]> git.proxmox.com Git - mirror_ubuntu-focal-kernel.git/blame - include/linux/sched.h
[PATCH] libata: add a function to decide if we need iordy
[mirror_ubuntu-focal-kernel.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>
2ff678b8 108#include <linux/hrtimer.h>
1da177e4
LT
109
110#include <asm/processor.h>
111
4a8342d2
LT
112/*
113 * Task state bitmask. NOTE! These bits are also
114 * encoded in fs/proc/array.c: get_task_state().
115 *
116 * We have two separate sets of flags: task->state
117 * is about runnability, while task->exit_state are
118 * about the task exiting. Confusing, but this way
119 * modifying one set can't modify the other one by
120 * mistake.
121 */
1da177e4
LT
122#define TASK_RUNNING 0
123#define TASK_INTERRUPTIBLE 1
124#define TASK_UNINTERRUPTIBLE 2
4a8342d2
LT
125#define TASK_STOPPED 4
126#define TASK_TRACED 8
127/* in tsk->exit_state */
128#define EXIT_ZOMBIE 16
129#define EXIT_DEAD 32
130/* in tsk->state again */
131#define TASK_NONINTERACTIVE 64
1da177e4
LT
132
133#define __set_task_state(tsk, state_value) \
134 do { (tsk)->state = (state_value); } while (0)
135#define set_task_state(tsk, state_value) \
136 set_mb((tsk)->state, (state_value))
137
498d0c57
AM
138/*
139 * set_current_state() includes a barrier so that the write of current->state
140 * is correctly serialised wrt the caller's subsequent test of whether to
141 * actually sleep:
142 *
143 * set_current_state(TASK_UNINTERRUPTIBLE);
144 * if (do_i_need_to_sleep())
145 * schedule();
146 *
147 * If the caller does not need such serialisation then use __set_current_state()
148 */
1da177e4
LT
149#define __set_current_state(state_value) \
150 do { current->state = (state_value); } while (0)
151#define set_current_state(state_value) \
152 set_mb(current->state, (state_value))
153
154/* Task command name length */
155#define TASK_COMM_LEN 16
156
157/*
158 * Scheduling policies
159 */
160#define SCHED_NORMAL 0
161#define SCHED_FIFO 1
162#define SCHED_RR 2
b0a9499c 163#define SCHED_BATCH 3
1da177e4
LT
164
165struct sched_param {
166 int sched_priority;
167};
168
169#ifdef __KERNEL__
170
171#include <linux/spinlock.h>
172
173/*
174 * This serializes "schedule()" and also protects
175 * the run-queue from deletions/modifications (but
176 * _adding_ to the beginning of the run-queue has
177 * a separate lock).
178 */
179extern rwlock_t tasklist_lock;
180extern spinlock_t mmlist_lock;
181
182typedef struct task_struct task_t;
183
184extern void sched_init(void);
185extern void sched_init_smp(void);
186extern void init_idle(task_t *idle, int cpu);
187
188extern cpumask_t nohz_cpu_mask;
189
190extern void show_state(void);
191extern void show_regs(struct pt_regs *);
192
193/*
194 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
195 * task), SP is the stack pointer of the first frame that should be shown in the back
196 * trace (or NULL if the entire call-chain of the task should be shown).
197 */
198extern void show_stack(struct task_struct *task, unsigned long *sp);
199
200void io_schedule(void);
201long io_schedule_timeout(long timeout);
202
203extern void cpu_init (void);
204extern void trap_init(void);
205extern void update_process_times(int user);
206extern void scheduler_tick(void);
207
8446f1d3
IM
208#ifdef CONFIG_DETECT_SOFTLOCKUP
209extern void softlockup_tick(struct pt_regs *regs);
210extern void spawn_softlockup_task(void);
211extern void touch_softlockup_watchdog(void);
212#else
213static inline void softlockup_tick(struct pt_regs *regs)
214{
215}
216static inline void spawn_softlockup_task(void)
217{
218}
219static inline void touch_softlockup_watchdog(void)
220{
221}
222#endif
223
224
1da177e4
LT
225/* Attach to any functions which should be ignored in wchan output. */
226#define __sched __attribute__((__section__(".sched.text")))
227/* Is this address in the __sched functions? */
228extern int in_sched_functions(unsigned long addr);
229
230#define MAX_SCHEDULE_TIMEOUT LONG_MAX
231extern signed long FASTCALL(schedule_timeout(signed long timeout));
64ed93a2
NA
232extern signed long schedule_timeout_interruptible(signed long timeout);
233extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
234asmlinkage void schedule(void);
235
236struct namespace;
237
238/* Maximum number of active map areas.. This is a random (large) number */
239#define DEFAULT_MAX_MAP_COUNT 65536
240
241extern int sysctl_max_map_count;
242
243#include <linux/aio.h>
244
245extern unsigned long
246arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
247 unsigned long, unsigned long);
248extern unsigned long
249arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
250 unsigned long len, unsigned long pgoff,
251 unsigned long flags);
1363c3cd
WW
252extern void arch_unmap_area(struct mm_struct *, unsigned long);
253extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 254
f412ac08
HD
255#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
256/*
257 * The mm counters are not protected by its page_table_lock,
258 * so must be incremented atomically.
259 */
d3cb4871
CL
260#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
261#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
262#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
263#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
264#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
265typedef atomic_long_t mm_counter_t;
f412ac08
HD
266
267#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
268/*
269 * The mm counters are protected by its page_table_lock,
270 * so can be incremented directly.
