]> bbs.cooldavid.org Git - net-next-2.6.git/blame - include/linux/sched.h
signals: unify send_sigqueue/send_group_sigqueue completely
[net-next-2.6.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4/*
5 * cloning flags:
6 */
7#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8#define CLONE_VM 0x00000100 /* set if VM shared between processes */
9#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15#define CLONE_THREAD 0x00010000 /* Same thread group? */
16#define CLONE_NEWNS 0x00020000 /* New namespace group? */
17#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
24#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
071df104 25#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 26#define CLONE_NEWIPC 0x08000000 /* New ipcs */
77ec739d 27#define CLONE_NEWUSER 0x10000000 /* New user namespace */
30e49c26 28#define CLONE_NEWPID 0x20000000 /* New pid namespace */
169e3674 29#define CLONE_NEWNET 0x40000000 /* New network namespace */
fadad878 30#define CLONE_IO 0x80000000 /* Clone io context */
b7b3c76a
DW
31
32/*
33 * Scheduling policies
34 */
35#define SCHED_NORMAL 0
36#define SCHED_FIFO 1
37#define SCHED_RR 2
38#define SCHED_BATCH 3
0e6aca43
IM
39/* SCHED_ISO: reserved but not implemented yet */
40#define SCHED_IDLE 5
b7b3c76a 41
a3b6714e 42#ifdef __KERNEL__
b7b3c76a
DW
43
44struct sched_param {
45 int sched_priority;
46};
47
1da177e4
LT
48#include <asm/param.h> /* for HZ */
49
1da177e4
LT
50#include <linux/capability.h>
51#include <linux/threads.h>
52#include <linux/kernel.h>
53#include <linux/types.h>
54#include <linux/timex.h>
55#include <linux/jiffies.h>
56#include <linux/rbtree.h>
57#include <linux/thread_info.h>
58#include <linux/cpumask.h>
59#include <linux/errno.h>
60#include <linux/nodemask.h>
c92ff1bd 61#include <linux/mm_types.h>
1da177e4
LT
62
63#include <asm/system.h>
1da177e4
LT
64#include <asm/page.h>
65#include <asm/ptrace.h>
1da177e4
LT
66#include <asm/cputime.h>
67
68#include <linux/smp.h>
69#include <linux/sem.h>
70#include <linux/signal.h>
1da177e4
LT
71#include <linux/fs_struct.h>
72#include <linux/compiler.h>
73#include <linux/completion.h>
74#include <linux/pid.h>
75#include <linux/percpu.h>
76#include <linux/topology.h>
3e26c149 77#include <linux/proportions.h>
1da177e4 78#include <linux/seccomp.h>
e56d0903 79#include <linux/rcupdate.h>
23f78d4a 80#include <linux/rtmutex.h>
1da177e4 81
a3b6714e
DW
82#include <linux/time.h>
83#include <linux/param.h>
84#include <linux/resource.h>
85#include <linux/timer.h>
86#include <linux/hrtimer.h>
7c3ab738 87#include <linux/task_io_accounting.h>
5cb350ba 88#include <linux/kobject.h>
9745512c 89#include <linux/latencytop.h>
a3b6714e
DW
90
91#include <asm/processor.h>
36d57ac4 92
78fb7466 93struct mem_cgroup;
1da177e4 94struct exec_domain;
c87e2837 95struct futex_pi_state;
286100a6 96struct robust_list_head;
d89d8796 97struct bio;
1da177e4 98
1da177e4
LT
99/*
100 * List of flags we want to share for kernel threads,
101 * if only because they are not used by them anyway.
102 */
103#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
104
105/*
106 * These are the constant used to fake the fixed-point load-average
107 * counting. Some notes:
108 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
109 * a load-average precision of 10 bits integer + 11 bits fractional
110 * - if you want to count load-averages more often, you need more
111 * precision, or rounding will get you. With 2-second counting freq,
112 * the EXP_n values would be 1981, 2034 and 2043 if still using only
113 * 11 bit fractions.
114 */
115extern unsigned long avenrun[]; /* Load averages */
116
117#define FSHIFT 11 /* nr of bits of precision */
118#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 119#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
120#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
121#define EXP_5 2014 /* 1/exp(5sec/5min) */
122#define EXP_15 2037 /* 1/exp(5sec/15min) */
123
124#define CALC_LOAD(load,exp,n) \
125 load *= exp; \
126 load += n*(FIXED_1-exp); \
127 load >>= FSHIFT;
128
129extern unsigned long total_forks;
130extern int nr_threads;
1da177e4
LT
131DECLARE_PER_CPU(unsigned long, process_counts);
132extern int nr_processes(void);
133extern unsigned long nr_running(void);
134extern unsigned long nr_uninterruptible(void);
db1b1fef 135extern unsigned long nr_active(void);
1da177e4 136extern unsigned long nr_iowait(void);
2dd73a4f 137extern unsigned long weighted_cpuload(const int cpu);
1da177e4 138
43ae34cb
IM
139struct seq_file;
140struct cfs_rq;
4cf86d77 141struct task_group;
43ae34cb
IM
142#ifdef CONFIG_SCHED_DEBUG
143extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
144extern void proc_sched_set_task(struct task_struct *p);
145extern void
5cef9eca 146print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
147#else
148static inline void
149proc_sched_show_task(struct task_struct *p, struct seq_file *m)
150{
151}
152static inline void proc_sched_set_task(struct task_struct *p)
153{
154}
155static inline void
5cef9eca 156print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
157{
158}
159#endif
1da177e4 160
4a8342d2
LT
161/*
162 * Task state bitmask. NOTE! These bits are also
163 * encoded in fs/proc/array.c: get_task_state().
164 *
165 * We have two separate sets of flags: task->state
166 * is about runnability, while task->exit_state are
167 * about the task exiting. Confusing, but this way
168 * modifying one set can't modify the other one by
169 * mistake.
170 */
1da177e4
LT
171#define TASK_RUNNING 0
172#define TASK_INTERRUPTIBLE 1
173#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
174#define __TASK_STOPPED 4
175#define __TASK_TRACED 8
4a8342d2
LT
176/* in tsk->exit_state */
177#define EXIT_ZOMBIE 16
178#define EXIT_DEAD 32
179/* in tsk->state again */
af927232 180#define TASK_DEAD 64
f021a3c2
MW
181#define TASK_WAKEKILL 128
182
183/* Convenience macros for the sake of set_task_state */
184#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
185#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
186#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 187
92a1f4bc
MW
188/* Convenience macros for the sake of wake_up */
189#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 190#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
191
192/* get_task_state() */
193#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
194 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
195 __TASK_TRACED)
92a1f4bc 196
f021a3c2
MW
197#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
198#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 199#define task_is_stopped_or_traced(task) \
f021a3c2 200 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc
MW
201#define task_contributes_to_load(task) \
202 ((task->state & TASK_UNINTERRUPTIBLE) != 0)
1da177e4
LT
203
204#define __set_task_state(tsk, state_value) \
205 do { (tsk)->state = (state_value); } while (0)
206#define set_task_state(tsk, state_value) \
207 set_mb((tsk)->state, (state_value))
208
498d0c57
AM
209/*
210 * set_current_state() includes a barrier so that the write of current->state
211 * is correctly serialised wrt the caller's subsequent test of whether to
212 * actually sleep:
213 *
214 * set_current_state(TASK_UNINTERRUPTIBLE);
215 * if (do_i_need_to_sleep())
216 * schedule();
217 *
218 * If the caller does not need such serialisation then use __set_current_state()
219 */
1da177e4
LT
220#define __set_current_state(state_value) \
221 do { current->state = (state_value); } while (0)
222#define set_current_state(state_value) \
223 set_mb(current->state, (state_value))
224
225/* Task command name length */
226#define TASK_COMM_LEN 16
227
1da177e4
LT
228#include <linux/spinlock.h>
229
230/*
231 * This serializes "schedule()" and also protects
232 * the run-queue from deletions/modifications (but
233 * _adding_ to the beginning of the run-queue has
234 * a separate lock).
235 */
236extern rwlock_t tasklist_lock;
237extern spinlock_t mmlist_lock;
238
36c8b586 239struct task_struct;
1da177e4
LT
240
241extern void sched_init(void);
242extern void sched_init_smp(void);
2d07b255 243extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 244extern void init_idle(struct task_struct *idle, int cpu);
1df21055 245extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4
LT
246
247extern cpumask_t nohz_cpu_mask;
46cb4b7c
SS
248#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
249extern int select_nohz_load_balancer(int cpu);
250#else
251static inline int select_nohz_load_balancer(int cpu)
252{
253 return 0;
254}
255#endif
1da177e4 256
48d5e258
PZ
257extern unsigned long rt_needs_cpu(int cpu);
258
e59e2ae2 259/*
39bc89fd 260 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
261 */
262extern void show_state_filter(unsigned long state_filter);
263
264static inline void show_state(void)
265{
39bc89fd 266 show_state_filter(0);
e59e2ae2
IM
267}
268
1da177e4
LT
269extern void show_regs(struct pt_regs *);
270
271/*
272 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
273 * task), SP is the stack pointer of the first frame that should be shown in the back
274 * trace (or NULL if the entire call-chain of the task should be shown).
