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