271 */
1da177e4
LT
272#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
273#define get_mm_counter(mm, member) ((mm)->_##member)
274#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
275#define inc_mm_counter(mm, member) (mm)->_##member++
276#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
277typedef unsigned long mm_counter_t;
278
279#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 280
f412ac08
HD
281#define get_mm_rss(mm) \
282 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
283#define update_hiwater_rss(mm) do { \
284 unsigned long _rss = get_mm_rss(mm); \
285 if ((mm)->hiwater_rss < _rss) \
286 (mm)->hiwater_rss = _rss; \
287} while (0)
288#define update_hiwater_vm(mm) do { \
289 if ((mm)->hiwater_vm < (mm)->total_vm) \
290 (mm)->hiwater_vm = (mm)->total_vm; \
291} while (0)
292
1da177e4
LT
293struct mm_struct {
294 struct vm_area_struct * mmap; /* list of VMAs */
295 struct rb_root mm_rb;
296 struct vm_area_struct * mmap_cache; /* last find_vma result */
297 unsigned long (*get_unmapped_area) (struct file *filp,
298 unsigned long addr, unsigned long len,
299 unsigned long pgoff, unsigned long flags);
1363c3cd
WW
300 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
301 unsigned long mmap_base; /* base of mmap area */
302 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
303 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
1da177e4
LT
304 pgd_t * pgd;
305 atomic_t mm_users; /* How many users with user space? */
306 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
307 int map_count; /* number of VMAs */
308 struct rw_semaphore mmap_sem;
309 spinlock_t page_table_lock; /* Protects page tables and some counters */
310
311 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
312 * together off init_mm.mmlist, and are protected
313 * by mmlist_lock
314 */
315
f412ac08
HD
316 /* Special counters, in some configurations protected by the
317 * page_table_lock, in other configurations by being atomic.
318 */
4294621f 319 mm_counter_t _file_rss;
1da177e4
LT
320 mm_counter_t _anon_rss;
321
f449952b
HD
322 unsigned long hiwater_rss; /* High-watermark of RSS usage */
323 unsigned long hiwater_vm; /* High-water virtual memory usage */
324
325 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
326 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
327 unsigned long start_code, end_code, start_data, end_data;
328 unsigned long start_brk, brk, start_stack;
329 unsigned long arg_start, arg_end, env_start, env_end;
330
36d57ac4 331 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
1da177e4 332
d6e71144 333 unsigned dumpable:2;
1da177e4
LT
334 cpumask_t cpu_vm_mask;
335
336 /* Architecture-specific MM context */
337 mm_context_t context;
338
339 /* Token based thrashing protection. */
340 unsigned long swap_token_time;
341 char recent_pagein;
342
343 /* coredumping support */
344 int core_waiters;
345 struct completion *core_startup_done, core_done;
346
347 /* aio bits */
348 rwlock_t ioctx_list_lock;
349 struct kioctx *ioctx_list;
1da177e4
LT
350};
351
352struct sighand_struct {
353 atomic_t count;
354 struct k_sigaction action[_NSIG];
355 spinlock_t siglock;
e56d0903 356 struct rcu_head rcu;
1da177e4
LT
357};
358
e56d0903
IM
359extern void sighand_free_cb(struct rcu_head *rhp);
360
361static inline void sighand_free(struct sighand_struct *sp)
362{
363 call_rcu(&sp->rcu, sighand_free_cb);
364}
365
1da177e4
LT
366/*
367 * NOTE! "signal_struct" does not have it's own
368 * locking, because a shared signal_struct always
369 * implies a shared sighand_struct, so locking
370 * sighand_struct is always a proper superset of
371 * the locking of signal_struct.
372 */
373struct signal_struct {
374 atomic_t count;
375 atomic_t live;
376
377 wait_queue_head_t wait_chldexit; /* for wait4() */
378
379 /* current thread group signal load-balancing target: */
380 task_t *curr_target;
381
382 /* shared signal handling: */
383 struct sigpending shared_pending;
384
385 /* thread group exit support */
386 int group_exit_code;
387 /* overloaded:
388 * - notify group_exit_task when ->count is equal to notify_count
389 * - everyone except group_exit_task is stopped during signal delivery
390 * of fatal signals, group_exit_task processes the signal.
391 */
392 struct task_struct *group_exit_task;
393 int notify_count;
394
395 /* thread group stop support, overloads group_exit_code too */
396 int group_stop_count;
397 unsigned int flags; /* see SIGNAL_* flags below */
398
399 /* POSIX.1b Interval Timers */
400 struct list_head posix_timers;
401
402 /* ITIMER_REAL timer for the process */
2ff678b8
TG
403 struct hrtimer real_timer;
404 ktime_t it_real_incr;
1da177e4
LT
405
406 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
407 cputime_t it_prof_expires, it_virt_expires;
408 cputime_t it_prof_incr, it_virt_incr;
409
410 /* job control IDs */
411 pid_t pgrp;
412 pid_t tty_old_pgrp;
413 pid_t session;
414 /* boolean value for session group leader */
415 int leader;
416
417 struct tty_struct *tty; /* NULL if no tty */
418
419 /*
420 * Cumulative resource counters for dead threads in the group,
421 * and for reaped dead child processes forked by this group.
422 * Live threads maintain their own counters and add to these
423 * in __exit_signal, except for the group leader.
424 */
425 cputime_t utime, stime, cutime, cstime;
426 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
427 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
428
429 /*
430 * Cumulative ns of scheduled CPU time for dead threads in the
431 * group, not including a zombie group leader. (This only differs
432 * from jiffies_to_ns(utime + stime) if sched_clock uses something
433 * other than jiffies.)
