7 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
9 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15 #define CLONE_THREAD 0x00010000 /* Same thread group? */
16 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
17 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
24 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
25 #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
26 #define CLONE_NEWIPC 0x08000000 /* New ipcs */
27 #define CLONE_NEWUSER 0x10000000 /* New user namespace */
28 #define CLONE_NEWPID 0x20000000 /* New pid namespace */
29 #define CLONE_NEWNET 0x40000000 /* New network namespace */
30 #define CLONE_IO 0x80000000 /* Clone io context */
35 #define SCHED_NORMAL 0
39 /* SCHED_ISO: reserved but not implemented yet */
48 #include <asm/param.h> /* for HZ */
50 #include <linux/capability.h>
51 #include <linux/threads.h>
52 #include <linux/kernel.h>
53 #include <linux/types.h>
54 #include <linux/timex.h>
55 #include <linux/jiffies.h>
56 #include <linux/rbtree.h>
57 #include <linux/thread_info.h>
58 #include <linux/cpumask.h>
59 #include <linux/errno.h>
60 #include <linux/nodemask.h>
61 #include <linux/mm_types.h>
63 #include <asm/system.h>
65 #include <asm/ptrace.h>
66 #include <asm/cputime.h>
68 #include <linux/smp.h>
69 #include <linux/sem.h>
70 #include <linux/signal.h>
71 #include <linux/fs_struct.h>
72 #include <linux/compiler.h>
73 #include <linux/completion.h>
74 #include <linux/pid.h>
75 #include <linux/percpu.h>
76 #include <linux/topology.h>
77 #include <linux/proportions.h>
78 #include <linux/seccomp.h>
79 #include <linux/rcupdate.h>
80 #include <linux/rtmutex.h>
82 #include <linux/time.h>
83 #include <linux/param.h>
84 #include <linux/resource.h>
85 #include <linux/timer.h>
86 #include <linux/hrtimer.h>
87 #include <linux/task_io_accounting.h>
88 #include <linux/kobject.h>
89 #include <linux/latencytop.h>
90 #include <linux/cred.h>
92 #include <asm/processor.h>
96 struct futex_pi_state
;
97 struct robust_list_head
;
102 * List of flags we want to share for kernel threads,
103 * if only because they are not used by them anyway.
105 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
108 * These are the constant used to fake the fixed-point load-average
109 * counting. Some notes:
110 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
111 * a load-average precision of 10 bits integer + 11 bits fractional
112 * - if you want to count load-averages more often, you need more
113 * precision, or rounding will get you. With 2-second counting freq,
114 * the EXP_n values would be 1981, 2034 and 2043 if still using only
117 extern unsigned long avenrun
[]; /* Load averages */
119 #define FSHIFT 11 /* nr of bits of precision */
120 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
121 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
122 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
123 #define EXP_5 2014 /* 1/exp(5sec/5min) */
124 #define EXP_15 2037 /* 1/exp(5sec/15min) */
126 #define CALC_LOAD(load,exp,n) \
128 load += n*(FIXED_1-exp); \
131 extern unsigned long total_forks
;
132 extern int nr_threads
;
133 DECLARE_PER_CPU(unsigned long, process_counts
);
134 extern int nr_processes(void);
135 extern unsigned long nr_running(void);
136 extern unsigned long nr_uninterruptible(void);
137 extern unsigned long nr_active(void);
138 extern unsigned long nr_iowait(void);
143 #ifdef CONFIG_SCHED_DEBUG
144 extern void proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
);
145 extern void proc_sched_set_task(struct task_struct
*p
);
147 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
);
150 proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
)
153 static inline void proc_sched_set_task(struct task_struct
*p
)
157 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
)
162 extern unsigned long long time_sync_thresh
;
165 * Task state bitmask. NOTE! These bits are also
166 * encoded in fs/proc/array.c: get_task_state().
168 * We have two separate sets of flags: task->state
169 * is about runnability, while task->exit_state are
170 * about the task exiting. Confusing, but this way
171 * modifying one set can't modify the other one by
174 #define TASK_RUNNING 0
175 #define TASK_INTERRUPTIBLE 1
176 #define TASK_UNINTERRUPTIBLE 2
177 #define __TASK_STOPPED 4
178 #define __TASK_TRACED 8
179 /* in tsk->exit_state */
180 #define EXIT_ZOMBIE 16
182 /* in tsk->state again */
184 #define TASK_WAKEKILL 128
186 /* Convenience macros for the sake of set_task_state */
187 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
188 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
189 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
191 /* Convenience macros for the sake of wake_up */
192 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
193 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
195 /* get_task_state() */
196 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
197 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
200 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
201 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
202 #define task_is_stopped_or_traced(task) \
203 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
204 #define task_contributes_to_load(task) \
205 ((task->state & TASK_UNINTERRUPTIBLE) != 0)
207 #define __set_task_state(tsk, state_value) \
208 do { (tsk)->state = (state_value); } while (0)
209 #define set_task_state(tsk, state_value) \
210 set_mb((tsk)->state, (state_value))
213 * set_current_state() includes a barrier so that the write of current->state
214 * is correctly serialised wrt the caller's subsequent test of whether to
217 * set_current_state(TASK_UNINTERRUPTIBLE);
218 * if (do_i_need_to_sleep())
221 * If the caller does not need such serialisation then use __set_current_state()
223 #define __set_current_state(state_value) \
224 do { current->state = (state_value); } while (0)
225 #define set_current_state(state_value) \
226 set_mb(current->state, (state_value))
228 /* Task command name length */
229 #define TASK_COMM_LEN 16
231 #include <linux/spinlock.h>
234 * This serializes "schedule()" and also protects
235 * the run-queue from deletions/modifications (but
236 * _adding_ to the beginning of the run-queue has
239 extern rwlock_t tasklist_lock
;
240 extern spinlock_t mmlist_lock
;
244 extern void sched_init(void);
245 extern void sched_init_smp(void);
246 extern asmlinkage
void schedule_tail(struct task_struct
*prev
);
247 extern void init_idle(struct task_struct
*idle
, int cpu
);
248 extern void init_idle_bootup_task(struct task_struct
*idle
);
250 extern int runqueue_is_locked(void);
251 extern void task_rq_unlock_wait(struct task_struct
*p
);
253 extern cpumask_t nohz_cpu_mask
;
254 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
255 extern int select_nohz_load_balancer(int cpu
);
257 static inline int select_nohz_load_balancer(int cpu
)
263 extern unsigned long rt_needs_cpu(int cpu
);
266 * Only dump TASK_* tasks. (0 for all tasks)
268 extern void show_state_filter(unsigned long state_filter
);
270 static inline void show_state(void)
272 show_state_filter(0);
275 extern void show_regs(struct pt_regs
*);
278 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
279 * task), SP is the stack pointer of the first frame that should be shown in the back
280 * trace (or NULL if the entire call-chain of the task should be shown).
