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 */
33 #define SCHED_NORMAL 0
37 /* SCHED_ISO: reserved but not implemented yet */
46 #include <asm/param.h> /* for HZ */
48 #include <linux/capability.h>
49 #include <linux/threads.h>
50 #include <linux/kernel.h>
51 #include <linux/types.h>
52 #include <linux/timex.h>
53 #include <linux/jiffies.h>
54 #include <linux/rbtree.h>
55 #include <linux/thread_info.h>
56 #include <linux/cpumask.h>
57 #include <linux/errno.h>
58 #include <linux/nodemask.h>
60 #include <asm/system.h>
61 #include <asm/semaphore.h>
63 #include <asm/ptrace.h>
65 #include <asm/cputime.h>
67 #include <linux/smp.h>
68 #include <linux/sem.h>
69 #include <linux/signal.h>
70 #include <linux/securebits.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/seccomp.h>
78 #include <linux/rcupdate.h>
79 #include <linux/futex.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>
89 #include <asm/processor.h>
92 struct futex_pi_state
;
96 * List of flags we want to share for kernel threads,
97 * if only because they are not used by them anyway.
99 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
102 * These are the constant used to fake the fixed-point load-average
103 * counting. Some notes:
104 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
105 * a load-average precision of 10 bits integer + 11 bits fractional
106 * - if you want to count load-averages more often, you need more
107 * precision, or rounding will get you. With 2-second counting freq,
108 * the EXP_n values would be 1981, 2034 and 2043 if still using only
111 extern unsigned long avenrun
[]; /* Load averages */
113 #define FSHIFT 11 /* nr of bits of precision */
114 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
115 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
116 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
117 #define EXP_5 2014 /* 1/exp(5sec/5min) */
118 #define EXP_15 2037 /* 1/exp(5sec/15min) */
120 #define CALC_LOAD(load,exp,n) \
122 load += n*(FIXED_1-exp); \
125 extern unsigned long total_forks
;
126 extern int nr_threads
;
127 DECLARE_PER_CPU(unsigned long, process_counts
);
128 extern int nr_processes(void);
129 extern unsigned long nr_running(void);
130 extern unsigned long nr_uninterruptible(void);
131 extern unsigned long nr_active(void);
132 extern unsigned long nr_iowait(void);
133 extern unsigned long weighted_cpuload(const int cpu
);
137 * Task state bitmask. NOTE! These bits are also
138 * encoded in fs/proc/array.c: get_task_state().
140 * We have two separate sets of flags: task->state
141 * is about runnability, while task->exit_state are
142 * about the task exiting. Confusing, but this way
143 * modifying one set can't modify the other one by
146 #define TASK_RUNNING 0
147 #define TASK_INTERRUPTIBLE 1
148 #define TASK_UNINTERRUPTIBLE 2
149 #define TASK_STOPPED 4
150 #define TASK_TRACED 8
151 /* in tsk->exit_state */
152 #define EXIT_ZOMBIE 16
154 /* in tsk->state again */
155 #define TASK_NONINTERACTIVE 64
156 #define TASK_DEAD 128
158 #define __set_task_state(tsk, state_value) \
159 do { (tsk)->state = (state_value); } while (0)
160 #define set_task_state(tsk, state_value) \
161 set_mb((tsk)->state, (state_value))
164 * set_current_state() includes a barrier so that the write of current->state
165 * is correctly serialised wrt the caller's subsequent test of whether to
168 * set_current_state(TASK_UNINTERRUPTIBLE);
169 * if (do_i_need_to_sleep())
172 * If the caller does not need such serialisation then use __set_current_state()
174 #define __set_current_state(state_value) \
175 do { current->state = (state_value); } while (0)
176 #define set_current_state(state_value) \
177 set_mb(current->state, (state_value))
179 /* Task command name length */
180 #define TASK_COMM_LEN 16
182 #include <linux/spinlock.h>
185 * This serializes "schedule()" and also protects
186 * the run-queue from deletions/modifications (but
187 * _adding_ to the beginning of the run-queue has
190 extern rwlock_t tasklist_lock
;
191 extern spinlock_t mmlist_lock
;
195 extern void sched_init(void);
196 extern void sched_init_smp(void);
197 extern void init_idle(struct task_struct
*idle
, int cpu
);
198 extern void init_idle_bootup_task(struct task_struct
*idle
);
200 extern cpumask_t nohz_cpu_mask
;
201 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
202 extern int select_nohz_load_balancer(int cpu
);
204 static inline int select_nohz_load_balancer(int cpu
)
211 * Only dump TASK_* tasks. (0 for all tasks)
213 extern void show_state_filter(unsigned long state_filter
);
215 static inline void show_state(void)
217 show_state_filter(0);
220 extern void show_regs(struct pt_regs
*);
223 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
224 * task), SP is the stack pointer of the first frame that should be shown in the back
225 * trace (or NULL if the entire call-chain of the task should be shown).
227 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
229 void io_schedule(void);
230 long io_schedule_timeout(long timeout
);
232 extern void cpu_init (void);
233 extern void trap_init(void);
234 extern void update_process_times(int user
);
235 extern void scheduler_tick(void);
237 #ifdef CONFIG_DETECT_SOFTLOCKUP
238 extern void softlockup_tick(void);
239 extern void spawn_softlockup_task(void);
240 extern void touch_softlockup_watchdog(void);
241 extern void touch_all_softlockup_watchdogs(void);
243 static inline void softlockup_tick(void)
246 static inline void spawn_softlockup_task(void)
249 static inline void touch_softlockup_watchdog(void)
252 static inline void touch_all_softlockup_watchdogs(void)
258 /* Attach to any functions which should be ignored in wchan output. */
259 #define __sched __attribute__((__section__(".sched.text")))
260 /* Is this address in the __sched functions? */
261 extern int in_sched_functions(unsigned long addr
);
263 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
264 extern signed long FASTCALL(schedule_timeout(signed long timeout
));
265 extern signed long schedule_timeout_interruptible(signed long timeout
);
266 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
267 asmlinkage
void schedule(void);
271 /* Maximum number of active map areas.. This is a random (large) number */
272 #define DEFAULT_MAX_MAP_COUNT 65536
274 extern int sysctl_max_map_count
;
276 #include <linux/aio.h>
279 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
280 unsigned long, unsigned long);
282 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
283 unsigned long len
, unsigned long pgoff
,
284 unsigned long flags
);
285 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
286 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
288 #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
290 * The mm counters are not protected by its page_table_lock,
291 * so must be incremented atomically.