434 */
435 unsigned long long sched_time;
436
437 /*
438 * We don't bother to synchronize most readers of this at all,
439 * because there is no reader checking a limit that actually needs
440 * to get both rlim_cur and rlim_max atomically, and either one
441 * alone is a single word that can safely be read normally.
442 * getrlimit/setrlimit use task_lock(current->group_leader) to
443 * protect this instead of the siglock, because they really
444 * have no need to disable irqs.
445 */
446 struct rlimit rlim[RLIM_NLIMITS];
447
448 struct list_head cpu_timers[3];
449
450 /* keep the process-shared keyrings here so that they do the right
451 * thing in threads created with CLONE_THREAD */
452#ifdef CONFIG_KEYS
453 struct key *session_keyring; /* keyring inherited over fork */
454 struct key *process_keyring; /* keyring private to this process */
455#endif
456};
457
4866cde0
NP
458/* Context switch must be unlocked if interrupts are to be enabled */
459#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
460# define __ARCH_WANT_UNLOCKED_CTXSW
461#endif
462
1da177e4
LT
463/*
464 * Bits in flags field of signal_struct.
465 */
466#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
467#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
468#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
469#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
470
471
472/*
473 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
b0a9499c
IM
474 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
475 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
476 * values are inverted: lower p->prio value means higher priority.
1da177e4
LT
477 *
478 * The MAX_USER_RT_PRIO value allows the actual maximum
479 * RT priority to be separate from the value exported to
480 * user-space. This allows kernel threads to set their
481 * priority to a value higher than any user task. Note:
482 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
483 */
484
485#define MAX_USER_RT_PRIO 100
486#define MAX_RT_PRIO MAX_USER_RT_PRIO
487
488#define MAX_PRIO (MAX_RT_PRIO + 40)
489
490#define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
491
492/*
493 * Some day this will be a full-fledged user tracking system..
494 */
495struct user_struct {
496 atomic_t __count; /* reference count */
497 atomic_t processes; /* How many processes does this user have? */
498 atomic_t files; /* How many open files does this user have? */
499 atomic_t sigpending; /* How many pending signals does this user have? */
0eeca283
RL
500#ifdef CONFIG_INOTIFY
501 atomic_t inotify_watches; /* How many inotify watches does this user have? */
502 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
503#endif
1da177e4
LT
504 /* protected by mq_lock */
505 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
506 unsigned long locked_shm; /* How many pages of mlocked shm ? */
507
508#ifdef CONFIG_KEYS
509 struct key *uid_keyring; /* UID specific keyring */
510 struct key *session_keyring; /* UID's default session keyring */
511#endif
512
513 /* Hash table maintenance information */
514 struct list_head uidhash_list;
515 uid_t uid;
516};
517
518extern struct user_struct *find_user(uid_t);
519
520extern struct user_struct root_user;
521#define INIT_USER (&root_user)
522
523typedef struct prio_array prio_array_t;
524struct backing_dev_info;
525struct reclaim_state;
526
527#ifdef CONFIG_SCHEDSTATS
528struct sched_info {
529 /* cumulative counters */
530 unsigned long cpu_time, /* time spent on the cpu */
531 run_delay, /* time spent waiting on a runqueue */
532 pcnt; /* # of timeslices run on this cpu */
533
534 /* timestamps */
535 unsigned long last_arrival, /* when we last ran on a cpu */
536 last_queued; /* when we were last queued to run */
537};
538
539extern struct file_operations proc_schedstat_operations;
540#endif
541
542enum idle_type
543{
544 SCHED_IDLE,
545 NOT_IDLE,
546 NEWLY_IDLE,
547 MAX_IDLE_TYPES
548};
549
550/*
551 * sched-domains (multiprocessor balancing) declarations:
552 */
553#ifdef CONFIG_SMP
554#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
555
556#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
557#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
558#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
559#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
560#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
561#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
562#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
563#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
1da177e4
LT
564
565struct sched_group {
566 struct sched_group *next; /* Must be a circular list */
567 cpumask_t cpumask;
568
569 /*
570 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
571 * single CPU. This is read only (except for setup, hotplug CPU).
