xref: /linux-6.15/include/linux/sched.h (revision a1e58bbd)
1 #ifndef _LINUX_SCHED_H
2 #define _LINUX_SCHED_H
3 
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 */
25 #define CLONE_NEWUTS		0x04000000	/* New utsname group? */
26 #define CLONE_NEWIPC		0x08000000	/* New ipcs */
27 #define CLONE_NEWUSER		0x10000000	/* New user namespace */
28 #define CLONE_NEWPID		0x20000000	/* New pid namespace */
29 #define CLONE_NEWNET		0x40000000	/* New network namespace */
30 #define CLONE_IO		0x80000000	/* Clone io context */
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
39 /* SCHED_ISO: reserved but not implemented yet */
40 #define SCHED_IDLE		5
41 
42 #ifdef __KERNEL__
43 
44 struct sched_param {
45 	int sched_priority;
46 };
47 
48 #include <asm/param.h>	/* for HZ */
49 
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>
61 #include <linux/mm_types.h>
62 
63 #include <asm/system.h>
64 #include <asm/semaphore.h>
65 #include <asm/page.h>
66 #include <asm/ptrace.h>
67 #include <asm/cputime.h>
68 
69 #include <linux/smp.h>
70 #include <linux/sem.h>
71 #include <linux/signal.h>
72 #include <linux/securebits.h>
73 #include <linux/fs_struct.h>
74 #include <linux/compiler.h>
75 #include <linux/completion.h>
76 #include <linux/pid.h>
77 #include <linux/percpu.h>
78 #include <linux/topology.h>
79 #include <linux/proportions.h>
80 #include <linux/seccomp.h>
81 #include <linux/rcupdate.h>
82 #include <linux/rtmutex.h>
83 
84 #include <linux/time.h>
85 #include <linux/param.h>
86 #include <linux/resource.h>
87 #include <linux/timer.h>
88 #include <linux/hrtimer.h>
89 #include <linux/task_io_accounting.h>
90 #include <linux/kobject.h>
91 #include <linux/latencytop.h>
92 
93 #include <asm/processor.h>
94 
95 struct mem_cgroup;
96 struct exec_domain;
97 struct futex_pi_state;
98 struct robust_list_head;
99 struct bio;
100 
101 /*
102  * List of flags we want to share for kernel threads,
103  * if only because they are not used by them anyway.
104  */
105 #define CLONE_KERNEL	(CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
106 
107 /*
108  * These are the constant used to fake the fixed-point load-average
109  * counting. Some notes:
110  *  - 11 bit fractions expand to 22 bits by the multiplies: this gives
111  *    a load-average precision of 10 bits integer + 11 bits fractional
112  *  - if you want to count load-averages more often, you need more
113  *    precision, or rounding will get you. With 2-second counting freq,
114  *    the EXP_n values would be 1981, 2034 and 2043 if still using only
115  *    11 bit fractions.
116  */
117 extern unsigned long avenrun[];		/* Load averages */
118 
119 #define FSHIFT		11		/* nr of bits of precision */
120 #define FIXED_1		(1<<FSHIFT)	/* 1.0 as fixed-point */
121 #define LOAD_FREQ	(5*HZ+1)	/* 5 sec intervals */
122 #define EXP_1		1884		/* 1/exp(5sec/1min) as fixed-point */
123 #define EXP_5		2014		/* 1/exp(5sec/5min) */
124 #define EXP_15		2037		/* 1/exp(5sec/15min) */
125 
126 #define CALC_LOAD(load,exp,n) \
127 	load *= exp; \
128 	load += n*(FIXED_1-exp); \
129 	load >>= FSHIFT;
130 
131 extern unsigned long total_forks;
132 extern int nr_threads;
133 DECLARE_PER_CPU(unsigned long, process_counts);
134 extern int nr_processes(void);
135 extern unsigned long nr_running(void);
136 extern unsigned long nr_uninterruptible(void);
137 extern unsigned long nr_active(void);
138 extern unsigned long nr_iowait(void);
139 extern unsigned long weighted_cpuload(const int cpu);
140 
141 struct seq_file;
142 struct cfs_rq;
143 struct task_group;
144 #ifdef CONFIG_SCHED_DEBUG
145 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
146 extern void proc_sched_set_task(struct task_struct *p);
147 extern void
148 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
149 #else
150 static inline void
151 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
152 {
153 }
154 static inline void proc_sched_set_task(struct task_struct *p)
155 {
156 }
157 static inline void
158 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
159 {
160 }
161 #endif
162 
163 /*
164  * Task state bitmask. NOTE! These bits are also
165  * encoded in fs/proc/array.c: get_task_state().
166  *
167  * We have two separate sets of flags: task->state
168  * is about runnability, while task->exit_state are
169  * about the task exiting. Confusing, but this way
170  * modifying one set can't modify the other one by
171  * mistake.
172  */
173 #define TASK_RUNNING		0
174 #define TASK_INTERRUPTIBLE	1
175 #define TASK_UNINTERRUPTIBLE	2
176 #define __TASK_STOPPED		4
177 #define __TASK_TRACED		8
178 /* in tsk->exit_state */
179 #define EXIT_ZOMBIE		16
180 #define EXIT_DEAD		32
181 /* in tsk->state again */
182 #define TASK_DEAD		64
183 #define TASK_WAKEKILL		128
184 
185 /* Convenience macros for the sake of set_task_state */
186 #define TASK_KILLABLE		(TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
187 #define TASK_STOPPED		(TASK_WAKEKILL | __TASK_STOPPED)
188 #define TASK_TRACED		(TASK_WAKEKILL | __TASK_TRACED)
189 
190 /* Convenience macros for the sake of wake_up */
191 #define TASK_NORMAL		(TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
192 #define TASK_ALL		(TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
193 
194 /* get_task_state() */
195 #define TASK_REPORT		(TASK_RUNNING | TASK_INTERRUPTIBLE | \
196 				 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
197 				 __TASK_TRACED)
198 
199 #define task_is_traced(task)	((task->state & __TASK_TRACED) != 0)
200 #define task_is_stopped(task)	((task->state & __TASK_STOPPED) != 0)
201 #define task_is_stopped_or_traced(task)	\
202 			((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
203 #define task_contributes_to_load(task)	\
204 				((task->state & TASK_UNINTERRUPTIBLE) != 0)
205 
206 #define __set_task_state(tsk, state_value)		\
207 	do { (tsk)->state = (state_value); } while (0)
208 #define set_task_state(tsk, state_value)		\
209 	set_mb((tsk)->state, (state_value))
210 
211 /*
212  * set_current_state() includes a barrier so that the write of current->state
213  * is correctly serialised wrt the caller's subsequent test of whether to
214  * actually sleep:
215  *
216  *	set_current_state(TASK_UNINTERRUPTIBLE);
217  *	if (do_i_need_to_sleep())
218  *		schedule();
219  *
220  * If the caller does not need such serialisation then use __set_current_state()
221  */
222 #define __set_current_state(state_value)			\
223 	do { current->state = (state_value); } while (0)
224 #define set_current_state(state_value)		\
225 	set_mb(current->state, (state_value))
226 
227 /* Task command name length */
228 #define TASK_COMM_LEN 16
229 
230 #include <linux/spinlock.h>
231 
232 /*
233  * This serializes "schedule()" and also protects
234  * the run-queue from deletions/modifications (but
235  * _adding_ to the beginning of the run-queue has
236  * a separate lock).
237  */
238 extern rwlock_t tasklist_lock;
239 extern spinlock_t mmlist_lock;
240 
241 struct task_struct;
242 
243 extern void sched_init(void);
244 extern void sched_init_smp(void);
245 extern asmlinkage void schedule_tail(struct task_struct *prev);
246 extern void init_idle(struct task_struct *idle, int cpu);
247 extern void init_idle_bootup_task(struct task_struct *idle);
248 
249 extern cpumask_t nohz_cpu_mask;
250 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
251 extern int select_nohz_load_balancer(int cpu);
252 #else
253 static inline int select_nohz_load_balancer(int cpu)
254 {
255 	return 0;
256 }
257 #endif
258 
259 extern unsigned long rt_needs_cpu(int cpu);
260 
261 /*
262  * Only dump TASK_* tasks. (0 for all tasks)
263  */
264 extern void show_state_filter(unsigned long state_filter);
265 
266 static inline void show_state(void)
267 {
268 	show_state_filter(0);
269 }
270 
271 extern void show_regs(struct pt_regs *);
272 
273 /*
274  * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
275  * task), SP is the stack pointer of the first frame that should be shown in the back
276  * trace (or NULL if the entire call-chain of the task should be shown).