275 */
276extern void show_stack(struct task_struct *task, unsigned long *sp);
277
278void io_schedule(void);
279long io_schedule_timeout(long timeout);
280
281extern void cpu_init (void);
282extern void trap_init(void);
fa13a5a1 283extern void account_process_tick(struct task_struct *task, int user);
1da177e4
LT
284extern void update_process_times(int user);
285extern void scheduler_tick(void);
8f4d37ec 286extern void hrtick_resched(void);
1da177e4 287
82a1fcb9
IM
288extern void sched_show_task(struct task_struct *p);
289
8446f1d3 290#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 291extern void softlockup_tick(void);
8446f1d3
IM
292extern void spawn_softlockup_task(void);
293extern void touch_softlockup_watchdog(void);
04c9167f 294extern void touch_all_softlockup_watchdogs(void);
90739081 295extern unsigned long softlockup_thresh;
82a1fcb9
IM
296extern unsigned long sysctl_hung_task_check_count;
297extern unsigned long sysctl_hung_task_timeout_secs;
90739081 298extern unsigned long sysctl_hung_task_warnings;
8446f1d3 299#else
6687a97d 300static inline void softlockup_tick(void)
8446f1d3
IM
301{
302}
303static inline void spawn_softlockup_task(void)
304{
305}
306static inline void touch_softlockup_watchdog(void)
307{
308}
04c9167f
JF
309static inline void touch_all_softlockup_watchdogs(void)
310{
311}
8446f1d3
IM
312#endif
313
314
1da177e4
LT
315/* Attach to any functions which should be ignored in wchan output. */
316#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
317
318/* Linker adds these: start and end of __sched functions */
319extern char __sched_text_start[], __sched_text_end[];
320
1da177e4
LT
321/* Is this address in the __sched functions? */
322extern int in_sched_functions(unsigned long addr);
323
324#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 325extern signed long schedule_timeout(signed long timeout);
64ed93a2 326extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 327extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 328extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
329asmlinkage void schedule(void);
330
ab516013 331struct nsproxy;
acce292c 332struct user_namespace;
1da177e4
LT
333
334/* Maximum number of active map areas.. This is a random (large) number */
335#define DEFAULT_MAX_MAP_COUNT 65536
336
337extern int sysctl_max_map_count;
338
339#include <linux/aio.h>
340
341extern unsigned long
342arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
343 unsigned long, unsigned long);
344extern unsigned long
345arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
346 unsigned long len, unsigned long pgoff,
347 unsigned long flags);
1363c3cd
WW
348extern void arch_unmap_area(struct mm_struct *, unsigned long);
349extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 350
f412ac08
HD
351#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
352/*
353 * The mm counters are not protected by its page_table_lock,
354 * so must be incremented atomically.
355 */
d3cb4871
CL
356#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
357#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
358#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
359#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
360#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
f412ac08
HD
361
362#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
363/*
364 * The mm counters are protected by its page_table_lock,
365 * so can be incremented directly.
366 */
1da177e4
LT
367#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
368#define get_mm_counter(mm, member) ((mm)->_##member)
369#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
370#define inc_mm_counter(mm, member) (mm)->_##member++
371#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
372
373#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 374
f412ac08
HD
375#define get_mm_rss(mm) \
376 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
377#define update_hiwater_rss(mm) do { \
378 unsigned long _rss = get_mm_rss(mm); \
379 if ((mm)->hiwater_rss < _rss) \
380 (mm)->hiwater_rss = _rss; \
381} while (0)
382#define update_hiwater_vm(mm) do { \
383 if ((mm)->hiwater_vm < (mm)->total_vm) \
384 (mm)->hiwater_vm = (mm)->total_vm; \
385} while (0)
386
6c5d5238
KH
387extern void set_dumpable(struct mm_struct *mm, int value);
388extern int get_dumpable(struct mm_struct *mm);
389
390/* mm flags */
3cb4a0bb 391/* dumpable bits */
6c5d5238
KH
392#define MMF_DUMPABLE 0 /* core dump is permitted */
393#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
3cb4a0bb
KH
394#define MMF_DUMPABLE_BITS 2
395
396/* coredump filter bits */
397#define MMF_DUMP_ANON_PRIVATE 2
398#define MMF_DUMP_ANON_SHARED 3
399#define MMF_DUMP_MAPPED_PRIVATE 4
400#define MMF_DUMP_MAPPED_SHARED 5
82df3973 401#define MMF_DUMP_ELF_HEADERS 6
3cb4a0bb 402#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
82df3973 403#define MMF_DUMP_FILTER_BITS 5
3cb4a0bb
KH
404#define MMF_DUMP_FILTER_MASK \
405 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
406#define MMF_DUMP_FILTER_DEFAULT \
407 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED))
6c5d5238 408
1da177e4
LT
409struct sighand_struct {
410 atomic_t count;
411 struct k_sigaction action[_NSIG];
412 spinlock_t siglock;
b8fceee1 413 wait_queue_head_t signalfd_wqh;
1da177e4
LT
414};
415
0e464814 416struct pacct_struct {
f6ec29a4
KK
417 int ac_flag;
418 long ac_exitcode;
0e464814 419 unsigned long ac_mem;
77787bfb
KK
420 cputime_t ac_utime, ac_stime;
421 unsigned long ac_minflt, ac_majflt;
0e464814
KK
422};
423
1da177e4
LT
424/*
425 * NOTE! "signal_struct" does not have it's own
426 * locking, because a shared signal_struct always
427 * implies a shared sighand_struct, so locking
428 * sighand_struct is always a proper superset of
429 * the locking of signal_struct.
430 */
431struct signal_struct {
432 atomic_t count;
433 atomic_t live;
434
435 wait_queue_head_t wait_chldexit; /* for wait4() */
436
437 /* current thread group signal load-balancing target: */
36c8b586 438 struct task_struct *curr_target;
1da177e4
LT
439
440 /* shared signal handling: */
441 struct sigpending shared_pending;
442
443 /* thread group exit support */
444 int group_exit_code;
445 /* overloaded:
446 * - notify group_exit_task when ->count is equal to notify_count
447 * - everyone except group_exit_task is stopped during signal delivery
448 * of fatal signals, group_exit_task processes the signal.
449 */
450 struct task_struct *group_exit_task;
451 int notify_count;
452
453 /* thread group stop support, overloads group_exit_code too */
454 int group_stop_count;
455 unsigned int flags; /* see SIGNAL_* flags below */
456
457 /* POSIX.1b Interval Timers */
458 struct list_head posix_timers;
459
460 /* ITIMER_REAL timer for the process */
2ff678b8 461 struct hrtimer real_timer;
fea9d175 462 struct pid *leader_pid;
2ff678b8 463 ktime_t it_real_incr;
1da177e4
LT
464
465 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
466 cputime_t it_prof_expires, it_virt_expires;
467 cputime_t it_prof_incr, it_virt_incr;
468
469 /* job control IDs */
9a2e7057
PE
470
471 /*
472 * pgrp and session fields are deprecated.
473 * use the task_session_Xnr and task_pgrp_Xnr routines below
474 */
475
476 union {
477 pid_t pgrp __deprecated;
478 pid_t __pgrp;
479 };
480
ab521dc0 481 struct pid *tty_old_pgrp;
1ec320af
CLG
482
483 union {
484 pid_t session __deprecated;
485 pid_t __session;
486 };
487
1da177e4
LT
488 /* boolean value for session group leader */
489 int leader;
490
491 struct tty_struct *tty; /* NULL if no tty */
492
493 /*
494 * Cumulative resource counters for dead threads in the group,
495 * and for reaped dead child processes forked by this group.
496 * Live threads maintain their own counters and add to these
497 * in __exit_signal, except for the group leader.
498 */
499 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
500 cputime_t gtime;
501 cputime_t cgtime;
1da177e4
LT
502 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
503 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 504 unsigned long inblock, oublock, cinblock, coublock;
1da177e4
LT
505
506 /*
507 * Cumulative ns of scheduled CPU time for dead threads in the
508 * group, not including a zombie group leader. (This only differs
509 * from jiffies_to_ns(utime + stime) if sched_clock uses something
510 * other than jiffies.)
511 */
41b86e9c 512 unsigned long long sum_sched_runtime;
1da177e4
LT
513
514 /*
515 * We don't bother to synchronize most readers of this at all,
516 * because there is no reader checking a limit that actually needs
517 * to get both rlim_cur and rlim_max atomically, and either one
518 * alone is a single word that can safely be read normally.
519 * getrlimit/setrlimit use task_lock(current->group_leader) to
520 * protect this instead of the siglock, because they really
521 * have no need to disable irqs.
522 */
523 struct rlimit rlim[RLIM_NLIMITS];
524
525 struct list_head cpu_timers[3];
526
527 /* keep the process-shared keyrings here so that they do the right
528 * thing in threads created with CLONE_THREAD */
529#ifdef CONFIG_KEYS
530 struct key *session_keyring; /* keyring inherited over fork */
531 struct key *process_keyring; /* keyring private to this process */
532#endif
0e464814
KK
533#ifdef CONFIG_BSD_PROCESS_ACCT
534 struct pacct_struct pacct; /* per-process accounting information */
535#endif
ad4ecbcb 536#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
537 struct taskstats *stats;
538#endif
522ed776
MT
539#ifdef CONFIG_AUDIT
540 unsigned audit_tty;
541 struct tty_audit_buf *tty_audit_buf;
542#endif
1da177e4
LT
543};
544
4866cde0
NP
545/* Context switch must be unlocked if interrupts are to be enabled */
546#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
547# define __ARCH_WANT_UNLOCKED_CTXSW
548#endif
549
1da177e4
LT
550/*
551 * Bits in flags field of signal_struct.