293 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
294 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
295 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
296 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
297 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
298 typedef atomic_long_t mm_counter_t
;
300 #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
302 * The mm counters are protected by its page_table_lock,
303 * so can be incremented directly.
305 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
306 #define get_mm_counter(mm, member) ((mm)->_##member)
307 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
308 #define inc_mm_counter(mm, member) (mm)->_##member++
309 #define dec_mm_counter(mm, member) (mm)->_##member--
310 typedef unsigned long mm_counter_t
;
312 #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
314 #define get_mm_rss(mm) \
315 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
316 #define update_hiwater_rss(mm) do { \
317 unsigned long _rss = get_mm_rss(mm); \
318 if ((mm)->hiwater_rss < _rss) \
319 (mm)->hiwater_rss = _rss; \
321 #define update_hiwater_vm(mm) do { \
322 if ((mm)->hiwater_vm < (mm)->total_vm) \
323 (mm)->hiwater_vm = (mm)->total_vm; \
327 struct vm_area_struct
* mmap
; /* list of VMAs */
328 struct rb_root mm_rb
;
329 struct vm_area_struct
* mmap_cache
; /* last find_vma result */
330 unsigned long (*get_unmapped_area
) (struct file
*filp
,
331 unsigned long addr
, unsigned long len
,
332 unsigned long pgoff
, unsigned long flags
);
333 void (*unmap_area
) (struct mm_struct
*mm
, unsigned long addr
);
334 unsigned long mmap_base
; /* base of mmap area */
335 unsigned long task_size
; /* size of task vm space */
336 unsigned long cached_hole_size
; /* if non-zero, the largest hole below free_area_cache */
337 unsigned long free_area_cache
; /* first hole of size cached_hole_size or larger */
339 atomic_t mm_users
; /* How many users with user space? */
340 atomic_t mm_count
; /* How many references to "struct mm_struct" (users count as 1) */
341 int map_count
; /* number of VMAs */
342 struct rw_semaphore mmap_sem
;
343 spinlock_t page_table_lock
; /* Protects page tables and some counters */
345 struct list_head mmlist
; /* List of maybe swapped mm's. These are globally strung
346 * together off init_mm.mmlist, and are protected
350 /* Special counters, in some configurations protected by the
351 * page_table_lock, in other configurations by being atomic.
353 mm_counter_t _file_rss
;
354 mm_counter_t _anon_rss
;
356 unsigned long hiwater_rss
; /* High-watermark of RSS usage */
357 unsigned long hiwater_vm
; /* High-water virtual memory usage */
359 unsigned long total_vm
, locked_vm
, shared_vm
, exec_vm
;
360 unsigned long stack_vm
, reserved_vm
, def_flags
, nr_ptes
;
361 unsigned long start_code
, end_code
, start_data
, end_data
;
362 unsigned long start_brk
, brk
, start_stack
;
363 unsigned long arg_start
, arg_end
, env_start
, env_end
;
365 unsigned long saved_auxv
[AT_VECTOR_SIZE
]; /* for /proc/PID/auxv */
367 cpumask_t cpu_vm_mask
;
369 /* Architecture-specific MM context */
370 mm_context_t context
;
372 /* Swap token stuff */
374 * Last value of global fault stamp as seen by this process.
375 * In other words, this value gives an indication of how long
376 * it has been since this task got the token.
377 * Look at mm/thrash.c
379 unsigned int faultstamp
;
380 unsigned int token_priority
;
381 unsigned int last_interval
;
383 unsigned char dumpable
:2;
385 /* coredumping support */
387 struct completion
*core_startup_done
, core_done
;
390 rwlock_t ioctx_list_lock
;
391 struct kioctx
*ioctx_list
;
394 struct sighand_struct
{
396 struct k_sigaction action
[_NSIG
];
398 struct list_head signalfd_list
;
401 struct pacct_struct
{
404 unsigned long ac_mem
;
405 cputime_t ac_utime
, ac_stime
;
406 unsigned long ac_minflt
, ac_majflt
;
410 * NOTE! "signal_struct" does not have it's own
411 * locking, because a shared signal_struct always
412 * implies a shared sighand_struct, so locking
413 * sighand_struct is always a proper superset of
414 * the locking of signal_struct.
416 struct signal_struct
{
420 wait_queue_head_t wait_chldexit
; /* for wait4() */
422 /* current thread group signal load-balancing target: */
423 struct task_struct
*curr_target
;
425 /* shared signal handling: */
426 struct sigpending shared_pending
;
428 /* thread group exit support */
431 * - notify group_exit_task when ->count is equal to notify_count
432 * - everyone except group_exit_task is stopped during signal delivery
433 * of fatal signals, group_exit_task processes the signal.
435 struct task_struct
*group_exit_task
;
438 /* thread group stop support, overloads group_exit_code too */
439 int group_stop_count
;
440 unsigned int flags
; /* see SIGNAL_* flags below */
442 /* POSIX.1b Interval Timers */
443 struct list_head posix_timers
;
445 /* ITIMER_REAL timer for the process */
446 struct hrtimer real_timer
;
447 struct task_struct
*tsk
;
448 ktime_t it_real_incr
;
450 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
451 cputime_t it_prof_expires
, it_virt_expires
;
452 cputime_t it_prof_incr
, it_virt_incr
;
454 /* job control IDs */
456 struct pid
*tty_old_pgrp
;
459 pid_t session __deprecated
;
463 /* boolean value for session group leader */
466 struct tty_struct
*tty
; /* NULL if no tty */
469 * Cumulative resource counters for dead threads in the group,
470 * and for reaped dead child processes forked by this group.