277  */
278 extern void show_stack(struct task_struct *task, unsigned long *sp);
279 
280 void io_schedule(void);
281 long io_schedule_timeout(long timeout);
282 
283 extern void cpu_init (void);
284 extern void trap_init(void);
285 extern void account_process_tick(struct task_struct *task, int user);
286 extern void update_process_times(int user);
287 extern void scheduler_tick(void);
288 extern void hrtick_resched(void);
289 
290 extern void sched_show_task(struct task_struct *p);
291 
292 #ifdef CONFIG_DETECT_SOFTLOCKUP
293 extern void softlockup_tick(void);
294 extern void spawn_softlockup_task(void);
295 extern void touch_softlockup_watchdog(void);
296 extern void touch_all_softlockup_watchdogs(void);
297 extern unsigned long  softlockup_thresh;
298 extern unsigned long sysctl_hung_task_check_count;
299 extern unsigned long sysctl_hung_task_timeout_secs;
300 extern unsigned long sysctl_hung_task_warnings;
301 #else
302 static inline void softlockup_tick(void)
303 {
304 }
305 static inline void spawn_softlockup_task(void)
306 {
307 }
308 static inline void touch_softlockup_watchdog(void)
309 {
310 }
311 static inline void touch_all_softlockup_watchdogs(void)
312 {
313 }
314 #endif
315 
316 
317 /* Attach to any functions which should be ignored in wchan output. */
318 #define __sched		__attribute__((__section__(".sched.text")))
319 
320 /* Linker adds these: start and end of __sched functions */
321 extern char __sched_text_start[], __sched_text_end[];
322 
323 /* Is this address in the __sched functions? */
324 extern int in_sched_functions(unsigned long addr);
325 
326 #define	MAX_SCHEDULE_TIMEOUT	LONG_MAX
327 extern signed long schedule_timeout(signed long timeout);
328 extern signed long schedule_timeout_interruptible(signed long timeout);
329 extern signed long schedule_timeout_killable(signed long timeout);
330 extern signed long schedule_timeout_uninterruptible(signed long timeout);
331 asmlinkage void schedule(void);
332 
333 struct nsproxy;
334 struct user_namespace;
335 
336 /* Maximum number of active map areas.. This is a random (large) number */
337 #define DEFAULT_MAX_MAP_COUNT	65536
338 
339 extern int sysctl_max_map_count;
340 
341 #include <linux/aio.h>
342 
343 extern unsigned long
344 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
345 		       unsigned long, unsigned long);
346 extern unsigned long
347 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
348 			  unsigned long len, unsigned long pgoff,
349 			  unsigned long flags);
350 extern void arch_unmap_area(struct mm_struct *, unsigned long);
351 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
352 
353 #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
354 /*
355  * The mm counters are not protected by its page_table_lock,
356  * so must be incremented atomically.
357  */
358 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
359 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
360 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
361 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
362 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
363 
364 #else  /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
365 /*
366  * The mm counters are protected by its page_table_lock,
367  * so can be incremented directly.
368  */
369 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
370 #define get_mm_counter(mm, member) ((mm)->_##member)
371 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
372 #define inc_mm_counter(mm, member) (mm)->_##member++
373 #define dec_mm_counter(mm, member) (mm)->_##member--
374 
375 #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
376 
377 #define get_mm_rss(mm)					\
378 	(get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
379 #define update_hiwater_rss(mm)	do {			\
380 	unsigned long _rss = get_mm_rss(mm);		\
381 	if ((mm)->hiwater_rss < _rss)			\
382 		(mm)->hiwater_rss = _rss;		\
383 } while (0)
384 #define update_hiwater_vm(mm)	do {			\
385 	if ((mm)->hiwater_vm < (mm)->total_vm)		\
386 		(mm)->hiwater_vm = (mm)->total_vm;	\
387 } while (0)
388 
389 extern void set_dumpable(struct mm_struct *mm, int value);
390 extern int get_dumpable(struct mm_struct *mm);
391 
392 /* mm flags */
393 /* dumpable bits */
394 #define MMF_DUMPABLE      0  /* core dump is permitted */
395 #define MMF_DUMP_SECURELY 1  /* core file is readable only by root */
396 #define MMF_DUMPABLE_BITS 2
397 
398 /* coredump filter bits */
399 #define MMF_DUMP_ANON_PRIVATE	2
400 #define MMF_DUMP_ANON_SHARED	3
401 #define MMF_DUMP_MAPPED_PRIVATE	4
402 #define MMF_DUMP_MAPPED_SHARED	5
403 #define MMF_DUMP_ELF_HEADERS	6
404 #define MMF_DUMP_FILTER_SHIFT	MMF_DUMPABLE_BITS
405 #define MMF_DUMP_FILTER_BITS	5
406 #define MMF_DUMP_FILTER_MASK \
407 	(((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
408 #define MMF_DUMP_FILTER_DEFAULT \
409 	((1 << MMF_DUMP_ANON_PRIVATE) |	(1 << MMF_DUMP_ANON_SHARED))
410 
411 struct sighand_struct {
412 	atomic_t		count;
413 	struct k_sigaction	action[_NSIG];
414 	spinlock_t		siglock;
415 	wait_queue_head_t	signalfd_wqh;
416 };
417 
418 struct pacct_struct {
419 	int			ac_flag;
420 	long			ac_exitcode;
421 	unsigned long		ac_mem;
422 	cputime_t		ac_utime, ac_stime;
423 	unsigned long		ac_minflt, ac_majflt;
424 };
425 
426 /*
427  * NOTE! "signal_struct" does not have it's own
428  * locking, because a shared signal_struct always
429  * implies a shared sighand_struct, so locking
430  * sighand_struct is always a proper superset of
431  * the locking of signal_struct.
432  */
433 struct signal_struct {
434 	atomic_t		count;
435 	atomic_t		live;
436 
437 	wait_queue_head_t	wait_chldexit;	/* for wait4() */
438 
439 	/* current thread group signal load-balancing target: */
440 	struct task_struct	*curr_target;
441 
442 	/* shared signal handling: */
443 	struct sigpending	shared_pending;
444 
445 	/* thread group exit support */
446 	int			group_exit_code;
447 	/* overloaded:
448 	 * - notify group_exit_task when ->count is equal to notify_count
449 	 * - everyone except group_exit_task is stopped during signal delivery
450 	 *   of fatal signals, group_exit_task processes the signal.
451 	 */
452 	struct task_struct	*group_exit_task;
453 	int			notify_count;
454 
455 	/* thread group stop support, overloads group_exit_code too */
456 	int			group_stop_count;
457 	unsigned int		flags; /* see SIGNAL_* flags below */
458 
459 	/* POSIX.1b Interval Timers */
460 	struct list_head posix_timers;
461 
462 	/* ITIMER_REAL timer for the process */
463 	struct hrtimer real_timer;
464 	struct pid *leader_pid;
465 	ktime_t it_real_incr;
466 
467 	/* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
468 	cputime_t it_prof_expires, it_virt_expires;
469 	cputime_t it_prof_incr, it_virt_incr;
470 
471 	/* job control IDs */
472 
473 	/*
474 	 * pgrp and session fields are deprecated.
475 	 * use the task_session_Xnr and task_pgrp_Xnr routines below
476 	 */
477 
478 	union {
479 		pid_t pgrp __deprecated;
480 		pid_t __pgrp;
481 	};
482 
483 	struct pid *tty_old_pgrp;
484 
485 	union {
486 		pid_t session __deprecated;
487 		pid_t __session;
488 	};
489 
490 	/* boolean value for session group leader */
491 	int leader;
492 
493 	struct tty_struct *tty; /* NULL if no tty */
494 
495 	/*
496 	 * Cumulative resource counters for dead threads in the group,
497 	 * and for reaped dead child processes forked by this group.
498 	 * Live threads maintain their own counters and add to these
499 	 * in __exit_signal, except for the group leader.
500 	 */
501 	cputime_t utime, stime, cutime, cstime;
502 	cputime_t gtime;
503 	cputime_t cgtime;
504 	unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
505 	unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
506 	unsigned long inblock, oublock, cinblock, coublock;
507 
508 	/*
509 	 * Cumulative ns of scheduled CPU time for dead threads in the
510 	 * group, not including a zombie group leader.  (This only differs
511 	 * from jiffies_to_ns(utime + stime) if sched_clock uses something
512 	 * other than jiffies.)
513 	 */
514 	unsigned long long sum_sched_runtime;
515 
516 	/*
517 	 * We don't bother to synchronize most readers of this at all,
518 	 * because there is no reader checking a limit that actually needs
519 	 * to get both rlim_cur and rlim_max atomically, and either one
520 	 * alone is a single word that can safely be read normally.
521 	 * getrlimit/setrlimit use task_lock(current->group_leader) to
522 	 * protect this instead of the siglock, because they really
523 	 * have no need to disable irqs.
524 	 */
525 	struct rlimit rlim[RLIM_NLIMITS];
526 
527 	struct list_head cpu_timers[3];
528 
529 	/* keep the process-shared keyrings here so that they do the right
530 	 * thing in threads created with CLONE_THREAD */
531 #ifdef CONFIG_KEYS
532 	struct key *session_keyring;	/* keyring inherited over fork */
533 	struct key *process_keyring;	/* keyring private to this process */
534 #endif
535 #ifdef CONFIG_BSD_PROCESS_ACCT
536 	struct pacct_struct pacct;	/* per-process accounting information */
537 #endif
538 #ifdef CONFIG_TASKSTATS
539 	struct taskstats *stats;
540 #endif
541 #ifdef CONFIG_AUDIT
542 	unsigned audit_tty;
543 	struct tty_audit_buf *tty_audit_buf;
544 #endif
545 };
546 
547 /* Context switch must be unlocked if interrupts are to be enabled */
548 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
549 # define __ARCH_WANT_UNLOCKED_CTXSW
550 #endif
551 
552 /*
553  * Bits in flags field of signal_struct.
