4 #include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
9 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
10 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
11 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
12 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
13 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
14 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
15 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
16 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
17 #define CLONE_THREAD 0x00010000 /* Same thread group? */
18 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
19 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
20 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
21 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
22 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
23 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
24 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
25 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
26 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
27 #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
28 #define CLONE_NEWIPC 0x08000000 /* New ipcs */
29 #define CLONE_NEWUSER 0x10000000 /* New user namespace */
34 #define SCHED_NORMAL 0
38 /* SCHED_ISO: reserved but not implemented yet */
47 #include <asm/param.h> /* for HZ */
49 #include <linux/capability.h>
50 #include <linux/threads.h>
51 #include <linux/kernel.h>
52 #include <linux/types.h>
53 #include <linux/timex.h>
54 #include <linux/jiffies.h>
55 #include <linux/rbtree.h>
56 #include <linux/thread_info.h>
57 #include <linux/cpumask.h>
58 #include <linux/errno.h>
59 #include <linux/nodemask.h>
61 #include <asm/system.h>
62 #include <asm/semaphore.h>
64 #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/securebits.h>
72 #include <linux/fs_struct.h>
73 #include <linux/compiler.h>
74 #include <linux/completion.h>
75 #include <linux/pid.h>
76 #include <linux/percpu.h>
77 #include <linux/topology.h>
78 #include <linux/seccomp.h>
79 #include <linux/rcupdate.h>
80 #include <linux/futex.h>
81 #include <linux/rtmutex.h>
83 #include <linux/time.h>
84 #include <linux/param.h>
85 #include <linux/resource.h>
86 #include <linux/timer.h>
87 #include <linux/hrtimer.h>
88 #include <linux/task_io_accounting.h>
90 #include <asm/processor.h>
93 struct futex_pi_state
;
97 * List of flags we want to share for kernel threads,
98 * if only because they are not used by them anyway.
100 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
103 * These are the constant used to fake the fixed-point load-average
104 * counting. Some notes:
105 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
106 * a load-average precision of 10 bits integer + 11 bits fractional
107 * - if you want to count load-averages more often, you need more
108 * precision, or rounding will get you. With 2-second counting freq,
109 * the EXP_n values would be 1981, 2034 and 2043 if still using only
112 extern unsigned long avenrun
[]; /* Load averages */
114 #define FSHIFT 11 /* nr of bits of precision */
115 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
116 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
117 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
118 #define EXP_5 2014 /* 1/exp(5sec/5min) */
119 #define EXP_15 2037 /* 1/exp(5sec/15min) */
121 #define CALC_LOAD(load,exp,n) \
123 load += n*(FIXED_1-exp); \
126 extern unsigned long total_forks
;
127 extern int nr_threads
;
128 DECLARE_PER_CPU(unsigned long, process_counts
);
129 extern int nr_processes(void);
130 extern unsigned long nr_running(void);
131 extern unsigned long nr_uninterruptible(void);
132 extern unsigned long nr_active(void);
133 extern unsigned long nr_iowait(void);
134 extern unsigned long weighted_cpuload(const int cpu
);
138 #ifdef CONFIG_SCHED_DEBUG
139 extern void proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
);
140 extern void proc_sched_set_task(struct task_struct
*p
);
142 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
, u64 now
);
145 proc_sched_show_task(struct task_struct
*p
, struct seq_file
*m
)
148 static inline void proc_sched_set_task(struct task_struct
*p
)
152 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
, u64 now
)
158 * Task state bitmask. NOTE! These bits are also
159 * encoded in fs/proc/array.c: get_task_state().
161 * We have two separate sets of flags: task->state
162 * is about runnability, while task->exit_state are
163 * about the task exiting. Confusing, but this way
164 * modifying one set can't modify the other one by
167 #define TASK_RUNNING 0
168 #define TASK_INTERRUPTIBLE 1
169 #define TASK_UNINTERRUPTIBLE 2
170 #define TASK_STOPPED 4
171 #define TASK_TRACED 8
172 /* in tsk->exit_state */
173 #define EXIT_ZOMBIE 16
175 /* in tsk->state again */
176 #define TASK_NONINTERACTIVE 64
177 #define TASK_DEAD 128
179 #define __set_task_state(tsk, state_value) \
180 do { (tsk)->state = (state_value); } while (0)
181 #define set_task_state(tsk, state_value) \
182 set_mb((tsk)->state, (state_value))
185 * set_current_state() includes a barrier so that the write of current->state
186 * is correctly serialised wrt the caller's subsequent test of whether to
189 * set_current_state(TASK_UNINTERRUPTIBLE);
190 * if (do_i_need_to_sleep())
193 * If the caller does not need such serialisation then use __set_current_state()
195 #define __set_current_state(state_value) \
196 do { current->state = (state_value); } while (0)
197 #define set_current_state(state_value) \
198 set_mb(current->state, (state_value))
200 /* Task command name length */
201 #define TASK_COMM_LEN 16
203 #include <linux/spinlock.h>
206 * This serializes "schedule()" and also protects
207 * the run-queue from deletions/modifications (but
208 * _adding_ to the beginning of the run-queue has
211 extern rwlock_t tasklist_lock
;
212 extern spinlock_t mmlist_lock
;
216 extern void sched_init(void);
217 extern void sched_init_smp(void);
218 extern void init_idle(struct task_struct
*idle
, int cpu
);
219 extern void init_idle_bootup_task(struct task_struct
*idle
);
221 extern cpumask_t nohz_cpu_mask
;
222 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
223 extern int select_nohz_load_balancer(int cpu
);
225 static inline int select_nohz_load_balancer(int cpu
)
232 * Only dump TASK_* tasks. (0 for all tasks)
234 extern void show_state_filter(unsigned long state_filter
);
236 static inline void show_state(void)
238 show_state_filter(0);
241 extern void show_regs(struct pt_regs
*);
244 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
245 * task), SP is the stack pointer of the first frame that should be shown in the back
246 * trace (or NULL if the entire call-chain of the task should be shown).