471 * Live threads maintain their own counters and add to these
472 * in __exit_signal, except for the group leader.
474 cputime_t utime
, stime
, cutime
, cstime
;
475 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
476 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
477 unsigned long inblock
, oublock
, cinblock
, coublock
;
480 * Cumulative ns of scheduled CPU time for dead threads in the
481 * group, not including a zombie group leader. (This only differs
482 * from jiffies_to_ns(utime + stime) if sched_clock uses something
483 * other than jiffies.)
485 unsigned long sched_time
;
486 unsigned long long sum_sched_runtime
;
489 * We don't bother to synchronize most readers of this at all,
490 * because there is no reader checking a limit that actually needs
491 * to get both rlim_cur and rlim_max atomically, and either one
492 * alone is a single word that can safely be read normally.
493 * getrlimit/setrlimit use task_lock(current->group_leader) to
494 * protect this instead of the siglock, because they really
495 * have no need to disable irqs.
497 struct rlimit rlim
[RLIM_NLIMITS
];
499 struct list_head cpu_timers
[3];
501 /* keep the process-shared keyrings here so that they do the right
502 * thing in threads created with CLONE_THREAD */
504 struct key
*session_keyring
; /* keyring inherited over fork */
505 struct key
*process_keyring
; /* keyring private to this process */
507 #ifdef CONFIG_BSD_PROCESS_ACCT
508 struct pacct_struct pacct
; /* per-process accounting information */
510 #ifdef CONFIG_TASKSTATS
511 struct taskstats
*stats
;
515 /* Context switch must be unlocked if interrupts are to be enabled */
516 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
517 # define __ARCH_WANT_UNLOCKED_CTXSW
521 * Bits in flags field of signal_struct.
523 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
524 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
525 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
526 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
529 * Some day this will be a full-fledged user tracking system..
532 atomic_t __count
; /* reference count */
533 atomic_t processes
; /* How many processes does this user have? */
534 atomic_t files
; /* How many open files does this user have? */
535 atomic_t sigpending
; /* How many pending signals does this user have? */
536 #ifdef CONFIG_INOTIFY_USER
537 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
538 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
540 /* protected by mq_lock */
541 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
542 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
545 struct key
*uid_keyring
; /* UID specific keyring */
546 struct key
*session_keyring
; /* UID's default session keyring */
549 /* Hash table maintenance information */
550 struct list_head uidhash_list
;
554 extern struct user_struct
*find_user(uid_t
);
556 extern struct user_struct root_user
;
557 #define INIT_USER (&root_user)
559 struct backing_dev_info
;
560 struct reclaim_state
;
562 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
564 /* cumulative counters */
565 unsigned long cpu_time
, /* time spent on the cpu */
566 run_delay
, /* time spent waiting on a runqueue */
567 pcnt
; /* # of timeslices run on this cpu */
570 unsigned long last_arrival
, /* when we last ran on a cpu */
571 last_queued
; /* when we were last queued to run */
573 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
575 #ifdef CONFIG_SCHEDSTATS
576 extern const struct file_operations proc_schedstat_operations
;
577 #endif /* CONFIG_SCHEDSTATS */
579 #ifdef CONFIG_TASK_DELAY_ACCT
580 struct task_delay_info
{
582 unsigned int flags
; /* Private per-task flags */
584 /* For each stat XXX, add following, aligned appropriately
586 * struct timespec XXX_start, XXX_end;
590 * Atomicity of updates to XXX_delay, XXX_count protected by
591 * single lock above (split into XXX_lock if contention is an issue).
595 * XXX_count is incremented on every XXX operation, the delay
596 * associated with the operation is added to XXX_delay.
597 * XXX_delay contains the accumulated delay time in nanoseconds.
599 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
600 u64 blkio_delay
; /* wait for sync block io completion */
601 u64 swapin_delay
; /* wait for swapin block io completion */
602 u32 blkio_count
; /* total count of the number of sync block */
603 /* io operations performed */
604 u32 swapin_count
; /* total count of the number of swapin block */
605 /* io operations performed */
607 #endif /* CONFIG_TASK_DELAY_ACCT */
609 static inline int sched_info_on(void)
611 #ifdef CONFIG_SCHEDSTATS
613 #elif defined(CONFIG_TASK_DELAY_ACCT)
614 extern int delayacct_on
;
629 * sched-domains (multiprocessor balancing) declarations:
633 * Increase resolution of nice-level calculations:
635 #define SCHED_LOAD_SHIFT 10
636 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
638 #define SCHED_LOAD_SCALE_FUZZ (SCHED_LOAD_SCALE >> 5)
641 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
642 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
643 #define SD_BALANCE_EXEC 4 /* Balance on exec */
644 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
645 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
646 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
647 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
648 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
649 #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
650 #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
651 #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
653 #define BALANCE_FOR_MC_POWER \
654 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
656 #define BALANCE_FOR_PKG_POWER \
657 ((sched_mc_power_savings || sched_smt_power_savings) ? \
658 SD_POWERSAVINGS_BALANCE : 0)
660 #define test_sd_parent(sd, flag) ((sd->parent && \
661 (sd->parent->flags & flag)) ? 1 : 0)
665 struct sched_group
*next
; /* Must be a circular list */
669 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
670 * single CPU. This is read only (except for setup, hotplug CPU).
