xref: /linux-6.15/include/linux/sched.h (revision 652afc27)
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/page.h>
65 #include <asm/ptrace.h>
66 #include <asm/cputime.h>
67 
68 #include <linux/smp.h>
69 #include <linux/sem.h>
70 #include <linux/signal.h>
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>
77 #include <linux/proportions.h>
78 #include <linux/seccomp.h>
79 #include <linux/rcupdate.h>
80 #include <linux/rtmutex.h>
81 
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>
87 #include <linux/task_io_accounting.h>
88 #include <linux/kobject.h>
89 #include <linux/latencytop.h>
90 #include <linux/cred.h>
91 
92 #include <asm/processor.h>
93 
94 struct mem_cgroup;
95 struct exec_domain;
96 struct futex_pi_state;
97 struct robust_list_head;
98 struct bio;
99 struct bts_tracer;
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 
140 struct seq_file;
141 struct cfs_rq;
142 struct task_group;
143 #ifdef CONFIG_SCHED_DEBUG
144 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
145 extern void proc_sched_set_task(struct task_struct *p);
146 extern void
147 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
148 #else
149 static inline void
150 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
151 {
152 }
153 static inline void proc_sched_set_task(struct task_struct *p)
154 {
155 }
156 static inline void
157 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
158 {
159 }
160 #endif
161 
162 extern unsigned long long time_sync_thresh;
163 
164 /*
165  * Task state bitmask. NOTE! These bits are also
166  * encoded in fs/proc/array.c: get_task_state().
167  *
168  * We have two separate sets of flags: task->state
169  * is about runnability, while task->exit_state are
170  * about the task exiting. Confusing, but this way
171  * modifying one set can't modify the other one by
172  * mistake.
173  */
174 #define TASK_RUNNING		0
175 #define TASK_INTERRUPTIBLE	1
176 #define TASK_UNINTERRUPTIBLE	2
177 #define __TASK_STOPPED		4
178 #define __TASK_TRACED		8
179 /* in tsk->exit_state */
180 #define EXIT_ZOMBIE		16
181 #define EXIT_DEAD		32
182 /* in tsk->state again */
183 #define TASK_DEAD		64
184 #define TASK_WAKEKILL		128
185 
186 /* Convenience macros for the sake of set_task_state */
187 #define TASK_KILLABLE		(TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
188 #define TASK_STOPPED		(TASK_WAKEKILL | __TASK_STOPPED)
189 #define TASK_TRACED		(TASK_WAKEKILL | __TASK_TRACED)
190 
191 /* Convenience macros for the sake of wake_up */
192 #define TASK_NORMAL		(TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
193 #define TASK_ALL		(TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
194 
195 /* get_task_state() */
196 #define TASK_REPORT		(TASK_RUNNING | TASK_INTERRUPTIBLE | \
197 				 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
198 				 __TASK_TRACED)
199 
200 #define task_is_traced(task)	((task->state & __TASK_TRACED) != 0)
201 #define task_is_stopped(task)	((task->state & __TASK_STOPPED) != 0)
202 #define task_is_stopped_or_traced(task)	\
203 			((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
204 #define task_contributes_to_load(task)	\
205 				((task->state & TASK_UNINTERRUPTIBLE) != 0)
206 
207 #define __set_task_state(tsk, state_value)		\
208 	do { (tsk)->state = (state_value); } while (0)
209 #define set_task_state(tsk, state_value)		\
210 	set_mb((tsk)->state, (state_value))
211 
212 /*
213  * set_current_state() includes a barrier so that the write of current->state
214  * is correctly serialised wrt the caller's subsequent test of whether to
215  * actually sleep:
216  *
217  *	set_current_state(TASK_UNINTERRUPTIBLE);
218  *	if (do_i_need_to_sleep())
219  *		schedule();
220  *
221  * If the caller does not need such serialisation then use __set_current_state()
222  */
223 #define __set_current_state(state_value)			\
224 	do { current->state = (state_value); } while (0)
225 #define set_current_state(state_value)		\
226 	set_mb(current->state, (state_value))
227 
228 /* Task command name length */
229 #define TASK_COMM_LEN 16
230 
231 #include <linux/spinlock.h>
232 
233 /*
234  * This serializes "schedule()" and also protects
235  * the run-queue from deletions/modifications (but
236  * _adding_ to the beginning of the run-queue has
237  * a separate lock).
238  */
239 extern rwlock_t tasklist_lock;
240 extern spinlock_t mmlist_lock;
241 
242 struct task_struct;
243 
244 extern void sched_init(void);
245 extern void sched_init_smp(void);
246 extern asmlinkage void schedule_tail(struct task_struct *prev);
247 extern void init_idle(struct task_struct *idle, int cpu);
248 extern void init_idle_bootup_task(struct task_struct *idle);
249 
250 extern int runqueue_is_locked(void);
251 extern void task_rq_unlock_wait(struct task_struct *p);
252 
253 extern cpumask_var_t nohz_cpu_mask;
254 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
255 extern int select_nohz_load_balancer(int cpu);
256 #else
257 static inline int select_nohz_load_balancer(int cpu)
258 {
259 	return 0;
260 }
261 #endif
262 
263 /*
264  * Only dump TASK_* tasks. (0 for all tasks)
265  */
266 extern void show_state_filter(unsigned long state_filter);
267 
268 static inline void show_state(void)
269 {
270 	show_state_filter(0);
271 }
272 
273 extern void show_regs(struct pt_regs *);
274 
275 /*
276  * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
277  * task), SP is the stack pointer of the first frame that should be shown in the back
278  * trace (or NULL if the entire call-chain of the task should be shown).
279  */
280 extern void show_stack(struct task_struct *task, unsigned long *sp);
281 
282 void io_schedule(void);
283 long io_schedule_timeout(long timeout);
284 
285 extern void cpu_init (void);
286 extern void trap_init(void);
287 extern void update_process_times(int user);
288 extern void scheduler_tick(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 touch_softlockup_watchdog(void);
295 extern void touch_all_softlockup_watchdogs(void);
296 extern unsigned int  softlockup_panic;
297 extern unsigned long sysctl_hung_task_check_count;
298 extern unsigned long sysctl_hung_task_timeout_secs;
299 extern unsigned long sysctl_hung_task_warnings;
300 extern int softlockup_thresh;
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 USE_SPLIT_PTLOCKS
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  /* !USE_SPLIT_PTLOCKS */
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 /* !USE_SPLIT_PTLOCKS */
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_HUGETLB_PRIVATE 7
405 #define MMF_DUMP_HUGETLB_SHARED  8
406 #define MMF_DUMP_FILTER_SHIFT	MMF_DUMPABLE_BITS
407 #define MMF_DUMP_FILTER_BITS	7
408 #define MMF_DUMP_FILTER_MASK \
409 	(((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
410 #define MMF_DUMP_FILTER_DEFAULT \
411 	((1 << MMF_DUMP_ANON_PRIVATE) |	(1 << MMF_DUMP_ANON_SHARED) |\
412 	 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
413 
414 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
415 # define MMF_DUMP_MASK_DEFAULT_ELF	(1 << MMF_DUMP_ELF_HEADERS)
416 #else
417 # define MMF_DUMP_MASK_DEFAULT_ELF	0
418 #endif
419 
420 struct sighand_struct {
421 	atomic_t		count;
422 	struct k_sigaction	action[_NSIG];
423 	spinlock_t		siglock;
424 	wait_queue_head_t	signalfd_wqh;
425 };
426 
427 struct pacct_struct {
428 	int			ac_flag;
429 	long			ac_exitcode;
430 	unsigned long		ac_mem;
431 	cputime_t		ac_utime, ac_stime;
432 	unsigned long		ac_minflt, ac_majflt;
433 };
434 
435 /**
436  * struct task_cputime - collected CPU time counts
437  * @utime:		time spent in user mode, in &cputime_t units
438  * @stime:		time spent in kernel mode, in &cputime_t units
439  * @sum_exec_runtime:	total time spent on the CPU, in nanoseconds
440  *
441  * This structure groups together three kinds of CPU time that are
442  * tracked for threads and thread groups.  Most things considering
443  * CPU time want to group these counts together and treat all three
444  * of them in parallel.
445  */
446 struct task_cputime {
447 	cputime_t utime;
448 	cputime_t stime;
449 	unsigned long long sum_exec_runtime;
450 };
451 /* Alternate field names when used to cache expirations. */
452 #define prof_exp	stime
453 #define virt_exp	utime
454 #define sched_exp	sum_exec_runtime
455 
456 /**
457  * struct thread_group_cputime - thread group interval timer counts
458  * @totals:		thread group interval timers; substructure for
459  *			uniprocessor kernel, per-cpu for SMP kernel.
460  *
461  * This structure contains the version of task_cputime, above, that is
462  * used for thread group CPU clock calculations.
463  */
464 struct thread_group_cputime {
465 	struct task_cputime *totals;
466 };
467 
468 /*
469  * NOTE! "signal_struct" does not have it's own
470  * locking, because a shared signal_struct always
471  * implies a shared sighand_struct, so locking
472  * sighand_struct is always a proper superset of
473  * the locking of signal_struct.
