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 */
41 /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
42 #define SCHED_RESET_ON_FORK 0x40000000
50 #include <asm/param.h> /* for HZ */
52 #include <linux/capability.h>
53 #include <linux/threads.h>
54 #include <linux/kernel.h>
55 #include <linux/types.h>
56 #include <linux/timex.h>
57 #include <linux/jiffies.h>
58 #include <linux/rbtree.h>
59 #include <linux/thread_info.h>
60 #include <linux/cpumask.h>
61 #include <linux/errno.h>
62 #include <linux/nodemask.h>
63 #include <linux/mm_types.h>
65 #include <asm/system.h>
67 #include <asm/ptrace.h>
68 #include <asm/cputime.h>
70 #include <linux/smp.h>
71 #include <linux/sem.h>
72 #include <linux/signal.h>
73 #include <linux/path.h>
74 #include <linux/compiler.h>
75 #include <linux/completion.h>
76 #include <linux/pid.h>
77 #include <linux/percpu.h>
78 #include <linux/topology.h>
79 #include <linux/proportions.h>
80 #include <linux/seccomp.h>
81 #include <linux/rcupdate.h>
82 #include <linux/rculist.h>
83 #include <linux/rtmutex.h>
85 #include <linux/time.h>
86 #include <linux/param.h>
87 #include <linux/resource.h>
88 #include <linux/timer.h>
89 #include <linux/hrtimer.h>
90 #include <linux/task_io_accounting.h>
91 #include <linux/kobject.h>
92 #include <linux/latencytop.h>
93 #include <linux/cred.h>
95 #include <asm/processor.h>
98 struct futex_pi_state
;
99 struct robust_list_head
;
102 struct perf_event_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 unsigned long nr_iowait_cpu(void);
143 extern unsigned long this_cpu_load(void);
146 extern void calc_global_load(void);
148 extern unsigned long get_parent_ip(unsigned long addr
);
153 #ifdef CONFIG_SCHED_DEBUG
154 extern void proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
);
155 extern void proc_sched_set_task(struct task_struct
*p
);
157 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
);
160 proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
)
163 static inline void proc_sched_set_task(struct task_struct
*p
)
167 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
)
173 * Task state bitmask. NOTE! These bits are also
174 * encoded in fs/proc/array.c: get_task_state().
176 * We have two separate sets of flags: task->state
177 * is about runnability, while task->exit_state are
178 * about the task exiting. Confusing, but this way
179 * modifying one set can't modify the other one by
182 #define TASK_RUNNING 0
183 #define TASK_INTERRUPTIBLE 1
184 #define TASK_UNINTERRUPTIBLE 2
185 #define __TASK_STOPPED 4
186 #define __TASK_TRACED 8
187 /* in tsk->exit_state */
188 #define EXIT_ZOMBIE 16
190 /* in tsk->state again */
192 #define TASK_WAKEKILL 128
193 #define TASK_WAKING 256
194 #define TASK_STATE_MAX 512
196 #define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
198 extern char ___assert_task_state
[1 - 2*!!(
199 sizeof(TASK_STATE_TO_CHAR_STR
)-1 != ilog2(TASK_STATE_MAX
)+1)];
201 /* Convenience macros for the sake of set_task_state */
202 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
203 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
204 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
206 /* Convenience macros for the sake of wake_up */
207 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
208 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
210 /* get_task_state() */
211 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
212 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
215 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
216 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
217 #define task_is_stopped_or_traced(task) \
218 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
219 #define task_contributes_to_load(task) \
220 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
221 (task->flags & PF_FREEZING) == 0)
223 #define __set_task_state(tsk, state_value) \
224 do { (tsk)->state = (state_value); } while (0)
225 #define set_task_state(tsk, state_value) \
226 set_mb((tsk)->state, (state_value))
229 * set_current_state() includes a barrier so that the write of current->state
230 * is correctly serialised wrt the caller's subsequent test of whether to
233 * set_current_state(TASK_UNINTERRUPTIBLE);
234 * if (do_i_need_to_sleep())
237 * If the caller does not need such serialisation then use __set_current_state()
239 #define __set_current_state(state_value) \
240 do { current->state = (state_value); } while (0)
241 #define set_current_state(state_value) \
242 set_mb(current->state, (state_value))
244 /* Task command name length */
245 #define TASK_COMM_LEN 16
247 #include <linux/spinlock.h>
250 * This serializes "schedule()" and also protects
251 * the run-queue from deletions/modifications (but
252 * _adding_ to the beginning of the run-queue has
255 extern rwlock_t tasklist_lock
;
256 extern spinlock_t mmlist_lock
;
260 #ifdef CONFIG_PROVE_RCU
261 extern int lockdep_tasklist_lock_is_held(void);
262 #endif /* #ifdef CONFIG_PROVE_RCU */
264 extern void sched_init(void);
265 extern void sched_init_smp(void);
266 extern asmlinkage
void schedule_tail(struct task_struct
*prev
);
267 extern void init_idle(struct task_struct
*idle
, int cpu
);
268 extern void init_idle_bootup_task(struct task_struct
*idle
);
270 extern int runqueue_is_locked(int cpu
);
271 extern void task_rq_unlock_wait(struct task_struct
*p
);
273 extern cpumask_var_t nohz_cpu_mask
;
274 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
275 extern int select_nohz_load_balancer(int cpu
);
276 extern int get_nohz_load_balancer(void);
277 extern int nohz_ratelimit(int cpu
);
279 static inline int select_nohz_load_balancer(int cpu
)
284 static inline int nohz_ratelimit(int cpu
)
291 * Only dump TASK_* tasks. (0 for all tasks)
293 extern void show_state_filter(unsigned long state_filter
);
295 static inline void show_state(void)
297 show_state_filter(0);
300 extern void show_regs(struct pt_regs
*);
303 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
304 * task), SP is the stack pointer of the first frame that should be shown in the back
305 * trace (or NULL if the entire call-chain of the task should be shown).
307 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
309 void io_schedule(void);
310 long io_schedule_timeout(long timeout
);
312 extern void cpu_init (void);
313 extern void trap_init(void);
314 extern void update_process_times(int user
);
315 extern void scheduler_tick(void);
317 extern void sched_show_task(struct task_struct
*p
);
319 #ifdef CONFIG_DETECT_SOFTLOCKUP
320 extern void softlockup_tick(void);
321 extern void touch_softlockup_watchdog(void);
322 extern void touch_softlockup_watchdog_sync(void);
323 extern void touch_all_softlockup_watchdogs(void);
324 extern int proc_dosoftlockup_thresh(struct ctl_table
*table
, int write
,
326 size_t *lenp
, loff_t
*ppos
);
327 extern unsigned int softlockup_panic
;
328 extern int softlockup_thresh
;
330 static inline void softlockup_tick(void)
333 static inline void touch_softlockup_watchdog(void)
336 static inline void touch_softlockup_watchdog_sync(void)
339 static inline void touch_all_softlockup_watchdogs(void)
344 #ifdef CONFIG_DETECT_HUNG_TASK
345 extern unsigned int sysctl_hung_task_panic
;
346 extern unsigned long sysctl_hung_task_check_count
;
347 extern unsigned long sysctl_hung_task_timeout_secs
;
348 extern unsigned long sysctl_hung_task_warnings
;
349 extern int proc_dohung_task_timeout_secs(struct ctl_table
*table
, int write
,
351 size_t *lenp
, loff_t
*ppos
);
354 /* Attach to any functions which should be ignored in wchan output. */
355 #define __sched __attribute__((__section__(".sched.text")))
357 /* Linker adds these: start and end of __sched functions */
358 extern char __sched_text_start
[], __sched_text_end
[];
360 /* Is this address in the __sched functions? */
361 extern int in_sched_functions(unsigned long addr
);
363 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
364 extern signed long schedule_timeout(signed long timeout
);
365 extern signed long schedule_timeout_interruptible(signed long timeout
);
366 extern signed long schedule_timeout_killable(signed long timeout
);
367 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
368 asmlinkage
void schedule(void);
369 extern int mutex_spin_on_owner(struct mutex
*lock
, struct thread_info
*owner
);
372 struct user_namespace
;
375 * Default maximum number of active map areas, this limits the number of vmas
376 * per mm struct. Users can overwrite this number by sysctl but there is a
379 * When a program's coredump is generated as ELF format, a section is created
380 * per a vma. In ELF, the number of sections is represented in unsigned short.