572 */
573 unsigned long cpu_power;
574};
575
576struct sched_domain {
577 /* These fields must be setup */
578 struct sched_domain *parent; /* top domain must be null terminated */
579 struct sched_group *groups; /* the balancing groups of the domain */
580 cpumask_t span; /* span of all CPUs in this domain */
581 unsigned long min_interval; /* Minimum balance interval ms */
582 unsigned long max_interval; /* Maximum balance interval ms */
583 unsigned int busy_factor; /* less balancing by factor if busy */
584 unsigned int imbalance_pct; /* No balance until over watermark */
585 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
586 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
587 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
7897986b
NP
588 unsigned int busy_idx;
589 unsigned int idle_idx;
590 unsigned int newidle_idx;
591 unsigned int wake_idx;
147cbb4b 592 unsigned int forkexec_idx;
1da177e4
LT
593 int flags; /* See SD_* */
594
595 /* Runtime fields. */
596 unsigned long last_balance; /* init to jiffies. units in jiffies */
597 unsigned int balance_interval; /* initialise to 1. units in ms. */
598 unsigned int nr_balance_failed; /* initialise to 0 */
599
600#ifdef CONFIG_SCHEDSTATS
601 /* load_balance() stats */
602 unsigned long lb_cnt[MAX_IDLE_TYPES];
603 unsigned long lb_failed[MAX_IDLE_TYPES];
604 unsigned long lb_balanced[MAX_IDLE_TYPES];
605 unsigned long lb_imbalance[MAX_IDLE_TYPES];
606 unsigned long lb_gained[MAX_IDLE_TYPES];
607 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
608 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
609 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
610
611 /* Active load balancing */
612 unsigned long alb_cnt;
613 unsigned long alb_failed;
614 unsigned long alb_pushed;
615
68767a0a
NP
616 /* SD_BALANCE_EXEC stats */
617 unsigned long sbe_cnt;
618 unsigned long sbe_balanced;
1da177e4
LT
619 unsigned long sbe_pushed;
620
68767a0a
NP
621 /* SD_BALANCE_FORK stats */
622 unsigned long sbf_cnt;
623 unsigned long sbf_balanced;
624 unsigned long sbf_pushed;
625
1da177e4
LT
626 /* try_to_wake_up() stats */
627 unsigned long ttwu_wake_remote;
628 unsigned long ttwu_move_affine;
629 unsigned long ttwu_move_balance;
630#endif
631};
632
1a20ff27
DG
633extern void partition_sched_domains(cpumask_t *partition1,
634 cpumask_t *partition2);
198e2f18 635
636/*
637 * Maximum cache size the migration-costs auto-tuning code will
638 * search from:
639 */
640extern unsigned int max_cache_size;
641
642#endif /* CONFIG_SMP */
1da177e4
LT
643
644
645struct io_context; /* See blkdev.h */
646void exit_io_context(void);
647struct cpuset;
648
649#define NGROUPS_SMALL 32
650#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
651struct group_info {
652 int ngroups;
653 atomic_t usage;
654 gid_t small_block[NGROUPS_SMALL];
655 int nblocks;
656 gid_t *blocks[0];
657};
658
659/*
660 * get_group_info() must be called with the owning task locked (via task_lock())
661 * when task != current. The reason being that the vast majority of callers are
662 * looking at current->group_info, which can not be changed except by the
663 * current task. Changing current->group_info requires the task lock, too.
664 */
665#define get_group_info(group_info) do { \
666 atomic_inc(&(group_info)->usage); \
667} while (0)
668
669#define put_group_info(group_info) do { \
670 if (atomic_dec_and_test(&(group_info)->usage)) \
671 groups_free(group_info); \
672} while (0)
673
3e30148c
DH
674extern struct group_info *groups_alloc(int gidsetsize);
675extern void groups_free(struct group_info *group_info);
676extern int set_current_groups(struct group_info *group_info);
677extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
678/* access the groups "array" with this macro */
679#define GROUP_AT(gi, i) \
680 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
681
383f2835
KC
682#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
683extern void prefetch_stack(struct task_struct*);
684#else
685static inline void prefetch_stack(struct task_struct *t) { }
686#endif
1da177e4
LT
687
688struct audit_context; /* See audit.c */
689struct mempolicy;
690
691struct task_struct {
692 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
693 struct thread_info *thread_info;
694 atomic_t usage;
695 unsigned long flags; /* per process flags, defined below */
696 unsigned long ptrace;
697
36772092 698 int lock_depth; /* BKL lock depth */
1da177e4 699
d7102e95 700#if defined(CONFIG_SMP)
701 int last_waker_cpu; /* CPU that last woke this task up */
702#if defined(__ARCH_WANT_UNLOCKED_CTXSW)
4866cde0 703 int oncpu;
d7102e95 704#endif
4866cde0 705#endif
1da177e4
LT
706 int prio, static_prio;
707 struct list_head run_list;
708 prio_array_t *array;
709
22e2c507
JA
710 unsigned short ioprio;
711
1da177e4
LT
712 unsigned long sleep_avg;
713 unsigned long long timestamp, last_ran;
714 unsigned long long sched_time; /* sched_clock time spent running */
715 int activated;
716
717 unsigned long policy;
718 cpumask_t cpus_allowed;
719 unsigned int time_slice, first_time_slice;
720
721#ifdef CONFIG_SCHEDSTATS
722 struct sched_info sched_info;
723#endif
724
725 struct list_head tasks;
726 /*
727 * ptrace_list/ptrace_children forms the list of my children
728 * that were stolen by a ptracer.
729 */
730 struct list_head ptrace_children;
731 struct list_head ptrace_list;
732
733 struct mm_struct *mm, *active_mm;
734
735/* task state */
736 struct linux_binfmt *binfmt;
737 long exit_state;
738 int exit_code, exit_signal;
739 int pdeath_signal; /* The signal sent when the parent dies */
740 /* ??? */
741 unsigned long personality;
742 unsigned did_exec:1;
743 pid_t pid;
744 pid_t tgid;
745 /*
746 * pointers to (original) parent process, youngest child, younger sibling,
747 * older sibling, respectively. (p->father can be replaced with
748 * p->parent->pid)
749 */
750 struct task_struct *real_parent; /* real parent process (when being debugged) */
751 struct task_struct *parent; /* parent process */
752 /*
753 * children/sibling forms the list of my children plus the
754 * tasks I'm ptracing.
