[PATCH] lightweight robust futexes: docs
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
4#include <asm/param.h> /* for HZ */
5
6#include <linux/config.h>
7#include <linux/capability.h>
8#include <linux/threads.h>
9#include <linux/kernel.h>
10#include <linux/types.h>
11#include <linux/timex.h>
12#include <linux/jiffies.h>
13#include <linux/rbtree.h>
14#include <linux/thread_info.h>
15#include <linux/cpumask.h>
16#include <linux/errno.h>
17#include <linux/nodemask.h>
18
19#include <asm/system.h>
20#include <asm/semaphore.h>
21#include <asm/page.h>
22#include <asm/ptrace.h>
23#include <asm/mmu.h>
24#include <asm/cputime.h>
25
26#include <linux/smp.h>
27#include <linux/sem.h>
28#include <linux/signal.h>
29#include <linux/securebits.h>
30#include <linux/fs_struct.h>
31#include <linux/compiler.h>
32#include <linux/completion.h>
33#include <linux/pid.h>
34#include <linux/percpu.h>
35#include <linux/topology.h>
36#include <linux/seccomp.h>
e56d0903 37#include <linux/rcupdate.h>
0771dfef 38#include <linux/futex.h>
1da177e4 39
36d57ac4
L
40#include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
41
1da177e4
LT
42struct exec_domain;
43
44/*
45 * cloning flags:
46 */
47#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
48#define CLONE_VM 0x00000100 /* set if VM shared between processes */
49#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
50#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
51#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
52#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
53#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
54#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
55#define CLONE_THREAD 0x00010000 /* Same thread group? */
56#define CLONE_NEWNS 0x00020000 /* New namespace group? */
57#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
58#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
59#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
60#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
61#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
62#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
63#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
64#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
65
66/*
67 * List of flags we want to share for kernel threads,
68 * if only because they are not used by them anyway.
69 */
70#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
71
72/*
73 * These are the constant used to fake the fixed-point load-average
74 * counting. Some notes:
75 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
76 * a load-average precision of 10 bits integer + 11 bits fractional
77 * - if you want to count load-averages more often, you need more
78 * precision, or rounding will get you. With 2-second counting freq,
79 * the EXP_n values would be 1981, 2034 and 2043 if still using only
80 * 11 bit fractions.
81 */
82extern unsigned long avenrun[]; /* Load averages */
83
84#define FSHIFT 11 /* nr of bits of precision */
85#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
86#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
87#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
88#define EXP_5 2014 /* 1/exp(5sec/5min) */
89#define EXP_15 2037 /* 1/exp(5sec/15min) */
90
91#define CALC_LOAD(load,exp,n) \
92 load *= exp; \
93 load += n*(FIXED_1-exp); \
94 load >>= FSHIFT;
95
96extern unsigned long total_forks;
97extern int nr_threads;
98extern int last_pid;
99DECLARE_PER_CPU(unsigned long, process_counts);
100extern int nr_processes(void);
101extern unsigned long nr_running(void);
102extern unsigned long nr_uninterruptible(void);
103extern unsigned long nr_iowait(void);
104
105#include <linux/time.h>
106#include <linux/param.h>
107#include <linux/resource.h>
108#include <linux/timer.h>
2ff678b8 109#include <linux/hrtimer.h>
1da177e4
LT
110
111#include <asm/processor.h>
112
4a8342d2
LT
113/*
114 * Task state bitmask. NOTE! These bits are also
115 * encoded in fs/proc/array.c: get_task_state().
116 *
117 * We have two separate sets of flags: task->state
118 * is about runnability, while task->exit_state are
119 * about the task exiting. Confusing, but this way
120 * modifying one set can't modify the other one by
121 * mistake.
122 */
1da177e4
LT
123#define TASK_RUNNING 0
124#define TASK_INTERRUPTIBLE 1
125#define TASK_UNINTERRUPTIBLE 2
4a8342d2
LT
126#define TASK_STOPPED 4
127#define TASK_TRACED 8
128/* in tsk->exit_state */
129#define EXIT_ZOMBIE 16
130#define EXIT_DEAD 32
131/* in tsk->state again */
132#define TASK_NONINTERACTIVE 64
1da177e4
LT
133
134#define __set_task_state(tsk, state_value) \
135 do { (tsk)->state = (state_value); } while (0)
136#define set_task_state(tsk, state_value) \
137 set_mb((tsk)->state, (state_value))
138
498d0c57
AM
139/*
140 * set_current_state() includes a barrier so that the write of current->state
141 * is correctly serialised wrt the caller's subsequent test of whether to
142 * actually sleep:
143 *
144 * set_current_state(TASK_UNINTERRUPTIBLE);
145 * if (do_i_need_to_sleep())
146 * schedule();
147 *
148 * If the caller does not need such serialisation then use __set_current_state()
149 */
1da177e4
LT
150#define __set_current_state(state_value) \
151 do { current->state = (state_value); } while (0)
152#define set_current_state(state_value) \
153 set_mb(current->state, (state_value))
154
155/* Task command name length */
156#define TASK_COMM_LEN 16
157
158/*
159 * Scheduling policies
160 */
161#define SCHED_NORMAL 0
162#define SCHED_FIFO 1
163#define SCHED_RR 2
b0a9499c 164#define SCHED_BATCH 3
1da177e4
LT
165
166struct sched_param {
167 int sched_priority;
168};
169
170#ifdef __KERNEL__
171
172#include <linux/spinlock.h>
173
174/*
175 * This serializes "schedule()" and also protects
176 * the run-queue from deletions/modifications (but
177 * _adding_ to the beginning of the run-queue has
178 * a separate lock).
179 */
180extern rwlock_t tasklist_lock;
181extern spinlock_t mmlist_lock;
182
183typedef struct task_struct task_t;
184
185extern void sched_init(void);
186extern void sched_init_smp(void);
187extern void init_idle(task_t *idle, int cpu);
188
189extern cpumask_t nohz_cpu_mask;
190
191extern void show_state(void);
192extern void show_regs(struct pt_regs *);
193
194/*
195 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
196 * task), SP is the stack pointer of the first frame that should be shown in the back
197 * trace (or NULL if the entire call-chain of the task should be shown).