381 * This means the number of sections should be smaller than 65535 at coredump.
382 * Because the kernel adds some informative sections to a image of program at
383 * generating coredump, we need some margin. The number of extra sections is
384 * 1-3 now and depends on arch. We use "5" as safe margin, here.
386 #define MAPCOUNT_ELF_CORE_MARGIN (5)
387 #define DEFAULT_MAX_MAP_COUNT (USHRT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
389 extern int sysctl_max_map_count
;
391 #include <linux/aio.h>
394 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
396 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
397 unsigned long, unsigned long);
399 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
400 unsigned long len
, unsigned long pgoff
,
401 unsigned long flags
);
402 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
403 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
405 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
) {}
409 extern void set_dumpable(struct mm_struct
*mm
, int value
);
410 extern int get_dumpable(struct mm_struct
*mm
);
414 #define MMF_DUMPABLE 0 /* core dump is permitted */
415 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
417 #define MMF_DUMPABLE_BITS 2
418 #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
420 /* coredump filter bits */
421 #define MMF_DUMP_ANON_PRIVATE 2
422 #define MMF_DUMP_ANON_SHARED 3
423 #define MMF_DUMP_MAPPED_PRIVATE 4
424 #define MMF_DUMP_MAPPED_SHARED 5
425 #define MMF_DUMP_ELF_HEADERS 6
426 #define MMF_DUMP_HUGETLB_PRIVATE 7
427 #define MMF_DUMP_HUGETLB_SHARED 8
429 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
430 #define MMF_DUMP_FILTER_BITS 7
431 #define MMF_DUMP_FILTER_MASK \
432 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
433 #define MMF_DUMP_FILTER_DEFAULT \
434 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
435 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
437 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
438 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
440 # define MMF_DUMP_MASK_DEFAULT_ELF 0
442 /* leave room for more dump flags */
443 #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
445 #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
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
;
470 * struct task_cputime - collected CPU time counts
471 * @utime: time spent in user mode, in &cputime_t units
472 * @stime: time spent in kernel mode, in &cputime_t units
473 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
475 * This structure groups together three kinds of CPU time that are
476 * tracked for threads and thread groups. Most things considering
477 * CPU time want to group these counts together and treat all three
478 * of them in parallel.
480 struct task_cputime
{
483 unsigned long long sum_exec_runtime
;
485 /* Alternate field names when used to cache expirations. */
486 #define prof_exp stime
487 #define virt_exp utime
488 #define sched_exp sum_exec_runtime
490 #define INIT_CPUTIME \
491 (struct task_cputime) { \
492 .utime = cputime_zero, \
493 .stime = cputime_zero, \
494 .sum_exec_runtime = 0, \
498 * Disable preemption until the scheduler is running.
499 * Reset by start_kernel()->sched_init()->init_idle().
501 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
502 * before the scheduler is active -- see should_resched().
504 #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
507 * struct thread_group_cputimer - thread group interval timer counts
508 * @cputime: thread group interval timers.
509 * @running: non-zero when there are timers running and
510 * @cputime receives updates.
511 * @lock: lock for fields in this struct.
513 * This structure contains the version of task_cputime, above, that is
514 * used for thread group CPU timer calculations.
516 struct thread_group_cputimer
{
517 struct task_cputime cputime
;
523 * NOTE! "signal_struct" does not have it's own
524 * locking, because a shared signal_struct always
525 * implies a shared sighand_struct, so locking
526 * sighand_struct is always a proper superset of
527 * the locking of signal_struct.
529 struct signal_struct
{
534 wait_queue_head_t wait_chldexit
; /* for wait4() */
536 /* current thread group signal load-balancing target: */
537 struct task_struct
*curr_target
;
539 /* shared signal handling: */
540 struct sigpending shared_pending
;
542 /* thread group exit support */
545 * - notify group_exit_task when ->count is equal to notify_count
546 * - everyone except group_exit_task is stopped during signal delivery
547 * of fatal signals, group_exit_task processes the signal.
550 struct task_struct
*group_exit_task
;
552 /* thread group stop support, overloads group_exit_code too */
553 int group_stop_count
;
554 unsigned int flags
; /* see SIGNAL_* flags below */
556 /* POSIX.1b Interval Timers */
557 struct list_head posix_timers
;
559 /* ITIMER_REAL timer for the process */
560 struct hrtimer real_timer
;
561 struct pid
*leader_pid
;
562 ktime_t it_real_incr
;
565 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
566 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
567 * values are defined to 0 and 1 respectively
569 struct cpu_itimer it
[2];
572 * Thread group totals for process CPU timers.
573 * See thread_group_cputimer(), et al, for details.
575 struct thread_group_cputimer cputimer
;
577 /* Earliest-expiration cache. */
578 struct task_cputime cputime_expires
;
580 struct list_head cpu_timers
[3];
582 struct pid
*tty_old_pgrp
;
584 /* boolean value for session group leader */
587 struct tty_struct
*tty
; /* NULL if no tty */
590 * Cumulative resource counters for dead threads in the group,
591 * and for reaped dead child processes forked by this group.
592 * Live threads maintain their own counters and add to these
593 * in __exit_signal, except for the group leader.
595 cputime_t utime
, stime
, cutime
, cstime
;
598 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
599 cputime_t prev_utime
, prev_stime
;
601 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
602 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
603 unsigned long inblock
, oublock
, cinblock
, coublock
;
604 unsigned long maxrss
, cmaxrss
;
605 struct task_io_accounting ioac
;
608 * Cumulative ns of schedule CPU time fo dead threads in the
609 * group, not including a zombie group leader, (This only differs
610 * from jiffies_to_ns(utime + stime) if sched_clock uses something
611 * other than jiffies.)
613 unsigned long long sum_sched_runtime
;
616 * We don't bother to synchronize most readers of this at all,
617 * because there is no reader checking a limit that actually needs
618 * to get both rlim_cur and rlim_max atomically, and either one
619 * alone is a single word that can safely be read normally.
620 * getrlimit/setrlimit use task_lock(current->group_leader) to
621 * protect this instead of the siglock, because they really
622 * have no need to disable irqs.
624 struct rlimit rlim
[RLIM_NLIMITS
];
626 #ifdef CONFIG_BSD_PROCESS_ACCT
627 struct pacct_struct pacct
; /* per-process accounting information */
629 #ifdef CONFIG_TASKSTATS
630 struct taskstats
*stats
;
634 struct tty_audit_buf
*tty_audit_buf
;
637 int oom_adj
; /* OOM kill score adjustment (bit shift) */
640 /* Context switch must be unlocked if interrupts are to be enabled */
641 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
642 # define __ARCH_WANT_UNLOCKED_CTXSW
646 * Bits in flags field of signal_struct.
648 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
649 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
650 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
651 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
653 * Pending notifications to parent.
655 #define SIGNAL_CLD_STOPPED 0x00000010
656 #define SIGNAL_CLD_CONTINUED 0x00000020
657 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
659 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
661 /* If true, all threads except ->group_exit_task have pending SIGKILL */
662 static inline int signal_group_exit(const struct signal_struct
*sig
)
664 return (sig
->flags
& SIGNAL_GROUP_EXIT
) ||
665 (sig
->group_exit_task
!= NULL
);
669 * Some day this will be a full-fledged user tracking system..
672 atomic_t __count
; /* reference count */
673 atomic_t processes
; /* How many processes does this user have? */
674 atomic_t files
; /* How many open files does this user have? */
675 atomic_t sigpending
; /* How many pending signals does this user have? */
676 #ifdef CONFIG_INOTIFY_USER
677 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
678 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
681 atomic_t epoll_watches
; /* The number of file descriptors currently watched */
683 #ifdef CONFIG_POSIX_MQUEUE
684 /* protected by mq_lock */
685 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
687 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
690 struct key
*uid_keyring
; /* UID specific keyring */
691 struct key
*session_keyring
; /* UID's default session keyring */
694 /* Hash table maintenance information */
695 struct hlist_node uidhash_node
;
697 struct user_namespace
*user_ns
;
699 #ifdef CONFIG_PERF_EVENTS
700 atomic_long_t locked_vm
;
704 extern int uids_sysfs_init(void);
706 extern struct user_struct
*find_user(uid_t
);
708 extern struct user_struct root_user
;
709 #define INIT_USER (&root_user)
712 struct backing_dev_info
;
713 struct reclaim_state
;
715 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
717 /* cumulative counters */
718 unsigned long pcount
; /* # of times run on this cpu */
719 unsigned long long run_delay
; /* time spent waiting on a runqueue */
722 unsigned long long last_arrival
,/* when we last ran on a cpu */
723 last_queued
; /* when we were last queued to run */
724 #ifdef CONFIG_SCHEDSTATS
726 unsigned int bkl_count
;
729 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
731 #ifdef CONFIG_TASK_DELAY_ACCT
732 struct task_delay_info
{
734 unsigned int flags
; /* Private per-task flags */
736 /* For each stat XXX, add following, aligned appropriately
738 * struct timespec XXX_start, XXX_end;
742 * Atomicity of updates to XXX_delay, XXX_count protected by
743 * single lock above (split into XXX_lock if contention is an issue).
