[PATCH] pid: replace is_orphaned_pgrp with is_current_pgrp_orphaned
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4#include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
5
6/*
7 * cloning flags:
8 */
9#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
10#define CLONE_VM 0x00000100 /* set if VM shared between processes */
11#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
12#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
13#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
14#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
15#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
16#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
17#define CLONE_THREAD 0x00010000 /* Same thread group? */
18#define CLONE_NEWNS 0x00020000 /* New namespace group? */
19#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
20#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
21#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
22#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
23#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
24#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
25#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
26#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
071df104 27#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 28#define CLONE_NEWIPC 0x08000000 /* New ipcs */
b7b3c76a
DW
29
30/*
31 * Scheduling policies
32 */
33#define SCHED_NORMAL 0
34#define SCHED_FIFO 1
35#define SCHED_RR 2
36#define SCHED_BATCH 3
37
a3b6714e 38#ifdef __KERNEL__
b7b3c76a
DW
39
40struct sched_param {
41 int sched_priority;
42};
43
1da177e4
LT
44#include <asm/param.h> /* for HZ */
45
1da177e4
LT
46#include <linux/capability.h>
47#include <linux/threads.h>
48#include <linux/kernel.h>
49#include <linux/types.h>
50#include <linux/timex.h>
51#include <linux/jiffies.h>
52#include <linux/rbtree.h>
53#include <linux/thread_info.h>
54#include <linux/cpumask.h>
55#include <linux/errno.h>
56#include <linux/nodemask.h>
57
58#include <asm/system.h>
59#include <asm/semaphore.h>
60#include <asm/page.h>
61#include <asm/ptrace.h>
62#include <asm/mmu.h>
63#include <asm/cputime.h>
64
65#include <linux/smp.h>
66#include <linux/sem.h>
67#include <linux/signal.h>
68#include <linux/securebits.h>
69#include <linux/fs_struct.h>
70#include <linux/compiler.h>
71#include <linux/completion.h>
72#include <linux/pid.h>
73#include <linux/percpu.h>
74#include <linux/topology.h>
75#include <linux/seccomp.h>
e56d0903 76#include <linux/rcupdate.h>
0771dfef 77#include <linux/futex.h>
23f78d4a 78#include <linux/rtmutex.h>
1da177e4 79
a3b6714e
DW
80#include <linux/time.h>
81#include <linux/param.h>
82#include <linux/resource.h>
83#include <linux/timer.h>
84#include <linux/hrtimer.h>
7c3ab738 85#include <linux/task_io_accounting.h>
a3b6714e
DW
86
87#include <asm/processor.h>
36d57ac4 88
1da177e4 89struct exec_domain;
c87e2837 90struct futex_pi_state;
1da177e4 91
1da177e4
LT
92/*
93 * List of flags we want to share for kernel threads,
94 * if only because they are not used by them anyway.
95 */
96#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
97
98/*
99 * These are the constant used to fake the fixed-point load-average
100 * counting. Some notes:
101 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
102 * a load-average precision of 10 bits integer + 11 bits fractional
103 * - if you want to count load-averages more often, you need more
104 * precision, or rounding will get you. With 2-second counting freq,
105 * the EXP_n values would be 1981, 2034 and 2043 if still using only
106 * 11 bit fractions.
107 */
108extern unsigned long avenrun[]; /* Load averages */
109
110#define FSHIFT 11 /* nr of bits of precision */
111#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
112#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
113#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
114#define EXP_5 2014 /* 1/exp(5sec/5min) */
115#define EXP_15 2037 /* 1/exp(5sec/15min) */
116
117#define CALC_LOAD(load,exp,n) \
118 load *= exp; \
119 load += n*(FIXED_1-exp); \
120 load >>= FSHIFT;
121
122extern unsigned long total_forks;
123extern int nr_threads;
1da177e4
LT
124DECLARE_PER_CPU(unsigned long, process_counts);
125extern int nr_processes(void);
126extern unsigned long nr_running(void);
127extern unsigned long nr_uninterruptible(void);
db1b1fef 128extern unsigned long nr_active(void);
1da177e4 129extern unsigned long nr_iowait(void);
2dd73a4f 130extern unsigned long weighted_cpuload(const int cpu);
1da177e4 131
1da177e4 132
4a8342d2
LT
133/*
134 * Task state bitmask. NOTE! These bits are also
135 * encoded in fs/proc/array.c: get_task_state().
136 *
137 * We have two separate sets of flags: task->state
138 * is about runnability, while task->exit_state are
139 * about the task exiting. Confusing, but this way
140 * modifying one set can't modify the other one by
141 * mistake.
142 */
1da177e4
LT
143#define TASK_RUNNING 0
144#define TASK_INTERRUPTIBLE 1
145#define TASK_UNINTERRUPTIBLE 2
4a8342d2
LT
146#define TASK_STOPPED 4
147#define TASK_TRACED 8
148/* in tsk->exit_state */
149#define EXIT_ZOMBIE 16
150#define EXIT_DEAD 32
151/* in tsk->state again */
152#define TASK_NONINTERACTIVE 64
c394cc9f 153#define TASK_DEAD 128
1da177e4
LT
154
155#define __set_task_state(tsk, state_value) \
156 do { (tsk)->state = (state_value); } while (0)
157#define set_task_state(tsk, state_value) \
158 set_mb((tsk)->state, (state_value))
159
498d0c57
AM
160/*
161 * set_current_state() includes a barrier so that the write of current->state
162 * is correctly serialised wrt the caller's subsequent test of whether to
163 * actually sleep:
164 *
165 * set_current_state(TASK_UNINTERRUPTIBLE);
166 * if (do_i_need_to_sleep())
167 * schedule();
168 *
169 * If the caller does not need such serialisation then use __set_current_state()
170 */
1da177e4
LT
171#define __set_current_state(state_value) \
172 do { current->state = (state_value); } while (0)
173#define set_current_state(state_value) \
174 set_mb(current->state, (state_value))
175
176/* Task command name length */
177#define TASK_COMM_LEN 16
178
1da177e4
LT
179#include <linux/spinlock.h>
180
181/*
182 * This serializes "schedule()" and also protects
183 * the run-queue from deletions/modifications (but
184 * _adding_ to the beginning of the run-queue has
185 * a separate lock).
186 */
187extern rwlock_t tasklist_lock;
188extern spinlock_t mmlist_lock;
189
36c8b586 190struct task_struct;
1da177e4
LT
191
192extern void sched_init(void);
193extern void sched_init_smp(void);
36c8b586 194extern void init_idle(struct task_struct *idle, int cpu);
1da177e4
LT
195
196extern cpumask_t nohz_cpu_mask;
197
e59e2ae2
IM
198/*
199 * Only dump TASK_* tasks. (-1 for all tasks)
200 */
201extern void show_state_filter(unsigned long state_filter);
202
203static inline void show_state(void)
204{
205 show_state_filter(-1);
206}
207
1da177e4
LT
208extern void show_regs(struct pt_regs *);
209
210/*
211 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
212 * task), SP is the stack pointer of the first frame that should be shown in the back
213 * trace (or NULL if the entire call-chain of the task should be shown).