282 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
284 void io_schedule(void);
285 long io_schedule_timeout(long timeout
);
287 extern void cpu_init (void);
288 extern void trap_init(void);
289 extern void account_process_tick(struct task_struct
*task
, int user
);
290 extern void update_process_times(int user
);
291 extern void scheduler_tick(void);
293 extern void sched_show_task(struct task_struct
*p
);
295 #ifdef CONFIG_DETECT_SOFTLOCKUP
296 extern void softlockup_tick(void);
297 extern void touch_softlockup_watchdog(void);
298 extern void touch_all_softlockup_watchdogs(void);
299 extern unsigned int softlockup_panic
;
300 extern unsigned long sysctl_hung_task_check_count
;
301 extern unsigned long sysctl_hung_task_timeout_secs
;
302 extern unsigned long sysctl_hung_task_warnings
;
303 extern int softlockup_thresh
;
305 static inline void softlockup_tick(void)
308 static inline void spawn_softlockup_task(void)
311 static inline void touch_softlockup_watchdog(void)
314 static inline void touch_all_softlockup_watchdogs(void)
320 /* Attach to any functions which should be ignored in wchan output. */
321 #define __sched __attribute__((__section__(".sched.text")))
323 /* Linker adds these: start and end of __sched functions */
324 extern char __sched_text_start
[], __sched_text_end
[];
326 /* Is this address in the __sched functions? */
327 extern int in_sched_functions(unsigned long addr
);
329 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
330 extern signed long schedule_timeout(signed long timeout
);
331 extern signed long schedule_timeout_interruptible(signed long timeout
);
332 extern signed long schedule_timeout_killable(signed long timeout
);
333 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
334 asmlinkage
void schedule(void);
337 struct user_namespace
;
339 /* Maximum number of active map areas.. This is a random (large) number */
340 #define DEFAULT_MAX_MAP_COUNT 65536
342 extern int sysctl_max_map_count
;
344 #include <linux/aio.h>
347 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
348 unsigned long, unsigned long);
350 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
351 unsigned long len
, unsigned long pgoff
,
352 unsigned long flags
);
353 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
354 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
356 #if USE_SPLIT_PTLOCKS
358 * The mm counters are not protected by its page_table_lock,
359 * so must be incremented atomically.
361 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
362 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
363 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
364 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
365 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
367 #else /* !USE_SPLIT_PTLOCKS */
369 * The mm counters are protected by its page_table_lock,
370 * so can be incremented directly.
372 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
373 #define get_mm_counter(mm, member) ((mm)->_##member)
374 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
375 #define inc_mm_counter(mm, member) (mm)->_##member++
376 #define dec_mm_counter(mm, member) (mm)->_##member--
378 #endif /* !USE_SPLIT_PTLOCKS */
380 #define get_mm_rss(mm) \
381 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
382 #define update_hiwater_rss(mm) do { \
383 unsigned long _rss = get_mm_rss(mm); \
384 if ((mm)->hiwater_rss < _rss) \
385 (mm)->hiwater_rss = _rss; \
387 #define update_hiwater_vm(mm) do { \
388 if ((mm)->hiwater_vm < (mm)->total_vm) \
389 (mm)->hiwater_vm = (mm)->total_vm; \
392 extern void set_dumpable(struct mm_struct
*mm
, int value
);
393 extern int get_dumpable(struct mm_struct
*mm
);
397 #define MMF_DUMPABLE 0 /* core dump is permitted */
398 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
399 #define MMF_DUMPABLE_BITS 2
401 /* coredump filter bits */
402 #define MMF_DUMP_ANON_PRIVATE 2
403 #define MMF_DUMP_ANON_SHARED 3
404 #define MMF_DUMP_MAPPED_PRIVATE 4
405 #define MMF_DUMP_MAPPED_SHARED 5
406 #define MMF_DUMP_ELF_HEADERS 6
407 #define MMF_DUMP_HUGETLB_PRIVATE 7
408 #define MMF_DUMP_HUGETLB_SHARED 8
409 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
410 #define MMF_DUMP_FILTER_BITS 7
411 #define MMF_DUMP_FILTER_MASK \
412 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
413 #define MMF_DUMP_FILTER_DEFAULT \
414 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
415 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
417 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
418 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
420 # define MMF_DUMP_MASK_DEFAULT_ELF 0
423 struct sighand_struct
{
425 struct k_sigaction action
[_NSIG
];
427 wait_queue_head_t signalfd_wqh
;
430 struct pacct_struct
{
433 unsigned long ac_mem
;
434 cputime_t ac_utime
, ac_stime
;
435 unsigned long ac_minflt
, ac_majflt
;
439 * struct task_cputime - collected CPU time counts
440 * @utime: time spent in user mode, in &cputime_t units
441 * @stime: time spent in kernel mode, in &cputime_t units
442 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
444 * This structure groups together three kinds of CPU time that are
445 * tracked for threads and thread groups. Most things considering
446 * CPU time want to group these counts together and treat all three
447 * of them in parallel.
449 struct task_cputime
{
452 unsigned long long sum_exec_runtime
;
454 /* Alternate field names when used to cache expirations. */
455 #define prof_exp stime
456 #define virt_exp utime
457 #define sched_exp sum_exec_runtime
460 * struct thread_group_cputime - thread group interval timer counts
461 * @totals: thread group interval timers; substructure for
462 * uniprocessor kernel, per-cpu for SMP kernel.
464 * This structure contains the version of task_cputime, above, that is
465 * used for thread group CPU clock calculations.
467 struct thread_group_cputime
{
468 struct task_cputime
*totals
;
472 * NOTE! "signal_struct" does not have it's own
473 * locking, because a shared signal_struct always
474 * implies a shared sighand_struct, so locking
475 * sighand_struct is always a proper superset of
476 * the locking of signal_struct.
478 struct signal_struct
{
482 wait_queue_head_t wait_chldexit
; /* for wait4() */
484 /* current thread group signal load-balancing target: */
485 struct task_struct
*curr_target
;
487 /* shared signal handling: */
488 struct sigpending shared_pending
;
490 /* thread group exit support */
493 * - notify group_exit_task when ->count is equal to notify_count
494 * - everyone except group_exit_task is stopped during signal delivery
495 * of fatal signals, group_exit_task processes the signal.
498 struct task_struct
*group_exit_task
;
500 /* thread group stop support, overloads group_exit_code too */
501 int group_stop_count
;
502 unsigned int flags
; /* see SIGNAL_* flags below */
504 /* POSIX.1b Interval Timers */
505 struct list_head posix_timers
;
507 /* ITIMER_REAL timer for the process */
508 struct hrtimer real_timer
;
509 struct pid
*leader_pid
;
510 ktime_t it_real_incr
;
512 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
513 cputime_t it_prof_expires
, it_virt_expires
;
514 cputime_t it_prof_incr
, it_virt_incr
;
517 * Thread group totals for process CPU clocks.
518 * See thread_group_cputime(), et al, for details.
520 struct thread_group_cputime cputime
;
522 /* Earliest-expiration cache. */
523 struct task_cputime cputime_expires
;
525 struct list_head cpu_timers
[3];
527 /* job control IDs */
530 * pgrp and session fields are deprecated.
531 * use the task_session_Xnr and task_pgrp_Xnr routines below
535 pid_t pgrp __deprecated
;
539 struct pid
*tty_old_pgrp
;
542 pid_t session __deprecated
;
546 /* boolean value for session group leader */
549 struct tty_struct
*tty
; /* NULL if no tty */
552 * Cumulative resource counters for dead threads in the group,
553 * and for reaped dead child processes forked by this group.
554 * Live threads maintain their own counters and add to these
555 * in __exit_signal, except for the group leader.
557 cputime_t cutime
, cstime
;
560 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
561 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
562 unsigned long inblock
, oublock
, cinblock
, coublock
;
563 struct task_io_accounting ioac
;
566 * We don't bother to synchronize most readers of this at all,
567 * because there is no reader checking a limit that actually needs
568 * to get both rlim_cur and rlim_max atomically, and either one
569 * alone is a single word that can safely be read normally.
570 * getrlimit/setrlimit use task_lock(current->group_leader) to
571 * protect this instead of the siglock, because they really
572 * have no need to disable irqs.
574 struct rlimit rlim
[RLIM_NLIMITS
];
576 /* keep the process-shared keyrings here so that they do the right
577 * thing in threads created with CLONE_THREAD */
579 struct key
*session_keyring
; /* keyring inherited over fork */
580 struct key
*process_keyring
; /* keyring private to this process */
582 #ifdef CONFIG_BSD_PROCESS_ACCT
583 struct pacct_struct pacct
; /* per-process accounting information */
585 #ifdef CONFIG_TASKSTATS
586 struct taskstats
*stats
;
590 struct tty_audit_buf
*tty_audit_buf
;
594 /* Context switch must be unlocked if interrupts are to be enabled */
595 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
596 # define __ARCH_WANT_UNLOCKED_CTXSW
600 * Bits in flags field of signal_struct.