474  */
475 struct signal_struct {
476 	atomic_t		count;
477 	atomic_t		live;
478 
479 	wait_queue_head_t	wait_chldexit;	/* for wait4() */
480 
481 	/* current thread group signal load-balancing target: */
482 	struct task_struct	*curr_target;
483 
484 	/* shared signal handling: */
485 	struct sigpending	shared_pending;
486 
487 	/* thread group exit support */
488 	int			group_exit_code;
489 	/* overloaded:
490 	 * - notify group_exit_task when ->count is equal to notify_count
491 	 * - everyone except group_exit_task is stopped during signal delivery
492 	 *   of fatal signals, group_exit_task processes the signal.
493 	 */
494 	int			notify_count;
495 	struct task_struct	*group_exit_task;
496 
497 	/* thread group stop support, overloads group_exit_code too */
498 	int			group_stop_count;
499 	unsigned int		flags; /* see SIGNAL_* flags below */
500 
501 	/* POSIX.1b Interval Timers */
502 	struct list_head posix_timers;
503 
504 	/* ITIMER_REAL timer for the process */
505 	struct hrtimer real_timer;
506 	struct pid *leader_pid;
507 	ktime_t it_real_incr;
508 
509 	/* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
510 	cputime_t it_prof_expires, it_virt_expires;
511 	cputime_t it_prof_incr, it_virt_incr;
512 
513 	/*
514 	 * Thread group totals for process CPU clocks.
515 	 * See thread_group_cputime(), et al, for details.
516 	 */
517 	struct thread_group_cputime cputime;
518 
519 	/* Earliest-expiration cache. */
520 	struct task_cputime cputime_expires;
521 
522 	struct list_head cpu_timers[3];
523 
524 	/* job control IDs */
525 
526 	/*
527 	 * pgrp and session fields are deprecated.
528 	 * use the task_session_Xnr and task_pgrp_Xnr routines below
529 	 */
530 
531 	union {
532 		pid_t pgrp __deprecated;
533 		pid_t __pgrp;
534 	};
535 
536 	struct pid *tty_old_pgrp;
537 
538 	union {
539 		pid_t session __deprecated;
540 		pid_t __session;
541 	};
542 
543 	/* boolean value for session group leader */
544 	int leader;
545 
546 	struct tty_struct *tty; /* NULL if no tty */
547 
548 	/*
549 	 * Cumulative resource counters for dead threads in the group,
550 	 * and for reaped dead child processes forked by this group.
551 	 * Live threads maintain their own counters and add to these
552 	 * in __exit_signal, except for the group leader.
553 	 */
554 	cputime_t cutime, cstime;
555 	cputime_t gtime;
556 	cputime_t cgtime;
557 	unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
558 	unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
559 	unsigned long inblock, oublock, cinblock, coublock;
560 	struct task_io_accounting ioac;
561 
562 	/*
563 	 * We don't bother to synchronize most readers of this at all,
564 	 * because there is no reader checking a limit that actually needs
565 	 * to get both rlim_cur and rlim_max atomically, and either one
566 	 * alone is a single word that can safely be read normally.
567 	 * getrlimit/setrlimit use task_lock(current->group_leader) to
568 	 * protect this instead of the siglock, because they really
569 	 * have no need to disable irqs.
570 	 */
571 	struct rlimit rlim[RLIM_NLIMITS];
572 
573 #ifdef CONFIG_BSD_PROCESS_ACCT
574 	struct pacct_struct pacct;	/* per-process accounting information */
575 #endif
576 #ifdef CONFIG_TASKSTATS
577 	struct taskstats *stats;
578 #endif
579 #ifdef CONFIG_AUDIT
580 	unsigned audit_tty;
581 	struct tty_audit_buf *tty_audit_buf;
582 #endif
583 };
584 
585 /* Context switch must be unlocked if interrupts are to be enabled */
586 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
587 # define __ARCH_WANT_UNLOCKED_CTXSW
588 #endif
589 
590 /*
591  * Bits in flags field of signal_struct.
592  */
593 #define SIGNAL_STOP_STOPPED	0x00000001 /* job control stop in effect */
594 #define SIGNAL_STOP_DEQUEUED	0x00000002 /* stop signal dequeued */
595 #define SIGNAL_STOP_CONTINUED	0x00000004 /* SIGCONT since WCONTINUED reap */
596 #define SIGNAL_GROUP_EXIT	0x00000008 /* group exit in progress */
597 /*
598  * Pending notifications to parent.
599  */
600 #define SIGNAL_CLD_STOPPED	0x00000010
601 #define SIGNAL_CLD_CONTINUED	0x00000020
602 #define SIGNAL_CLD_MASK		(SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
603 
604 #define SIGNAL_UNKILLABLE	0x00000040 /* for init: ignore fatal signals */
605 
606 /* If true, all threads except ->group_exit_task have pending SIGKILL */
607 static inline int signal_group_exit(const struct signal_struct *sig)
608 {
609 	return	(sig->flags & SIGNAL_GROUP_EXIT) ||
610 		(sig->group_exit_task != NULL);
611 }
612 
613 /*
614  * Some day this will be a full-fledged user tracking system..
615  */
616 struct user_struct {
617 	atomic_t __count;	/* reference count */
618 	atomic_t processes;	/* How many processes does this user have? */
619 	atomic_t files;		/* How many open files does this user have? */
620 	atomic_t sigpending;	/* How many pending signals does this user have? */
621 #ifdef CONFIG_INOTIFY_USER
622 	atomic_t inotify_watches; /* How many inotify watches does this user have? */
623 	atomic_t inotify_devs;	/* How many inotify devs does this user have opened? */
624 #endif
625 #ifdef CONFIG_EPOLL
626 	atomic_t epoll_devs;	/* The number of epoll descriptors currently open */
627 	atomic_t epoll_watches;	/* The number of file descriptors currently watched */
628 #endif
629 #ifdef CONFIG_POSIX_MQUEUE
630 	/* protected by mq_lock	*/
631 	unsigned long mq_bytes;	/* How many bytes can be allocated to mqueue? */
632 #endif
633 	unsigned long locked_shm; /* How many pages of mlocked shm ? */
634 
635 #ifdef CONFIG_KEYS
636 	struct key *uid_keyring;	/* UID specific keyring */
637 	struct key *session_keyring;	/* UID's default session keyring */
638 #endif
639 
640 	/* Hash table maintenance information */
641 	struct hlist_node uidhash_node;
642 	uid_t uid;
643 	struct user_namespace *user_ns;
644 
645 #ifdef CONFIG_USER_SCHED
646 	struct task_group *tg;
647 #ifdef CONFIG_SYSFS
648 	struct kobject kobj;
649 	struct work_struct work;
650 #endif
651 #endif
652 };
653 
654 extern int uids_sysfs_init(void);
655 
656 extern struct user_struct *find_user(uid_t);
657 
658 extern struct user_struct root_user;
659 #define INIT_USER (&root_user)
660 
661 
662 struct backing_dev_info;
663 struct reclaim_state;
664 
665 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
666 struct sched_info {
667 	/* cumulative counters */
668 	unsigned long pcount;	      /* # of times run on this cpu */
669 	unsigned long long run_delay; /* time spent waiting on a runqueue */
670 
671 	/* timestamps */
672 	unsigned long long last_arrival,/* when we last ran on a cpu */
673 			   last_queued;	/* when we were last queued to run */
674 #ifdef CONFIG_SCHEDSTATS
675 	/* BKL stats */
676 	unsigned int bkl_count;
677 #endif
678 };
679 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
680 
681 #ifdef CONFIG_TASK_DELAY_ACCT
682 struct task_delay_info {
683 	spinlock_t	lock;
684 	unsigned int	flags;	/* Private per-task flags */
685 
686 	/* For each stat XXX, add following, aligned appropriately
687 	 *
688 	 * struct timespec XXX_start, XXX_end;
689 	 * u64 XXX_delay;
690 	 * u32 XXX_count;
691 	 *
692 	 * Atomicity of updates to XXX_delay, XXX_count protected by
693 	 * single lock above (split into XXX_lock if contention is an issue).
694 	 */
695 
696 	/*
697 	 * XXX_count is incremented on every XXX operation, the delay
698 	 * associated with the operation is added to XXX_delay.
699 	 * XXX_delay contains the accumulated delay time in nanoseconds.