198 */
199extern void show_stack(struct task_struct *task, unsigned long *sp);
200
201void io_schedule(void);
202long io_schedule_timeout(long timeout);
203
204extern void cpu_init (void);
205extern void trap_init(void);
206extern void update_process_times(int user);
207extern void scheduler_tick(void);
208
8446f1d3 209#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 210extern void softlockup_tick(void);
8446f1d3
IM
211extern void spawn_softlockup_task(void);
212extern void touch_softlockup_watchdog(void);
213#else
6687a97d 214static inline void softlockup_tick(void)
8446f1d3
IM
215{
216}
217static inline void spawn_softlockup_task(void)
218{
219}
220static inline void touch_softlockup_watchdog(void)
221{
222}
223#endif
224
225
1da177e4
LT
226/* Attach to any functions which should be ignored in wchan output. */
227#define __sched __attribute__((__section__(".sched.text")))
228/* Is this address in the __sched functions? */
229extern int in_sched_functions(unsigned long addr);
230
231#define MAX_SCHEDULE_TIMEOUT LONG_MAX
232extern signed long FASTCALL(schedule_timeout(signed long timeout));
64ed93a2
NA
233extern signed long schedule_timeout_interruptible(signed long timeout);
234extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
235asmlinkage void schedule(void);
236
237struct namespace;
238
239/* Maximum number of active map areas.. This is a random (large) number */
240#define DEFAULT_MAX_MAP_COUNT 65536
241
242extern int sysctl_max_map_count;
243
244#include <linux/aio.h>
245
246extern unsigned long
247arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
248 unsigned long, unsigned long);
249extern unsigned long
250arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
251 unsigned long len, unsigned long pgoff,
252 unsigned long flags);
1363c3cd
WW
253extern void arch_unmap_area(struct mm_struct *, unsigned long);
254extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 255
f412ac08
HD
256#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
257/*
258 * The mm counters are not protected by its page_table_lock,
259 * so must be incremented atomically.
260 */
d3cb4871
CL
261#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
262#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
263#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
264#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
265#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
266typedef atomic_long_t mm_counter_t;
f412ac08
HD
267
268#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
269/*
270 * The mm counters are protected by its page_table_lock,
271 * so can be incremented directly.
272 */
1da177e4
LT
273#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
274#define get_mm_counter(mm, member) ((mm)->_##member)
275#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
276#define inc_mm_counter(mm, member) (mm)->_##member++
277#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
278typedef unsigned long mm_counter_t;
279
280#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 281
f412ac08
HD
282#define get_mm_rss(mm) \
283 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
284#define update_hiwater_rss(mm) do { \
285 unsigned long _rss = get_mm_rss(mm); \
286 if ((mm)->hiwater_rss < _rss) \
287 (mm)->hiwater_rss = _rss; \
288} while (0)
289#define update_hiwater_vm(mm) do { \
290 if ((mm)->hiwater_vm < (mm)->total_vm) \
291 (mm)->hiwater_vm = (mm)->total_vm; \
292} while (0)
293
1da177e4
LT
294struct mm_struct {
295 struct vm_area_struct * mmap; /* list of VMAs */
296 struct rb_root mm_rb;
297 struct vm_area_struct * mmap_cache; /* last find_vma result */
298 unsigned long (*get_unmapped_area) (struct file *filp,
299 unsigned long addr, unsigned long len,
300 unsigned long pgoff, unsigned long flags);
1363c3cd 301 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
0551fbd2
BH
302 unsigned long mmap_base; /* base of mmap area */
303 unsigned long task_size; /* size of task vm space */
304 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
1363c3cd 305 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
1da177e4
LT
306 pgd_t * pgd;
307 atomic_t mm_users; /* How many users with user space? */
308 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
309 int map_count; /* number of VMAs */
310 struct rw_semaphore mmap_sem;
311 spinlock_t page_table_lock; /* Protects page tables and some counters */
312
313 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
314 * together off init_mm.mmlist, and are protected
315 * by mmlist_lock
316 */
317
f412ac08
HD
318 /* Special counters, in some configurations protected by the
319 * page_table_lock, in other configurations by being atomic.
320 */
4294621f 321 mm_counter_t _file_rss;
1da177e4
LT
322 mm_counter_t _anon_rss;
323
f449952b
HD
324 unsigned long hiwater_rss; /* High-watermark of RSS usage */
325 unsigned long hiwater_vm; /* High-water virtual memory usage */
326
327 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
328 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
329 unsigned long start_code, end_code, start_data, end_data;
330 unsigned long start_brk, brk, start_stack;
331 unsigned long arg_start, arg_end, env_start, env_end;
332
36d57ac4 333 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
1da177e4 334
d6e71144 335 unsigned dumpable:2;
1da177e4
LT
336 cpumask_t cpu_vm_mask;
337
338 /* Architecture-specific MM context */
339 mm_context_t context;
340
341 /* Token based thrashing protection. */
342 unsigned long swap_token_time;
343 char recent_pagein;
344
345 /* coredumping support */
346 int core_waiters;
347 struct completion *core_startup_done, core_done;
348
349 /* aio bits */
350 rwlock_t ioctx_list_lock;
351 struct kioctx *ioctx_list;
1da177e4
LT
352};
353
354struct sighand_struct {
355 atomic_t count;
356 struct k_sigaction action[_NSIG];
357 spinlock_t siglock;
e56d0903 358 struct rcu_head rcu;
1da177e4
LT
359};
360
e56d0903
IM
361extern void sighand_free_cb(struct rcu_head *rhp);
362
363static inline void sighand_free(struct sighand_struct *sp)
364{
365 call_rcu(&sp->rcu, sighand_free_cb);
366}
367
1da177e4
LT
368/*
369 * NOTE! "signal_struct" does not have it's own
370 * locking, because a shared signal_struct always
371 * implies a shared sighand_struct, so locking
372 * sighand_struct is always a proper superset of
373 * the locking of signal_struct.