747 * XXX_count is incremented on every XXX operation, the delay
748 * associated with the operation is added to XXX_delay.
749 * XXX_delay contains the accumulated delay time in nanoseconds.
751 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
752 u64 blkio_delay
; /* wait for sync block io completion */
753 u64 swapin_delay
; /* wait for swapin block io completion */
754 u32 blkio_count
; /* total count of the number of sync block */
755 /* io operations performed */
756 u32 swapin_count
; /* total count of the number of swapin block */
757 /* io operations performed */
759 struct timespec freepages_start
, freepages_end
;
760 u64 freepages_delay
; /* wait for memory reclaim */
761 u32 freepages_count
; /* total count of memory reclaim */
763 #endif /* CONFIG_TASK_DELAY_ACCT */
765 static inline int sched_info_on(void)
767 #ifdef CONFIG_SCHEDSTATS
769 #elif defined(CONFIG_TASK_DELAY_ACCT)
770 extern int delayacct_on
;
785 * sched-domains (multiprocessor balancing) declarations:
789 * Increase resolution of nice-level calculations:
791 #define SCHED_LOAD_SHIFT 10
792 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
794 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
797 #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
798 #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
799 #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
800 #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
801 #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
802 #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
803 #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
804 #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
805 #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
806 #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
807 #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
809 #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
811 enum powersavings_balance_level
{
812 POWERSAVINGS_BALANCE_NONE
= 0, /* No power saving load balance */
813 POWERSAVINGS_BALANCE_BASIC
, /* Fill one thread/core/package
814 * first for long running threads
816 POWERSAVINGS_BALANCE_WAKEUP
, /* Also bias task wakeups to semi-idle
817 * cpu package for power savings
819 MAX_POWERSAVINGS_BALANCE_LEVELS
822 extern int sched_mc_power_savings
, sched_smt_power_savings
;
824 static inline int sd_balance_for_mc_power(void)
826 if (sched_smt_power_savings
)
827 return SD_POWERSAVINGS_BALANCE
;
829 if (!sched_mc_power_savings
)
830 return SD_PREFER_SIBLING
;
835 static inline int sd_balance_for_package_power(void)
837 if (sched_mc_power_savings
| sched_smt_power_savings
)
838 return SD_POWERSAVINGS_BALANCE
;
840 return SD_PREFER_SIBLING
;
844 * Optimise SD flags for power savings:
845 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
846 * Keep default SD flags if sched_{smt,mc}_power_saving=0
849 static inline int sd_power_saving_flags(void)
851 if (sched_mc_power_savings
| sched_smt_power_savings
)
852 return SD_BALANCE_NEWIDLE
;
858 struct sched_group
*next
; /* Must be a circular list */
861 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
864 unsigned int cpu_power
;
867 * The CPUs this group covers.
869 * NOTE: this field is variable length. (Allocated dynamically
870 * by attaching extra space to the end of the structure,
871 * depending on how many CPUs the kernel has booted up with)
873 * It is also be embedded into static data structures at build
874 * time. (See 'struct static_sched_group' in kernel/sched.c)
876 unsigned long cpumask
[0];
879 static inline struct cpumask
*sched_group_cpus(struct sched_group
*sg
)
881 return to_cpumask(sg
->cpumask
);
884 enum sched_domain_level
{
894 struct sched_domain_attr
{
895 int relax_domain_level
;
898 #define SD_ATTR_INIT (struct sched_domain_attr) { \
899 .relax_domain_level = -1, \
902 struct sched_domain
{
903 /* These fields must be setup */
904 struct sched_domain
*parent
; /* top domain must be null terminated */
905 struct sched_domain
*child
; /* bottom domain must be null terminated */
906 struct sched_group
*groups
; /* the balancing groups of the domain */
907 unsigned long min_interval
; /* Minimum balance interval ms */
908 unsigned long max_interval
; /* Maximum balance interval ms */
909 unsigned int busy_factor
; /* less balancing by factor if busy */
910 unsigned int imbalance_pct
; /* No balance until over watermark */
911 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
912 unsigned int busy_idx
;
913 unsigned int idle_idx
;
914 unsigned int newidle_idx
;
915 unsigned int wake_idx
;
916 unsigned int forkexec_idx
;
917 unsigned int smt_gain
;
918 int flags
; /* See SD_* */
919 enum sched_domain_level level
;
921 /* Runtime fields. */
922 unsigned long last_balance
; /* init to jiffies. units in jiffies */
923 unsigned int balance_interval
; /* initialise to 1. units in ms. */
924 unsigned int nr_balance_failed
; /* initialise to 0 */
928 #ifdef CONFIG_SCHEDSTATS
929 /* load_balance() stats */
930 unsigned int lb_count
[CPU_MAX_IDLE_TYPES
];
931 unsigned int lb_failed
[CPU_MAX_IDLE_TYPES
];
932 unsigned int lb_balanced
[CPU_MAX_IDLE_TYPES
];
933 unsigned int lb_imbalance
[CPU_MAX_IDLE_TYPES
];
934 unsigned int lb_gained
[CPU_MAX_IDLE_TYPES
];
935 unsigned int lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
936 unsigned int lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
937 unsigned int lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
939 /* Active load balancing */
940 unsigned int alb_count
;
941 unsigned int alb_failed
;
942 unsigned int alb_pushed
;
944 /* SD_BALANCE_EXEC stats */
945 unsigned int sbe_count
;
946 unsigned int sbe_balanced
;
947 unsigned int sbe_pushed
;
949 /* SD_BALANCE_FORK stats */
950 unsigned int sbf_count
;
951 unsigned int sbf_balanced
;
952 unsigned int sbf_pushed
;
954 /* try_to_wake_up() stats */
955 unsigned int ttwu_wake_remote
;
956 unsigned int ttwu_move_affine
;
957 unsigned int ttwu_move_balance
;
959 #ifdef CONFIG_SCHED_DEBUG
963 unsigned int span_weight
;
965 * Span of all CPUs in this domain.