671 * Note : Never change cpu_power without recompute its reciprocal
673 unsigned int __cpu_power
;
675 * reciprocal value of cpu_power to avoid expensive divides
676 * (see include/linux/reciprocal_div.h)
678 u32 reciprocal_cpu_power
;
681 struct sched_domain
{
682 /* These fields must be setup */
683 struct sched_domain
*parent
; /* top domain must be null terminated */
684 struct sched_domain
*child
; /* bottom domain must be null terminated */
685 struct sched_group
*groups
; /* the balancing groups of the domain */
686 cpumask_t span
; /* span of all CPUs in this domain */
687 unsigned long min_interval
; /* Minimum balance interval ms */
688 unsigned long max_interval
; /* Maximum balance interval ms */
689 unsigned int busy_factor
; /* less balancing by factor if busy */
690 unsigned int imbalance_pct
; /* No balance until over watermark */
691 unsigned long long cache_hot_time
; /* Task considered cache hot (ns) */
692 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
693 unsigned int busy_idx
;
694 unsigned int idle_idx
;
695 unsigned int newidle_idx
;
696 unsigned int wake_idx
;
697 unsigned int forkexec_idx
;
698 int flags
; /* See SD_* */
700 /* Runtime fields. */
701 unsigned long last_balance
; /* init to jiffies. units in jiffies */
702 unsigned int balance_interval
; /* initialise to 1. units in ms. */
703 unsigned int nr_balance_failed
; /* initialise to 0 */
705 #ifdef CONFIG_SCHEDSTATS
706 /* load_balance() stats */
707 unsigned long lb_cnt
[CPU_MAX_IDLE_TYPES
];
708 unsigned long lb_failed
[CPU_MAX_IDLE_TYPES
];
709 unsigned long lb_balanced
[CPU_MAX_IDLE_TYPES
];
710 unsigned long lb_imbalance
[CPU_MAX_IDLE_TYPES
];
711 unsigned long lb_gained
[CPU_MAX_IDLE_TYPES
];
712 unsigned long lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
713 unsigned long lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
714 unsigned long lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
716 /* Active load balancing */
717 unsigned long alb_cnt
;
718 unsigned long alb_failed
;
719 unsigned long alb_pushed
;
721 /* SD_BALANCE_EXEC stats */
722 unsigned long sbe_cnt
;
723 unsigned long sbe_balanced
;
724 unsigned long sbe_pushed
;
726 /* SD_BALANCE_FORK stats */
727 unsigned long sbf_cnt
;
728 unsigned long sbf_balanced
;
729 unsigned long sbf_pushed
;
731 /* try_to_wake_up() stats */
732 unsigned long ttwu_wake_remote
;
733 unsigned long ttwu_move_affine
;
734 unsigned long ttwu_move_balance
;
738 extern int partition_sched_domains(cpumask_t
*partition1
,
739 cpumask_t
*partition2
);
741 #endif /* CONFIG_SMP */
744 struct io_context
; /* See blkdev.h */
747 #define NGROUPS_SMALL 32
748 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
752 gid_t small_block
[NGROUPS_SMALL
];
758 * get_group_info() must be called with the owning task locked (via task_lock())
759 * when task != current. The reason being that the vast majority of callers are
760 * looking at current->group_info, which can not be changed except by the
761 * current task. Changing current->group_info requires the task lock, too.
763 #define get_group_info(group_info) do { \
764 atomic_inc(&(group_info)->usage); \
767 #define put_group_info(group_info) do { \
768 if (atomic_dec_and_test(&(group_info)->usage)) \
769 groups_free(group_info); \
772 extern struct group_info
*groups_alloc(int gidsetsize
);
773 extern void groups_free(struct group_info
*group_info
);
774 extern int set_current_groups(struct group_info
*group_info
);
775 extern int groups_search(struct group_info
*group_info
, gid_t grp
);
776 /* access the groups "array" with this macro */
777 #define GROUP_AT(gi, i) \
778 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
780 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
781 extern void prefetch_stack(struct task_struct
*t
);
783 static inline void prefetch_stack(struct task_struct
*t
) { }
786 struct audit_context
; /* See audit.c */
788 struct pipe_inode_info
;
789 struct uts_namespace
;
796 struct sched_class
*next
;
798 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
,
799 int wakeup
, u64 now
);
800 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
,
802 void (*yield_task
) (struct rq
*rq
, struct task_struct
*p
);
804 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
);
806 struct task_struct
* (*pick_next_task
) (struct rq
*rq
, u64 now
);
807 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
, u64 now
);
809 int (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
811 unsigned long max_nr_move
, unsigned long max_load_move
,
812 struct sched_domain
*sd
, enum cpu_idle_type idle
,
813 int *all_pinned
, unsigned long *total_load_moved
);
815 void (*set_curr_task
) (struct rq
*rq
);
816 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
);
817 void (*task_new
) (struct rq
*rq
, struct task_struct
*p
);
821 unsigned long weight
, inv_weight
;
825 * CFS stats for a schedulable entity (task, task-group etc)
827 * Current field usage histogram:
832 * 4 se->sleep_start_fair
835 * 15 se->wait_runtime
837 struct sched_entity
{
839 unsigned long delta_fair_run
;
840 unsigned long delta_fair_sleep
;
841 unsigned long delta_exec
;
843 struct load_weight load
; /* for load-balancing */
844 struct rb_node run_node
;
851 u64 sleep_start_fair
;
859 u64 sum_exec_runtime
;
860 s64 sum_wait_runtime
;
861 s64 sum_sleep_runtime
;
862 unsigned long wait_runtime_overruns
;
863 unsigned long wait_runtime_underruns
;
864 #ifdef CONFIG_FAIR_GROUP_SCHED
865 struct sched_entity
*parent
;
866 /* rq on which this entity is (to be) queued: */
867 struct cfs_rq
*cfs_rq
;
868 /* rq "owned" by this entity/group: */
874 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
877 unsigned int flags
; /* per process flags, defined below */
880 int lock_depth
; /* BKL lock depth */
883 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
887 int load_weight
; /* for niceness load balancing purposes */
888 int prio
, static_prio
, normal_prio
;
889 struct list_head run_list
;
890 struct prio_array
*array
;
891 struct sched_class
*sched_class
;
892 struct sched_entity se
;
894 unsigned short ioprio
;
895 #ifdef CONFIG_BLK_DEV_IO_TRACE
896 unsigned int btrace_seq
;
898 unsigned long sleep_avg
;
899 unsigned long long timestamp
, last_ran
;
900 unsigned long long sched_time
; /* sched_clock time spent running */
903 cpumask_t cpus_allowed
;
904 unsigned int time_slice
, first_time_slice
;
906 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
907 struct sched_info sched_info
;
910 struct list_head tasks
;
912 * ptrace_list/ptrace_children forms the list of my children
913 * that were stolen by a ptracer.