602 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
603 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
604 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
605 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
607 * Pending notifications to parent.
609 #define SIGNAL_CLD_STOPPED 0x00000010
610 #define SIGNAL_CLD_CONTINUED 0x00000020
611 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
613 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
615 /* If true, all threads except ->group_exit_task have pending SIGKILL */
616 static inline int signal_group_exit(const struct signal_struct
*sig
)
618 return (sig
->flags
& SIGNAL_GROUP_EXIT
) ||
619 (sig
->group_exit_task
!= NULL
);
623 * Some day this will be a full-fledged user tracking system..
626 atomic_t __count
; /* reference count */
627 atomic_t processes
; /* How many processes does this user have? */
628 atomic_t files
; /* How many open files does this user have? */
629 atomic_t sigpending
; /* How many pending signals does this user have? */
630 #ifdef CONFIG_INOTIFY_USER
631 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
632 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
635 atomic_t epoll_devs
; /* The number of epoll descriptors currently open */
636 atomic_t epoll_watches
; /* The number of file descriptors currently watched */
638 #ifdef CONFIG_POSIX_MQUEUE
639 /* protected by mq_lock */
640 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
642 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
645 struct key
*uid_keyring
; /* UID specific keyring */
646 struct key
*session_keyring
; /* UID's default session keyring */
649 /* Hash table maintenance information */
650 struct hlist_node uidhash_node
;
653 #ifdef CONFIG_USER_SCHED
654 struct task_group
*tg
;
657 struct work_struct work
;
662 extern int uids_sysfs_init(void);
664 extern struct user_struct
*find_user(uid_t
);
666 extern struct user_struct root_user
;
667 #define INIT_USER (&root_user)
669 struct backing_dev_info
;
670 struct reclaim_state
;
672 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
674 /* cumulative counters */
675 unsigned long pcount
; /* # of times run on this cpu */
676 unsigned long long cpu_time
, /* time spent on the cpu */
677 run_delay
; /* time spent waiting on a runqueue */
680 unsigned long long last_arrival
,/* when we last ran on a cpu */
681 last_queued
; /* when we were last queued to run */
682 #ifdef CONFIG_SCHEDSTATS
684 unsigned int bkl_count
;
687 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
689 #ifdef CONFIG_TASK_DELAY_ACCT
690 struct task_delay_info
{
692 unsigned int flags
; /* Private per-task flags */
694 /* For each stat XXX, add following, aligned appropriately
696 * struct timespec XXX_start, XXX_end;
700 * Atomicity of updates to XXX_delay, XXX_count protected by
701 * single lock above (split into XXX_lock if contention is an issue).
705 * XXX_count is incremented on every XXX operation, the delay
706 * associated with the operation is added to XXX_delay.
707 * XXX_delay contains the accumulated delay time in nanoseconds.
709 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
710 u64 blkio_delay
; /* wait for sync block io completion */
711 u64 swapin_delay
; /* wait for swapin block io completion */
712 u32 blkio_count
; /* total count of the number of sync block */
713 /* io operations performed */
714 u32 swapin_count
; /* total count of the number of swapin block */
715 /* io operations performed */
717 struct timespec freepages_start
, freepages_end
;
718 u64 freepages_delay
; /* wait for memory reclaim */
719 u32 freepages_count
; /* total count of memory reclaim */
721 #endif /* CONFIG_TASK_DELAY_ACCT */
723 static inline int sched_info_on(void)
725 #ifdef CONFIG_SCHEDSTATS
727 #elif defined(CONFIG_TASK_DELAY_ACCT)
728 extern int delayacct_on
;
743 * sched-domains (multiprocessor balancing) declarations:
747 * Increase resolution of nice-level calculations:
749 #define SCHED_LOAD_SHIFT 10
750 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
752 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
755 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
756 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
757 #define SD_BALANCE_EXEC 4 /* Balance on exec */
758 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
759 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
760 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
761 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
762 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
763 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
764 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
765 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
766 #define SD_WAKE_IDLE_FAR 2048 /* Gain latency sacrificing cache hit */
768 #define BALANCE_FOR_MC_POWER \
769 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
771 #define BALANCE_FOR_PKG_POWER \
772 ((sched_mc_power_savings || sched_smt_power_savings) ? \
773 SD_POWERSAVINGS_BALANCE : 0)
775 #define test_sd_parent(sd, flag) ((sd->parent && \
776 (sd->parent->flags & flag)) ? 1 : 0)
780 struct sched_group
*next
; /* Must be a circular list */
784 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
785 * single CPU. This is read only (except for setup, hotplug CPU).
786 * Note : Never change cpu_power without recompute its reciprocal
788 unsigned int __cpu_power
;
790 * reciprocal value of cpu_power to avoid expensive divides
791 * (see include/linux/reciprocal_div.h)
793 u32 reciprocal_cpu_power
;
796 enum sched_domain_level
{
806 struct sched_domain_attr
{
807 int relax_domain_level
;
810 #define SD_ATTR_INIT (struct sched_domain_attr) { \
811 .relax_domain_level = -1, \
814 struct sched_domain
{
815 /* These fields must be setup */
816 struct sched_domain
*parent
; /* top domain must be null terminated */
817 struct sched_domain
*child
; /* bottom domain must be null terminated */
818 struct sched_group
*groups
; /* the balancing groups of the domain */
819 cpumask_t span
; /* span of all CPUs in this domain */
820 unsigned long min_interval
; /* Minimum balance interval ms */
821 unsigned long max_interval
; /* Maximum balance interval ms */
822 unsigned int busy_factor
; /* less balancing by factor if busy */
823 unsigned int imbalance_pct
; /* No balance until over watermark */
824 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
825 unsigned int busy_idx
;
826 unsigned int idle_idx
;
827 unsigned int newidle_idx
;
828 unsigned int wake_idx
;
829 unsigned int forkexec_idx
;
830 int flags
; /* See SD_* */
831 enum sched_domain_level level
;
833 /* Runtime fields. */
834 unsigned long last_balance
; /* init to jiffies. units in jiffies */
835 unsigned int balance_interval
; /* initialise to 1. units in ms. */
836 unsigned int nr_balance_failed
; /* initialise to 0 */
840 #ifdef CONFIG_SCHEDSTATS
841 /* load_balance() stats */
842 unsigned int lb_count
[CPU_MAX_IDLE_TYPES
];
843 unsigned int lb_failed
[CPU_MAX_IDLE_TYPES
];
844 unsigned int lb_balanced
[CPU_MAX_IDLE_TYPES
];
845 unsigned int lb_imbalance
[CPU_MAX_IDLE_TYPES
];
846 unsigned int lb_gained
[CPU_MAX_IDLE_TYPES
];
847 unsigned int lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
848 unsigned int lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
849 unsigned int lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
851 /* Active load balancing */
852 unsigned int alb_count
;
853 unsigned int alb_failed
;
854 unsigned int alb_pushed
;
856 /* SD_BALANCE_EXEC stats */
857 unsigned int sbe_count
;
858 unsigned int sbe_balanced
;
859 unsigned int sbe_pushed
;
861 /* SD_BALANCE_FORK stats */
862 unsigned int sbf_count
;
863 unsigned int sbf_balanced
;
864 unsigned int sbf_pushed
;
866 /* try_to_wake_up() stats */
867 unsigned int ttwu_wake_remote
;
868 unsigned int ttwu_move_affine
;
869 unsigned int ttwu_move_balance
;
871 #ifdef CONFIG_SCHED_DEBUG
876 extern void partition_sched_domains(int ndoms_new
, cpumask_t
*doms_new
,
877 struct sched_domain_attr
*dattr_new
);
878 extern int arch_reinit_sched_domains(void);
880 #else /* CONFIG_SMP */
882 struct sched_domain_attr
;
885 partition_sched_domains(int ndoms_new
, cpumask_t
*doms_new
,
886 struct sched_domain_attr
*dattr_new
)
889 #endif /* !CONFIG_SMP */
891 struct io_context
; /* See blkdev.h */
892 #define NGROUPS_SMALL 32
893 #define NGROUPS_PER_BLOCK ((unsigned int)(PAGE_SIZE / sizeof(gid_t)))
897 gid_t small_block
[NGROUPS_SMALL
];
903 * get_group_info() must be called with the owning task locked (via task_lock())
904 * when task != current. The reason being that the vast majority of callers are
905 * looking at current->group_info, which can not be changed except by the
906 * current task. Changing current->group_info requires the task lock, too.