214 */
215extern void show_stack(struct task_struct *task, unsigned long *sp);
216
217void io_schedule(void);
218long io_schedule_timeout(long timeout);
219
220extern void cpu_init (void);
221extern void trap_init(void);
222extern void update_process_times(int user);
223extern void scheduler_tick(void);
224
8446f1d3 225#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 226extern void softlockup_tick(void);
8446f1d3
IM
227extern void spawn_softlockup_task(void);
228extern void touch_softlockup_watchdog(void);
229#else
6687a97d 230static inline void softlockup_tick(void)
8446f1d3
IM
231{
232}
233static inline void spawn_softlockup_task(void)
234{
235}
236static inline void touch_softlockup_watchdog(void)
237{
238}
239#endif
240
241
1da177e4
LT
242/* Attach to any functions which should be ignored in wchan output. */
243#define __sched __attribute__((__section__(".sched.text")))
244/* Is this address in the __sched functions? */
245extern int in_sched_functions(unsigned long addr);
246
247#define MAX_SCHEDULE_TIMEOUT LONG_MAX
248extern signed long FASTCALL(schedule_timeout(signed long timeout));
64ed93a2
NA
249extern signed long schedule_timeout_interruptible(signed long timeout);
250extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
251asmlinkage void schedule(void);
252
ab516013 253struct nsproxy;
1da177e4
LT
254
255/* Maximum number of active map areas.. This is a random (large) number */
256#define DEFAULT_MAX_MAP_COUNT 65536
257
258extern int sysctl_max_map_count;
259
260#include <linux/aio.h>
261
262extern unsigned long
263arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
264 unsigned long, unsigned long);
265extern unsigned long
266arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
267 unsigned long len, unsigned long pgoff,
268 unsigned long flags);
1363c3cd
WW
269extern void arch_unmap_area(struct mm_struct *, unsigned long);
270extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 271
f412ac08
HD
272#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
273/*
274 * The mm counters are not protected by its page_table_lock,
275 * so must be incremented atomically.
276 */
d3cb4871
CL
277#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
278#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
279#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
280#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
281#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
282typedef atomic_long_t mm_counter_t;
f412ac08
HD
283
284#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
285/*
286 * The mm counters are protected by its page_table_lock,
287 * so can be incremented directly.
288 */
1da177e4
LT
289#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
290#define get_mm_counter(mm, member) ((mm)->_##member)
291#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
292#define inc_mm_counter(mm, member) (mm)->_##member++
293#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
294typedef unsigned long mm_counter_t;
295
296#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 297
f412ac08
HD
298#define get_mm_rss(mm) \
299 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
300#define update_hiwater_rss(mm) do { \
301 unsigned long _rss = get_mm_rss(mm); \
302 if ((mm)->hiwater_rss < _rss) \
303 (mm)->hiwater_rss = _rss; \
304} while (0)
305#define update_hiwater_vm(mm) do { \
306 if ((mm)->hiwater_vm < (mm)->total_vm) \
307 (mm)->hiwater_vm = (mm)->total_vm; \
308} while (0)
309
1da177e4
LT
310struct mm_struct {
311 struct vm_area_struct * mmap; /* list of VMAs */
312 struct rb_root mm_rb;
313 struct vm_area_struct * mmap_cache; /* last find_vma result */
314 unsigned long (*get_unmapped_area) (struct file *filp,
315 unsigned long addr, unsigned long len,
316 unsigned long pgoff, unsigned long flags);
1363c3cd 317 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
0551fbd2
BH
318 unsigned long mmap_base; /* base of mmap area */
319 unsigned long task_size; /* size of task vm space */
320 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
1363c3cd 321 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
1da177e4
LT
322 pgd_t * pgd;
323 atomic_t mm_users; /* How many users with user space? */
324 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
325 int map_count; /* number of VMAs */
326 struct rw_semaphore mmap_sem;
327 spinlock_t page_table_lock; /* Protects page tables and some counters */
328
329 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
330 * together off init_mm.mmlist, and are protected
331 * by mmlist_lock
332 */
333
f412ac08
HD
334 /* Special counters, in some configurations protected by the
335 * page_table_lock, in other configurations by being atomic.
336 */
4294621f 337 mm_counter_t _file_rss;
1da177e4
LT
338 mm_counter_t _anon_rss;
339
f449952b
HD
340 unsigned long hiwater_rss; /* High-watermark of RSS usage */
341 unsigned long hiwater_vm; /* High-water virtual memory usage */
342
343 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
344 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
345 unsigned long start_code, end_code, start_data, end_data;
346 unsigned long start_brk, brk, start_stack;
347 unsigned long arg_start, arg_end, env_start, env_end;
348
36d57ac4 349 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
1da177e4 350
1da177e4
LT
351 cpumask_t cpu_vm_mask;
352
353 /* Architecture-specific MM context */
354 mm_context_t context;
355
7602bdf2
AC
356 /* Swap token stuff */
357 /*
358 * Last value of global fault stamp as seen by this process.
359 * In other words, this value gives an indication of how long
360 * it has been since this task got the token.
361 * Look at mm/thrash.c
362 */
363 unsigned int faultstamp;
364 unsigned int token_priority;
365 unsigned int last_interval;
1da177e4 366
36de6437
ACM
367 unsigned char dumpable:2;
368
1da177e4
LT
369 /* coredumping support */
370 int core_waiters;
371 struct completion *core_startup_done, core_done;
372
373 /* aio bits */
374 rwlock_t ioctx_list_lock;
375 struct kioctx *ioctx_list;
1da177e4
LT
376};
377
378struct sighand_struct {
379 atomic_t count;
380 struct k_sigaction action[_NSIG];
381 spinlock_t siglock;
382};
383
0e464814 384struct pacct_struct {
f6ec29a4
KK
385 int ac_flag;
386 long ac_exitcode;
0e464814 387 unsigned long ac_mem;
77787bfb
KK
388 cputime_t ac_utime, ac_stime;
389 unsigned long ac_minflt, ac_majflt;
0e464814
KK
390};
391
1da177e4
LT
392/*
393 * NOTE! "signal_struct" does not have it's own
394 * locking, because a shared signal_struct always
395 * implies a shared sighand_struct, so locking
396 * sighand_struct is always a proper superset of
397 * the locking of signal_struct.
398 */
399struct signal_struct {
400 atomic_t count;
401 atomic_t live;
402
403 wait_queue_head_t wait_chldexit; /* for wait4() */
404
405 /* current thread group signal load-balancing target: */
36c8b586 406 struct task_struct *curr_target;
1da177e4
LT
407
408 /* shared signal handling: */
409 struct sigpending shared_pending;
410
411 /* thread group exit support */
412 int group_exit_code;
413 /* overloaded:
414 * - notify group_exit_task when ->count is equal to notify_count
415 * - everyone except group_exit_task is stopped during signal delivery
416 * of fatal signals, group_exit_task processes the signal.
