code style scripts/checkpatch.pl (linux-3.9-rc1) formatting
[linux-2.6.32.60-moxart.git] / include / linux / sched.h
blob71849bf135bc8fe45efd2a0e0219dacd46555f5d
1 #ifndef _LINUX_SCHED_H
2 #define _LINUX_SCHED_H
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 */
33 * Scheduling policies
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 /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
42 #define SCHED_RESET_ON_FORK 0x40000000
44 #ifdef __KERNEL__
46 struct sched_param {
47 int sched_priority;
50 #include <asm/param.h> /* for HZ */
52 #include <linux/capability.h>
53 #include <linux/threads.h>
54 #include <linux/kernel.h>
55 #include <linux/types.h>
56 #include <linux/timex.h>
57 #include <linux/jiffies.h>
58 #include <linux/rbtree.h>
59 #include <linux/thread_info.h>
60 #include <linux/cpumask.h>
61 #include <linux/errno.h>
62 #include <linux/nodemask.h>
63 #include <linux/mm_types.h>
65 #include <asm/system.h>
66 #include <asm/page.h>
67 #include <asm/ptrace.h>
68 #include <asm/cputime.h>
70 #include <linux/smp.h>
71 #include <linux/sem.h>
72 #include <linux/signal.h>
73 #include <linux/path.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/rculist.h>
83 #include <linux/rtmutex.h>
85 #include <linux/time.h>
86 #include <linux/param.h>
87 #include <linux/resource.h>
88 #include <linux/timer.h>
89 #include <linux/hrtimer.h>
90 #include <linux/task_io_accounting.h>
91 #include <linux/kobject.h>
92 #include <linux/latencytop.h>
93 #include <linux/cred.h>
95 #include <asm/processor.h>
97 struct exec_domain;
98 struct futex_pi_state;
99 struct robust_list_head;
100 struct bio;
101 struct fs_struct;
102 struct bts_context;
103 struct perf_event_context;
106 * List of flags we want to share for kernel threads,
107 * if only because they are not used by them anyway.
109 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
112 * These are the constant used to fake the fixed-point load-average
113 * counting. Some notes:
114 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
115 * a load-average precision of 10 bits integer + 11 bits fractional
116 * - if you want to count load-averages more often, you need more
117 * precision, or rounding will get you. With 2-second counting freq,
118 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 * 11 bit fractions.
121 extern unsigned long avenrun[]; /* Load averages */
122 extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
124 #define FSHIFT 11 /* nr of bits of precision */
125 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
126 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
127 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
128 #define EXP_5 2014 /* 1/exp(5sec/5min) */
129 #define EXP_15 2037 /* 1/exp(5sec/15min) */
131 #define CALC_LOAD(load,exp,n) \
132 load *= exp; \
133 load += n*(FIXED_1-exp); \
134 load >>= FSHIFT;
136 extern unsigned long total_forks;
137 extern int nr_threads;
138 DECLARE_PER_CPU(unsigned long, process_counts);
139 extern int nr_processes(void);
140 extern unsigned long nr_running(void);
141 extern unsigned long nr_uninterruptible(void);
142 extern unsigned long nr_iowait(void);
143 extern unsigned long nr_iowait_cpu(void);
144 extern unsigned long this_cpu_load(void);
147 extern void calc_global_load(void);
149 extern unsigned long get_parent_ip(unsigned long addr);
151 struct seq_file;
152 struct cfs_rq;
153 struct task_group;
154 #ifdef CONFIG_SCHED_DEBUG
155 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
156 extern void proc_sched_set_task(struct task_struct *p);
157 extern void
158 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
159 #else
160 static inline void
161 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
164 static inline void proc_sched_set_task(struct task_struct *p)
167 static inline void
168 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
171 #endif
173 extern unsigned long long time_sync_thresh;
176 * Task state bitmask. NOTE! These bits are also
177 * encoded in fs/proc/array.c: get_task_state().
179 * We have two separate sets of flags: task->state
180 * is about runnability, while task->exit_state are
181 * about the task exiting. Confusing, but this way
182 * modifying one set can't modify the other one by
183 * mistake.
185 #define TASK_RUNNING 0
186 #define TASK_INTERRUPTIBLE 1
187 #define TASK_UNINTERRUPTIBLE 2
188 #define __TASK_STOPPED 4
189 #define __TASK_TRACED 8
190 /* in tsk->exit_state */
191 #define EXIT_ZOMBIE 16
192 #define EXIT_DEAD 32
193 /* in tsk->state again */
194 #define TASK_DEAD 64
195 #define TASK_WAKEKILL 128
196 #define TASK_WAKING 256
198 /* Convenience macros for the sake of set_task_state */
199 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
200 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
201 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
203 /* Convenience macros for the sake of wake_up */
204 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
205 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
207 /* get_task_state() */
208 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
209 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
210 __TASK_TRACED)
212 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
213 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
214 #define task_is_stopped_or_traced(task) \
215 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
216 #define task_contributes_to_load(task) \
217 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
218 (task->flags & PF_FREEZING) == 0)
220 #define __set_task_state(tsk, state_value) \
221 do { (tsk)->state = (state_value); } while (0)
222 #define set_task_state(tsk, state_value) \
223 set_mb((tsk)->state, (state_value))
226 * set_current_state() includes a barrier so that the write of current->state
227 * is correctly serialised wrt the caller's subsequent test of whether to
228 * actually sleep:
230 * set_current_state(TASK_UNINTERRUPTIBLE);
231 * if (do_i_need_to_sleep())
232 * schedule();
234 * If the caller does not need such serialisation then use __set_current_state()
236 #define __set_current_state(state_value) \
237 do { current->state = (state_value); } while (0)
238 #define set_current_state(state_value) \
239 set_mb(current->state, (state_value))
241 /* Task command name length */
242 #define TASK_COMM_LEN 16
244 #include <linux/spinlock.h>
247 * This serializes "schedule()" and also protects
248 * the run-queue from deletions/modifications (but
249 * _adding_ to the beginning of the run-queue has
250 * a separate lock).
252 extern rwlock_t tasklist_lock;
253 extern spinlock_t mmlist_lock;
255 struct task_struct;
257 extern void sched_init(void);
258 extern void sched_init_smp(void);
259 extern asmlinkage void schedule_tail(struct task_struct *prev);
260 extern void init_idle(struct task_struct *idle, int cpu);
261 extern void init_idle_bootup_task(struct task_struct *idle);
263 extern int runqueue_is_locked(int cpu);
264 extern void task_rq_unlock_wait(struct task_struct *p);
266 extern cpumask_var_t nohz_cpu_mask;
267 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
268 extern int select_nohz_load_balancer(int cpu);
269 extern int get_nohz_load_balancer(void);
270 #else
271 static inline int select_nohz_load_balancer(int cpu)
273 return 0;
275 #endif
278 * Only dump TASK_* tasks. (0 for all tasks)
280 extern void show_state_filter(unsigned long state_filter);
282 static inline void show_state(void)
284 show_state_filter(0);
287 extern void show_regs(struct pt_regs *);
290 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
291 * task), SP is the stack pointer of the first frame that should be shown in the back
292 * trace (or NULL if the entire call-chain of the task should be shown).