915 struct list_head ptrace_children
;
916 struct list_head ptrace_list
;
918 struct mm_struct
*mm
, *active_mm
;
921 struct linux_binfmt
*binfmt
;
923 int exit_code
, exit_signal
;
924 int pdeath_signal
; /* The signal sent when the parent dies */
926 unsigned int personality
;
931 #ifdef CONFIG_CC_STACKPROTECTOR
932 /* Canary value for the -fstack-protector gcc feature */
933 unsigned long stack_canary
;
936 * pointers to (original) parent process, youngest child, younger sibling,
937 * older sibling, respectively. (p->father can be replaced with
940 struct task_struct
*real_parent
; /* real parent process (when being debugged) */
941 struct task_struct
*parent
; /* parent process */
943 * children/sibling forms the list of my children plus the
944 * tasks I'm ptracing.
946 struct list_head children
; /* list of my children */
947 struct list_head sibling
; /* linkage in my parent's children list */
948 struct task_struct
*group_leader
; /* threadgroup leader */
950 /* PID/PID hash table linkage. */
951 struct pid_link pids
[PIDTYPE_MAX
];
952 struct list_head thread_group
;
954 struct completion
*vfork_done
; /* for vfork() */
955 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
956 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
958 unsigned int rt_priority
;
959 cputime_t utime
, stime
;
960 unsigned long nvcsw
, nivcsw
; /* context switch counts */
961 struct timespec start_time
;
962 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
963 unsigned long min_flt
, maj_flt
;
965 cputime_t it_prof_expires
, it_virt_expires
;
966 unsigned long long it_sched_expires
;
967 struct list_head cpu_timers
[3];
969 /* process credentials */
970 uid_t uid
,euid
,suid
,fsuid
;
971 gid_t gid
,egid
,sgid
,fsgid
;
972 struct group_info
*group_info
;
973 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
;
974 unsigned keep_capabilities
:1;
975 struct user_struct
*user
;
977 struct key
*request_key_auth
; /* assumed request_key authority */
978 struct key
*thread_keyring
; /* keyring private to this thread */
979 unsigned char jit_keyring
; /* default keyring to attach requested keys to */
982 * fpu_counter contains the number of consecutive context switches
983 * that the FPU is used. If this is over a threshold, the lazy fpu
984 * saving becomes unlazy to save the trap. This is an unsigned char
985 * so that after 256 times the counter wraps and the behavior turns
986 * lazy again; this to deal with bursty apps that only use FPU for
989 unsigned char fpu_counter
;
990 int oomkilladj
; /* OOM kill score adjustment (bit shift). */
991 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
992 - access with [gs]et_task_comm (which lock
994 - initialized normally by flush_old_exec */
995 /* file system info */
996 int link_count
, total_link_count
;
997 #ifdef CONFIG_SYSVIPC
999 struct sysv_sem sysvsem
;
1001 /* CPU-specific state of this task */
1002 struct thread_struct thread
;
1003 /* filesystem information */
1004 struct fs_struct
*fs
;
1005 /* open file information */
1006 struct files_struct
*files
;
1008 struct nsproxy
*nsproxy
;
1009 /* signal handlers */
1010 struct signal_struct
*signal
;
1011 struct sighand_struct
*sighand
;
1013 sigset_t blocked
, real_blocked
;
1014 sigset_t saved_sigmask
; /* To be restored with TIF_RESTORE_SIGMASK */
1015 struct sigpending pending
;
1017 unsigned long sas_ss_sp
;
1019 int (*notifier
)(void *priv
);
1020 void *notifier_data
;
1021 sigset_t
*notifier_mask
;
1024 struct audit_context
*audit_context
;
1027 /* Thread group tracking */
1030 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1031 spinlock_t alloc_lock
;
1033 /* Protection of the PI data structures: */
1036 #ifdef CONFIG_RT_MUTEXES
1037 /* PI waiters blocked on a rt_mutex held by this task */
1038 struct plist_head pi_waiters
;
1039 /* Deadlock detection and priority inheritance handling */
1040 struct rt_mutex_waiter
*pi_blocked_on
;
1043 #ifdef CONFIG_DEBUG_MUTEXES
1044 /* mutex deadlock detection */
1045 struct mutex_waiter
*blocked_on
;
1047 #ifdef CONFIG_TRACE_IRQFLAGS
1048 unsigned int irq_events
;
1049 int hardirqs_enabled
;
1050 unsigned long hardirq_enable_ip
;
1051 unsigned int hardirq_enable_event
;
1052 unsigned long hardirq_disable_ip
;
1053 unsigned int hardirq_disable_event
;
1054 int softirqs_enabled
;
1055 unsigned long softirq_disable_ip
;
1056 unsigned int softirq_disable_event
;
1057 unsigned long softirq_enable_ip
;
1058 unsigned int softirq_enable_event
;
1059 int hardirq_context
;
1060 int softirq_context
;
1062 #ifdef CONFIG_LOCKDEP
1063 # define MAX_LOCK_DEPTH 30UL
1066 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1067 unsigned int lockdep_recursion
;
1070 /* journalling filesystem info */
1073 /* stacked block device info */
1074 struct bio
*bio_list
, **bio_tail
;
1077 struct reclaim_state
*reclaim_state
;
1079 struct backing_dev_info
*backing_dev_info
;
1081 struct io_context
*io_context
;
1083 unsigned long ptrace_message
;
1084 siginfo_t
*last_siginfo
; /* For ptrace use. */
1086 * current io wait handle: wait queue entry to use for io waits
1087 * If this thread is processing aio, this points at the waitqueue
1088 * inside the currently handled kiocb. It may be NULL (i.e. default
1089 * to a stack based synchronous wait) if its doing sync IO.
