Sanitise linux/audit.h for userspace consumption, split elf-em.h from elf.h
[linux-2.6-block.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
a3b6714e 4#ifdef __KERNEL__
1da177e4
LT
5#include <asm/param.h> /* for HZ */
6
1da177e4
LT
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
a3b6714e
DW
40#include <linux/time.h>
41#include <linux/param.h>
42#include <linux/resource.h>
43#include <linux/timer.h>
44#include <linux/hrtimer.h>
45
46#include <asm/processor.h>
47#endif
48
36d57ac4
L
49#include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
50
1da177e4
LT
51struct exec_domain;
52
53/*
54 * cloning flags:
55 */
56#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
57#define CLONE_VM 0x00000100 /* set if VM shared between processes */
58#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
59#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
60#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
61#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
62#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
63#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
64#define CLONE_THREAD 0x00010000 /* Same thread group? */
65#define CLONE_NEWNS 0x00020000 /* New namespace group? */
66#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
67#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
68#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
69#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
70#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
71#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
72#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
73#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
74
75/*
76 * List of flags we want to share for kernel threads,
77 * if only because they are not used by them anyway.
78 */
79#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
80
81/*
82 * These are the constant used to fake the fixed-point load-average
83 * counting. Some notes:
84 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
85 * a load-average precision of 10 bits integer + 11 bits fractional
86 * - if you want to count load-averages more often, you need more
87 * precision, or rounding will get you. With 2-second counting freq,
88 * the EXP_n values would be 1981, 2034 and 2043 if still using only
89 * 11 bit fractions.
90 */
91extern unsigned long avenrun[]; /* Load averages */
92
93#define FSHIFT 11 /* nr of bits of precision */
94#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
95#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
96#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
97#define EXP_5 2014 /* 1/exp(5sec/5min) */
98#define EXP_15 2037 /* 1/exp(5sec/15min) */
99
100#define CALC_LOAD(load,exp,n) \
101 load *= exp; \
102 load += n*(FIXED_1-exp); \
103 load >>= FSHIFT;
104
105extern unsigned long total_forks;
106extern int nr_threads;
107extern int last_pid;
108DECLARE_PER_CPU(unsigned long, process_counts);
109extern int nr_processes(void);
110extern unsigned long nr_running(void);
111extern unsigned long nr_uninterruptible(void);
db1b1fef 112extern unsigned long nr_active(void);
1da177e4
LT
113extern unsigned long nr_iowait(void);
114
1da177e4 115
4a8342d2
LT
116/*
117 * Task state bitmask. NOTE! These bits are also
118 * encoded in fs/proc/array.c: get_task_state().
119 *
120 * We have two separate sets of flags: task->state
121 * is about runnability, while task->exit_state are
122 * about the task exiting. Confusing, but this way
123 * modifying one set can't modify the other one by
124 * mistake.
125 */
1da177e4
LT
126#define TASK_RUNNING 0
127#define TASK_INTERRUPTIBLE 1
128#define TASK_UNINTERRUPTIBLE 2
4a8342d2
LT
129#define TASK_STOPPED 4
130#define TASK_TRACED 8
131/* in tsk->exit_state */
132#define EXIT_ZOMBIE 16
133#define EXIT_DEAD 32
134/* in tsk->state again */
135#define TASK_NONINTERACTIVE 64
1da177e4
LT
136
137#define __set_task_state(tsk, state_value) \
138 do { (tsk)->state = (state_value); } while (0)
139#define set_task_state(tsk, state_value) \
140 set_mb((tsk)->state, (state_value))
141
498d0c57
AM
142/*
143 * set_current_state() includes a barrier so that the write of current->state
144 * is correctly serialised wrt the caller's subsequent test of whether to
145 * actually sleep:
146 *
147 * set_current_state(TASK_UNINTERRUPTIBLE);
148 * if (do_i_need_to_sleep())
149 * schedule();
150 *
151 * If the caller does not need such serialisation then use __set_current_state()
152 */
1da177e4
LT
153#define __set_current_state(state_value) \
154 do { current->state = (state_value); } while (0)
155#define set_current_state(state_value) \
156 set_mb(current->state, (state_value))
157
158/* Task command name length */
159#define TASK_COMM_LEN 16
160
161/*
162 * Scheduling policies
163 */
164#define SCHED_NORMAL 0
165#define SCHED_FIFO 1
166#define SCHED_RR 2
b0a9499c 167#define SCHED_BATCH 3
1da177e4
LT
168
169struct sched_param {
170 int sched_priority;
171};
172
173#ifdef __KERNEL__
174
175#include <linux/spinlock.h>
176
177/*
178 * This serializes "schedule()" and also protects
179 * the run-queue from deletions/modifications (but
180 * _adding_ to the beginning of the run-queue has
181 * a separate lock).
182 */
183extern rwlock_t tasklist_lock;
184extern spinlock_t mmlist_lock;
185
186typedef struct task_struct task_t;
187
188extern void sched_init(void);
189extern void sched_init_smp(void);
190extern void init_idle(task_t *idle, int cpu);
191
192extern cpumask_t nohz_cpu_mask;
193
194extern void show_state(void);
195extern void show_regs(struct pt_regs *);
196
197/*
198 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
199 * task), SP is the stack pointer of the first frame that should be shown in the back
200 * trace (or NULL if the entire call-chain of the task should be shown).