755 */
756 struct list_head children; /* list of my children */
757 struct list_head sibling; /* linkage in my parent's children list */
758 struct task_struct *group_leader; /* threadgroup leader */
759
760 /* PID/PID hash table linkage. */
761 struct pid pids[PIDTYPE_MAX];
762
763 struct completion *vfork_done; /* for vfork() */
764 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
765 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
766
767 unsigned long rt_priority;
768 cputime_t utime, stime;
769 unsigned long nvcsw, nivcsw; /* context switch counts */
770 struct timespec start_time;
771/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
772 unsigned long min_flt, maj_flt;
773
774 cputime_t it_prof_expires, it_virt_expires;
775 unsigned long long it_sched_expires;
776 struct list_head cpu_timers[3];
777
778/* process credentials */
779 uid_t uid,euid,suid,fsuid;
780 gid_t gid,egid,sgid,fsgid;
781 struct group_info *group_info;
782 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
783 unsigned keep_capabilities:1;
784 struct user_struct *user;
785#ifdef CONFIG_KEYS
b5f545c8 786 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 787 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 788 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4
LT
789#endif
790 int oomkilladj; /* OOM kill score adjustment (bit shift). */
36772092
PBG
791 char comm[TASK_COMM_LEN]; /* executable name excluding path
792 - access with [gs]et_task_comm (which lock
793 it with task_lock())
794 - initialized normally by flush_old_exec */
1da177e4
LT
795/* file system info */
796 int link_count, total_link_count;
797/* ipc stuff */
798 struct sysv_sem sysvsem;
799/* CPU-specific state of this task */
800 struct thread_struct thread;
801/* filesystem information */
802 struct fs_struct *fs;
803/* open file information */
804 struct files_struct *files;
805/* namespace */
806 struct namespace *namespace;
807/* signal handlers */
808 struct signal_struct *signal;
809 struct sighand_struct *sighand;
810
811 sigset_t blocked, real_blocked;
812 struct sigpending pending;
813
814 unsigned long sas_ss_sp;
815 size_t sas_ss_size;
816 int (*notifier)(void *priv);
817 void *notifier_data;
818 sigset_t *notifier_mask;
819
820 void *security;
821 struct audit_context *audit_context;
822 seccomp_t seccomp;
823
824/* Thread group tracking */
825 u32 parent_exec_id;
826 u32 self_exec_id;
827/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
828 spinlock_t alloc_lock;
829/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
830 spinlock_t proc_lock;
1da177e4 831
408894ee
IM
832#ifdef CONFIG_DEBUG_MUTEXES
833 /* mutex deadlock detection */
834 struct mutex_waiter *blocked_on;
835#endif
836
1da177e4
LT
837/* journalling filesystem info */
838 void *journal_info;
839
840/* VM state */
841 struct reclaim_state *reclaim_state;
842
843 struct dentry *proc_dentry;
844 struct backing_dev_info *backing_dev_info;
845
846 struct io_context *io_context;
847
848 unsigned long ptrace_message;
849 siginfo_t *last_siginfo; /* For ptrace use. */
850/*
851 * current io wait handle: wait queue entry to use for io waits
852 * If this thread is processing aio, this points at the waitqueue
853 * inside the currently handled kiocb. It may be NULL (i.e. default
854 * to a stack based synchronous wait) if its doing sync IO.
855 */
856 wait_queue_t *io_wait;
857/* i/o counters(bytes read/written, #syscalls */
858 u64 rchar, wchar, syscr, syscw;
859#if defined(CONFIG_BSD_PROCESS_ACCT)
860 u64 acct_rss_mem1; /* accumulated rss usage */
861 u64 acct_vm_mem1; /* accumulated virtual memory usage */
862 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
863#endif
864#ifdef CONFIG_NUMA
865 struct mempolicy *mempolicy;
866 short il_next;
867#endif
868#ifdef CONFIG_CPUSETS
869 struct cpuset *cpuset;
870 nodemask_t mems_allowed;
871 int cpuset_mems_generation;
872#endif
22e2c507 873 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 874 struct rcu_head rcu;
1da177e4
LT
875};
876
877static inline pid_t process_group(struct task_struct *tsk)
878{
879 return tsk->signal->pgrp;
880}
881
882/**
883 * pid_alive - check that a task structure is not stale
884 * @p: Task structure to be checked.
885 *
886 * Test if a process is not yet dead (at most zombie state)
887 * If pid_alive fails, then pointers within the task structure
888 * can be stale and must not be dereferenced.
889 */
890static inline int pid_alive(struct task_struct *p)
891{
892 return p->pids[PIDTYPE_PID].nr != 0;
893}
894
895extern void free_task(struct task_struct *tsk);
896extern void __put_task_struct(struct task_struct *tsk);
897#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 898
e56d0903
IM
899extern void __put_task_struct_cb(struct rcu_head *rhp);
900
901static inline void put_task_struct(struct task_struct *t)
902{
903 if (atomic_dec_and_test(&t->usage))
904 call_rcu(&t->rcu, __put_task_struct_cb);
905}
1da177e4
LT
906
907/*
908 * Per process flags
909 */
910#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
911 /* Not implemented yet, only for 486*/
912#define PF_STARTING 0x00000002 /* being created */
913#define PF_EXITING 0x00000004 /* getting shut down */
914#define PF_DEAD 0x00000008 /* Dead */
915#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
916#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
917#define PF_DUMPCORE 0x00000200 /* dumped core */
918#define PF_SIGNALED 0x00000400 /* killed by a signal */
919#define PF_MEMALLOC 0x00000800 /* Allocating memory */
920#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
921#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
922#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
923#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
924#define PF_FROZEN 0x00010000 /* frozen for system suspend */
925#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
926#define PF_KSWAPD 0x00040000 /* I am kswapd */
927#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
928#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
929#define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
930#define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
931#define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
930d9152 932#define PF_SWAPWRITE 0x01000000 /* Allowed to write to swap */
1da177e4
LT
933
934/*
935 * Only the _current_ task can read/write to tsk->flags, but other
936 * tasks can access tsk->flags in readonly mode for example
937 * with tsk_used_math (like during threaded core dumping).