294 extern void show_stack(struct task_struct *task, unsigned long *sp);
296 void io_schedule(void);
297 long io_schedule_timeout(long timeout);
299 extern void cpu_init (void);
300 extern void trap_init(void);
301 extern void update_process_times(int user);
302 extern void scheduler_tick(void);
304 extern void sched_show_task(struct task_struct *p);
306 #ifdef CONFIG_DETECT_SOFTLOCKUP
307 extern void softlockup_tick(void);
308 extern void touch_softlockup_watchdog(void);
309 extern void touch_all_softlockup_watchdogs(void);
310 extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
311 void __user *buffer,
312 size_t *lenp, loff_t *ppos);
313 extern unsigned int softlockup_panic;
314 extern int softlockup_thresh;
315 #else
316 static inline void softlockup_tick(void)
319 static inline void touch_softlockup_watchdog(void)
322 static inline void touch_all_softlockup_watchdogs(void)
325 #endif
327 #ifdef CONFIG_DETECT_HUNG_TASK
328 extern unsigned int sysctl_hung_task_panic;
329 extern unsigned long sysctl_hung_task_check_count;
330 extern unsigned long sysctl_hung_task_timeout_secs;
331 extern unsigned long sysctl_hung_task_warnings;
332 extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
333 void __user *buffer,
334 size_t *lenp, loff_t *ppos);
335 #endif
337 /* Attach to any functions which should be ignored in wchan output. */
338 #define __sched __attribute__((__section__(".sched.text")))
340 /* Linker adds these: start and end of __sched functions */
341 extern char __sched_text_start[], __sched_text_end[];
343 /* Is this address in the __sched functions? */
344 extern int in_sched_functions(unsigned long addr);
346 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
347 extern signed long schedule_timeout(signed long timeout);
348 extern signed long schedule_timeout_interruptible(signed long timeout);
349 extern signed long schedule_timeout_killable(signed long timeout);
350 extern signed long schedule_timeout_uninterruptible(signed long timeout);
351 asmlinkage void __schedule(void);
352 asmlinkage void schedule(void);
353 extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
355 struct nsproxy;
356 struct user_namespace;
359 * Default maximum number of active map areas, this limits the number of vmas
360 * per mm struct. Users can overwrite this number by sysctl but there is a
361 * problem.
363 * When a program's coredump is generated as ELF format, a section is created
364 * per a vma. In ELF, the number of sections is represented in unsigned short.
365 * This means the number of sections should be smaller than 65535 at coredump.
366 * Because the kernel adds some informative sections to a image of program at
367 * generating coredump, we need some margin. The number of extra sections is
368 * 1-3 now and depends on arch. We use "5" as safe margin, here.
370 #define MAPCOUNT_ELF_CORE_MARGIN (5)
371 #define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
373 extern int sysctl_max_map_count;
375 #include <linux/aio.h>
377 extern unsigned long
378 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
379 unsigned long, unsigned long);
380 extern unsigned long
381 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
382 unsigned long len, unsigned long pgoff,
383 unsigned long flags);
384 extern void arch_unmap_area(struct mm_struct *, unsigned long);
385 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
387 #if USE_SPLIT_PTLOCKS
389 * The mm counters are not protected by its page_table_lock,
390 * so must be incremented atomically.
392 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
393 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
394 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
395 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
396 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
398 #else /* !USE_SPLIT_PTLOCKS */
400 * The mm counters are protected by its page_table_lock,
401 * so can be incremented directly.
403 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
404 #define get_mm_counter(mm, member) ((mm)->_##member)
405 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
406 #define inc_mm_counter(mm, member) (mm)->_##member++
407 #define dec_mm_counter(mm, member) (mm)->_##member--
409 #endif /* !USE_SPLIT_PTLOCKS */
411 #define get_mm_rss(mm) \
412 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
413 #define update_hiwater_rss(mm) do { \
414 unsigned long _rss = get_mm_rss(mm); \
415 if ((mm)->hiwater_rss < _rss) \
416 (mm)->hiwater_rss = _rss; \
417 } while (0)
418 #define update_hiwater_vm(mm) do { \
419 if ((mm)->hiwater_vm < (mm)->total_vm) \
420 (mm)->hiwater_vm = (mm)->total_vm; \
421 } while (0)
423 static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
425 return max(mm->hiwater_rss, get_mm_rss(mm));
428 static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
429 struct mm_struct *mm)
431 unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
433 if (*maxrss < hiwater_rss)
434 *maxrss = hiwater_rss;
437 static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
439 return max(mm->hiwater_vm, mm->total_vm);
442 extern void set_dumpable(struct mm_struct *mm, int value);
443 extern int get_dumpable(struct mm_struct *mm);
445 /* mm flags */
446 /* dumpable bits */
447 #define MMF_DUMPABLE 0 /* core dump is permitted */
448 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
450 #define MMF_DUMPABLE_BITS 2
451 #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
453 /* coredump filter bits */
454 #define MMF_DUMP_ANON_PRIVATE 2
455 #define MMF_DUMP_ANON_SHARED 3
456 #define MMF_DUMP_MAPPED_PRIVATE 4
457 #define MMF_DUMP_MAPPED_SHARED 5
458 #define MMF_DUMP_ELF_HEADERS 6
459 #define MMF_DUMP_HUGETLB_PRIVATE 7
460 #define MMF_DUMP_HUGETLB_SHARED 8
462 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
463 #define MMF_DUMP_FILTER_BITS 7
464 #define MMF_DUMP_FILTER_MASK \
465 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
466 #define MMF_DUMP_FILTER_DEFAULT \
467 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
468 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
470 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
471 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
472 #else
473 # define MMF_DUMP_MASK_DEFAULT_ELF 0
474 #endif
475 /* leave room for more dump flags */
476 #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
478 #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
480 struct sighand_struct {
481 atomic_t count;
482 struct k_sigaction action[_NSIG];
483 spinlock_t siglock;
484 wait_queue_head_t signalfd_wqh;
487 struct pacct_struct {
488 int ac_flag;
489 long ac_exitcode;
490 unsigned long ac_mem;
491 cputime_t ac_utime, ac_stime;
492 unsigned long ac_minflt, ac_majflt;
495 struct cpu_itimer {
496 cputime_t expires;
497 cputime_t incr;
498 u32 error;
499 u32 incr_error;
503 * struct task_cputime - collected CPU time counts
504 * @utime: time spent in user mode, in &cputime_t units
505 * @stime: time spent in kernel mode, in &cputime_t units
506 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
508 * This structure groups together three kinds of CPU time that are
509 * tracked for threads and thread groups. Most things considering
510 * CPU time want to group these counts together and treat all three
511 * of them in parallel.
513 struct task_cputime {
514 cputime_t utime;
515 cputime_t stime;
516 unsigned long long sum_exec_runtime;
518 /* Alternate field names when used to cache expirations. */
519 #define prof_exp stime
520 #define virt_exp utime
521 #define sched_exp sum_exec_runtime
523 #define INIT_CPUTIME \
524 (struct task_cputime) { \
525 .utime = cputime_zero, \
526 .stime = cputime_zero, \
527 .sum_exec_runtime = 0, \
531 * Disable preemption until the scheduler is running.
532 * Reset by start_kernel()->sched_init()->init_idle().
534 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
535 * before the scheduler is active -- see should_resched().
537 #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
540 * struct thread_group_cputimer - thread group interval timer counts
541 * @cputime: thread group interval timers.
542 * @running: non-zero when there are timers running and
543 * @cputime receives updates.
544 * @lock: lock for fields in this struct.
546 * This structure contains the version of task_cputime, above, that is
547 * used for thread group CPU timer calculations.
549 struct thread_group_cputimer {
550 struct task_cputime cputime;
551 int running;
552 spinlock_t lock;
556 * NOTE! "signal_struct" does not have it's own
557 * locking, because a shared signal_struct always
558 * implies a shared sighand_struct, so locking
559 * sighand_struct is always a proper superset of
560 * the locking of signal_struct.