700 	 */
701 	struct timespec blkio_start, blkio_end;	/* Shared by blkio, swapin */
702 	u64 blkio_delay;	/* wait for sync block io completion */
703 	u64 swapin_delay;	/* wait for swapin block io completion */
704 	u32 blkio_count;	/* total count of the number of sync block */
705 				/* io operations performed */
706 	u32 swapin_count;	/* total count of the number of swapin block */
707 				/* io operations performed */
708 
709 	struct timespec freepages_start, freepages_end;
710 	u64 freepages_delay;	/* wait for memory reclaim */
711 	u32 freepages_count;	/* total count of memory reclaim */
712 };
713 #endif	/* CONFIG_TASK_DELAY_ACCT */
714 
715 static inline int sched_info_on(void)
716 {
717 #ifdef CONFIG_SCHEDSTATS
718 	return 1;
719 #elif defined(CONFIG_TASK_DELAY_ACCT)
720 	extern int delayacct_on;
721 	return delayacct_on;
722 #else
723 	return 0;
724 #endif
725 }
726 
727 enum cpu_idle_type {
728 	CPU_IDLE,
729 	CPU_NOT_IDLE,
730 	CPU_NEWLY_IDLE,
731 	CPU_MAX_IDLE_TYPES
732 };
733 
734 /*
735  * sched-domains (multiprocessor balancing) declarations:
736  */
737 
738 /*
739  * Increase resolution of nice-level calculations:
740  */
741 #define SCHED_LOAD_SHIFT	10
742 #define SCHED_LOAD_SCALE	(1L << SCHED_LOAD_SHIFT)
743 
744 #define SCHED_LOAD_SCALE_FUZZ	SCHED_LOAD_SCALE
745 
746 #ifdef CONFIG_SMP
747 #define SD_LOAD_BALANCE		1	/* Do load balancing on this domain. */
748 #define SD_BALANCE_NEWIDLE	2	/* Balance when about to become idle */
749 #define SD_BALANCE_EXEC		4	/* Balance on exec */
750 #define SD_BALANCE_FORK		8	/* Balance on fork, clone */
751 #define SD_WAKE_IDLE		16	/* Wake to idle CPU on task wakeup */
752 #define SD_WAKE_AFFINE		32	/* Wake task to waking CPU */
753 #define SD_WAKE_BALANCE		64	/* Perform balancing at task wakeup */
754 #define SD_SHARE_CPUPOWER	128	/* Domain members share cpu power */
755 #define SD_POWERSAVINGS_BALANCE	256	/* Balance for power savings */
756 #define SD_SHARE_PKG_RESOURCES	512	/* Domain members share cpu pkg resources */
757 #define SD_SERIALIZE		1024	/* Only a single load balancing instance */
758 #define SD_WAKE_IDLE_FAR	2048	/* Gain latency sacrificing cache hit */
759 
760 enum powersavings_balance_level {
761 	POWERSAVINGS_BALANCE_NONE = 0,  /* No power saving load balance */
762 	POWERSAVINGS_BALANCE_BASIC,	/* Fill one thread/core/package
763 					 * first for long running threads
764 					 */
765 	POWERSAVINGS_BALANCE_WAKEUP,	/* Also bias task wakeups to semi-idle
766 					 * cpu package for power savings
767 					 */
768 	MAX_POWERSAVINGS_BALANCE_LEVELS
769 };
770 
771 extern int sched_mc_power_savings, sched_smt_power_savings;
772 
773 static inline int sd_balance_for_mc_power(void)
774 {
775 	if (sched_smt_power_savings)
776 		return SD_POWERSAVINGS_BALANCE;
777 
778 	return 0;
779 }
780 
781 static inline int sd_balance_for_package_power(void)
782 {
783 	if (sched_mc_power_savings | sched_smt_power_savings)
784 		return SD_POWERSAVINGS_BALANCE;
785 
786 	return 0;
787 }
788 
789 /*
790  * Optimise SD flags for power savings:
791  * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
792  * Keep default SD flags if sched_{smt,mc}_power_saving=0
793  */
794 
795 static inline int sd_power_saving_flags(void)
796 {
797 	if (sched_mc_power_savings | sched_smt_power_savings)
798 		return SD_BALANCE_NEWIDLE;
799 
800 	return 0;
801 }
802 
803 struct sched_group {
804 	struct sched_group *next;	/* Must be a circular list */
805 
806 	/*
807 	 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
808 	 * single CPU. This is read only (except for setup, hotplug CPU).
809 	 * Note : Never change cpu_power without recompute its reciprocal
810 	 */
811 	unsigned int __cpu_power;
812 	/*
813 	 * reciprocal value of cpu_power to avoid expensive divides
814 	 * (see include/linux/reciprocal_div.h)
815 	 */
816 	u32 reciprocal_cpu_power;
817 
818 	unsigned long cpumask[];
819 };
820 
821 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
822 {
823 	return to_cpumask(sg->cpumask);
824 }
825 
826 enum sched_domain_level {
827 	SD_LV_NONE = 0,
828 	SD_LV_SIBLING,
829 	SD_LV_MC,
830 	SD_LV_CPU,
831 	SD_LV_NODE,
832 	SD_LV_ALLNODES,
833 	SD_LV_MAX
834 };
835 
836 struct sched_domain_attr {
837 	int relax_domain_level;
838 };
839 
840 #define SD_ATTR_INIT	(struct sched_domain_attr) {	\
841 	.relax_domain_level = -1,			\
842 }
843 
844 struct sched_domain {
845 	/* These fields must be setup */
846 	struct sched_domain *parent;	/* top domain must be null terminated */
847 	struct sched_domain *child;	/* bottom domain must be null terminated */
848 	struct sched_group *groups;	/* the balancing groups of the domain */
849 	unsigned long min_interval;	/* Minimum balance interval ms */
850 	unsigned long max_interval;	/* Maximum balance interval ms */
851 	unsigned int busy_factor;	/* less balancing by factor if busy */
852 	unsigned int imbalance_pct;	/* No balance until over watermark */
853 	unsigned int cache_nice_tries;	/* Leave cache hot tasks for # tries */
854 	unsigned int busy_idx;
855 	unsigned int idle_idx;
856 	unsigned int newidle_idx;
857 	unsigned int wake_idx;
858 	unsigned int forkexec_idx;
859 	int flags;			/* See SD_* */
860 	enum sched_domain_level level;
861 
862 	/* Runtime fields. */
863 	unsigned long last_balance;	/* init to jiffies. units in jiffies */
864 	unsigned int balance_interval;	/* initialise to 1. units in ms. */
865 	unsigned int nr_balance_failed; /* initialise to 0 */
866 
867 	u64 last_update;
868 
869 #ifdef CONFIG_SCHEDSTATS
870 	/* load_balance() stats */
871 	unsigned int lb_count[CPU_MAX_IDLE_TYPES];
872 	unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
873 	unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
874 	unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
875 	unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
876 	unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
877 	unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
878 	unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
879 
880 	/* Active load balancing */
881 	unsigned int alb_count;
882 	unsigned int alb_failed;
883 	unsigned int alb_pushed;
884 
885 	/* SD_BALANCE_EXEC stats */
886 	unsigned int sbe_count;
887 	unsigned int sbe_balanced;
888 	unsigned int sbe_pushed;
889 
890 	/* SD_BALANCE_FORK stats */
891 	unsigned int sbf_count;
892 	unsigned int sbf_balanced;
893 	unsigned int sbf_pushed;
894 
895 	/* try_to_wake_up() stats */
896 	unsigned int ttwu_wake_remote;
897 	unsigned int ttwu_move_affine;
898 	unsigned int ttwu_move_balance;
899 #endif
900 #ifdef CONFIG_SCHED_DEBUG
901 	char *name;
902 #endif
903 
904 	/* span of all CPUs in this domain */
905 	unsigned long span[];
906 };
907 
908 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
909 {
910 	return to_cpumask(sd->span);
911 }
912 
913 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
914 				    struct sched_domain_attr *dattr_new);
915 extern int arch_reinit_sched_domains(void);
916 
917 /* Test a flag in parent sched domain */
918 static inline int test_sd_parent(struct sched_domain *sd, int flag)
919 {
920 	if (sd->parent && (sd->parent->flags & flag))
921 		return 1;
922 
923 	return 0;
924 }
925 
926 #else /* CONFIG_SMP */
927 
928 struct sched_domain_attr;
929 
930 static inline void
931 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
932 			struct sched_domain_attr *dattr_new)
933 {
934 }
935 #endif	/* !CONFIG_SMP */
936 
937 struct io_context;			/* See blkdev.h */
938 
939 
940 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
941 extern void prefetch_stack(struct task_struct *t);
942 #else
943 static inline void prefetch_stack(struct task_struct *t) { }