938 * There is however an exception to this rule during ptrace
939 * or during fork: the ptracer task is allowed to write to the
940 * child->flags of its traced child (same goes for fork, the parent
941 * can write to the child->flags), because we're guaranteed the
942 * child is not running and in turn not changing child->flags
943 * at the same time the parent does it.
944 */
945#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
946#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
947#define clear_used_math() clear_stopped_child_used_math(current)
948#define set_used_math() set_stopped_child_used_math(current)
949#define conditional_stopped_child_used_math(condition, child) \
950 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
951#define conditional_used_math(condition) \
952 conditional_stopped_child_used_math(condition, current)
953#define copy_to_stopped_child_used_math(child) \
954 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
955/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
956#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
957#define used_math() tsk_used_math(current)
958
959#ifdef CONFIG_SMP
960extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
961#else
962static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
963{
4098f991 964 if (!cpu_isset(0, new_mask))
1da177e4
LT
965 return -EINVAL;
966 return 0;
967}
968#endif
969
970extern unsigned long long sched_clock(void);
971extern unsigned long long current_sched_time(const task_t *current_task);
972
973/* sched_exec is called by processes performing an exec */
974#ifdef CONFIG_SMP
975extern void sched_exec(void);
976#else
977#define sched_exec() {}
978#endif
979
980#ifdef CONFIG_HOTPLUG_CPU
981extern void idle_task_exit(void);
982#else
983static inline void idle_task_exit(void) {}
984#endif
985
986extern void sched_idle_next(void);
987extern void set_user_nice(task_t *p, long nice);
988extern int task_prio(const task_t *p);
989extern int task_nice(const task_t *p);
e43379f1 990extern int can_nice(const task_t *p, const int nice);
1da177e4
LT
991extern int task_curr(const task_t *p);
992extern int idle_cpu(int cpu);
993extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
994extern task_t *idle_task(int cpu);
a2a97982
KO
995extern task_t *curr_task(int cpu);
996extern void set_curr_task(int cpu, task_t *p);
1da177e4
LT
997
998void yield(void);
999
1000/*
1001 * The default (Linux) execution domain.
1002 */
1003extern struct exec_domain default_exec_domain;
1004
1005union thread_union {
1006 struct thread_info thread_info;
1007 unsigned long stack[THREAD_SIZE/sizeof(long)];
1008};
1009
1010#ifndef __HAVE_ARCH_KSTACK_END
1011static inline int kstack_end(void *addr)
1012{
1013 /* Reliable end of stack detection:
1014 * Some APM bios versions misalign the stack
1015 */
1016 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1017}
1018#endif
1019
1020extern union thread_union init_thread_union;
1021extern struct task_struct init_task;
1022
1023extern struct mm_struct init_mm;
1024
1025#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1026extern struct task_struct *find_task_by_pid_type(int type, int pid);
1027extern void set_special_pids(pid_t session, pid_t pgrp);
1028extern void __set_special_pids(pid_t session, pid_t pgrp);
1029
1030/* per-UID process charging. */
1031extern struct user_struct * alloc_uid(uid_t);
1032static inline struct user_struct *get_uid(struct user_struct *u)
1033{
1034 atomic_inc(&u->__count);
1035 return u;
1036}
1037extern void free_uid(struct user_struct *);
1038extern void switch_uid(struct user_struct *);
1039
1040#include <asm/current.h>
1041
1042extern void do_timer(struct pt_regs *);
1043
1044extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1045extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1046extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1047 unsigned long clone_flags));
1048#ifdef CONFIG_SMP
1049 extern void kick_process(struct task_struct *tsk);
1050#else
1051 static inline void kick_process(struct task_struct *tsk) { }
1052#endif
476d139c 1053extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
1da177e4
LT
1054extern void FASTCALL(sched_exit(task_t * p));
1055
1056extern int in_group_p(gid_t);
1057extern int in_egroup_p(gid_t);
1058
1059extern void proc_caches_init(void);
1060extern void flush_signals(struct task_struct *);
1061extern void flush_signal_handlers(struct task_struct *, int force_default);
1062extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1063
1064static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1065{
1066 unsigned long flags;
1067 int ret;
1068
1069 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1070 ret = dequeue_signal(tsk, mask, info);
1071 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1072
1073 return ret;
1074}
1075
1076extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1077 sigset_t *mask);
1078extern void unblock_all_signals(void);
1079extern void release_task(struct task_struct * p);
1080extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1081extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1082extern int force_sigsegv(int, struct task_struct *);
1083extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1084extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1085extern int kill_pg_info(int, struct siginfo *, pid_t);
1086extern int kill_proc_info(int, struct siginfo *, pid_t);
46113830 1087extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t);
1da177e4
LT
1088extern void do_notify_parent(struct task_struct *, int);
1089extern void force_sig(int, struct task_struct *);
1090extern void force_sig_specific(int, struct task_struct *);
1091extern int send_sig(int, struct task_struct *, int);
1092extern void zap_other_threads(struct task_struct *p);
1093extern int kill_pg(pid_t, int, int);
1094extern int kill_sl(pid_t, int, int);
1095extern int kill_proc(pid_t, int, int);
1096extern struct sigqueue *sigqueue_alloc(void);
1097extern void sigqueue_free(struct sigqueue *);
1098extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1099extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
1100extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
1101extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1102
1103/* These can be the second arg to send_sig_info/send_group_sig_info. */
1104#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1105#define SEND_SIG_PRIV ((struct siginfo *) 1)
1106#define SEND_SIG_FORCED ((struct siginfo *) 2)
1107
621d3121
ON
1108static inline int is_si_special(const struct siginfo *info)