554  */
555 #define SIGNAL_STOP_STOPPED	0x00000001 /* job control stop in effect */
556 #define SIGNAL_STOP_DEQUEUED	0x00000002 /* stop signal dequeued */
557 #define SIGNAL_STOP_CONTINUED	0x00000004 /* SIGCONT since WCONTINUED reap */
558 #define SIGNAL_GROUP_EXIT	0x00000008 /* group exit in progress */
559 
560 /* If true, all threads except ->group_exit_task have pending SIGKILL */
561 static inline int signal_group_exit(const struct signal_struct *sig)
562 {
563 	return	(sig->flags & SIGNAL_GROUP_EXIT) ||
564 		(sig->group_exit_task != NULL);
565 }
566 
567 /*
568  * Some day this will be a full-fledged user tracking system..
569  */
570 struct user_struct {
571 	atomic_t __count;	/* reference count */
572 	atomic_t processes;	/* How many processes does this user have? */
573 	atomic_t files;		/* How many open files does this user have? */
574 	atomic_t sigpending;	/* How many pending signals does this user have? */
575 #ifdef CONFIG_INOTIFY_USER
576 	atomic_t inotify_watches; /* How many inotify watches does this user have? */
577 	atomic_t inotify_devs;	/* How many inotify devs does this user have opened? */
578 #endif
579 #ifdef CONFIG_POSIX_MQUEUE
580 	/* protected by mq_lock	*/
581 	unsigned long mq_bytes;	/* How many bytes can be allocated to mqueue? */
582 #endif
583 	unsigned long locked_shm; /* How many pages of mlocked shm ? */
584 
585 #ifdef CONFIG_KEYS
586 	struct key *uid_keyring;	/* UID specific keyring */
587 	struct key *session_keyring;	/* UID's default session keyring */
588 #endif
589 
590 	/* Hash table maintenance information */
591 	struct hlist_node uidhash_node;
592 	uid_t uid;
593 
594 #ifdef CONFIG_USER_SCHED
595 	struct task_group *tg;
596 #ifdef CONFIG_SYSFS
597 	struct kobject kobj;
598 	struct work_struct work;
599 #endif
600 #endif
601 };
602 
603 extern int uids_sysfs_init(void);
604 
605 extern struct user_struct *find_user(uid_t);
606 
607 extern struct user_struct root_user;
608 #define INIT_USER (&root_user)
609 
610 struct backing_dev_info;
611 struct reclaim_state;
612 
613 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
614 struct sched_info {
615 	/* cumulative counters */
616 	unsigned long pcount;	      /* # of times run on this cpu */
617 	unsigned long long cpu_time,  /* time spent on the cpu */
618 			   run_delay; /* time spent waiting on a runqueue */
619 
620 	/* timestamps */
621 	unsigned long long last_arrival,/* when we last ran on a cpu */
622 			   last_queued;	/* when we were last queued to run */
623 #ifdef CONFIG_SCHEDSTATS
624 	/* BKL stats */
625 	unsigned int bkl_count;
626 #endif
627 };
628 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
629 
630 #ifdef CONFIG_SCHEDSTATS
631 extern const struct file_operations proc_schedstat_operations;
632 #endif /* CONFIG_SCHEDSTATS */
633 
634 #ifdef CONFIG_TASK_DELAY_ACCT
635 struct task_delay_info {
636 	spinlock_t	lock;
637 	unsigned int	flags;	/* Private per-task flags */
638 
639 	/* For each stat XXX, add following, aligned appropriately
640 	 *
641 	 * struct timespec XXX_start, XXX_end;
642 	 * u64 XXX_delay;
643 	 * u32 XXX_count;
644 	 *
645 	 * Atomicity of updates to XXX_delay, XXX_count protected by
646 	 * single lock above (split into XXX_lock if contention is an issue).
647 	 */
648 
649 	/*
650 	 * XXX_count is incremented on every XXX operation, the delay
651 	 * associated with the operation is added to XXX_delay.
652 	 * XXX_delay contains the accumulated delay time in nanoseconds.
653 	 */
654 	struct timespec blkio_start, blkio_end;	/* Shared by blkio, swapin */
655 	u64 blkio_delay;	/* wait for sync block io completion */
656 	u64 swapin_delay;	/* wait for swapin block io completion */
657 	u32 blkio_count;	/* total count of the number of sync block */
658 				/* io operations performed */
659 	u32 swapin_count;	/* total count of the number of swapin block */
660 				/* io operations performed */
661 };
662 #endif	/* CONFIG_TASK_DELAY_ACCT */
663 
664 static inline int sched_info_on(void)
665 {
666 #ifdef CONFIG_SCHEDSTATS
667 	return 1;
668 #elif defined(CONFIG_TASK_DELAY_ACCT)
669 	extern int delayacct_on;
670 	return delayacct_on;
671 #else
672 	return 0;
673 #endif
674 }
675 
676 enum cpu_idle_type {
677 	CPU_IDLE,
678 	CPU_NOT_IDLE,
679 	CPU_NEWLY_IDLE,
680 	CPU_MAX_IDLE_TYPES
681 };
682 
683 /*
684  * sched-domains (multiprocessor balancing) declarations:
685  */
686 
687 /*
688  * Increase resolution of nice-level calculations:
689  */
690 #define SCHED_LOAD_SHIFT	10
691 #define SCHED_LOAD_SCALE	(1L << SCHED_LOAD_SHIFT)
692 
693 #define SCHED_LOAD_SCALE_FUZZ	SCHED_LOAD_SCALE
694 
695 #ifdef CONFIG_SMP
696 #define SD_LOAD_BALANCE		1	/* Do load balancing on this domain. */
697 #define SD_BALANCE_NEWIDLE	2	/* Balance when about to become idle */
698 #define SD_BALANCE_EXEC		4	/* Balance on exec */
699 #define SD_BALANCE_FORK		8	/* Balance on fork, clone */
700 #define SD_WAKE_IDLE		16	/* Wake to idle CPU on task wakeup */
701 #define SD_WAKE_AFFINE		32	/* Wake task to waking CPU */
702 #define SD_WAKE_BALANCE		64	/* Perform balancing at task wakeup */
703 #define SD_SHARE_CPUPOWER	128	/* Domain members share cpu power */
704 #define SD_POWERSAVINGS_BALANCE	256	/* Balance for power savings */
705 #define SD_SHARE_PKG_RESOURCES	512	/* Domain members share cpu pkg resources */
706 #define SD_SERIALIZE		1024	/* Only a single load balancing instance */
707 
708 #define BALANCE_FOR_MC_POWER	\
709 	(sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
710 
711 #define BALANCE_FOR_PKG_POWER	\
712 	((sched_mc_power_savings || sched_smt_power_savings) ?	\
713 	 SD_POWERSAVINGS_BALANCE : 0)
714 
715 #define test_sd_parent(sd, flag)	((sd->parent &&		\
716 					 (sd->parent->flags & flag)) ? 1 : 0)
717 
718 
719 struct sched_group {
720 	struct sched_group *next;	/* Must be a circular list */
721 	cpumask_t cpumask;
722 
723 	/*
724 	 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
725 	 * single CPU. This is read only (except for setup, hotplug CPU).
726 	 * Note : Never change cpu_power without recompute its reciprocal
727 	 */
728 	unsigned int __cpu_power;
729 	/*
730 	 * reciprocal value of cpu_power to avoid expensive divides
731 	 * (see include/linux/reciprocal_div.h)
732 	 */
733 	u32 reciprocal_cpu_power;
734 };
735 
736 struct sched_domain {
737 	/* These fields must be setup */
738 	struct sched_domain *parent;	/* top domain must be null terminated */
739 	struct sched_domain *child;	/* bottom domain must be null terminated */
740 	struct sched_group *groups;	/* the balancing groups of the domain */
741 	cpumask_t span;			/* span of all CPUs in this domain */
742 	unsigned long min_interval;	/* Minimum balance interval ms */
743 	unsigned long max_interval;	/* Maximum balance interval ms */
744 	unsigned int busy_factor;	/* less balancing by factor if busy */
745 	unsigned int imbalance_pct;	/* No balance until over watermark */
746 	unsigned int cache_nice_tries;	/* Leave cache hot tasks for # tries */
747 	unsigned int busy_idx;
748 	unsigned int idle_idx;
749 	unsigned int newidle_idx;
750 	unsigned int wake_idx;
751 	unsigned int forkexec_idx;
752 	int flags;			/* See SD_* */
753 
754 	/* Runtime fields. */
755 	unsigned long last_balance;	/* init to jiffies. units in jiffies */
756 	unsigned int balance_interval;	/* initialise to 1. units in ms. */
757 	unsigned int nr_balance_failed; /* initialise to 0 */
758 
759 #ifdef CONFIG_SCHEDSTATS
760 	/* load_balance() stats */
761 	unsigned int lb_count[CPU_MAX_IDLE_TYPES];
762 	unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
763 	unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
764 	unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
765 	unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
766 	unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
767 	unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
768 	unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
769 
770 	/* Active load balancing */
771 	unsigned int alb_count;
772 	unsigned int alb_failed;
773 	unsigned int alb_pushed;
774 
775 	/* SD_BALANCE_EXEC stats */
776 	unsigned int sbe_count;
777 	unsigned int sbe_balanced;
778 	unsigned int sbe_pushed;
779 
780 	/* SD_BALANCE_FORK stats */
781 	unsigned int sbf_count;
782 	unsigned int sbf_balanced;
783 	unsigned int sbf_pushed;
784 
785 	/* try_to_wake_up() stats */
786 	unsigned int ttwu_wake_remote;
787 	unsigned int ttwu_move_affine;
788 	unsigned int ttwu_move_balance;
789 #endif
790 };
791 
792 extern void partition_sched_domains(int ndoms_new, cpumask_t *doms_new);
793 extern int arch_reinit_sched_domains(void);
794 
795 #endif	/* CONFIG_SMP */
796 
797 /*
798  * A runqueue laden with a single nice 0 task scores a weighted_cpuload of
799  * SCHED_LOAD_SCALE. This function returns 1 if any cpu is laden with a
800  * task of nice 0 or enough lower priority tasks to bring up the
801  * weighted_cpuload
802  */
803 static inline int above_background_load(void)
804 {
805 	unsigned long cpu;
806 
807 	for_each_online_cpu(cpu) {
808 		if (weighted_cpuload(cpu) >= SCHED_LOAD_SCALE)
809 			return 1;
810 	}
811 	return 0;
812 }
813 
814 struct io_context;			/* See blkdev.h */
815 #define NGROUPS_SMALL		32
816 #define NGROUPS_PER_BLOCK	((unsigned int)(PAGE_SIZE / sizeof(gid_t)))
817 struct group_info {
818 	int ngroups;
819 	atomic_t usage;
820 	gid_t small_block[NGROUPS_SMALL];
821 	int nblocks;
822 	gid_t *blocks[0];
823 };
824 
825 /*
826  * get_group_info() must be called with the owning task locked (via task_lock())
827  * when task != current.  The reason being that the vast majority of callers are
828  * looking at current->group_info, which can not be changed except by the
829  * current task.  Changing current->group_info requires the task lock, too.