374 */
375struct signal_struct {
376 atomic_t count;
377 atomic_t live;
378
379 wait_queue_head_t wait_chldexit; /* for wait4() */
380
381 /* current thread group signal load-balancing target: */
382 task_t *curr_target;
383
384 /* shared signal handling: */
385 struct sigpending shared_pending;
386
387 /* thread group exit support */
388 int group_exit_code;
389 /* overloaded:
390 * - notify group_exit_task when ->count is equal to notify_count
391 * - everyone except group_exit_task is stopped during signal delivery
392 * of fatal signals, group_exit_task processes the signal.
393 */
394 struct task_struct *group_exit_task;
395 int notify_count;
396
397 /* thread group stop support, overloads group_exit_code too */
398 int group_stop_count;
399 unsigned int flags; /* see SIGNAL_* flags below */
400
401 /* POSIX.1b Interval Timers */
402 struct list_head posix_timers;
403
404 /* ITIMER_REAL timer for the process */
2ff678b8 405 struct hrtimer real_timer;
05cfb614 406 struct task_struct *tsk;
2ff678b8 407 ktime_t it_real_incr;
1da177e4
LT
408
409 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
410 cputime_t it_prof_expires, it_virt_expires;
411 cputime_t it_prof_incr, it_virt_incr;
412
413 /* job control IDs */
414 pid_t pgrp;
415 pid_t tty_old_pgrp;
416 pid_t session;
417 /* boolean value for session group leader */
418 int leader;
419
420 struct tty_struct *tty; /* NULL if no tty */
421
422 /*
423 * Cumulative resource counters for dead threads in the group,
424 * and for reaped dead child processes forked by this group.
425 * Live threads maintain their own counters and add to these
426 * in __exit_signal, except for the group leader.
427 */
428 cputime_t utime, stime, cutime, cstime;
429 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
430 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
431
432 /*
433 * Cumulative ns of scheduled CPU time for dead threads in the
434 * group, not including a zombie group leader. (This only differs
435 * from jiffies_to_ns(utime + stime) if sched_clock uses something
436 * other than jiffies.)
437 */
438 unsigned long long sched_time;
439
440 /*
441 * We don't bother to synchronize most readers of this at all,
442 * because there is no reader checking a limit that actually needs
443 * to get both rlim_cur and rlim_max atomically, and either one
444 * alone is a single word that can safely be read normally.
445 * getrlimit/setrlimit use task_lock(current->group_leader) to
446 * protect this instead of the siglock, because they really
447 * have no need to disable irqs.
448 */
449 struct rlimit rlim[RLIM_NLIMITS];
450
451 struct list_head cpu_timers[3];
452
453 /* keep the process-shared keyrings here so that they do the right
454 * thing in threads created with CLONE_THREAD */
455#ifdef CONFIG_KEYS
456 struct key *session_keyring; /* keyring inherited over fork */
457 struct key *process_keyring; /* keyring private to this process */
458#endif
459};
460
4866cde0
NP
461/* Context switch must be unlocked if interrupts are to be enabled */
462#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
463# define __ARCH_WANT_UNLOCKED_CTXSW
464#endif
465
1da177e4
LT
466/*
467 * Bits in flags field of signal_struct.
468 */
469#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
470#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
471#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
472#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
473
474
475/*
476 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
b0a9499c
IM
477 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
478 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
479 * values are inverted: lower p->prio value means higher priority.
1da177e4
LT
480 *
481 * The MAX_USER_RT_PRIO value allows the actual maximum
482 * RT priority to be separate from the value exported to
483 * user-space. This allows kernel threads to set their
484 * priority to a value higher than any user task. Note:
485 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
486 */
487
488#define MAX_USER_RT_PRIO 100
489#define MAX_RT_PRIO MAX_USER_RT_PRIO
490
491#define MAX_PRIO (MAX_RT_PRIO + 40)
492
493#define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
494
495/*
496 * Some day this will be a full-fledged user tracking system..
497 */
498struct user_struct {
499 atomic_t __count; /* reference count */
500 atomic_t processes; /* How many processes does this user have? */
501 atomic_t files; /* How many open files does this user have? */
502 atomic_t sigpending; /* How many pending signals does this user have? */
0eeca283
RL
503#ifdef CONFIG_INOTIFY
504 atomic_t inotify_watches; /* How many inotify watches does this user have? */
505 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
506#endif
1da177e4
LT
507 /* protected by mq_lock */
508 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
509 unsigned long locked_shm; /* How many pages of mlocked shm ? */
510
511#ifdef CONFIG_KEYS
512 struct key *uid_keyring; /* UID specific keyring */
513 struct key *session_keyring; /* UID's default session keyring */
514#endif
515
516 /* Hash table maintenance information */
517 struct list_head uidhash_list;
518 uid_t uid;
519};
520
521extern struct user_struct *find_user(uid_t);
522
523extern struct user_struct root_user;
524#define INIT_USER (&root_user)
525
526typedef struct prio_array prio_array_t;
527struct backing_dev_info;
528struct reclaim_state;
529
530#ifdef CONFIG_SCHEDSTATS
531struct sched_info {
532 /* cumulative counters */
533 unsigned long cpu_time, /* time spent on the cpu */
534 run_delay, /* time spent waiting on a runqueue */
535 pcnt; /* # of timeslices run on this cpu */
536
537 /* timestamps */
538 unsigned long last_arrival, /* when we last ran on a cpu */
539 last_queued; /* when we were last queued to run */
540};
541
542extern struct file_operations proc_schedstat_operations;
543#endif
544
545enum idle_type
546{
547 SCHED_IDLE,
548 NOT_IDLE,
549 NEWLY_IDLE,
550 MAX_IDLE_TYPES
551};
552
553/*
554 * sched-domains (multiprocessor balancing) declarations:
555 */
556#ifdef CONFIG_SMP
557#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
558
559#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
560#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
561#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
562#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
563#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
564#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
565#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
566#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
1da177e4
LT
567
568struct sched_group {
569 struct sched_group *next; /* Must be a circular list */
570 cpumask_t cpumask;
571
572 /*
573 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
574 * single CPU. This is read only (except for setup, hotplug CPU).