417 */
418 struct task_struct *group_exit_task;
419 int notify_count;
420
421 /* thread group stop support, overloads group_exit_code too */
422 int group_stop_count;
423 unsigned int flags; /* see SIGNAL_* flags below */
424
425 /* POSIX.1b Interval Timers */
426 struct list_head posix_timers;
427
428 /* ITIMER_REAL timer for the process */
2ff678b8 429 struct hrtimer real_timer;
05cfb614 430 struct task_struct *tsk;
2ff678b8 431 ktime_t it_real_incr;
1da177e4
LT
432
433 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
434 cputime_t it_prof_expires, it_virt_expires;
435 cputime_t it_prof_incr, it_virt_incr;
436
437 /* job control IDs */
438 pid_t pgrp;
439 pid_t tty_old_pgrp;
1ec320af
CLG
440
441 union {
442 pid_t session __deprecated;
443 pid_t __session;
444 };
445
1da177e4
LT
446 /* boolean value for session group leader */
447 int leader;
448
449 struct tty_struct *tty; /* NULL if no tty */
450
451 /*
452 * Cumulative resource counters for dead threads in the group,
453 * and for reaped dead child processes forked by this group.
454 * Live threads maintain their own counters and add to these
455 * in __exit_signal, except for the group leader.
456 */
457 cputime_t utime, stime, cutime, cstime;
458 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
459 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
460
461 /*
462 * Cumulative ns of scheduled CPU time for dead threads in the
463 * group, not including a zombie group leader. (This only differs
464 * from jiffies_to_ns(utime + stime) if sched_clock uses something
465 * other than jiffies.)
466 */
467 unsigned long long sched_time;
468
469 /*
470 * We don't bother to synchronize most readers of this at all,
471 * because there is no reader checking a limit that actually needs
472 * to get both rlim_cur and rlim_max atomically, and either one
473 * alone is a single word that can safely be read normally.
474 * getrlimit/setrlimit use task_lock(current->group_leader) to
475 * protect this instead of the siglock, because they really
476 * have no need to disable irqs.
477 */
478 struct rlimit rlim[RLIM_NLIMITS];
479
480 struct list_head cpu_timers[3];
481
482 /* keep the process-shared keyrings here so that they do the right
483 * thing in threads created with CLONE_THREAD */
484#ifdef CONFIG_KEYS
485 struct key *session_keyring; /* keyring inherited over fork */
486 struct key *process_keyring; /* keyring private to this process */
487#endif
0e464814
KK
488#ifdef CONFIG_BSD_PROCESS_ACCT
489 struct pacct_struct pacct; /* per-process accounting information */
490#endif
ad4ecbcb 491#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
492 struct taskstats *stats;
493#endif
1da177e4
LT
494};
495
4866cde0
NP
496/* Context switch must be unlocked if interrupts are to be enabled */
497#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
498# define __ARCH_WANT_UNLOCKED_CTXSW
499#endif
500
1da177e4
LT
501/*
502 * Bits in flags field of signal_struct.
503 */
504#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
505#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
506#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
507#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
508
509
510/*
511 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
b0a9499c
IM
512 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
513 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
514 * values are inverted: lower p->prio value means higher priority.
1da177e4
LT
515 *
516 * The MAX_USER_RT_PRIO value allows the actual maximum
517 * RT priority to be separate from the value exported to
518 * user-space. This allows kernel threads to set their
519 * priority to a value higher than any user task. Note:
520 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
521 */
522
523#define MAX_USER_RT_PRIO 100
524#define MAX_RT_PRIO MAX_USER_RT_PRIO
525
526#define MAX_PRIO (MAX_RT_PRIO + 40)
527
b29739f9
IM
528#define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
529#define rt_task(p) rt_prio((p)->prio)
d425b274 530#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
57a6f51c
ON
531#define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
532#define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
1da177e4
LT
533
534/*
535 * Some day this will be a full-fledged user tracking system..
536 */
537struct user_struct {
538 atomic_t __count; /* reference count */
539 atomic_t processes; /* How many processes does this user have? */
540 atomic_t files; /* How many open files does this user have? */
541 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 542#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
543 atomic_t inotify_watches; /* How many inotify watches does this user have? */
544 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
545#endif
1da177e4
LT
546 /* protected by mq_lock */
547 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
548 unsigned long locked_shm; /* How many pages of mlocked shm ? */
549
550#ifdef CONFIG_KEYS
551 struct key *uid_keyring; /* UID specific keyring */
552 struct key *session_keyring; /* UID's default session keyring */
553#endif
554
555 /* Hash table maintenance information */
556 struct list_head uidhash_list;
557 uid_t uid;
558};
559
560extern struct user_struct *find_user(uid_t);
561
562extern struct user_struct root_user;
563#define INIT_USER (&root_user)
564
1da177e4
LT
565struct backing_dev_info;
566struct reclaim_state;
567
52f17b6c 568#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
569struct sched_info {
570 /* cumulative counters */
571 unsigned long cpu_time, /* time spent on the cpu */
572 run_delay, /* time spent waiting on a runqueue */
573 pcnt; /* # of timeslices run on this cpu */
574
575 /* timestamps */
576 unsigned long last_arrival, /* when we last ran on a cpu */
577 last_queued; /* when we were last queued to run */
578};
52f17b6c 579#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 580
52f17b6c 581#ifdef CONFIG_SCHEDSTATS
15ad7cdc 582extern const struct file_operations proc_schedstat_operations;
52f17b6c 583#endif /* CONFIG_SCHEDSTATS */
1da177e4 584
ca74e92b
SN
585#ifdef CONFIG_TASK_DELAY_ACCT
586struct task_delay_info {
587 spinlock_t lock;
588 unsigned int flags; /* Private per-task flags */
589
590 /* For each stat XXX, add following, aligned appropriately
591 *
592 * struct timespec XXX_start, XXX_end;
593 * u64 XXX_delay;
594 * u32 XXX_count;
595 *
596 * Atomicity of updates to XXX_delay, XXX_count protected by
597 * single lock above (split into XXX_lock if contention is an issue).
598 */
0ff92245
SN
599
600 /*
601 * XXX_count is incremented on every XXX operation, the delay
602 * associated with the operation is added to XXX_delay.
603 * XXX_delay contains the accumulated delay time in nanoseconds.