248 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
250 void io_schedule(void);
251 long io_schedule_timeout(long timeout
);
253 extern void cpu_init (void);
254 extern void trap_init(void);
255 extern void update_process_times(int user
);
256 extern void scheduler_tick(void);
258 #ifdef CONFIG_DETECT_SOFTLOCKUP
259 extern void softlockup_tick(void);
260 extern void spawn_softlockup_task(void);
261 extern void touch_softlockup_watchdog(void);
262 extern void touch_all_softlockup_watchdogs(void);
264 static inline void softlockup_tick(void)
267 static inline void spawn_softlockup_task(void)
270 static inline void touch_softlockup_watchdog(void)
273 static inline void touch_all_softlockup_watchdogs(void)
279 /* Attach to any functions which should be ignored in wchan output. */
280 #define __sched __attribute__((__section__(".sched.text")))
281 /* Is this address in the __sched functions? */
282 extern int in_sched_functions(unsigned long addr
);
284 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
285 extern signed long FASTCALL(schedule_timeout(signed long timeout
));
286 extern signed long schedule_timeout_interruptible(signed long timeout
);
287 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
288 asmlinkage
void schedule(void);
291 struct user_namespace
;
293 /* Maximum number of active map areas.. This is a random (large) number */
294 #define DEFAULT_MAX_MAP_COUNT 65536
296 extern int sysctl_max_map_count
;
298 #include <linux/aio.h>
301 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
302 unsigned long, unsigned long);
304 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
305 unsigned long len
, unsigned long pgoff
,
306 unsigned long flags
);
307 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
308 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
310 #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
312 * The mm counters are not protected by its page_table_lock,
313 * so must be incremented atomically.
315 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
316 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
317 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
318 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
319 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
320 typedef atomic_long_t mm_counter_t
;
322 #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
324 * The mm counters are protected by its page_table_lock,
325 * so can be incremented directly.
327 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
328 #define get_mm_counter(mm, member) ((mm)->_##member)
329 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
330 #define inc_mm_counter(mm, member) (mm)->_##member++
331 #define dec_mm_counter(mm, member) (mm)->_##member--
332 typedef unsigned long mm_counter_t
;
334 #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
336 #define get_mm_rss(mm) \
337 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
338 #define update_hiwater_rss(mm) do { \
339 unsigned long _rss = get_mm_rss(mm); \
340 if ((mm)->hiwater_rss < _rss) \
341 (mm)->hiwater_rss = _rss; \
343 #define update_hiwater_vm(mm) do { \
344 if ((mm)->hiwater_vm < (mm)->total_vm) \
345 (mm)->hiwater_vm = (mm)->total_vm; \
349 struct vm_area_struct
* mmap
; /* list of VMAs */
350 struct rb_root mm_rb
;
351 struct vm_area_struct
* mmap_cache
; /* last find_vma result */
352 unsigned long (*get_unmapped_area
) (struct file
*filp
,
353 unsigned long addr
, unsigned long len
,
354 unsigned long pgoff
, unsigned long flags
);
355 void (*unmap_area
) (struct mm_struct
*mm
, unsigned long addr
);
356 unsigned long mmap_base
; /* base of mmap area */
357 unsigned long task_size
; /* size of task vm space */
358 unsigned long cached_hole_size
; /* if non-zero, the largest hole below free_area_cache */
359 unsigned long free_area_cache
; /* first hole of size cached_hole_size or larger */
361 atomic_t mm_users
; /* How many users with user space? */
362 atomic_t mm_count
; /* How many references to "struct mm_struct" (users count as 1) */
363 int map_count
; /* number of VMAs */
364 struct rw_semaphore mmap_sem
;
365 spinlock_t page_table_lock
; /* Protects page tables and some counters */
367 struct list_head mmlist
; /* List of maybe swapped mm's. These are globally strung
368 * together off init_mm.mmlist, and are protected
372 /* Special counters, in some configurations protected by the
373 * page_table_lock, in other configurations by being atomic.
375 mm_counter_t _file_rss
;
376 mm_counter_t _anon_rss
;
378 unsigned long hiwater_rss
; /* High-watermark of RSS usage */
379 unsigned long hiwater_vm
; /* High-water virtual memory usage */
381 unsigned long total_vm
, locked_vm
, shared_vm
, exec_vm
;
382 unsigned long stack_vm
, reserved_vm
, def_flags
, nr_ptes
;
383 unsigned long start_code
, end_code
, start_data
, end_data
;
384 unsigned long start_brk
, brk
, start_stack
;
385 unsigned long arg_start
, arg_end
, env_start
, env_end
;
387 unsigned long saved_auxv
[AT_VECTOR_SIZE
]; /* for /proc/PID/auxv */
389 cpumask_t cpu_vm_mask
;
391 /* Architecture-specific MM context */
392 mm_context_t context
;
394 /* Swap token stuff */
396 * Last value of global fault stamp as seen by this process.
397 * In other words, this value gives an indication of how long
398 * it has been since this task got the token.
399 * Look at mm/thrash.c
401 unsigned int faultstamp
;
402 unsigned int token_priority
;
403 unsigned int last_interval
;
405 unsigned char dumpable
:2;
407 /* coredumping support */
409 struct completion
*core_startup_done
, core_done
;
412 rwlock_t ioctx_list_lock
;
413 struct kioctx
*ioctx_list
;
416 struct sighand_struct
{
418 struct k_sigaction action
[_NSIG
];
420 struct list_head signalfd_list
;
423 struct pacct_struct
{
426 unsigned long ac_mem
;
427 cputime_t ac_utime
, ac_stime
;
428 unsigned long ac_minflt
, ac_majflt
;
432 * NOTE! "signal_struct" does not have it's own
433 * locking, because a shared signal_struct always
434 * implies a shared sighand_struct, so locking
435 * sighand_struct is always a proper superset of
436 * the locking of signal_struct.
438 struct signal_struct
{
442 wait_queue_head_t wait_chldexit
; /* for wait4() */
444 /* current thread group signal load-balancing target: */
445 struct task_struct
*curr_target
;
447 /* shared signal handling: */
448 struct sigpending shared_pending
;
450 /* thread group exit support */
453 * - notify group_exit_task when ->count is equal to notify_count
454 * - everyone except group_exit_task is stopped during signal delivery
455 * of fatal signals, group_exit_task processes the signal.