575 */
576 unsigned long cpu_power;
577};
578
579struct sched_domain {
580 /* These fields must be setup */
581 struct sched_domain *parent; /* top domain must be null terminated */
582 struct sched_group *groups; /* the balancing groups of the domain */
583 cpumask_t span; /* span of all CPUs in this domain */
584 unsigned long min_interval; /* Minimum balance interval ms */
585 unsigned long max_interval; /* Maximum balance interval ms */
586 unsigned int busy_factor; /* less balancing by factor if busy */
587 unsigned int imbalance_pct; /* No balance until over watermark */
588 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
589 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
590 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
7897986b
NP
591 unsigned int busy_idx;
592 unsigned int idle_idx;
593 unsigned int newidle_idx;
594 unsigned int wake_idx;
147cbb4b 595 unsigned int forkexec_idx;
1da177e4
LT
596 int flags; /* See SD_* */
597
598 /* Runtime fields. */
599 unsigned long last_balance; /* init to jiffies. units in jiffies */
600 unsigned int balance_interval; /* initialise to 1. units in ms. */
601 unsigned int nr_balance_failed; /* initialise to 0 */
602
603#ifdef CONFIG_SCHEDSTATS
604 /* load_balance() stats */
605 unsigned long lb_cnt[MAX_IDLE_TYPES];
606 unsigned long lb_failed[MAX_IDLE_TYPES];
607 unsigned long lb_balanced[MAX_IDLE_TYPES];
608 unsigned long lb_imbalance[MAX_IDLE_TYPES];
609 unsigned long lb_gained[MAX_IDLE_TYPES];
610 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
611 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
612 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
613
614 /* Active load balancing */
615 unsigned long alb_cnt;
616 unsigned long alb_failed;
617 unsigned long alb_pushed;
618
68767a0a
NP
619 /* SD_BALANCE_EXEC stats */
620 unsigned long sbe_cnt;
621 unsigned long sbe_balanced;
1da177e4
LT
622 unsigned long sbe_pushed;
623
68767a0a
NP
624 /* SD_BALANCE_FORK stats */
625 unsigned long sbf_cnt;
626 unsigned long sbf_balanced;
627 unsigned long sbf_pushed;
628
1da177e4
LT
629 /* try_to_wake_up() stats */
630 unsigned long ttwu_wake_remote;
631 unsigned long ttwu_move_affine;
632 unsigned long ttwu_move_balance;
633#endif
634};
635
1a20ff27
DG
636extern void partition_sched_domains(cpumask_t *partition1,
637 cpumask_t *partition2);
198e2f18
AM
638
639/*
640 * Maximum cache size the migration-costs auto-tuning code will
641 * search from:
642 */
643extern unsigned int max_cache_size;
644
645#endif /* CONFIG_SMP */
1da177e4
LT
646
647
648struct io_context; /* See blkdev.h */
649void exit_io_context(void);
650struct cpuset;
651
652#define NGROUPS_SMALL 32
653#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
654struct group_info {
655 int ngroups;
656 atomic_t usage;
657 gid_t small_block[NGROUPS_SMALL];
658 int nblocks;
659 gid_t *blocks[0];
660};
661
662/*
663 * get_group_info() must be called with the owning task locked (via task_lock())
664 * when task != current. The reason being that the vast majority of callers are
665 * looking at current->group_info, which can not be changed except by the
666 * current task. Changing current->group_info requires the task lock, too.
667 */
668#define get_group_info(group_info) do { \
669 atomic_inc(&(group_info)->usage); \
670} while (0)
671
672#define put_group_info(group_info) do { \
673 if (atomic_dec_and_test(&(group_info)->usage)) \
674 groups_free(group_info); \
675} while (0)
676
3e30148c
DH
677extern struct group_info *groups_alloc(int gidsetsize);
678extern void groups_free(struct group_info *group_info);
679extern int set_current_groups(struct group_info *group_info);
680extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
681/* access the groups "array" with this macro */
682#define GROUP_AT(gi, i) \
683 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
684
383f2835
KC
685#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
686extern void prefetch_stack(struct task_struct*);
687#else
688static inline void prefetch_stack(struct task_struct *t) { }
689#endif
1da177e4
LT
690
691struct audit_context; /* See audit.c */
692struct mempolicy;
693
694struct task_struct {
695 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
696 struct thread_info *thread_info;
697 atomic_t usage;
698 unsigned long flags; /* per process flags, defined below */
699 unsigned long ptrace;
700
36772092 701 int lock_depth; /* BKL lock depth */
1da177e4 702
d6077cb8 703#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
4866cde0
NP
704 int oncpu;
705#endif
1da177e4
LT
706 int prio, static_prio;
707 struct list_head run_list;
708 prio_array_t *array;
709
22e2c507 710 unsigned short ioprio;
2056a782 711 unsigned int btrace_seq;
22e2c507 712
1da177e4
LT
713 unsigned long sleep_avg;
714 unsigned long long timestamp, last_ran;
715 unsigned long long sched_time; /* sched_clock time spent running */
716 int activated;
717
718 unsigned long policy;
719 cpumask_t cpus_allowed;
720 unsigned int time_slice, first_time_slice;
721
722#ifdef CONFIG_SCHEDSTATS
723 struct sched_info sched_info;
724#endif
725
726 struct list_head tasks;
727 /*
728 * ptrace_list/ptrace_children forms the list of my children
729 * that were stolen by a ptracer.