944 #endif
945 
946 struct audit_context;		/* See audit.c */
947 struct mempolicy;
948 struct pipe_inode_info;
949 struct uts_namespace;
950 
951 struct rq;
952 struct sched_domain;
953 
954 struct sched_class {
955 	const struct sched_class *next;
956 
957 	void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
958 	void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
959 	void (*yield_task) (struct rq *rq);
960 
961 	void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
962 
963 	struct task_struct * (*pick_next_task) (struct rq *rq);
964 	void (*put_prev_task) (struct rq *rq, struct task_struct *p);
965 
966 #ifdef CONFIG_SMP
967 	int  (*select_task_rq)(struct task_struct *p, int sync);
968 
969 	unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
970 			struct rq *busiest, unsigned long max_load_move,
971 			struct sched_domain *sd, enum cpu_idle_type idle,
972 			int *all_pinned, int *this_best_prio);
973 
974 	int (*move_one_task) (struct rq *this_rq, int this_cpu,
975 			      struct rq *busiest, struct sched_domain *sd,
976 			      enum cpu_idle_type idle);
977 	void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
978 	void (*post_schedule) (struct rq *this_rq);
979 	void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
980 
981 	void (*set_cpus_allowed)(struct task_struct *p,
982 				 const struct cpumask *newmask);
983 
984 	void (*rq_online)(struct rq *rq);
985 	void (*rq_offline)(struct rq *rq);
986 #endif
987 
988 	void (*set_curr_task) (struct rq *rq);
989 	void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
990 	void (*task_new) (struct rq *rq, struct task_struct *p);
991 
992 	void (*switched_from) (struct rq *this_rq, struct task_struct *task,
993 			       int running);
994 	void (*switched_to) (struct rq *this_rq, struct task_struct *task,
995 			     int running);
996 	void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
997 			     int oldprio, int running);
998 
999 #ifdef CONFIG_FAIR_GROUP_SCHED
1000 	void (*moved_group) (struct task_struct *p);
1001 #endif
1002 };
1003 
1004 struct load_weight {
1005 	unsigned long weight, inv_weight;
1006 };
1007 
1008 /*
1009  * CFS stats for a schedulable entity (task, task-group etc)
1010  *
1011  * Current field usage histogram:
1012  *
1013  *     4 se->block_start
1014  *     4 se->run_node
1015  *     4 se->sleep_start
1016  *     6 se->load.weight
1017  */
1018 struct sched_entity {
1019 	struct load_weight	load;		/* for load-balancing */
1020 	struct rb_node		run_node;
1021 	struct list_head	group_node;
1022 	unsigned int		on_rq;
1023 
1024 	u64			exec_start;
1025 	u64			sum_exec_runtime;
1026 	u64			vruntime;
1027 	u64			prev_sum_exec_runtime;
1028 
1029 	u64			last_wakeup;
1030 	u64			avg_overlap;
1031 
1032 #ifdef CONFIG_SCHEDSTATS
1033 	u64			wait_start;
1034 	u64			wait_max;
1035 	u64			wait_count;
1036 	u64			wait_sum;
1037 
1038 	u64			sleep_start;
1039 	u64			sleep_max;
1040 	s64			sum_sleep_runtime;
1041 
1042 	u64			block_start;
1043 	u64			block_max;
1044 	u64			exec_max;
1045 	u64			slice_max;
1046 
1047 	u64			nr_migrations;
1048 	u64			nr_migrations_cold;
1049 	u64			nr_failed_migrations_affine;
1050 	u64			nr_failed_migrations_running;
1051 	u64			nr_failed_migrations_hot;
1052 	u64			nr_forced_migrations;
1053 	u64			nr_forced2_migrations;
1054 
1055 	u64			nr_wakeups;
1056 	u64			nr_wakeups_sync;
1057 	u64			nr_wakeups_migrate;
1058 	u64			nr_wakeups_local;
1059 	u64			nr_wakeups_remote;
1060 	u64			nr_wakeups_affine;
1061 	u64			nr_wakeups_affine_attempts;
1062 	u64			nr_wakeups_passive;
1063 	u64			nr_wakeups_idle;
1064 #endif
1065 
1066 #ifdef CONFIG_FAIR_GROUP_SCHED
1067 	struct sched_entity	*parent;
1068 	/* rq on which this entity is (to be) queued: */
1069 	struct cfs_rq		*cfs_rq;
1070 	/* rq "owned" by this entity/group: */
1071 	struct cfs_rq		*my_q;
1072 #endif
1073 };
1074 
1075 struct sched_rt_entity {
1076 	struct list_head run_list;
1077 	unsigned long timeout;
1078 	unsigned int time_slice;
1079 	int nr_cpus_allowed;
1080 
1081 	struct sched_rt_entity *back;
1082 #ifdef CONFIG_RT_GROUP_SCHED
1083 	struct sched_rt_entity	*parent;
1084 	/* rq on which this entity is (to be) queued: */
1085 	struct rt_rq		*rt_rq;
1086 	/* rq "owned" by this entity/group: */
1087 	struct rt_rq		*my_q;
1088 #endif
1089 };
1090 
1091 struct task_struct {
1092 	volatile long state;	/* -1 unrunnable, 0 runnable, >0 stopped */
1093 	void *stack;
1094 	atomic_t usage;
1095 	unsigned int flags;	/* per process flags, defined below */
1096 	unsigned int ptrace;
1097 
1098 	int lock_depth;		/* BKL lock depth */
1099 
1100 #ifdef CONFIG_SMP
1101 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1102 	int oncpu;
1103 #endif
1104 #endif
1105 
1106 	int prio, static_prio, normal_prio;
1107 	unsigned int rt_priority;
1108 	const struct sched_class *sched_class;
1109 	struct sched_entity se;
1110 	struct sched_rt_entity rt;
1111 
1112 #ifdef CONFIG_PREEMPT_NOTIFIERS
1113 	/* list of struct preempt_notifier: */
1114 	struct hlist_head preempt_notifiers;
1115 #endif
1116 
1117 	/*
1118 	 * fpu_counter contains the number of consecutive context switches
1119 	 * that the FPU is used. If this is over a threshold, the lazy fpu
1120 	 * saving becomes unlazy to save the trap. This is an unsigned char
1121 	 * so that after 256 times the counter wraps and the behavior turns
1122 	 * lazy again; this to deal with bursty apps that only use FPU for
1123 	 * a short time
1124 	 */
1125 	unsigned char fpu_counter;
1126 	s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1127 #ifdef CONFIG_BLK_DEV_IO_TRACE
1128 	unsigned int btrace_seq;
1129 #endif
1130 
1131 	unsigned int policy;
1132 	cpumask_t cpus_allowed;
1133 
1134 #ifdef CONFIG_PREEMPT_RCU
1135 	int rcu_read_lock_nesting;
1136 	int rcu_flipctr_idx;
1137 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1138 
1139 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1140 	struct sched_info sched_info;
1141 #endif
1142 
1143 	struct list_head tasks;
1144 
1145 	struct mm_struct *mm, *active_mm;
1146 
1147 /* task state */
1148 	struct linux_binfmt *binfmt;
1149 	int exit_state;
1150 	int exit_code, exit_signal;
1151 	int pdeath_signal;  /*  The signal sent when the parent dies  */
1152 	/* ??? */
1153 	unsigned int personality;
1154 	unsigned did_exec:1;
1155 	pid_t pid;
1156 	pid_t tgid;
1157 
1158 #ifdef CONFIG_CC_STACKPROTECTOR
1159 	/* Canary value for the -fstack-protector gcc feature */
1160 	unsigned long stack_canary;
1161 #endif
1162 	/*
1163 	 * pointers to (original) parent process, youngest child, younger sibling,
1164 	 * older sibling, respectively.  (p->father can be replaced with
1165 	 * p->real_parent->pid)
1166 	 */
1167 	struct task_struct *real_parent; /* real parent process */
1168 	struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1169 	/*
1170 	 * children/sibling forms the list of my natural children
1171 	 */
1172 	struct list_head children;	/* list of my children */
1173 	struct list_head sibling;	/* linkage in my parent's children list */
1174 	struct task_struct *group_leader;	/* threadgroup leader */
1175 
1176 	/*
1177 	 * ptraced is the list of tasks this task is using ptrace on.
1178 	 * This includes both natural children and PTRACE_ATTACH targets.
1179 	 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1180 	 */
1181 	struct list_head ptraced;
1182 	struct list_head ptrace_entry;
1183 
1184 #ifdef CONFIG_X86_PTRACE_BTS
1185 	/*
1186 	 * This is the tracer handle for the ptrace BTS extension.
1187 	 * This field actually belongs to the ptracer task.
1188 	 */
1189 	struct bts_tracer *bts;
1190 	/*
1191 	 * The buffer to hold the BTS data.