1091 wait_queue_t
*io_wait
;
1092 #ifdef CONFIG_TASK_XACCT
1093 /* i/o counters(bytes read/written, #syscalls */
1094 u64 rchar
, wchar
, syscr
, syscw
;
1096 struct task_io_accounting ioac
;
1097 #if defined(CONFIG_TASK_XACCT)
1098 u64 acct_rss_mem1
; /* accumulated rss usage */
1099 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1100 cputime_t acct_stimexpd
;/* stime since last update */
1103 struct mempolicy
*mempolicy
;
1106 #ifdef CONFIG_CPUSETS
1107 struct cpuset
*cpuset
;
1108 nodemask_t mems_allowed
;
1109 int cpuset_mems_generation
;
1110 int cpuset_mem_spread_rotor
;
1112 struct robust_list_head __user
*robust_list
;
1113 #ifdef CONFIG_COMPAT
1114 struct compat_robust_list_head __user
*compat_robust_list
;
1116 struct list_head pi_state_list
;
1117 struct futex_pi_state
*pi_state_cache
;
1119 atomic_t fs_excl
; /* holding fs exclusive resources */
1120 struct rcu_head rcu
;
1123 * cache last used pipe for splice
1125 struct pipe_inode_info
*splice_pipe
;
1126 #ifdef CONFIG_TASK_DELAY_ACCT
1127 struct task_delay_info
*delays
;
1129 #ifdef CONFIG_FAULT_INJECTION
1135 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1136 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1137 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1138 * values are inverted: lower p->prio value means higher priority.
1140 * The MAX_USER_RT_PRIO value allows the actual maximum
1141 * RT priority to be separate from the value exported to
1142 * user-space. This allows kernel threads to set their
1143 * priority to a value higher than any user task. Note:
1144 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1147 #define MAX_USER_RT_PRIO 100
1148 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1150 #define MAX_PRIO (MAX_RT_PRIO + 40)
1151 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1153 static inline int rt_prio(int prio
)
1155 if (unlikely(prio
< MAX_RT_PRIO
))
1160 static inline int rt_task(struct task_struct
*p
)
1162 return rt_prio(p
->prio
);
1165 static inline int batch_task(struct task_struct
*p
)
1167 return p
->policy
== SCHED_BATCH
;
1170 static inline pid_t
process_group(struct task_struct
*tsk
)
1172 return tsk
->signal
->pgrp
;
1175 static inline pid_t
signal_session(struct signal_struct
*sig
)
1177 return sig
->__session
;
1180 static inline pid_t
process_session(struct task_struct
*tsk
)
1182 return signal_session(tsk
->signal
);
1185 static inline void set_signal_session(struct signal_struct
*sig
, pid_t session
)
1187 sig
->__session
= session
;
1190 static inline struct pid
*task_pid(struct task_struct
*task
)
1192 return task
->pids
[PIDTYPE_PID
].pid
;
1195 static inline struct pid
*task_tgid(struct task_struct
*task
)
1197 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1200 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1202 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1205 static inline struct pid
*task_session(struct task_struct
*task
)
1207 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1211 * pid_alive - check that a task structure is not stale
1212 * @p: Task structure to be checked.
1214 * Test if a process is not yet dead (at most zombie state)
1215 * If pid_alive fails, then pointers within the task structure
1216 * can be stale and must not be dereferenced.
1218 static inline int pid_alive(struct task_struct
*p
)
1220 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1224 * is_init - check if a task structure is init
1225 * @tsk: Task structure to be checked.
1227 * Check if a task structure is the first user space task the kernel created.
1229 static inline int is_init(struct task_struct
*tsk
)
1231 return tsk
->pid
== 1;
1234 extern struct pid
*cad_pid
;
1236 extern void free_task(struct task_struct
*tsk
);
1237 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1239 extern void __put_task_struct(struct task_struct
*t
);
1241 static inline void put_task_struct(struct task_struct
*t
)
1243 if (atomic_dec_and_test(&t
->usage
))
1244 __put_task_struct(t
);
1250 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1251 /* Not implemented yet, only for 486*/
1252 #define PF_STARTING 0x00000002 /* being created */
1253 #define PF_EXITING 0x00000004 /* getting shut down */
1254 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1255 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1256 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1257 #define PF_DUMPCORE 0x00000200 /* dumped core */
1258 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1259 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1260 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1261 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1262 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1263 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1264 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1265 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1266 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1267 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1268 #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1269 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1270 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1271 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1272 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1273 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1274 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1275 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1278 * Only the _current_ task can read/write to tsk->flags, but other
1279 * tasks can access tsk->flags in readonly mode for example
1280 * with tsk_used_math (like during threaded core dumping).
1281 * There is however an exception to this rule during ptrace
1282 * or during fork: the ptracer task is allowed to write to the
1283 * child->flags of its traced child (same goes for fork, the parent
1284 * can write to the child->flags), because we're guaranteed the
1285 * child is not running and in turn not changing child->flags
1286 * at the same time the parent does it.