730 */
731 struct list_head ptrace_children;
732 struct list_head ptrace_list;
733
734 struct mm_struct *mm, *active_mm;
735
736/* task state */
737 struct linux_binfmt *binfmt;
738 long exit_state;
739 int exit_code, exit_signal;
740 int pdeath_signal; /* The signal sent when the parent dies */
741 /* ??? */
742 unsigned long personality;
743 unsigned did_exec:1;
744 pid_t pid;
745 pid_t tgid;
746 /*
747 * pointers to (original) parent process, youngest child, younger sibling,
748 * older sibling, respectively. (p->father can be replaced with
749 * p->parent->pid)
750 */
751 struct task_struct *real_parent; /* real parent process (when being debugged) */
752 struct task_struct *parent; /* parent process */
753 /*
754 * children/sibling forms the list of my children plus the
755 * tasks I'm ptracing.
756 */
757 struct list_head children; /* list of my children */
758 struct list_head sibling; /* linkage in my parent's children list */
759 struct task_struct *group_leader; /* threadgroup leader */
760
761 /* PID/PID hash table linkage. */
762 struct pid pids[PIDTYPE_MAX];
763
764 struct completion *vfork_done; /* for vfork() */
765 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
766 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
767
768 unsigned long rt_priority;
769 cputime_t utime, stime;
770 unsigned long nvcsw, nivcsw; /* context switch counts */
771 struct timespec start_time;
772/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
773 unsigned long min_flt, maj_flt;
774
775 cputime_t it_prof_expires, it_virt_expires;
776 unsigned long long it_sched_expires;
777 struct list_head cpu_timers[3];
778
779/* process credentials */
780 uid_t uid,euid,suid,fsuid;
781 gid_t gid,egid,sgid,fsgid;
782 struct group_info *group_info;
783 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
784 unsigned keep_capabilities:1;
785 struct user_struct *user;
786#ifdef CONFIG_KEYS
b5f545c8 787 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 788 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 789 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4
LT
790#endif
791 int oomkilladj; /* OOM kill score adjustment (bit shift). */
36772092
PBG
792 char comm[TASK_COMM_LEN]; /* executable name excluding path
793 - access with [gs]et_task_comm (which lock
794 it with task_lock())
795 - initialized normally by flush_old_exec */
1da177e4
LT
796/* file system info */
797 int link_count, total_link_count;
798/* ipc stuff */
799 struct sysv_sem sysvsem;
800/* CPU-specific state of this task */
801 struct thread_struct thread;
802/* filesystem information */
803 struct fs_struct *fs;
804/* open file information */
805 struct files_struct *files;
806/* namespace */
807 struct namespace *namespace;
808/* signal handlers */
809 struct signal_struct *signal;
810 struct sighand_struct *sighand;
811
812 sigset_t blocked, real_blocked;
150256d8 813 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
1da177e4
LT
814 struct sigpending pending;
815
816 unsigned long sas_ss_sp;
817 size_t sas_ss_size;
818 int (*notifier)(void *priv);
819 void *notifier_data;
820 sigset_t *notifier_mask;
821
822 void *security;
823 struct audit_context *audit_context;
824 seccomp_t seccomp;
825
826/* Thread group tracking */
827 u32 parent_exec_id;
828 u32 self_exec_id;
829/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
830 spinlock_t alloc_lock;
831/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
832 spinlock_t proc_lock;
1da177e4 833
408894ee
IM
834#ifdef CONFIG_DEBUG_MUTEXES
835 /* mutex deadlock detection */
836 struct mutex_waiter *blocked_on;
837#endif
838
1da177e4
LT
839/* journalling filesystem info */
840 void *journal_info;
841
842/* VM state */
843 struct reclaim_state *reclaim_state;
844
845 struct dentry *proc_dentry;
846 struct backing_dev_info *backing_dev_info;
847
848 struct io_context *io_context;
849
850 unsigned long ptrace_message;
851 siginfo_t *last_siginfo; /* For ptrace use. */
852/*
853 * current io wait handle: wait queue entry to use for io waits
854 * If this thread is processing aio, this points at the waitqueue
855 * inside the currently handled kiocb. It may be NULL (i.e. default
856 * to a stack based synchronous wait) if its doing sync IO.
857 */
858 wait_queue_t *io_wait;
859/* i/o counters(bytes read/written, #syscalls */
860 u64 rchar, wchar, syscr, syscw;
861#if defined(CONFIG_BSD_PROCESS_ACCT)
862 u64 acct_rss_mem1; /* accumulated rss usage */
863 u64 acct_vm_mem1; /* accumulated virtual memory usage */
864 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
865#endif
866#ifdef CONFIG_NUMA
867 struct mempolicy *mempolicy;
868 short il_next;
869#endif
870#ifdef CONFIG_CPUSETS
871 struct cpuset *cpuset;
872 nodemask_t mems_allowed;
873 int cpuset_mems_generation;
825a46af 874 int cpuset_mem_spread_rotor;
1da177e4 875#endif
0771dfef
IM
876 struct robust_list_head __user *robust_list;
877
22e2c507 878 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 879 struct rcu_head rcu;
1da177e4
LT
880};
881
882static inline pid_t process_group(struct task_struct *tsk)
883{
884 return tsk->signal->pgrp;
885}
886
887/**
888 * pid_alive - check that a task structure is not stale
889 * @p: Task structure to be checked.
890 *
891 * Test if a process is not yet dead (at most zombie state)
892 * If pid_alive fails, then pointers within the task structure
893 * can be stale and must not be dereferenced.