201 */
202extern void show_stack(struct task_struct *task, unsigned long *sp);
203
204void io_schedule(void);
205long io_schedule_timeout(long timeout);
206
207extern void cpu_init (void);
208extern void trap_init(void);
209extern void update_process_times(int user);
210extern void scheduler_tick(void);
211
8446f1d3 212#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 213extern void softlockup_tick(void);
8446f1d3
IM
214extern void spawn_softlockup_task(void);
215extern void touch_softlockup_watchdog(void);
216#else
6687a97d 217static inline void softlockup_tick(void)
8446f1d3
IM
218{
219}
220static inline void spawn_softlockup_task(void)
221{
222}
223static inline void touch_softlockup_watchdog(void)
224{
225}
226#endif
227
228
1da177e4
LT
229/* Attach to any functions which should be ignored in wchan output. */
230#define __sched __attribute__((__section__(".sched.text")))
231/* Is this address in the __sched functions? */
232extern int in_sched_functions(unsigned long addr);
233
234#define MAX_SCHEDULE_TIMEOUT LONG_MAX
235extern signed long FASTCALL(schedule_timeout(signed long timeout));
64ed93a2
NA
236extern signed long schedule_timeout_interruptible(signed long timeout);
237extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
238asmlinkage void schedule(void);
239
240struct namespace;
241
242/* Maximum number of active map areas.. This is a random (large) number */
243#define DEFAULT_MAX_MAP_COUNT 65536
244
245extern int sysctl_max_map_count;
246
247#include <linux/aio.h>
248
249extern unsigned long
250arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
251 unsigned long, unsigned long);
252extern unsigned long
253arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
254 unsigned long len, unsigned long pgoff,
255 unsigned long flags);
1363c3cd
WW
256extern void arch_unmap_area(struct mm_struct *, unsigned long);
257extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 258
f412ac08
HD
259#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
260/*
261 * The mm counters are not protected by its page_table_lock,
262 * so must be incremented atomically.
263 */
d3cb4871
CL
264#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
265#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
266#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
267#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
268#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
269typedef atomic_long_t mm_counter_t;
f412ac08
HD
270
271#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
272/*
273 * The mm counters are protected by its page_table_lock,
274 * so can be incremented directly.
275 */
1da177e4
LT
276#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
277#define get_mm_counter(mm, member) ((mm)->_##member)
278#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
279#define inc_mm_counter(mm, member) (mm)->_##member++
280#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
281typedef unsigned long mm_counter_t;
282
283#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 284
f412ac08
HD
285#define get_mm_rss(mm) \
286 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
287#define update_hiwater_rss(mm) do { \
288 unsigned long _rss = get_mm_rss(mm); \
289 if ((mm)->hiwater_rss < _rss) \
290 (mm)->hiwater_rss = _rss; \
291} while (0)
292#define update_hiwater_vm(mm) do { \
293 if ((mm)->hiwater_vm < (mm)->total_vm) \
294 (mm)->hiwater_vm = (mm)->total_vm; \
295} while (0)
296
1da177e4
LT
297struct mm_struct {
298 struct vm_area_struct * mmap; /* list of VMAs */
299 struct rb_root mm_rb;
300 struct vm_area_struct * mmap_cache; /* last find_vma result */
301 unsigned long (*get_unmapped_area) (struct file *filp,
302 unsigned long addr, unsigned long len,
303 unsigned long pgoff, unsigned long flags);
1363c3cd 304 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
0551fbd2
BH
305 unsigned long mmap_base; /* base of mmap area */
306 unsigned long task_size; /* size of task vm space */
307 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
1363c3cd 308 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
1da177e4
LT
309 pgd_t * pgd;
310 atomic_t mm_users; /* How many users with user space? */
311 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
312 int map_count; /* number of VMAs */
313 struct rw_semaphore mmap_sem;
314 spinlock_t page_table_lock; /* Protects page tables and some counters */
315
316 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
317 * together off init_mm.mmlist, and are protected
318 * by mmlist_lock
319 */
320
f412ac08
HD
321 /* Special counters, in some configurations protected by the
322 * page_table_lock, in other configurations by being atomic.
323 */
4294621f 324 mm_counter_t _file_rss;
1da177e4
LT
325 mm_counter_t _anon_rss;
326
f449952b
HD
327 unsigned long hiwater_rss; /* High-watermark of RSS usage */
328 unsigned long hiwater_vm; /* High-water virtual memory usage */
329
330 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
331 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
332 unsigned long start_code, end_code, start_data, end_data;
333 unsigned long start_brk, brk, start_stack;
334 unsigned long arg_start, arg_end, env_start, env_end;
335
36d57ac4 336 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
1da177e4 337
d6e71144 338 unsigned dumpable:2;
1da177e4
LT
339 cpumask_t cpu_vm_mask;
340
341 /* Architecture-specific MM context */
342 mm_context_t context;
343
344 /* Token based thrashing protection. */
345 unsigned long swap_token_time;
346 char recent_pagein;
347
348 /* coredumping support */
349 int core_waiters;
350 struct completion *core_startup_done, core_done;
351
352 /* aio bits */
353 rwlock_t ioctx_list_lock;
354 struct kioctx *ioctx_list;
1da177e4
LT
355};
356
357struct sighand_struct {
358 atomic_t count;
359 struct k_sigaction action[_NSIG];
360 spinlock_t siglock;
361};
362
363/*
364 * NOTE! "signal_struct" does not have it's own
365 * locking, because a shared signal_struct always
366 * implies a shared sighand_struct, so locking
367 * sighand_struct is always a proper superset of
368 * the locking of signal_struct.