830  */
831 #define get_group_info(group_info) do { \
832 	atomic_inc(&(group_info)->usage); \
833 } while (0)
834 
835 #define put_group_info(group_info) do { \
836 	if (atomic_dec_and_test(&(group_info)->usage)) \
837 		groups_free(group_info); \
838 } while (0)
839 
840 extern struct group_info *groups_alloc(int gidsetsize);
841 extern void groups_free(struct group_info *group_info);
842 extern int set_current_groups(struct group_info *group_info);
843 extern int groups_search(struct group_info *group_info, gid_t grp);
844 /* access the groups "array" with this macro */
845 #define GROUP_AT(gi, i) \
846     ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
847 
848 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
849 extern void prefetch_stack(struct task_struct *t);
850 #else
851 static inline void prefetch_stack(struct task_struct *t) { }
852 #endif
853 
854 struct audit_context;		/* See audit.c */
855 struct mempolicy;
856 struct pipe_inode_info;
857 struct uts_namespace;
858 
859 struct rq;
860 struct sched_domain;
861 
862 struct sched_class {
863 	const struct sched_class *next;
864 
865 	void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
866 	void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
867 	void (*yield_task) (struct rq *rq);
868 	int  (*select_task_rq)(struct task_struct *p, int sync);
869 
870 	void (*check_preempt_curr) (struct rq *rq, struct task_struct *p);
871 
872 	struct task_struct * (*pick_next_task) (struct rq *rq);
873 	void (*put_prev_task) (struct rq *rq, struct task_struct *p);
874 
875 #ifdef CONFIG_SMP
876 	unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
877 			struct rq *busiest, unsigned long max_load_move,
878 			struct sched_domain *sd, enum cpu_idle_type idle,
879 			int *all_pinned, int *this_best_prio);
880 
881 	int (*move_one_task) (struct rq *this_rq, int this_cpu,
882 			      struct rq *busiest, struct sched_domain *sd,
883 			      enum cpu_idle_type idle);
884 	void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
885 	void (*post_schedule) (struct rq *this_rq);
886 	void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
887 #endif
888 
889 	void (*set_curr_task) (struct rq *rq);
890 	void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
891 	void (*task_new) (struct rq *rq, struct task_struct *p);
892 	void (*set_cpus_allowed)(struct task_struct *p, cpumask_t *newmask);
893 
894 	void (*join_domain)(struct rq *rq);
895 	void (*leave_domain)(struct rq *rq);
896 
897 	void (*switched_from) (struct rq *this_rq, struct task_struct *task,
898 			       int running);
899 	void (*switched_to) (struct rq *this_rq, struct task_struct *task,
900 			     int running);
901 	void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
902 			     int oldprio, int running);
903 
904 #ifdef CONFIG_FAIR_GROUP_SCHED
905 	void (*moved_group) (struct task_struct *p);
906 #endif
907 };
908 
909 struct load_weight {
910 	unsigned long weight, inv_weight;
911 };
912 
913 /*
914  * CFS stats for a schedulable entity (task, task-group etc)
915  *
916  * Current field usage histogram:
917  *
918  *     4 se->block_start
919  *     4 se->run_node
920  *     4 se->sleep_start
921  *     6 se->load.weight
922  */
923 struct sched_entity {
924 	struct load_weight	load;		/* for load-balancing */
925 	struct rb_node		run_node;
926 	unsigned int		on_rq;
927 
928 	u64			exec_start;
929 	u64			sum_exec_runtime;
930 	u64			vruntime;
931 	u64			prev_sum_exec_runtime;
932 
933 	u64			last_wakeup;
934 	u64			avg_overlap;
935 
936 #ifdef CONFIG_SCHEDSTATS
937 	u64			wait_start;
938 	u64			wait_max;
939 	u64			wait_count;
940 	u64			wait_sum;
941 
942 	u64			sleep_start;
943 	u64			sleep_max;
944 	s64			sum_sleep_runtime;
945 
946 	u64			block_start;
947 	u64			block_max;
948 	u64			exec_max;
949 	u64			slice_max;
950 
951 	u64			nr_migrations;
952 	u64			nr_migrations_cold;
953 	u64			nr_failed_migrations_affine;
954 	u64			nr_failed_migrations_running;
955 	u64			nr_failed_migrations_hot;
956 	u64			nr_forced_migrations;
957 	u64			nr_forced2_migrations;
958 
959 	u64			nr_wakeups;
960 	u64			nr_wakeups_sync;
961 	u64			nr_wakeups_migrate;
962 	u64			nr_wakeups_local;
963 	u64			nr_wakeups_remote;
964 	u64			nr_wakeups_affine;
965 	u64			nr_wakeups_affine_attempts;
966 	u64			nr_wakeups_passive;
967 	u64			nr_wakeups_idle;
968 #endif
969 
970 #ifdef CONFIG_FAIR_GROUP_SCHED
971 	struct sched_entity	*parent;
972 	/* rq on which this entity is (to be) queued: */
973 	struct cfs_rq		*cfs_rq;
974 	/* rq "owned" by this entity/group: */
975 	struct cfs_rq		*my_q;
976 #endif
977 };
978 
979 struct sched_rt_entity {
980 	struct list_head run_list;
981 	unsigned int time_slice;
982 	unsigned long timeout;
983 	int nr_cpus_allowed;
984 
985 #ifdef CONFIG_RT_GROUP_SCHED
986 	struct sched_rt_entity	*parent;
987 	/* rq on which this entity is (to be) queued: */
988 	struct rt_rq		*rt_rq;
989 	/* rq "owned" by this entity/group: */
990 	struct rt_rq		*my_q;
991 #endif
992 };
993 
994 struct task_struct {
995 	volatile long state;	/* -1 unrunnable, 0 runnable, >0 stopped */
996 	void *stack;
997 	atomic_t usage;
998 	unsigned int flags;	/* per process flags, defined below */
999 	unsigned int ptrace;
1000 
1001 	int lock_depth;		/* BKL lock depth */
1002 
1003 #ifdef CONFIG_SMP
1004 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1005 	int oncpu;
1006 #endif
1007 #endif
1008 
1009 	int prio, static_prio, normal_prio;
1010 	const struct sched_class *sched_class;
1011 	struct sched_entity se;
1012 	struct sched_rt_entity rt;
1013 
1014 #ifdef CONFIG_PREEMPT_NOTIFIERS
1015 	/* list of struct preempt_notifier: */
1016 	struct hlist_head preempt_notifiers;
1017 #endif
1018 
1019 	/*
1020 	 * fpu_counter contains the number of consecutive context switches
1021 	 * that the FPU is used. If this is over a threshold, the lazy fpu
1022 	 * saving becomes unlazy to save the trap. This is an unsigned char
1023 	 * so that after 256 times the counter wraps and the behavior turns
1024 	 * lazy again; this to deal with bursty apps that only use FPU for
1025 	 * a short time
1026 	 */
1027 	unsigned char fpu_counter;
1028 	s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1029 #ifdef CONFIG_BLK_DEV_IO_TRACE
1030 	unsigned int btrace_seq;
1031 #endif
1032 
1033 	unsigned int policy;
1034 	cpumask_t cpus_allowed;
1035 
1036 #ifdef CONFIG_PREEMPT_RCU
1037 	int rcu_read_lock_nesting;
1038 	int rcu_flipctr_idx;
1039 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1040 
1041 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1042 	struct sched_info sched_info;
1043 #endif
1044 
1045 	struct list_head tasks;
1046 	/*
1047 	 * ptrace_list/ptrace_children forms the list of my children
1048 	 * that were stolen by a ptracer.
