not implemented yet, only for 486*/
1653 #define PF_STARTING 0x00000002 /* being created */
1654 #define PF_EXITING 0x00000004 /* getting shut down */
1655 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1656 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1657 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1658 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1659 #define PF_DUMPCORE 0x00000200 /* dumped core */
1660 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1661 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1662 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1663 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1664 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1665 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1666 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1667 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1668 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1669 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1670 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1671 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1672 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1673 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1674 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1675 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1676 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1677 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1678 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1679 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1682 * Only the _current_ task can read/write to tsk->flags, but other
1683 * tasks can access tsk->flags in readonly mode for example
1684 * with tsk_used_math (like during threaded core dumping).
1685 * There is however an exception to this rule during ptrace
1686 * or during fork: the ptracer task is allowed to write to the
1687 * child->flags of its traced child (same goes for fork, the parent
1688 * can write to the child->flags), because we're guaranteed the
1689 * child is not running and in turn not changing child->flags
1690 * at the same time the parent does it.
1692 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1693 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1694 #define clear_used_math() clear_stopped_child_used_math(current)
1695 #define set_used_math() set_stopped_child_used_math(current)
1696 #define conditional_stopped_child_used_math(condition, child) \
1697 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1698 #define conditional_used_math(condition) \
1699 conditional_stopped_child_used_math(condition, current)
1700 #define copy_to_stopped_child_used_math(child) \
1701 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1702 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1703 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1704 #define used_math() tsk_used_math(current)
1707 extern int set_cpus_allowed_ptr(struct task_struct *p,
1708 const struct cpumask *new_mask);
1710 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1711 const struct cpumask *new_mask)
1713 if (!cpumask_test_cpu(0, new_mask))
1718 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1720 return set_cpus_allowed_ptr(p, &new_mask);
1724 * Architectures can set this to 1 if they have specified
1725 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1726 * but then during bootup it turns out that sched_clock()
1727 * is reliable after all:
1729 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1730 extern int sched_clock_stable;
1733 extern unsigned long long sched_clock(void);
1735 extern void sched_clock_init(void);
1736 extern u64 sched_clock_cpu(int cpu);
1738 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1739 static inline void sched_clock_tick(void)
1743 static inline void sched_clock_idle_sleep_event(void)
1747 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1751 extern void sched_clock_tick(void);
1752 extern void sched_clock_idle_sleep_event(void);
1753 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1757 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1758 * clock constructed from sched_clock():
1760 extern unsigned long long cpu_clock(int cpu);
1762 extern unsigned long long
1763 task_sched_runtime(struct task_struct *task);
1764 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1766 /* sched_exec is called by processes performing an exec */
1768 extern void sched_exec(void);
1770 #define sched_exec() {}
1773 extern void sched_clock_idle_sleep_event(void);
1774 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1776 #ifdef CONFIG_HOTPLUG_CPU
1777 extern void idle_task_exit(void);
1779 static inline void idle_task_exit(void) {}
1782 extern void sched_idle_next(void);
1784 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1785 extern void wake_up_idle_cpu(int cpu);
1787 static inline void wake_up_idle_cpu(int cpu) { }