457 struct task_struct
*group_exit_task
;
460 /* thread group stop support, overloads group_exit_code too */
461 int group_stop_count
;
462 unsigned int flags
; /* see SIGNAL_* flags below */
464 /* POSIX.1b Interval Timers */
465 struct list_head posix_timers
;
467 /* ITIMER_REAL timer for the process */
468 struct hrtimer real_timer
;
469 struct task_struct
*tsk
;
470 ktime_t it_real_incr
;
472 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
473 cputime_t it_prof_expires
, it_virt_expires
;
474 cputime_t it_prof_incr
, it_virt_incr
;
476 /* job control IDs */
478 struct pid
*tty_old_pgrp
;
481 pid_t session __deprecated
;
485 /* boolean value for session group leader */
488 struct tty_struct
*tty
; /* NULL if no tty */
491 * Cumulative resource counters for dead threads in the group,
492 * and for reaped dead child processes forked by this group.
493 * Live threads maintain their own counters and add to these
494 * in __exit_signal, except for the group leader.
496 cputime_t utime
, stime
, cutime
, cstime
;
497 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
498 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
499 unsigned long inblock
, oublock
, cinblock
, coublock
;
502 * Cumulative ns of scheduled CPU time for dead threads in the
503 * group, not including a zombie group leader. (This only differs
504 * from jiffies_to_ns(utime + stime) if sched_clock uses something
505 * other than jiffies.)
507 unsigned long long sum_sched_runtime
;
510 * We don't bother to synchronize most readers of this at all,
511 * because there is no reader checking a limit that actually needs
512 * to get both rlim_cur and rlim_max atomically, and either one
513 * alone is a single word that can safely be read normally.
514 * getrlimit/setrlimit use task_lock(current->group_leader) to
515 * protect this instead of the siglock, because they really
516 * have no need to disable irqs.
518 struct rlimit rlim
[RLIM_NLIMITS
];
520 struct list_head cpu_timers
[3];
522 /* keep the process-shared keyrings here so that they do the right
523 * thing in threads created with CLONE_THREAD */
525 struct key
*session_keyring
; /* keyring inherited over fork */
526 struct key
*process_keyring
; /* keyring private to this process */
528 #ifdef CONFIG_BSD_PROCESS_ACCT
529 struct pacct_struct pacct
; /* per-process accounting information */
531 #ifdef CONFIG_TASKSTATS
532 struct taskstats
*stats
;
536 struct tty_audit_buf
*tty_audit_buf
;
540 /* Context switch must be unlocked if interrupts are to be enabled */
541 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
542 # define __ARCH_WANT_UNLOCKED_CTXSW
546 * Bits in flags field of signal_struct.
548 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
549 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
550 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
551 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
554 * Some day this will be a full-fledged user tracking system..
557 atomic_t __count
; /* reference count */
558 atomic_t processes
; /* How many processes does this user have? */
559 atomic_t files
; /* How many open files does this user have? */
560 atomic_t sigpending
; /* How many pending signals does this user have? */
561 #ifdef CONFIG_INOTIFY_USER
562 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
563 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
565 /* protected by mq_lock */
566 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
567 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
570 struct key
*uid_keyring
; /* UID specific keyring */
571 struct key
*session_keyring
; /* UID's default session keyring */
574 /* Hash table maintenance information */
575 struct list_head uidhash_list
;
579 extern struct user_struct
*find_user(uid_t
);
581 extern struct user_struct root_user
;
582 #define INIT_USER (&root_user)
584 struct backing_dev_info
;
585 struct reclaim_state
;
587 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
589 /* cumulative counters */
590 unsigned long pcnt
; /* # of times run on this cpu */
591 unsigned long long cpu_time
, /* time spent on the cpu */
592 run_delay
; /* time spent waiting on a runqueue */
595 unsigned long long last_arrival
,/* when we last ran on a cpu */
596 last_queued
; /* when we were last queued to run */
598 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
600 #ifdef CONFIG_SCHEDSTATS
601 extern const struct file_operations proc_schedstat_operations
;
602 #endif /* CONFIG_SCHEDSTATS */
604 #ifdef CONFIG_TASK_DELAY_ACCT
605 struct task_delay_info
{
607 unsigned int flags
; /* Private per-task flags */
609 /* For each stat XXX, add following, aligned appropriately
611 * struct timespec XXX_start, XXX_end;
615 * Atomicity of updates to XXX_delay, XXX_count protected by
616 * single lock above (split into XXX_lock if contention is an issue).
620 * XXX_count is incremented on every XXX operation, the delay
621 * associated with the operation is added to XXX_delay.
622 * XXX_delay contains the accumulated delay time in nanoseconds.
624 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
625 u64 blkio_delay
; /* wait for sync block io completion */
626 u64 swapin_delay
; /* wait for swapin block io completion */
627 u32 blkio_count
; /* total count of the number of sync block */
628 /* io operations performed */
629 u32 swapin_count
; /* total count of the number of swapin block */
630 /* io operations performed */
632 #endif /* CONFIG_TASK_DELAY_ACCT */
634 static inline int sched_info_on(void)
636 #ifdef CONFIG_SCHEDSTATS
638 #elif defined(CONFIG_TASK_DELAY_ACCT)
639 extern int delayacct_on
;
654 * sched-domains (multiprocessor balancing) declarations:
658 * Increase resolution of nice-level calculations:
660 #define SCHED_LOAD_SHIFT 10
661 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
663 #define SCHED_LOAD_SCALE_FUZZ (SCHED_LOAD_SCALE >> 5)
666 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
667 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
668 #define SD_BALANCE_EXEC 4 /* Balance on exec */
669 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
670 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
671 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
672 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
673 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
674 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
675 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
676 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
678 #define BALANCE_FOR_MC_POWER \
679 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
681 #define BALANCE_FOR_PKG_POWER \
682 ((sched_mc_power_savings || sched_smt_power_savings) ? \
683 SD_POWERSAVINGS_BALANCE : 0)
685 #define test_sd_parent(sd, flag) ((sd->parent && \
686 (sd->parent->flags & flag)) ? 1 : 0)
690 struct sched_group
*next
; /* Must be a circular list */
694 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
695 * single CPU. This is read only (except for setup, hotplug CPU).