1192 	 */
1193 	void *bts_buffer;
1194 	size_t bts_size;
1195 #endif /* CONFIG_X86_PTRACE_BTS */
1196 
1197 	/* PID/PID hash table linkage. */
1198 	struct pid_link pids[PIDTYPE_MAX];
1199 	struct list_head thread_group;
1200 
1201 	struct completion *vfork_done;		/* for vfork() */
1202 	int __user *set_child_tid;		/* CLONE_CHILD_SETTID */
1203 	int __user *clear_child_tid;		/* CLONE_CHILD_CLEARTID */
1204 
1205 	cputime_t utime, stime, utimescaled, stimescaled;
1206 	cputime_t gtime;
1207 	cputime_t prev_utime, prev_stime;
1208 	unsigned long nvcsw, nivcsw; /* context switch counts */
1209 	struct timespec start_time; 		/* monotonic time */
1210 	struct timespec real_start_time;	/* boot based time */
1211 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1212 	unsigned long min_flt, maj_flt;
1213 
1214 	struct task_cputime cputime_expires;
1215 	struct list_head cpu_timers[3];
1216 
1217 /* process credentials */
1218 	const struct cred *real_cred;	/* objective and real subjective task
1219 					 * credentials (COW) */
1220 	const struct cred *cred;	/* effective (overridable) subjective task
1221 					 * credentials (COW) */
1222 	struct mutex cred_exec_mutex;	/* execve vs ptrace cred calculation mutex */
1223 
1224 	char comm[TASK_COMM_LEN]; /* executable name excluding path
1225 				     - access with [gs]et_task_comm (which lock
1226 				       it with task_lock())
1227 				     - initialized normally by flush_old_exec */
1228 /* file system info */
1229 	int link_count, total_link_count;
1230 #ifdef CONFIG_SYSVIPC
1231 /* ipc stuff */
1232 	struct sysv_sem sysvsem;
1233 #endif
1234 #ifdef CONFIG_DETECT_SOFTLOCKUP
1235 /* hung task detection */
1236 	unsigned long last_switch_timestamp;
1237 	unsigned long last_switch_count;
1238 #endif
1239 /* CPU-specific state of this task */
1240 	struct thread_struct thread;
1241 /* filesystem information */
1242 	struct fs_struct *fs;
1243 /* open file information */
1244 	struct files_struct *files;
1245 /* namespaces */
1246 	struct nsproxy *nsproxy;
1247 /* signal handlers */
1248 	struct signal_struct *signal;
1249 	struct sighand_struct *sighand;
1250 
1251 	sigset_t blocked, real_blocked;
1252 	sigset_t saved_sigmask;	/* restored if set_restore_sigmask() was used */
1253 	struct sigpending pending;
1254 
1255 	unsigned long sas_ss_sp;
1256 	size_t sas_ss_size;
1257 	int (*notifier)(void *priv);
1258 	void *notifier_data;
1259 	sigset_t *notifier_mask;
1260 	struct audit_context *audit_context;
1261 #ifdef CONFIG_AUDITSYSCALL
1262 	uid_t loginuid;
1263 	unsigned int sessionid;
1264 #endif
1265 	seccomp_t seccomp;
1266 
1267 /* Thread group tracking */
1268    	u32 parent_exec_id;
1269    	u32 self_exec_id;
1270 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1271 	spinlock_t alloc_lock;
1272 
1273 	/* Protection of the PI data structures: */
1274 	spinlock_t pi_lock;
1275 
1276 #ifdef CONFIG_RT_MUTEXES
1277 	/* PI waiters blocked on a rt_mutex held by this task */
1278 	struct plist_head pi_waiters;
1279 	/* Deadlock detection and priority inheritance handling */
1280 	struct rt_mutex_waiter *pi_blocked_on;
1281 #endif
1282 
1283 #ifdef CONFIG_DEBUG_MUTEXES
1284 	/* mutex deadlock detection */
1285 	struct mutex_waiter *blocked_on;
1286 #endif
1287 #ifdef CONFIG_TRACE_IRQFLAGS
1288 	unsigned int irq_events;
1289 	int hardirqs_enabled;
1290 	unsigned long hardirq_enable_ip;
1291 	unsigned int hardirq_enable_event;
1292 	unsigned long hardirq_disable_ip;
1293 	unsigned int hardirq_disable_event;
1294 	int softirqs_enabled;
1295 	unsigned long softirq_disable_ip;
1296 	unsigned int softirq_disable_event;
1297 	unsigned long softirq_enable_ip;
1298 	unsigned int softirq_enable_event;
1299 	int hardirq_context;
1300 	int softirq_context;
1301 #endif
1302 #ifdef CONFIG_LOCKDEP
1303 # define MAX_LOCK_DEPTH 48UL
1304 	u64 curr_chain_key;
1305 	int lockdep_depth;
1306 	unsigned int lockdep_recursion;
1307 	struct held_lock held_locks[MAX_LOCK_DEPTH];
1308 #endif
1309 
1310 /* journalling filesystem info */
1311 	void *journal_info;
1312 
1313 /* stacked block device info */
1314 	struct bio *bio_list, **bio_tail;
1315 
1316 /* VM state */
1317 	struct reclaim_state *reclaim_state;
1318 
1319 	struct backing_dev_info *backing_dev_info;
1320 
1321 	struct io_context *io_context;
1322 
1323 	unsigned long ptrace_message;
1324 	siginfo_t *last_siginfo; /* For ptrace use.  */
1325 	struct task_io_accounting ioac;
1326 #if defined(CONFIG_TASK_XACCT)
1327 	u64 acct_rss_mem1;	/* accumulated rss usage */
1328 	u64 acct_vm_mem1;	/* accumulated virtual memory usage */
1329 	cputime_t acct_timexpd;	/* stime + utime since last update */
1330 #endif
1331 #ifdef CONFIG_CPUSETS
1332 	nodemask_t mems_allowed;
1333 	int cpuset_mems_generation;
1334 	int cpuset_mem_spread_rotor;
1335 #endif
1336 #ifdef CONFIG_CGROUPS
1337 	/* Control Group info protected by css_set_lock */
1338 	struct css_set *cgroups;
1339 	/* cg_list protected by css_set_lock and tsk->alloc_lock */
1340 	struct list_head cg_list;
1341 #endif
1342 #ifdef CONFIG_FUTEX
1343 	struct robust_list_head __user *robust_list;
1344 #ifdef CONFIG_COMPAT
1345 	struct compat_robust_list_head __user *compat_robust_list;
1346 #endif
1347 	struct list_head pi_state_list;
1348 	struct futex_pi_state *pi_state_cache;
1349 #endif
1350 #ifdef CONFIG_NUMA
1351 	struct mempolicy *mempolicy;
1352 	short il_next;
1353 #endif
1354 	atomic_t fs_excl;	/* holding fs exclusive resources */
1355 	struct rcu_head rcu;
1356 
1357 	/*
1358 	 * cache last used pipe for splice
1359 	 */
1360 	struct pipe_inode_info *splice_pipe;
1361 #ifdef	CONFIG_TASK_DELAY_ACCT
1362 	struct task_delay_info *delays;
1363 #endif
1364 #ifdef CONFIG_FAULT_INJECTION
1365 	int make_it_fail;
1366 #endif
1367 	struct prop_local_single dirties;
1368 #ifdef CONFIG_LATENCYTOP
1369 	int latency_record_count;
1370 	struct latency_record latency_record[LT_SAVECOUNT];
1371 #endif
1372 	/*
1373 	 * time slack values; these are used to round up poll() and
1374 	 * select() etc timeout values. These are in nanoseconds.
1375 	 */
1376 	unsigned long timer_slack_ns;
1377 	unsigned long default_timer_slack_ns;
1378 
1379 	struct list_head	*scm_work_list;
1380 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1381 	/* Index of current stored adress in ret_stack */
1382 	int curr_ret_stack;
1383 	/* Stack of return addresses for return function tracing */
1384 	struct ftrace_ret_stack	*ret_stack;
1385 	/*
1386 	 * Number of functions that haven't been traced
1387 	 * because of depth overrun.
1388 	 */
1389 	atomic_t trace_overrun;
1390 	/* Pause for the tracing */
1391 	atomic_t tracing_graph_pause;
1392 #endif
1393 #ifdef CONFIG_TRACING
1394 	/* state flags for use by tracers */
1395 	unsigned long trace;
1396 #endif
1397 };
1398 
1399 /*
1400  * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1401  * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1402  * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1403  * values are inverted: lower p->prio value means higher priority.
1404  *
1405  * The MAX_USER_RT_PRIO value allows the actual maximum
1406  * RT priority to be separate from the value exported to
1407  * user-space.  This allows kernel threads to set their
1408  * priority to a value higher than any user task. Note:
1409  * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1410  */
1411 
1412 #define MAX_USER_RT_PRIO	100
1413 #define MAX_RT_PRIO		MAX_USER_RT_PRIO
1414 
1415 #define MAX_PRIO		(MAX_RT_PRIO + 40)
1416 #define DEFAULT_PRIO		(MAX_RT_PRIO + 20)
1417 
1418 static inline int rt_prio(int prio)
1419 {
1420 	if (unlikely(prio < MAX_RT_PRIO))
1421 		return 1;
1422 	return 0;
1423 }
1424 
1425 static inline int rt_task(struct task_struct *p)
1426 {
1427 	return rt_prio(p->prio);
1428 }
1429 
1430 static inline void set_task_session(struct task_struct *tsk, pid_t session)
1431 {
1432 	tsk->signal->__session = session;
1433 }
1434 
1435 static inline void set_task_pgrp(struct task_struct *tsk, pid_t pgrp)
1436 {
1437 	tsk->signal->__pgrp = pgrp;
1438 }
1439 
1440 static inline struct pid *task_pid(struct task_struct *task)
1441 {
1442 	return task->pids[PIDTYPE_PID].pid;
1443 }
1444 
1445 static inline struct pid *task_tgid(struct task_struct *task)
1446 {
1447 	return task->group_leader->pids[PIDTYPE_PID].pid;
1448 }
1449 
1450 static inline struct pid *task_pgrp(struct task_struct *task)
1451 {
1452 	return task->group_leader->pids[PIDTYPE_PGID].pid;
1453 }
1454 
1455 static inline struct pid *task_session(struct task_struct *task)
1456 {
1457 	return task->group_leader->pids[PIDTYPE_SID].pid;
1458 }
1459 
1460 struct pid_namespace;
1461 
1462 /*
1463  * the helpers to get the task's different pids as they are seen
1464  * from various namespaces
1465  *
1466  * task_xid_nr()     : global id, i.e. the id seen from the init namespace;
1467  * task_xid_vnr()    : virtual id, i.e. the id seen from the pid namespace of
1468  *                     current.
1469  * task_xid_nr_ns()  : id seen from the ns specified;
1470  *
1471  * set_task_vxid()   : assigns a virtual id to a task;
1472  *
1473  * see also pid_nr() etc in include/linux/pid.h
1474  */
1475 
1476 static inline pid_t task_pid_nr(struct task_struct *tsk)
1477 {
1478 	return tsk->pid;
1479 }
1480 
1481 pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1482 
1483 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1484 {
1485 	return pid_vnr(task_pid(tsk));
1486 }
1487 
1488 
1489 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1490 {
1491 	return tsk->tgid;
1492 }
1493 
1494 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1495 
1496 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1497 {
1498 	return pid_vnr(task_tgid(tsk));
1499 }
1500 
1501 
1502 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1503 {
1504 	return tsk->signal->__pgrp;
1505 }
1506 
1507 pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1508 
1509 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1510 {
1511 	return pid_vnr(task_pgrp(tsk));
1512 }
1513 
1514 
1515 static inline pid_t task_session_nr(struct task_struct *tsk)
1516 {
1517 	return tsk->signal->__session;
1518 }
1519 
1520 pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1521 
1522 static inline pid_t task_session_vnr(struct task_struct *tsk)
1523 {
1524 	return pid_vnr(task_session(tsk));
1525 }
1526 
1527 
1528 /**
1529  * pid_alive - check that a task structure is not stale
1530  * @p: Task structure to be checked.