604 */
605 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
606 u64 blkio_delay; /* wait for sync block io completion */
607 u64 swapin_delay; /* wait for swapin block io completion */
608 u32 blkio_count; /* total count of the number of sync block */
609 /* io operations performed */
610 u32 swapin_count; /* total count of the number of swapin block */
611 /* io operations performed */
ca74e92b 612};
52f17b6c
CS
613#endif /* CONFIG_TASK_DELAY_ACCT */
614
615static inline int sched_info_on(void)
616{
617#ifdef CONFIG_SCHEDSTATS
618 return 1;
619#elif defined(CONFIG_TASK_DELAY_ACCT)
620 extern int delayacct_on;
621 return delayacct_on;
622#else
623 return 0;
ca74e92b 624#endif
52f17b6c 625}
ca74e92b 626
1da177e4
LT
627enum idle_type
628{
629 SCHED_IDLE,
630 NOT_IDLE,
631 NEWLY_IDLE,
632 MAX_IDLE_TYPES
633};
634
635/*
636 * sched-domains (multiprocessor balancing) declarations:
637 */
1da177e4
LT
638#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
639
2dd73a4f 640#ifdef CONFIG_SMP
1da177e4
LT
641#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
642#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
643#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
644#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
645#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
646#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
647#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
648#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
5c45bf27 649#define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
89c4710e 650#define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
08c183f3 651#define SD_SERIALIZE 1024 /* Only a single load balancing instance */
5c45bf27 652
89c4710e
SS
653#define BALANCE_FOR_MC_POWER \
654 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
655
656#define BALANCE_FOR_PKG_POWER \
657 ((sched_mc_power_savings || sched_smt_power_savings) ? \
658 SD_POWERSAVINGS_BALANCE : 0)
659
660#define test_sd_parent(sd, flag) ((sd->parent && \
661 (sd->parent->flags & flag)) ? 1 : 0)
5c45bf27 662
1da177e4
LT
663
664struct sched_group {
665 struct sched_group *next; /* Must be a circular list */
666 cpumask_t cpumask;
667
668 /*
669 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
670 * single CPU. This is read only (except for setup, hotplug CPU).
671 */
672 unsigned long cpu_power;
673};
674
675struct sched_domain {
676 /* These fields must be setup */
677 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 678 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4
LT
679 struct sched_group *groups; /* the balancing groups of the domain */
680 cpumask_t span; /* span of all CPUs in this domain */
681 unsigned long min_interval; /* Minimum balance interval ms */
682 unsigned long max_interval; /* Maximum balance interval ms */
683 unsigned int busy_factor; /* less balancing by factor if busy */
684 unsigned int imbalance_pct; /* No balance until over watermark */
685 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
686 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
687 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
7897986b
NP
688 unsigned int busy_idx;
689 unsigned int idle_idx;
690 unsigned int newidle_idx;
691 unsigned int wake_idx;
147cbb4b 692 unsigned int forkexec_idx;
1da177e4
LT
693 int flags; /* See SD_* */
694
695 /* Runtime fields. */
696 unsigned long last_balance; /* init to jiffies. units in jiffies */
697 unsigned int balance_interval; /* initialise to 1. units in ms. */
698 unsigned int nr_balance_failed; /* initialise to 0 */
699
700#ifdef CONFIG_SCHEDSTATS
701 /* load_balance() stats */
702 unsigned long lb_cnt[MAX_IDLE_TYPES];
703 unsigned long lb_failed[MAX_IDLE_TYPES];
704 unsigned long lb_balanced[MAX_IDLE_TYPES];
705 unsigned long lb_imbalance[MAX_IDLE_TYPES];
706 unsigned long lb_gained[MAX_IDLE_TYPES];
707 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
708 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
709 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
710
711 /* Active load balancing */
712 unsigned long alb_cnt;
713 unsigned long alb_failed;
714 unsigned long alb_pushed;
715
68767a0a
NP
716 /* SD_BALANCE_EXEC stats */
717 unsigned long sbe_cnt;
718 unsigned long sbe_balanced;
1da177e4
LT
719 unsigned long sbe_pushed;
720
68767a0a
NP
721 /* SD_BALANCE_FORK stats */
722 unsigned long sbf_cnt;
723 unsigned long sbf_balanced;
724 unsigned long sbf_pushed;
725
1da177e4
LT
726 /* try_to_wake_up() stats */
727 unsigned long ttwu_wake_remote;
728 unsigned long ttwu_move_affine;
729 unsigned long ttwu_move_balance;
730#endif
731};
732
51888ca2 733extern int partition_sched_domains(cpumask_t *partition1,
1a20ff27 734 cpumask_t *partition2);
198e2f18
AM
735
736/*
737 * Maximum cache size the migration-costs auto-tuning code will
738 * search from:
739 */
740extern unsigned int max_cache_size;
741
742#endif /* CONFIG_SMP */
1da177e4
LT
743
744
745struct io_context; /* See blkdev.h */
1da177e4
LT
746struct cpuset;
747
748#define NGROUPS_SMALL 32
749#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
750struct group_info {
751 int ngroups;
752 atomic_t usage;
753 gid_t small_block[NGROUPS_SMALL];
754 int nblocks;
755 gid_t *blocks[0];
756};
757
758/*
759 * get_group_info() must be called with the owning task locked (via task_lock())
760 * when task != current. The reason being that the vast majority of callers are
761 * looking at current->group_info, which can not be changed except by the
762 * current task. Changing current->group_info requires the task lock, too.
763 */
764#define get_group_info(group_info) do { \
765 atomic_inc(&(group_info)->usage); \
766} while (0)
767
768#define put_group_info(group_info) do { \
769 if (atomic_dec_and_test(&(group_info)->usage)) \
770 groups_free(group_info); \
771} while (0)
772
3e30148c
DH
773extern struct group_info *groups_alloc(int gidsetsize);
774extern void groups_free(struct group_info *group_info);
775extern int set_current_groups(struct group_info *group_info);
776extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
777/* access the groups "array" with this macro */
778#define GROUP_AT(gi, i) \
779 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
780
383f2835 781#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 782extern void prefetch_stack(struct task_struct *t);
383f2835
KC
783#else
784static inline void prefetch_stack(struct task_struct *t) { }
785#endif
1da177e4
LT
786
787struct audit_context; /* See audit.c */
788struct mempolicy;
b92ce558 789struct pipe_inode_info;
4865ecf1 790struct uts_namespace;
1da177e4 791
3dee386e
CK
792enum sleep_type {
793 SLEEP_NORMAL,
794 SLEEP_NONINTERACTIVE,
795 SLEEP_INTERACTIVE,
796 SLEEP_INTERRUPTED,
797};
798
70b97a7f
IM
799struct prio_array;
800
1da177e4
LT
801struct task_struct {
802 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
803 struct thread_info *thread_info;
804 atomic_t usage;
805 unsigned long flags; /* per process flags, defined below */
806 unsigned long ptrace;
807
36772092 808 int lock_depth; /* BKL lock depth */
1da177e4 809
2dd73a4f
PW
810#ifdef CONFIG_SMP
811#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
812 int oncpu;
813#endif
2dd73a4f
PW
814#endif
815 int load_weight; /* for niceness load balancing purposes */
b29739f9 816 int prio, static_prio, normal_prio;
1da177e4 817 struct list_head run_list;
70b97a7f 818 struct prio_array *array;
1da177e4 819
22e2c507 820 unsigned short ioprio;
6c5c9341 821#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 822 unsigned int btrace_seq;
6c5c9341 823#endif
1da177e4
LT
824 unsigned long sleep_avg;
825 unsigned long long timestamp, last_ran;
826 unsigned long long sched_time; /* sched_clock time spent running */
3dee386e 827 enum sleep_type sleep_type;
1da177e4
LT
828
829 unsigned long policy;
830 cpumask_t cpus_allowed;
831 unsigned int time_slice, first_time_slice;
832
52f17b6c 833#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
834 struct sched_info sched_info;
835#endif
836
837 struct list_head tasks;
838 /*
839 * ptrace_list/ptrace_children forms the list of my children
840 * that were stolen by a ptracer.