1049 	 */
1050 	struct list_head ptrace_children;
1051 	struct list_head ptrace_list;
1052 
1053 	struct mm_struct *mm, *active_mm;
1054 
1055 /* task state */
1056 	struct linux_binfmt *binfmt;
1057 	int exit_state;
1058 	int exit_code, exit_signal;
1059 	int pdeath_signal;  /*  The signal sent when the parent dies  */
1060 	/* ??? */
1061 	unsigned int personality;
1062 	unsigned did_exec:1;
1063 	pid_t pid;
1064 	pid_t tgid;
1065 
1066 #ifdef CONFIG_CC_STACKPROTECTOR
1067 	/* Canary value for the -fstack-protector gcc feature */
1068 	unsigned long stack_canary;
1069 #endif
1070 	/*
1071 	 * pointers to (original) parent process, youngest child, younger sibling,
1072 	 * older sibling, respectively.  (p->father can be replaced with
1073 	 * p->parent->pid)
1074 	 */
1075 	struct task_struct *real_parent; /* real parent process (when being debugged) */
1076 	struct task_struct *parent;	/* parent process */
1077 	/*
1078 	 * children/sibling forms the list of my children plus the
1079 	 * tasks I'm ptracing.
1080 	 */
1081 	struct list_head children;	/* list of my children */
1082 	struct list_head sibling;	/* linkage in my parent's children list */
1083 	struct task_struct *group_leader;	/* threadgroup leader */
1084 
1085 	/* PID/PID hash table linkage. */
1086 	struct pid_link pids[PIDTYPE_MAX];
1087 	struct list_head thread_group;
1088 
1089 	struct completion *vfork_done;		/* for vfork() */
1090 	int __user *set_child_tid;		/* CLONE_CHILD_SETTID */
1091 	int __user *clear_child_tid;		/* CLONE_CHILD_CLEARTID */
1092 
1093 	unsigned int rt_priority;
1094 	cputime_t utime, stime, utimescaled, stimescaled;
1095 	cputime_t gtime;
1096 	cputime_t prev_utime, prev_stime;
1097 	unsigned long nvcsw, nivcsw; /* context switch counts */
1098 	struct timespec start_time; 		/* monotonic time */
1099 	struct timespec real_start_time;	/* boot based time */
1100 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1101 	unsigned long min_flt, maj_flt;
1102 
1103   	cputime_t it_prof_expires, it_virt_expires;
1104 	unsigned long long it_sched_expires;
1105 	struct list_head cpu_timers[3];
1106 
1107 /* process credentials */
1108 	uid_t uid,euid,suid,fsuid;
1109 	gid_t gid,egid,sgid,fsgid;
1110 	struct group_info *group_info;
1111 	kernel_cap_t   cap_effective, cap_inheritable, cap_permitted, cap_bset;
1112 	unsigned keep_capabilities:1;
1113 	struct user_struct *user;
1114 #ifdef CONFIG_KEYS
1115 	struct key *request_key_auth;	/* assumed request_key authority */
1116 	struct key *thread_keyring;	/* keyring private to this thread */
1117 	unsigned char jit_keyring;	/* default keyring to attach requested keys to */
1118 #endif
1119 	char comm[TASK_COMM_LEN]; /* executable name excluding path
1120 				     - access with [gs]et_task_comm (which lock
1121 				       it with task_lock())
1122 				     - initialized normally by flush_old_exec */
1123 /* file system info */
1124 	int link_count, total_link_count;
1125 #ifdef CONFIG_SYSVIPC
1126 /* ipc stuff */
1127 	struct sysv_sem sysvsem;
1128 #endif
1129 #ifdef CONFIG_DETECT_SOFTLOCKUP
1130 /* hung task detection */
1131 	unsigned long last_switch_timestamp;
1132 	unsigned long last_switch_count;
1133 #endif
1134 /* CPU-specific state of this task */
1135 	struct thread_struct thread;
1136 /* filesystem information */
1137 	struct fs_struct *fs;
1138 /* open file information */
1139 	struct files_struct *files;
1140 /* namespaces */
1141 	struct nsproxy *nsproxy;
1142 /* signal handlers */
1143 	struct signal_struct *signal;
1144 	struct sighand_struct *sighand;
1145 
1146 	sigset_t blocked, real_blocked;
1147 	sigset_t saved_sigmask;		/* To be restored with TIF_RESTORE_SIGMASK */
1148 	struct sigpending pending;
1149 
1150 	unsigned long sas_ss_sp;
1151 	size_t sas_ss_size;
1152 	int (*notifier)(void *priv);
1153 	void *notifier_data;
1154 	sigset_t *notifier_mask;
1155 #ifdef CONFIG_SECURITY
1156 	void *security;
1157 #endif
1158 	struct audit_context *audit_context;
1159 #ifdef CONFIG_AUDITSYSCALL
1160 	uid_t loginuid;
1161 	unsigned int sessionid;
1162 #endif
1163 	seccomp_t seccomp;
1164 
1165 /* Thread group tracking */
1166    	u32 parent_exec_id;
1167    	u32 self_exec_id;
1168 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1169 	spinlock_t alloc_lock;
1170 
1171 	/* Protection of the PI data structures: */
1172 	spinlock_t pi_lock;
1173 
1174 #ifdef CONFIG_RT_MUTEXES
1175 	/* PI waiters blocked on a rt_mutex held by this task */
1176 	struct plist_head pi_waiters;
1177 	/* Deadlock detection and priority inheritance handling */
1178 	struct rt_mutex_waiter *pi_blocked_on;
1179 #endif
1180 
1181 #ifdef CONFIG_DEBUG_MUTEXES
1182 	/* mutex deadlock detection */
1183 	struct mutex_waiter *blocked_on;
1184 #endif
1185 #ifdef CONFIG_TRACE_IRQFLAGS
1186 	unsigned int irq_events;
1187 	int hardirqs_enabled;
1188 	unsigned long hardirq_enable_ip;
1189 	unsigned int hardirq_enable_event;
1190 	unsigned long hardirq_disable_ip;
1191 	unsigned int hardirq_disable_event;
1192 	int softirqs_enabled;
1193 	unsigned long softirq_disable_ip;
1194 	unsigned int softirq_disable_event;
1195 	unsigned long softirq_enable_ip;
1196 	unsigned int softirq_enable_event;
1197 	int hardirq_context;
1198 	int softirq_context;
1199 #endif
1200 #ifdef CONFIG_LOCKDEP
1201 # define MAX_LOCK_DEPTH 48UL
1202 	u64 curr_chain_key;
1203 	int lockdep_depth;
1204 	struct held_lock held_locks[MAX_LOCK_DEPTH];
1205 	unsigned int lockdep_recursion;
1206 #endif
1207 
1208 /* journalling filesystem info */
1209 	void *journal_info;
1210 
1211 /* stacked block device info */
1212 	struct bio *bio_list, **bio_tail;
1213 
1214 /* VM state */
1215 	struct reclaim_state *reclaim_state;
1216 
1217 	struct backing_dev_info *backing_dev_info;
1218 
1219 	struct io_context *io_context;
1220 
1221 	unsigned long ptrace_message;
1222 	siginfo_t *last_siginfo; /* For ptrace use.  */
1223 #ifdef CONFIG_TASK_XACCT
1224 /* i/o counters(bytes read/written, #syscalls */
1225 	u64 rchar, wchar, syscr, syscw;
1226 #endif
1227 	struct task_io_accounting ioac;
1228 #if defined(CONFIG_TASK_XACCT)
1229 	u64 acct_rss_mem1;	/* accumulated rss usage */
1230 	u64 acct_vm_mem1;	/* accumulated virtual memory usage */
1231 	cputime_t acct_stimexpd;/* stime since last update */
1232 #endif
1233 #ifdef CONFIG_NUMA
1234   	struct mempolicy *mempolicy;
1235 	short il_next;
1236 #endif
1237 #ifdef CONFIG_CPUSETS
1238 	nodemask_t mems_allowed;
1239 	int cpuset_mems_generation;
1240 	int cpuset_mem_spread_rotor;
1241 #endif
1242 #ifdef CONFIG_CGROUPS
1243 	/* Control Group info protected by css_set_lock */
1244 	struct css_set *cgroups;
1245 	/* cg_list protected by css_set_lock and tsk->alloc_lock */
1246 	struct list_head cg_list;
1247 #endif
1248 #ifdef CONFIG_FUTEX
1249 	struct robust_list_head __user *robust_list;
1250 #ifdef CONFIG_COMPAT
1251 	struct compat_robust_list_head __user *compat_robust_list;
1252 #endif
1253 	struct list_head pi_state_list;
1254 	struct futex_pi_state *pi_state_cache;
1255 #endif
1256 	atomic_t fs_excl;	/* holding fs exclusive resources */
1257 	struct rcu_head rcu;
1258 
1259 	/*
1260 	 * cache last used pipe for splice
1261 	 */
1262 	struct pipe_inode_info *splice_pipe;
1263 #ifdef	CONFIG_TASK_DELAY_ACCT
1264 	struct task_delay_info *delays;
1265 #endif
1266 #ifdef CONFIG_FAULT_INJECTION
1267 	int make_it_fail;
1268 #endif
1269 	struct prop_local_single dirties;
1270 #ifdef CONFIG_LATENCYTOP
1271 	int latency_record_count;
1272 	struct latency_record latency_record[LT_SAVECOUNT];
1273 #endif
1274 };
1275 
1276 /*
1277  * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1278  * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1279  * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1280  * values are inverted: lower p->prio value means higher priority.