908 #define get_group_info(group_info) do { \
909 atomic_inc(&(group_info)->usage); \
912 #define put_group_info(group_info) do { \
913 if (atomic_dec_and_test(&(group_info)->usage)) \
914 groups_free(group_info); \
917 extern struct group_info
*groups_alloc(int gidsetsize
);
918 extern void groups_free(struct group_info
*group_info
);
919 extern int set_current_groups(struct group_info
*group_info
);
920 extern int groups_search(struct group_info
*group_info
, gid_t grp
);
921 /* access the groups "array" with this macro */
922 #define GROUP_AT(gi, i) \
923 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
925 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
926 extern void prefetch_stack(struct task_struct
*t
);
928 static inline void prefetch_stack(struct task_struct
*t
) { }
931 struct audit_context
; /* See audit.c */
933 struct pipe_inode_info
;
934 struct uts_namespace
;
940 const struct sched_class
*next
;
942 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
, int wakeup
);
943 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
, int sleep
);
944 void (*yield_task
) (struct rq
*rq
);
946 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
, int sync
);
948 struct task_struct
* (*pick_next_task
) (struct rq
*rq
);
949 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
);
952 int (*select_task_rq
)(struct task_struct
*p
, int sync
);
954 unsigned long (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
955 struct rq
*busiest
, unsigned long max_load_move
,
956 struct sched_domain
*sd
, enum cpu_idle_type idle
,
957 int *all_pinned
, int *this_best_prio
);
959 int (*move_one_task
) (struct rq
*this_rq
, int this_cpu
,
960 struct rq
*busiest
, struct sched_domain
*sd
,
961 enum cpu_idle_type idle
);
962 void (*pre_schedule
) (struct rq
*this_rq
, struct task_struct
*task
);
963 void (*post_schedule
) (struct rq
*this_rq
);
964 void (*task_wake_up
) (struct rq
*this_rq
, struct task_struct
*task
);
966 void (*set_cpus_allowed
)(struct task_struct
*p
,
967 const cpumask_t
*newmask
);
969 void (*rq_online
)(struct rq
*rq
);
970 void (*rq_offline
)(struct rq
*rq
);
973 void (*set_curr_task
) (struct rq
*rq
);
974 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
, int queued
);
975 void (*task_new
) (struct rq
*rq
, struct task_struct
*p
);
977 void (*switched_from
) (struct rq
*this_rq
, struct task_struct
*task
,
979 void (*switched_to
) (struct rq
*this_rq
, struct task_struct
*task
,
981 void (*prio_changed
) (struct rq
*this_rq
, struct task_struct
*task
,
982 int oldprio
, int running
);
984 #ifdef CONFIG_FAIR_GROUP_SCHED
985 void (*moved_group
) (struct task_struct
*p
);
990 unsigned long weight
, inv_weight
;
994 * CFS stats for a schedulable entity (task, task-group etc)
996 * Current field usage histogram:
1003 struct sched_entity
{
1004 struct load_weight load
; /* for load-balancing */
1005 struct rb_node run_node
;
1006 struct list_head group_node
;
1010 u64 sum_exec_runtime
;
1012 u64 prev_sum_exec_runtime
;
1017 #ifdef CONFIG_SCHEDSTATS
1025 s64 sum_sleep_runtime
;
1033 u64 nr_migrations_cold
;
1034 u64 nr_failed_migrations_affine
;
1035 u64 nr_failed_migrations_running
;
1036 u64 nr_failed_migrations_hot
;
1037 u64 nr_forced_migrations
;
1038 u64 nr_forced2_migrations
;
1041 u64 nr_wakeups_sync
;
1042 u64 nr_wakeups_migrate
;
1043 u64 nr_wakeups_local
;
1044 u64 nr_wakeups_remote
;
1045 u64 nr_wakeups_affine
;
1046 u64 nr_wakeups_affine_attempts
;
1047 u64 nr_wakeups_passive
;
1048 u64 nr_wakeups_idle
;
1051 #ifdef CONFIG_FAIR_GROUP_SCHED
1052 struct sched_entity
*parent
;
1053 /* rq on which this entity is (to be) queued: */
1054 struct cfs_rq
*cfs_rq
;
1055 /* rq "owned" by this entity/group: */
1056 struct cfs_rq
*my_q
;
1060 struct sched_rt_entity
{
1061 struct list_head run_list
;
1062 unsigned long timeout
;
1063 unsigned int time_slice
;
1064 int nr_cpus_allowed
;
1066 struct sched_rt_entity
*back
;
1067 #ifdef CONFIG_RT_GROUP_SCHED
1068 struct sched_rt_entity
*parent
;
1069 /* rq on which this entity is (to be) queued: */
1070 struct rt_rq
*rt_rq
;
1071 /* rq "owned" by this entity/group: */
1076 struct task_struct
{
1077 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
1080 unsigned int flags
; /* per process flags, defined below */
1081 unsigned int ptrace
;
1083 int lock_depth
; /* BKL lock depth */
1086 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1091 int prio
, static_prio
, normal_prio
;
1092 unsigned int rt_priority
;
1093 const struct sched_class
*sched_class
;
1094 struct sched_entity se
;
1095 struct sched_rt_entity rt
;
1097 #ifdef CONFIG_PREEMPT_NOTIFIERS
1098 /* list of struct preempt_notifier: */
1099 struct hlist_head preempt_notifiers
;
1103 * fpu_counter contains the number of consecutive context switches
1104 * that the FPU is used. If this is over a threshold, the lazy fpu
1105 * saving becomes unlazy to save the trap. This is an unsigned char
1106 * so that after 256 times the counter wraps and the behavior turns
1107 * lazy again; this to deal with bursty apps that only use FPU for
1110 unsigned char fpu_counter
;
1111 s8 oomkilladj
; /* OOM kill score adjustment (bit shift). */
1112 #ifdef CONFIG_BLK_DEV_IO_TRACE
1113 unsigned int btrace_seq
;
1116 unsigned int policy
;
1117 cpumask_t cpus_allowed
;
1119 #ifdef CONFIG_PREEMPT_RCU
1120 int rcu_read_lock_nesting
;
1121 int rcu_flipctr_idx
;
1122 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1124 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1125 struct sched_info sched_info
;
1128 struct list_head tasks
;
1130 struct mm_struct
*mm
, *active_mm
;
1133 struct linux_binfmt
*binfmt
;
1135 int exit_code
, exit_signal
;
1136 int pdeath_signal
; /* The signal sent when the parent dies */
1138 unsigned int personality
;
1139 unsigned did_exec
:1;
1143 #ifdef CONFIG_CC_STACKPROTECTOR
1144 /* Canary value for the -fstack-protector gcc feature */
1145 unsigned long stack_canary
;
1148 * pointers to (original) parent process, youngest child, younger sibling,
1149 * older sibling, respectively. (p->father can be replaced with
1150 * p->real_parent->pid)
1152 struct task_struct
*real_parent
; /* real parent process */
1153 struct task_struct
*parent
; /* recipient of SIGCHLD, wait4() reports */
1155 * children/sibling forms the list of my natural children
1157 struct list_head children
; /* list of my children */
1158 struct list_head sibling
; /* linkage in my parent's children list */
1159 struct task_struct
*group_leader
; /* threadgroup leader */
1162 * ptraced is the list of tasks this task is using ptrace on.