562 struct signal_struct {
563 atomic_t count;
564 atomic_t live;
566 wait_queue_head_t wait_chldexit; /* for wait4() */
568 /* current thread group signal load-balancing target: */
569 struct task_struct *curr_target;
571 /* shared signal handling: */
572 struct sigpending shared_pending;
574 /* thread group exit support */
575 int group_exit_code;
576 /* overloaded:
577 * - notify group_exit_task when ->count is equal to notify_count
578 * - everyone except group_exit_task is stopped during signal delivery
579 * of fatal signals, group_exit_task processes the signal.
581 int notify_count;
582 struct task_struct *group_exit_task;
584 /* thread group stop support, overloads group_exit_code too */
585 int group_stop_count;
586 unsigned int flags; /* see SIGNAL_* flags below */
588 /* POSIX.1b Interval Timers */
589 struct list_head posix_timers;
591 /* ITIMER_REAL timer for the process */
592 struct hrtimer real_timer;
593 struct pid *leader_pid;
594 ktime_t it_real_incr;
597 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
598 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
599 * values are defined to 0 and 1 respectively
601 struct cpu_itimer it[2];
604 * Thread group totals for process CPU timers.
605 * See thread_group_cputimer(), et al, for details.
607 struct thread_group_cputimer cputimer;
609 /* Earliest-expiration cache. */
610 struct task_cputime cputime_expires;
612 struct list_head cpu_timers[3];
614 struct pid *tty_old_pgrp;
616 /* boolean value for session group leader */
617 int leader;
619 struct tty_struct *tty; /* NULL if no tty */
622 * Cumulative resource counters for dead threads in the group,
623 * and for reaped dead child processes forked by this group.
624 * Live threads maintain their own counters and add to these
625 * in __exit_signal, except for the group leader.
627 cputime_t utime, stime, cutime, cstime;
628 cputime_t gtime;
629 cputime_t cgtime;
630 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
631 cputime_t prev_utime, prev_stime;
632 #endif
633 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
634 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
635 unsigned long inblock, oublock, cinblock, coublock;
636 unsigned long maxrss, cmaxrss;
637 struct task_io_accounting ioac;
640 * Cumulative ns of schedule CPU time fo dead threads in the
641 * group, not including a zombie group leader, (This only differs
642 * from jiffies_to_ns(utime + stime) if sched_clock uses something
643 * other than jiffies.)
645 unsigned long long sum_sched_runtime;
648 * We don't bother to synchronize most readers of this at all,
649 * because there is no reader checking a limit that actually needs
650 * to get both rlim_cur and rlim_max atomically, and either one
651 * alone is a single word that can safely be read normally.
652 * getrlimit/setrlimit use task_lock(current->group_leader) to
653 * protect this instead of the siglock, because they really
654 * have no need to disable irqs.
656 struct rlimit rlim[RLIM_NLIMITS];
658 #ifdef CONFIG_BSD_PROCESS_ACCT
659 struct pacct_struct pacct; /* per-process accounting information */
660 #endif
661 #ifdef CONFIG_TASKSTATS
662 struct taskstats *stats;
663 #endif
664 #ifdef CONFIG_AUDIT
665 unsigned audit_tty;
666 struct tty_audit_buf *tty_audit_buf;
667 #endif
669 int oom_adj; /* OOM kill score adjustment (bit shift) */
672 /* Context switch must be unlocked if interrupts are to be enabled */
673 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
674 # define __ARCH_WANT_UNLOCKED_CTXSW
675 #endif
678 * Bits in flags field of signal_struct.
680 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
681 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
682 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
683 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
685 * Pending notifications to parent.
687 #define SIGNAL_CLD_STOPPED 0x00000010
688 #define SIGNAL_CLD_CONTINUED 0x00000020
689 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
691 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
693 /* If true, all threads except ->group_exit_task have pending SIGKILL */
694 static inline int signal_group_exit(const struct signal_struct *sig)
696 return (sig->flags & SIGNAL_GROUP_EXIT) ||
697 (sig->group_exit_task != NULL);
701 * Some day this will be a full-fledged user tracking system..
703 struct user_struct {
704 atomic_t __count; /* reference count */
705 atomic_t processes; /* How many processes does this user have? */
706 atomic_t files; /* How many open files does this user have? */
707 atomic_t sigpending; /* How many pending signals does this user have? */
708 #ifdef CONFIG_INOTIFY_USER
709 atomic_t inotify_watches; /* How many inotify watches does this user have? */
710 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
711 #endif
712 #ifdef CONFIG_EPOLL
713 atomic_t epoll_watches; /* The number of file descriptors currently watched */
714 #endif
715 #ifdef CONFIG_POSIX_MQUEUE
716 /* protected by mq_lock */
717 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
718 #endif
719 unsigned long locked_shm; /* How many pages of mlocked shm ? */
721 #ifdef CONFIG_KEYS
722 struct key *uid_keyring; /* UID specific keyring */
723 struct key *session_keyring; /* UID's default session keyring */
724 #endif
726 /* Hash table maintenance information */
727 struct hlist_node uidhash_node;
728 uid_t uid;
729 struct user_namespace *user_ns;
731 #ifdef CONFIG_PERF_EVENTS
732 atomic_long_t locked_vm;
733 #endif
736 extern int uids_sysfs_init(void);
738 extern struct user_struct *find_user(uid_t);
740 extern struct user_struct root_user;
741 #define INIT_USER (&root_user)
744 struct backing_dev_info;
745 struct reclaim_state;
747 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
748 struct sched_info {
749 /* cumulative counters */
750 unsigned long pcount; /* # of times run on this cpu */
751 unsigned long long run_delay; /* time spent waiting on a runqueue */
753 /* timestamps */
754 unsigned long long last_arrival,/* when we last ran on a cpu */
755 last_queued; /* when we were last queued to run */
756 #ifdef CONFIG_SCHEDSTATS
757 /* BKL stats */
758 unsigned int bkl_count;
759 #endif
761 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
763 #ifdef CONFIG_TASK_DELAY_ACCT
764 struct task_delay_info {
765 spinlock_t lock;
766 unsigned int flags; /* Private per-task flags */
768 /* For each stat XXX, add following, aligned appropriately
770 * struct timespec XXX_start, XXX_end;
771 * u64 XXX_delay;
772 * u32 XXX_count;
774 * Atomicity of updates to XXX_delay, XXX_count protected by
775 * single lock above (split into XXX_lock if contention is an issue).
779 * XXX_count is incremented on every XXX operation, the delay
780 * associated with the operation is added to XXX_delay.
781 * XXX_delay contains the accumulated delay time in nanoseconds.