369 */
370struct signal_struct {
371 atomic_t count;
372 atomic_t live;
373
374 wait_queue_head_t wait_chldexit; /* for wait4() */
375
376 /* current thread group signal load-balancing target: */
377 task_t *curr_target;
378
379 /* shared signal handling: */
380 struct sigpending shared_pending;
381
382 /* thread group exit support */
383 int group_exit_code;
384 /* overloaded:
385 * - notify group_exit_task when ->count is equal to notify_count
386 * - everyone except group_exit_task is stopped during signal delivery
387 * of fatal signals, group_exit_task processes the signal.
388 */
389 struct task_struct *group_exit_task;
390 int notify_count;
391
392 /* thread group stop support, overloads group_exit_code too */
393 int group_stop_count;
394 unsigned int flags; /* see SIGNAL_* flags below */
395
396 /* POSIX.1b Interval Timers */
397 struct list_head posix_timers;
398
399 /* ITIMER_REAL timer for the process */
2ff678b8 400 struct hrtimer real_timer;
05cfb614 401 struct task_struct *tsk;
2ff678b8 402 ktime_t it_real_incr;
1da177e4
LT
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
b0a9499c
IM
472 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
473 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
474 * values are inverted: lower p->prio value means higher priority.
1da177e4
LT
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))
d425b274 489#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
1da177e4
LT
490
491/*
492 * Some day this will be a full-fledged user tracking system..
493 */
494struct user_struct {
495 atomic_t __count; /* reference count */
496 atomic_t processes; /* How many processes does this user have? */
497 atomic_t files; /* How many open files does this user have? */
498 atomic_t sigpending; /* How many pending signals does this user have? */
0eeca283
RL
499#ifdef CONFIG_INOTIFY
500 atomic_t inotify_watches; /* How many inotify watches does this user have? */
501 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
502#endif
1da177e4
LT
503 /* protected by mq_lock */
504 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
505 unsigned long locked_shm; /* How many pages of mlocked shm ? */
506
507#ifdef CONFIG_KEYS
508 struct key *uid_keyring; /* UID specific keyring */
509 struct key *session_keyring; /* UID's default session keyring */
510#endif
511
512 /* Hash table maintenance information */
513 struct list_head uidhash_list;
514 uid_t uid;
515};
516
517extern struct user_struct *find_user(uid_t);
518
519extern struct user_struct root_user;
520#define INIT_USER (&root_user)
521
522typedef struct prio_array prio_array_t;
523struct backing_dev_info;
524struct reclaim_state;
525
526#ifdef CONFIG_SCHEDSTATS
527struct sched_info {
528 /* cumulative counters */
529 unsigned long cpu_time, /* time spent on the cpu */
530 run_delay, /* time spent waiting on a runqueue */
531 pcnt; /* # of timeslices run on this cpu */
532
533 /* timestamps */
534 unsigned long last_arrival, /* when we last ran on a cpu */
535 last_queued; /* when we were last queued to run */
536};
537
538extern struct file_operations proc_schedstat_operations;
539#endif
540
541enum idle_type
542{
543 SCHED_IDLE,
544 NOT_IDLE,
545 NEWLY_IDLE,
546 MAX_IDLE_TYPES
547};
548
549/*
550 * sched-domains (multiprocessor balancing) declarations:
551 */
552#ifdef CONFIG_SMP
553#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
554
555#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
556#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
557#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
558#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
559#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
560#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
561#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
562#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
1da177e4
LT
563
564struct sched_group {
565 struct sched_group *next; /* Must be a circular list */
566 cpumask_t cpumask;
567
568 /*
569 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
570 * single CPU. This is read only (except for setup, hotplug CPU).
571 */
572 unsigned long cpu_power;
573};
574
575struct sched_domain {
576 /* These fields must be setup */
577 struct sched_domain *parent; /* top domain must be null terminated */
578 struct sched_group *groups; /* the balancing groups of the domain */
579 cpumask_t span; /* span of all CPUs in this domain */
580 unsigned long min_interval; /* Minimum balance interval ms */
581 unsigned long max_interval; /* Maximum balance interval ms */
582 unsigned int busy_factor; /* less balancing by factor if busy */
583 unsigned int imbalance_pct; /* No balance until over watermark */
584 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
585 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
586 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
7897986b
NP
587 unsigned int busy_idx;
588 unsigned int idle_idx;
589 unsigned int newidle_idx;
590 unsigned int wake_idx;
147cbb4b 591 unsigned int forkexec_idx;
1da177e4
LT
592 int flags; /* See SD_* */
593
594 /* Runtime fields. */
595 unsigned long last_balance; /* init to jiffies. units in jiffies */
596 unsigned int balance_interval; /* initialise to 1. units in ms. */
597 unsigned int nr_balance_failed; /* initialise to 0 */
598
599#ifdef CONFIG_SCHEDSTATS
600 /* load_balance() stats */
601 unsigned long lb_cnt[MAX_IDLE_TYPES];
602 unsigned long lb_failed[MAX_IDLE_TYPES];
603 unsigned long lb_balanced[MAX_IDLE_TYPES];
604 unsigned long lb_imbalance[MAX_IDLE_TYPES];
605 unsigned long lb_gained[MAX_IDLE_TYPES];
606 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
607 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
608 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
609
610 /* Active load balancing */
611 unsigned long alb_cnt;
612 unsigned long alb_failed;
613 unsigned long alb_pushed;
614
68767a0a
NP
615 /* SD_BALANCE_EXEC stats */
616 unsigned long sbe_cnt;
617 unsigned long sbe_balanced;
1da177e4
LT
618 unsigned long sbe_pushed;
619
68767a0a
NP
620 /* SD_BALANCE_FORK stats */
621 unsigned long sbf_cnt;
622 unsigned long sbf_balanced;
623 unsigned long sbf_pushed;
624
1da177e4
LT
625 /* try_to_wake_up() stats */
626 unsigned long ttwu_wake_remote;
627 unsigned long ttwu_move_affine;
628 unsigned long ttwu_move_balance;
629#endif
630};
631
1a20ff27
DG
632extern void partition_sched_domains(cpumask_t *partition1,
633 cpumask_t *partition2);
198e2f18 634
635/*
636 * Maximum cache size the migration-costs auto-tuning code will
637 * search from:
638 */
639extern unsigned int max_cache_size;
640
641#endif /* CONFIG_SMP */
1da177e4
LT
642
643
644struct io_context; /* See blkdev.h */
645void exit_io_context(void);
646struct cpuset;
647
648#define NGROUPS_SMALL 32
649#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
650struct group_info {
651 int ngroups;
652 atomic_t usage;
653 gid_t small_block[NGROUPS_SMALL];
654 int nblocks;
655 gid_t *blocks[0];
656};
657
658/*
659 * get_group_info() must be called with the owning task locked (via task_lock())
660 * when task != current. The reason being that the vast majority of callers are
661 * looking at current->group_info, which can not be changed except by the
662 * current task. Changing current->group_info requires the task lock, too.