1288 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1289 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1290 #define clear_used_math() clear_stopped_child_used_math(current)
1291 #define set_used_math() set_stopped_child_used_math(current)
1292 #define conditional_stopped_child_used_math(condition, child) \
1293 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1294 #define conditional_used_math(condition) \
1295 conditional_stopped_child_used_math(condition, current)
1296 #define copy_to_stopped_child_used_math(child) \
1297 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1298 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1299 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1300 #define used_math() tsk_used_math(current)
1303 extern int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
);
1305 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1307 if (!cpu_isset(0, new_mask
))
1313 extern unsigned long long sched_clock(void);
1314 extern unsigned long long
1315 task_sched_runtime(struct task_struct
*task
);
1317 /* sched_exec is called by processes performing an exec */
1319 extern void sched_exec(void);
1321 #define sched_exec() {}
1324 #ifdef CONFIG_HOTPLUG_CPU
1325 extern void idle_task_exit(void);
1327 static inline void idle_task_exit(void) {}
1330 extern void sched_idle_next(void);
1332 extern unsigned int sysctl_sched_granularity
;
1333 extern unsigned int sysctl_sched_wakeup_granularity
;
1334 extern unsigned int sysctl_sched_batch_wakeup_granularity
;
1335 extern unsigned int sysctl_sched_stat_granularity
;
1336 extern unsigned int sysctl_sched_runtime_limit
;
1337 extern unsigned int sysctl_sched_child_runs_first
;
1338 extern unsigned int sysctl_sched_features
;
1340 #ifdef CONFIG_RT_MUTEXES
1341 extern int rt_mutex_getprio(struct task_struct
*p
);
1342 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1343 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1345 static inline int rt_mutex_getprio(struct task_struct
*p
)
1347 return p
->normal_prio
;
1349 # define rt_mutex_adjust_pi(p) do { } while (0)
1352 extern void set_user_nice(struct task_struct
*p
, long nice
);
1353 extern int task_prio(const struct task_struct
*p
);
1354 extern int task_nice(const struct task_struct
*p
);
1355 extern int can_nice(const struct task_struct
*p
, const int nice
);
1356 extern int task_curr(const struct task_struct
*p
);
1357 extern int idle_cpu(int cpu
);
1358 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1359 extern struct task_struct
*idle_task(int cpu
);
1360 extern struct task_struct
*curr_task(int cpu
);
1361 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1366 * The default (Linux) execution domain.
1368 extern struct exec_domain default_exec_domain
;
1370 union thread_union
{
1371 struct thread_info thread_info
;
1372 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1375 #ifndef __HAVE_ARCH_KSTACK_END
1376 static inline int kstack_end(void *addr
)
1378 /* Reliable end of stack detection:
1379 * Some APM bios versions misalign the stack
1381 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1385 extern union thread_union init_thread_union
;
1386 extern struct task_struct init_task
;
1388 extern struct mm_struct init_mm
;
1390 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1391 extern struct task_struct
*find_task_by_pid_type(int type
, int pid
);
1392 extern void __set_special_pids(pid_t session
, pid_t pgrp
);
1394 /* per-UID process charging. */
1395 extern struct user_struct
* alloc_uid(uid_t
);
1396 static inline struct user_struct
*get_uid(struct user_struct
*u
)
1398 atomic_inc(&u
->__count
);
1401 extern void free_uid(struct user_struct
*);
1402 extern void switch_uid(struct user_struct
*);
1404 #include <asm/current.h>
1406 extern void do_timer(unsigned long ticks
);
1408 extern int FASTCALL(wake_up_state(struct task_struct
* tsk
, unsigned int state
));
1409 extern int FASTCALL(wake_up_process(struct task_struct
* tsk
));
1410 extern void FASTCALL(wake_up_new_task(struct task_struct
* tsk
,
1411 unsigned long clone_flags
));
1413 extern void kick_process(struct task_struct
*tsk
);
1415 static inline void kick_process(struct task_struct
*tsk
) { }
1417 extern void FASTCALL(sched_fork(struct task_struct
* p
, int clone_flags
));
1418 extern void FASTCALL(sched_exit(struct task_struct
* p
));
1420 extern int in_group_p(gid_t
);
1421 extern int in_egroup_p(gid_t
);
1423 extern void proc_caches_init(void);
1424 extern void flush_signals(struct task_struct
*);
1425 extern void ignore_signals(struct task_struct
*);
1426 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
1427 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
1429 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
1431 unsigned long flags
;
1434 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
1435 ret
= dequeue_signal(tsk
, mask
, info
);
1436 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
1441 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
1443 extern void unblock_all_signals(void);
1444 extern void release_task(struct task_struct
* p
);
1445 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
1446 extern int send_group_sig_info(int, struct siginfo
*, struct task_struct
*);
1447 extern int force_sigsegv(int, struct task_struct
*);
1448 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
1449 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1450 extern int kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
1451 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
1452 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
1453 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
1454 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
1455 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
1456 extern void do_notify_parent(struct task_struct
*, int);
1457 extern void force_sig(int, struct task_struct
*);
1458 extern void force_sig_specific(int, struct task_struct
*);
1459 extern int send_sig(int, struct task_struct
*, int);
1460 extern void zap_other_threads(struct task_struct
*p
);
1461 extern int kill_proc(pid_t
, int, int);
1462 extern struct sigqueue
*sigqueue_alloc(void);
1463 extern void sigqueue_free(struct sigqueue
*);
1464 extern int send_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1465 extern int send_group_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1466 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
1467 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
1469 static inline int kill_cad_pid(int sig
, int priv
)
1471 return kill_pid(cad_pid
, sig
, priv
);
1474 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1475 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1476 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1477 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1479 static inline int is_si_special(const struct siginfo
*info
)
1481 return info
<= SEND_SIG_FORCED
;
1484 /* True if we are on the alternate signal stack. */
1486 static inline int on_sig_stack(unsigned long sp
)
1488 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
1491 static inline int sas_ss_flags(unsigned long sp
)
1493 return (current
->sas_ss_size
== 0 ? SS_DISABLE
1494 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
1498 * Routines for handling mm_structs
1500 extern struct mm_struct
* mm_alloc(void);
1502 /* mmdrop drops the mm and the page tables */
1503 extern void FASTCALL(__mmdrop(struct mm_struct
*));
1504 static inline void mmdrop(struct mm_struct
* mm
)
1506 if (atomic_dec_and_test(&mm
->mm_count
))
1510 /* mmput gets rid of the mappings and all user-space */
1511 extern void mmput(struct mm_struct
*);
1512 /* Grab a reference to a task's mm, if it is not already going away */
1513 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1514 /* Remove the current tasks stale references to the old mm_struct */
1515 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1517 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1518 extern void flush_thread(void);
1519 extern void exit_thread(void);
1521 extern void exit_files(struct task_struct
*);
1522 extern void __cleanup_signal(struct signal_struct
*);
1523 extern void __cleanup_sighand(struct sighand_struct
*);
1524 extern void exit_itimers(struct signal_struct
*);
1526 extern NORET_TYPE
void do_group_exit(int);
1528 extern void daemonize(const char *, ...);
1529 extern int allow_signal(int);
1530 extern int disallow_signal(int);
1532 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1533 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1534 struct task_struct
*fork_idle(int);
1536 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1537 extern void get_task_comm(char *to
, struct task_struct
*tsk
);
1540 extern void wait_task_inactive(struct task_struct
* p
);
1542 #define wait_task_inactive(p) do { } while (0)
1545 #define remove_parent(p) list_del_init(&(p)->sibling)
1546 #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1548 #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1550 #define for_each_process(p) \
1551 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1554 * Careful: do_each_thread/while_each_thread is a double loop so
1555 * 'break' will not work as expected - use goto instead.