894 */
895static inline int pid_alive(struct task_struct *p)
896{
897 return p->pids[PIDTYPE_PID].nr != 0;
898}
899
900extern void free_task(struct task_struct *tsk);
1da177e4 901#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 902
e56d0903
IM
903extern void __put_task_struct_cb(struct rcu_head *rhp);
904
905static inline void put_task_struct(struct task_struct *t)
906{
907 if (atomic_dec_and_test(&t->usage))
908 call_rcu(&t->rcu, __put_task_struct_cb);
909}
1da177e4
LT
910
911/*
912 * Per process flags
913 */
914#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
915 /* Not implemented yet, only for 486*/
916#define PF_STARTING 0x00000002 /* being created */
917#define PF_EXITING 0x00000004 /* getting shut down */
918#define PF_DEAD 0x00000008 /* Dead */
919#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
920#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
921#define PF_DUMPCORE 0x00000200 /* dumped core */
922#define PF_SIGNALED 0x00000400 /* killed by a signal */
923#define PF_MEMALLOC 0x00000800 /* Allocating memory */
924#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
925#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
926#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
927#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
928#define PF_FROZEN 0x00010000 /* frozen for system suspend */
929#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
930#define PF_KSWAPD 0x00040000 /* I am kswapd */
931#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
932#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
933#define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
934#define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
935#define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
930d9152 936#define PF_SWAPWRITE 0x01000000 /* Allowed to write to swap */
825a46af
PJ
937#define PF_SPREAD_PAGE 0x04000000 /* Spread page cache over cpuset */
938#define PF_SPREAD_SLAB 0x08000000 /* Spread some slab caches over cpuset */
c61afb18 939#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1da177e4
LT
940
941/*
942 * Only the _current_ task can read/write to tsk->flags, but other
943 * tasks can access tsk->flags in readonly mode for example
944 * with tsk_used_math (like during threaded core dumping).
945 * There is however an exception to this rule during ptrace
946 * or during fork: the ptracer task is allowed to write to the
947 * child->flags of its traced child (same goes for fork, the parent
948 * can write to the child->flags), because we're guaranteed the
949 * child is not running and in turn not changing child->flags
950 * at the same time the parent does it.
951 */
952#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
953#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
954#define clear_used_math() clear_stopped_child_used_math(current)
955#define set_used_math() set_stopped_child_used_math(current)
956#define conditional_stopped_child_used_math(condition, child) \
957 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
958#define conditional_used_math(condition) \
959 conditional_stopped_child_used_math(condition, current)
960#define copy_to_stopped_child_used_math(child) \
961 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
962/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
963#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
964#define used_math() tsk_used_math(current)
965
966#ifdef CONFIG_SMP
967extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
968#else
969static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
970{
4098f991 971 if (!cpu_isset(0, new_mask))
1da177e4
LT
972 return -EINVAL;
973 return 0;
974}
975#endif
976
977extern unsigned long long sched_clock(void);
978extern unsigned long long current_sched_time(const task_t *current_task);
979
980/* sched_exec is called by processes performing an exec */
981#ifdef CONFIG_SMP
982extern void sched_exec(void);
983#else
984#define sched_exec() {}
985#endif
986
987#ifdef CONFIG_HOTPLUG_CPU
988extern void idle_task_exit(void);
989#else
990static inline void idle_task_exit(void) {}
991#endif
992
993extern void sched_idle_next(void);
994extern void set_user_nice(task_t *p, long nice);
995extern int task_prio(const task_t *p);
996extern int task_nice(const task_t *p);
e43379f1 997extern int can_nice(const task_t *p, const int nice);
1da177e4
LT
998extern int task_curr(const task_t *p);
999extern int idle_cpu(int cpu);
1000extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1001extern task_t *idle_task(int cpu);
a2a97982
KO
1002extern task_t *curr_task(int cpu);
1003extern void set_curr_task(int cpu, task_t *p);
1da177e4
LT
1004
1005void yield(void);
1006
1007/*
1008 * The default (Linux) execution domain.
1009 */
1010extern struct exec_domain default_exec_domain;
1011
1012union thread_union {
1013 struct thread_info thread_info;
1014 unsigned long stack[THREAD_SIZE/sizeof(long)];
1015};
1016
1017#ifndef __HAVE_ARCH_KSTACK_END
1018static inline int kstack_end(void *addr)
1019{
1020 /* Reliable end of stack detection:
1021 * Some APM bios versions misalign the stack
1022 */
1023 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1024}
1025#endif
1026
1027extern union thread_union init_thread_union;
1028extern struct task_struct init_task;
1029
1030extern struct mm_struct init_mm;
1031
1032#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1033extern struct task_struct *find_task_by_pid_type(int type, int pid);
1034extern void set_special_pids(pid_t session, pid_t pgrp);
1035extern void __set_special_pids(pid_t session, pid_t pgrp);
1036
1037/* per-UID process charging. */
1038extern struct user_struct * alloc_uid(uid_t);
1039static inline struct user_struct *get_uid(struct user_struct *u)
1040{
1041 atomic_inc(&u->__count);
1042 return u;
1043}
1044extern void free_uid(struct user_struct *);
1045extern void switch_uid(struct user_struct *);
1046
1047#include <asm/current.h>
1048
1049extern void do_timer(struct pt_regs *);
1050
1051extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1052extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1053extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1054 unsigned long clone_flags));
1055#ifdef CONFIG_SMP
1056 extern void kick_process(struct task_struct *tsk);
1057#else
1058 static inline void kick_process(struct task_struct *tsk) { }
1059#endif
476d139c 1060extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
1da177e4
LT
1061extern void FASTCALL(sched_exit(task_t * p));
1062