663 */
664#define get_group_info(group_info) do { \
665 atomic_inc(&(group_info)->usage); \
666} while (0)
667
668#define put_group_info(group_info) do { \
669 if (atomic_dec_and_test(&(group_info)->usage)) \
670 groups_free(group_info); \
671} while (0)
672
3e30148c
DH
673extern struct group_info *groups_alloc(int gidsetsize);
674extern void groups_free(struct group_info *group_info);
675extern int set_current_groups(struct group_info *group_info);
676extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
677/* access the groups "array" with this macro */
678#define GROUP_AT(gi, i) \
679 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
680
383f2835
CK
681#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
682extern void prefetch_stack(struct task_struct*);
683#else
684static inline void prefetch_stack(struct task_struct *t) { }
685#endif
1da177e4
LT
686
687struct audit_context; /* See audit.c */
688struct mempolicy;
b92ce558 689struct pipe_inode_info;
1da177e4 690
3dee386e
CK
691enum sleep_type {
692 SLEEP_NORMAL,
693 SLEEP_NONINTERACTIVE,
694 SLEEP_INTERACTIVE,
695 SLEEP_INTERRUPTED,
696};
697
1da177e4
LT
698struct task_struct {
699 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
700 struct thread_info *thread_info;
701 atomic_t usage;
702 unsigned long flags; /* per process flags, defined below */
703 unsigned long ptrace;
704
36772092 705 int lock_depth; /* BKL lock depth */
1da177e4 706
d6077cb8 707#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
4866cde0
NP
708 int oncpu;
709#endif
1da177e4
LT
710 int prio, static_prio;
711 struct list_head run_list;
712 prio_array_t *array;
713
22e2c507 714 unsigned short ioprio;
2056a782 715 unsigned int btrace_seq;
22e2c507 716
1da177e4
LT
717 unsigned long sleep_avg;
718 unsigned long long timestamp, last_ran;
719 unsigned long long sched_time; /* sched_clock time spent running */
3dee386e 720 enum sleep_type sleep_type;
1da177e4
LT
721
722 unsigned long policy;
723 cpumask_t cpus_allowed;
724 unsigned int time_slice, first_time_slice;
725
726#ifdef CONFIG_SCHEDSTATS
727 struct sched_info sched_info;
728#endif
729
730 struct list_head tasks;
731 /*
732 * ptrace_list/ptrace_children forms the list of my children
733 * that were stolen by a ptracer.
734 */
735 struct list_head ptrace_children;
736 struct list_head ptrace_list;
737
738 struct mm_struct *mm, *active_mm;
739
740/* task state */
741 struct linux_binfmt *binfmt;
742 long exit_state;
743 int exit_code, exit_signal;
744 int pdeath_signal; /* The signal sent when the parent dies */
745 /* ??? */
746 unsigned long personality;
747 unsigned did_exec:1;
748 pid_t pid;
749 pid_t tgid;
750 /*
751 * pointers to (original) parent process, youngest child, younger sibling,
752 * older sibling, respectively. (p->father can be replaced with
753 * p->parent->pid)
754 */
755 struct task_struct *real_parent; /* real parent process (when being debugged) */
756 struct task_struct *parent; /* parent process */
757 /*
758 * children/sibling forms the list of my children plus the
759 * tasks I'm ptracing.