841 */
842 struct list_head ptrace_children;
843 struct list_head ptrace_list;
844
845 struct mm_struct *mm, *active_mm;
846
847/* task state */
848 struct linux_binfmt *binfmt;
849 long exit_state;
850 int exit_code, exit_signal;
851 int pdeath_signal; /* The signal sent when the parent dies */
852 /* ??? */
853 unsigned long personality;
854 unsigned did_exec:1;
855 pid_t pid;
856 pid_t tgid;
0a425405
AV
857
858#ifdef CONFIG_CC_STACKPROTECTOR
859 /* Canary value for the -fstack-protector gcc feature */
860 unsigned long stack_canary;
861#endif
1da177e4
LT
862 /*
863 * pointers to (original) parent process, youngest child, younger sibling,
864 * older sibling, respectively. (p->father can be replaced with
865 * p->parent->pid)
866 */
867 struct task_struct *real_parent; /* real parent process (when being debugged) */
868 struct task_struct *parent; /* parent process */
869 /*
870 * children/sibling forms the list of my children plus the
871 * tasks I'm ptracing.
872 */
873 struct list_head children; /* list of my children */
874 struct list_head sibling; /* linkage in my parent's children list */
875 struct task_struct *group_leader; /* threadgroup leader */
876
877 /* PID/PID hash table linkage. */
92476d7f 878 struct pid_link pids[PIDTYPE_MAX];
47e65328 879 struct list_head thread_group;
1da177e4
LT
880
881 struct completion *vfork_done; /* for vfork() */
882 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
883 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
884
885 unsigned long rt_priority;
886 cputime_t utime, stime;
887 unsigned long nvcsw, nivcsw; /* context switch counts */
888 struct timespec start_time;
889/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
890 unsigned long min_flt, maj_flt;
891
892 cputime_t it_prof_expires, it_virt_expires;
893 unsigned long long it_sched_expires;
894 struct list_head cpu_timers[3];
895
896/* process credentials */
897 uid_t uid,euid,suid,fsuid;
898 gid_t gid,egid,sgid,fsgid;
899 struct group_info *group_info;
900 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
901 unsigned keep_capabilities:1;
902 struct user_struct *user;
903#ifdef CONFIG_KEYS
b5f545c8 904 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 905 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 906 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4 907#endif
e07e23e1
AV
908 /*
909 * fpu_counter contains the number of consecutive context switches
910 * that the FPU is used. If this is over a threshold, the lazy fpu
911 * saving becomes unlazy to save the trap. This is an unsigned char
912 * so that after 256 times the counter wraps and the behavior turns
913 * lazy again; this to deal with bursty apps that only use FPU for
914 * a short time
915 */
916 unsigned char fpu_counter;
1da177e4 917 int oomkilladj; /* OOM kill score adjustment (bit shift). */
36772092
PBG
918 char comm[TASK_COMM_LEN]; /* executable name excluding path
919 - access with [gs]et_task_comm (which lock
920 it with task_lock())
921 - initialized normally by flush_old_exec */
1da177e4
LT
922/* file system info */
923 int link_count, total_link_count;
3d5b6fcc 924#ifdef CONFIG_SYSVIPC
1da177e4
LT
925/* ipc stuff */
926 struct sysv_sem sysvsem;
3d5b6fcc 927#endif
1da177e4
LT
928/* CPU-specific state of this task */
929 struct thread_struct thread;
930/* filesystem information */
931 struct fs_struct *fs;
932/* open file information */
933 struct files_struct *files;
1651e14e 934/* namespaces */
ab516013 935 struct nsproxy *nsproxy;
1da177e4
LT
936/* signal handlers */
937 struct signal_struct *signal;
938 struct sighand_struct *sighand;
939
940 sigset_t blocked, real_blocked;
150256d8 941 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
1da177e4
LT
942 struct sigpending pending;
943
944 unsigned long sas_ss_sp;
945 size_t sas_ss_size;
946 int (*notifier)(void *priv);
947 void *notifier_data;
948 sigset_t *notifier_mask;
949
950 void *security;
951 struct audit_context *audit_context;
952 seccomp_t seccomp;
953
954/* Thread group tracking */
955 u32 parent_exec_id;
956 u32 self_exec_id;
957/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
958 spinlock_t alloc_lock;
1da177e4 959
b29739f9
IM
960 /* Protection of the PI data structures: */
961 spinlock_t pi_lock;
962
23f78d4a
IM
963#ifdef CONFIG_RT_MUTEXES
964 /* PI waiters blocked on a rt_mutex held by this task */
965 struct plist_head pi_waiters;
966 /* Deadlock detection and priority inheritance handling */
967 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
968#endif
969
408894ee
IM
970#ifdef CONFIG_DEBUG_MUTEXES
971 /* mutex deadlock detection */
972 struct mutex_waiter *blocked_on;
973#endif
de30a2b3
IM
974#ifdef CONFIG_TRACE_IRQFLAGS
975 unsigned int irq_events;
976 int hardirqs_enabled;
977 unsigned long hardirq_enable_ip;
978 unsigned int hardirq_enable_event;
979 unsigned long hardirq_disable_ip;
980 unsigned int hardirq_disable_event;
981 int softirqs_enabled;
982 unsigned long softirq_disable_ip;
983 unsigned int softirq_disable_event;
984 unsigned long softirq_enable_ip;
985 unsigned int softirq_enable_event;
986 int hardirq_context;
987 int softirq_context;
988#endif
fbb9ce95
IM
989#ifdef CONFIG_LOCKDEP
990# define MAX_LOCK_DEPTH 30UL
991 u64 curr_chain_key;
992 int lockdep_depth;
993 struct held_lock held_locks[MAX_LOCK_DEPTH];
994 unsigned int lockdep_recursion;
995#endif
408894ee 996
1da177e4
LT
997/* journalling filesystem info */
998 void *journal_info;
999
1000/* VM state */
1001 struct reclaim_state *reclaim_state;
1002
1da177e4
LT
1003 struct backing_dev_info *backing_dev_info;
1004
1005 struct io_context *io_context;
1006
1007 unsigned long ptrace_message;
1008 siginfo_t *last_siginfo; /* For ptrace use. */
1009/*
1010 * current io wait handle: wait queue entry to use for io waits
1011 * If this thread is processing aio, this points at the waitqueue
1012 * inside the currently handled kiocb. It may be NULL (i.e. default
1013 * to a stack based synchronous wait) if its doing sync IO.