552 */
553#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
554#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
555#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
556#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
557/*
558 * Pending notifications to parent.
559 */
560#define SIGNAL_CLD_STOPPED 0x00000010
561#define SIGNAL_CLD_CONTINUED 0x00000020
562#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 563
ed5d2cac
ON
564/* If true, all threads except ->group_exit_task have pending SIGKILL */
565static inline int signal_group_exit(const struct signal_struct *sig)
566{
567 return (sig->flags & SIGNAL_GROUP_EXIT) ||
568 (sig->group_exit_task != NULL);
569}
570
1da177e4
LT
571/*
572 * Some day this will be a full-fledged user tracking system..
573 */
574struct user_struct {
575 atomic_t __count; /* reference count */
576 atomic_t processes; /* How many processes does this user have? */
577 atomic_t files; /* How many open files does this user have? */
578 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 579#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
580 atomic_t inotify_watches; /* How many inotify watches does this user have? */
581 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
582#endif
970a8645 583#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
584 /* protected by mq_lock */
585 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 586#endif
1da177e4
LT
587 unsigned long locked_shm; /* How many pages of mlocked shm ? */
588
589#ifdef CONFIG_KEYS
590 struct key *uid_keyring; /* UID specific keyring */
591 struct key *session_keyring; /* UID's default session keyring */
592#endif
593
594 /* Hash table maintenance information */
735de223 595 struct hlist_node uidhash_node;
1da177e4 596 uid_t uid;
24e377a8 597
052f1dc7 598#ifdef CONFIG_USER_SCHED
4cf86d77 599 struct task_group *tg;
b1a8c172 600#ifdef CONFIG_SYSFS
eb41d946 601 struct kobject kobj;
5cb350ba 602 struct work_struct work;
24e377a8 603#endif
b1a8c172 604#endif
1da177e4
LT
605};
606
eb41d946 607extern int uids_sysfs_init(void);
5cb350ba 608
1da177e4
LT
609extern struct user_struct *find_user(uid_t);
610
611extern struct user_struct root_user;
612#define INIT_USER (&root_user)
613
1da177e4
LT
614struct backing_dev_info;
615struct reclaim_state;
616
52f17b6c 617#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
618struct sched_info {
619 /* cumulative counters */
2d72376b 620 unsigned long pcount; /* # of times run on this cpu */
172ba844
BS
621 unsigned long long cpu_time, /* time spent on the cpu */
622 run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
623
624 /* timestamps */
172ba844
BS
625 unsigned long long last_arrival,/* when we last ran on a cpu */
626 last_queued; /* when we were last queued to run */
b8efb561
IM
627#ifdef CONFIG_SCHEDSTATS
628 /* BKL stats */
480b9434 629 unsigned int bkl_count;
b8efb561 630#endif
1da177e4 631};
52f17b6c 632#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 633
52f17b6c 634#ifdef CONFIG_SCHEDSTATS
15ad7cdc 635extern const struct file_operations proc_schedstat_operations;
52f17b6c 636#endif /* CONFIG_SCHEDSTATS */
1da177e4 637
ca74e92b
SN
638#ifdef CONFIG_TASK_DELAY_ACCT
639struct task_delay_info {
640 spinlock_t lock;
641 unsigned int flags; /* Private per-task flags */
642
643 /* For each stat XXX, add following, aligned appropriately
644 *
645 * struct timespec XXX_start, XXX_end;
646 * u64 XXX_delay;
647 * u32 XXX_count;
648 *
649 * Atomicity of updates to XXX_delay, XXX_count protected by
650 * single lock above (split into XXX_lock if contention is an issue).
651 */
0ff92245
SN
652
653 /*
654 * XXX_count is incremented on every XXX operation, the delay
655 * associated with the operation is added to XXX_delay.
656 * XXX_delay contains the accumulated delay time in nanoseconds.
657 */
658 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
659 u64 blkio_delay; /* wait for sync block io completion */
660 u64 swapin_delay; /* wait for swapin block io completion */
661 u32 blkio_count; /* total count of the number of sync block */
662 /* io operations performed */
663 u32 swapin_count; /* total count of the number of swapin block */
664 /* io operations performed */
ca74e92b 665};
52f17b6c
CS
666#endif /* CONFIG_TASK_DELAY_ACCT */
667
668static inline int sched_info_on(void)
669{
670#ifdef CONFIG_SCHEDSTATS
671 return 1;
672#elif defined(CONFIG_TASK_DELAY_ACCT)
673 extern int delayacct_on;
674 return delayacct_on;
675#else
676 return 0;
ca74e92b 677#endif
52f17b6c 678}
ca74e92b 679
d15bcfdb
IM
680enum cpu_idle_type {
681 CPU_IDLE,
682 CPU_NOT_IDLE,
683 CPU_NEWLY_IDLE,
684 CPU_MAX_IDLE_TYPES
1da177e4
LT
685};
686
687/*
688 * sched-domains (multiprocessor balancing) declarations:
689 */
9aa7b369
IM
690
691/*
692 * Increase resolution of nice-level calculations:
693 */
694#define SCHED_LOAD_SHIFT 10
695#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
696
f8700df7 697#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 698
2dd73a4f 699#ifdef CONFIG_SMP
1da177e4
LT
700#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
701#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
702#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
703#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
704#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
705#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
706#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
707#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
5c45bf27 708#define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
89c4710e 709#define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
08c183f3 710#define SD_SERIALIZE 1024 /* Only a single load balancing instance */
1d3504fc 711#define SD_WAKE_IDLE_FAR 2048 /* Gain latency sacrificing cache hit */
5c45bf27 712
89c4710e
SS
713#define BALANCE_FOR_MC_POWER \
714 (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
715
716#define BALANCE_FOR_PKG_POWER \
717 ((sched_mc_power_savings || sched_smt_power_savings) ? \
718 SD_POWERSAVINGS_BALANCE : 0)
719
720#define test_sd_parent(sd, flag) ((sd->parent && \
721 (sd->parent->flags & flag)) ? 1 : 0)
5c45bf27 722
1da177e4
LT
723
724struct sched_group {
725 struct sched_group *next; /* Must be a circular list */
726 cpumask_t cpumask;
727
728 /*
729 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
730 * single CPU. This is read only (except for setup, hotplug CPU).
5517d86b 731 * Note : Never change cpu_power without recompute its reciprocal
1da177e4 732 */
5517d86b
ED
733 unsigned int __cpu_power;
734 /*
735 * reciprocal value of cpu_power to avoid expensive divides
736 * (see include/linux/reciprocal_div.h)
737 */
738 u32 reciprocal_cpu_power;
1da177e4
LT
739};
740
1d3504fc
HS
741enum sched_domain_level {
742 SD_LV_NONE = 0,
743 SD_LV_SIBLING,
744 SD_LV_MC,
745 SD_LV_CPU,
746 SD_LV_NODE,
747 SD_LV_ALLNODES,
748 SD_LV_MAX
749};
750
751struct sched_domain_attr {
752 int relax_domain_level;
753};
754
755#define SD_ATTR_INIT (struct sched_domain_attr) { \
756 .relax_domain_level = -1, \
757}
758
1da177e4
LT
759struct sched_domain {
760 /* These fields must be setup */
761 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 762 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4
LT
763 struct sched_group *groups; /* the balancing groups of the domain */
764 cpumask_t span; /* span of all CPUs in this domain */
18d95a28 765 int first_cpu; /* cache of the first cpu in this domain */
1da177e4
LT
766 unsigned long min_interval; /* Minimum balance interval ms */
767 unsigned long max_interval; /* Maximum balance interval ms */
768 unsigned int busy_factor; /* less balancing by factor if busy */
769 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 770 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
771 unsigned int busy_idx;
772 unsigned int idle_idx;
773 unsigned int newidle_idx;
774 unsigned int wake_idx;
147cbb4b 775 unsigned int forkexec_idx;
1da177e4 776 int flags; /* See SD_* */
1d3504fc 777 enum sched_domain_level level;
1da177e4
LT
778
779 /* Runtime fields. */
780 unsigned long last_balance; /* init to jiffies. units in jiffies */
781 unsigned int balance_interval; /* initialise to 1. units in ms. */
782 unsigned int nr_balance_failed; /* initialise to 0 */
783
784#ifdef CONFIG_SCHEDSTATS
785 /* load_balance() stats */
480b9434
KC
786 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
787 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
788 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
789 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
790 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
791 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
792 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
793 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
794
795 /* Active load balancing */
480b9434
KC
796 unsigned int alb_count;
797 unsigned int alb_failed;
798 unsigned int alb_pushed;
1da177e4 799
68767a0a 800 /* SD_BALANCE_EXEC stats */
480b9434
KC
801 unsigned int sbe_count;
802 unsigned int sbe_balanced;
803 unsigned int sbe_pushed;
1da177e4 804
68767a0a 805 /* SD_BALANCE_FORK stats */
480b9434
KC
806 unsigned int sbf_count;
807 unsigned int sbf_balanced;
808 unsigned int sbf_pushed;
68767a0a 809
1da177e4 810 /* try_to_wake_up() stats */
480b9434
KC
811 unsigned int ttwu_wake_remote;
812 unsigned int ttwu_move_affine;
813 unsigned int ttwu_move_balance;
1da177e4
LT
814#endif
815};
816
1d3504fc
HS
817extern void partition_sched_domains(int ndoms_new, cpumask_t *doms_new,
818 struct sched_domain_attr *dattr_new);
9aefd0ab 819extern int arch_reinit_sched_domains(void);
029190c5 820
198e2f18 821#endif /* CONFIG_SMP */
1da177e4 822
d02c7a8c
CK
823/*
824 * A runqueue laden with a single nice 0 task scores a weighted_cpuload of
825 * SCHED_LOAD_SCALE. This function returns 1 if any cpu is laden with a
826 * task of nice 0 or enough lower priority tasks to bring up the
827 * weighted_cpuload
828 */
829static inline int above_background_load(void)
830{
831 unsigned long cpu;
832
833 for_each_online_cpu(cpu) {
834 if (weighted_cpuload(cpu) >= SCHED_LOAD_SCALE)
835 return 1;
836 }
837 return 0;
838}
1da177e4
LT
839
840struct io_context; /* See blkdev.h */
1da177e4 841#define NGROUPS_SMALL 32
1bf47346 842#define NGROUPS_PER_BLOCK ((unsigned int)(PAGE_SIZE / sizeof(gid_t)))
1da177e4
LT
843struct group_info {
844 int ngroups;
845 atomic_t usage;
846 gid_t small_block[NGROUPS_SMALL];
847 int nblocks;
848 gid_t *blocks[0];
849};
850
851/*
852 * get_group_info() must be called with the owning task locked (via task_lock())
853 * when task != current. The reason being that the vast majority of callers are
854 * looking at current->group_info, which can not be changed except by the
855 * current task. Changing current->group_info requires the task lock, too.