760 */
761 struct list_head children; /* list of my children */
762 struct list_head sibling; /* linkage in my parent's children list */
763 struct task_struct *group_leader; /* threadgroup leader */
764
765 /* PID/PID hash table linkage. */
92476d7f 766 struct pid_link pids[PIDTYPE_MAX];
47e65328 767 struct list_head thread_group;
1da177e4
LT
768
769 struct completion *vfork_done; /* for vfork() */
770 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
771 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
772
773 unsigned long rt_priority;
774 cputime_t utime, stime;
775 unsigned long nvcsw, nivcsw; /* context switch counts */
776 struct timespec start_time;
777/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
778 unsigned long min_flt, maj_flt;
779
780 cputime_t it_prof_expires, it_virt_expires;
781 unsigned long long it_sched_expires;
782 struct list_head cpu_timers[3];
783
784/* process credentials */
785 uid_t uid,euid,suid,fsuid;
786 gid_t gid,egid,sgid,fsgid;
787 struct group_info *group_info;
788 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
789 unsigned keep_capabilities:1;
790 struct user_struct *user;
791#ifdef CONFIG_KEYS
b5f545c8 792 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 793 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 794 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4
LT
795#endif
796 int oomkilladj; /* OOM kill score adjustment (bit shift). */
36772092
PBG
797 char comm[TASK_COMM_LEN]; /* executable name excluding path
798 - access with [gs]et_task_comm (which lock
799 it with task_lock())
800 - initialized normally by flush_old_exec */
1da177e4
LT
801/* file system info */
802 int link_count, total_link_count;
803/* ipc stuff */
804 struct sysv_sem sysvsem;
805/* CPU-specific state of this task */
806 struct thread_struct thread;
807/* filesystem information */
808 struct fs_struct *fs;
809/* open file information */
810 struct files_struct *files;
811/* namespace */
812 struct namespace *namespace;
813/* signal handlers */
814 struct signal_struct *signal;
815 struct sighand_struct *sighand;
816
817 sigset_t blocked, real_blocked;
150256d8 818 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
1da177e4
LT
819 struct sigpending pending;
820
821 unsigned long sas_ss_sp;
822 size_t sas_ss_size;
823 int (*notifier)(void *priv);
824 void *notifier_data;
825 sigset_t *notifier_mask;
826
827 void *security;
828 struct audit_context *audit_context;
829 seccomp_t seccomp;
830
831/* Thread group tracking */
832 u32 parent_exec_id;
833 u32 self_exec_id;
834/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
835 spinlock_t alloc_lock;
836/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
837 spinlock_t proc_lock;
1da177e4 838
408894ee
IM
839#ifdef CONFIG_DEBUG_MUTEXES
840 /* mutex deadlock detection */
841 struct mutex_waiter *blocked_on;
842#endif
843
1da177e4
LT
844/* journalling filesystem info */
845 void *journal_info;
846
847/* VM state */
848 struct reclaim_state *reclaim_state;
849
850 struct dentry *proc_dentry;
851 struct backing_dev_info *backing_dev_info;
852
853 struct io_context *io_context;
854
855 unsigned long ptrace_message;
856 siginfo_t *last_siginfo; /* For ptrace use. */
857/*
858 * current io wait handle: wait queue entry to use for io waits
859 * If this thread is processing aio, this points at the waitqueue
860 * inside the currently handled kiocb. It may be NULL (i.e. default
861 * to a stack based synchronous wait) if its doing sync IO.
862 */
863 wait_queue_t *io_wait;
864/* i/o counters(bytes read/written, #syscalls */
865 u64 rchar, wchar, syscr, syscw;
866#if defined(CONFIG_BSD_PROCESS_ACCT)
867 u64 acct_rss_mem1; /* accumulated rss usage */
868 u64 acct_vm_mem1; /* accumulated virtual memory usage */
869 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
870#endif
871#ifdef CONFIG_NUMA
872 struct mempolicy *mempolicy;
873 short il_next;
874#endif
875#ifdef CONFIG_CPUSETS
876 struct cpuset *cpuset;
877 nodemask_t mems_allowed;
878 int cpuset_mems_generation;
825a46af 879 int cpuset_mem_spread_rotor;
1da177e4 880#endif
0771dfef 881 struct robust_list_head __user *robust_list;
34f192c6
IM
882#ifdef CONFIG_COMPAT
883 struct compat_robust_list_head __user *compat_robust_list;
884#endif
0771dfef 885
22e2c507 886 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 887 struct rcu_head rcu;
b92ce558
JA
888
889 /*
890 * cache last used pipe for splice
891 */
892 struct pipe_inode_info *splice_pipe;
1da177e4
LT
893};
894
895static inline pid_t process_group(struct task_struct *tsk)
896{
897 return tsk->signal->pgrp;
898}
899
900/**
901 * pid_alive - check that a task structure is not stale
902 * @p: Task structure to be checked.
903 *
904 * Test if a process is not yet dead (at most zombie state)
905 * If pid_alive fails, then pointers within the task structure
906 * can be stale and must not be dereferenced.
907 */
908static inline int pid_alive(struct task_struct *p)
909{
92476d7f 910 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
911}
912
913extern void free_task(struct task_struct *tsk);
1da177e4 914#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 915
158d9ebd 916extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
917
918static inline void put_task_struct(struct task_struct *t)
919{
920 if (atomic_dec_and_test(&t->usage))
8c7904a0 921 __put_task_struct(t);
e56d0903 922}
1da177e4
LT
923
924/*
925 * Per process flags
926 */
927#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
928 /* Not implemented yet, only for 486*/
929#define PF_STARTING 0x00000002 /* being created */
930#define PF_EXITING 0x00000004 /* getting shut down */
931#define PF_DEAD 0x00000008 /* Dead */
932#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
933#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
934#define PF_DUMPCORE 0x00000200 /* dumped core */
935#define PF_SIGNALED 0x00000400 /* killed by a signal */
936#define PF_MEMALLOC 0x00000800 /* Allocating memory */
937#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
938#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
939#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
940#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
941#define PF_FROZEN 0x00010000 /* frozen for system suspend */
942#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
943#define PF_KSWAPD 0x00040000 /* I am kswapd */
944#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
945#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
946#define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
947#define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
948#define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
930d9152 949#define PF_SWAPWRITE 0x01000000 /* Allowed to write to swap */
825a46af
PJ
950#define PF_SPREAD_PAGE 0x04000000 /* Spread page cache over cpuset */
951#define PF_SPREAD_SLAB 0x08000000 /* Spread some slab caches over cpuset */
c61afb18 952#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1da177e4
LT
953
954/*
955 * Only the _current_ task can read/write to tsk->flags, but other
956 * tasks can access tsk->flags in readonly mode for example
957 * with tsk_used_math (like during threaded core dumping).