696 * Note : Never change cpu_power without recompute its reciprocal
698 unsigned int __cpu_power
;
700 * reciprocal value of cpu_power to avoid expensive divides
701 * (see include/linux/reciprocal_div.h)
703 u32 reciprocal_cpu_power
;
706 struct sched_domain
{
707 /* These fields must be setup */
708 struct sched_domain
*parent
; /* top domain must be null terminated */
709 struct sched_domain
*child
; /* bottom domain must be null terminated */
710 struct sched_group
*groups
; /* the balancing groups of the domain */
711 cpumask_t span
; /* span of all CPUs in this domain */
712 unsigned long min_interval
; /* Minimum balance interval ms */
713 unsigned long max_interval
; /* Maximum balance interval ms */
714 unsigned int busy_factor
; /* less balancing by factor if busy */
715 unsigned int imbalance_pct
; /* No balance until over watermark */
716 unsigned long long cache_hot_time
; /* Task considered cache hot (ns) */
717 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
718 unsigned int busy_idx
;
719 unsigned int idle_idx
;
720 unsigned int newidle_idx
;
721 unsigned int wake_idx
;
722 unsigned int forkexec_idx
;
723 int flags
; /* See SD_* */
725 /* Runtime fields. */
726 unsigned long last_balance
; /* init to jiffies. units in jiffies */
727 unsigned int balance_interval
; /* initialise to 1. units in ms. */
728 unsigned int nr_balance_failed
; /* initialise to 0 */
730 #ifdef CONFIG_SCHEDSTATS
731 /* load_balance() stats */
732 unsigned long lb_cnt
[CPU_MAX_IDLE_TYPES
];
733 unsigned long lb_failed
[CPU_MAX_IDLE_TYPES
];
734 unsigned long lb_balanced
[CPU_MAX_IDLE_TYPES
];
735 unsigned long lb_imbalance
[CPU_MAX_IDLE_TYPES
];
736 unsigned long lb_gained
[CPU_MAX_IDLE_TYPES
];
737 unsigned long lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
738 unsigned long lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
739 unsigned long lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
741 /* Active load balancing */
742 unsigned long alb_cnt
;
743 unsigned long alb_failed
;
744 unsigned long alb_pushed
;
746 /* SD_BALANCE_EXEC stats */
747 unsigned long sbe_cnt
;
748 unsigned long sbe_balanced
;
749 unsigned long sbe_pushed
;
751 /* SD_BALANCE_FORK stats */
752 unsigned long sbf_cnt
;
753 unsigned long sbf_balanced
;
754 unsigned long sbf_pushed
;
756 /* try_to_wake_up() stats */
757 unsigned long ttwu_wake_remote
;
758 unsigned long ttwu_move_affine
;
759 unsigned long ttwu_move_balance
;
763 extern int partition_sched_domains(cpumask_t
*partition1
,
764 cpumask_t
*partition2
);
766 #endif /* CONFIG_SMP */
769 struct io_context
; /* See blkdev.h */
772 #define NGROUPS_SMALL 32
773 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
777 gid_t small_block
[NGROUPS_SMALL
];
783 * get_group_info() must be called with the owning task locked (via task_lock())
784 * when task != current. The reason being that the vast majority of callers are
785 * looking at current->group_info, which can not be changed except by the
786 * current task. Changing current->group_info requires the task lock, too.
788 #define get_group_info(group_info) do { \
789 atomic_inc(&(group_info)->usage); \
792 #define put_group_info(group_info) do { \
793 if (atomic_dec_and_test(&(group_info)->usage)) \
794 groups_free(group_info); \
797 extern struct group_info
*groups_alloc(int gidsetsize
);
798 extern void groups_free(struct group_info
*group_info
);
799 extern int set_current_groups(struct group_info
*group_info
);
800 extern int groups_search(struct group_info
*group_info
, gid_t grp
);
801 /* access the groups "array" with this macro */
802 #define GROUP_AT(gi, i) \
803 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
805 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
806 extern void prefetch_stack(struct task_struct
*t
);
808 static inline void prefetch_stack(struct task_struct
*t
) { }
811 struct audit_context
; /* See audit.c */
813 struct pipe_inode_info
;
814 struct uts_namespace
;
820 struct sched_class
*next
;
822 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
,
823 int wakeup
, u64 now
);
824 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
,
826 void (*yield_task
) (struct rq
*rq
, struct task_struct
*p
);
828 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
);
830 struct task_struct
* (*pick_next_task
) (struct rq
*rq
, u64 now
);
831 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
, u64 now
);
833 int (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
835 unsigned long max_nr_move
, unsigned long max_load_move
,
836 struct sched_domain
*sd
, enum cpu_idle_type idle
,
837 int *all_pinned
, unsigned long *total_load_moved
);
839 void (*set_curr_task
) (struct rq
*rq
);
840 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
);
841 void (*task_new
) (struct rq
*rq
, struct task_struct
*p
);
845 unsigned long weight
, inv_weight
;
849 * CFS stats for a schedulable entity (task, task-group etc)
851 * Current field usage histogram:
856 * 4 se->sleep_start_fair
859 * 15 se->wait_runtime
861 struct sched_entity
{
863 unsigned long delta_fair_run
;
864 unsigned long delta_fair_sleep
;
865 unsigned long delta_exec
;
867 struct load_weight load
; /* for load-balancing */
868 struct rb_node run_node
;
875 u64 sleep_start_fair
;
883 u64 sum_exec_runtime
;
884 s64 sum_wait_runtime
;
885 s64 sum_sleep_runtime
;
886 unsigned long wait_runtime_overruns
;
887 unsigned long wait_runtime_underruns
;
888 #ifdef CONFIG_FAIR_GROUP_SCHED
889 struct sched_entity
*parent
;
890 /* rq on which this entity is (to be) queued: */
891 struct cfs_rq
*cfs_rq
;
892 /* rq "owned" by this entity/group: */
898 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
901 unsigned int flags
; /* per process flags, defined below */
904 int lock_depth
; /* BKL lock depth */
907 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
912 int prio
, static_prio
, normal_prio
;
913 struct list_head run_list
;
914 struct sched_class
*sched_class
;
915 struct sched_entity se
;
917 unsigned short ioprio
;
918 #ifdef CONFIG_BLK_DEV_IO_TRACE
919 unsigned int btrace_seq
;
923 cpumask_t cpus_allowed
;
924 unsigned int time_slice
;
926 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
927 struct sched_info sched_info
;
930 struct list_head tasks
;
932 * ptrace_list/ptrace_children forms the list of my children
933 * that were stolen by a ptracer.