967 * NOTE: this field is variable length. (Allocated dynamically
968 * by attaching extra space to the end of the structure,
969 * depending on how many CPUs the kernel has booted up with)
971 * It is also be embedded into static data structures at build
972 * time. (See 'struct static_sched_domain' in kernel/sched.c)
974 unsigned long span
[0];
977 static inline struct cpumask
*sched_domain_span(struct sched_domain
*sd
)
979 return to_cpumask(sd
->span
);
982 extern void partition_sched_domains(int ndoms_new
, cpumask_var_t doms_new
[],
983 struct sched_domain_attr
*dattr_new
);
985 /* Allocate an array of sched domains, for partition_sched_domains(). */
986 cpumask_var_t
*alloc_sched_domains(unsigned int ndoms
);
987 void free_sched_domains(cpumask_var_t doms
[], unsigned int ndoms
);
989 /* Test a flag in parent sched domain */
990 static inline int test_sd_parent(struct sched_domain
*sd
, int flag
)
992 if (sd
->parent
&& (sd
->parent
->flags
& flag
))
998 unsigned long default_scale_freq_power(struct sched_domain
*sd
, int cpu
);
999 unsigned long default_scale_smt_power(struct sched_domain
*sd
, int cpu
);
1001 #else /* CONFIG_SMP */
1003 struct sched_domain_attr
;
1006 partition_sched_domains(int ndoms_new
, cpumask_var_t doms_new
[],
1007 struct sched_domain_attr
*dattr_new
)
1010 #endif /* !CONFIG_SMP */
1013 struct io_context
; /* See blkdev.h */
1016 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
1017 extern void prefetch_stack(struct task_struct
*t
);
1019 static inline void prefetch_stack(struct task_struct
*t
) { }
1022 struct audit_context
; /* See audit.c */
1024 struct pipe_inode_info
;
1025 struct uts_namespace
;
1028 struct sched_domain
;
1033 #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
1034 #define WF_FORK 0x02 /* child wakeup after fork */
1036 #define ENQUEUE_WAKEUP 1
1037 #define ENQUEUE_WAKING 2
1038 #define ENQUEUE_HEAD 4
1040 #define DEQUEUE_SLEEP 1
1042 struct sched_class
{
1043 const struct sched_class
*next
;
1045 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
, int flags
);
1046 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
, int flags
);
1047 void (*yield_task
) (struct rq
*rq
);
1049 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
, int flags
);
1051 struct task_struct
* (*pick_next_task
) (struct rq
*rq
);
1052 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
);
1055 int (*select_task_rq
)(struct rq
*rq
, struct task_struct
*p
,
1056 int sd_flag
, int flags
);
1058 void (*pre_schedule
) (struct rq
*this_rq
, struct task_struct
*task
);
1059 void (*post_schedule
) (struct rq
*this_rq
);
1060 void (*task_waking
) (struct rq
*this_rq
, struct task_struct
*task
);
1061 void (*task_woken
) (struct rq
*this_rq
, struct task_struct
*task
);
1063 void (*set_cpus_allowed
)(struct task_struct
*p
,
1064 const struct cpumask
*newmask
);
1066 void (*rq_online
)(struct rq
*rq
);
1067 void (*rq_offline
)(struct rq
*rq
);
1070 void (*set_curr_task
) (struct rq
*rq
);
1071 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
, int queued
);
1072 void (*task_fork
) (struct task_struct
*p
);
1074 void (*switched_from
) (struct rq
*this_rq
, struct task_struct
*task
,
1076 void (*switched_to
) (struct rq
*this_rq
, struct task_struct
*task
,
1078 void (*prio_changed
) (struct rq
*this_rq
, struct task_struct
*task
,
1079 int oldprio
, int running
);
1081 unsigned int (*get_rr_interval
) (struct rq
*rq
,
1082 struct task_struct
*task
);
1084 #ifdef CONFIG_FAIR_GROUP_SCHED
1085 void (*moved_group
) (struct task_struct
*p
, int on_rq
);
1089 struct load_weight
{
1090 unsigned long weight
, inv_weight
;
1093 #ifdef CONFIG_SCHEDSTATS
1094 struct sched_statistics
{
1104 s64 sum_sleep_runtime
;
1111 u64 nr_migrations_cold
;
1112 u64 nr_failed_migrations_affine
;
1113 u64 nr_failed_migrations_running
;
1114 u64 nr_failed_migrations_hot
;
1115 u64 nr_forced_migrations
;
1118 u64 nr_wakeups_sync
;
1119 u64 nr_wakeups_migrate
;
1120 u64 nr_wakeups_local
;
1121 u64 nr_wakeups_remote
;
1122 u64 nr_wakeups_affine
;
1123 u64 nr_wakeups_affine_attempts
;
1124 u64 nr_wakeups_passive
;
1125 u64 nr_wakeups_idle
;
1129 struct sched_entity
{
1130 struct load_weight load
; /* for load-balancing */
1131 struct rb_node run_node
;
1132 struct list_head group_node
;
1136 u64 sum_exec_runtime
;
1138 u64 prev_sum_exec_runtime
;
1142 #ifdef CONFIG_SCHEDSTATS
1143 struct sched_statistics statistics
;
1146 #ifdef CONFIG_FAIR_GROUP_SCHED
1147 struct sched_entity
*parent
;
1148 /* rq on which this entity is (to be) queued: */
1149 struct cfs_rq
*cfs_rq
;
1150 /* rq "owned" by this entity/group: */
1151 struct cfs_rq
*my_q
;
1155 struct sched_rt_entity
{
1156 struct list_head run_list
;
1157 unsigned long timeout
;
1158 unsigned int time_slice
;
1159 int nr_cpus_allowed
;
1161 struct sched_rt_entity
*back
;
1162 #ifdef CONFIG_RT_GROUP_SCHED
1163 struct sched_rt_entity
*parent
;
1164 /* rq on which this entity is (to be) queued: */
1165 struct rt_rq
*rt_rq
;
1166 /* rq "owned" by this entity/group: */
1173 struct task_struct
{
1174 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
1177 unsigned int flags
; /* per process flags, defined below */
1178 unsigned int ptrace
;
1180 int lock_depth
; /* BKL lock depth */
1183 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1188 int prio
, static_prio
, normal_prio
;
1189 unsigned int rt_priority
;
1190 const struct sched_class
*sched_class
;
1191 struct sched_entity se
;
1192 struct sched_rt_entity rt
;
1194 #ifdef CONFIG_PREEMPT_NOTIFIERS
1195 /* list of struct preempt_notifier: */
1196 struct hlist_head preempt_notifiers
;
1200 * fpu_counter contains the number of consecutive context switches
1201 * that the FPU is used. If this is over a threshold, the lazy fpu
1202 * saving becomes unlazy to save the trap. This is an unsigned char
1203 * so that after 256 times the counter wraps and the behavior turns
1204 * lazy again; this to deal with bursty apps that only use FPU for
1207 unsigned char fpu_counter
;
1208 #ifdef CONFIG_BLK_DEV_IO_TRACE
1209 unsigned int btrace_seq
;
1212 unsigned int policy
;
1213 cpumask_t cpus_allowed
;
1215 #ifdef CONFIG_TREE_PREEMPT_RCU
1216 int rcu_read_lock_nesting
;
1217 char rcu_read_unlock_special
;
1218 struct rcu_node
*rcu_blocked_node
;
1219 struct list_head rcu_node_entry
;
1220 #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1222 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1223 struct sched_info sched_info
;
1226 struct list_head tasks
;
1227 struct plist_node pushable_tasks
;
1229 struct mm_struct
*mm
, *active_mm
;
1230 #if defined(SPLIT_RSS_COUNTING)
1231 struct task_rss_stat rss_stat
;
1235 int exit_code
, exit_signal
;
1236 int pdeath_signal
; /* The signal sent when the parent dies */
1238 unsigned int personality
;
1239 unsigned did_exec
:1;
1240 unsigned in_execve
:1; /* Tell the LSMs that the process is doing an
1242 unsigned in_iowait
:1;
1245 /* Revert to default priority/policy when forking */
1246 unsigned sched_reset_on_fork
:1;
1251 #ifdef CONFIG_CC_STACKPROTECTOR
1252 /* Canary value for the -fstack-protector gcc feature */
1253 unsigned long stack_canary
;
1257 * pointers to (original) parent process, youngest child, younger sibling,
1258 * older sibling, respectively. (p->father can be replaced with
1259 * p->real_parent->pid)
1261 struct task_struct
*real_parent
; /* real parent process */
1262 struct task_struct
*parent
; /* recipient of SIGCHLD, wait4() reports */
1264 * children/sibling forms the list of my natural children
1266 struct list_head children
; /* list of my children */
1267 struct list_head sibling
; /* linkage in my parent's children list */
1268 struct task_struct
*group_leader
; /* threadgroup leader */
1271 * ptraced is the list of tasks this task is using ptrace on.
1272 * This includes both natural children and PTRACE_ATTACH targets.