1014 */
1015 wait_queue_t *io_wait;
4b98d11b 1016#ifdef CONFIG_TASK_XACCT
1da177e4
LT
1017/* i/o counters(bytes read/written, #syscalls */
1018 u64 rchar, wchar, syscr, syscw;
4b98d11b 1019#endif
7c3ab738 1020 struct task_io_accounting ioac;
8f0ab514 1021#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1022 u64 acct_rss_mem1; /* accumulated rss usage */
1023 u64 acct_vm_mem1; /* accumulated virtual memory usage */
db5fed26 1024 cputime_t acct_stimexpd;/* stime since last update */
1da177e4
LT
1025#endif
1026#ifdef CONFIG_NUMA
1027 struct mempolicy *mempolicy;
1028 short il_next;
1029#endif
1030#ifdef CONFIG_CPUSETS
1031 struct cpuset *cpuset;
1032 nodemask_t mems_allowed;
1033 int cpuset_mems_generation;
825a46af 1034 int cpuset_mem_spread_rotor;
1da177e4 1035#endif
0771dfef 1036 struct robust_list_head __user *robust_list;
34f192c6
IM
1037#ifdef CONFIG_COMPAT
1038 struct compat_robust_list_head __user *compat_robust_list;
1039#endif
c87e2837
IM
1040 struct list_head pi_state_list;
1041 struct futex_pi_state *pi_state_cache;
0771dfef 1042
22e2c507 1043 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1044 struct rcu_head rcu;
b92ce558
JA
1045
1046 /*
1047 * cache last used pipe for splice
1048 */
1049 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1050#ifdef CONFIG_TASK_DELAY_ACCT
1051 struct task_delay_info *delays;
f4f154fd
AM
1052#endif
1053#ifdef CONFIG_FAULT_INJECTION
1054 int make_it_fail;
ca74e92b 1055#endif
1da177e4
LT
1056};
1057
1058static inline pid_t process_group(struct task_struct *tsk)
1059{
1060 return tsk->signal->pgrp;
1061}
1062
1ec320af
CLG
1063static inline pid_t signal_session(struct signal_struct *sig)
1064{
1065 return sig->__session;
1066}
1067
937949d9
CLG
1068static inline pid_t process_session(struct task_struct *tsk)
1069{
1ec320af
CLG
1070 return signal_session(tsk->signal);
1071}
1072
1073static inline void set_signal_session(struct signal_struct *sig, pid_t session)
1074{
1075 sig->__session = session;
937949d9
CLG
1076}
1077
22c935f4
EB
1078static inline struct pid *task_pid(struct task_struct *task)
1079{
1080 return task->pids[PIDTYPE_PID].pid;
1081}
1082
1083static inline struct pid *task_tgid(struct task_struct *task)
1084{
1085 return task->group_leader->pids[PIDTYPE_PID].pid;
1086}
1087
1088static inline struct pid *task_pgrp(struct task_struct *task)
1089{
1090 return task->group_leader->pids[PIDTYPE_PGID].pid;
1091}
1092
1093static inline struct pid *task_session(struct task_struct *task)
1094{
1095 return task->group_leader->pids[PIDTYPE_SID].pid;
1096}
1097
1da177e4
LT
1098/**
1099 * pid_alive - check that a task structure is not stale
1100 * @p: Task structure to be checked.
1101 *
1102 * Test if a process is not yet dead (at most zombie state)
1103 * If pid_alive fails, then pointers within the task structure
1104 * can be stale and must not be dereferenced.
1105 */
1106static inline int pid_alive(struct task_struct *p)
1107{
92476d7f 1108 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1109}
1110
f400e198 1111/**
3260259f
HK
1112 * is_init - check if a task structure is init
1113 * @tsk: Task structure to be checked.
1114 *
1115 * Check if a task structure is the first user space task the kernel created.
f400e198
SB
1116 */
1117static inline int is_init(struct task_struct *tsk)
1118{
1119 return tsk->pid == 1;
1120}
1121
9ec52099
CLG
1122extern struct pid *cad_pid;
1123
1da177e4 1124extern void free_task(struct task_struct *tsk);
1da177e4 1125#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1126
158d9ebd 1127extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1128
1129static inline void put_task_struct(struct task_struct *t)
1130{
1131 if (atomic_dec_and_test(&t->usage))
8c7904a0 1132 __put_task_struct(t);
e56d0903 1133}
1da177e4
LT
1134
1135/*
1136 * Per process flags
1137 */
1138#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1139 /* Not implemented yet, only for 486*/
1140#define PF_STARTING 0x00000002 /* being created */
1141#define PF_EXITING 0x00000004 /* getting shut down */
1da177e4
LT
1142#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1143#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1144#define PF_DUMPCORE 0x00000200 /* dumped core */
1145#define PF_SIGNALED 0x00000400 /* killed by a signal */
1146#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1147#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1148#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1da177e4
LT
1149#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1150#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1151#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1152#define PF_KSWAPD 0x00040000 /* I am kswapd */
1153#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1154#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
b31dc66a
JA
1155#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1156#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1157#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1158#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1159#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
c61afb18 1160#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1161#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1da177e4
LT
1162
1163/*
1164 * Only the _current_ task can read/write to tsk->flags, but other
1165 * tasks can access tsk->flags in readonly mode for example
1166 * with tsk_used_math (like during threaded core dumping).
1167 * There is however an exception to this rule during ptrace
1168 * or during fork: the ptracer task is allowed to write to the
1169 * child->flags of its traced child (same goes for fork, the parent
1170 * can write to the child->flags), because we're guaranteed the
1171 * child is not running and in turn not changing child->flags
1172 * at the same time the parent does it.
1173 */
1174#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1175#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1176#define clear_used_math() clear_stopped_child_used_math(current)
1177#define set_used_math() set_stopped_child_used_math(current)
1178#define conditional_stopped_child_used_math(condition, child) \
1179 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1180#define conditional_used_math(condition) \
1181 conditional_stopped_child_used_math(condition, current)
1182#define copy_to_stopped_child_used_math(child) \
1183 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1184/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1185#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1186#define used_math() tsk_used_math(current)
1187
1188#ifdef CONFIG_SMP
36c8b586 1189extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
1da177e4 1190#else
36c8b586 1191static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1da177e4 1192{
4098f991 1193 if (!cpu_isset(0, new_mask))
1da177e4
LT
1194 return -EINVAL;
1195 return 0;
1196}
1197#endif
1198
1199extern unsigned long long sched_clock(void);
36c8b586
IM
1200extern unsigned long long
1201current_sched_time(const struct task_struct *current_task);
1da177e4
LT
1202
1203/* sched_exec is called by processes performing an exec */
1204#ifdef CONFIG_SMP
1205extern void sched_exec(void);
1206#else
1207#define sched_exec() {}
1208#endif
1209
1210#ifdef CONFIG_HOTPLUG_CPU
1211extern void idle_task_exit(void);
1212#else
1213static inline void idle_task_exit(void) {}
1214#endif
1215
1216extern void sched_idle_next(void);
b29739f9
IM
1217
1218#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1219extern int rt_mutex_getprio(struct task_struct *p);
1220extern void rt_mutex_setprio(struct task_struct *p, int prio);
1221extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1222#else
36c8b586 1223static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1224{
1225 return p->normal_prio;
1226}
95e02ca9 1227# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1228#endif
1229
36c8b586
IM
1230extern void set_user_nice(struct task_struct *p, long nice);
1231extern int task_prio(const struct task_struct *p);
1232extern int task_nice(const struct task_struct *p);
1233extern int can_nice(const struct task_struct *p, const int nice);
1234extern int task_curr(const struct task_struct *p);
1da177e4
LT
1235extern int idle_cpu(int cpu);
1236extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
36c8b586
IM
1237extern struct task_struct *idle_task(int cpu);
1238extern struct task_struct *curr_task(int cpu);
1239extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1240
1241void yield(void);
1242
1243/*
1244 * The default (Linux) execution domain.