1790 extern unsigned int sysctl_sched_latency;
1791 extern unsigned int sysctl_sched_min_granularity;
1792 extern unsigned int sysctl_sched_wakeup_granularity;
1793 extern unsigned int sysctl_sched_shares_ratelimit;
1794 extern unsigned int sysctl_sched_shares_thresh;
1795 #ifdef CONFIG_SCHED_DEBUG
1796 extern unsigned int sysctl_sched_child_runs_first;
1797 extern unsigned int sysctl_sched_features;
1798 extern unsigned int sysctl_sched_migration_cost;
1799 extern unsigned int sysctl_sched_nr_migrate;
1800 extern unsigned int sysctl_timer_migration;
1802 int sched_nr_latency_handler(struct ctl_table *table, int write,
1803 struct file *file, void __user *buffer, size_t *length,
1806 #ifdef CONFIG_SCHED_DEBUG
1807 static inline unsigned int get_sysctl_timer_migration(void)
1809 return sysctl_timer_migration;
1812 static inline unsigned int get_sysctl_timer_migration(void)
1817 extern unsigned int sysctl_sched_rt_period;
1818 extern int sysctl_sched_rt_runtime;
1820 int sched_rt_handler(struct ctl_table *table, int write,
1821 struct file *filp, void __user *buffer, size_t *lenp,
1824 extern unsigned int sysctl_sched_compat_yield;
1826 #ifdef CONFIG_RT_MUTEXES
1827 extern int rt_mutex_getprio(struct task_struct *p);
1828 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1829 extern void rt_mutex_adjust_pi(struct task_struct *p);
1831 static inline int rt_mutex_getprio(struct task_struct *p)
1833 return p->normal_prio;
1835 # define rt_mutex_adjust_pi(p) do { } while (0)
1838 extern void set_user_nice(struct task_struct *p, long nice);
1839 extern int task_prio(const struct task_struct *p);
1840 extern int task_nice(const struct task_struct *p);
1841 extern int can_nice(const struct task_struct *p, const int nice);
1842 extern int task_curr(const struct task_struct *p);
1843 extern int idle_cpu(int cpu);
1844 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1845 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1846 struct sched_param *);
1847 extern struct task_struct *idle_task(int cpu);
1848 extern struct task_struct *curr_task(int cpu);
1849 extern void set_curr_task(int cpu, struct task_struct *p);
1854 * The default (Linux) execution domain.
1856 extern struct exec_domain default_exec_domain;
1858 union thread_union {
1859 struct thread_info thread_info;
1860 unsigned long stack[THREAD_SIZE/sizeof(long)];
1863 #ifndef __HAVE_ARCH_KSTACK_END
1864 static inline int kstack_end(void *addr)
1866 /* Reliable end of stack detection:
1867 * Some APM bios versions misalign the stack
1869 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1873 extern union thread_union init_thread_union;
1874 extern struct task_struct init_task;
1876 extern struct mm_struct init_mm;
1878 extern struct pid_namespace init_pid_ns;
1881 * find a task by one of its numerical ids
1883 * find_task_by_pid_type_ns():
1884 * it is the most generic call - it finds a task by all id,
1885 * type and namespace specified
1886 * find_task_by_pid_ns():
1887 * finds a task by its pid in the specified namespace
1888 * find_task_by_vpid():
1889 * finds a task by its virtual pid
1891 * see also find_vpid() etc in include/linux/pid.h
1894 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1895 struct pid_namespace *ns);
1897 extern struct task_struct *find_task_by_vpid(pid_t nr);
1898 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1899 struct pid_namespace *ns);
1901 extern void __set_special_pids(struct pid *pid);
1903 /* per-UID process charging. */
1904 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1905 static inline struct user_struct *get_uid(struct user_struct *u)
1907 atomic_inc(&u->__count);
1910 extern void free_uid(struct user_struct *);
1911 extern void release_uids(struct user_namespace *ns);
1913 #include <asm/current.h>
1915 extern void do_timer(unsigned long ticks);
1917 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1918 extern int wake_up_process(struct task_struct *tsk);
1919 extern void wake_up_new_task(struct task_struct *tsk,
1920 unsigned long clone_flags);
1922 extern void kick_process(struct task_struct *tsk);
1924 static inline void kick_process(struct task_struct *tsk) { }
1926 extern void sched_fork(struct task_struct *p, int clone_flags);
1927 extern void sched_dead(struct task_struct *p);
1929 extern void proc_caches_init(void);
1930 extern void flush_signals(struct task_struct *);
1931 extern void __flush_signals(struct task_struct *);
1932 extern void ignore_signals(struct task_struct *);
1933 extern void flush_signal_handlers(struct task_struct *, int force_default);
1934 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1936 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1938 unsigned long flags;
1941 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1942 ret = dequeue_signal(tsk, mask, info);
1943 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1948 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1950 extern void unblock_all_signals(void);