958 * There is however an exception to this rule during ptrace
959 * or during fork: the ptracer task is allowed to write to the
960 * child->flags of its traced child (same goes for fork, the parent
961 * can write to the child->flags), because we're guaranteed the
962 * child is not running and in turn not changing child->flags
963 * at the same time the parent does it.
964 */
965#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
966#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
967#define clear_used_math() clear_stopped_child_used_math(current)
968#define set_used_math() set_stopped_child_used_math(current)
969#define conditional_stopped_child_used_math(condition, child) \
970 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
971#define conditional_used_math(condition) \
972 conditional_stopped_child_used_math(condition, current)
973#define copy_to_stopped_child_used_math(child) \
974 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
975/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
976#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
977#define used_math() tsk_used_math(current)
978
979#ifdef CONFIG_SMP
980extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
981#else
982static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
983{
4098f991 984 if (!cpu_isset(0, new_mask))
1da177e4
LT
985 return -EINVAL;
986 return 0;
987}
988#endif
989
990extern unsigned long long sched_clock(void);
991extern unsigned long long current_sched_time(const task_t *current_task);
992
993/* sched_exec is called by processes performing an exec */
994#ifdef CONFIG_SMP
995extern void sched_exec(void);
996#else
997#define sched_exec() {}
998#endif
999
1000#ifdef CONFIG_HOTPLUG_CPU
1001extern void idle_task_exit(void);
1002#else
1003static inline void idle_task_exit(void) {}
1004#endif
1005
1006extern void sched_idle_next(void);
1007extern void set_user_nice(task_t *p, long nice);
1008extern int task_prio(const task_t *p);
1009extern int task_nice(const task_t *p);
e43379f1 1010extern int can_nice(const task_t *p, const int nice);
1da177e4
LT
1011extern int task_curr(const task_t *p);
1012extern int idle_cpu(int cpu);
1013extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1014extern task_t *idle_task(int cpu);
a2a97982
KO
1015extern task_t *curr_task(int cpu);
1016extern void set_curr_task(int cpu, task_t *p);
1da177e4
LT
1017
1018void yield(void);
1019
1020/*
1021 * The default (Linux) execution domain.
1022 */
1023extern struct exec_domain default_exec_domain;
1024
1025union thread_union {
1026 struct thread_info thread_info;
1027 unsigned long stack[THREAD_SIZE/sizeof(long)];
1028};
1029
1030#ifndef __HAVE_ARCH_KSTACK_END
1031static inline int kstack_end(void *addr)
1032{
1033 /* Reliable end of stack detection:
1034 * Some APM bios versions misalign the stack
1035 */
1036 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1037}
1038#endif
1039
1040extern union thread_union init_thread_union;
1041extern struct task_struct init_task;
1042
1043extern struct mm_struct init_mm;
1044
1045#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1046extern struct task_struct *find_task_by_pid_type(int type, int pid);
1047extern void set_special_pids(pid_t session, pid_t pgrp);
1048extern void __set_special_pids(pid_t session, pid_t pgrp);
1049
1050/* per-UID process charging. */
1051extern struct user_struct * alloc_uid(uid_t);
1052static inline struct user_struct *get_uid(struct user_struct *u)
1053{
1054 atomic_inc(&u->__count);
1055 return u;
1056}
1057extern void free_uid(struct user_struct *);
1058extern void switch_uid(struct user_struct *);
1059
1060#include <asm/current.h>
1061
1062extern void do_timer(struct pt_regs *);
1063
1064extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1065extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1066extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1067 unsigned long clone_flags));
1068#ifdef CONFIG_SMP
1069 extern void kick_process(struct task_struct *tsk);
1070#else
1071 static inline void kick_process(struct task_struct *tsk) { }
1072#endif
476d139c 1073extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
1da177e4
LT
1074extern void FASTCALL(sched_exit(task_t * p));
1075
1076extern int in_group_p(gid_t);
1077extern int in_egroup_p(gid_t);
1078
1079extern void proc_caches_init(void);
1080extern void flush_signals(struct task_struct *);
1081extern void flush_signal_handlers(struct task_struct *, int force_default);
1082extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1083
1084static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1085{
1086 unsigned long flags;
1087 int ret;
1088
1089 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1090 ret = dequeue_signal(tsk, mask, info);
1091 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1092
1093 return ret;
1094}
1095
1096extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1097 sigset_t *mask);
1098extern void unblock_all_signals(void);
1099extern void release_task(struct task_struct * p);
1100extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1101extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1102extern int force_sigsegv(int, struct task_struct *);
1103extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1104extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1105extern int kill_pg_info(int, struct siginfo *, pid_t);
1106extern int kill_proc_info(int, struct siginfo *, pid_t);
46113830 1107extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t);
1da177e4
LT
1108extern void do_notify_parent(struct task_struct *, int);
1109extern void force_sig(int, struct task_struct *);
1110extern void force_sig_specific(int, struct task_struct *);
1111extern int send_sig(int, struct task_struct *, int);
1112extern void zap_other_threads(struct task_struct *p);
1113extern int kill_pg(pid_t, int, int);
1da177e4
LT
1114extern int kill_proc(pid_t, int, int);
1115extern struct sigqueue *sigqueue_alloc(void);
1116extern void sigqueue_free(struct sigqueue *);
1117extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1118extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
9ac95f2f 1119extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
1120extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1121