1273 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1275 struct list_head ptraced
;
1276 struct list_head ptrace_entry
;
1278 /* PID/PID hash table linkage. */
1279 struct pid_link pids
[PIDTYPE_MAX
];
1280 struct list_head thread_group
;
1282 struct completion
*vfork_done
; /* for vfork() */
1283 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
1284 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
1286 cputime_t utime
, stime
, utimescaled
, stimescaled
;
1288 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
1289 cputime_t prev_utime
, prev_stime
;
1291 unsigned long nvcsw
, nivcsw
; /* context switch counts */
1292 struct timespec start_time
; /* monotonic time */
1293 struct timespec real_start_time
; /* boot based time */
1294 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1295 unsigned long min_flt
, maj_flt
;
1297 struct task_cputime cputime_expires
;
1298 struct list_head cpu_timers
[3];
1300 /* process credentials */
1301 const struct cred
*real_cred
; /* objective and real subjective task
1302 * credentials (COW) */
1303 const struct cred
*cred
; /* effective (overridable) subjective task
1304 * credentials (COW) */
1305 struct mutex cred_guard_mutex
; /* guard against foreign influences on
1306 * credential calculations
1307 * (notably. ptrace) */
1308 struct cred
*replacement_session_keyring
; /* for KEYCTL_SESSION_TO_PARENT */
1310 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1311 - access with [gs]et_task_comm (which lock
1312 it with task_lock())
1313 - initialized normally by setup_new_exec */
1314 /* file system info */
1315 int link_count
, total_link_count
;
1316 #ifdef CONFIG_SYSVIPC
1318 struct sysv_sem sysvsem
;
1320 #ifdef CONFIG_DETECT_HUNG_TASK
1321 /* hung task detection */
1322 unsigned long last_switch_count
;
1324 /* CPU-specific state of this task */
1325 struct thread_struct thread
;
1326 /* filesystem information */
1327 struct fs_struct
*fs
;
1328 /* open file information */
1329 struct files_struct
*files
;
1331 struct nsproxy
*nsproxy
;
1332 /* signal handlers */
1333 struct signal_struct
*signal
;
1334 struct sighand_struct
*sighand
;
1336 sigset_t blocked
, real_blocked
;
1337 sigset_t saved_sigmask
; /* restored if set_restore_sigmask() was used */
1338 struct sigpending pending
;
1340 unsigned long sas_ss_sp
;
1342 int (*notifier
)(void *priv
);
1343 void *notifier_data
;
1344 sigset_t
*notifier_mask
;
1345 struct audit_context
*audit_context
;
1346 #ifdef CONFIG_AUDITSYSCALL
1348 unsigned int sessionid
;
1352 /* Thread group tracking */
1355 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1357 spinlock_t alloc_lock
;
1359 #ifdef CONFIG_GENERIC_HARDIRQS
1360 /* IRQ handler threads */
1361 struct irqaction
*irqaction
;
1364 /* Protection of the PI data structures: */
1365 raw_spinlock_t pi_lock
;
1367 #ifdef CONFIG_RT_MUTEXES
1368 /* PI waiters blocked on a rt_mutex held by this task */
1369 struct plist_head pi_waiters
;
1370 /* Deadlock detection and priority inheritance handling */
1371 struct rt_mutex_waiter
*pi_blocked_on
;
1374 #ifdef CONFIG_DEBUG_MUTEXES
1375 /* mutex deadlock detection */
1376 struct mutex_waiter
*blocked_on
;
1378 #ifdef CONFIG_TRACE_IRQFLAGS
1379 unsigned int irq_events
;
1380 unsigned long hardirq_enable_ip
;
1381 unsigned long hardirq_disable_ip
;
1382 unsigned int hardirq_enable_event
;
1383 unsigned int hardirq_disable_event
;
1384 int hardirqs_enabled
;
1385 int hardirq_context
;
1386 unsigned long softirq_disable_ip
;
1387 unsigned long softirq_enable_ip
;
1388 unsigned int softirq_disable_event
;
1389 unsigned int softirq_enable_event
;
1390 int softirqs_enabled
;
1391 int softirq_context
;
1393 #ifdef CONFIG_LOCKDEP
1394 # define MAX_LOCK_DEPTH 48UL
1397 unsigned int lockdep_recursion
;
1398 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1399 gfp_t lockdep_reclaim_gfp
;
1402 /* journalling filesystem info */
1405 /* stacked block device info */
1406 struct bio_list
*bio_list
;
1409 struct reclaim_state
*reclaim_state
;
1411 struct backing_dev_info
*backing_dev_info
;
1413 struct io_context
*io_context
;
1415 unsigned long ptrace_message
;
1416 siginfo_t
*last_siginfo
; /* For ptrace use. */
1417 struct task_io_accounting ioac
;
1418 #if defined(CONFIG_TASK_XACCT)
1419 u64 acct_rss_mem1
; /* accumulated rss usage */
1420 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1421 cputime_t acct_timexpd
; /* stime + utime since last update */
1423 #ifdef CONFIG_CPUSETS
1424 nodemask_t mems_allowed
; /* Protected by alloc_lock */
1425 int mems_allowed_change_disable
;
1426 int cpuset_mem_spread_rotor
;
1427 int cpuset_slab_spread_rotor
;
1429 #ifdef CONFIG_CGROUPS
1430 /* Control Group info protected by css_set_lock */
1431 struct css_set
*cgroups
;
1432 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1433 struct list_head cg_list
;
1436 struct robust_list_head __user
*robust_list
;
1437 #ifdef CONFIG_COMPAT
1438 struct compat_robust_list_head __user
*compat_robust_list
;
1440 struct list_head pi_state_list
;
1441 struct futex_pi_state
*pi_state_cache
;
1443 #ifdef CONFIG_PERF_EVENTS
1444 struct perf_event_context
*perf_event_ctxp
;
1445 struct mutex perf_event_mutex
;
1446 struct list_head perf_event_list
;
1449 struct mempolicy
*mempolicy
; /* Protected by alloc_lock */
1452 atomic_t fs_excl
; /* holding fs exclusive resources */
1453 struct rcu_head rcu
;
1456 * cache last used pipe for splice
1458 struct pipe_inode_info
*splice_pipe
;
1459 #ifdef CONFIG_TASK_DELAY_ACCT
1460 struct task_delay_info
*delays
;
1462 #ifdef CONFIG_FAULT_INJECTION
1465 struct prop_local_single dirties
;
1466 #ifdef CONFIG_LATENCYTOP
1467 int latency_record_count
;
1468 struct latency_record latency_record
[LT_SAVECOUNT
];
1471 * time slack values; these are used to round up poll() and
1472 * select() etc timeout values. These are in nanoseconds.
1474 unsigned long timer_slack_ns
;
1475 unsigned long default_timer_slack_ns
;
1477 struct list_head
*scm_work_list
;
1478 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1479 /* Index of current stored address in ret_stack */
1481 /* Stack of return addresses for return function tracing */
1482 struct ftrace_ret_stack
*ret_stack
;
1483 /* time stamp for last schedule */
1484 unsigned long long ftrace_timestamp
;
1486 * Number of functions that haven't been traced
1487 * because of depth overrun.
1489 atomic_t trace_overrun
;
1490 /* Pause for the tracing */
1491 atomic_t tracing_graph_pause
;
1493 #ifdef CONFIG_TRACING
1494 /* state flags for use by tracers */
1495 unsigned long trace
;
1496 /* bitmask of trace recursion */
1497 unsigned long trace_recursion
;
1498 #endif /* CONFIG_TRACING */
1499 #ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1500 struct memcg_batch_info
{
1501 int do_batch
; /* incremented when batch uncharge started */
1502 struct mem_cgroup
*memcg
; /* target memcg of uncharge */
1503 unsigned long bytes
; /* uncharged usage */
1504 unsigned long memsw_bytes
; /* uncharged mem+swap usage */
1509 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1510 #define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
1513 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1514 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1515 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1516 * values are inverted: lower p->prio value means higher priority.
1518 * The MAX_USER_RT_PRIO value allows the actual maximum
1519 * RT priority to be separate from the value exported to
1520 * user-space. This allows kernel threads to set their
1521 * priority to a value higher than any user task. Note:
1522 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1525 #define MAX_USER_RT_PRIO 100
1526 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1528 #define MAX_PRIO (MAX_RT_PRIO + 40)
1529 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1531 static inline int rt_prio(int prio
)
1533 if (unlikely(prio
< MAX_RT_PRIO
))
1538 static inline int rt_task(struct task_struct
*p
)
1540 return rt_prio(p
->prio
);
1543 static inline struct pid
*task_pid(struct task_struct
*task
)
1545 return task
->pids
[PIDTYPE_PID
].pid
;
1548 static inline struct pid
*task_tgid(struct task_struct
*task
)
1550 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1554 * Without tasklist or rcu lock it is not safe to dereference
1555 * the result of task_pgrp/task_session even if task == current,
1556 * we can race with another thread doing sys_setsid/sys_setpgid.