1951 extern void release_task(struct task_struct * p);
1952 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1953 extern int force_sigsegv(int, struct task_struct *);
1954 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1955 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1956 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1957 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1958 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1959 extern int kill_pid(struct pid *pid, int sig, int priv);
1960 extern int kill_proc_info(int, struct siginfo *, pid_t);
1961 extern int do_notify_parent(struct task_struct *, int);
1962 extern void force_sig(int, struct task_struct *);
1963 extern void force_sig_specific(int, struct task_struct *);
1964 extern int send_sig(int, struct task_struct *, int);
1965 extern void zap_other_threads(struct task_struct *p);
1966 extern struct sigqueue *sigqueue_alloc(void);
1967 extern void sigqueue_free(struct sigqueue *);
1968 extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
1969 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1970 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1972 static inline int kill_cad_pid(int sig, int priv)
1974 return kill_pid(cad_pid, sig, priv);
1977 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1978 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1979 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1980 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1982 static inline int is_si_special(const struct siginfo *info)
1984 return info <= SEND_SIG_FORCED;
1987 /* True if we are on the alternate signal stack. */
1989 static inline int on_sig_stack(unsigned long sp)
1991 return (sp - current->sas_ss_sp < current->sas_ss_size);
1994 static inline int sas_ss_flags(unsigned long sp)
1996 return (current->sas_ss_size == 0 ? SS_DISABLE
1997 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2001 * Routines for handling mm_structs
2003 extern struct mm_struct * mm_alloc(void);
2005 /* mmdrop drops the mm and the page tables */
2006 extern void __mmdrop(struct mm_struct *);
2007 static inline void mmdrop(struct mm_struct * mm)
2009 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
2013 /* mmput gets rid of the mappings and all user-space */
2014 extern void mmput(struct mm_struct *);
2015 /* Grab a reference to a task's mm, if it is not already going away */
2016 extern struct mm_struct *get_task_mm(struct task_struct *task);
2017 /* Remove the current tasks stale references to the old mm_struct */
2018 extern void mm_release(struct task_struct *, struct mm_struct *);
2019 /* Allocate a new mm structure and copy contents from tsk->mm */
2020 extern struct mm_struct *dup_mm(struct task_struct *tsk);
2022 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2023 struct task_struct *, struct pt_regs *);
2024 extern void flush_thread(void);
2025 extern void exit_thread(void);
2027 extern void exit_files(struct task_struct *);
2028 extern void __cleanup_signal(struct signal_struct *);
2029 extern void __cleanup_sighand(struct sighand_struct *);
2031 extern void exit_itimers(struct signal_struct *);
2032 extern void flush_itimer_signals(void);
2034 extern NORET_TYPE void do_group_exit(int);
2036 extern void daemonize(const char *, ...);
2037 extern int allow_signal(int);
2038 extern int disallow_signal(int);
2040 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2041 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
2042 struct task_struct *fork_idle(int);
2044 extern void set_task_comm(struct task_struct *tsk, char *from);
2045 extern char *get_task_comm(char *to, struct task_struct *tsk);
2048 extern void wait_task_context_switch(struct task_struct *p);
2049 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
2051 static inline void wait_task_context_switch(struct task_struct *p) {}
2052 static inline unsigned long wait_task_inactive(struct task_struct *p,
2059 #define next_task(p) \
2060 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2062 #define for_each_process(p) \
2063 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2065 extern bool is_single_threaded(struct task_struct *);
2068 * Careful: do_each_thread/while_each_thread is a double loop so
2069 * 'break' will not work as expected - use goto instead.
2071 #define do_each_thread(g, t) \
2072 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2074 #define while_each_thread(g, t) \
2075 while ((t = next_thread(t)) != g)
2077 /* de_thread depends on thread_group_leader not being a pid based check */
2078 #define thread_group_leader(p) (p == p->group_leader)
2080 /* Do to the insanities of de_thread it is possible for a process
2081 * to have the pid of the thread group leader without actually being
2082 * the thread group leader. For iteration through the pids in proc
2083 * all we care about is that we have a task with the appropriate
2084 * pid, we don't actually care if we have the right task.