1281  *
1282  * The MAX_USER_RT_PRIO value allows the actual maximum
1283  * RT priority to be separate from the value exported to
1284  * user-space.  This allows kernel threads to set their
1285  * priority to a value higher than any user task. Note:
1286  * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1287  */
1288 
1289 #define MAX_USER_RT_PRIO	100
1290 #define MAX_RT_PRIO		MAX_USER_RT_PRIO
1291 
1292 #define MAX_PRIO		(MAX_RT_PRIO + 40)
1293 #define DEFAULT_PRIO		(MAX_RT_PRIO + 20)
1294 
1295 static inline int rt_prio(int prio)
1296 {
1297 	if (unlikely(prio < MAX_RT_PRIO))
1298 		return 1;
1299 	return 0;
1300 }
1301 
1302 static inline int rt_task(struct task_struct *p)
1303 {
1304 	return rt_prio(p->prio);
1305 }
1306 
1307 static inline void set_task_session(struct task_struct *tsk, pid_t session)
1308 {
1309 	tsk->signal->__session = session;
1310 }
1311 
1312 static inline void set_task_pgrp(struct task_struct *tsk, pid_t pgrp)
1313 {
1314 	tsk->signal->__pgrp = pgrp;
1315 }
1316 
1317 static inline struct pid *task_pid(struct task_struct *task)
1318 {
1319 	return task->pids[PIDTYPE_PID].pid;
1320 }
1321 
1322 static inline struct pid *task_tgid(struct task_struct *task)
1323 {
1324 	return task->group_leader->pids[PIDTYPE_PID].pid;
1325 }
1326 
1327 static inline struct pid *task_pgrp(struct task_struct *task)
1328 {
1329 	return task->group_leader->pids[PIDTYPE_PGID].pid;
1330 }
1331 
1332 static inline struct pid *task_session(struct task_struct *task)
1333 {
1334 	return task->group_leader->pids[PIDTYPE_SID].pid;
1335 }
1336 
1337 struct pid_namespace;
1338 
1339 /*
1340  * the helpers to get the task's different pids as they are seen
1341  * from various namespaces
1342  *
1343  * task_xid_nr()     : global id, i.e. the id seen from the init namespace;
1344  * task_xid_vnr()    : virtual id, i.e. the id seen from the pid namespace of
1345  *                     current.
1346  * task_xid_nr_ns()  : id seen from the ns specified;
1347  *
1348  * set_task_vxid()   : assigns a virtual id to a task;
1349  *
1350  * see also pid_nr() etc in include/linux/pid.h
1351  */
1352 
1353 static inline pid_t task_pid_nr(struct task_struct *tsk)
1354 {
1355 	return tsk->pid;
1356 }
1357 
1358 pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1359 
1360 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1361 {
1362 	return pid_vnr(task_pid(tsk));
1363 }
1364 
1365 
1366 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1367 {
1368 	return tsk->tgid;
1369 }
1370 
1371 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1372 
1373 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1374 {
1375 	return pid_vnr(task_tgid(tsk));
1376 }
1377 
1378 
1379 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1380 {
1381 	return tsk->signal->__pgrp;
1382 }
1383 
1384 pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1385 
1386 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1387 {
1388 	return pid_vnr(task_pgrp(tsk));
1389 }
1390 
1391 
1392 static inline pid_t task_session_nr(struct task_struct *tsk)
1393 {
1394 	return tsk->signal->__session;
1395 }
1396 
1397 pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1398 
1399 static inline pid_t task_session_vnr(struct task_struct *tsk)
1400 {
1401 	return pid_vnr(task_session(tsk));
1402 }
1403 
1404 
1405 /**
1406  * pid_alive - check that a task structure is not stale
1407  * @p: Task structure to be checked.
1408  *
1409  * Test if a process is not yet dead (at most zombie state)
1410  * If pid_alive fails, then pointers within the task structure
1411  * can be stale and must not be dereferenced.
1412  */
1413 static inline int pid_alive(struct task_struct *p)
1414 {
1415 	return p->pids[PIDTYPE_PID].pid != NULL;
1416 }
1417 
1418 /**
1419  * is_global_init - check if a task structure is init
1420  * @tsk: Task structure to be checked.
1421  *
1422  * Check if a task structure is the first user space task the kernel created.
1423  */
1424 static inline int is_global_init(struct task_struct *tsk)
1425 {
1426 	return tsk->pid == 1;
1427 }
1428 
1429 /*
1430  * is_container_init:
1431  * check whether in the task is init in its own pid namespace.
1432  */
1433 extern int is_container_init(struct task_struct *tsk);
1434 
1435 extern struct pid *cad_pid;
1436 
1437 extern void free_task(struct task_struct *tsk);
1438 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1439 
1440 extern void __put_task_struct(struct task_struct *t);
1441 
1442 static inline void put_task_struct(struct task_struct *t)
1443 {
1444 	if (atomic_dec_and_test(&t->usage))
1445 		__put_task_struct(t);
1446 }
1447 
1448 /*
1449  * Per process flags
1450  */
1451 #define PF_ALIGNWARN	0x00000001	/* Print alignment warning msgs */
1452 					/* Not implemented yet, only for 486*/
1453 #define PF_STARTING	0x00000002	/* being created */
1454 #define PF_EXITING	0x00000004	/* getting shut down */
1455 #define PF_EXITPIDONE	0x00000008	/* pi exit done on shut down */
1456 #define PF_VCPU		0x00000010	/* I'm a virtual CPU */
1457 #define PF_FORKNOEXEC	0x00000040	/* forked but didn't exec */
1458 #define PF_SUPERPRIV	0x00000100	/* used super-user privileges */
1459 #define PF_DUMPCORE	0x00000200	/* dumped core */
1460 #define PF_SIGNALED	0x00000400	/* killed by a signal */
1461 #define PF_MEMALLOC	0x00000800	/* Allocating memory */
1462 #define PF_FLUSHER	0x00001000	/* responsible for disk writeback */
1463 #define PF_USED_MATH	0x00002000	/* if unset the fpu must be initialized before use */
1464 #define PF_NOFREEZE	0x00008000	/* this thread should not be frozen */
1465 #define PF_FROZEN	0x00010000	/* frozen for system suspend */
1466 #define PF_FSTRANS	0x00020000	/* inside a filesystem transaction */
1467 #define PF_KSWAPD	0x00040000	/* I am kswapd */
1468 #define PF_SWAPOFF	0x00080000	/* I am in swapoff */
1469 #define PF_LESS_THROTTLE 0x00100000	/* Throttle me less: I clean memory */
1470 #define PF_BORROWED_MM	0x00200000	/* I am a kthread doing use_mm */
1471 #define PF_RANDOMIZE	0x00400000	/* randomize virtual address space */
1472 #define PF_SWAPWRITE	0x00800000	/* Allowed to write to swap */
1473 #define PF_SPREAD_PAGE	0x01000000	/* Spread page cache over cpuset */
1474 #define PF_SPREAD_SLAB	0x02000000	/* Spread some slab caches over cpuset */
1475 #define PF_MEMPOLICY	0x10000000	/* Non-default NUMA mempolicy */
1476 #define PF_MUTEX_TESTER	0x20000000	/* Thread belongs to the rt mutex tester */
1477 #define PF_FREEZER_SKIP	0x40000000	/* Freezer should not count it as freezeable */
1478 
1479 /*
1480  * Only the _current_ task can read/write to tsk->flags, but other
1481  * tasks can access tsk->flags in readonly mode for example
1482  * with tsk_used_math (like during threaded core dumping).
1483  * There is however an exception to this rule during ptrace
1484  * or during fork: the ptracer task is allowed to write to the
1485  * child->flags of its traced child (same goes for fork, the parent
1486  * can write to the child->flags), because we're guaranteed the
1487  * child is not running and in turn not changing child->flags
1488  * at the same time the parent does it.