1109{
1110 return info <= SEND_SIG_FORCED;
1111}
1112
1da177e4
LT
1113/* True if we are on the alternate signal stack. */
1114
1115static inline int on_sig_stack(unsigned long sp)
1116{
1117 return (sp - current->sas_ss_sp < current->sas_ss_size);
1118}
1119
1120static inline int sas_ss_flags(unsigned long sp)
1121{
1122 return (current->sas_ss_size == 0 ? SS_DISABLE
1123 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1124}
1125
1da177e4
LT
1126/*
1127 * Routines for handling mm_structs
1128 */
1129extern struct mm_struct * mm_alloc(void);
1130
1131/* mmdrop drops the mm and the page tables */
1132extern void FASTCALL(__mmdrop(struct mm_struct *));
1133static inline void mmdrop(struct mm_struct * mm)
1134{
1135 if (atomic_dec_and_test(&mm->mm_count))
1136 __mmdrop(mm);
1137}
1138
1139/* mmput gets rid of the mappings and all user-space */
1140extern void mmput(struct mm_struct *);
1141/* Grab a reference to a task's mm, if it is not already going away */
1142extern struct mm_struct *get_task_mm(struct task_struct *task);
1143/* Remove the current tasks stale references to the old mm_struct */
1144extern void mm_release(struct task_struct *, struct mm_struct *);
1145
1146extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1147extern void flush_thread(void);
1148extern void exit_thread(void);
1149
1da177e4
LT
1150extern void exit_files(struct task_struct *);
1151extern void exit_signal(struct task_struct *);
1152extern void __exit_signal(struct task_struct *);
1153extern void exit_sighand(struct task_struct *);
1154extern void __exit_sighand(struct task_struct *);
1155extern void exit_itimers(struct signal_struct *);
1156
1157extern NORET_TYPE void do_group_exit(int);
1158
1da177e4
LT
1159extern void daemonize(const char *, ...);
1160extern int allow_signal(int);
1161extern int disallow_signal(int);
1162extern task_t *child_reaper;
1163
1164extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1165extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1166task_t *fork_idle(int);
1167
1168extern void set_task_comm(struct task_struct *tsk, char *from);
1169extern void get_task_comm(char *to, struct task_struct *tsk);
1170
1171#ifdef CONFIG_SMP
1172extern void wait_task_inactive(task_t * p);
1173#else
1174#define wait_task_inactive(p) do { } while (0)
1175#endif
1176
1177#define remove_parent(p) list_del_init(&(p)->sibling)
1178#define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1179
1180#define REMOVE_LINKS(p) do { \
1181 if (thread_group_leader(p)) \
1182 list_del_init(&(p)->tasks); \
1183 remove_parent(p); \
1184 } while (0)
1185
1186#define SET_LINKS(p) do { \
1187 if (thread_group_leader(p)) \
1188 list_add_tail(&(p)->tasks,&init_task.tasks); \
1189 add_parent(p, (p)->parent); \
1190 } while (0)
1191
1192#define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1193#define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1194
1195#define for_each_process(p) \
1196 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1197
1198/*
1199 * Careful: do_each_thread/while_each_thread is a double loop so
1200 * 'break' will not work as expected - use goto instead.
1201 */
1202#define do_each_thread(g, t) \
1203 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1204
1205#define while_each_thread(g, t) \
1206 while ((t = next_thread(t)) != g)
1207
1208extern task_t * FASTCALL(next_thread(const task_t *p));
1209
1210#define thread_group_leader(p) (p->pid == p->tgid)
1211
1212static inline int thread_group_empty(task_t *p)
1213{
1214 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1215}
1216
1217#define delay_group_leader(p) \
1218 (thread_group_leader(p) && !thread_group_empty(p))
1219
1220extern void unhash_process(struct task_struct *p);
1221
1222/*
1223 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
22e2c507 1224 * subscriptions and synchronises with wait4(). Also used in procfs. Also
053199ed 1225 * pins the final release of task.io_context. Also protects ->cpuset.
1da177e4
LT
1226 *
1227 * Nests both inside and outside of read_lock(&tasklist_lock).
1228 * It must not be nested with write_lock_irq(&tasklist_lock),
1229 * neither inside nor outside.
1230 */
1231static inline void task_lock(struct task_struct *p)
1232{
1233 spin_lock(&p->alloc_lock);
1234}
1235
1236static inline void task_unlock(struct task_struct *p)
1237{
1238 spin_unlock(&p->alloc_lock);
1239}
1240
f037360f
AV
1241#ifndef __HAVE_THREAD_FUNCTIONS
1242
a1261f54 1243#define task_thread_info(task) (task)->thread_info
9fc65876 1244#define task_stack_page(task) ((void*)((task)->thread_info))
a1261f54 1245
10ebffde
AV
1246static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1247{
1248 *task_thread_info(p) = *task_thread_info(org);
1249 task_thread_info(p)->task = p;
1250}
1251
1252static inline unsigned long *end_of_stack(struct task_struct *p)
1253{
1254 return (unsigned long *)(p->thread_info + 1);
1255}
1256
f037360f
AV
1257#endif
1258
1da177e4
LT
1259/* set thread flags in other task's structures
1260 * - see asm/thread_info.h for TIF_xxxx flags available
1261 */
1262static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1263{
a1261f54 1264 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1265}
1266
1267static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1268{
a1261f54 1269 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1270}
1271
1272static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1273{
a1261f54 1274 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1275}
1276
1277static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1278{
a1261f54 1279 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1280}
1281
1282static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1283{
a1261f54 1284 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1285}
1286
1287static inline void set_tsk_need_resched(struct task_struct *tsk)
1288{
1289 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1290}
1291
1292static inline void clear_tsk_need_resched(struct task_struct *tsk)
1293{
1294 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1295}
1296
1297static inline int signal_pending(struct task_struct *p)
1298{
1299 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1300}
1301
1302static inline int need_resched(void)
1303{
1304 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1305}
1306
1307/*
1308 * cond_resched() and cond_resched_lock(): latency reduction via
1309 * explicit rescheduling in places that are safe. The return
1310 * value indicates whether a reschedule was done in fact.