1063extern int in_group_p(gid_t);
1064extern int in_egroup_p(gid_t);
1065
1066extern void proc_caches_init(void);
1067extern void flush_signals(struct task_struct *);
1068extern void flush_signal_handlers(struct task_struct *, int force_default);
1069extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1070
1071static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1072{
1073 unsigned long flags;
1074 int ret;
1075
1076 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1077 ret = dequeue_signal(tsk, mask, info);
1078 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1079
1080 return ret;
1081}
1082
1083extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1084 sigset_t *mask);
1085extern void unblock_all_signals(void);
1086extern void release_task(struct task_struct * p);
1087extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1088extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1089extern int force_sigsegv(int, struct task_struct *);
1090extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1091extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1092extern int kill_pg_info(int, struct siginfo *, pid_t);
1093extern int kill_proc_info(int, struct siginfo *, pid_t);
46113830 1094extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t);
1da177e4
LT
1095extern void do_notify_parent(struct task_struct *, int);
1096extern void force_sig(int, struct task_struct *);
1097extern void force_sig_specific(int, struct task_struct *);
1098extern int send_sig(int, struct task_struct *, int);
1099extern void zap_other_threads(struct task_struct *p);
1100extern int kill_pg(pid_t, int, int);
1101extern int kill_sl(pid_t, int, int);
1102extern int kill_proc(pid_t, int, int);
1103extern struct sigqueue *sigqueue_alloc(void);
1104extern void sigqueue_free(struct sigqueue *);
1105extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1106extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
9ac95f2f 1107extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
1108extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1109
1110/* These can be the second arg to send_sig_info/send_group_sig_info. */
1111#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1112#define SEND_SIG_PRIV ((struct siginfo *) 1)
1113#define SEND_SIG_FORCED ((struct siginfo *) 2)
1114
621d3121
ON
1115static inline int is_si_special(const struct siginfo *info)
1116{
1117 return info <= SEND_SIG_FORCED;
1118}
1119
1da177e4
LT
1120/* True if we are on the alternate signal stack. */
1121
1122static inline int on_sig_stack(unsigned long sp)
1123{
1124 return (sp - current->sas_ss_sp < current->sas_ss_size);
1125}
1126
1127static inline int sas_ss_flags(unsigned long sp)
1128{
1129 return (current->sas_ss_size == 0 ? SS_DISABLE
1130 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1131}
1132
1da177e4
LT
1133/*
1134 * Routines for handling mm_structs
1135 */
1136extern struct mm_struct * mm_alloc(void);
1137
1138/* mmdrop drops the mm and the page tables */
1139extern void FASTCALL(__mmdrop(struct mm_struct *));
1140static inline void mmdrop(struct mm_struct * mm)
1141{
1142 if (atomic_dec_and_test(&mm->mm_count))
1143 __mmdrop(mm);
1144}
1145
1146/* mmput gets rid of the mappings and all user-space */
1147extern void mmput(struct mm_struct *);
1148/* Grab a reference to a task's mm, if it is not already going away */
1149extern struct mm_struct *get_task_mm(struct task_struct *task);
1150/* Remove the current tasks stale references to the old mm_struct */
1151extern void mm_release(struct task_struct *, struct mm_struct *);
1152
1153extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1154extern void flush_thread(void);
1155extern void exit_thread(void);
1156
1da177e4
LT
1157extern void exit_files(struct task_struct *);
1158extern void exit_signal(struct task_struct *);
1159extern void __exit_signal(struct task_struct *);
1160extern void exit_sighand(struct task_struct *);
1161extern void __exit_sighand(struct task_struct *);
1162extern void exit_itimers(struct signal_struct *);
1163
1164extern NORET_TYPE void do_group_exit(int);
1165
1da177e4
LT
1166extern void daemonize(const char *, ...);
1167extern int allow_signal(int);
1168extern int disallow_signal(int);
1169extern task_t *child_reaper;
1170
1171extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1172extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1173task_t *fork_idle(int);
1174
1175extern void set_task_comm(struct task_struct *tsk, char *from);
1176extern void get_task_comm(char *to, struct task_struct *tsk);
1177
1178#ifdef CONFIG_SMP
1179extern void wait_task_inactive(task_t * p);
1180#else
1181#define wait_task_inactive(p) do { } while (0)
1182#endif
1183
1184#define remove_parent(p) list_del_init(&(p)->sibling)
1185#define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1186
1187#define REMOVE_LINKS(p) do { \
1188 if (thread_group_leader(p)) \
1189 list_del_init(&(p)->tasks); \
1190 remove_parent(p); \
1191 } while (0)
1192
1193#define SET_LINKS(p) do { \
1194 if (thread_group_leader(p)) \
1195 list_add_tail(&(p)->tasks,&init_task.tasks); \
1196 add_parent(p, (p)->parent); \
1197 } while (0)
1198
1199#define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1200#define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1201
1202#define for_each_process(p) \
1203 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1204
1205/*
1206 * Careful: do_each_thread/while_each_thread is a double loop so
1207 * 'break' will not work as expected - use goto instead.
1208 */
1209#define do_each_thread(g, t) \
1210 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1211
1212#define while_each_thread(g, t) \
1213 while ((t = next_thread(t)) != g)
1214
1215extern task_t * FASTCALL(next_thread(const task_t *p));
1216
1217#define thread_group_leader(p) (p->pid == p->tgid)
1218
1219static inline int thread_group_empty(task_t *p)
1220{
1221 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1222}
1223
1224#define delay_group_leader(p) \
1225 (thread_group_leader(p) && !thread_group_empty(p))
1226
1227extern void unhash_process(struct task_struct *p);
1228
1229/*
1230 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
22e2c507 1231 * subscriptions and synchronises with wait4(). Also used in procfs. Also
053199ed 1232 * pins the final release of task.io_context. Also protects ->cpuset.
1da177e4
LT
1233 *
1234 * Nests both inside and outside of read_lock(&tasklist_lock).
1235 * It must not be nested with write_lock_irq(&tasklist_lock),
1236 * neither inside nor outside.