1558 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1560 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1563 static inline struct pid
*task_session(struct task_struct
*task
)
1565 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1568 struct pid_namespace
;
1571 * the helpers to get the task's different pids as they are seen
1572 * from various namespaces
1574 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1575 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1577 * task_xid_nr_ns() : id seen from the ns specified;
1579 * set_task_vxid() : assigns a virtual id to a task;
1581 * see also pid_nr() etc in include/linux/pid.h
1583 pid_t
__task_pid_nr_ns(struct task_struct
*task
, enum pid_type type
,
1584 struct pid_namespace
*ns
);
1586 static inline pid_t
task_pid_nr(struct task_struct
*tsk
)
1591 static inline pid_t
task_pid_nr_ns(struct task_struct
*tsk
,
1592 struct pid_namespace
*ns
)
1594 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, ns
);
1597 static inline pid_t
task_pid_vnr(struct task_struct
*tsk
)
1599 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, NULL
);
1603 static inline pid_t
task_tgid_nr(struct task_struct
*tsk
)
1608 pid_t
task_tgid_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1610 static inline pid_t
task_tgid_vnr(struct task_struct
*tsk
)
1612 return pid_vnr(task_tgid(tsk
));
1616 static inline pid_t
task_pgrp_nr_ns(struct task_struct
*tsk
,
1617 struct pid_namespace
*ns
)
1619 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, ns
);
1622 static inline pid_t
task_pgrp_vnr(struct task_struct
*tsk
)
1624 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, NULL
);
1628 static inline pid_t
task_session_nr_ns(struct task_struct
*tsk
,
1629 struct pid_namespace
*ns
)
1631 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, ns
);
1634 static inline pid_t
task_session_vnr(struct task_struct
*tsk
)
1636 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, NULL
);
1639 /* obsolete, do not use */
1640 static inline pid_t
task_pgrp_nr(struct task_struct
*tsk
)
1642 return task_pgrp_nr_ns(tsk
, &init_pid_ns
);
1646 * pid_alive - check that a task structure is not stale
1647 * @p: Task structure to be checked.
1649 * Test if a process is not yet dead (at most zombie state)
1650 * If pid_alive fails, then pointers within the task structure
1651 * can be stale and must not be dereferenced.
1653 static inline int pid_alive(struct task_struct
*p
)
1655 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1659 * is_global_init - check if a task structure is init
1660 * @tsk: Task structure to be checked.
1662 * Check if a task structure is the first user space task the kernel created.
1664 static inline int is_global_init(struct task_struct
*tsk
)
1666 return tsk
->pid
== 1;
1670 * is_container_init:
1671 * check whether in the task is init in its own pid namespace.
1673 extern int is_container_init(struct task_struct
*tsk
);
1675 extern struct pid
*cad_pid
;
1677 extern void free_task(struct task_struct
*tsk
);
1678 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1680 extern void __put_task_struct(struct task_struct
*t
);
1682 static inline void put_task_struct(struct task_struct
*t
)
1684 if (atomic_dec_and_test(&t
->usage
))
1685 __put_task_struct(t
);
1688 extern void task_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1689 extern void thread_group_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1694 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1695 /* Not implemented yet, only for 486*/
1696 #define PF_STARTING 0x00000002 /* being created */
1697 #define PF_EXITING 0x00000004 /* getting shut down */
1698 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1699 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1700 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1701 #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1702 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1703 #define PF_DUMPCORE 0x00000200 /* dumped core */
1704 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1705 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1706 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1707 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1708 #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1709 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1710 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1711 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1712 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1713 #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1714 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1715 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1716 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1717 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1718 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1719 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1720 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1721 #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
1722 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1723 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1724 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1725 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1728 * Only the _current_ task can read/write to tsk->flags, but other
1729 * tasks can access tsk->flags in readonly mode for example
1730 * with tsk_used_math (like during threaded core dumping).
1731 * There is however an exception to this rule during ptrace
1732 * or during fork: the ptracer task is allowed to write to the
1733 * child->flags of its traced child (same goes for fork, the parent
1734 * can write to the child->flags), because we're guaranteed the
1735 * child is not running and in turn not changing child->flags
1736 * at the same time the parent does it.
1738 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1739 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1740 #define clear_used_math() clear_stopped_child_used_math(current)
1741 #define set_used_math() set_stopped_child_used_math(current)
1742 #define conditional_stopped_child_used_math(condition, child) \
1743 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1744 #define conditional_used_math(condition) \
1745 conditional_stopped_child_used_math(condition, current)
1746 #define copy_to_stopped_child_used_math(child) \
1747 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1748 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1749 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1750 #define used_math() tsk_used_math(current)
1752 #ifdef CONFIG_TREE_PREEMPT_RCU
1754 #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1755 #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1757 static inline void rcu_copy_process(struct task_struct
*p
)
1759 p
->rcu_read_lock_nesting
= 0;
1760 p
->rcu_read_unlock_special
= 0;
1761 p
->rcu_blocked_node
= NULL
;
1762 INIT_LIST_HEAD(&p
->rcu_node_entry
);
1767 static inline void rcu_copy_process(struct task_struct
*p
)
1774 extern int set_cpus_allowed_ptr(struct task_struct
*p
,
1775 const struct cpumask
*new_mask
);
1777 static inline int set_cpus_allowed_ptr(struct task_struct
*p
,
1778 const struct cpumask
*new_mask
)
1780 if (!cpumask_test_cpu(0, new_mask
))
1786 #ifndef CONFIG_CPUMASK_OFFSTACK
1787 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1789 return set_cpus_allowed_ptr(p
, &new_mask
);
1794 * Architectures can set this to 1 if they have specified
1795 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1796 * but then during bootup it turns out that sched_clock()
1797 * is reliable after all:
1799 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1800 extern int sched_clock_stable
;
1803 /* ftrace calls sched_clock() directly */
1804 extern unsigned long long notrace
sched_clock(void);
1806 extern void sched_clock_init(void);
1807 extern u64
sched_clock_cpu(int cpu
);
1809 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1810 static inline void sched_clock_tick(void)
1814 static inline void sched_clock_idle_sleep_event(void)
1818 static inline void sched_clock_idle_wakeup_event(u64 delta_ns
)
1822 extern void sched_clock_tick(void);
1823 extern void sched_clock_idle_sleep_event(void);
1824 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1828 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1829 * clock constructed from sched_clock():
1831 extern unsigned long long cpu_clock(int cpu
);
1833 extern unsigned long long
1834 task_sched_runtime(struct task_struct
*task
);
1835 extern unsigned long long thread_group_sched_runtime(struct task_struct
*task
);
1837 /* sched_exec is called by processes performing an exec */
1839 extern void sched_exec(void);
1841 #define sched_exec() {}
1844 extern void sched_clock_idle_sleep_event(void);
1845 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1847 #ifdef CONFIG_HOTPLUG_CPU
1848 extern void move_task_off_dead_cpu(int dead_cpu
, struct task_struct
*p
);
1849 extern void idle_task_exit(void);
1851 static inline void idle_task_exit(void) {}
1854 extern void sched_idle_next(void);
1856 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1857 extern void wake_up_idle_cpu(int cpu
);
1859 static inline void wake_up_idle_cpu(int cpu
) { }
1862 extern unsigned int sysctl_sched_latency
;
1863 extern unsigned int sysctl_sched_min_granularity
;
1864 extern unsigned int sysctl_sched_wakeup_granularity
;
1865 extern unsigned int sysctl_sched_shares_ratelimit
;
1866 extern unsigned int sysctl_sched_shares_thresh
;
1867 extern unsigned int sysctl_sched_child_runs_first
;
1869 enum sched_tunable_scaling
{
1870 SCHED_TUNABLESCALING_NONE
,
1871 SCHED_TUNABLESCALING_LOG
,
1872 SCHED_TUNABLESCALING_LINEAR
,
1873 SCHED_TUNABLESCALING_END
,
1875 extern enum sched_tunable_scaling sysctl_sched_tunable_scaling
;
1877 #ifdef CONFIG_SCHED_DEBUG
1878 extern unsigned int sysctl_sched_migration_cost
;
1879 extern unsigned int sysctl_sched_nr_migrate
;
1880 extern unsigned int sysctl_sched_time_avg
;
1881 extern unsigned int sysctl_timer_migration
;
1883 int sched_proc_update_handler(struct ctl_table
*table
, int write
,
1884 void __user
*buffer
, size_t *length
,
1887 #ifdef CONFIG_SCHED_DEBUG
1888 static inline unsigned int get_sysctl_timer_migration(void)
1890 return sysctl_timer_migration
;
1893 static inline unsigned int get_sysctl_timer_migration(void)
1898 extern unsigned int sysctl_sched_rt_period
;
1899 extern int sysctl_sched_rt_runtime
;
1901 int sched_rt_handler(struct ctl_table
*table
, int write
,
1902 void __user
*buffer
, size_t *lenp
,
1905 extern unsigned int sysctl_sched_compat_yield
;
1907 #ifdef CONFIG_RT_MUTEXES
1908 extern int rt_mutex_getprio(struct task_struct
*p
);
1909 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1910 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1912 static inline int rt_mutex_getprio(struct task_struct
*p
)
1914 return p
->normal_prio
;
1916 # define rt_mutex_adjust_pi(p) do { } while (0)
1919 extern void set_user_nice(struct task_struct
*p
, long nice
);
1920 extern int task_prio(const struct task_struct
*p
);
1921 extern int task_nice(const struct task_struct
*p
);
1922 extern int can_nice(const struct task_struct
*p
, const int nice
);
1923 extern int task_curr(const struct task_struct
*p
);
1924 extern int idle_cpu(int cpu
);
1925 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1926 extern int sched_setscheduler_nocheck(struct task_struct
*, int,
1927 struct sched_param
*);
1928 extern struct task_struct
*idle_task(int cpu
);
1929 extern struct task_struct
*curr_task(int cpu
);
1930 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1935 * The default (Linux) execution domain.