1245 */
1246extern struct exec_domain default_exec_domain;
1247
1248union thread_union {
1249 struct thread_info thread_info;
1250 unsigned long stack[THREAD_SIZE/sizeof(long)];
1251};
1252
1253#ifndef __HAVE_ARCH_KSTACK_END
1254static inline int kstack_end(void *addr)
1255{
1256 /* Reliable end of stack detection:
1257 * Some APM bios versions misalign the stack
1258 */
1259 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1260}
1261#endif
1262
1263extern union thread_union init_thread_union;
1264extern struct task_struct init_task;
1265
1266extern struct mm_struct init_mm;
1267
1268#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1269extern struct task_struct *find_task_by_pid_type(int type, int pid);
1da177e4
LT
1270extern void __set_special_pids(pid_t session, pid_t pgrp);
1271
1272/* per-UID process charging. */
1273extern struct user_struct * alloc_uid(uid_t);
1274static inline struct user_struct *get_uid(struct user_struct *u)
1275{
1276 atomic_inc(&u->__count);
1277 return u;
1278}
1279extern void free_uid(struct user_struct *);
1280extern void switch_uid(struct user_struct *);
1281
1282#include <asm/current.h>
1283
3171a030 1284extern void do_timer(unsigned long ticks);
1da177e4
LT
1285
1286extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1287extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1288extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1289 unsigned long clone_flags));
1290#ifdef CONFIG_SMP
1291 extern void kick_process(struct task_struct *tsk);
1292#else
1293 static inline void kick_process(struct task_struct *tsk) { }
1294#endif
36c8b586
IM
1295extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
1296extern void FASTCALL(sched_exit(struct task_struct * p));
1da177e4
LT
1297
1298extern int in_group_p(gid_t);
1299extern int in_egroup_p(gid_t);
1300
1301extern void proc_caches_init(void);
1302extern void flush_signals(struct task_struct *);
1303extern void flush_signal_handlers(struct task_struct *, int force_default);
1304extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1305
1306static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1307{
1308 unsigned long flags;
1309 int ret;
1310
1311 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1312 ret = dequeue_signal(tsk, mask, info);
1313 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1314
1315 return ret;
1316}
1317
1318extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1319 sigset_t *mask);
1320extern void unblock_all_signals(void);
1321extern void release_task(struct task_struct * p);
1322extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1323extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1324extern int force_sigsegv(int, struct task_struct *);
1325extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1
EB
1326extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1327extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1328extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 1329extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
1330extern int kill_pgrp(struct pid *pid, int sig, int priv);
1331extern int kill_pid(struct pid *pid, int sig, int priv);
1da177e4
LT
1332extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1333extern int kill_pg_info(int, struct siginfo *, pid_t);
1da177e4
LT
1334extern void do_notify_parent(struct task_struct *, int);
1335extern void force_sig(int, struct task_struct *);
1336extern void force_sig_specific(int, struct task_struct *);
1337extern int send_sig(int, struct task_struct *, int);
1338extern void zap_other_threads(struct task_struct *p);
1339extern int kill_pg(pid_t, int, int);
1da177e4
LT
1340extern int kill_proc(pid_t, int, int);
1341extern struct sigqueue *sigqueue_alloc(void);
1342extern void sigqueue_free(struct sigqueue *);
1343extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1344extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
9ac95f2f 1345extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
1346extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1347
9ec52099
CLG
1348static inline int kill_cad_pid(int sig, int priv)
1349{
1350 return kill_pid(cad_pid, sig, priv);
1351}
1352
1da177e4
LT
1353/* These can be the second arg to send_sig_info/send_group_sig_info. */
1354#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1355#define SEND_SIG_PRIV ((struct siginfo *) 1)
1356#define SEND_SIG_FORCED ((struct siginfo *) 2)
1357
621d3121
ON
1358static inline int is_si_special(const struct siginfo *info)
1359{
1360 return info <= SEND_SIG_FORCED;
1361}
1362
1da177e4
LT
1363/* True if we are on the alternate signal stack. */
1364
1365static inline int on_sig_stack(unsigned long sp)
1366{
1367 return (sp - current->sas_ss_sp < current->sas_ss_size);
1368}
1369
1370static inline int sas_ss_flags(unsigned long sp)
1371{
1372 return (current->sas_ss_size == 0 ? SS_DISABLE
1373 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1374}
1375
1da177e4
LT
1376/*
1377 * Routines for handling mm_structs
1378 */
1379extern struct mm_struct * mm_alloc(void);
1380
1381/* mmdrop drops the mm and the page tables */
1382extern void FASTCALL(__mmdrop(struct mm_struct *));
1383static inline void mmdrop(struct mm_struct * mm)
1384{
1385 if (atomic_dec_and_test(&mm->mm_count))
1386 __mmdrop(mm);
1387}
1388
1389/* mmput gets rid of the mappings and all user-space */
1390extern void mmput(struct mm_struct *);
1391/* Grab a reference to a task's mm, if it is not already going away */
1392extern struct mm_struct *get_task_mm(struct task_struct *task);
1393/* Remove the current tasks stale references to the old mm_struct */
1394extern void mm_release(struct task_struct *, struct mm_struct *);
1395
1396extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1397extern void flush_thread(void);
1398extern void exit_thread(void);
1399
1da177e4 1400extern void exit_files(struct task_struct *);
6b3934ef 1401extern void __cleanup_signal(struct signal_struct *);
a7e5328a 1402extern void __cleanup_sighand(struct sighand_struct *);
1da177e4
LT
1403extern void exit_itimers(struct signal_struct *);
1404
1405extern NORET_TYPE void do_group_exit(int);
1406
1da177e4
LT
1407extern void daemonize(const char *, ...);
1408extern int allow_signal(int);
1409extern int disallow_signal(int);
1da177e4
LT
1410
1411extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1412extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 1413struct task_struct *fork_idle(int);
1da177e4
LT
1414
1415extern void set_task_comm(struct task_struct *tsk, char *from);
1416extern void get_task_comm(char *to, struct task_struct *tsk);
1417
1418#ifdef CONFIG_SMP
36c8b586 1419extern void wait_task_inactive(struct task_struct * p);
1da177e4
LT
1420#else
1421#define wait_task_inactive(p) do { } while (0)
1422#endif
1423
1424#define remove_parent(p) list_del_init(&(p)->sibling)
8fafabd8 1425#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1da177e4 1426
5e85d4ab 1427#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1da177e4
LT
1428
1429#define for_each_process(p) \
1430 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1431
1432/*
1433 * Careful: do_each_thread/while_each_thread is a double loop so
1434 * 'break' will not work as expected - use goto instead.