935 struct list_head ptrace_children
;
936 struct list_head ptrace_list
;
938 struct mm_struct
*mm
, *active_mm
;
941 struct linux_binfmt
*binfmt
;
943 int exit_code
, exit_signal
;
944 int pdeath_signal
; /* The signal sent when the parent dies */
946 unsigned int personality
;
951 #ifdef CONFIG_CC_STACKPROTECTOR
952 /* Canary value for the -fstack-protector gcc feature */
953 unsigned long stack_canary
;
956 * pointers to (original) parent process, youngest child, younger sibling,
957 * older sibling, respectively. (p->father can be replaced with
960 struct task_struct
*real_parent
; /* real parent process (when being debugged) */
961 struct task_struct
*parent
; /* parent process */
963 * children/sibling forms the list of my children plus the
964 * tasks I'm ptracing.
966 struct list_head children
; /* list of my children */
967 struct list_head sibling
; /* linkage in my parent's children list */
968 struct task_struct
*group_leader
; /* threadgroup leader */
970 /* PID/PID hash table linkage. */
971 struct pid_link pids
[PIDTYPE_MAX
];
972 struct list_head thread_group
;
974 struct completion
*vfork_done
; /* for vfork() */
975 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
976 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
978 unsigned int rt_priority
;
979 cputime_t utime
, stime
;
980 unsigned long nvcsw
, nivcsw
; /* context switch counts */
981 struct timespec start_time
; /* monotonic time */
982 struct timespec real_start_time
; /* boot based time */
983 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
984 unsigned long min_flt
, maj_flt
;
986 cputime_t it_prof_expires
, it_virt_expires
;
987 unsigned long long it_sched_expires
;
988 struct list_head cpu_timers
[3];
990 /* process credentials */
991 uid_t uid
,euid
,suid
,fsuid
;
992 gid_t gid
,egid
,sgid
,fsgid
;
993 struct group_info
*group_info
;
994 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
;
995 unsigned keep_capabilities
:1;
996 struct user_struct
*user
;
998 struct key
*request_key_auth
; /* assumed request_key authority */
999 struct key
*thread_keyring
; /* keyring private to this thread */
1000 unsigned char jit_keyring
; /* default keyring to attach requested keys to */
1003 * fpu_counter contains the number of consecutive context switches
1004 * that the FPU is used. If this is over a threshold, the lazy fpu
1005 * saving becomes unlazy to save the trap. This is an unsigned char
1006 * so that after 256 times the counter wraps and the behavior turns
1007 * lazy again; this to deal with bursty apps that only use FPU for
1010 unsigned char fpu_counter
;
1011 int oomkilladj
; /* OOM kill score adjustment (bit shift). */
1012 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1013 - access with [gs]et_task_comm (which lock
1014 it with task_lock())
1015 - initialized normally by flush_old_exec */
1016 /* file system info */
1017 int link_count
, total_link_count
;
1018 #ifdef CONFIG_SYSVIPC
1020 struct sysv_sem sysvsem
;
1022 /* CPU-specific state of this task */
1023 struct thread_struct thread
;
1024 /* filesystem information */
1025 struct fs_struct
*fs
;
1026 /* open file information */
1027 struct files_struct
*files
;
1029 struct nsproxy
*nsproxy
;
1030 /* signal handlers */
1031 struct signal_struct
*signal
;
1032 struct sighand_struct
*sighand
;
1034 sigset_t blocked
, real_blocked
;
1035 sigset_t saved_sigmask
; /* To be restored with TIF_RESTORE_SIGMASK */
1036 struct sigpending pending
;
1038 unsigned long sas_ss_sp
;
1040 int (*notifier
)(void *priv
);
1041 void *notifier_data
;
1042 sigset_t
*notifier_mask
;
1045 struct audit_context
*audit_context
;
1048 /* Thread group tracking */
1051 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1052 spinlock_t alloc_lock
;
1054 /* Protection of the PI data structures: */
1057 #ifdef CONFIG_RT_MUTEXES
1058 /* PI waiters blocked on a rt_mutex held by this task */
1059 struct plist_head pi_waiters
;
1060 /* Deadlock detection and priority inheritance handling */
1061 struct rt_mutex_waiter
*pi_blocked_on
;
1064 #ifdef CONFIG_DEBUG_MUTEXES
1065 /* mutex deadlock detection */
1066 struct mutex_waiter
*blocked_on
;
1068 #ifdef CONFIG_TRACE_IRQFLAGS
1069 unsigned int irq_events
;
1070 int hardirqs_enabled
;
1071 unsigned long hardirq_enable_ip
;
1072 unsigned int hardirq_enable_event
;
1073 unsigned long hardirq_disable_ip
;
1074 unsigned int hardirq_disable_event
;
1075 int softirqs_enabled
;
1076 unsigned long softirq_disable_ip
;
1077 unsigned int softirq_disable_event
;
1078 unsigned long softirq_enable_ip
;
1079 unsigned int softirq_enable_event
;
1080 int hardirq_context
;
1081 int softirq_context
;
1083 #ifdef CONFIG_LOCKDEP
1084 # define MAX_LOCK_DEPTH 30UL
1087 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1088 unsigned int lockdep_recursion
;
1091 /* journalling filesystem info */
1094 /* stacked block device info */
1095 struct bio
*bio_list
, **bio_tail
;
1098 struct reclaim_state
*reclaim_state
;
1100 struct backing_dev_info
*backing_dev_info
;
1102 struct io_context
*io_context
;
1104 unsigned long ptrace_message
;
1105 siginfo_t
*last_siginfo
; /* For ptrace use. */
1107 * current io wait handle: wait queue entry to use for io waits
1108 * If this thread is processing aio, this points at the waitqueue
1109 * inside the currently handled kiocb. It may be NULL (i.e. default
1110 * to a stack based synchronous wait) if its doing sync IO.