1531  *
1532  * Test if a process is not yet dead (at most zombie state)
1533  * If pid_alive fails, then pointers within the task structure
1534  * can be stale and must not be dereferenced.
1535  */
1536 static inline int pid_alive(struct task_struct *p)
1537 {
1538 	return p->pids[PIDTYPE_PID].pid != NULL;
1539 }
1540 
1541 /**
1542  * is_global_init - check if a task structure is init
1543  * @tsk: Task structure to be checked.
1544  *
1545  * Check if a task structure is the first user space task the kernel created.
1546  */
1547 static inline int is_global_init(struct task_struct *tsk)
1548 {
1549 	return tsk->pid == 1;
1550 }
1551 
1552 /*
1553  * is_container_init:
1554  * check whether in the task is init in its own pid namespace.
1555  */
1556 extern int is_container_init(struct task_struct *tsk);
1557 
1558 extern struct pid *cad_pid;
1559 
1560 extern void free_task(struct task_struct *tsk);
1561 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1562 
1563 extern void __put_task_struct(struct task_struct *t);
1564 
1565 static inline void put_task_struct(struct task_struct *t)
1566 {
1567 	if (atomic_dec_and_test(&t->usage))
1568 		__put_task_struct(t);
1569 }
1570 
1571 extern cputime_t task_utime(struct task_struct *p);
1572 extern cputime_t task_stime(struct task_struct *p);
1573 extern cputime_t task_gtime(struct task_struct *p);
1574 
1575 /*
1576  * Per process flags
1577  */
1578 #define PF_ALIGNWARN	0x00000001	/* Print alignment warning msgs */
1579 					/* Not implemented yet, only for 486*/
1580 #define PF_STARTING	0x00000002	/* being created */
1581 #define PF_EXITING	0x00000004	/* getting shut down */
1582 #define PF_EXITPIDONE	0x00000008	/* pi exit done on shut down */
1583 #define PF_VCPU		0x00000010	/* I'm a virtual CPU */
1584 #define PF_FORKNOEXEC	0x00000040	/* forked but didn't exec */
1585 #define PF_SUPERPRIV	0x00000100	/* used super-user privileges */
1586 #define PF_DUMPCORE	0x00000200	/* dumped core */
1587 #define PF_SIGNALED	0x00000400	/* killed by a signal */
1588 #define PF_MEMALLOC	0x00000800	/* Allocating memory */
1589 #define PF_FLUSHER	0x00001000	/* responsible for disk writeback */
1590 #define PF_USED_MATH	0x00002000	/* if unset the fpu must be initialized before use */
1591 #define PF_NOFREEZE	0x00008000	/* this thread should not be frozen */
1592 #define PF_FROZEN	0x00010000	/* frozen for system suspend */
1593 #define PF_FSTRANS	0x00020000	/* inside a filesystem transaction */
1594 #define PF_KSWAPD	0x00040000	/* I am kswapd */
1595 #define PF_SWAPOFF	0x00080000	/* I am in swapoff */
1596 #define PF_LESS_THROTTLE 0x00100000	/* Throttle me less: I clean memory */
1597 #define PF_KTHREAD	0x00200000	/* I am a kernel thread */
1598 #define PF_RANDOMIZE	0x00400000	/* randomize virtual address space */
1599 #define PF_SWAPWRITE	0x00800000	/* Allowed to write to swap */
1600 #define PF_SPREAD_PAGE	0x01000000	/* Spread page cache over cpuset */
1601 #define PF_SPREAD_SLAB	0x02000000	/* Spread some slab caches over cpuset */
1602 #define PF_THREAD_BOUND	0x04000000	/* Thread bound to specific cpu */
1603 #define PF_MEMPOLICY	0x10000000	/* Non-default NUMA mempolicy */
1604 #define PF_MUTEX_TESTER	0x20000000	/* Thread belongs to the rt mutex tester */
1605 #define PF_FREEZER_SKIP	0x40000000	/* Freezer should not count it as freezeable */
1606 #define PF_FREEZER_NOSIG 0x80000000	/* Freezer won't send signals to it */
1607 
1608 /*
1609  * Only the _current_ task can read/write to tsk->flags, but other
1610  * tasks can access tsk->flags in readonly mode for example
1611  * with tsk_used_math (like during threaded core dumping).
1612  * There is however an exception to this rule during ptrace
1613  * or during fork: the ptracer task is allowed to write to the
1614  * child->flags of its traced child (same goes for fork, the parent
1615  * can write to the child->flags), because we're guaranteed the
1616  * child is not running and in turn not changing child->flags
1617  * at the same time the parent does it.
1618  */
1619 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1620 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1621 #define clear_used_math() clear_stopped_child_used_math(current)
1622 #define set_used_math() set_stopped_child_used_math(current)
1623 #define conditional_stopped_child_used_math(condition, child) \
1624 	do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1625 #define conditional_used_math(condition) \
1626 	conditional_stopped_child_used_math(condition, current)
1627 #define copy_to_stopped_child_used_math(child) \
1628 	do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1629 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1630 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1631 #define used_math() tsk_used_math(current)
1632 
1633 #ifdef CONFIG_SMP
1634 extern int set_cpus_allowed_ptr(struct task_struct *p,
1635 				const struct cpumask *new_mask);
1636 #else
1637 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1638 				       const struct cpumask *new_mask)
1639 {
1640 	if (!cpumask_test_cpu(0, new_mask))
1641 		return -EINVAL;
1642 	return 0;
1643 }
1644 #endif
1645 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1646 {
1647 	return set_cpus_allowed_ptr(p, &new_mask);
1648 }
1649 
1650 extern unsigned long long sched_clock(void);
1651 
1652 extern void sched_clock_init(void);
1653 extern u64 sched_clock_cpu(int cpu);
1654 
1655 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1656 static inline void sched_clock_tick(void)
1657 {
1658 }
1659 
1660 static inline void sched_clock_idle_sleep_event(void)
1661 {
1662 }
1663 
1664 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1665 {
1666 }
1667 #else
1668 extern void sched_clock_tick(void);
1669 extern void sched_clock_idle_sleep_event(void);
1670 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1671 #endif
1672 
1673 /*
1674  * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1675  * clock constructed from sched_clock():
1676  */
1677 extern unsigned long long cpu_clock(int cpu);
1678 
1679 extern unsigned long long
1680 task_sched_runtime(struct task_struct *task);
1681 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1682 
1683 /* sched_exec is called by processes performing an exec */
1684 #ifdef CONFIG_SMP
1685 extern void sched_exec(void);
1686 #else
1687 #define sched_exec()   {}
1688 #endif
1689 
1690 extern void sched_clock_idle_sleep_event(void);
1691 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1692 
1693 #ifdef CONFIG_HOTPLUG_CPU
1694 extern void idle_task_exit(void);
1695 #else
1696 static inline void idle_task_exit(void) {}
1697 #endif
1698 
1699 extern void sched_idle_next(void);
1700 
1701 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1702 extern void wake_up_idle_cpu(int cpu);
1703 #else
1704 static inline void wake_up_idle_cpu(int cpu) { }
1705 #endif
1706 
1707 #ifdef CONFIG_SCHED_DEBUG
1708 extern unsigned int sysctl_sched_latency;
1709 extern unsigned int sysctl_sched_min_granularity;
1710 extern unsigned int sysctl_sched_wakeup_granularity;
1711 extern unsigned int sysctl_sched_child_runs_first;
1712 extern unsigned int sysctl_sched_features;
1713 extern unsigned int sysctl_sched_migration_cost;
1714 extern unsigned int sysctl_sched_nr_migrate;
1715 extern unsigned int sysctl_sched_shares_ratelimit;
1716 extern unsigned int sysctl_sched_shares_thresh;
1717 
1718 int sched_nr_latency_handler(struct ctl_table *table, int write,
1719 		struct file *file, void __user *buffer, size_t *length,
1720 		loff_t *ppos);
1721 #endif
1722 extern unsigned int sysctl_sched_rt_period;
1723 extern int sysctl_sched_rt_runtime;
1724 
1725 int sched_rt_handler(struct ctl_table *table, int write,
1726 		struct file *filp, void __user *buffer, size_t *lenp,
1727 		loff_t *ppos);
1728 
1729 extern unsigned int sysctl_sched_compat_yield;
1730 
1731 #ifdef CONFIG_RT_MUTEXES
1732 extern int rt_mutex_getprio(struct task_struct *p);
1733 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1734 extern void rt_mutex_adjust_pi(struct task_struct *p);
1735 #else
1736 static inline int rt_mutex_getprio(struct task_struct *p)
1737 {
1738 	return p->normal_prio;
1739 }
1740 # define rt_mutex_adjust_pi(p)		do { } while (0)
1741 #endif
1742 
1743 extern void set_user_nice(struct task_struct *p, long nice);
1744 extern int task_prio(const struct task_struct *p);
1745 extern int task_nice(const struct task_struct *p);
1746 extern int can_nice(const struct task_struct *p, const int nice);
1747 extern int task_curr(const struct task_struct *p);
1748 extern int idle_cpu(int cpu);
1749 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1750 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1751 				      struct sched_param *);
1752 extern struct task_struct *idle_task(int cpu);
1753 extern struct task_struct *curr_task(int cpu);
1754 extern void set_curr_task(int cpu, struct task_struct *p);
1755 
1756 void yield(void);
1757 
1758 /*
1759  * The default (Linux) execution domain.