1557 #define do_each_thread(g, t) \
1558 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1560 #define while_each_thread(g, t) \
1561 while ((t = next_thread(t)) != g)
1563 /* de_thread depends on thread_group_leader not being a pid based check */
1564 #define thread_group_leader(p) (p == p->group_leader)
1566 /* Do to the insanities of de_thread it is possible for a process
1567 * to have the pid of the thread group leader without actually being
1568 * the thread group leader. For iteration through the pids in proc
1569 * all we care about is that we have a task with the appropriate
1570 * pid, we don't actually care if we have the right task.
1572 static inline int has_group_leader_pid(struct task_struct
*p
)
1574 return p
->pid
== p
->tgid
;
1577 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
1579 return list_entry(rcu_dereference(p
->thread_group
.next
),
1580 struct task_struct
, thread_group
);
1583 static inline int thread_group_empty(struct task_struct
*p
)
1585 return list_empty(&p
->thread_group
);
1588 #define delay_group_leader(p) \
1589 (thread_group_leader(p) && !thread_group_empty(p))
1592 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
1593 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1594 * pins the final release of task.io_context. Also protects ->cpuset.
1596 * Nests both inside and outside of read_lock(&tasklist_lock).
1597 * It must not be nested with write_lock_irq(&tasklist_lock),
1598 * neither inside nor outside.
1600 static inline void task_lock(struct task_struct
*p
)
1602 spin_lock(&p
->alloc_lock
);
1605 static inline void task_unlock(struct task_struct
*p
)
1607 spin_unlock(&p
->alloc_lock
);
1610 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
1611 unsigned long *flags
);
1613 static inline void unlock_task_sighand(struct task_struct
*tsk
,
1614 unsigned long *flags
)
1616 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
1619 #ifndef __HAVE_THREAD_FUNCTIONS
1621 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
1622 #define task_stack_page(task) ((task)->stack)
1624 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
1626 *task_thread_info(p
) = *task_thread_info(org
);
1627 task_thread_info(p
)->task
= p
;
1630 static inline unsigned long *end_of_stack(struct task_struct
*p
)
1632 return (unsigned long *)(task_thread_info(p
) + 1);
1637 /* set thread flags in other task's structures
1638 * - see asm/thread_info.h for TIF_xxxx flags available
1640 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1642 set_ti_thread_flag(task_thread_info(tsk
), flag
);
1645 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1647 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1650 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1652 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
1655 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1657 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
1660 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1662 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
1665 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
1667 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1670 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
1672 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1675 static inline int signal_pending(struct task_struct
*p
)
1677 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
1680 static inline int need_resched(void)
1682 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
1686 * cond_resched() and cond_resched_lock(): latency reduction via
1687 * explicit rescheduling in places that are safe. The return
1688 * value indicates whether a reschedule was done in fact.
1689 * cond_resched_lock() will drop the spinlock before scheduling,
1690 * cond_resched_softirq() will enable bhs before scheduling.
1692 extern int cond_resched(void);
1693 extern int cond_resched_lock(spinlock_t
* lock
);
1694 extern int cond_resched_softirq(void);
1697 * Does a critical section need to be broken due to another
1700 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1701 # define need_lockbreak(lock) ((lock)->break_lock)
1703 # define need_lockbreak(lock) 0
1707 * Does a critical section need to be broken due to another
1708 * task waiting or preemption being signalled:
1710 static inline int lock_need_resched(spinlock_t
*lock
)
1712 if (need_lockbreak(lock
) || need_resched())
1718 * Reevaluate whether the task has signals pending delivery.
1719 * Wake the task if so.
1720 * This is required every time the blocked sigset_t changes.
1721 * callers must hold sighand->siglock.
1723 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
1724 extern void recalc_sigpending(void);
1726 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
1729 * Wrappers for p->thread_info->cpu access. No-op on UP.
1733 static inline unsigned int task_cpu(const struct task_struct
*p
)
1735 return task_thread_info(p
)->cpu
;
1738 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
1742 static inline unsigned int task_cpu(const struct task_struct
*p
)
1747 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1751 #endif /* CONFIG_SMP */
1753 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1754 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
1756 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
)
1758 mm
->mmap_base
= TASK_UNMAPPED_BASE
;
1759 mm
->get_unmapped_area
= arch_get_unmapped_area
;
1760 mm
->unmap_area
= arch_unmap_area
;
1764 extern long sched_setaffinity(pid_t pid
, cpumask_t new_mask
);
1765 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
1767 extern int sched_mc_power_savings
, sched_smt_power_savings
;
1769 extern void normalize_rt_tasks(void);
1771 #ifdef CONFIG_TASK_XACCT
1772 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1777 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1782 static inline void inc_syscr(struct task_struct
*tsk
)
1787 static inline void inc_syscw(struct task_struct
*tsk
)
1792 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
1796 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
1800 static inline void inc_syscr(struct task_struct
*tsk
)
1804 static inline void inc_syscw(struct task_struct
*tsk
)
1809 #endif /* __KERNEL__ */