783 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
784 u64 blkio_delay; /* wait for sync block io completion */
785 u64 swapin_delay; /* wait for swapin block io completion */
786 u32 blkio_count; /* total count of the number of sync block */
787 /* io operations performed */
788 u32 swapin_count; /* total count of the number of swapin block */
789 /* io operations performed */
791 struct timespec freepages_start, freepages_end;
792 u64 freepages_delay; /* wait for memory reclaim */
793 u32 freepages_count; /* total count of memory reclaim */
795 #endif /* CONFIG_TASK_DELAY_ACCT */
797 static inline int sched_info_on(void)
799 #ifdef CONFIG_SCHEDSTATS
800 return 1;
801 #elif defined(CONFIG_TASK_DELAY_ACCT)
802 extern int delayacct_on;
803 return delayacct_on;
804 #else
805 return 0;
806 #endif
809 enum cpu_idle_type {
810 CPU_IDLE,
811 CPU_NOT_IDLE,
812 CPU_NEWLY_IDLE,
813 CPU_MAX_IDLE_TYPES
817 * sched-domains (multiprocessor balancing) declarations:
821 * Increase resolution of nice-level calculations:
823 #define SCHED_LOAD_SHIFT 10
824 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
826 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
828 #ifdef CONFIG_SMP
829 #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
830 #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
831 #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
832 #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
833 #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
834 #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
835 #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
836 #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
837 #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
838 #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
839 #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
841 #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
843 enum powersavings_balance_level {
844 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
845 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
846 * first for long running threads
848 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
849 * cpu package for power savings
851 MAX_POWERSAVINGS_BALANCE_LEVELS
854 extern int sched_mc_power_savings, sched_smt_power_savings;
856 static inline int sd_balance_for_mc_power(void)
858 if (sched_smt_power_savings)
859 return SD_POWERSAVINGS_BALANCE;
861 if (!sched_mc_power_savings)
862 return SD_PREFER_SIBLING;
864 return 0;
867 static inline int sd_balance_for_package_power(void)
869 if (sched_mc_power_savings | sched_smt_power_savings)
870 return SD_POWERSAVINGS_BALANCE;
872 return SD_PREFER_SIBLING;
876 * Optimise SD flags for power savings:
877 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
878 * Keep default SD flags if sched_{smt,mc}_power_saving=0
881 static inline int sd_power_saving_flags(void)
883 if (sched_mc_power_savings | sched_smt_power_savings)
884 return SD_BALANCE_NEWIDLE;
886 return 0;
889 struct sched_group {
890 struct sched_group *next; /* Must be a circular list */
893 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
894 * single CPU.
896 unsigned int cpu_power;
897 unsigned int group_weight;
900 * The CPUs this group covers.
902 * NOTE: this field is variable length. (Allocated dynamically
903 * by attaching extra space to the end of the structure,
904 * depending on how many CPUs the kernel has booted up with)
906 * It is also be embedded into static data structures at build
907 * time. (See 'struct static_sched_group' in kernel/sched.c)
909 unsigned long cpumask[0];
912 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
914 return to_cpumask(sg->cpumask);
917 enum sched_domain_level {
918 SD_LV_NONE = 0,
919 SD_LV_SIBLING,
920 SD_LV_MC,
921 SD_LV_CPU,
922 SD_LV_NODE,
923 SD_LV_ALLNODES,
924 SD_LV_MAX
927 struct sched_domain_attr {
928 int relax_domain_level;
931 #define SD_ATTR_INIT (struct sched_domain_attr) { \
932 .relax_domain_level = -1, \
935 struct sched_domain {
936 /* These fields must be setup */
937 struct sched_domain *parent; /* top domain must be null terminated */
938 struct sched_domain *child; /* bottom domain must be null terminated */
939 struct sched_group *groups; /* the balancing groups of the domain */
940 unsigned long min_interval; /* Minimum balance interval ms */
941 unsigned long max_interval; /* Maximum balance interval ms */
942 unsigned int busy_factor; /* less balancing by factor if busy */
943 unsigned int imbalance_pct; /* No balance until over watermark */
944 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
945 unsigned int busy_idx;
946 unsigned int idle_idx;
947 unsigned int newidle_idx;
948 unsigned int wake_idx;
949 unsigned int forkexec_idx;
950 unsigned int smt_gain;
951 int flags; /* See SD_* */
952 enum sched_domain_level level;
954 /* Runtime fields. */
955 unsigned long last_balance; /* init to jiffies. units in jiffies */
956 unsigned int balance_interval; /* initialise to 1. units in ms. */
957 unsigned int nr_balance_failed; /* initialise to 0 */
959 u64 last_update;
961 #ifdef CONFIG_SCHEDSTATS
962 /* load_balance() stats */
963 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
964 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
965 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
966 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
967 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
968 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
969 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
970 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
972 /* Active load balancing */
973 unsigned int alb_count;
974 unsigned int alb_failed;
975 unsigned int alb_pushed;
977 /* SD_BALANCE_EXEC stats */
978 unsigned int sbe_count;
979 unsigned int sbe_balanced;
980 unsigned int sbe_pushed;
982 /* SD_BALANCE_FORK stats */
983 unsigned int sbf_count;
984 unsigned int sbf_balanced;
985 unsigned int sbf_pushed;
987 /* try_to_wake_up() stats */
988 unsigned int ttwu_wake_remote;
989 unsigned int ttwu_move_affine;
990 unsigned int ttwu_move_balance;
991 #endif
992 #ifdef CONFIG_SCHED_DEBUG
993 char *name;
994 #endif
996 unsigned int span_weight;
998 * Span of all CPUs in this domain.
1000 * NOTE: this field is variable length. (Allocated dynamically
1001 * by attaching extra space to the end of the structure,
1002 * depending on how many CPUs the kernel has booted up with)
1004 * It is also be embedded into static data structures at build
1005 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1007 unsigned long span[0];
1010 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1012 return to_cpumask(sd->span);
1015 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1016 struct sched_domain_attr *dattr_new);
1018 /* Test a flag in parent sched domain */
1019 static inline int test_sd_parent(struct sched_domain *sd, int flag)
1021 if (sd->parent && (sd->parent->flags & flag))
1022 return 1;
1024 return 0;
1027 unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1028 unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1030 #else /* CONFIG_SMP */
1032 struct sched_domain_attr;
1034 static inline void
1035 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1036 struct sched_domain_attr *dattr_new)
1039 #endif /* !CONFIG_SMP */
1042 struct io_context; /* See blkdev.h */
1045 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
1046 extern void prefetch_stack(struct task_struct *t);
1047 #else
1048 static inline void prefetch_stack(struct task_struct *t) { }