1122/* These can be the second arg to send_sig_info/send_group_sig_info. */
1123#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1124#define SEND_SIG_PRIV ((struct siginfo *) 1)
1125#define SEND_SIG_FORCED ((struct siginfo *) 2)
1126
621d3121
ON
1127static inline int is_si_special(const struct siginfo *info)
1128{
1129 return info <= SEND_SIG_FORCED;
1130}
1131
1da177e4
LT
1132/* True if we are on the alternate signal stack. */
1133
1134static inline int on_sig_stack(unsigned long sp)
1135{
1136 return (sp - current->sas_ss_sp < current->sas_ss_size);
1137}
1138
1139static inline int sas_ss_flags(unsigned long sp)
1140{
1141 return (current->sas_ss_size == 0 ? SS_DISABLE
1142 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1143}
1144
1da177e4
LT
1145/*
1146 * Routines for handling mm_structs
1147 */
1148extern struct mm_struct * mm_alloc(void);
1149
1150/* mmdrop drops the mm and the page tables */
1151extern void FASTCALL(__mmdrop(struct mm_struct *));
1152static inline void mmdrop(struct mm_struct * mm)
1153{
1154 if (atomic_dec_and_test(&mm->mm_count))
1155 __mmdrop(mm);
1156}
1157
1158/* mmput gets rid of the mappings and all user-space */
1159extern void mmput(struct mm_struct *);
1160/* Grab a reference to a task's mm, if it is not already going away */
1161extern struct mm_struct *get_task_mm(struct task_struct *task);
1162/* Remove the current tasks stale references to the old mm_struct */
1163extern void mm_release(struct task_struct *, struct mm_struct *);
1164
1165extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1166extern void flush_thread(void);
1167extern void exit_thread(void);
1168
1da177e4 1169extern void exit_files(struct task_struct *);
6b3934ef 1170extern void __cleanup_signal(struct signal_struct *);
a7e5328a 1171extern void __cleanup_sighand(struct sighand_struct *);
1da177e4
LT
1172extern void exit_itimers(struct signal_struct *);
1173
1174extern NORET_TYPE void do_group_exit(int);
1175
1da177e4
LT
1176extern void daemonize(const char *, ...);
1177extern int allow_signal(int);
1178extern int disallow_signal(int);
1179extern task_t *child_reaper;
1180
1181extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1182extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1183task_t *fork_idle(int);
1184
1185extern void set_task_comm(struct task_struct *tsk, char *from);
1186extern void get_task_comm(char *to, struct task_struct *tsk);
1187
1188#ifdef CONFIG_SMP
1189extern void wait_task_inactive(task_t * p);
1190#else
1191#define wait_task_inactive(p) do { } while (0)
1192#endif
1193
1194#define remove_parent(p) list_del_init(&(p)->sibling)
8fafabd8 1195#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1da177e4 1196
5e85d4ab 1197#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1da177e4
LT
1198
1199#define for_each_process(p) \
1200 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1201
1202/*
1203 * Careful: do_each_thread/while_each_thread is a double loop so
1204 * 'break' will not work as expected - use goto instead.
1205 */
1206#define do_each_thread(g, t) \
1207 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1208
1209#define while_each_thread(g, t) \
1210 while ((t = next_thread(t)) != g)
1211
de12a787
EB
1212/* de_thread depends on thread_group_leader not being a pid based check */
1213#define thread_group_leader(p) (p == p->group_leader)
1da177e4 1214
a9cdf410 1215static inline task_t *next_thread(const task_t *p)
47e65328
ON
1216{
1217 return list_entry(rcu_dereference(p->thread_group.next),
1218 task_t, thread_group);
1219}
1220
1da177e4
LT
1221static inline int thread_group_empty(task_t *p)
1222{
47e65328 1223 return list_empty(&p->thread_group);
1da177e4
LT
1224}
1225
1226#define delay_group_leader(p) \
1227 (thread_group_leader(p) && !thread_group_empty(p))
1228
1da177e4
LT
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
f63ee72e
ON
1248extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1249 unsigned long *flags);
1250
1251static inline void unlock_task_sighand(struct task_struct *tsk,
1252 unsigned long *flags)
1253{
1254 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1255}
1256
f037360f
AV
1257#ifndef __HAVE_THREAD_FUNCTIONS
1258
a1261f54 1259#define task_thread_info(task) (task)->thread_info
9fc65876 1260#define task_stack_page(task) ((void*)((task)->thread_info))
a1261f54 1261
10ebffde
AV
1262static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1263{
1264 *task_thread_info(p) = *task_thread_info(org);
1265 task_thread_info(p)->task = p;
1266}
1267
1268static inline unsigned long *end_of_stack(struct task_struct *p)
1269{
1270 return (unsigned long *)(p->thread_info + 1);
1271}
1272
f037360f
AV
1273#endif
1274
1da177e4
LT
1275/* set thread flags in other task's structures
1276 * - see asm/thread_info.h for TIF_xxxx flags available
1277 */
1278static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1279{
a1261f54 1280 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1281}
1282
1283static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1284{
a1261f54 1285 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1286}
1287
1288static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1289{
a1261f54 1290 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1291}
1292
1293static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1294{
a1261f54 1295 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1296}
1297
1298static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1299{
a1261f54 1300 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1301}
1302
1303static inline void set_tsk_need_resched(struct task_struct *tsk)
1304{
1305 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1306}
1307
1308static inline void clear_tsk_need_resched(struct task_struct *tsk)
1309{
1310 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1311}
1312
1313static inline int signal_pending(struct task_struct *p)
1314{
1315 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1316}
1317
1318static inline int need_resched(void)
1319{
1320 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1321}
1322
1323/*
1324 * cond_resched() and cond_resched_lock(): latency reduction via
1325 * explicit rescheduling in places that are safe. The return
1326 * value indicates whether a reschedule was done in fact.