1937 extern struct exec_domain default_exec_domain
;
1939 union thread_union
{
1940 struct thread_info thread_info
;
1941 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1944 #ifndef __HAVE_ARCH_KSTACK_END
1945 static inline int kstack_end(void *addr
)
1947 /* Reliable end of stack detection:
1948 * Some APM bios versions misalign the stack
1950 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1954 extern union thread_union init_thread_union
;
1955 extern struct task_struct init_task
;
1957 extern struct mm_struct init_mm
;
1959 extern struct pid_namespace init_pid_ns
;
1962 * find a task by one of its numerical ids
1964 * find_task_by_pid_ns():
1965 * finds a task by its pid in the specified namespace
1966 * find_task_by_vpid():
1967 * finds a task by its virtual pid
1969 * see also find_vpid() etc in include/linux/pid.h
1972 extern struct task_struct
*find_task_by_vpid(pid_t nr
);
1973 extern struct task_struct
*find_task_by_pid_ns(pid_t nr
,
1974 struct pid_namespace
*ns
);
1976 extern void __set_special_pids(struct pid
*pid
);
1978 /* per-UID process charging. */
1979 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
1980 static inline struct user_struct
*get_uid(struct user_struct
*u
)
1982 atomic_inc(&u
->__count
);
1985 extern void free_uid(struct user_struct
*);
1986 extern void release_uids(struct user_namespace
*ns
);
1988 #include <asm/current.h>
1990 extern void do_timer(unsigned long ticks
);
1992 extern int wake_up_state(struct task_struct
*tsk
, unsigned int state
);
1993 extern int wake_up_process(struct task_struct
*tsk
);
1994 extern void wake_up_new_task(struct task_struct
*tsk
,
1995 unsigned long clone_flags
);
1997 extern void kick_process(struct task_struct
*tsk
);
1999 static inline void kick_process(struct task_struct
*tsk
) { }
2001 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
2002 extern void sched_dead(struct task_struct
*p
);
2004 extern void proc_caches_init(void);
2005 extern void flush_signals(struct task_struct
*);
2006 extern void __flush_signals(struct task_struct
*);
2007 extern void ignore_signals(struct task_struct
*);
2008 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
2009 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
2011 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
2013 unsigned long flags
;
2016 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
2017 ret
= dequeue_signal(tsk
, mask
, info
);
2018 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
2023 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
2025 extern void unblock_all_signals(void);
2026 extern void release_task(struct task_struct
* p
);
2027 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
2028 extern int force_sigsegv(int, struct task_struct
*);
2029 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
2030 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
2031 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
2032 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
2033 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
2034 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
2035 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
2036 extern int do_notify_parent(struct task_struct
*, int);
2037 extern void __wake_up_parent(struct task_struct
*p
, struct task_struct
*parent
);
2038 extern void force_sig(int, struct task_struct
*);
2039 extern int send_sig(int, struct task_struct
*, int);
2040 extern int zap_other_threads(struct task_struct
*p
);
2041 extern struct sigqueue
*sigqueue_alloc(void);
2042 extern void sigqueue_free(struct sigqueue
*);
2043 extern int send_sigqueue(struct sigqueue
*, struct task_struct
*, int group
);
2044 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
2045 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
2047 static inline int kill_cad_pid(int sig
, int priv
)
2049 return kill_pid(cad_pid
, sig
, priv
);
2052 /* These can be the second arg to send_sig_info/send_group_sig_info. */
2053 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
2054 #define SEND_SIG_PRIV ((struct siginfo *) 1)
2055 #define SEND_SIG_FORCED ((struct siginfo *) 2)
2058 * True if we are on the alternate signal stack.
2060 static inline int on_sig_stack(unsigned long sp
)
2062 #ifdef CONFIG_STACK_GROWSUP
2063 return sp
>= current
->sas_ss_sp
&&
2064 sp
- current
->sas_ss_sp
< current
->sas_ss_size
;
2066 return sp
> current
->sas_ss_sp
&&
2067 sp
- current
->sas_ss_sp
<= current
->sas_ss_size
;
2071 static inline int sas_ss_flags(unsigned long sp
)
2073 return (current
->sas_ss_size
== 0 ? SS_DISABLE
2074 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
2078 * Routines for handling mm_structs
2080 extern struct mm_struct
* mm_alloc(void);
2082 /* mmdrop drops the mm and the page tables */
2083 extern void __mmdrop(struct mm_struct
*);
2084 static inline void mmdrop(struct mm_struct
* mm
)
2086 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
2090 /* mmput gets rid of the mappings and all user-space */
2091 extern void mmput(struct mm_struct
*);
2092 /* Grab a reference to a task's mm, if it is not already going away */
2093 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
2094 /* Remove the current tasks stale references to the old mm_struct */
2095 extern void mm_release(struct task_struct
*, struct mm_struct
*);
2096 /* Allocate a new mm structure and copy contents from tsk->mm */
2097 extern struct mm_struct
*dup_mm(struct task_struct
*tsk
);
2099 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2100 struct task_struct
*, struct pt_regs
*);
2101 extern void flush_thread(void);
2102 extern void exit_thread(void);
2104 extern void exit_files(struct task_struct
*);
2105 extern void __cleanup_sighand(struct sighand_struct
*);
2107 extern void exit_itimers(struct signal_struct
*);
2108 extern void flush_itimer_signals(void);
2110 extern NORET_TYPE
void do_group_exit(int);
2112 extern void daemonize(const char *, ...);
2113 extern int allow_signal(int);
2114 extern int disallow_signal(int);
2116 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
2117 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
2118 struct task_struct
*fork_idle(int);
2120 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
2121 extern char *get_task_comm(char *to
, struct task_struct
*tsk
);
2124 extern unsigned long wait_task_inactive(struct task_struct
*, long match_state
);
2126 static inline unsigned long wait_task_inactive(struct task_struct
*p
,
2133 #define next_task(p) \
2134 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2136 #define for_each_process(p) \
2137 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2139 extern bool current_is_single_threaded(void);
2142 * Careful: do_each_thread/while_each_thread is a double loop so
2143 * 'break' will not work as expected - use goto instead.
2145 #define do_each_thread(g, t) \
2146 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2148 #define while_each_thread(g, t) \
2149 while ((t = next_thread(t)) != g)
2151 /* de_thread depends on thread_group_leader not being a pid based check */
2152 #define thread_group_leader(p) (p == p->group_leader)
2154 /* Do to the insanities of de_thread it is possible for a process
2155 * to have the pid of the thread group leader without actually being
2156 * the thread group leader. For iteration through the pids in proc
2157 * all we care about is that we have a task with the appropriate
2158 * pid, we don't actually care if we have the right task.
2160 static inline int has_group_leader_pid(struct task_struct
*p
)
2162 return p
->pid
== p
->tgid
;
2166 int same_thread_group(struct task_struct
*p1
, struct task_struct
*p2
)
2168 return p1
->tgid
== p2
->tgid
;
2171 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
2173 return list_entry_rcu(p
->thread_group
.next
,
2174 struct task_struct
, thread_group
);
2177 static inline int thread_group_empty(struct task_struct
*p
)
2179 return list_empty(&p
->thread_group
);
2182 #define delay_group_leader(p) \
2183 (thread_group_leader(p) && !thread_group_empty(p))
2185 static inline int task_detached(struct task_struct
*p
)
2187 return p
->exit_signal
== -1;
2191 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2192 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2193 * pins the final release of task.io_context. Also protects ->cpuset and
2194 * ->cgroup.subsys[].
2196 * Nests both inside and outside of read_lock(&tasklist_lock).