1049 #endif
1051 struct audit_context; /* See audit.c */
1052 struct mempolicy;
1053 struct pipe_inode_info;
1054 struct uts_namespace;
1056 struct rq;
1057 struct sched_domain;
1060 * wake flags
1062 #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
1063 #define WF_FORK 0x02 /* child wakeup after fork */
1065 struct sched_class {
1066 const struct sched_class *next;
1068 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup,
1069 bool head);
1070 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
1071 void (*yield_task) (struct rq *rq);
1073 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
1075 struct task_struct * (*pick_next_task) (struct rq *rq);
1076 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
1078 #ifdef CONFIG_SMP
1079 int (*select_task_rq)(struct rq *rq, struct task_struct *p,
1080 int sd_flag, int flags);
1082 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
1083 struct rq *busiest, unsigned long max_load_move,
1084 struct sched_domain *sd, enum cpu_idle_type idle,
1085 int *all_pinned, int *this_best_prio);
1087 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1088 struct rq *busiest, struct sched_domain *sd,
1089 enum cpu_idle_type idle);
1090 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1091 void (*post_schedule) (struct rq *this_rq);
1092 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1093 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
1095 void (*set_cpus_allowed)(struct task_struct *p,
1096 const struct cpumask *newmask);
1098 void (*rq_online)(struct rq *rq);
1099 void (*rq_offline)(struct rq *rq);
1100 #endif
1102 void (*set_curr_task) (struct rq *rq);
1103 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1104 void (*task_fork) (struct task_struct *p);
1106 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1107 int running);
1108 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1109 int running);
1110 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1111 int oldprio, int running);
1113 unsigned int (*get_rr_interval) (struct rq *rq,
1114 struct task_struct *task);
1116 #ifdef CONFIG_FAIR_GROUP_SCHED
1117 void (*task_move_group) (struct task_struct *p, int on_rq);
1118 #endif
1121 struct load_weight {
1122 unsigned long weight, inv_weight;
1126 * CFS stats for a schedulable entity (task, task-group etc)
1128 * Current field usage histogram:
1130 * 4 se->block_start
1131 * 4 se->run_node
1132 * 4 se->sleep_start
1133 * 6 se->load.weight
1135 struct sched_entity {
1136 struct load_weight load; /* for load-balancing */
1137 struct rb_node run_node;
1138 struct list_head group_node;
1139 unsigned int on_rq;
1141 u64 exec_start;
1142 u64 sum_exec_runtime;
1143 u64 vruntime;
1144 u64 prev_sum_exec_runtime;
1146 u64 last_wakeup;
1147 u64 avg_overlap;
1149 u64 nr_migrations;
1151 u64 start_runtime;
1152 u64 avg_wakeup;
1154 u64 avg_running;
1156 #ifdef CONFIG_SCHEDSTATS
1157 u64 wait_start;
1158 u64 wait_max;
1159 u64 wait_count;
1160 u64 wait_sum;
1161 u64 iowait_count;
1162 u64 iowait_sum;
1164 u64 sleep_start;
1165 u64 sleep_max;
1166 s64 sum_sleep_runtime;
1168 u64 block_start;
1169 u64 block_max;
1170 u64 exec_max;
1171 u64 slice_max;
1173 u64 nr_migrations_cold;
1174 u64 nr_failed_migrations_affine;
1175 u64 nr_failed_migrations_running;
1176 u64 nr_failed_migrations_hot;
1177 u64 nr_forced_migrations;
1179 u64 nr_wakeups;
1180 u64 nr_wakeups_sync;
1181 u64 nr_wakeups_migrate;
1182 u64 nr_wakeups_local;
1183 u64 nr_wakeups_remote;
1184 u64 nr_wakeups_affine;
1185 u64 nr_wakeups_affine_attempts;
1186 u64 nr_wakeups_passive;
1187 u64 nr_wakeups_idle;
1188 #endif
1190 #ifdef CONFIG_FAIR_GROUP_SCHED
1191 struct sched_entity *parent;
1192 /* rq on which this entity is (to be) queued: */
1193 struct cfs_rq *cfs_rq;
1194 /* rq "owned" by this entity/group: */
1195 struct cfs_rq *my_q;
1196 #endif
1199 struct sched_rt_entity {
1200 struct list_head run_list;
1201 unsigned long timeout;
1202 unsigned int time_slice;
1203 int nr_cpus_allowed;
1205 struct sched_rt_entity *back;
1206 #ifdef CONFIG_RT_GROUP_SCHED
1207 struct sched_rt_entity *parent;
1208 /* rq on which this entity is (to be) queued: */
1209 struct rt_rq *rt_rq;
1210 /* rq "owned" by this entity/group: */
1211 struct rt_rq *my_q;
1212 #endif
1215 struct rcu_node;
1217 struct task_struct {
1218 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
1219 void *stack;
1220 atomic_t usage;
1221 unsigned int flags; /* per process flags, defined below */
1222 unsigned int ptrace;
1224 int lock_depth; /* BKL lock depth */
1226 #ifdef CONFIG_SMP
1227 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1228 int oncpu;
1229 #endif
1230 #endif
1232 int prio, static_prio, normal_prio;
1233 unsigned int rt_priority;
1234 const struct sched_class *sched_class;
1235 struct sched_entity se;
1236 struct sched_rt_entity rt;
1238 #ifdef CONFIG_PREEMPT_NOTIFIERS
1239 /* list of struct preempt_notifier: */
1240 struct hlist_head preempt_notifiers;
1241 #endif
1244 * fpu_counter contains the number of consecutive context switches
1245 * that the FPU is used. If this is over a threshold, the lazy fpu
1246 * saving becomes unlazy to save the trap. This is an unsigned char
1247 * so that after 256 times the counter wraps and the behavior turns
1248 * lazy again; this to deal with bursty apps that only use FPU for
1249 * a short time
1251 unsigned char fpu_counter;
1252 #ifdef CONFIG_BLK_DEV_IO_TRACE
1253 unsigned int btrace_seq;
1254 #endif
1256 unsigned int policy;
1257 cpumask_t cpus_allowed;
1259 #ifdef CONFIG_TREE_PREEMPT_RCU
1260 int rcu_read_lock_nesting;
1261 char rcu_read_unlock_special;
1262 struct rcu_node *rcu_blocked_node;
1263 struct list_head rcu_node_entry;
1264 #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1266 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1267 struct sched_info sched_info;
1268 #endif
1270 struct list_head tasks;
1271 struct plist_node pushable_tasks;
1273 struct mm_struct *mm, *active_mm;
1275 /* task state */
1276 int exit_state;
1277 int exit_code, exit_signal;
1278 int pdeath_signal; /* The signal sent when the parent dies */
1279 /* ??? */
1280 unsigned int personality;
1281 unsigned did_exec:1;
1282 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1283 * execve */
1284 unsigned in_iowait:1;
1287 /* Revert to default priority/policy when forking */
1288 unsigned sched_reset_on_fork:1;
1290 pid_t pid;
1291 pid_t tgid;
1293 #ifdef CONFIG_CC_STACKPROTECTOR
1294 /* Canary value for the -fstack-protector gcc feature */
1295 unsigned long stack_canary;
1296 #endif
1299 * pointers to (original) parent process, youngest child, younger sibling,
1300 * older sibling, respectively. (p->father can be replaced with
1301 * p->real_parent->pid)
1303 struct task_struct *real_parent; /* real parent process */
1304 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1306 * children/sibling forms the list of my natural children
1308 struct list_head children; /* list of my children */
1309 struct list_head sibling; /* linkage in my parent's children list */
1310 struct task_struct *group_leader; /* threadgroup leader */
1313 * ptraced is the list of tasks this task is using ptrace on.
1314 * This includes both natural children and PTRACE_ATTACH targets.
1315 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1317 struct list_head ptraced;
1318 struct list_head ptrace_entry;
1321 * This is the tracer handle for the ptrace BTS extension.
1322 * This field actually belongs to the ptracer task.