2086 static inline int has_group_leader_pid(struct task_struct *p)
2088 return p->pid == p->tgid;
2092 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2094 return p1->tgid == p2->tgid;
2097 static inline struct task_struct *next_thread(const struct task_struct *p)
2099 return list_entry_rcu(p->thread_group.next,
2100 struct task_struct, thread_group);
2103 static inline int thread_group_empty(struct task_struct *p)
2105 return list_empty(&p->thread_group);
2108 #define delay_group_leader(p) \
2109 (thread_group_leader(p) && !thread_group_empty(p))
2111 static inline int task_detached(struct task_struct *p)
2113 return p->exit_signal == -1;
2117 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2118 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2119 * pins the final release of task.io_context. Also protects ->cpuset and
2120 * ->cgroup.subsys[].
2122 * Nests both inside and outside of read_lock(&tasklist_lock).
2123 * It must not be nested with write_lock_irq(&tasklist_lock),
2124 * neither inside nor outside.
2126 static inline void task_lock(struct task_struct *p)
2128 spin_lock(&p->alloc_lock);
2131 static inline void task_unlock(struct task_struct *p)
2133 spin_unlock(&p->alloc_lock);
2136 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2137 unsigned long *flags);
2139 static inline void unlock_task_sighand(struct task_struct *tsk,
2140 unsigned long *flags)
2142 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2145 #ifndef __HAVE_THREAD_FUNCTIONS
2147 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2148 #define task_stack_page(task) ((task)->stack)
2150 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2152 *task_thread_info(p) = *task_thread_info(org);
2153 task_thread_info(p)->task = p;
2156 static inline unsigned long *end_of_stack(struct task_struct *p)
2158 return (unsigned long *)(task_thread_info(p) + 1);
2163 static inline int object_is_on_stack(void *obj)
2165 void *stack = task_stack_page(current);
2167 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2170 extern void thread_info_cache_init(void);
2172 #ifdef CONFIG_DEBUG_STACK_USAGE
2173 static inline unsigned long stack_not_used(struct task_struct *p)
2175 unsigned long *n = end_of_stack(p);
2177 do { /* Skip over canary */
2181 return (unsigned long)n - (unsigned long)end_of_stack(p);
2185 /* set thread flags in other task's structures
2186 * - see asm/thread_info.h for TIF_xxxx flags available
2188 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2190 set_ti_thread_flag(task_thread_info(tsk), flag);
2193 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2195 clear_ti_thread_flag(task_thread_info(tsk), flag);
2198 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2200 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2203 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2205 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2208 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2210 return test_ti_thread_flag(task_thread_info(tsk), flag);
2213 static inline void set_tsk_need_resched(struct task_struct *tsk)
2215 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2218 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2220 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2223 static inline int test_tsk_need_resched(struct task_struct *tsk)
2225 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2228 static inline int restart_syscall(void)
2230 set_tsk_thread_flag(current, TIF_SIGPENDING);
2231 return -ERESTARTNOINTR;
2234 static inline int signal_pending(struct task_struct *p)
2236 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2239 extern int __fatal_signal_pending(struct task_struct *p);
2241 static inline int fatal_signal_pending(struct task_struct *p)
2243 return signal_pending(p) && __fatal_signal_pending(p);
2246 static inline int signal_pending_state(long state, struct task_struct *p)
2248 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2250 if (!signal_pending(p))
2253 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2256 static inline int need_resched(void)
2258 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2262 * cond_resched() and cond_resched_lock(): latency reduction via
2263 * explicit rescheduling in places that are safe. The return
2264 * value indicates whether a reschedule was done in fact.
2265 * cond_resched_lock() will drop the spinlock before scheduling,
2266 * cond_resched_softirq() will enable bhs before scheduling.