1163 * This includes both natural children and PTRACE_ATTACH targets.
1164 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1166 struct list_head ptraced
;
1167 struct list_head ptrace_entry
;
1169 #ifdef CONFIG_X86_PTRACE_BTS
1171 * This is the tracer handle for the ptrace BTS extension.
1172 * This field actually belongs to the ptracer task.
1174 struct bts_tracer
*bts
;
1176 * The buffer to hold the BTS data.
1179 #endif /* CONFIG_X86_PTRACE_BTS */
1181 /* PID/PID hash table linkage. */
1182 struct pid_link pids
[PIDTYPE_MAX
];
1183 struct list_head thread_group
;
1185 struct completion
*vfork_done
; /* for vfork() */
1186 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
1187 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
1189 cputime_t utime
, stime
, utimescaled
, stimescaled
;
1191 cputime_t prev_utime
, prev_stime
;
1192 unsigned long nvcsw
, nivcsw
; /* context switch counts */
1193 struct timespec start_time
; /* monotonic time */
1194 struct timespec real_start_time
; /* boot based time */
1195 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1196 unsigned long min_flt
, maj_flt
;
1198 struct task_cputime cputime_expires
;
1199 struct list_head cpu_timers
[3];
1201 /* process credentials */
1202 uid_t uid
,euid
,suid
,fsuid
;
1203 gid_t gid
,egid
,sgid
,fsgid
;
1204 struct group_info
*group_info
;
1205 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
, cap_bset
;
1206 struct user_struct
*user
;
1207 unsigned securebits
;
1209 unsigned char jit_keyring
; /* default keyring to attach requested keys to */
1210 struct key
*request_key_auth
; /* assumed request_key authority */
1211 struct key
*thread_keyring
; /* keyring private to this thread */
1213 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1214 - access with [gs]et_task_comm (which lock
1215 it with task_lock())
1216 - initialized normally by flush_old_exec */
1217 /* file system info */
1218 int link_count
, total_link_count
;
1219 #ifdef CONFIG_SYSVIPC
1221 struct sysv_sem sysvsem
;
1223 #ifdef CONFIG_DETECT_SOFTLOCKUP
1224 /* hung task detection */
1225 unsigned long last_switch_timestamp
;
1226 unsigned long last_switch_count
;
1228 /* CPU-specific state of this task */
1229 struct thread_struct thread
;
1230 /* filesystem information */
1231 struct fs_struct
*fs
;
1232 /* open file information */
1233 struct files_struct
*files
;
1235 struct nsproxy
*nsproxy
;
1236 /* signal handlers */
1237 struct signal_struct
*signal
;
1238 struct sighand_struct
*sighand
;
1240 sigset_t blocked
, real_blocked
;
1241 sigset_t saved_sigmask
; /* restored if set_restore_sigmask() was used */
1242 struct sigpending pending
;
1244 unsigned long sas_ss_sp
;
1246 int (*notifier
)(void *priv
);
1247 void *notifier_data
;
1248 sigset_t
*notifier_mask
;
1249 #ifdef CONFIG_SECURITY
1252 struct audit_context
*audit_context
;
1253 #ifdef CONFIG_AUDITSYSCALL
1255 unsigned int sessionid
;
1259 /* Thread group tracking */
1262 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1263 spinlock_t alloc_lock
;
1265 /* Protection of the PI data structures: */
1268 #ifdef CONFIG_RT_MUTEXES
1269 /* PI waiters blocked on a rt_mutex held by this task */
1270 struct plist_head pi_waiters
;
1271 /* Deadlock detection and priority inheritance handling */
1272 struct rt_mutex_waiter
*pi_blocked_on
;
1275 #ifdef CONFIG_DEBUG_MUTEXES
1276 /* mutex deadlock detection */
1277 struct mutex_waiter
*blocked_on
;
1279 #ifdef CONFIG_TRACE_IRQFLAGS
1280 unsigned int irq_events
;
1281 int hardirqs_enabled
;
1282 unsigned long hardirq_enable_ip
;
1283 unsigned int hardirq_enable_event
;
1284 unsigned long hardirq_disable_ip
;
1285 unsigned int hardirq_disable_event
;
1286 int softirqs_enabled
;
1287 unsigned long softirq_disable_ip
;
1288 unsigned int softirq_disable_event
;
1289 unsigned long softirq_enable_ip
;
1290 unsigned int softirq_enable_event
;
1291 int hardirq_context
;
1292 int softirq_context
;
1294 #ifdef CONFIG_LOCKDEP
1295 # define MAX_LOCK_DEPTH 48UL
1298 unsigned int lockdep_recursion
;
1299 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1302 /* journalling filesystem info */
1305 /* stacked block device info */
1306 struct bio
*bio_list
, **bio_tail
;
1309 struct reclaim_state
*reclaim_state
;
1311 struct backing_dev_info
*backing_dev_info
;
1313 struct io_context
*io_context
;
1315 unsigned long ptrace_message
;
1316 siginfo_t
*last_siginfo
; /* For ptrace use. */
1317 struct task_io_accounting ioac
;
1318 #if defined(CONFIG_TASK_XACCT)
1319 u64 acct_rss_mem1
; /* accumulated rss usage */
1320 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1321 cputime_t acct_timexpd
; /* stime + utime since last update */
1323 #ifdef CONFIG_CPUSETS
1324 nodemask_t mems_allowed
;
1325 int cpuset_mems_generation
;
1326 int cpuset_mem_spread_rotor
;
1328 #ifdef CONFIG_CGROUPS
1329 /* Control Group info protected by css_set_lock */
1330 struct css_set
*cgroups
;
1331 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1332 struct list_head cg_list
;
1335 struct robust_list_head __user
*robust_list
;
1336 #ifdef CONFIG_COMPAT
1337 struct compat_robust_list_head __user
*compat_robust_list
;
1339 struct list_head pi_state_list
;
1340 struct futex_pi_state
*pi_state_cache
;
1343 struct mempolicy
*mempolicy
;
1346 atomic_t fs_excl
; /* holding fs exclusive resources */
1347 struct rcu_head rcu
;
1350 * cache last used pipe for splice
1352 struct pipe_inode_info
*splice_pipe
;
1353 #ifdef CONFIG_TASK_DELAY_ACCT
1354 struct task_delay_info
*delays
;
1356 #ifdef CONFIG_FAULT_INJECTION
1359 struct prop_local_single dirties
;
1360 #ifdef CONFIG_LATENCYTOP
1361 int latency_record_count
;
1362 struct latency_record latency_record
[LT_SAVECOUNT
];
1365 * time slack values; these are used to round up poll() and
1366 * select() etc timeout values. These are in nanoseconds.
1368 unsigned long timer_slack_ns
;
1369 unsigned long default_timer_slack_ns
;
1371 struct list_head
*scm_work_list
;
1372 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1373 /* Index of current stored adress in ret_stack */
1375 /* Stack of return addresses for return function tracing */
1376 struct ftrace_ret_stack
*ret_stack
;
1378 * Number of functions that haven't been traced
1379 * because of depth overrun.
1381 atomic_t trace_overrun
;
1383 #ifdef CONFIG_TRACING
1384 /* state flags for use by tracers */
1385 unsigned long trace
;
1390 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1391 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1392 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1393 * values are inverted: lower p->prio value means higher priority.