1435 */
1436#define do_each_thread(g, t) \
1437 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1438
1439#define while_each_thread(g, t) \
1440 while ((t = next_thread(t)) != g)
1441
de12a787
EB
1442/* de_thread depends on thread_group_leader not being a pid based check */
1443#define thread_group_leader(p) (p == p->group_leader)
1da177e4 1444
0804ef4b
EB
1445/* Do to the insanities of de_thread it is possible for a process
1446 * to have the pid of the thread group leader without actually being
1447 * the thread group leader. For iteration through the pids in proc
1448 * all we care about is that we have a task with the appropriate
1449 * pid, we don't actually care if we have the right task.
1450 */
1451static inline int has_group_leader_pid(struct task_struct *p)
1452{
1453 return p->pid == p->tgid;
1454}
1455
36c8b586 1456static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328
ON
1457{
1458 return list_entry(rcu_dereference(p->thread_group.next),
36c8b586 1459 struct task_struct, thread_group);
47e65328
ON
1460}
1461
36c8b586 1462static inline int thread_group_empty(struct task_struct *p)
1da177e4 1463{
47e65328 1464 return list_empty(&p->thread_group);
1da177e4
LT
1465}
1466
1467#define delay_group_leader(p) \
1468 (thread_group_leader(p) && !thread_group_empty(p))
1469
1da177e4 1470/*
260ea101 1471 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 1472 * subscriptions and synchronises with wait4(). Also used in procfs. Also
053199ed 1473 * pins the final release of task.io_context. Also protects ->cpuset.
1da177e4
LT
1474 *
1475 * Nests both inside and outside of read_lock(&tasklist_lock).
1476 * It must not be nested with write_lock_irq(&tasklist_lock),
1477 * neither inside nor outside.
1478 */
1479static inline void task_lock(struct task_struct *p)
1480{
1481 spin_lock(&p->alloc_lock);
1482}
1483
1484static inline void task_unlock(struct task_struct *p)
1485{
1486 spin_unlock(&p->alloc_lock);
1487}
1488
f63ee72e
ON
1489extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1490 unsigned long *flags);
1491
1492static inline void unlock_task_sighand(struct task_struct *tsk,
1493 unsigned long *flags)
1494{
1495 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1496}
1497
f037360f
AV
1498#ifndef __HAVE_THREAD_FUNCTIONS
1499
a1261f54 1500#define task_thread_info(task) (task)->thread_info
9fc65876 1501#define task_stack_page(task) ((void*)((task)->thread_info))
a1261f54 1502
10ebffde
AV
1503static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1504{
1505 *task_thread_info(p) = *task_thread_info(org);
1506 task_thread_info(p)->task = p;
1507}
1508
1509static inline unsigned long *end_of_stack(struct task_struct *p)
1510{
1511 return (unsigned long *)(p->thread_info + 1);
1512}
1513
f037360f
AV
1514#endif
1515
1da177e4
LT
1516/* set thread flags in other task's structures
1517 * - see asm/thread_info.h for TIF_xxxx flags available
1518 */
1519static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1520{
a1261f54 1521 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1522}
1523
1524static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1525{
a1261f54 1526 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1527}
1528
1529static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1530{
a1261f54 1531 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1532}
1533
1534static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1535{
a1261f54 1536 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1537}
1538
1539static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1540{
a1261f54 1541 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1542}
1543
1544static inline void set_tsk_need_resched(struct task_struct *tsk)
1545{
1546 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1547}
1548
1549static inline void clear_tsk_need_resched(struct task_struct *tsk)
1550{
1551 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1552}
1553
1554static inline int signal_pending(struct task_struct *p)
1555{
1556 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1557}
1558
1559static inline int need_resched(void)
1560{
1561 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1562}
1563
1564/*
1565 * cond_resched() and cond_resched_lock(): latency reduction via
1566 * explicit rescheduling in places that are safe. The return
1567 * value indicates whether a reschedule was done in fact.
1568 * cond_resched_lock() will drop the spinlock before scheduling,
1569 * cond_resched_softirq() will enable bhs before scheduling.
1570 */
1571extern int cond_resched(void);
1572extern int cond_resched_lock(spinlock_t * lock);
1573extern int cond_resched_softirq(void);
1574
1575/*
1576 * Does a critical section need to be broken due to another
1577 * task waiting?:
1578 */
1579#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1580# define need_lockbreak(lock) ((lock)->break_lock)
1581#else
1582# define need_lockbreak(lock) 0
1583#endif
1584
1585/*
1586 * Does a critical section need to be broken due to another
1587 * task waiting or preemption being signalled:
1588 */
1589static inline int lock_need_resched(spinlock_t *lock)
1590{
1591 if (need_lockbreak(lock) || need_resched())
1592 return 1;
1593 return 0;
1594}
1595
1596/* Reevaluate whether the task has signals pending delivery.
1597 This is required every time the blocked sigset_t changes.
1598 callers must hold sighand->siglock. */
1599
1600extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1601extern void recalc_sigpending(void);
1602
1603extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1604
1605/*
1606 * Wrappers for p->thread_info->cpu access. No-op on UP.
1607 */
1608#ifdef CONFIG_SMP
1609
1610static inline unsigned int task_cpu(const struct task_struct *p)
1611{
a1261f54 1612 return task_thread_info(p)->cpu;
1da177e4
LT
1613}
1614
1615static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1616{
a1261f54 1617 task_thread_info(p)->cpu = cpu;
1da177e4
LT
1618}
1619
1620#else
1621
1622static inline unsigned int task_cpu(const struct task_struct *p)
1623{
1624 return 0;
1625}
1626
1627static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1628{
1629}
1630
1631#endif /* CONFIG_SMP */
1632
1633#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1634extern void arch_pick_mmap_layout(struct mm_struct *mm);
1635#else
1636static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1637{
1638 mm->mmap_base = TASK_UNMAPPED_BASE;
1639 mm->get_unmapped_area = arch_get_unmapped_area;
1640 mm->unmap_area = arch_unmap_area;
1641}
1642#endif
1643
1644extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1645extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1646
5c45bf27
SS
1647#include <linux/sysdev.h>
1648extern int sched_mc_power_savings, sched_smt_power_savings;
1649extern struct sysdev_attribute attr_sched_mc_power_savings, attr_sched_smt_power_savings;
1650extern int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls);
1651
1da177e4
LT
1652extern void normalize_rt_tasks(void);
1653
4b98d11b
AD
1654#ifdef CONFIG_TASK_XACCT
1655static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
1656{
1657 tsk->rchar += amt;
1658}
1659
1660static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
1661{
1662 tsk->wchar += amt;
1663}
1664
1665static inline void inc_syscr(struct task_struct *tsk)
1666{
1667 tsk->syscr++;
1668}
1669
1670static inline void inc_syscw(struct task_struct *tsk)
1671{
1672 tsk->syscw++;
1673}
1674#else
1675static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
1676{
1677}
1678
1679static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
1680{
1681}
1682
1683static inline void inc_syscr(struct task_struct *tsk)
1684{
1685}
1686
1687static inline void inc_syscw(struct task_struct *tsk)
1688{
1689}
1690#endif
1691
1da177e4
LT
1692#endif /* __KERNEL__ */
1693
1694#endif