1760  */
1761 extern struct exec_domain	default_exec_domain;
1762 
1763 union thread_union {
1764 	struct thread_info thread_info;
1765 	unsigned long stack[THREAD_SIZE/sizeof(long)];
1766 };
1767 
1768 #ifndef __HAVE_ARCH_KSTACK_END
1769 static inline int kstack_end(void *addr)
1770 {
1771 	/* Reliable end of stack detection:
1772 	 * Some APM bios versions misalign the stack
1773 	 */
1774 	return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1775 }
1776 #endif
1777 
1778 extern union thread_union init_thread_union;
1779 extern struct task_struct init_task;
1780 
1781 extern struct   mm_struct init_mm;
1782 
1783 extern struct pid_namespace init_pid_ns;
1784 
1785 /*
1786  * find a task by one of its numerical ids
1787  *
1788  * find_task_by_pid_type_ns():
1789  *      it is the most generic call - it finds a task by all id,
1790  *      type and namespace specified
1791  * find_task_by_pid_ns():
1792  *      finds a task by its pid in the specified namespace
1793  * find_task_by_vpid():
1794  *      finds a task by its virtual pid
1795  *
1796  * see also find_vpid() etc in include/linux/pid.h
1797  */
1798 
1799 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1800 		struct pid_namespace *ns);
1801 
1802 extern struct task_struct *find_task_by_vpid(pid_t nr);
1803 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1804 		struct pid_namespace *ns);
1805 
1806 extern void __set_special_pids(struct pid *pid);
1807 
1808 /* per-UID process charging. */
1809 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1810 static inline struct user_struct *get_uid(struct user_struct *u)
1811 {
1812 	atomic_inc(&u->__count);
1813 	return u;
1814 }
1815 extern void free_uid(struct user_struct *);
1816 extern void release_uids(struct user_namespace *ns);
1817 
1818 #include <asm/current.h>
1819 
1820 extern void do_timer(unsigned long ticks);
1821 
1822 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1823 extern int wake_up_process(struct task_struct *tsk);
1824 extern void wake_up_new_task(struct task_struct *tsk,
1825 				unsigned long clone_flags);
1826 #ifdef CONFIG_SMP
1827  extern void kick_process(struct task_struct *tsk);
1828 #else
1829  static inline void kick_process(struct task_struct *tsk) { }
1830 #endif
1831 extern void sched_fork(struct task_struct *p, int clone_flags);
1832 extern void sched_dead(struct task_struct *p);
1833 
1834 extern void proc_caches_init(void);
1835 extern void flush_signals(struct task_struct *);
1836 extern void ignore_signals(struct task_struct *);
1837 extern void flush_signal_handlers(struct task_struct *, int force_default);
1838 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1839 
1840 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1841 {
1842 	unsigned long flags;
1843 	int ret;
1844 
1845 	spin_lock_irqsave(&tsk->sighand->siglock, flags);
1846 	ret = dequeue_signal(tsk, mask, info);
1847 	spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1848 
1849 	return ret;
1850 }
1851 
1852 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1853 			      sigset_t *mask);
1854 extern void unblock_all_signals(void);
1855 extern void release_task(struct task_struct * p);
1856 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1857 extern int force_sigsegv(int, struct task_struct *);
1858 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1859 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1860 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1861 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1862 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1863 extern int kill_pid(struct pid *pid, int sig, int priv);
1864 extern int kill_proc_info(int, struct siginfo *, pid_t);
1865 extern int do_notify_parent(struct task_struct *, int);
1866 extern void force_sig(int, struct task_struct *);
1867 extern void force_sig_specific(int, struct task_struct *);
1868 extern int send_sig(int, struct task_struct *, int);
1869 extern void zap_other_threads(struct task_struct *p);
1870 extern struct sigqueue *sigqueue_alloc(void);
1871 extern void sigqueue_free(struct sigqueue *);
1872 extern int send_sigqueue(struct sigqueue *,  struct task_struct *, int group);
1873 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1874 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1875 
1876 static inline int kill_cad_pid(int sig, int priv)
1877 {
1878 	return kill_pid(cad_pid, sig, priv);
1879 }
1880 
1881 /* These can be the second arg to send_sig_info/send_group_sig_info.  */
1882 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1883 #define SEND_SIG_PRIV	((struct siginfo *) 1)
1884 #define SEND_SIG_FORCED	((struct siginfo *) 2)
1885 
1886 static inline int is_si_special(const struct siginfo *info)
1887 {
1888 	return info <= SEND_SIG_FORCED;
1889 }
1890 
1891 /* True if we are on the alternate signal stack.  */
1892 
1893 static inline int on_sig_stack(unsigned long sp)
1894 {
1895 	return (sp - current->sas_ss_sp < current->sas_ss_size);
1896 }
1897 
1898 static inline int sas_ss_flags(unsigned long sp)
1899 {
1900 	return (current->sas_ss_size == 0 ? SS_DISABLE
1901 		: on_sig_stack(sp) ? SS_ONSTACK : 0);
1902 }
1903 
1904 /*
1905  * Routines for handling mm_structs
1906  */
1907 extern struct mm_struct * mm_alloc(void);
1908 
1909 /* mmdrop drops the mm and the page tables */
1910 extern void __mmdrop(struct mm_struct *);
1911 static inline void mmdrop(struct mm_struct * mm)
1912 {
1913 	if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1914 		__mmdrop(mm);
1915 }
1916 
1917 /* mmput gets rid of the mappings and all user-space */
1918 extern void mmput(struct mm_struct *);
1919 /* Grab a reference to a task's mm, if it is not already going away */
1920 extern struct mm_struct *get_task_mm(struct task_struct *task);
1921 /* Remove the current tasks stale references to the old mm_struct */
1922 extern void mm_release(struct task_struct *, struct mm_struct *);
1923 /* Allocate a new mm structure and copy contents from tsk->mm */
1924 extern struct mm_struct *dup_mm(struct task_struct *tsk);
1925 
1926 extern int  copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1927 extern void flush_thread(void);
1928 extern void exit_thread(void);
1929 
1930 extern void exit_files(struct task_struct *);
1931 extern void __cleanup_signal(struct signal_struct *);
1932 extern void __cleanup_sighand(struct sighand_struct *);
1933 
1934 extern void exit_itimers(struct signal_struct *);
1935 extern void flush_itimer_signals(void);
1936 
1937 extern NORET_TYPE void do_group_exit(int);
1938 
1939 extern void daemonize(const char *, ...);
1940 extern int allow_signal(int);
1941 extern int disallow_signal(int);
1942 
1943 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1944 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1945 struct task_struct *fork_idle(int);
1946 
1947 extern void set_task_comm(struct task_struct *tsk, char *from);
1948 extern char *get_task_comm(char *to, struct task_struct *tsk);
1949 
1950 #ifdef CONFIG_SMP
1951 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1952 #else
1953 static inline unsigned long wait_task_inactive(struct task_struct *p,
1954 					       long match_state)
1955 {
1956 	return 1;
1957 }
1958 #endif
1959 
1960 #define next_task(p)	list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1961 
1962 #define for_each_process(p) \
1963 	for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1964 
1965 extern bool is_single_threaded(struct task_struct *);
1966 
1967 /*
1968  * Careful: do_each_thread/while_each_thread is a double loop so
1969  *          'break' will not work as expected - use goto instead.
1970  */
1971 #define do_each_thread(g, t) \
1972 	for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1973 
1974 #define while_each_thread(g, t) \
1975 	while ((t = next_thread(t)) != g)
1976 
1977 /* de_thread depends on thread_group_leader not being a pid based check */
1978 #define thread_group_leader(p)	(p == p->group_leader)
1979 
1980 /* Do to the insanities of de_thread it is possible for a process
1981  * to have the pid of the thread group leader without actually being
1982  * the thread group leader.  For iteration through the pids in proc
1983  * all we care about is that we have a task with the appropriate
1984  * pid, we don't actually care if we have the right task.
1985  */
1986 static inline int has_group_leader_pid(struct task_struct *p)
1987 {
1988 	return p->pid == p->tgid;
1989 }
1990 
1991 static inline
1992 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
1993 {
1994 	return p1->tgid == p2->tgid;
1995 }
1996 
1997 static inline struct task_struct *next_thread(const struct task_struct *p)
1998 {
1999 	return list_entry(rcu_dereference(p->thread_group.next),
2000 			  struct task_struct, thread_group);
2001 }
2002 
2003 static inline int thread_group_empty(struct task_struct *p)
2004 {
2005 	return list_empty(&p->thread_group);
2006 }
2007 
2008 #define delay_group_leader(p) \
2009 		(thread_group_leader(p) && !thread_group_empty(p))
2010 
2011 /*
2012  * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2013  * subscriptions and synchronises with wait4().  Also used in procfs.  Also
2014  * pins the final release of task.io_context.  Also protects ->cpuset and
2015  * ->cgroup.subsys[].