1327 * cond_resched_lock() will drop the spinlock before scheduling,
1328 * cond_resched_softirq() will enable bhs before scheduling.
1329 */
1330extern int cond_resched(void);
1331extern int cond_resched_lock(spinlock_t * lock);
1332extern int cond_resched_softirq(void);
1333
1334/*
1335 * Does a critical section need to be broken due to another
1336 * task waiting?:
1337 */
1338#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1339# define need_lockbreak(lock) ((lock)->break_lock)
1340#else
1341# define need_lockbreak(lock) 0
1342#endif
1343
1344/*
1345 * Does a critical section need to be broken due to another
1346 * task waiting or preemption being signalled:
1347 */
1348static inline int lock_need_resched(spinlock_t *lock)
1349{
1350 if (need_lockbreak(lock) || need_resched())
1351 return 1;
1352 return 0;
1353}
1354
1355/* Reevaluate whether the task has signals pending delivery.
1356 This is required every time the blocked sigset_t changes.
1357 callers must hold sighand->siglock. */
1358
1359extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1360extern void recalc_sigpending(void);
1361
1362extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1363
1364/*
1365 * Wrappers for p->thread_info->cpu access. No-op on UP.
1366 */
1367#ifdef CONFIG_SMP
1368
1369static inline unsigned int task_cpu(const struct task_struct *p)
1370{
a1261f54 1371 return task_thread_info(p)->cpu;
1da177e4
LT
1372}
1373
1374static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1375{
a1261f54 1376 task_thread_info(p)->cpu = cpu;
1da177e4
LT
1377}
1378
1379#else
1380
1381static inline unsigned int task_cpu(const struct task_struct *p)
1382{
1383 return 0;
1384}
1385
1386static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1387{
1388}
1389
1390#endif /* CONFIG_SMP */
1391
1392#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1393extern void arch_pick_mmap_layout(struct mm_struct *mm);
1394#else
1395static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1396{
1397 mm->mmap_base = TASK_UNMAPPED_BASE;
1398 mm->get_unmapped_area = arch_get_unmapped_area;
1399 mm->unmap_area = arch_unmap_area;
1400}
1401#endif
1402
1403extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1404extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1405
1da177e4
LT
1406extern void normalize_rt_tasks(void);
1407
1da177e4 1408#ifdef CONFIG_PM
3e1d1d28
CL
1409/*
1410 * Check if a process has been frozen
1411 */
1412static inline int frozen(struct task_struct *p)
1413{
1414 return p->flags & PF_FROZEN;
1415}
1416
1417/*
1418 * Check if there is a request to freeze a process
1419 */
1420static inline int freezing(struct task_struct *p)
1421{
1422 return p->flags & PF_FREEZE;
1423}
1424
1425/*
1426 * Request that a process be frozen
1427 * FIXME: SMP problem. We may not modify other process' flags!
1428 */
1429static inline void freeze(struct task_struct *p)
1430{
1431 p->flags |= PF_FREEZE;
1432}
1433
1434/*
1435 * Wake up a frozen process
1436 */
1437static inline int thaw_process(struct task_struct *p)
1438{
1439 if (frozen(p)) {
1440 p->flags &= ~PF_FROZEN;
1441 wake_up_process(p);
1442 return 1;
1443 }
1444 return 0;
1445}
1446
1447/*
1448 * freezing is complete, mark process as frozen
1449 */
1450static inline void frozen_process(struct task_struct *p)
1451{
1452 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1453}
1454
1455extern void refrigerator(void);
1da177e4
LT
1456extern int freeze_processes(void);
1457extern void thaw_processes(void);
1458
3e1d1d28 1459static inline int try_to_freeze(void)
1da177e4 1460{
3e1d1d28
CL
1461 if (freezing(current)) {
1462 refrigerator();
1da177e4
LT
1463 return 1;
1464 } else
1465 return 0;
1466}
1467#else
3e1d1d28
CL
1468static inline int frozen(struct task_struct *p) { return 0; }
1469static inline int freezing(struct task_struct *p) { return 0; }
1470static inline void freeze(struct task_struct *p) { BUG(); }
1471static inline int thaw_process(struct task_struct *p) { return 1; }
1472static inline void frozen_process(struct task_struct *p) { BUG(); }
1473
1474static inline void refrigerator(void) {}
1da177e4
LT
1475static inline int freeze_processes(void) { BUG(); return 0; }
1476static inline void thaw_processes(void) {}
1477
3e1d1d28
CL
1478static inline int try_to_freeze(void) { return 0; }
1479
1da177e4
LT
1480#endif /* CONFIG_PM */
1481#endif /* __KERNEL__ */
1482
1483#endif