1112 wait_queue_t
*io_wait
;
1113 #ifdef CONFIG_TASK_XACCT
1114 /* i/o counters(bytes read/written, #syscalls */
1115 u64 rchar
, wchar
, syscr
, syscw
;
1117 struct task_io_accounting ioac
;
1118 #if defined(CONFIG_TASK_XACCT)
1119 u64 acct_rss_mem1
; /* accumulated rss usage */
1120 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1121 cputime_t acct_stimexpd
;/* stime since last update */
1124 struct mempolicy
*mempolicy
;
1127 #ifdef CONFIG_CPUSETS
1128 struct cpuset
*cpuset
;
1129 nodemask_t mems_allowed
;
1130 int cpuset_mems_generation
;
1131 int cpuset_mem_spread_rotor
;
1133 struct robust_list_head __user
*robust_list
;
1134 #ifdef CONFIG_COMPAT
1135 struct compat_robust_list_head __user
*compat_robust_list
;
1137 struct list_head pi_state_list
;
1138 struct futex_pi_state
*pi_state_cache
;
1140 atomic_t fs_excl
; /* holding fs exclusive resources */
1141 struct rcu_head rcu
;
1144 * cache last used pipe for splice
1146 struct pipe_inode_info
*splice_pipe
;
1147 #ifdef CONFIG_TASK_DELAY_ACCT
1148 struct task_delay_info
*delays
;
1150 #ifdef CONFIG_FAULT_INJECTION
1156 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1157 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1158 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1159 * values are inverted: lower p->prio value means higher priority.
1161 * The MAX_USER_RT_PRIO value allows the actual maximum
1162 * RT priority to be separate from the value exported to
1163 * user-space. This allows kernel threads to set their
1164 * priority to a value higher than any user task. Note:
1165 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1168 #define MAX_USER_RT_PRIO 100
1169 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1171 #define MAX_PRIO (MAX_RT_PRIO + 40)
1172 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1174 static inline int rt_prio(int prio
)
1176 if (unlikely(prio
< MAX_RT_PRIO
))
1181 static inline int rt_task(struct task_struct
*p
)
1183 return rt_prio(p
->prio
);
1186 static inline pid_t
process_group(struct task_struct
*tsk
)
1188 return tsk
->signal
->pgrp
;
1191 static inline pid_t
signal_session(struct signal_struct
*sig
)
1193 return sig
->__session
;
1196 static inline pid_t
process_session(struct task_struct
*tsk
)
1198 return signal_session(tsk
->signal
);
1201 static inline void set_signal_session(struct signal_struct
*sig
, pid_t session
)
1203 sig
->__session
= session
;
1206 static inline struct pid
*task_pid(struct task_struct
*task
)
1208 return task
->pids
[PIDTYPE_PID
].pid
;
1211 static inline struct pid
*task_tgid(struct task_struct
*task
)
1213 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1216 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1218 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1221 static inline struct pid
*task_session(struct task_struct
*task
)
1223 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1227 * pid_alive - check that a task structure is not stale
1228 * @p: Task structure to be checked.
1230 * Test if a process is not yet dead (at most zombie state)
1231 * If pid_alive fails, then pointers within the task structure
1232 * can be stale and must not be dereferenced.
1234 static inline int pid_alive(struct task_struct
*p
)
1236 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1240 * is_init - check if a task structure is init
1241 * @tsk: Task structure to be checked.
1243 * Check if a task structure is the first user space task the kernel created.
1245 static inline int is_init(struct task_struct
*tsk
)
1247 return tsk
->pid
== 1;
1250 extern struct pid
*cad_pid
;
1252 extern void free_task(struct task_struct
*tsk
);
1253 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1255 extern void __put_task_struct(struct task_struct
*t
);
1257 static inline void put_task_struct(struct task_struct
*t
)
1259 if (atomic_dec_and_test(&t
->usage
))
1260 __put_task_struct(t
);
1266 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1267 /* Not implemented yet, only for 486*/
1268 #define PF_STARTING 0x00000002 /* being created */
1269 #define PF_EXITING 0x00000004 /* getting shut down */
1270 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1271 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1272 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1273 #define PF_DUMPCORE 0x00000200 /* dumped core */
1274 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1275 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1276 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1277 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1278 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1279 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1280 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1281 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1282 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1283 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1284 #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1285 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1286 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1287 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1288 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1289 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1290 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1291 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1294 * Only the _current_ task can read/write to tsk->flags, but other
1295 * tasks can access tsk->flags in readonly mode for example
1296 * with tsk_used_math (like during threaded core dumping).
1297 * There is however an exception to this rule during ptrace
1298 * or during fork: the ptracer task is allowed to write to the
1299 * child->flags of its traced child (same goes for fork, the parent
1300 * can write to the child->flags), because we're guaranteed the
1301 * child is not running and in turn not changing child->flags
1302 * at the same time the parent does it.