1324 struct bts_context *bts;
1326 /* PID/PID hash table linkage. */
1327 struct pid_link pids[PIDTYPE_MAX];
1328 struct list_head thread_group;
1330 struct completion *vfork_done; /* for vfork() */
1331 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1332 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1334 cputime_t utime, stime, utimescaled, stimescaled;
1335 cputime_t gtime;
1336 cputime_t prev_utime, prev_stime;
1337 unsigned long nvcsw, nivcsw; /* context switch counts */
1338 struct timespec start_time; /* monotonic time */
1339 struct timespec real_start_time; /* boot based time */
1340 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1341 unsigned long min_flt, maj_flt;
1343 struct task_cputime cputime_expires;
1344 struct list_head cpu_timers[3];
1346 /* process credentials */
1347 const struct cred *real_cred; /* objective and real subjective task
1348 * credentials (COW) */
1349 const struct cred *cred; /* effective (overridable) subjective task
1350 * credentials (COW) */
1351 struct mutex cred_guard_mutex; /* guard against foreign influences on
1352 * credential calculations
1353 * (notably. ptrace) */
1354 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
1356 char comm[TASK_COMM_LEN]; /* executable name excluding path
1357 - access with [gs]et_task_comm (which lock
1358 it with task_lock())
1359 - initialized normally by setup_new_exec */
1360 /* file system info */
1361 int link_count, total_link_count;
1362 #ifdef CONFIG_SYSVIPC
1363 /* ipc stuff */
1364 struct sysv_sem sysvsem;
1365 #endif
1366 #ifdef CONFIG_DETECT_HUNG_TASK
1367 /* hung task detection */
1368 unsigned long last_switch_count;
1369 #endif
1370 /* CPU-specific state of this task */
1371 struct thread_struct thread;
1372 /* filesystem information */
1373 struct fs_struct *fs;
1374 /* open file information */
1375 struct files_struct *files;
1376 /* namespaces */
1377 struct nsproxy *nsproxy;
1378 /* signal handlers */
1379 struct signal_struct *signal;
1380 struct sighand_struct *sighand;
1382 sigset_t blocked, real_blocked;
1383 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1384 struct sigpending pending;
1386 unsigned long sas_ss_sp;
1387 size_t sas_ss_size;
1388 int (*notifier)(void *priv);
1389 void *notifier_data;
1390 sigset_t *notifier_mask;
1391 struct audit_context *audit_context;
1392 #ifdef CONFIG_AUDITSYSCALL
1393 uid_t loginuid;
1394 unsigned int sessionid;
1395 #endif
1396 seccomp_t seccomp;
1398 /* Thread group tracking */
1399 u32 parent_exec_id;
1400 u32 self_exec_id;
1401 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1402 * mempolicy */
1403 spinlock_t alloc_lock;
1405 #ifdef CONFIG_GENERIC_HARDIRQS
1406 /* IRQ handler threads */
1407 struct irqaction *irqaction;
1408 #endif
1410 /* Protection of the PI data structures: */
1411 spinlock_t pi_lock;
1413 #ifdef CONFIG_RT_MUTEXES
1414 /* PI waiters blocked on a rt_mutex held by this task */
1415 struct plist_head pi_waiters;
1416 /* Deadlock detection and priority inheritance handling */
1417 struct rt_mutex_waiter *pi_blocked_on;
1418 #endif
1420 #ifdef CONFIG_DEBUG_MUTEXES
1421 /* mutex deadlock detection */
1422 struct mutex_waiter *blocked_on;
1423 #endif
1424 #ifdef CONFIG_TRACE_IRQFLAGS
1425 unsigned int irq_events;
1426 int hardirqs_enabled;
1427 unsigned long hardirq_enable_ip;
1428 unsigned int hardirq_enable_event;
1429 unsigned long hardirq_disable_ip;
1430 unsigned int hardirq_disable_event;
1431 int softirqs_enabled;
1432 unsigned long softirq_disable_ip;
1433 unsigned int softirq_disable_event;
1434 unsigned long softirq_enable_ip;
1435 unsigned int softirq_enable_event;
1436 int hardirq_context;
1437 int softirq_context;
1438 #endif
1439 #ifdef CONFIG_LOCKDEP
1440 # define MAX_LOCK_DEPTH 48UL
1441 u64 curr_chain_key;
1442 int lockdep_depth;
1443 unsigned int lockdep_recursion;
1444 struct held_lock held_locks[MAX_LOCK_DEPTH];
1445 gfp_t lockdep_reclaim_gfp;
1446 #endif
1448 /* journalling filesystem info */
1449 void *journal_info;
1451 /* stacked block device info */
1452 struct bio *bio_list, **bio_tail;
1454 /* VM state */
1455 struct reclaim_state *reclaim_state;
1457 struct backing_dev_info *backing_dev_info;
1459 struct io_context *io_context;
1461 unsigned long ptrace_message;
1462 siginfo_t *last_siginfo; /* For ptrace use. */
1463 struct task_io_accounting ioac;
1464 #if defined(CONFIG_TASK_XACCT)
1465 u64 acct_rss_mem1; /* accumulated rss usage */
1466 u64 acct_vm_mem1; /* accumulated virtual memory usage */
1467 cputime_t acct_timexpd; /* stime + utime since last update */
1468 #endif
1469 #ifdef CONFIG_CPUSETS
1470 nodemask_t mems_allowed; /* Protected by alloc_lock */
1471 int cpuset_mem_spread_rotor;
1472 #endif
1473 #ifdef CONFIG_CGROUPS
1474 /* Control Group info protected by css_set_lock */
1475 struct css_set *cgroups;
1476 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1477 struct list_head cg_list;
1478 #endif
1479 #ifdef CONFIG_FUTEX
1480 struct robust_list_head __user *robust_list;
1481 #ifdef CONFIG_COMPAT
1482 struct compat_robust_list_head __user *compat_robust_list;
1483 #endif
1484 struct list_head pi_state_list;
1485 struct futex_pi_state *pi_state_cache;
1486 #endif
1487 #ifdef CONFIG_PERF_EVENTS
1488 struct perf_event_context *perf_event_ctxp;
1489 struct mutex perf_event_mutex;
1490 struct list_head perf_event_list;
1491 #endif
1492 #ifdef CONFIG_NUMA
1493 struct mempolicy *mempolicy; /* Protected by alloc_lock */
1494 short il_next;
1495 #endif
1496 atomic_t fs_excl; /* holding fs exclusive resources */
1497 struct rcu_head rcu;
1500 * cache last used pipe for splice
1502 struct pipe_inode_info *splice_pipe;
1503 #ifdef CONFIG_TASK_DELAY_ACCT
1504 struct task_delay_info *delays;
1505 #endif
1506 #ifdef CONFIG_FAULT_INJECTION
1507 int make_it_fail;
1508 #endif
1509 struct prop_local_single dirties;
1510 #ifdef CONFIG_LATENCYTOP
1511 int latency_record_count;
1512 struct latency_record latency_record[LT_SAVECOUNT];
1513 #endif
1515 * time slack values; these are used to round up poll() and
1516 * select() etc timeout values. These are in nanoseconds.
1518 unsigned long timer_slack_ns;
1519 unsigned long default_timer_slack_ns;
1521 struct list_head *scm_work_list;
1522 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1523 /* Index of current stored adress in ret_stack */
1524 int curr_ret_stack;
1525 /* Stack of return addresses for return function tracing */
1526 struct ftrace_ret_stack *ret_stack;
1527 /* time stamp for last schedule */
1528 unsigned long long ftrace_timestamp;
1530 * Number of functions that haven't been traced
1531 * because of depth overrun.
1533 atomic_t trace_overrun;
1534 /* Pause for the tracing */
1535 atomic_t tracing_graph_pause;
1536 #endif
1537 #ifdef CONFIG_TRACING
1538 /* state flags for use by tracers */
1539 unsigned long trace;
1540 /* bitmask of trace recursion */
1541 unsigned long trace_recursion;
1542 #endif /* CONFIG_TRACING */
1545 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1546 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1549 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1550 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1551 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1552 * values are inverted: lower p->prio value means higher priority.
1554 * The MAX_USER_RT_PRIO value allows the actual maximum
1555 * RT priority to be separate from the value exported to
1556 * user-space. This allows kernel threads to set their
1557 * priority to a value higher than any user task. Note:
1558 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1561 #define MAX_USER_RT_PRIO 100
1562 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1564 #define MAX_PRIO (MAX_RT_PRIO + 40)
1565 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1567 static inline int rt_prio(int prio)
1569 if (unlikely(prio < MAX_RT_PRIO))
1570 return 1;
1571 return 0;
1574 static inline int rt_task(struct task_struct *p)
1576 return rt_prio(p->prio);
1579 static inline struct pid *task_pid(struct task_struct *task)
1581 return task->pids[PIDTYPE_PID].pid;
1584 static inline struct pid *task_tgid(struct task_struct *task)
1586 return task->group_leader->pids[PIDTYPE_PID].pid;
1590 * Without tasklist or rcu lock it is not safe to dereference
1591 * the result of task_pgrp/task_session even if task == current,
1592 * we can race with another thread doing sys_setsid/sys_setpgid.