1489  */
1490 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1491 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1492 #define clear_used_math() clear_stopped_child_used_math(current)
1493 #define set_used_math() set_stopped_child_used_math(current)
1494 #define conditional_stopped_child_used_math(condition, child) \
1495 	do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1496 #define conditional_used_math(condition) \
1497 	conditional_stopped_child_used_math(condition, current)
1498 #define copy_to_stopped_child_used_math(child) \
1499 	do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1500 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1501 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1502 #define used_math() tsk_used_math(current)
1503 
1504 #ifdef CONFIG_SMP
1505 extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
1506 #else
1507 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1508 {
1509 	if (!cpu_isset(0, new_mask))
1510 		return -EINVAL;
1511 	return 0;
1512 }
1513 #endif
1514 
1515 extern unsigned long long sched_clock(void);
1516 
1517 /*
1518  * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1519  * clock constructed from sched_clock():
1520  */
1521 extern unsigned long long cpu_clock(int cpu);
1522 
1523 extern unsigned long long
1524 task_sched_runtime(struct task_struct *task);
1525 
1526 /* sched_exec is called by processes performing an exec */
1527 #ifdef CONFIG_SMP
1528 extern void sched_exec(void);
1529 #else
1530 #define sched_exec()   {}
1531 #endif
1532 
1533 extern void sched_clock_idle_sleep_event(void);
1534 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1535 
1536 #ifdef CONFIG_HOTPLUG_CPU
1537 extern void idle_task_exit(void);
1538 #else
1539 static inline void idle_task_exit(void) {}
1540 #endif
1541 
1542 extern void sched_idle_next(void);
1543 
1544 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1545 extern void wake_up_idle_cpu(int cpu);
1546 #else
1547 static inline void wake_up_idle_cpu(int cpu) { }
1548 #endif
1549 
1550 #ifdef CONFIG_SCHED_DEBUG
1551 extern unsigned int sysctl_sched_latency;
1552 extern unsigned int sysctl_sched_min_granularity;
1553 extern unsigned int sysctl_sched_wakeup_granularity;
1554 extern unsigned int sysctl_sched_batch_wakeup_granularity;
1555 extern unsigned int sysctl_sched_child_runs_first;
1556 extern unsigned int sysctl_sched_features;
1557 extern unsigned int sysctl_sched_migration_cost;
1558 extern unsigned int sysctl_sched_nr_migrate;
1559 
1560 int sched_nr_latency_handler(struct ctl_table *table, int write,
1561 		struct file *file, void __user *buffer, size_t *length,
1562 		loff_t *ppos);
1563 #endif
1564 extern unsigned int sysctl_sched_rt_period;
1565 extern int sysctl_sched_rt_runtime;
1566 
1567 extern unsigned int sysctl_sched_compat_yield;
1568 
1569 #ifdef CONFIG_RT_MUTEXES
1570 extern int rt_mutex_getprio(struct task_struct *p);
1571 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1572 extern void rt_mutex_adjust_pi(struct task_struct *p);
1573 #else
1574 static inline int rt_mutex_getprio(struct task_struct *p)
1575 {
1576 	return p->normal_prio;
1577 }
1578 # define rt_mutex_adjust_pi(p)		do { } while (0)
1579 #endif
1580 
1581 extern void set_user_nice(struct task_struct *p, long nice);
1582 extern int task_prio(const struct task_struct *p);
1583 extern int task_nice(const struct task_struct *p);
1584 extern int can_nice(const struct task_struct *p, const int nice);
1585 extern int task_curr(const struct task_struct *p);
1586 extern int idle_cpu(int cpu);
1587 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1588 extern struct task_struct *idle_task(int cpu);
1589 extern struct task_struct *curr_task(int cpu);
1590 extern void set_curr_task(int cpu, struct task_struct *p);
1591 
1592 void yield(void);
1593 
1594 /*
1595  * The default (Linux) execution domain.
1596  */
1597 extern struct exec_domain	default_exec_domain;
1598 
1599 union thread_union {
1600 	struct thread_info thread_info;
1601 	unsigned long stack[THREAD_SIZE/sizeof(long)];
1602 };
1603 
1604 #ifndef __HAVE_ARCH_KSTACK_END
1605 static inline int kstack_end(void *addr)
1606 {
1607 	/* Reliable end of stack detection:
1608 	 * Some APM bios versions misalign the stack
1609 	 */
1610 	return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1611 }
1612 #endif
1613 
1614 extern union thread_union init_thread_union;
1615 extern struct task_struct init_task;
1616 
1617 extern struct   mm_struct init_mm;
1618 
1619 extern struct pid_namespace init_pid_ns;
1620 
1621 /*
1622  * find a task by one of its numerical ids
1623  *
1624  * find_task_by_pid_type_ns():
1625  *      it is the most generic call - it finds a task by all id,
1626  *      type and namespace specified
1627  * find_task_by_pid_ns():
1628  *      finds a task by its pid in the specified namespace
1629  * find_task_by_vpid():
1630  *      finds a task by its virtual pid
1631  * find_task_by_pid():
1632  *      finds a task by its global pid
1633  *
1634  * see also find_pid() etc in include/linux/pid.h
1635  */
1636 
1637 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1638 		struct pid_namespace *ns);
1639 
1640 extern struct task_struct *find_task_by_pid(pid_t nr);
1641 extern struct task_struct *find_task_by_vpid(pid_t nr);
1642 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1643 		struct pid_namespace *ns);
1644 
1645 extern void __set_special_pids(struct pid *pid);
1646 
1647 /* per-UID process charging. */
1648 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1649 static inline struct user_struct *get_uid(struct user_struct *u)
1650 {
1651 	atomic_inc(&u->__count);
1652 	return u;
1653 }
1654 extern void free_uid(struct user_struct *);
1655 extern void switch_uid(struct user_struct *);
1656 extern void release_uids(struct user_namespace *ns);
1657 
1658 #include <asm/current.h>
1659 
1660 extern void do_timer(unsigned long ticks);
1661 
1662 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1663 extern int wake_up_process(struct task_struct *tsk);
1664 extern void wake_up_new_task(struct task_struct *tsk,
1665 				unsigned long clone_flags);
1666 #ifdef CONFIG_SMP
1667  extern void kick_process(struct task_struct *tsk);
1668 #else
1669  static inline void kick_process(struct task_struct *tsk) { }
1670 #endif
1671 extern void sched_fork(struct task_struct *p, int clone_flags);
1672 extern void sched_dead(struct task_struct *p);
1673 
1674 extern int in_group_p(gid_t);
1675 extern int in_egroup_p(gid_t);
1676 
1677 extern void proc_caches_init(void);
1678 extern void flush_signals(struct task_struct *);
1679 extern void ignore_signals(struct task_struct *);
1680 extern void flush_signal_handlers(struct task_struct *, int force_default);
1681 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1682 
1683 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1684 {
1685 	unsigned long flags;
1686 	int ret;
1687 
1688 	spin_lock_irqsave(&tsk->sighand->siglock, flags);
1689 	ret = dequeue_signal(tsk, mask, info);
1690 	spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1691 
1692 	return ret;
1693 }
1694 
1695 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1696 			      sigset_t *mask);
1697 extern void unblock_all_signals(void);
1698 extern void release_task(struct task_struct * p);
1699 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1700 extern int force_sigsegv(int, struct task_struct *);
1701 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1702 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1703 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1704 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1705 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1706 extern int kill_pid(struct pid *pid, int sig, int priv);
1707 extern int kill_proc_info(int, struct siginfo *, pid_t);
1708 extern void do_notify_parent(struct task_struct *, int);
1709 extern void force_sig(int, struct task_struct *);
1710 extern void force_sig_specific(int, struct task_struct *);
1711 extern int send_sig(int, struct task_struct *, int);
1712 extern void zap_other_threads(struct task_struct *p);
1713 extern int kill_proc(pid_t, int, int);
1714 extern struct sigqueue *sigqueue_alloc(void);
1715 extern void sigqueue_free(struct sigqueue *);
1716 extern int send_sigqueue(int, struct sigqueue *,  struct task_struct *);
1717 extern int send_group_sigqueue(int, struct sigqueue *,  struct task_struct *);
1718 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1719 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1720 
1721 static inline int kill_cad_pid(int sig, int priv)
1722 {
1723 	return kill_pid(cad_pid, sig, priv);
1724 }
1725 
1726 /* These can be the second arg to send_sig_info/send_group_sig_info.  */
1727 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1728 #define SEND_SIG_PRIV	((struct siginfo *) 1)
1729 #define SEND_SIG_FORCED	((struct siginfo *) 2)
1730 
1731 static inline int is_si_special(const struct siginfo *info)
1732 {
1733 	return info <= SEND_SIG_FORCED;
1734 }
1735 
1736 /* True if we are on the alternate signal stack.  */
1737 
1738 static inline int on_sig_stack(unsigned long sp)
1739 {
1740 	return (sp - current->sas_ss_sp < current->sas_ss_size);
1741 }
1742 
1743 static inline int sas_ss_flags(unsigned long sp)
1744 {
1745 	return (current->sas_ss_size == 0 ? SS_DISABLE
1746 		: on_sig_stack(sp) ? SS_ONSTACK : 0);
1747 }
1748 
1749 /*
1750  * Routines for handling mm_structs
1751  */
1752 extern struct mm_struct * mm_alloc(void);
1753 
1754 /* mmdrop drops the mm and the page tables */
1755 extern void __mmdrop(struct mm_struct *);
1756 static inline void mmdrop(struct mm_struct * mm)
1757 {
1758 	if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1759 		__mmdrop(mm);
1760 }
1761 
1762 /* mmput gets rid of the mappings and all user-space */
1763 extern void mmput(struct mm_struct *);
1764 /* Grab a reference to a task's mm, if it is not already going away */
1765 extern struct mm_struct *get_task_mm(struct task_struct *task);
1766 /* Remove the current tasks stale references to the old mm_struct */
1767 extern void mm_release(struct task_struct *, struct mm_struct *);
1768 
1769 extern int  copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1770 extern void flush_thread(void);
1771 extern void exit_thread(void);
1772 
1773 extern void exit_files(struct task_struct *);
1774 extern void __cleanup_signal(struct signal_struct *);
1775 extern void __cleanup_sighand(struct sighand_struct *);
1776 extern void exit_itimers(struct signal_struct *);
1777 
1778 extern NORET_TYPE void do_group_exit(int);
1779 
1780 extern void daemonize(const char *, ...);
1781 extern int allow_signal(int);
1782 extern int disallow_signal(int);
1783 
1784 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1785 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1786 struct task_struct *fork_idle(int);
1787 
1788 extern void set_task_comm(struct task_struct *tsk, char *from);
1789 extern char *get_task_comm(char *to, struct task_struct *tsk);
1790 
1791 #ifdef CONFIG_SMP
1792 extern void wait_task_inactive(struct task_struct * p);
1793 #else
1794 #define wait_task_inactive(p)	do { } while (0)
1795 #endif
1796 
1797 #define remove_parent(p)	list_del_init(&(p)->sibling)
1798 #define add_parent(p)		list_add_tail(&(p)->sibling,&(p)->parent->children)
1799 
1800 #define next_task(p)	list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1801 
1802 #define for_each_process(p) \
1803 	for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1804 
1805 /*
1806  * Careful: do_each_thread/while_each_thread is a double loop so
1807  *          'break' will not work as expected - use goto instead.