1304 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1305 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1306 #define clear_used_math() clear_stopped_child_used_math(current)
1307 #define set_used_math() set_stopped_child_used_math(current)
1308 #define conditional_stopped_child_used_math(condition, child) \
1309 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1310 #define conditional_used_math(condition) \
1311 conditional_stopped_child_used_math(condition, current)
1312 #define copy_to_stopped_child_used_math(child) \
1313 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1314 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1315 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1316 #define used_math() tsk_used_math(current)
1319 extern int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
);
1321 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1323 if (!cpu_isset(0, new_mask
))
1329 extern unsigned long long sched_clock(void);
1330 extern unsigned long long
1331 task_sched_runtime(struct task_struct
*task
);
1333 /* sched_exec is called by processes performing an exec */
1335 extern void sched_exec(void);
1337 #define sched_exec() {}
1340 extern void sched_clock_unstable_event(void);
1342 #ifdef CONFIG_HOTPLUG_CPU
1343 extern void idle_task_exit(void);
1345 static inline void idle_task_exit(void) {}
1348 extern void sched_idle_next(void);
1350 extern unsigned int sysctl_sched_granularity
;
1351 extern unsigned int sysctl_sched_wakeup_granularity
;
1352 extern unsigned int sysctl_sched_batch_wakeup_granularity
;
1353 extern unsigned int sysctl_sched_stat_granularity
;
1354 extern unsigned int sysctl_sched_runtime_limit
;
1355 extern unsigned int sysctl_sched_child_runs_first
;
1356 extern unsigned int sysctl_sched_features
;
1358 #ifdef CONFIG_RT_MUTEXES
1359 extern int rt_mutex_getprio(struct task_struct
*p
);
1360 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1361 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1363 static inline int rt_mutex_getprio(struct task_struct
*p
)
1365 return p
->normal_prio
;
1367 # define rt_mutex_adjust_pi(p) do { } while (0)
1370 extern void set_user_nice(struct task_struct
*p
, long nice
);
1371 extern int task_prio(const struct task_struct
*p
);
1372 extern int task_nice(const struct task_struct
*p
);
1373 extern int can_nice(const struct task_struct
*p
, const int nice
);
1374 extern int task_curr(const struct task_struct
*p
);
1375 extern int idle_cpu(int cpu
);
1376 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1377 extern struct task_struct
*idle_task(int cpu
);
1378 extern struct task_struct
*curr_task(int cpu
);
1379 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1384 * The default (Linux) execution domain.
1386 extern struct exec_domain default_exec_domain
;
1388 union thread_union
{
1389 struct thread_info thread_info
;
1390 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1393 #ifndef __HAVE_ARCH_KSTACK_END
1394 static inline int kstack_end(void *addr
)
1396 /* Reliable end of stack detection:
1397 * Some APM bios versions misalign the stack
1399 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1403 extern union thread_union init_thread_union
;
1404 extern struct task_struct init_task
;
1406 extern struct mm_struct init_mm
;
1408 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1409 extern struct task_struct
*find_task_by_pid_type(int type
, int pid
);
1410 extern void __set_special_pids(pid_t session
, pid_t pgrp
);
1412 /* per-UID process charging. */
1413 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
1414 static inline struct user_struct
*get_uid(struct user_struct
*u
)
1416 atomic_inc(&u
->__count
);
1419 extern void free_uid(struct user_struct
*);
1420 extern void switch_uid(struct user_struct
*);
1422 #include <asm/current.h>
1424 extern void do_timer(unsigned long ticks
);
1426 extern int FASTCALL(wake_up_state(struct task_struct
* tsk
, unsigned int state
));
1427 extern int FASTCALL(wake_up_process(struct task_struct
* tsk
));
1428 extern void FASTCALL(wake_up_new_task(struct task_struct
* tsk
,
1429 unsigned long clone_flags
));
1431 extern void kick_process(struct task_struct
*tsk
);
1433 static inline void kick_process(struct task_struct
*tsk
) { }
1435 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
1436 extern void sched_dead(struct task_struct
*p
);
1438 extern int in_group_p(gid_t
);
1439 extern int in_egroup_p(gid_t
);
1441 extern void proc_caches_init(void);
1442 extern void flush_signals(struct task_struct
*);
1443 extern void ignore_signals(struct task_struct
*);
1444 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
1445 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
1447 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
1449 unsigned long flags
;
1452 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
1453 ret
= dequeue_signal(tsk
, mask
, info
);
1454 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
1459 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
1461 extern void unblock_all_signals(void);
1462 extern void release_task(struct task_struct
* p
);
1463 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
1464 extern int send_group_sig_info(int, struct siginfo
*, struct task_struct
*);
1465 extern int force_sigsegv(int, struct task_struct
*);
1466 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
1467 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1468 extern int kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1469 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
1470 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
1471 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
1472 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
1473 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
1474 extern void do_notify_parent(struct task_struct
*, int);
1475 extern void force_sig(int, struct task_struct
*);
1476 extern void force_sig_specific(int, struct task_struct
*);
1477 extern int send_sig(int, struct task_struct
*, int);
1478 extern void zap_other_threads(struct task_struct
*p
);
1479 extern int kill_proc(pid_t
, int, int);
1480 extern struct sigqueue
*sigqueue_alloc(void);
1481 extern void sigqueue_free(struct sigqueue
*);
1482 extern int send_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1483 extern int send_group_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1484 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
1485 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
1487 static inline int kill_cad_pid(int sig
, int priv
)
1489 return kill_pid(cad_pid
, sig
, priv
);
1492 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1493 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1494 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1495 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1497 static inline int is_si_special(const struct siginfo
*info
)
1499 return info
<= SEND_SIG_FORCED
;
1502 /* True if we are on the alternate signal stack. */
1504 static inline int on_sig_stack(unsigned long sp
)
1506 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
1509 static inline int sas_ss_flags(unsigned long sp
)
1511 return (current
->sas_ss_size
== 0 ? SS_DISABLE
1512 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
1516 * Routines for handling mm_structs
1518 extern struct mm_struct
* mm_alloc(void);
1520 /* mmdrop drops the mm and the page tables */
1521 extern void FASTCALL(__mmdrop(struct mm_struct
*));
1522 static inline void mmdrop(struct mm_struct
* mm
)
1524 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
1528 /* mmput gets rid of the mappings and all user-space */
1529 extern void mmput(struct mm_struct
*);
1530 /* Grab a reference to a task's mm, if it is not already going away */
1531 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1532 /* Remove the current tasks stale references to the old mm_struct */
1533 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1535 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1536 extern void flush_thread(void);
1537 extern void exit_thread(void);
1539 extern void exit_files(struct task_struct
*);
1540 extern void __cleanup_signal(struct signal_struct
*);
1541 extern void __cleanup_sighand(struct sighand_struct
*);
1542 extern void exit_itimers(struct signal_struct
*);
1544 extern NORET_TYPE
void do_group_exit(int);
1546 extern void daemonize(const char *, ...);
1547 extern int allow_signal(int);
1548 extern int disallow_signal(int);
1550 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1551 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1552 struct task_struct
*fork_idle(int);
1554 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1555 extern void get_task_comm(char *to
, struct task_struct
*tsk
);
1558 extern void wait_task_inactive(struct task_struct
* p
);
1560 #define wait_task_inactive(p) do { } while (0)
1563 #define remove_parent(p) list_del_init(&(p)->sibling)
1564 #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1566 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1568 #define for_each_process(p) \
1569 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1572 * Careful: do_each_thread/while_each_thread is a double loop so
1573 * 'break' will not work as expected - use goto instead.