1311 * cond_resched_lock() will drop the spinlock before scheduling,
1312 * cond_resched_softirq() will enable bhs before scheduling.
1313 */
1314extern int cond_resched(void);
1315extern int cond_resched_lock(spinlock_t * lock);
1316extern int cond_resched_softirq(void);
1317
1318/*
1319 * Does a critical section need to be broken due to another
1320 * task waiting?:
1321 */
1322#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1323# define need_lockbreak(lock) ((lock)->break_lock)
1324#else
1325# define need_lockbreak(lock) 0
1326#endif
1327
1328/*
1329 * Does a critical section need to be broken due to another
1330 * task waiting or preemption being signalled:
1331 */
1332static inline int lock_need_resched(spinlock_t *lock)
1333{
1334 if (need_lockbreak(lock) || need_resched())
1335 return 1;
1336 return 0;
1337}
1338
1339/* Reevaluate whether the task has signals pending delivery.
1340 This is required every time the blocked sigset_t changes.
1341 callers must hold sighand->siglock. */
1342
1343extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1344extern void recalc_sigpending(void);
1345
1346extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1347
1348/*
1349 * Wrappers for p->thread_info->cpu access. No-op on UP.
1350 */
1351#ifdef CONFIG_SMP
1352
1353static inline unsigned int task_cpu(const struct task_struct *p)
1354{
a1261f54 1355 return task_thread_info(p)->cpu;
1da177e4
LT
1356}
1357
1358static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1359{
a1261f54 1360 task_thread_info(p)->cpu = cpu;
1da177e4
LT
1361}
1362
1363#else
1364
1365static inline unsigned int task_cpu(const struct task_struct *p)
1366{
1367 return 0;
1368}
1369
1370static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1371{
1372}
1373
1374#endif /* CONFIG_SMP */
1375
1376#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1377extern void arch_pick_mmap_layout(struct mm_struct *mm);
1378#else
1379static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1380{
1381 mm->mmap_base = TASK_UNMAPPED_BASE;
1382 mm->get_unmapped_area = arch_get_unmapped_area;
1383 mm->unmap_area = arch_unmap_area;
1384}
1385#endif
1386
1387extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1388extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1389
1da177e4
LT
1390extern void normalize_rt_tasks(void);
1391
1da177e4 1392#ifdef CONFIG_PM
3e1d1d28
CL
1393/*
1394 * Check if a process has been frozen
1395 */
1396static inline int frozen(struct task_struct *p)
1397{
1398 return p->flags & PF_FROZEN;
1399}
1400
1401/*
1402 * Check if there is a request to freeze a process
1403 */
1404static inline int freezing(struct task_struct *p)
1405{
1406 return p->flags & PF_FREEZE;
1407}
1408
1409/*
1410 * Request that a process be frozen
1411 * FIXME: SMP problem. We may not modify other process' flags!
1412 */
1413static inline void freeze(struct task_struct *p)
1414{
1415 p->flags |= PF_FREEZE;
1416}
1417
1418/*
1419 * Wake up a frozen process
1420 */
1421static inline int thaw_process(struct task_struct *p)
1422{
1423 if (frozen(p)) {
1424 p->flags &= ~PF_FROZEN;
1425 wake_up_process(p);
1426 return 1;
1427 }
1428 return 0;
1429}
1430
1431/*
1432 * freezing is complete, mark process as frozen
1433 */
1434static inline void frozen_process(struct task_struct *p)
1435{
1436 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1437}
1438
1439extern void refrigerator(void);
1da177e4
LT
1440extern int freeze_processes(void);
1441extern void thaw_processes(void);
1442
3e1d1d28 1443static inline int try_to_freeze(void)
1da177e4 1444{
3e1d1d28
CL
1445 if (freezing(current)) {
1446 refrigerator();
1da177e4
LT
1447 return 1;
1448 } else
1449 return 0;
1450}
1451#else
3e1d1d28
CL
1452static inline int frozen(struct task_struct *p) { return 0; }
1453static inline int freezing(struct task_struct *p) { return 0; }
1454static inline void freeze(struct task_struct *p) { BUG(); }
1455static inline int thaw_process(struct task_struct *p) { return 1; }
1456static inline void frozen_process(struct task_struct *p) { BUG(); }
1457
1458static inline void refrigerator(void) {}
1da177e4
LT
1459static inline int freeze_processes(void) { BUG(); return 0; }
1460static inline void thaw_processes(void) {}
1461
3e1d1d28
CL
1462static inline int try_to_freeze(void) { return 0; }
1463
1da177e4
LT
1464#endif /* CONFIG_PM */
1465#endif /* __KERNEL__ */
1466
1467#endif