2197 * It must not be nested with write_lock_irq(&tasklist_lock),
2198 * neither inside nor outside.
2200 static inline void task_lock(struct task_struct
*p
)
2202 spin_lock(&p
->alloc_lock
);
2205 static inline void task_unlock(struct task_struct
*p
)
2207 spin_unlock(&p
->alloc_lock
);
2210 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
2211 unsigned long *flags
);
2213 static inline void unlock_task_sighand(struct task_struct
*tsk
,
2214 unsigned long *flags
)
2216 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
2219 #ifndef __HAVE_THREAD_FUNCTIONS
2221 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2222 #define task_stack_page(task) ((task)->stack)
2224 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
2226 *task_thread_info(p
) = *task_thread_info(org
);
2227 task_thread_info(p
)->task
= p
;
2230 static inline unsigned long *end_of_stack(struct task_struct
*p
)
2232 return (unsigned long *)(task_thread_info(p
) + 1);
2237 static inline int object_is_on_stack(void *obj
)
2239 void *stack
= task_stack_page(current
);
2241 return (obj
>= stack
) && (obj
< (stack
+ THREAD_SIZE
));
2244 extern void thread_info_cache_init(void);
2246 #ifdef CONFIG_DEBUG_STACK_USAGE
2247 static inline unsigned long stack_not_used(struct task_struct
*p
)
2249 unsigned long *n
= end_of_stack(p
);
2251 do { /* Skip over canary */
2255 return (unsigned long)n
- (unsigned long)end_of_stack(p
);
2259 /* set thread flags in other task's structures
2260 * - see asm/thread_info.h for TIF_xxxx flags available
2262 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2264 set_ti_thread_flag(task_thread_info(tsk
), flag
);
2267 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2269 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2272 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2274 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
2277 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2279 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2282 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2284 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
2287 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
2289 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2292 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
2294 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2297 static inline int test_tsk_need_resched(struct task_struct
*tsk
)
2299 return unlikely(test_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
));
2302 static inline int restart_syscall(void)
2304 set_tsk_thread_flag(current
, TIF_SIGPENDING
);
2305 return -ERESTARTNOINTR
;
2308 static inline int signal_pending(struct task_struct
*p
)
2310 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
2313 static inline int __fatal_signal_pending(struct task_struct
*p
)
2315 return unlikely(sigismember(&p
->pending
.signal
, SIGKILL
));
2318 static inline int fatal_signal_pending(struct task_struct
*p
)
2320 return signal_pending(p
) && __fatal_signal_pending(p
);
2323 static inline int signal_pending_state(long state
, struct task_struct
*p
)
2325 if (!(state
& (TASK_INTERRUPTIBLE
| TASK_WAKEKILL
)))
2327 if (!signal_pending(p
))
2330 return (state
& TASK_INTERRUPTIBLE
) || __fatal_signal_pending(p
);
2333 static inline int need_resched(void)
2335 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
2339 * cond_resched() and cond_resched_lock(): latency reduction via
2340 * explicit rescheduling in places that are safe. The return
2341 * value indicates whether a reschedule was done in fact.
2342 * cond_resched_lock() will drop the spinlock before scheduling,
2343 * cond_resched_softirq() will enable bhs before scheduling.
2345 extern int _cond_resched(void);
2347 #define cond_resched() ({ \
2348 __might_sleep(__FILE__, __LINE__, 0); \
2352 extern int __cond_resched_lock(spinlock_t
*lock
);
2354 #ifdef CONFIG_PREEMPT
2355 #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
2357 #define PREEMPT_LOCK_OFFSET 0
2360 #define cond_resched_lock(lock) ({ \
2361 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
2362 __cond_resched_lock(lock); \
2365 extern int __cond_resched_softirq(void);
2367 #define cond_resched_softirq() ({ \
2368 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2369 __cond_resched_softirq(); \
2373 * Does a critical section need to be broken due to another
2374 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2375 * but a general need for low latency)
2377 static inline int spin_needbreak(spinlock_t
*lock
)
2379 #ifdef CONFIG_PREEMPT
2380 return spin_is_contended(lock
);
2387 * Thread group CPU time accounting.
2389 void thread_group_cputime(struct task_struct
*tsk
, struct task_cputime
*times
);
2390 void thread_group_cputimer(struct task_struct
*tsk
, struct task_cputime
*times
);
2392 static inline void thread_group_cputime_init(struct signal_struct
*sig
)
2394 spin_lock_init(&sig
->cputimer
.lock
);
2397 static inline void thread_group_cputime_free(struct signal_struct
*sig
)
2402 * Reevaluate whether the task has signals pending delivery.
2403 * Wake the task if so.
2404 * This is required every time the blocked sigset_t changes.
2405 * callers must hold sighand->siglock.
2407 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
2408 extern void recalc_sigpending(void);
2410 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
2413 * Wrappers for p->thread_info->cpu access. No-op on UP.
2417 static inline unsigned int task_cpu(const struct task_struct
*p
)
2419 return task_thread_info(p
)->cpu
;
2422 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
2426 static inline unsigned int task_cpu(const struct task_struct
*p
)
2431 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
2435 #endif /* CONFIG_SMP */
2437 #ifdef CONFIG_TRACING
2439 __trace_special(void *__tr
, void *__data
,
2440 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
);
2443 __trace_special(void *__tr
, void *__data
,
2444 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
)
2449 extern long sched_setaffinity(pid_t pid
, const struct cpumask
*new_mask
);
2450 extern long sched_getaffinity(pid_t pid
, struct cpumask
*mask
);
2452 extern void normalize_rt_tasks(void);
2454 #ifdef CONFIG_CGROUP_SCHED
2456 extern struct task_group init_task_group
;
2458 extern struct task_group
*sched_create_group(struct task_group
*parent
);
2459 extern void sched_destroy_group(struct task_group
*tg
);
2460 extern void sched_move_task(struct task_struct
*tsk
);
2461 #ifdef CONFIG_FAIR_GROUP_SCHED
2462 extern int sched_group_set_shares(struct task_group
*tg
, unsigned long shares
);
2463 extern unsigned long sched_group_shares(struct task_group
*tg
);
2465 #ifdef CONFIG_RT_GROUP_SCHED
2466 extern int sched_group_set_rt_runtime(struct task_group
*tg
,
2467 long rt_runtime_us
);
2468 extern long sched_group_rt_runtime(struct task_group
*tg
);
2469 extern int sched_group_set_rt_period(struct task_group
*tg
,
2471 extern long sched_group_rt_period(struct task_group
*tg
);
2472 extern int sched_rt_can_attach(struct task_group
*tg
, struct task_struct
*tsk
);
2476 extern int task_can_switch_user(struct user_struct
*up
,
2477 struct task_struct
*tsk
);
2479 #ifdef CONFIG_TASK_XACCT
2480 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2482 tsk
->ioac
.rchar
+= amt
;
2485 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2487 tsk
->ioac
.wchar
+= amt
;
2490 static inline void inc_syscr(struct task_struct
*tsk
)
2495 static inline void inc_syscw(struct task_struct
*tsk
)
2500 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2504 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2508 static inline void inc_syscr(struct task_struct
*tsk
)
2512 static inline void inc_syscw(struct task_struct
*tsk
)
2517 #ifndef TASK_SIZE_OF
2518 #define TASK_SIZE_OF(tsk) TASK_SIZE
2522 * Call the function if the target task is executing on a CPU right now:
2524 extern void task_oncpu_function_call(struct task_struct
*p
,
2525 void (*func
) (void *info
), void *info
);
2528 #ifdef CONFIG_MM_OWNER
2529 extern void mm_update_next_owner(struct mm_struct
*mm
);
2530 extern void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
);
2532 static inline void mm_update_next_owner(struct mm_struct
*mm
)
2536 static inline void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
)
2539 #endif /* CONFIG_MM_OWNER */
2541 static inline unsigned long task_rlimit(const struct task_struct
*tsk
,
2544 return ACCESS_ONCE(tsk
->signal
->rlim
[limit
].rlim_cur
);
2547 static inline unsigned long task_rlimit_max(const struct task_struct
*tsk
,
2550 return ACCESS_ONCE(tsk
->signal
->rlim
[limit
].rlim_max
);
2553 static inline unsigned long rlimit(unsigned int limit
)
2555 return task_rlimit(current
, limit
);
2558 static inline unsigned long rlimit_max(unsigned int limit
)
2560 return task_rlimit_max(current
, limit
);
2563 #endif /* __KERNEL__ */