1575 #define do_each_thread(g, t) \
1576 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1578 #define while_each_thread(g, t) \
1579 while ((t = next_thread(t)) != g)
1581 /* de_thread depends on thread_group_leader not being a pid based check */
1582 #define thread_group_leader(p) (p == p->group_leader)
1584 /* Do to the insanities of de_thread it is possible for a process
1585 * to have the pid of the thread group leader without actually being
1586 * the thread group leader. For iteration through the pids in proc
1587 * all we care about is that we have a task with the appropriate
1588 * pid, we don't actually care if we have the right task.
1590 static inline int has_group_leader_pid(struct task_struct
*p
)
1592 return p
->pid
== p
->tgid
;
1595 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
1597 return list_entry(rcu_dereference(p
->thread_group
.next
),
1598 struct task_struct
, thread_group
);
1601 static inline int thread_group_empty(struct task_struct
*p
)
1603 return list_empty(&p
->thread_group
);
1606 #define delay_group_leader(p) \
1607 (thread_group_leader(p) && !thread_group_empty(p))
1610 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
1611 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1612 * pins the final release of task.io_context. Also protects ->cpuset.
1614 * Nests both inside and outside of read_lock(&tasklist_lock).
1615 * It must not be nested with write_lock_irq(&tasklist_lock),
1616 * neither inside nor outside.
1618 static inline void task_lock(struct task_struct
*p
)
1620 spin_lock(&p
->alloc_lock
);
1623 static inline void task_unlock(struct task_struct
*p
)
1625 spin_unlock(&p
->alloc_lock
);
1628 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
1629 unsigned long *flags
);
1631 static inline void unlock_task_sighand(struct task_struct
*tsk
,
1632 unsigned long *flags
)
1634 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
1637 #ifndef __HAVE_THREAD_FUNCTIONS
1639 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
1640 #define task_stack_page(task) ((task)->stack)
1642 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
1644 *task_thread_info(p
) = *task_thread_info(org
);
1645 task_thread_info(p
)->task
= p
;
1648 static inline unsigned long *end_of_stack(struct task_struct
*p
)
1650 return (unsigned long *)(task_thread_info(p
) + 1);
1655 /* set thread flags in other task's structures
1656 * - see asm/thread_info.h for TIF_xxxx flags available
1658 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1660 set_ti_thread_flag(task_thread_info(tsk
), flag
);
1663 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1665 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1668 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1670 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
1673 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1675 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1678 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1680 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
1683 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
1685 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1688 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
1690 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1693 static inline int signal_pending(struct task_struct
*p
)
1695 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
1698 static inline int need_resched(void)
1700 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
1704 * cond_resched() and cond_resched_lock(): latency reduction via
1705 * explicit rescheduling in places that are safe. The return
1706 * value indicates whether a reschedule was done in fact.
1707 * cond_resched_lock() will drop the spinlock before scheduling,
1708 * cond_resched_softirq() will enable bhs before scheduling.
1710 extern int cond_resched(void);
1711 extern int cond_resched_lock(spinlock_t
* lock
);
1712 extern int cond_resched_softirq(void);
1715 * Does a critical section need to be broken due to another
1718 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1719 # define need_lockbreak(lock) ((lock)->break_lock)
1721 # define need_lockbreak(lock) 0
1725 * Does a critical section need to be broken due to another
1726 * task waiting or preemption being signalled:
1728 static inline int lock_need_resched(spinlock_t
*lock
)
1730 if (need_lockbreak(lock
) || need_resched())
1736 * Reevaluate whether the task has signals pending delivery.
1737 * Wake the task if so.
1738 * This is required every time the blocked sigset_t changes.
1739 * callers must hold sighand->siglock.
1741 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
1742 extern void recalc_sigpending(void);
1744 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
1747 * Wrappers for p->thread_info->cpu access. No-op on UP.
1751 static inline unsigned int task_cpu(const struct task_struct
*p
)
1753 return task_thread_info(p
)->cpu
;
1756 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
1760 static inline unsigned int task_cpu(const struct task_struct
*p
)
1765 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1769 #endif /* CONFIG_SMP */
1771 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1772 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
1774 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
)
1776 mm
->mmap_base
= TASK_UNMAPPED_BASE
;
1777 mm
->get_unmapped_area
= arch_get_unmapped_area
;
1778 mm
->unmap_area
= arch_unmap_area
;
1782 extern long sched_setaffinity(pid_t pid
, cpumask_t new_mask
);
1783 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
1785 extern int sched_mc_power_savings
, sched_smt_power_savings
;
1787 extern void normalize_rt_tasks(void);
1789 #ifdef CONFIG_TASK_XACCT
1790 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1795 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1800 static inline void inc_syscr(struct task_struct
*tsk
)
1805 static inline void inc_syscw(struct task_struct
*tsk
)
1810 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1814 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1818 static inline void inc_syscr(struct task_struct
*tsk
)
1822 static inline void inc_syscw(struct task_struct
*tsk
)
1827 #endif /* __KERNEL__ */