856 */
857#define get_group_info(group_info) do { \
858 atomic_inc(&(group_info)->usage); \
859} while (0)
860
861#define put_group_info(group_info) do { \
862 if (atomic_dec_and_test(&(group_info)->usage)) \
863 groups_free(group_info); \
864} while (0)
865
3e30148c
DH
866extern struct group_info *groups_alloc(int gidsetsize);
867extern void groups_free(struct group_info *group_info);
868extern int set_current_groups(struct group_info *group_info);
869extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
870/* access the groups "array" with this macro */
871#define GROUP_AT(gi, i) \
872 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
873
383f2835 874#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 875extern void prefetch_stack(struct task_struct *t);
383f2835
KC
876#else
877static inline void prefetch_stack(struct task_struct *t) { }
878#endif
1da177e4
LT
879
880struct audit_context; /* See audit.c */
881struct mempolicy;
b92ce558 882struct pipe_inode_info;
4865ecf1 883struct uts_namespace;
1da177e4 884
20b8a59f
IM
885struct rq;
886struct sched_domain;
887
888struct sched_class {
5522d5d5 889 const struct sched_class *next;
20b8a59f 890
fd390f6a 891 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
f02231e5 892 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
4530d7ab 893 void (*yield_task) (struct rq *rq);
e7693a36 894 int (*select_task_rq)(struct task_struct *p, int sync);
20b8a59f
IM
895
896 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p);
897
fb8d4724 898 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 899 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 900
681f3e68 901#ifdef CONFIG_SMP
43010659 902 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
e1d1484f 903 struct rq *busiest, unsigned long max_load_move,
20b8a59f 904 struct sched_domain *sd, enum cpu_idle_type idle,
a4ac01c3 905 int *all_pinned, int *this_best_prio);
20b8a59f 906
e1d1484f
PW
907 int (*move_one_task) (struct rq *this_rq, int this_cpu,
908 struct rq *busiest, struct sched_domain *sd,
909 enum cpu_idle_type idle);
9a897c5a
SR
910 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
911 void (*post_schedule) (struct rq *this_rq);
912 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
681f3e68 913#endif
e1d1484f 914
83b699ed 915 void (*set_curr_task) (struct rq *rq);
8f4d37ec 916 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
ee0827d8 917 void (*task_new) (struct rq *rq, struct task_struct *p);
cd8ba7cd
MT
918 void (*set_cpus_allowed)(struct task_struct *p,
919 const cpumask_t *newmask);
57d885fe
GH
920
921 void (*join_domain)(struct rq *rq);
922 void (*leave_domain)(struct rq *rq);
cb469845
SR
923
924 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
925 int running);
926 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
927 int running);
928 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
929 int oldprio, int running);
810b3817
PZ
930
931#ifdef CONFIG_FAIR_GROUP_SCHED
932 void (*moved_group) (struct task_struct *p);
933#endif
20b8a59f
IM
934};
935
936struct load_weight {
937 unsigned long weight, inv_weight;
938};
939
940/*
941 * CFS stats for a schedulable entity (task, task-group etc)
942 *
943 * Current field usage histogram:
944 *
945 * 4 se->block_start
946 * 4 se->run_node
947 * 4 se->sleep_start
20b8a59f 948 * 6 se->load.weight
20b8a59f
IM
949 */
950struct sched_entity {
20b8a59f
IM
951 struct load_weight load; /* for load-balancing */
952 struct rb_node run_node;
4a55bd5e 953 struct list_head group_node;
20b8a59f
IM
954 unsigned int on_rq;
955
94c18227
IM
956 u64 exec_start;
957 u64 sum_exec_runtime;
e9acbff6 958 u64 vruntime;
f6cf891c 959 u64 prev_sum_exec_runtime;
94c18227 960
4ae7d5ce
IM
961 u64 last_wakeup;
962 u64 avg_overlap;
963
94c18227 964#ifdef CONFIG_SCHEDSTATS
20b8a59f 965 u64 wait_start;
94c18227 966 u64 wait_max;
6d082592
AV
967 u64 wait_count;
968 u64 wait_sum;
94c18227 969
20b8a59f 970 u64 sleep_start;
20b8a59f 971 u64 sleep_max;
94c18227
IM
972 s64 sum_sleep_runtime;
973
974 u64 block_start;
20b8a59f
IM
975 u64 block_max;
976 u64 exec_max;
eba1ed4b 977 u64 slice_max;
cc367732
IM
978
979 u64 nr_migrations;
980 u64 nr_migrations_cold;
981 u64 nr_failed_migrations_affine;
982 u64 nr_failed_migrations_running;
983 u64 nr_failed_migrations_hot;
984 u64 nr_forced_migrations;
985 u64 nr_forced2_migrations;
986
987 u64 nr_wakeups;
988 u64 nr_wakeups_sync;
989 u64 nr_wakeups_migrate;
990 u64 nr_wakeups_local;
991 u64 nr_wakeups_remote;
992 u64 nr_wakeups_affine;
993 u64 nr_wakeups_affine_attempts;
994 u64 nr_wakeups_passive;
995 u64 nr_wakeups_idle;
94c18227
IM
996#endif
997
20b8a59f
IM
998#ifdef CONFIG_FAIR_GROUP_SCHED
999 struct sched_entity *parent;
1000 /* rq on which this entity is (to be) queued: */
1001 struct cfs_rq *cfs_rq;
1002 /* rq "owned" by this entity/group: */
1003 struct cfs_rq *my_q;
1004#endif
1005};
70b97a7f 1006
fa717060
PZ
1007struct sched_rt_entity {
1008 struct list_head run_list;
1009 unsigned int time_slice;
78f2c7db 1010 unsigned long timeout;
6f505b16
PZ
1011 int nr_cpus_allowed;
1012
58d6c2d7 1013 struct sched_rt_entity *back;
052f1dc7 1014#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1015 struct sched_rt_entity *parent;
1016 /* rq on which this entity is (to be) queued: */
1017 struct rt_rq *rt_rq;
1018 /* rq "owned" by this entity/group: */
1019 struct rt_rq *my_q;
1020#endif
fa717060
PZ
1021};
1022
1da177e4
LT
1023struct task_struct {
1024 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1025 void *stack;
1da177e4 1026 atomic_t usage;
97dc32cd
WC
1027 unsigned int flags; /* per process flags, defined below */
1028 unsigned int ptrace;
1da177e4 1029
36772092 1030 int lock_depth; /* BKL lock depth */
1da177e4 1031
2dd73a4f
PW
1032#ifdef CONFIG_SMP
1033#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1034 int oncpu;
1035#endif
2dd73a4f 1036#endif
50e645a8 1037
b29739f9 1038 int prio, static_prio, normal_prio;
5522d5d5 1039 const struct sched_class *sched_class;
20b8a59f 1040 struct sched_entity se;
fa717060 1041 struct sched_rt_entity rt;
1da177e4 1042
e107be36
AK
1043#ifdef CONFIG_PREEMPT_NOTIFIERS
1044 /* list of struct preempt_notifier: */
1045 struct hlist_head preempt_notifiers;
1046#endif
1047
18796aa0
AD
1048 /*
1049 * fpu_counter contains the number of consecutive context switches
1050 * that the FPU is used. If this is over a threshold, the lazy fpu
1051 * saving becomes unlazy to save the trap. This is an unsigned char
1052 * so that after 256 times the counter wraps and the behavior turns
1053 * lazy again; this to deal with bursty apps that only use FPU for
1054 * a short time
1055 */
1056 unsigned char fpu_counter;
1057 s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
6c5c9341 1058#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1059 unsigned int btrace_seq;
6c5c9341 1060#endif
1da177e4 1061
97dc32cd 1062 unsigned int policy;
1da177e4 1063 cpumask_t cpus_allowed;
1da177e4 1064
e260be67
PM
1065#ifdef CONFIG_PREEMPT_RCU
1066 int rcu_read_lock_nesting;
1067 int rcu_flipctr_idx;
1068#endif /* #ifdef CONFIG_PREEMPT_RCU */
1069
52f17b6c 1070#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1071 struct sched_info sched_info;
1072#endif
1073
1074 struct list_head tasks;
1075 /*
1076 * ptrace_list/ptrace_children forms the list of my children
1077 * that were stolen by a ptracer.