1594 static inline struct pid *task_pgrp(struct task_struct *task)
1596 return task->group_leader->pids[PIDTYPE_PGID].pid;
1599 static inline struct pid *task_session(struct task_struct *task)
1601 return task->group_leader->pids[PIDTYPE_SID].pid;
1604 struct pid_namespace;
1607 * the helpers to get the task's different pids as they are seen
1608 * from various namespaces
1610 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1611 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1612 * current.
1613 * task_xid_nr_ns() : id seen from the ns specified;
1615 * set_task_vxid() : assigns a virtual id to a task;
1617 * see also pid_nr() etc in include/linux/pid.h
1619 pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1620 struct pid_namespace *ns);
1622 static inline pid_t task_pid_nr(struct task_struct *tsk)
1624 return tsk->pid;
1627 static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1628 struct pid_namespace *ns)
1630 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1633 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1635 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
1639 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1641 return tsk->tgid;
1644 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1646 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1648 return pid_vnr(task_tgid(tsk));
1652 static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1653 struct pid_namespace *ns)
1655 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
1658 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1660 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
1664 static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1665 struct pid_namespace *ns)
1667 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
1670 static inline pid_t task_session_vnr(struct task_struct *tsk)
1672 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
1675 /* obsolete, do not use */
1676 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1678 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1682 * pid_alive - check that a task structure is not stale
1683 * @p: Task structure to be checked.
1685 * Test if a process is not yet dead (at most zombie state)
1686 * If pid_alive fails, then pointers within the task structure
1687 * can be stale and must not be dereferenced.
1689 static inline int pid_alive(struct task_struct *p)
1691 return p->pids[PIDTYPE_PID].pid != NULL;
1695 * is_global_init - check if a task structure is init
1696 * @tsk: Task structure to be checked.
1698 * Check if a task structure is the first user space task the kernel created.
1700 static inline int is_global_init(struct task_struct *tsk)
1702 return tsk->pid == 1;
1706 * is_container_init:
1707 * check whether in the task is init in its own pid namespace.
1709 extern int is_container_init(struct task_struct *tsk);
1711 extern struct pid *cad_pid;
1713 extern void free_task(struct task_struct *tsk);
1714 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1716 extern void __put_task_struct(struct task_struct *t);
1718 static inline void put_task_struct(struct task_struct *t)
1720 if (atomic_dec_and_test(&t->usage))
1721 __put_task_struct(t);
1724 extern cputime_t task_utime(struct task_struct *p);
1725 extern cputime_t task_stime(struct task_struct *p);
1726 extern cputime_t task_gtime(struct task_struct *p);
1727 extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
1730 * Per process flags
1732 #define PF_KSOFTIRQD 0x00000001 /* I am ksoftirqd */
1733 #define PF_STARTING 0x00000002 /* being created */
1734 #define PF_EXITING 0x00000004 /* getting shut down */
1735 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1736 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1737 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1738 #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1739 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1740 #define PF_DUMPCORE 0x00000200 /* dumped core */
1741 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1742 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1743 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1744 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1745 #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1746 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1747 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1748 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1749 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1750 #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1751 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1752 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1753 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1754 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1755 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1756 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1757 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1758 #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
1759 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1760 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1761 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1762 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1765 * Only the _current_ task can read/write to tsk->flags, but other
1766 * tasks can access tsk->flags in readonly mode for example
1767 * with tsk_used_math (like during threaded core dumping).
1768 * There is however an exception to this rule during ptrace
1769 * or during fork: the ptracer task is allowed to write to the
1770 * child->flags of its traced child (same goes for fork, the parent
1771 * can write to the child->flags), because we're guaranteed the
1772 * child is not running and in turn not changing child->flags
1773 * at the same time the parent does it.
1775 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1776 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1777 #define clear_used_math() clear_stopped_child_used_math(current)
1778 #define set_used_math() set_stopped_child_used_math(current)
1779 #define conditional_stopped_child_used_math(condition, child) \
1780 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1781 #define conditional_used_math(condition) \
1782 conditional_stopped_child_used_math(condition, current)
1783 #define copy_to_stopped_child_used_math(child) \
1784 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1785 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1786 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1787 #define used_math() tsk_used_math(current)
1789 #ifdef CONFIG_TREE_PREEMPT_RCU
1791 #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1792 #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1794 static inline void rcu_copy_process(struct task_struct *p)
1796 p->rcu_read_lock_nesting = 0;
1797 p->rcu_read_unlock_special = 0;
1798 p->rcu_blocked_node = NULL;
1799 INIT_LIST_HEAD(&p->rcu_node_entry);
1802 #else
1804 static inline void rcu_copy_process(struct task_struct *p)
1808 #endif
1810 #ifdef CONFIG_SMP
1811 extern int set_cpus_allowed_ptr(struct task_struct *p,
1812 const struct cpumask *new_mask);
1813 #else
1814 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1815 const struct cpumask *new_mask)
1817 if (!cpumask_test_cpu(0, new_mask))
1818 return -EINVAL;
1819 return 0;
1821 #endif
1823 #ifndef CONFIG_CPUMASK_OFFSTACK
1824 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1826 return set_cpus_allowed_ptr(p, &new_mask);
1828 #endif
1831 * Architectures can set this to 1 if they have specified
1832 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1833 * but then during bootup it turns out that sched_clock()
1834 * is reliable after all:
1836 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1837 extern int sched_clock_stable;
1838 #endif
1840 extern unsigned long long sched_clock(void);
1842 extern void sched_clock_init(void);
1843 extern u64 sched_clock_cpu(int cpu);
1845 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1846 static inline void sched_clock_tick(void)
1850 static inline void sched_clock_idle_sleep_event(void)
1854 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1857 #else
1858 extern void sched_clock_tick(void);
1859 extern void sched_clock_idle_sleep_event(void);
1860 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1861 #endif
1864 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1865 * clock constructed from sched_clock():
1867 extern unsigned long long cpu_clock(int cpu);
1869 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
1871 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
1872 * The reason for this explicit opt-in is not to have perf penalty with
1873 * slow sched_clocks.