1395 * The MAX_USER_RT_PRIO value allows the actual maximum
1396 * RT priority to be separate from the value exported to
1397 * user-space. This allows kernel threads to set their
1398 * priority to a value higher than any user task. Note:
1399 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1402 #define MAX_USER_RT_PRIO 100
1403 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1405 #define MAX_PRIO (MAX_RT_PRIO + 40)
1406 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1408 static inline int rt_prio(int prio
)
1410 if (unlikely(prio
< MAX_RT_PRIO
))
1415 static inline int rt_task(struct task_struct
*p
)
1417 return rt_prio(p
->prio
);
1420 static inline void set_task_session(struct task_struct
*tsk
, pid_t session
)
1422 tsk
->signal
->__session
= session
;
1425 static inline void set_task_pgrp(struct task_struct
*tsk
, pid_t pgrp
)
1427 tsk
->signal
->__pgrp
= pgrp
;
1430 static inline struct pid
*task_pid(struct task_struct
*task
)
1432 return task
->pids
[PIDTYPE_PID
].pid
;
1435 static inline struct pid
*task_tgid(struct task_struct
*task
)
1437 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1440 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1442 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1445 static inline struct pid
*task_session(struct task_struct
*task
)
1447 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1450 struct pid_namespace
;
1453 * the helpers to get the task's different pids as they are seen
1454 * from various namespaces
1456 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1457 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1459 * task_xid_nr_ns() : id seen from the ns specified;
1461 * set_task_vxid() : assigns a virtual id to a task;
1463 * see also pid_nr() etc in include/linux/pid.h
1466 static inline pid_t
task_pid_nr(struct task_struct
*tsk
)
1471 pid_t
task_pid_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1473 static inline pid_t
task_pid_vnr(struct task_struct
*tsk
)
1475 return pid_vnr(task_pid(tsk
));
1479 static inline pid_t
task_tgid_nr(struct task_struct
*tsk
)
1484 pid_t
task_tgid_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1486 static inline pid_t
task_tgid_vnr(struct task_struct
*tsk
)
1488 return pid_vnr(task_tgid(tsk
));
1492 static inline pid_t
task_pgrp_nr(struct task_struct
*tsk
)
1494 return tsk
->signal
->__pgrp
;
1497 pid_t
task_pgrp_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1499 static inline pid_t
task_pgrp_vnr(struct task_struct
*tsk
)
1501 return pid_vnr(task_pgrp(tsk
));
1505 static inline pid_t
task_session_nr(struct task_struct
*tsk
)
1507 return tsk
->signal
->__session
;
1510 pid_t
task_session_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1512 static inline pid_t
task_session_vnr(struct task_struct
*tsk
)
1514 return pid_vnr(task_session(tsk
));
1519 * pid_alive - check that a task structure is not stale
1520 * @p: Task structure to be checked.
1522 * Test if a process is not yet dead (at most zombie state)
1523 * If pid_alive fails, then pointers within the task structure
1524 * can be stale and must not be dereferenced.
1526 static inline int pid_alive(struct task_struct
*p
)
1528 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1532 * is_global_init - check if a task structure is init
1533 * @tsk: Task structure to be checked.
1535 * Check if a task structure is the first user space task the kernel created.
1537 static inline int is_global_init(struct task_struct
*tsk
)
1539 return tsk
->pid
== 1;
1543 * is_container_init:
1544 * check whether in the task is init in its own pid namespace.
1546 extern int is_container_init(struct task_struct
*tsk
);
1548 extern struct pid
*cad_pid
;
1550 extern void free_task(struct task_struct
*tsk
);
1551 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1553 extern void __put_task_struct(struct task_struct
*t
);
1555 static inline void put_task_struct(struct task_struct
*t
)
1557 if (atomic_dec_and_test(&t
->usage
))
1558 __put_task_struct(t
);
1561 extern cputime_t
task_utime(struct task_struct
*p
);
1562 extern cputime_t
task_stime(struct task_struct
*p
);
1563 extern cputime_t
task_gtime(struct task_struct
*p
);
1568 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1569 /* Not implemented yet, only for 486*/
1570 #define PF_STARTING 0x00000002 /* being created */
1571 #define PF_EXITING 0x00000004 /* getting shut down */
1572 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1573 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1574 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1575 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1576 #define PF_DUMPCORE 0x00000200 /* dumped core */
1577 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1578 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1579 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1580 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1581 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1582 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1583 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1584 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1585 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1586 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1587 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1588 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1589 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1590 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1591 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1592 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1593 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1594 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1595 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1596 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1599 * Only the _current_ task can read/write to tsk->flags, but other
1600 * tasks can access tsk->flags in readonly mode for example
1601 * with tsk_used_math (like during threaded core dumping).
1602 * There is however an exception to this rule during ptrace
1603 * or during fork: the ptracer task is allowed to write to the
1604 * child->flags of its traced child (same goes for fork, the parent
1605 * can write to the child->flags), because we're guaranteed the
1606 * child is not running and in turn not changing child->flags
1607 * at the same time the parent does it.
1609 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1610 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1611 #define clear_used_math() clear_stopped_child_used_math(current)
1612 #define set_used_math() set_stopped_child_used_math(current)
1613 #define conditional_stopped_child_used_math(condition, child) \
1614 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1615 #define conditional_used_math(condition) \
1616 conditional_stopped_child_used_math(condition, current)
1617 #define copy_to_stopped_child_used_math(child) \
1618 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1619 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1620 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1621 #define used_math() tsk_used_math(current)
1624 extern int set_cpus_allowed_ptr(struct task_struct
*p
,
1625 const cpumask_t
*new_mask
);
1627 static inline int set_cpus_allowed_ptr(struct task_struct
*p
,
1628 const cpumask_t
*new_mask
)
1630 if (!cpu_isset(0, *new_mask
))
1635 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1637 return set_cpus_allowed_ptr(p
, &new_mask
);
1640 extern unsigned long long sched_clock(void);
1642 extern void sched_clock_init(void);
1643 extern u64
sched_clock_cpu(int cpu
);
1645 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1646 static inline void sched_clock_tick(void)
1650 static inline void sched_clock_idle_sleep_event(void)
1654 static inline void sched_clock_idle_wakeup_event(u64 delta_ns
)
1658 extern void sched_clock_tick(void);
1659 extern void sched_clock_idle_sleep_event(void);
1660 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1664 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1665 * clock constructed from sched_clock():
1667 extern unsigned long long cpu_clock(int cpu
);
1669 extern unsigned long long
1670 task_sched_runtime(struct task_struct
*task
);
1671 extern unsigned long long thread_group_sched_runtime(struct task_struct
*task
);
1673 /* sched_exec is called by processes performing an exec */
1675 extern void sched_exec(void);
1677 #define sched_exec() {}
1680 extern void sched_clock_idle_sleep_event(void);
1681 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1683 #ifdef CONFIG_HOTPLUG_CPU
1684 extern void idle_task_exit(void);
1686 static inline void idle_task_exit(void) {}
1689 extern void sched_idle_next(void);
1691 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1692 extern void wake_up_idle_cpu(int cpu
);
1694 static inline void wake_up_idle_cpu(int cpu
) { }
1697 #ifdef CONFIG_SCHED_DEBUG
1698 extern unsigned int sysctl_sched_latency
;
1699 extern unsigned int sysctl_sched_min_granularity
;
1700 extern unsigned int sysctl_sched_wakeup_granularity
;
1701 extern unsigned int sysctl_sched_child_runs_first
;
1702 extern unsigned int sysctl_sched_features
;
1703 extern unsigned int sysctl_sched_migration_cost
;
1704 extern unsigned int sysctl_sched_nr_migrate
;
1705 extern unsigned int sysctl_sched_shares_ratelimit
;
1706 extern unsigned int sysctl_sched_shares_thresh
;
1708 int sched_nr_latency_handler(struct ctl_table
*table
, int write
,
1709 struct file
*file
, void __user
*buffer
, size_t *length
,
1712 extern unsigned int sysctl_sched_rt_period
;
1713 extern int sysctl_sched_rt_runtime
;
1715 int sched_rt_handler(struct ctl_table
*table
, int write
,
1716 struct file
*filp
, void __user
*buffer
, size_t *lenp
,
1719 extern unsigned int sysctl_sched_compat_yield
;
1721 #ifdef CONFIG_RT_MUTEXES
1722 extern int rt_mutex_getprio(struct task_struct
*p
);
1723 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1724 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1726 static inline int rt_mutex_getprio(struct task_struct
*p
)
1728 return p
->normal_prio
;
1730 # define rt_mutex_adjust_pi(p) do { } while (0)
1733 extern void set_user_nice(struct task_struct
*p
, long nice
);
1734 extern int task_prio(const struct task_struct
*p
);
1735 extern int task_nice(const struct task_struct
*p
);
1736 extern int can_nice(const struct task_struct
*p
, const int nice
);
1737 extern int task_curr(const struct task_struct
*p
);
1738 extern int idle_cpu(int cpu
);
1739 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1740 extern int sched_setscheduler_nocheck(struct task_struct
*, int,
1741 struct sched_param
*);
1742 extern struct task_struct
*idle_task(int cpu
);
1743 extern struct task_struct
*curr_task(int cpu
);
1744 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1749 * The default (Linux) execution domain.