1078 */
1079 struct list_head ptrace_children;
1080 struct list_head ptrace_list;
1081
1082 struct mm_struct *mm, *active_mm;
1083
1084/* task state */
1085 struct linux_binfmt *binfmt;
97dc32cd 1086 int exit_state;
1da177e4
LT
1087 int exit_code, exit_signal;
1088 int pdeath_signal; /* The signal sent when the parent dies */
1089 /* ??? */
97dc32cd 1090 unsigned int personality;
1da177e4
LT
1091 unsigned did_exec:1;
1092 pid_t pid;
1093 pid_t tgid;
0a425405
AV
1094
1095#ifdef CONFIG_CC_STACKPROTECTOR
1096 /* Canary value for the -fstack-protector gcc feature */
1097 unsigned long stack_canary;
1098#endif
1da177e4
LT
1099 /*
1100 * pointers to (original) parent process, youngest child, younger sibling,
1101 * older sibling, respectively. (p->father can be replaced with
1102 * p->parent->pid)
1103 */
1104 struct task_struct *real_parent; /* real parent process (when being debugged) */
1105 struct task_struct *parent; /* parent process */
1106 /*
1107 * children/sibling forms the list of my children plus the
1108 * tasks I'm ptracing.
1109 */
1110 struct list_head children; /* list of my children */
1111 struct list_head sibling; /* linkage in my parent's children list */
1112 struct task_struct *group_leader; /* threadgroup leader */
1113
1114 /* PID/PID hash table linkage. */
92476d7f 1115 struct pid_link pids[PIDTYPE_MAX];
47e65328 1116 struct list_head thread_group;
1da177e4
LT
1117
1118 struct completion *vfork_done; /* for vfork() */
1119 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1120 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1121
97dc32cd 1122 unsigned int rt_priority;
c66f08be 1123 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1124 cputime_t gtime;
9301899b 1125 cputime_t prev_utime, prev_stime;
1da177e4 1126 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1127 struct timespec start_time; /* monotonic time */
1128 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1129/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1130 unsigned long min_flt, maj_flt;
1131
1132 cputime_t it_prof_expires, it_virt_expires;
1133 unsigned long long it_sched_expires;
1134 struct list_head cpu_timers[3];
1135
1136/* process credentials */
1137 uid_t uid,euid,suid,fsuid;
1138 gid_t gid,egid,sgid,fsgid;
1139 struct group_info *group_info;
3b7391de 1140 kernel_cap_t cap_effective, cap_inheritable, cap_permitted, cap_bset;
3898b1b4 1141 unsigned securebits;
1da177e4
LT
1142 struct user_struct *user;
1143#ifdef CONFIG_KEYS
b5f545c8 1144 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 1145 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 1146 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4 1147#endif
36772092
PBG
1148 char comm[TASK_COMM_LEN]; /* executable name excluding path
1149 - access with [gs]et_task_comm (which lock
1150 it with task_lock())
1151 - initialized normally by flush_old_exec */
1da177e4
LT
1152/* file system info */
1153 int link_count, total_link_count;
3d5b6fcc 1154#ifdef CONFIG_SYSVIPC
1da177e4
LT
1155/* ipc stuff */
1156 struct sysv_sem sysvsem;
3d5b6fcc 1157#endif
82a1fcb9
IM
1158#ifdef CONFIG_DETECT_SOFTLOCKUP
1159/* hung task detection */
1160 unsigned long last_switch_timestamp;
1161 unsigned long last_switch_count;
1162#endif
1da177e4
LT
1163/* CPU-specific state of this task */
1164 struct thread_struct thread;
1165/* filesystem information */
1166 struct fs_struct *fs;
1167/* open file information */
1168 struct files_struct *files;
1651e14e 1169/* namespaces */
ab516013 1170 struct nsproxy *nsproxy;
1da177e4
LT
1171/* signal handlers */
1172 struct signal_struct *signal;
1173 struct sighand_struct *sighand;
1174
1175 sigset_t blocked, real_blocked;
150256d8 1176 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
1da177e4
LT
1177 struct sigpending pending;
1178
1179 unsigned long sas_ss_sp;
1180 size_t sas_ss_size;
1181 int (*notifier)(void *priv);
1182 void *notifier_data;
1183 sigset_t *notifier_mask;
57c521ce 1184#ifdef CONFIG_SECURITY
1da177e4 1185 void *security;
57c521ce 1186#endif
1da177e4 1187 struct audit_context *audit_context;
bfef93a5
AV
1188#ifdef CONFIG_AUDITSYSCALL
1189 uid_t loginuid;
4746ec5b 1190 unsigned int sessionid;
bfef93a5 1191#endif
1da177e4
LT
1192 seccomp_t seccomp;
1193
1194/* Thread group tracking */
1195 u32 parent_exec_id;
1196 u32 self_exec_id;
1197/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1198 spinlock_t alloc_lock;
1da177e4 1199
b29739f9
IM
1200 /* Protection of the PI data structures: */
1201 spinlock_t pi_lock;
1202
23f78d4a
IM
1203#ifdef CONFIG_RT_MUTEXES
1204 /* PI waiters blocked on a rt_mutex held by this task */
1205 struct plist_head pi_waiters;
1206 /* Deadlock detection and priority inheritance handling */
1207 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1208#endif
1209
408894ee
IM
1210#ifdef CONFIG_DEBUG_MUTEXES
1211 /* mutex deadlock detection */
1212 struct mutex_waiter *blocked_on;
1213#endif
de30a2b3
IM
1214#ifdef CONFIG_TRACE_IRQFLAGS
1215 unsigned int irq_events;
1216 int hardirqs_enabled;
1217 unsigned long hardirq_enable_ip;
1218 unsigned int hardirq_enable_event;
1219 unsigned long hardirq_disable_ip;
1220 unsigned int hardirq_disable_event;
1221 int softirqs_enabled;
1222 unsigned long softirq_disable_ip;
1223 unsigned int softirq_disable_event;
1224 unsigned long softirq_enable_ip;
1225 unsigned int softirq_enable_event;
1226 int hardirq_context;
1227 int softirq_context;
1228#endif
fbb9ce95 1229#ifdef CONFIG_LOCKDEP
bdb9441e 1230# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1231 u64 curr_chain_key;
1232 int lockdep_depth;
1233 struct held_lock held_locks[MAX_LOCK_DEPTH];
1234 unsigned int lockdep_recursion;
1235#endif
408894ee 1236
1da177e4
LT
1237/* journalling filesystem info */
1238 void *journal_info;
1239
d89d8796
NB
1240/* stacked block device info */
1241 struct bio *bio_list, **bio_tail;
1242
1da177e4
LT
1243/* VM state */
1244 struct reclaim_state *reclaim_state;
1245
1da177e4
LT
1246 struct backing_dev_info *backing_dev_info;
1247
1248 struct io_context *io_context;
1249
1250 unsigned long ptrace_message;
1251 siginfo_t *last_siginfo; /* For ptrace use. */
4b98d11b 1252#ifdef CONFIG_TASK_XACCT
1da177e4
LT
1253/* i/o counters(bytes read/written, #syscalls */
1254 u64 rchar, wchar, syscr, syscw;
4b98d11b 1255#endif
7c3ab738 1256 struct task_io_accounting ioac;
8f0ab514 1257#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1258 u64 acct_rss_mem1; /* accumulated rss usage */
1259 u64 acct_vm_mem1; /* accumulated virtual memory usage */
db5fed26 1260 cputime_t acct_stimexpd;/* stime since last update */
1da177e4
LT
1261#endif
1262#ifdef CONFIG_NUMA
1263 struct mempolicy *mempolicy;
1264 short il_next;
1265#endif
1266#ifdef CONFIG_CPUSETS
1da177e4
LT
1267 nodemask_t mems_allowed;
1268 int cpuset_mems_generation;
825a46af 1269 int cpuset_mem_spread_rotor;
1da177e4 1270#endif
ddbcc7e8 1271#ifdef CONFIG_CGROUPS
817929ec
PM
1272 /* Control Group info protected by css_set_lock */
1273 struct css_set *cgroups;
1274 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1275 struct list_head cg_list;
ddbcc7e8 1276#endif
42b2dd0a 1277#ifdef CONFIG_FUTEX
0771dfef 1278 struct robust_list_head __user *robust_list;
34f192c6
IM
1279#ifdef CONFIG_COMPAT
1280 struct compat_robust_list_head __user *compat_robust_list;
1281#endif
c87e2837
IM
1282 struct list_head pi_state_list;
1283 struct futex_pi_state *pi_state_cache;
42b2dd0a 1284#endif
22e2c507 1285 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1286 struct rcu_head rcu;
b92ce558
JA
1287
1288 /*
1289 * cache last used pipe for splice
1290 */
1291 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1292#ifdef CONFIG_TASK_DELAY_ACCT
1293 struct task_delay_info *delays;
f4f154fd
AM
1294#endif
1295#ifdef CONFIG_FAULT_INJECTION
1296 int make_it_fail;
ca74e92b 1297#endif
3e26c149 1298 struct prop_local_single dirties;
9745512c
AV
1299#ifdef CONFIG_LATENCYTOP
1300 int latency_record_count;
1301 struct latency_record latency_record[LT_SAVECOUNT];
1302#endif
1da177e4
LT
1303};
1304
e05606d3
IM
1305/*
1306 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1307 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1308 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1309 * values are inverted: lower p->prio value means higher priority.
