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