1751 extern struct exec_domain default_exec_domain
;
1753 union thread_union
{
1754 struct thread_info thread_info
;
1755 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1758 #ifndef __HAVE_ARCH_KSTACK_END
1759 static inline int kstack_end(void *addr
)
1761 /* Reliable end of stack detection:
1762 * Some APM bios versions misalign the stack
1764 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1768 extern union thread_union init_thread_union
;
1769 extern struct task_struct init_task
;
1771 extern struct mm_struct init_mm
;
1773 extern struct pid_namespace init_pid_ns
;
1776 * find a task by one of its numerical ids
1778 * find_task_by_pid_type_ns():
1779 * it is the most generic call - it finds a task by all id,
1780 * type and namespace specified
1781 * find_task_by_pid_ns():
1782 * finds a task by its pid in the specified namespace
1783 * find_task_by_vpid():
1784 * finds a task by its virtual pid
1786 * see also find_vpid() etc in include/linux/pid.h
1789 extern struct task_struct
*find_task_by_pid_type_ns(int type
, int pid
,
1790 struct pid_namespace
*ns
);
1792 extern struct task_struct
*find_task_by_vpid(pid_t nr
);
1793 extern struct task_struct
*find_task_by_pid_ns(pid_t nr
,
1794 struct pid_namespace
*ns
);
1796 extern void __set_special_pids(struct pid
*pid
);
1798 /* per-UID process charging. */
1799 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
1800 static inline struct user_struct
*get_uid(struct user_struct
*u
)
1802 atomic_inc(&u
->__count
);
1805 extern void free_uid(struct user_struct
*);
1806 extern void switch_uid(struct user_struct
*);
1807 extern void release_uids(struct user_namespace
*ns
);
1809 #include <asm/current.h>
1811 extern void do_timer(unsigned long ticks
);
1813 extern int wake_up_state(struct task_struct
*tsk
, unsigned int state
);
1814 extern int wake_up_process(struct task_struct
*tsk
);
1815 extern void wake_up_new_task(struct task_struct
*tsk
,
1816 unsigned long clone_flags
);
1818 extern void kick_process(struct task_struct
*tsk
);
1820 static inline void kick_process(struct task_struct
*tsk
) { }
1822 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
1823 extern void sched_dead(struct task_struct
*p
);
1825 extern int in_group_p(gid_t
);
1826 extern int in_egroup_p(gid_t
);
1828 extern void proc_caches_init(void);
1829 extern void flush_signals(struct task_struct
*);
1830 extern void ignore_signals(struct task_struct
*);
1831 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
1832 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
1834 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
1836 unsigned long flags
;
1839 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
1840 ret
= dequeue_signal(tsk
, mask
, info
);
1841 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
1846 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
1848 extern void unblock_all_signals(void);
1849 extern void release_task(struct task_struct
* p
);
1850 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
1851 extern int force_sigsegv(int, struct task_struct
*);
1852 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
1853 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1854 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
1855 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
1856 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
1857 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
1858 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
1859 extern int do_notify_parent(struct task_struct
*, int);
1860 extern void force_sig(int, struct task_struct
*);
1861 extern void force_sig_specific(int, struct task_struct
*);
1862 extern int send_sig(int, struct task_struct
*, int);
1863 extern void zap_other_threads(struct task_struct
*p
);
1864 extern struct sigqueue
*sigqueue_alloc(void);
1865 extern void sigqueue_free(struct sigqueue
*);
1866 extern int send_sigqueue(struct sigqueue
*, struct task_struct
*, int group
);
1867 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
1868 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
1870 static inline int kill_cad_pid(int sig
, int priv
)
1872 return kill_pid(cad_pid
, sig
, priv
);
1875 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1876 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1877 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1878 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1880 static inline int is_si_special(const struct siginfo
*info
)
1882 return info
<= SEND_SIG_FORCED
;
1885 /* True if we are on the alternate signal stack. */
1887 static inline int on_sig_stack(unsigned long sp
)
1889 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
1892 static inline int sas_ss_flags(unsigned long sp
)
1894 return (current
->sas_ss_size
== 0 ? SS_DISABLE
1895 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
1899 * Routines for handling mm_structs
1901 extern struct mm_struct
* mm_alloc(void);
1903 /* mmdrop drops the mm and the page tables */
1904 extern void __mmdrop(struct mm_struct
*);
1905 static inline void mmdrop(struct mm_struct
* mm
)
1907 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
1911 /* mmput gets rid of the mappings and all user-space */
1912 extern void mmput(struct mm_struct
*);
1913 /* Grab a reference to a task's mm, if it is not already going away */
1914 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1915 /* Remove the current tasks stale references to the old mm_struct */
1916 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1917 /* Allocate a new mm structure and copy contents from tsk->mm */
1918 extern struct mm_struct
*dup_mm(struct task_struct
*tsk
);
1920 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1921 extern void flush_thread(void);
1922 extern void exit_thread(void);
1924 extern void exit_files(struct task_struct
*);
1925 extern void __cleanup_signal(struct signal_struct
*);
1926 extern void __cleanup_sighand(struct sighand_struct
*);
1928 extern void exit_itimers(struct signal_struct
*);
1929 extern void flush_itimer_signals(void);
1931 extern NORET_TYPE
void do_group_exit(int);
1933 extern void daemonize(const char *, ...);
1934 extern int allow_signal(int);
1935 extern int disallow_signal(int);
1937 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1938 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1939 struct task_struct
*fork_idle(int);
1941 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1942 extern char *get_task_comm(char *to
, struct task_struct
*tsk
);
1945 extern unsigned long wait_task_inactive(struct task_struct
*, long match_state
);
1947 static inline unsigned long wait_task_inactive(struct task_struct
*p
,
1954 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1956 #define for_each_process(p) \
1957 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1960 * Careful: do_each_thread/while_each_thread is a double loop so
1961 * 'break' will not work as expected - use goto instead.
1963 #define do_each_thread(g, t) \
1964 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1966 #define while_each_thread(g, t) \
1967 while ((t = next_thread(t)) != g)
1969 /* de_thread depends on thread_group_leader not being a pid based check */
1970 #define thread_group_leader(p) (p == p->group_leader)
1972 /* Do to the insanities of de_thread it is possible for a process
1973 * to have the pid of the thread group leader without actually being
1974 * the thread group leader. For iteration through the pids in proc
1975 * all we care about is that we have a task with the appropriate
1976 * pid, we don't actually care if we have the right task.