1237 */
1238static inline void task_lock(struct task_struct *p)
1239{
1240 spin_lock(&p->alloc_lock);
1241}
1242
1243static inline void task_unlock(struct task_struct *p)
1244{
1245 spin_unlock(&p->alloc_lock);
1246}
1247
f037360f
AV
1248#ifndef __HAVE_THREAD_FUNCTIONS
1249
a1261f54 1250#define task_thread_info(task) (task)->thread_info
9fc65876 1251#define task_stack_page(task) ((void*)((task)->thread_info))
a1261f54 1252
10ebffde
AV
1253static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1254{
1255 *task_thread_info(p) = *task_thread_info(org);
1256 task_thread_info(p)->task = p;
1257}
1258
1259static inline unsigned long *end_of_stack(struct task_struct *p)
1260{
1261 return (unsigned long *)(p->thread_info + 1);
1262}
1263
f037360f
AV
1264#endif
1265
1da177e4
LT
1266/* set thread flags in other task's structures
1267 * - see asm/thread_info.h for TIF_xxxx flags available
1268 */
1269static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1270{
a1261f54 1271 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1272}
1273
1274static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1275{
a1261f54 1276 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1277}
1278
1279static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1280{
a1261f54 1281 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1282}
1283
1284static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1285{
a1261f54 1286 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1287}
1288
1289static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1290{
a1261f54 1291 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1292}
1293
1294static inline void set_tsk_need_resched(struct task_struct *tsk)
1295{
1296 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1297}
1298
1299static inline void clear_tsk_need_resched(struct task_struct *tsk)
1300{
1301 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1302}
1303
1304static inline int signal_pending(struct task_struct *p)
1305{
1306 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1307}
1308
1309static inline int need_resched(void)
1310{
1311 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1312}
1313
1314/*
1315 * cond_resched() and cond_resched_lock(): latency reduction via
1316 * explicit rescheduling in places that are safe. The return
1317 * value indicates whether a reschedule was done in fact.
1318 * cond_resched_lock() will drop the spinlock before scheduling,
1319 * cond_resched_softirq() will enable bhs before scheduling.
1320 */
1321extern int cond_resched(void);
1322extern int cond_resched_lock(spinlock_t * lock);
1323extern int cond_resched_softirq(void);
1324
1325/*
1326 * Does a critical section need to be broken due to another
1327 * task waiting?:
1328 */
1329#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1330# define need_lockbreak(lock) ((lock)->break_lock)
1331#else
1332# define need_lockbreak(lock) 0
1333#endif
1334
1335/*
1336 * Does a critical section need to be broken due to another
1337 * task waiting or preemption being signalled:
1338 */
1339static inline int lock_need_resched(spinlock_t *lock)
1340{
1341 if (need_lockbreak(lock) || need_resched())
1342 return 1;
1343 return 0;
1344}
1345
1346/* Reevaluate whether the task has signals pending delivery.
1347 This is required every time the blocked sigset_t changes.
1348 callers must hold sighand->siglock. */
1349
1350extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1351extern void recalc_sigpending(void);
1352
1353extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1354
1355/*
1356 * Wrappers for p->thread_info->cpu access. No-op on UP.
1357 */
1358#ifdef CONFIG_SMP
1359
1360static inline unsigned int task_cpu(const struct task_struct *p)
1361{
a1261f54 1362 return task_thread_info(p)->cpu;
1da177e4
LT
1363}
1364
1365static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1366{
a1261f54 1367 task_thread_info(p)->cpu = cpu;
1da177e4
LT
1368}
1369
1370#else
1371
1372static inline unsigned int task_cpu(const struct task_struct *p)
1373{
1374 return 0;
1375}
1376
1377static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1378{
1379}
1380
1381#endif /* CONFIG_SMP */
1382
1383#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1384extern void arch_pick_mmap_layout(struct mm_struct *mm);
1385#else
1386static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1387{
1388 mm->mmap_base = TASK_UNMAPPED_BASE;
1389 mm->get_unmapped_area = arch_get_unmapped_area;
1390 mm->unmap_area = arch_unmap_area;
1391}
1392#endif
1393
1394extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1395extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1396
1da177e4
LT
1397extern void normalize_rt_tasks(void);
1398
1da177e4 1399#ifdef CONFIG_PM
3e1d1d28
CL
1400/*
1401 * Check if a process has been frozen
1402 */
1403static inline int frozen(struct task_struct *p)
1404{
1405 return p->flags & PF_FROZEN;
1406}
1407
1408/*
1409 * Check if there is a request to freeze a process
1410 */
1411static inline int freezing(struct task_struct *p)
1412{
1413 return p->flags & PF_FREEZE;
1414}
1415
1416/*
1417 * Request that a process be frozen
1418 * FIXME: SMP problem. We may not modify other process' flags!
1419 */
1420static inline void freeze(struct task_struct *p)
1421{
1422 p->flags |= PF_FREEZE;
1423}
1424
1425/*
1426 * Wake up a frozen process
1427 */
1428static inline int thaw_process(struct task_struct *p)
1429{
1430 if (frozen(p)) {
1431 p->flags &= ~PF_FROZEN;
1432 wake_up_process(p);
1433 return 1;
1434 }
1435 return 0;
1436}
1437
1438/*
1439 * freezing is complete, mark process as frozen
1440 */
1441static inline void frozen_process(struct task_struct *p)
1442{
1443 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1444}
1445
1446extern void refrigerator(void);
1da177e4
LT
1447extern int freeze_processes(void);
1448extern void thaw_processes(void);
1449
3e1d1d28 1450static inline int try_to_freeze(void)
1da177e4 1451{
3e1d1d28
CL
1452 if (freezing(current)) {
1453 refrigerator();
1da177e4
LT
1454 return 1;
1455 } else
1456 return 0;
1457}
1458#else
3e1d1d28
CL
1459static inline int frozen(struct task_struct *p) { return 0; }
1460static inline int freezing(struct task_struct *p) { return 0; }
1461static inline void freeze(struct task_struct *p) { BUG(); }
1462static inline int thaw_process(struct task_struct *p) { return 1; }
1463static inline void frozen_process(struct task_struct *p) { BUG(); }
1464
1465static inline void refrigerator(void) {}
1da177e4
LT
1466static inline int freeze_processes(void) { BUG(); return 0; }
1467static inline void thaw_processes(void) {}
1468
3e1d1d28
CL
1469static inline int try_to_freeze(void) { return 0; }
1470
1da177e4
LT
1471#endif /* CONFIG_PM */
1472#endif /* __KERNEL__ */
1473
1474#endif