1310 *
1311 * The MAX_USER_RT_PRIO value allows the actual maximum
1312 * RT priority to be separate from the value exported to
1313 * user-space. This allows kernel threads to set their
1314 * priority to a value higher than any user task. Note:
1315 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1316 */
1317
1318#define MAX_USER_RT_PRIO 100
1319#define MAX_RT_PRIO MAX_USER_RT_PRIO
1320
1321#define MAX_PRIO (MAX_RT_PRIO + 40)
1322#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1323
1324static inline int rt_prio(int prio)
1325{
1326 if (unlikely(prio < MAX_RT_PRIO))
1327 return 1;
1328 return 0;
1329}
1330
e868171a 1331static inline int rt_task(struct task_struct *p)
e05606d3
IM
1332{
1333 return rt_prio(p->prio);
1334}
1335
a47afb0f 1336static inline void set_task_session(struct task_struct *tsk, pid_t session)
937949d9 1337{
a47afb0f 1338 tsk->signal->__session = session;
937949d9
CLG
1339}
1340
9a2e7057
PE
1341static inline void set_task_pgrp(struct task_struct *tsk, pid_t pgrp)
1342{
1343 tsk->signal->__pgrp = pgrp;
1344}
1345
e868171a 1346static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1347{
1348 return task->pids[PIDTYPE_PID].pid;
1349}
1350
e868171a 1351static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1352{
1353 return task->group_leader->pids[PIDTYPE_PID].pid;
1354}
1355
e868171a 1356static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1357{
1358 return task->group_leader->pids[PIDTYPE_PGID].pid;
1359}
1360
e868171a 1361static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1362{
1363 return task->group_leader->pids[PIDTYPE_SID].pid;
1364}
1365
7af57294
PE
1366struct pid_namespace;
1367
1368/*
1369 * the helpers to get the task's different pids as they are seen
1370 * from various namespaces
1371 *
1372 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1373 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1374 * current.
7af57294
PE
1375 * task_xid_nr_ns() : id seen from the ns specified;
1376 *
1377 * set_task_vxid() : assigns a virtual id to a task;
1378 *
7af57294
PE
1379 * see also pid_nr() etc in include/linux/pid.h
1380 */
1381
e868171a 1382static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1383{
1384 return tsk->pid;
1385}
1386
2f2a3a46 1387pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1388
1389static inline pid_t task_pid_vnr(struct task_struct *tsk)
1390{
1391 return pid_vnr(task_pid(tsk));
1392}
1393
1394
e868171a 1395static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1396{
1397 return tsk->tgid;
1398}
1399
2f2a3a46 1400pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1401
1402static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1403{
1404 return pid_vnr(task_tgid(tsk));
1405}
1406
1407
e868171a 1408static inline pid_t task_pgrp_nr(struct task_struct *tsk)
7af57294 1409{
9a2e7057 1410 return tsk->signal->__pgrp;
7af57294
PE
1411}
1412
2f2a3a46 1413pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1414
1415static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1416{
1417 return pid_vnr(task_pgrp(tsk));
1418}
1419
1420
e868171a 1421static inline pid_t task_session_nr(struct task_struct *tsk)
7af57294
PE
1422{
1423 return tsk->signal->__session;
1424}
1425
2f2a3a46 1426pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1427
1428static inline pid_t task_session_vnr(struct task_struct *tsk)
1429{
1430 return pid_vnr(task_session(tsk));
1431}
1432
1433
1da177e4
LT
1434/**
1435 * pid_alive - check that a task structure is not stale
1436 * @p: Task structure to be checked.
1437 *
1438 * Test if a process is not yet dead (at most zombie state)
1439 * If pid_alive fails, then pointers within the task structure
1440 * can be stale and must not be dereferenced.
1441 */
e868171a 1442static inline int pid_alive(struct task_struct *p)
1da177e4 1443{
92476d7f 1444 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1445}
1446
f400e198 1447/**
b460cbc5 1448 * is_global_init - check if a task structure is init
3260259f
HK
1449 * @tsk: Task structure to be checked.
1450 *
1451 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1452 */
e868171a 1453static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1454{
1455 return tsk->pid == 1;
1456}
b460cbc5
SH
1457
1458/*
1459 * is_container_init:
1460 * check whether in the task is init in its own pid namespace.
f400e198 1461 */
b461cc03 1462extern int is_container_init(struct task_struct *tsk);
f400e198 1463
9ec52099
CLG
1464extern struct pid *cad_pid;
1465
1da177e4 1466extern void free_task(struct task_struct *tsk);
1da177e4 1467#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1468
158d9ebd 1469extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1470
1471static inline void put_task_struct(struct task_struct *t)
1472{
1473 if (atomic_dec_and_test(&t->usage))
8c7904a0 1474 __put_task_struct(t);
e56d0903 1475}
1da177e4
LT
1476
1477/*
1478 * Per process flags
1479 */
1480#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1481 /* Not implemented yet, only for 486*/
1482#define PF_STARTING 0x00000002 /* being created */
1483#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1484#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1485#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1da177e4
LT
1486#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1487#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1488#define PF_DUMPCORE 0x00000200 /* dumped core */
1489#define PF_SIGNALED 0x00000400 /* killed by a signal */
1490#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1491#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1492#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1da177e4
LT
1493#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1494#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1495#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1496#define PF_KSWAPD 0x00040000 /* I am kswapd */
1497#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1498#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
b31dc66a
JA
1499#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1500#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1501#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1502#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1503#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
c61afb18 1504#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1505#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
ba96a0c8 1506#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1da177e4
LT
1507
1508/*
1509 * Only the _current_ task can read/write to tsk->flags, but other
1510 * tasks can access tsk->flags in readonly mode for example
1511 * with tsk_used_math (like during threaded core dumping).
1512 * There is however an exception to this rule during ptrace
1513 * or during fork: the ptracer task is allowed to write to the
1514 * child->flags of its traced child (same goes for fork, the parent
1515 * can write to the child->flags), because we're guaranteed the
1516 * child is not running and in turn not changing child->flags
1517 * at the same time the parent does it.
1518 */
1519#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1520#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1521#define clear_used_math() clear_stopped_child_used_math(current)
1522#define set_used_math() set_stopped_child_used_math(current)
1523#define conditional_stopped_child_used_math(condition, child) \
1524 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1525#define conditional_used_math(condition) \
1526 conditional_stopped_child_used_math(condition, current)
1527#define copy_to_stopped_child_used_math(child) \
1528 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1529/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1530#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1531#define used_math() tsk_used_math(current)
1532
1533#ifdef CONFIG_SMP
cd8ba7cd
MT
1534extern int set_cpus_allowed_ptr(struct task_struct *p,
1535 const cpumask_t *new_mask);
1da177e4 1536#else
cd8ba7cd
MT
1537static inline int set_cpus_allowed_ptr(struct task_struct *p,
1538 const cpumask_t *new_mask)
1da177e4 1539{
cd8ba7cd 1540 if (!cpu_isset(0, *new_mask))
1da177e4
LT
1541 return -EINVAL;
1542 return 0;
1543}
1544#endif
cd8ba7cd
MT
1545static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1546{
1547 return set_cpus_allowed_ptr(p, &new_mask);
1548}
1da177e4
LT
1549
1550extern unsigned long long sched_clock(void);
e436d800
IM
1551
1552/*
1553 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1554 * clock constructed from sched_clock():
1555 */
1556extern unsigned long long cpu_clock(int cpu);
1557
36c8b586 1558extern unsigned long long
41b86e9c 1559task_sched_runtime(struct task_struct *task);
1da177e4
LT
1560
1561/* sched_exec is called by processes performing an exec */
1562#ifdef CONFIG_SMP
1563extern void sched_exec(void);
1564#else
1565#define sched_exec() {}
1566#endif
1567
2aa44d05
IM
1568extern void sched_clock_idle_sleep_event(void);
1569extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1570
1da177e4
LT
1571#ifdef CONFIG_HOTPLUG_CPU
1572extern void idle_task_exit(void);
1573#else
1574static inline void idle_task_exit(void) {}
1575#endif
1576
1577extern void sched_idle_next(void);
b29739f9 1578
06d8308c
TG
1579#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1580extern void wake_up_idle_cpu(int cpu);
1581#else
1582static inline void wake_up_idle_cpu(int cpu) { }
1583#endif
1584
2bd8e6d4 1585#ifdef CONFIG_SCHED_DEBUG
21805085 1586extern unsigned int sysctl_sched_latency;
b2be5e96 1587extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1588extern unsigned int sysctl_sched_wakeup_granularity;
bf0f6f24
IM
1589extern unsigned int sysctl_sched_child_runs_first;
1590extern unsigned int sysctl_sched_features;
da84d961 1591extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1592extern unsigned int sysctl_sched_nr_migrate;
b2be5e96
PZ
1593
1594int sched_nr_latency_handler(struct ctl_table *table, int write,
1595 struct file *file, void __user *buffer, size_t *length,
1596 loff_t *ppos);
2bd8e6d4 1597#endif
9f0c1e56
PZ
1598extern unsigned int sysctl_sched_rt_period;
1599extern int sysctl_sched_rt_runtime;
2bd8e6d4 1600
d0b27fa7
PZ
1601int sched_rt_handler(struct ctl_table *table, int write,
1602 struct file *filp, void __user *buffer, size_t *lenp,
1603 loff_t *ppos);
1604
2bd8e6d4 1605extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1606
b29739f9 1607#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1608extern int rt_mutex_getprio(struct task_struct *p);
1609extern void rt_mutex_setprio(struct task_struct *p, int prio);
1610extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1611#else
e868171a 1612static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1613{
1614 return p->normal_prio;
1615}
95e02ca9 1616# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1617#endif
1618
36c8b586
IM
1619extern void set_user_nice(struct task_struct *p, long nice);
1620extern int task_prio(const struct task_struct *p);
1621extern int task_nice(const struct task_struct *p);
1622extern int can_nice(const struct task_struct *p, const int nice);
1623extern int task_curr(const struct task_struct *p);
1da177e4
LT
1624extern int idle_cpu(int cpu);
1625extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
36c8b586
IM
1626extern struct task_struct *idle_task(int cpu);
1627extern struct task_struct *curr_task(int cpu);
1628extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1629
1630void yield(void);
1631
1632/*
1633 * The default (Linux) execution domain.