1978 static inline int has_group_leader_pid(struct task_struct
*p
)
1980 return p
->pid
== p
->tgid
;
1984 int same_thread_group(struct task_struct
*p1
, struct task_struct
*p2
)
1986 return p1
->tgid
== p2
->tgid
;
1989 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
1991 return list_entry(rcu_dereference(p
->thread_group
.next
),
1992 struct task_struct
, thread_group
);
1995 static inline int thread_group_empty(struct task_struct
*p
)
1997 return list_empty(&p
->thread_group
);
2000 #define delay_group_leader(p) \
2001 (thread_group_leader(p) && !thread_group_empty(p))
2004 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2005 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2006 * pins the final release of task.io_context. Also protects ->cpuset and
2007 * ->cgroup.subsys[].
2009 * Nests both inside and outside of read_lock(&tasklist_lock).
2010 * It must not be nested with write_lock_irq(&tasklist_lock),
2011 * neither inside nor outside.
2013 static inline void task_lock(struct task_struct
*p
)
2015 spin_lock(&p
->alloc_lock
);
2018 static inline void task_unlock(struct task_struct
*p
)
2020 spin_unlock(&p
->alloc_lock
);
2023 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
2024 unsigned long *flags
);
2026 static inline void unlock_task_sighand(struct task_struct
*tsk
,
2027 unsigned long *flags
)
2029 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
2032 #ifndef __HAVE_THREAD_FUNCTIONS
2034 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2035 #define task_stack_page(task) ((task)->stack)
2037 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
2039 *task_thread_info(p
) = *task_thread_info(org
);
2040 task_thread_info(p
)->task
= p
;
2043 static inline unsigned long *end_of_stack(struct task_struct
*p
)
2045 return (unsigned long *)(task_thread_info(p
) + 1);
2050 static inline int object_is_on_stack(void *obj
)
2052 void *stack
= task_stack_page(current
);
2054 return (obj
>= stack
) && (obj
< (stack
+ THREAD_SIZE
));
2057 extern void thread_info_cache_init(void);
2059 /* set thread flags in other task's structures
2060 * - see asm/thread_info.h for TIF_xxxx flags available
2062 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2064 set_ti_thread_flag(task_thread_info(tsk
), flag
);
2067 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2069 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2072 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2074 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
2077 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2079 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2082 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2084 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
2087 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
2089 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2092 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
2094 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2097 static inline int test_tsk_need_resched(struct task_struct
*tsk
)
2099 return unlikely(test_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
));
2102 static inline int signal_pending(struct task_struct
*p
)
2104 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
2107 extern int __fatal_signal_pending(struct task_struct
*p
);
2109 static inline int fatal_signal_pending(struct task_struct
*p
)
2111 return signal_pending(p
) && __fatal_signal_pending(p
);
2114 static inline int signal_pending_state(long state
, struct task_struct
*p
)
2116 if (!(state
& (TASK_INTERRUPTIBLE
| TASK_WAKEKILL
)))
2118 if (!signal_pending(p
))
2121 return (state
& TASK_INTERRUPTIBLE
) || __fatal_signal_pending(p
);
2124 static inline int need_resched(void)
2126 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
2130 * cond_resched() and cond_resched_lock(): latency reduction via
2131 * explicit rescheduling in places that are safe. The return
2132 * value indicates whether a reschedule was done in fact.
2133 * cond_resched_lock() will drop the spinlock before scheduling,
2134 * cond_resched_softirq() will enable bhs before scheduling.
2136 extern int _cond_resched(void);
2137 #ifdef CONFIG_PREEMPT_BKL
2138 static inline int cond_resched(void)
2143 static inline int cond_resched(void)
2145 return _cond_resched();
2148 extern int cond_resched_lock(spinlock_t
* lock
);
2149 extern int cond_resched_softirq(void);
2150 static inline int cond_resched_bkl(void)
2152 return _cond_resched();
2156 * Does a critical section need to be broken due to another
2157 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2158 * but a general need for low latency)
2160 static inline int spin_needbreak(spinlock_t
*lock
)
2162 #ifdef CONFIG_PREEMPT
2163 return spin_is_contended(lock
);
2170 * Thread group CPU time accounting.
2173 extern int thread_group_cputime_alloc(struct task_struct
*);
2174 extern void thread_group_cputime(struct task_struct
*, struct task_cputime
*);
2176 static inline void thread_group_cputime_init(struct signal_struct
*sig
)
2178 sig
->cputime
.totals
= NULL
;
2181 static inline int thread_group_cputime_clone_thread(struct task_struct
*curr
)
2183 if (curr
->signal
->cputime
.totals
)
2185 return thread_group_cputime_alloc(curr
);
2188 static inline void thread_group_cputime_free(struct signal_struct
*sig
)
2190 free_percpu(sig
->cputime
.totals
);
2194 * Reevaluate whether the task has signals pending delivery.
2195 * Wake the task if so.
2196 * This is required every time the blocked sigset_t changes.
2197 * callers must hold sighand->siglock.
2199 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
2200 extern void recalc_sigpending(void);
2202 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
2205 * Wrappers for p->thread_info->cpu access. No-op on UP.
2209 static inline unsigned int task_cpu(const struct task_struct
*p
)
2211 return task_thread_info(p
)->cpu
;
2214 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
2218 static inline unsigned int task_cpu(const struct task_struct
*p
)
2223 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
2227 #endif /* CONFIG_SMP */
2229 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
2231 #ifdef CONFIG_TRACING
2233 __trace_special(void *__tr
, void *__data
,
2234 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
);
2237 __trace_special(void *__tr
, void *__data
,
2238 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
)
2243 extern long sched_setaffinity(pid_t pid
, const cpumask_t
*new_mask
);
2244 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
2246 extern int sched_mc_power_savings
, sched_smt_power_savings
;
2248 extern void normalize_rt_tasks(void);
2250 #ifdef CONFIG_GROUP_SCHED
2252 extern struct task_group init_task_group
;
2253 #ifdef CONFIG_USER_SCHED
2254 extern struct task_group root_task_group
;
2257 extern struct task_group
*sched_create_group(struct task_group
*parent
);
2258 extern void sched_destroy_group(struct task_group
*tg
);
2259 extern void sched_move_task(struct task_struct
*tsk
);
2260 #ifdef CONFIG_FAIR_GROUP_SCHED
2261 extern int sched_group_set_shares(struct task_group
*tg
, unsigned long shares
);
2262 extern unsigned long sched_group_shares(struct task_group
*tg
);
2264 #ifdef CONFIG_RT_GROUP_SCHED
2265 extern int sched_group_set_rt_runtime(struct task_group
*tg
,
2266 long rt_runtime_us
);
2267 extern long sched_group_rt_runtime(struct task_group
*tg
);
2268 extern int sched_group_set_rt_period(struct task_group
*tg
,
2270 extern long sched_group_rt_period(struct task_group
*tg
);
2274 #ifdef CONFIG_TASK_XACCT
2275 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2277 tsk
->ioac
.rchar
+= amt
;
2280 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2282 tsk
->ioac
.wchar
+= amt
;
2285 static inline void inc_syscr(struct task_struct
*tsk
)
2290 static inline void inc_syscw(struct task_struct
*tsk
)
2295 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2299 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2303 static inline void inc_syscr(struct task_struct
*tsk
)
2307 static inline void inc_syscw(struct task_struct
*tsk
)
2312 #ifndef TASK_SIZE_OF
2313 #define TASK_SIZE_OF(tsk) TASK_SIZE
2316 #ifdef CONFIG_MM_OWNER
2317 extern void mm_update_next_owner(struct mm_struct
*mm
);
2318 extern void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
);
2320 static inline void mm_update_next_owner(struct mm_struct
*mm
)
2324 static inline void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
)
2327 #endif /* CONFIG_MM_OWNER */
2329 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2331 #endif /* __KERNEL__ */