1808  */
1809 #define do_each_thread(g, t) \
1810 	for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1811 
1812 #define while_each_thread(g, t) \
1813 	while ((t = next_thread(t)) != g)
1814 
1815 /* de_thread depends on thread_group_leader not being a pid based check */
1816 #define thread_group_leader(p)	(p == p->group_leader)
1817 
1818 /* Do to the insanities of de_thread it is possible for a process
1819  * to have the pid of the thread group leader without actually being
1820  * the thread group leader.  For iteration through the pids in proc
1821  * all we care about is that we have a task with the appropriate
1822  * pid, we don't actually care if we have the right task.
1823  */
1824 static inline int has_group_leader_pid(struct task_struct *p)
1825 {
1826 	return p->pid == p->tgid;
1827 }
1828 
1829 static inline
1830 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
1831 {
1832 	return p1->tgid == p2->tgid;
1833 }
1834 
1835 static inline struct task_struct *next_thread(const struct task_struct *p)
1836 {
1837 	return list_entry(rcu_dereference(p->thread_group.next),
1838 			  struct task_struct, thread_group);
1839 }
1840 
1841 static inline int thread_group_empty(struct task_struct *p)
1842 {
1843 	return list_empty(&p->thread_group);
1844 }
1845 
1846 #define delay_group_leader(p) \
1847 		(thread_group_leader(p) && !thread_group_empty(p))
1848 
1849 /*
1850  * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
1851  * subscriptions and synchronises with wait4().  Also used in procfs.  Also
1852  * pins the final release of task.io_context.  Also protects ->cpuset and
1853  * ->cgroup.subsys[].
1854  *
1855  * Nests both inside and outside of read_lock(&tasklist_lock).
1856  * It must not be nested with write_lock_irq(&tasklist_lock),
1857  * neither inside nor outside.
1858  */
1859 static inline void task_lock(struct task_struct *p)
1860 {
1861 	spin_lock(&p->alloc_lock);
1862 }
1863 
1864 static inline void task_unlock(struct task_struct *p)
1865 {
1866 	spin_unlock(&p->alloc_lock);
1867 }
1868 
1869 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1870 							unsigned long *flags);
1871 
1872 static inline void unlock_task_sighand(struct task_struct *tsk,
1873 						unsigned long *flags)
1874 {
1875 	spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1876 }
1877 
1878 #ifndef __HAVE_THREAD_FUNCTIONS
1879 
1880 #define task_thread_info(task)	((struct thread_info *)(task)->stack)
1881 #define task_stack_page(task)	((task)->stack)
1882 
1883 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1884 {
1885 	*task_thread_info(p) = *task_thread_info(org);
1886 	task_thread_info(p)->task = p;
1887 }
1888 
1889 static inline unsigned long *end_of_stack(struct task_struct *p)
1890 {
1891 	return (unsigned long *)(task_thread_info(p) + 1);
1892 }
1893 
1894 #endif
1895 
1896 /* set thread flags in other task's structures
1897  * - see asm/thread_info.h for TIF_xxxx flags available
1898  */
1899 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1900 {
1901 	set_ti_thread_flag(task_thread_info(tsk), flag);
1902 }
1903 
1904 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1905 {
1906 	clear_ti_thread_flag(task_thread_info(tsk), flag);
1907 }
1908 
1909 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1910 {
1911 	return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1912 }
1913 
1914 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1915 {
1916 	return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1917 }
1918 
1919 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1920 {
1921 	return test_ti_thread_flag(task_thread_info(tsk), flag);
1922 }
1923 
1924 static inline void set_tsk_need_resched(struct task_struct *tsk)
1925 {
1926 	set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1927 }
1928 
1929 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1930 {
1931 	clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1932 }
1933 
1934 static inline int signal_pending(struct task_struct *p)
1935 {
1936 	return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1937 }
1938 
1939 extern int __fatal_signal_pending(struct task_struct *p);
1940 
1941 static inline int fatal_signal_pending(struct task_struct *p)
1942 {
1943 	return signal_pending(p) && __fatal_signal_pending(p);
1944 }
1945 
1946 static inline int need_resched(void)
1947 {
1948 	return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1949 }
1950 
1951 /*
1952  * cond_resched() and cond_resched_lock(): latency reduction via
1953  * explicit rescheduling in places that are safe. The return
1954  * value indicates whether a reschedule was done in fact.
1955  * cond_resched_lock() will drop the spinlock before scheduling,
1956  * cond_resched_softirq() will enable bhs before scheduling.
1957  */
1958 #ifdef CONFIG_PREEMPT
1959 static inline int cond_resched(void)
1960 {
1961 	return 0;
1962 }
1963 #else
1964 extern int _cond_resched(void);
1965 static inline int cond_resched(void)
1966 {
1967 	return _cond_resched();
1968 }
1969 #endif
1970 extern int cond_resched_lock(spinlock_t * lock);
1971 extern int cond_resched_softirq(void);
1972 
1973 /*
1974  * Does a critical section need to be broken due to another
1975  * task waiting?: (technically does not depend on CONFIG_PREEMPT,
1976  * but a general need for low latency)
1977  */
1978 static inline int spin_needbreak(spinlock_t *lock)
1979 {
1980 #ifdef CONFIG_PREEMPT
1981 	return spin_is_contended(lock);
1982 #else
1983 	return 0;
1984 #endif
1985 }
1986 
1987 /*
1988  * Reevaluate whether the task has signals pending delivery.
1989  * Wake the task if so.
1990  * This is required every time the blocked sigset_t changes.
1991  * callers must hold sighand->siglock.
1992  */
1993 extern void recalc_sigpending_and_wake(struct task_struct *t);
1994 extern void recalc_sigpending(void);
1995 
1996 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1997 
1998 /*
1999  * Wrappers for p->thread_info->cpu access. No-op on UP.
2000  */
2001 #ifdef CONFIG_SMP
2002 
2003 static inline unsigned int task_cpu(const struct task_struct *p)
2004 {
2005 	return task_thread_info(p)->cpu;
2006 }
2007 
2008 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2009 
2010 #else
2011 
2012 static inline unsigned int task_cpu(const struct task_struct *p)
2013 {
2014 	return 0;
2015 }
2016 
2017 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2018 {
2019 }
2020 
2021 #endif /* CONFIG_SMP */
2022 
2023 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
2024 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2025 #else
2026 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
2027 {
2028 	mm->mmap_base = TASK_UNMAPPED_BASE;
2029 	mm->get_unmapped_area = arch_get_unmapped_area;
2030 	mm->unmap_area = arch_unmap_area;
2031 }
2032 #endif
2033 
2034 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
2035 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
2036 
2037 extern int sched_mc_power_savings, sched_smt_power_savings;
2038 
2039 extern void normalize_rt_tasks(void);
2040 
2041 #ifdef CONFIG_GROUP_SCHED
2042 
2043 extern struct task_group init_task_group;
2044 
2045 extern struct task_group *sched_create_group(void);
2046 extern void sched_destroy_group(struct task_group *tg);
2047 extern void sched_move_task(struct task_struct *tsk);
2048 #ifdef CONFIG_FAIR_GROUP_SCHED
2049 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2050 extern unsigned long sched_group_shares(struct task_group *tg);
2051 #endif
2052 #ifdef CONFIG_RT_GROUP_SCHED
2053 extern int sched_group_set_rt_runtime(struct task_group *tg,
2054 				      long rt_runtime_us);
2055 extern long sched_group_rt_runtime(struct task_group *tg);
2056 #endif
2057 #endif
2058 
2059 #ifdef CONFIG_TASK_XACCT
2060 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2061 {
2062 	tsk->rchar += amt;
2063 }
2064 
2065 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2066 {
2067 	tsk->wchar += amt;
2068 }
2069 
2070 static inline void inc_syscr(struct task_struct *tsk)
2071 {
2072 	tsk->syscr++;
2073 }
2074 
2075 static inline void inc_syscw(struct task_struct *tsk)
2076 {
2077 	tsk->syscw++;
2078 }
2079 #else
2080 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2081 {
2082 }
2083 
2084 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2085 {
2086 }
2087 
2088 static inline void inc_syscr(struct task_struct *tsk)
2089 {
2090 }
2091 
2092 static inline void inc_syscw(struct task_struct *tsk)
2093 {
2094 }
2095 #endif
2096 
2097 #ifdef CONFIG_SMP
2098 void migration_init(void);
2099 #else
2100 static inline void migration_init(void)
2101 {
2102 }
2103 #endif
2104 
2105 #ifndef TASK_SIZE_OF
2106 #define TASK_SIZE_OF(tsk)	TASK_SIZE
2107 #endif
2108 
2109 #endif /* __KERNEL__ */
2110 
2111 #endif
2112