1634 */
1635extern struct exec_domain default_exec_domain;
1636
1637union thread_union {
1638 struct thread_info thread_info;
1639 unsigned long stack[THREAD_SIZE/sizeof(long)];
1640};
1641
1642#ifndef __HAVE_ARCH_KSTACK_END
1643static inline int kstack_end(void *addr)
1644{
1645 /* Reliable end of stack detection:
1646 * Some APM bios versions misalign the stack
1647 */
1648 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1649}
1650#endif
1651
1652extern union thread_union init_thread_union;
1653extern struct task_struct init_task;
1654
1655extern struct mm_struct init_mm;
1656
198fe21b
PE
1657extern struct pid_namespace init_pid_ns;
1658
1659/*
1660 * find a task by one of its numerical ids
1661 *
1662 * find_task_by_pid_type_ns():
1663 * it is the most generic call - it finds a task by all id,
1664 * type and namespace specified
1665 * find_task_by_pid_ns():
1666 * finds a task by its pid in the specified namespace
228ebcbe
PE
1667 * find_task_by_vpid():
1668 * finds a task by its virtual pid
198fe21b
PE
1669 * find_task_by_pid():
1670 * finds a task by its global pid
1671 *
1672 * see also find_pid() etc in include/linux/pid.h
1673 */
1674
1675extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1676 struct pid_namespace *ns);
1677
228ebcbe
PE
1678extern struct task_struct *find_task_by_pid(pid_t nr);
1679extern struct task_struct *find_task_by_vpid(pid_t nr);
1680extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1681 struct pid_namespace *ns);
198fe21b 1682
8520d7c7 1683extern void __set_special_pids(struct pid *pid);
1da177e4
LT
1684
1685/* per-UID process charging. */
acce292c 1686extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
1687static inline struct user_struct *get_uid(struct user_struct *u)
1688{
1689 atomic_inc(&u->__count);
1690 return u;
1691}
1692extern void free_uid(struct user_struct *);
1693extern void switch_uid(struct user_struct *);
28f300d2 1694extern void release_uids(struct user_namespace *ns);
1da177e4
LT
1695
1696#include <asm/current.h>
1697
3171a030 1698extern void do_timer(unsigned long ticks);
1da177e4 1699
b3c97528
HH
1700extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1701extern int wake_up_process(struct task_struct *tsk);
1702extern void wake_up_new_task(struct task_struct *tsk,
1703 unsigned long clone_flags);
1da177e4
LT
1704#ifdef CONFIG_SMP
1705 extern void kick_process(struct task_struct *tsk);
1706#else
1707 static inline void kick_process(struct task_struct *tsk) { }
1708#endif
ad46c2c4
IM
1709extern void sched_fork(struct task_struct *p, int clone_flags);
1710extern void sched_dead(struct task_struct *p);
1da177e4
LT
1711
1712extern int in_group_p(gid_t);
1713extern int in_egroup_p(gid_t);
1714
1715extern void proc_caches_init(void);
1716extern void flush_signals(struct task_struct *);
10ab825b 1717extern void ignore_signals(struct task_struct *);
1da177e4
LT
1718extern void flush_signal_handlers(struct task_struct *, int force_default);
1719extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1720
1721static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1722{
1723 unsigned long flags;
1724 int ret;
1725
1726 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1727 ret = dequeue_signal(tsk, mask, info);
1728 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1729
1730 return ret;
1731}
1732
1733extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1734 sigset_t *mask);
1735extern void unblock_all_signals(void);
1736extern void release_task(struct task_struct * p);
1737extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
1738extern int force_sigsegv(int, struct task_struct *);
1739extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 1740extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 1741extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 1742extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
1743extern int kill_pgrp(struct pid *pid, int sig, int priv);
1744extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 1745extern int kill_proc_info(int, struct siginfo *, pid_t);
1da177e4
LT
1746extern void do_notify_parent(struct task_struct *, int);
1747extern void force_sig(int, struct task_struct *);
1748extern void force_sig_specific(int, struct task_struct *);
1749extern int send_sig(int, struct task_struct *, int);
1750extern void zap_other_threads(struct task_struct *p);
1da177e4
LT
1751extern int kill_proc(pid_t, int, int);
1752extern struct sigqueue *sigqueue_alloc(void);
1753extern void sigqueue_free(struct sigqueue *);
1754extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1755extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
9ac95f2f 1756extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
1757extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1758
9ec52099
CLG
1759static inline int kill_cad_pid(int sig, int priv)
1760{
1761 return kill_pid(cad_pid, sig, priv);
1762}
1763
1da177e4
LT
1764/* These can be the second arg to send_sig_info/send_group_sig_info. */
1765#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1766#define SEND_SIG_PRIV ((struct siginfo *) 1)
1767#define SEND_SIG_FORCED ((struct siginfo *) 2)
1768
621d3121
ON
1769static inline int is_si_special(const struct siginfo *info)
1770{
1771 return info <= SEND_SIG_FORCED;
1772}
1773
1da177e4
LT
1774/* True if we are on the alternate signal stack. */
1775
1776static inline int on_sig_stack(unsigned long sp)
1777{
1778 return (sp - current->sas_ss_sp < current->sas_ss_size);
1779}
1780
1781static inline int sas_ss_flags(unsigned long sp)
1782{
1783 return (current->sas_ss_size == 0 ? SS_DISABLE
1784 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1785}
1786
1da177e4
LT
1787/*
1788 * Routines for handling mm_structs
1789 */
1790extern struct mm_struct * mm_alloc(void);
1791
1792/* mmdrop drops the mm and the page tables */
b3c97528 1793extern void __mmdrop(struct mm_struct *);
1da177e4
LT
1794static inline void mmdrop(struct mm_struct * mm)
1795{
6fb43d7b 1796 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
1797 __mmdrop(mm);
1798}
1799
1800/* mmput gets rid of the mappings and all user-space */
1801extern void mmput(struct mm_struct *);
1802/* Grab a reference to a task's mm, if it is not already going away */
1803extern struct mm_struct *get_task_mm(struct task_struct *task);
1804/* Remove the current tasks stale references to the old mm_struct */
1805extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
1806/* Allocate a new mm structure and copy contents from tsk->mm */
1807extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4
LT
1808
1809extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1810extern void flush_thread(void);
1811extern void exit_thread(void);
1812
1da177e4 1813extern void exit_files(struct task_struct *);
6b3934ef 1814extern void __cleanup_signal(struct signal_struct *);
a7e5328a 1815extern void __cleanup_sighand(struct sighand_struct *);
1da177e4
LT
1816extern void exit_itimers(struct signal_struct *);
1817
1818extern NORET_TYPE void do_group_exit(int);
1819
1da177e4
LT
1820extern void daemonize(const char *, ...);
1821extern int allow_signal(int);
1822extern int disallow_signal(int);
1da177e4
LT
1823
1824extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1825extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 1826struct task_struct *fork_idle(int);
1da177e4
LT
1827
1828extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 1829extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
1830
1831#ifdef CONFIG_SMP
36c8b586 1832extern void wait_task_inactive(struct task_struct * p);
1da177e4
LT
1833#else
1834#define wait_task_inactive(p) do { } while (0)
1835#endif
1836
1837#define remove_parent(p) list_del_init(&(p)->sibling)
8fafabd8 1838#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1da177e4 1839
5e85d4ab 1840#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1da177e4
LT
1841
1842#define for_each_process(p) \
1843 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1844
1845/*
1846 * Careful: do_each_thread/while_each_thread is a double loop so
1847 * 'break' will not work as expected - use goto instead.