2016  *
2017  * Nests both inside and outside of read_lock(&tasklist_lock).
2018  * It must not be nested with write_lock_irq(&tasklist_lock),
2019  * neither inside nor outside.
2020  */
2021 static inline void task_lock(struct task_struct *p)
2022 {
2023 	spin_lock(&p->alloc_lock);
2024 }
2025 
2026 static inline void task_unlock(struct task_struct *p)
2027 {
2028 	spin_unlock(&p->alloc_lock);
2029 }
2030 
2031 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2032 							unsigned long *flags);
2033 
2034 static inline void unlock_task_sighand(struct task_struct *tsk,
2035 						unsigned long *flags)
2036 {
2037 	spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2038 }
2039 
2040 #ifndef __HAVE_THREAD_FUNCTIONS
2041 
2042 #define task_thread_info(task)	((struct thread_info *)(task)->stack)
2043 #define task_stack_page(task)	((task)->stack)
2044 
2045 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2046 {
2047 	*task_thread_info(p) = *task_thread_info(org);
2048 	task_thread_info(p)->task = p;
2049 }
2050 
2051 static inline unsigned long *end_of_stack(struct task_struct *p)
2052 {
2053 	return (unsigned long *)(task_thread_info(p) + 1);
2054 }
2055 
2056 #endif
2057 
2058 static inline int object_is_on_stack(void *obj)
2059 {
2060 	void *stack = task_stack_page(current);
2061 
2062 	return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2063 }
2064 
2065 extern void thread_info_cache_init(void);
2066 
2067 /* set thread flags in other task's structures
2068  * - see asm/thread_info.h for TIF_xxxx flags available
2069  */
2070 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2071 {
2072 	set_ti_thread_flag(task_thread_info(tsk), flag);
2073 }
2074 
2075 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2076 {
2077 	clear_ti_thread_flag(task_thread_info(tsk), flag);
2078 }
2079 
2080 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2081 {
2082 	return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2083 }
2084 
2085 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2086 {
2087 	return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2088 }
2089 
2090 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2091 {
2092 	return test_ti_thread_flag(task_thread_info(tsk), flag);
2093 }
2094 
2095 static inline void set_tsk_need_resched(struct task_struct *tsk)
2096 {
2097 	set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2098 }
2099 
2100 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2101 {
2102 	clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2103 }
2104 
2105 static inline int test_tsk_need_resched(struct task_struct *tsk)
2106 {
2107 	return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2108 }
2109 
2110 static inline int signal_pending(struct task_struct *p)
2111 {
2112 	return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2113 }
2114 
2115 extern int __fatal_signal_pending(struct task_struct *p);
2116 
2117 static inline int fatal_signal_pending(struct task_struct *p)
2118 {
2119 	return signal_pending(p) && __fatal_signal_pending(p);
2120 }
2121 
2122 static inline int signal_pending_state(long state, struct task_struct *p)
2123 {
2124 	if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2125 		return 0;
2126 	if (!signal_pending(p))
2127 		return 0;
2128 
2129 	return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2130 }
2131 
2132 static inline int need_resched(void)
2133 {
2134 	return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2135 }
2136 
2137 /*
2138  * cond_resched() and cond_resched_lock(): latency reduction via
2139  * explicit rescheduling in places that are safe. The return
2140  * value indicates whether a reschedule was done in fact.
2141  * cond_resched_lock() will drop the spinlock before scheduling,
2142  * cond_resched_softirq() will enable bhs before scheduling.
2143  */
2144 extern int _cond_resched(void);
2145 #ifdef CONFIG_PREEMPT_BKL
2146 static inline int cond_resched(void)
2147 {
2148 	return 0;
2149 }
2150 #else
2151 static inline int cond_resched(void)
2152 {
2153 	return _cond_resched();
2154 }
2155 #endif
2156 extern int cond_resched_lock(spinlock_t * lock);
2157 extern int cond_resched_softirq(void);
2158 static inline int cond_resched_bkl(void)
2159 {
2160 	return _cond_resched();
2161 }
2162 
2163 /*
2164  * Does a critical section need to be broken due to another
2165  * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2166  * but a general need for low latency)
2167  */
2168 static inline int spin_needbreak(spinlock_t *lock)
2169 {
2170 #ifdef CONFIG_PREEMPT
2171 	return spin_is_contended(lock);
2172 #else
2173 	return 0;
2174 #endif
2175 }
2176 
2177 /*
2178  * Thread group CPU time accounting.
2179  */
2180 
2181 extern int thread_group_cputime_alloc(struct task_struct *);
2182 extern void thread_group_cputime(struct task_struct *, struct task_cputime *);
2183 
2184 static inline void thread_group_cputime_init(struct signal_struct *sig)
2185 {
2186 	sig->cputime.totals = NULL;
2187 }
2188 
2189 static inline int thread_group_cputime_clone_thread(struct task_struct *curr)
2190 {
2191 	if (curr->signal->cputime.totals)
2192 		return 0;
2193 	return thread_group_cputime_alloc(curr);
2194 }
2195 
2196 static inline void thread_group_cputime_free(struct signal_struct *sig)
2197 {
2198 	free_percpu(sig->cputime.totals);
2199 }
2200 
2201 /*
2202  * Reevaluate whether the task has signals pending delivery.
2203  * Wake the task if so.
2204  * This is required every time the blocked sigset_t changes.
2205  * callers must hold sighand->siglock.
2206  */
2207 extern void recalc_sigpending_and_wake(struct task_struct *t);
2208 extern void recalc_sigpending(void);
2209 
2210 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2211 
2212 /*
2213  * Wrappers for p->thread_info->cpu access. No-op on UP.
2214  */
2215 #ifdef CONFIG_SMP
2216 
2217 static inline unsigned int task_cpu(const struct task_struct *p)
2218 {
2219 	return task_thread_info(p)->cpu;
2220 }
2221 
2222 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2223 
2224 #else
2225 
2226 static inline unsigned int task_cpu(const struct task_struct *p)
2227 {
2228 	return 0;
2229 }
2230 
2231 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2232 {
2233 }
2234 
2235 #endif /* CONFIG_SMP */
2236 
2237 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2238 
2239 #ifdef CONFIG_TRACING
2240 extern void
2241 __trace_special(void *__tr, void *__data,
2242 		unsigned long arg1, unsigned long arg2, unsigned long arg3);
2243 #else
2244 static inline void
2245 __trace_special(void *__tr, void *__data,
2246 		unsigned long arg1, unsigned long arg2, unsigned long arg3)
2247 {
2248 }
2249 #endif
2250 
2251 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2252 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2253 
2254 extern void normalize_rt_tasks(void);
2255 
2256 #ifdef CONFIG_GROUP_SCHED
2257 
2258 extern struct task_group init_task_group;
2259 #ifdef CONFIG_USER_SCHED
2260 extern struct task_group root_task_group;
2261 extern void set_tg_uid(struct user_struct *user);
2262 #endif
2263 
2264 extern struct task_group *sched_create_group(struct task_group *parent);
2265 extern void sched_destroy_group(struct task_group *tg);
2266 extern void sched_move_task(struct task_struct *tsk);
2267 #ifdef CONFIG_FAIR_GROUP_SCHED
2268 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2269 extern unsigned long sched_group_shares(struct task_group *tg);
2270 #endif
2271 #ifdef CONFIG_RT_GROUP_SCHED
2272 extern int sched_group_set_rt_runtime(struct task_group *tg,
2273 				      long rt_runtime_us);
2274 extern long sched_group_rt_runtime(struct task_group *tg);
2275 extern int sched_group_set_rt_period(struct task_group *tg,
2276 				      long rt_period_us);
2277 extern long sched_group_rt_period(struct task_group *tg);
2278 #endif
2279 #endif
2280 
2281 #ifdef CONFIG_TASK_XACCT
2282 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2283 {
2284 	tsk->ioac.rchar += amt;
2285 }
2286 
2287 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2288 {
2289 	tsk->ioac.wchar += amt;
2290 }
2291 
2292 static inline void inc_syscr(struct task_struct *tsk)
2293 {
2294 	tsk->ioac.syscr++;
2295 }
2296 
2297 static inline void inc_syscw(struct task_struct *tsk)
2298 {
2299 	tsk->ioac.syscw++;
2300 }
2301 #else
2302 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2303 {
2304 }
2305 
2306 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2307 {
2308 }
2309 
2310 static inline void inc_syscr(struct task_struct *tsk)
2311 {
2312 }
2313 
2314 static inline void inc_syscw(struct task_struct *tsk)
2315 {
2316 }
2317 #endif
2318 
2319 #ifndef TASK_SIZE_OF
2320 #define TASK_SIZE_OF(tsk)	TASK_SIZE
2321 #endif
2322 
2323 #ifdef CONFIG_MM_OWNER
2324 extern void mm_update_next_owner(struct mm_struct *mm);
2325 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2326 #else
2327 static inline void mm_update_next_owner(struct mm_struct *mm)
2328 {
2329 }
2330 
2331 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2332 {
2333 }
2334 #endif /* CONFIG_MM_OWNER */
2335 
2336 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2337 
2338 #endif /* __KERNEL__ */
2339 
2340 #endif
2341