2268 extern int _cond_resched(void);
2269 #ifdef CONFIG_PREEMPT_BKL
2270 static inline int cond_resched(void)
2275 static inline int cond_resched(void)
2277 return _cond_resched();
2280 extern int cond_resched_lock(spinlock_t * lock);
2281 extern int cond_resched_softirq(void);
2282 static inline int cond_resched_bkl(void)
2284 return _cond_resched();
2288 * Does a critical section need to be broken due to another
2289 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2290 * but a general need for low latency)
2292 static inline int spin_needbreak(spinlock_t *lock)
2294 #ifdef CONFIG_PREEMPT
2295 return spin_is_contended(lock);
2302 * Thread group CPU time accounting.
2304 void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2305 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2307 static inline void thread_group_cputime_init(struct signal_struct *sig)
2309 sig->cputimer.cputime = INIT_CPUTIME;
2310 spin_lock_init(&sig->cputimer.lock);
2311 sig->cputimer.running = 0;
2314 static inline void thread_group_cputime_free(struct signal_struct *sig)
2319 * Reevaluate whether the task has signals pending delivery.
2320 * Wake the task if so.
2321 * This is required every time the blocked sigset_t changes.
2322 * callers must hold sighand->siglock.
2324 extern void recalc_sigpending_and_wake(struct task_struct *t);
2325 extern void recalc_sigpending(void);
2327 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2330 * Wrappers for p->thread_info->cpu access. No-op on UP.
2334 static inline unsigned int task_cpu(const struct task_struct *p)
2336 return task_thread_info(p)->cpu;
2339 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2343 static inline unsigned int task_cpu(const struct task_struct *p)
2348 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2352 #endif /* CONFIG_SMP */
2354 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2356 #ifdef CONFIG_TRACING
2358 __trace_special(void *__tr, void *__data,
2359 unsigned long arg1, unsigned long arg2, unsigned long arg3);
2362 __trace_special(void *__tr, void *__data,
2363 unsigned long arg1, unsigned long arg2, unsigned long arg3)
2368 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2369 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2371 extern void normalize_rt_tasks(void);
2373 #ifdef CONFIG_GROUP_SCHED
2375 extern struct task_group init_task_group;
2376 #ifdef CONFIG_USER_SCHED
2377 extern struct task_group root_task_group;
2378 extern void set_tg_uid(struct user_struct *user);
2381 extern struct task_group *sched_create_group(struct task_group *parent);
2382 extern void sched_destroy_group(struct task_group *tg);
2383 extern void sched_move_task(struct task_struct *tsk);
2384 #ifdef CONFIG_FAIR_GROUP_SCHED
2385 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2386 extern unsigned long sched_group_shares(struct task_group *tg);
2388 #ifdef CONFIG_RT_GROUP_SCHED
2389 extern int sched_group_set_rt_runtime(struct task_group *tg,
2390 long rt_runtime_us);
2391 extern long sched_group_rt_runtime(struct task_group *tg);
2392 extern int sched_group_set_rt_period(struct task_group *tg,
2394 extern long sched_group_rt_period(struct task_group *tg);
2395 extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
2399 extern int task_can_switch_user(struct user_struct *up,
2400 struct task_struct *tsk);
2402 #ifdef CONFIG_TASK_XACCT
2403 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2405 tsk->ioac.rchar += amt;
2408 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2410 tsk->ioac.wchar += amt;
2413 static inline void inc_syscr(struct task_struct *tsk)
2418 static inline void inc_syscw(struct task_struct *tsk)
2423 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2427 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2431 static inline void inc_syscr(struct task_struct *tsk)
2435 static inline void inc_syscw(struct task_struct *tsk)
2440 #ifndef TASK_SIZE_OF
2441 #define TASK_SIZE_OF(tsk) TASK_SIZE
2445 * Call the function if the target task is executing on a CPU right now:
2447 extern void task_oncpu_function_call(struct task_struct *p,
2448 void (*func) (void *info), void *info);
2451 #ifdef CONFIG_MM_OWNER
2452 extern void mm_update_next_owner(struct mm_struct *mm);
2453 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2455 static inline void mm_update_next_owner(struct mm_struct *mm)
2459 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2462 #endif /* CONFIG_MM_OWNER */
2464 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2466 #endif /* __KERNEL__ */