1848 */
1849#define do_each_thread(g, t) \
1850 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1851
1852#define while_each_thread(g, t) \
1853 while ((t = next_thread(t)) != g)
1854
de12a787
EB
1855/* de_thread depends on thread_group_leader not being a pid based check */
1856#define thread_group_leader(p) (p == p->group_leader)
1da177e4 1857
0804ef4b
EB
1858/* Do to the insanities of de_thread it is possible for a process
1859 * to have the pid of the thread group leader without actually being
1860 * the thread group leader. For iteration through the pids in proc
1861 * all we care about is that we have a task with the appropriate
1862 * pid, we don't actually care if we have the right task.
1863 */
e868171a 1864static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
1865{
1866 return p->pid == p->tgid;
1867}
1868
bac0abd6
PE
1869static inline
1870int same_thread_group(struct task_struct *p1, struct task_struct *p2)
1871{
1872 return p1->tgid == p2->tgid;
1873}
1874
36c8b586 1875static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328
ON
1876{
1877 return list_entry(rcu_dereference(p->thread_group.next),
36c8b586 1878 struct task_struct, thread_group);
47e65328
ON
1879}
1880
e868171a 1881static inline int thread_group_empty(struct task_struct *p)
1da177e4 1882{
47e65328 1883 return list_empty(&p->thread_group);
1da177e4
LT
1884}
1885
1886#define delay_group_leader(p) \
1887 (thread_group_leader(p) && !thread_group_empty(p))
1888
1da177e4 1889/*
260ea101 1890 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 1891 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
1892 * pins the final release of task.io_context. Also protects ->cpuset and
1893 * ->cgroup.subsys[].
1da177e4
LT
1894 *
1895 * Nests both inside and outside of read_lock(&tasklist_lock).
1896 * It must not be nested with write_lock_irq(&tasklist_lock),
1897 * neither inside nor outside.
1898 */
1899static inline void task_lock(struct task_struct *p)
1900{
1901 spin_lock(&p->alloc_lock);
1902}
1903
1904static inline void task_unlock(struct task_struct *p)
1905{
1906 spin_unlock(&p->alloc_lock);
1907}
1908
f63ee72e
ON
1909extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1910 unsigned long *flags);
1911
1912static inline void unlock_task_sighand(struct task_struct *tsk,
1913 unsigned long *flags)
1914{
1915 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1916}
1917
f037360f
AV
1918#ifndef __HAVE_THREAD_FUNCTIONS
1919
f7e4217b
RZ
1920#define task_thread_info(task) ((struct thread_info *)(task)->stack)
1921#define task_stack_page(task) ((task)->stack)
a1261f54 1922
10ebffde
AV
1923static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1924{
1925 *task_thread_info(p) = *task_thread_info(org);
1926 task_thread_info(p)->task = p;
1927}
1928
1929static inline unsigned long *end_of_stack(struct task_struct *p)
1930{
f7e4217b 1931 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
1932}
1933
f037360f
AV
1934#endif
1935
8c9843e5
BH
1936extern void thread_info_cache_init(void);
1937
1da177e4
LT
1938/* set thread flags in other task's structures
1939 * - see asm/thread_info.h for TIF_xxxx flags available
1940 */
1941static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1942{
a1261f54 1943 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1944}
1945
1946static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1947{
a1261f54 1948 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1949}
1950
1951static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1952{
a1261f54 1953 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1954}
1955
1956static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1957{
a1261f54 1958 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1959}
1960
1961static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1962{
a1261f54 1963 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1964}
1965
1966static inline void set_tsk_need_resched(struct task_struct *tsk)
1967{
1968 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1969}
1970
1971static inline void clear_tsk_need_resched(struct task_struct *tsk)
1972{
1973 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1974}
1975
1976static inline int signal_pending(struct task_struct *p)
1977{
1978 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1979}
f776d12d 1980
b3c97528 1981extern int __fatal_signal_pending(struct task_struct *p);
f776d12d
MW
1982
1983static inline int fatal_signal_pending(struct task_struct *p)
1984{
1985 return signal_pending(p) && __fatal_signal_pending(p);
1986}
1987
1da177e4
LT
1988static inline int need_resched(void)
1989{
1990 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1991}
1992
1993/*
1994 * cond_resched() and cond_resched_lock(): latency reduction via
1995 * explicit rescheduling in places that are safe. The return
1996 * value indicates whether a reschedule was done in fact.
1997 * cond_resched_lock() will drop the spinlock before scheduling,
1998 * cond_resched_softirq() will enable bhs before scheduling.
1999 */
02b67cc3
HX
2000#ifdef CONFIG_PREEMPT
2001static inline int cond_resched(void)
2002{
2003 return 0;
2004}
2005#else
2006extern int _cond_resched(void);
2007static inline int cond_resched(void)
2008{
2009 return _cond_resched();
2010}
2011#endif
1da177e4
LT
2012extern int cond_resched_lock(spinlock_t * lock);
2013extern int cond_resched_softirq(void);
2014
2015/*
2016 * Does a critical section need to be broken due to another
95c354fe
NP
2017 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2018 * but a general need for low latency)
1da177e4 2019 */
95c354fe 2020static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2021{
95c354fe
NP
2022#ifdef CONFIG_PREEMPT
2023 return spin_is_contended(lock);
2024#else
1da177e4 2025 return 0;
95c354fe 2026#endif
1da177e4
LT
2027}
2028
7bb44ade
RM
2029/*
2030 * Reevaluate whether the task has signals pending delivery.
2031 * Wake the task if so.
2032 * This is required every time the blocked sigset_t changes.
2033 * callers must hold sighand->siglock.
2034 */
2035extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2036extern void recalc_sigpending(void);
2037
2038extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2039
2040/*
2041 * Wrappers for p->thread_info->cpu access. No-op on UP.
2042 */
2043#ifdef CONFIG_SMP
2044
2045static inline unsigned int task_cpu(const struct task_struct *p)
2046{
a1261f54 2047 return task_thread_info(p)->cpu;
1da177e4
LT
2048}
2049
c65cc870 2050extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2051
2052#else
2053
2054static inline unsigned int task_cpu(const struct task_struct *p)
2055{
2056 return 0;
2057}
2058
2059static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2060{
2061}
2062
2063#endif /* CONFIG_SMP */
2064
2065#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
2066extern void arch_pick_mmap_layout(struct mm_struct *mm);
2067#else
2068static inline void arch_pick_mmap_layout(struct mm_struct *mm)
2069{
2070 mm->mmap_base = TASK_UNMAPPED_BASE;
2071 mm->get_unmapped_area = arch_get_unmapped_area;
2072 mm->unmap_area = arch_unmap_area;
2073}
2074#endif
2075
b53e921b 2076extern long sched_setaffinity(pid_t pid, const cpumask_t *new_mask);
1da177e4
LT
2077extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
2078
5c45bf27 2079extern int sched_mc_power_savings, sched_smt_power_savings;
5c45bf27 2080
1da177e4
LT
2081extern void normalize_rt_tasks(void);
2082
052f1dc7 2083#ifdef CONFIG_GROUP_SCHED
9b5b7751 2084
4cf86d77 2085extern struct task_group init_task_group;
eff766a6
PZ
2086#ifdef CONFIG_USER_SCHED
2087extern struct task_group root_task_group;
2088#endif
9b5b7751 2089
ec7dc8ac 2090extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2091extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2092extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2093#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2094extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2095extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2096#endif
2097#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2098extern int sched_group_set_rt_runtime(struct task_group *tg,
2099 long rt_runtime_us);
2100extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2101extern int sched_group_set_rt_period(struct task_group *tg,
2102 long rt_period_us);
2103extern long sched_group_rt_period(struct task_group *tg);
052f1dc7 2104#endif
9b5b7751
SV
2105#endif
2106
4b98d11b
AD
2107#ifdef CONFIG_TASK_XACCT
2108static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2109{
2110 tsk->rchar += amt;
2111}
2112
2113static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2114{
2115 tsk->wchar += amt;
2116}
2117
2118static inline void inc_syscr(struct task_struct *tsk)
2119{
2120 tsk->syscr++;
2121}
2122
2123static inline void inc_syscw(struct task_struct *tsk)
2124{
2125 tsk->syscw++;
2126}
2127#else
2128static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2129{
2130}
2131
2132static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2133{
2134}
2135
2136static inline void inc_syscr(struct task_struct *tsk)
2137{
2138}
2139
2140static inline void inc_syscw(struct task_struct *tsk)
2141{
2142}
2143#endif
2144
e6fe6649
AB
2145#ifdef CONFIG_SMP
2146void migration_init(void);
2147#else
2148static inline void migration_init(void)
2149{
2150}
2151#endif
2152
82455257
DH
2153#ifndef TASK_SIZE_OF
2154#define TASK_SIZE_OF(tsk) TASK_SIZE
2155#endif
2156
cf475ad2
BS
2157#ifdef CONFIG_MM_OWNER
2158extern void mm_update_next_owner(struct mm_struct *mm);
2159extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2160#else
2161static inline void mm_update_next_owner(struct mm_struct *mm)
2162{
2163}
2164
2165static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2166{
2167}
2168#endif /* CONFIG_MM_OWNER */
2169
1da177e4
LT
2170#endif /* __KERNEL__ */
2171
2172#endif