1875 extern void enable_sched_clock_irqtime(void);
1876 extern void disable_sched_clock_irqtime(void);
1877 #else
1878 static inline void enable_sched_clock_irqtime(void) {}
1879 static inline void disable_sched_clock_irqtime(void) {}
1880 #endif
1882 extern unsigned long long
1883 task_sched_runtime(struct task_struct *task);
1884 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1886 /* sched_exec is called by processes performing an exec */
1887 #ifdef CONFIG_SMP
1888 extern void sched_exec(void);
1889 #else
1890 #define sched_exec() {}
1891 #endif
1893 extern void sched_clock_idle_sleep_event(void);
1894 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1896 #ifdef CONFIG_HOTPLUG_CPU
1897 extern void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p);
1898 extern void idle_task_exit(void);
1899 #else
1900 static inline void idle_task_exit(void) {}
1901 #endif
1903 extern void sched_idle_next(void);
1905 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1906 extern void wake_up_idle_cpu(int cpu);
1907 #else
1908 static inline void wake_up_idle_cpu(int cpu) { }
1909 #endif
1911 extern unsigned int sysctl_sched_latency;
1912 extern unsigned int sysctl_sched_min_granularity;
1913 extern unsigned int sysctl_sched_wakeup_granularity;
1914 extern unsigned int sysctl_sched_shares_ratelimit;
1915 extern unsigned int sysctl_sched_shares_thresh;
1916 extern unsigned int sysctl_sched_child_runs_first;
1917 #ifdef CONFIG_SCHED_DEBUG
1918 extern unsigned int sysctl_sched_features;
1919 extern unsigned int sysctl_sched_migration_cost;
1920 extern unsigned int sysctl_sched_nr_migrate;
1921 extern unsigned int sysctl_sched_time_avg;
1922 extern unsigned int sysctl_timer_migration;
1924 int sched_nr_latency_handler(struct ctl_table *table, int write,
1925 void __user *buffer, size_t *length,
1926 loff_t *ppos);
1927 #endif
1928 #ifdef CONFIG_SCHED_DEBUG
1929 static inline unsigned int get_sysctl_timer_migration(void)
1931 return sysctl_timer_migration;
1933 #else
1934 static inline unsigned int get_sysctl_timer_migration(void)
1936 return 1;
1938 #endif
1939 extern unsigned int sysctl_sched_rt_period;
1940 extern int sysctl_sched_rt_runtime;
1942 int sched_rt_handler(struct ctl_table *table, int write,
1943 void __user *buffer, size_t *lenp,
1944 loff_t *ppos);
1946 extern unsigned int sysctl_sched_compat_yield;
1948 #ifdef CONFIG_RT_MUTEXES
1949 extern int rt_mutex_getprio(struct task_struct *p);
1950 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1951 extern void rt_mutex_adjust_pi(struct task_struct *p);
1952 #else
1953 static inline int rt_mutex_getprio(struct task_struct *p)
1955 return p->normal_prio;
1957 # define rt_mutex_adjust_pi(p) do { } while (0)
1958 #endif
1960 extern void set_user_nice(struct task_struct *p, long nice);
1961 extern int task_prio(const struct task_struct *p);
1962 extern int task_nice(const struct task_struct *p);
1963 extern int can_nice(const struct task_struct *p, const int nice);
1964 extern int task_curr(const struct task_struct *p);
1965 extern int idle_cpu(int cpu);
1966 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1967 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1968 struct sched_param *);
1969 extern struct task_struct *idle_task(int cpu);
1970 extern struct task_struct *curr_task(int cpu);
1971 extern void set_curr_task(int cpu, struct task_struct *p);
1973 void yield(void);
1976 * The default (Linux) execution domain.
1978 extern struct exec_domain default_exec_domain;
1980 union thread_union {
1981 struct thread_info thread_info;
1982 unsigned long stack[THREAD_SIZE/sizeof(long)];
1985 #ifndef __HAVE_ARCH_KSTACK_END
1986 static inline int kstack_end(void *addr)
1988 /* Reliable end of stack detection:
1989 * Some APM bios versions misalign the stack
1991 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1993 #endif
1995 extern union thread_union init_thread_union;
1996 extern struct task_struct init_task;
1998 extern struct mm_struct init_mm;
2000 extern struct pid_namespace init_pid_ns;
2003 * find a task by one of its numerical ids
2005 * find_task_by_pid_ns():
2006 * finds a task by its pid in the specified namespace
2007 * find_task_by_vpid():
2008 * finds a task by its virtual pid
2010 * see also find_vpid() etc in include/linux/pid.h
2013 extern struct task_struct *find_task_by_vpid(pid_t nr);
2014 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2015 struct pid_namespace *ns);
2017 extern void __set_special_pids(struct pid *pid);
2019 /* per-UID process charging. */
2020 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
2021 static inline struct user_struct *get_uid(struct user_struct *u)
2023 atomic_inc(&u->__count);
2024 return u;
2026 extern void free_uid(struct user_struct *);
2027 extern void release_uids(struct user_namespace *ns);
2029 #include <asm/current.h>
2031 extern void do_timer(unsigned long ticks);
2033 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2034 extern int wake_up_process(struct task_struct *tsk);
2035 extern void wake_up_new_task(struct task_struct *tsk,
2036 unsigned long clone_flags);
2037 #ifdef CONFIG_SMP
2038 extern void kick_process(struct task_struct *tsk);
2039 #else
2040 static inline void kick_process(struct task_struct *tsk) { }
2041 #endif
2042 extern void sched_fork(struct task_struct *p, int clone_flags);
2043 extern void sched_dead(struct task_struct *p);
2045 extern void proc_caches_init(void);
2046 extern void flush_signals(struct task_struct *);
2047 extern void __flush_signals(struct task_struct *);
2048 extern void ignore_signals(struct task_struct *);
2049 extern void flush_signal_handlers(struct task_struct *, int force_default);
2050 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2052 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2054 unsigned long flags;
2055 int ret;
2057 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2058 ret = dequeue_signal(tsk, mask, info);
2059 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2061 return ret;
2064 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2065 sigset_t *mask);
2066 extern void unblock_all_signals(void);
2067 extern void release_task(struct task_struct * p);
2068 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
2069 extern int force_sigsegv(int, struct task_struct *);
2070 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
2071 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
2072 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2073 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
2074 extern int kill_pgrp(struct pid *pid, int sig, int priv);
2075 extern int kill_pid(struct pid *pid, int sig, int priv);
2076 extern int kill_proc_info(int, struct siginfo *, pid_t);
2077 extern int do_notify_parent(struct task_struct *, int);
2078 extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
2079 extern void force_sig(int, struct task_struct *);
2080 extern void force_sig_specific(int, struct task_struct *);
2081 extern int send_sig(int, struct task_struct *, int);
2082 extern void zap_other_threads(struct task_struct *p);
2083 extern struct sigqueue *sigqueue_alloc(void);
2084 extern void sigqueue_free(struct sigqueue *);
2085 extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
2086 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
2087 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2089 static inline int kill_cad_pid(int sig, int priv)
2091 return kill_pid(cad_pid, sig, priv);
2094 /* These can be the second arg to send_sig_info/send_group_sig_info. */
2095 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
2096 #define SEND_SIG_PRIV ((struct siginfo *) 1)
2097 #define SEND_SIG_FORCED ((struct siginfo *) 2)
2099 static inline int is_si_special(const struct siginfo *info)
2101 return info <= SEND_SIG_FORCED;
2105 * True if we are on the alternate signal stack.
2107 static inline int on_sig_stack(unsigned long sp)
2109 #ifdef CONFIG_STACK_GROWSUP
2110 return sp >= current->sas_ss_sp &&
2111 sp - current->sas_ss_sp < current->sas_ss_size;
2112 #else
2113 return sp > current->sas_ss_sp &&
2114 sp - current->sas_ss_sp <= current->sas_ss_size;
2115 #endif
2118 static inline int sas_ss_flags(unsigned long sp)
2120 return (current->sas_ss_size == 0 ? SS_DISABLE
2121 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2125 * Routines for handling mm_structs
2127 extern struct mm_struct * mm_alloc(void);
2129 /* mmdrop drops the mm and the page tables */
2130 extern void __mmdrop(struct mm_struct *);
2131 static inline void mmdrop(struct mm_struct * mm)
2133 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
2134 __mmdrop(mm);
2137 /* mmput gets rid of the mappings and all user-space */
2138 extern void mmput(struct mm_struct *);
2139 /* Grab a reference to a task's mm, if it is not already going away */
2140 extern struct mm_struct *get_task_mm(struct task_struct *task);
2141 /* Remove the current tasks stale references to the old mm_struct */
2142 extern void mm_release(struct task_struct *, struct mm_struct *);
2143 /* Allocate a new mm structure and copy contents from tsk->mm */
2144 extern struct mm_struct *dup_mm(struct task_struct *tsk);
2146 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2147 struct task_struct *, struct pt_regs *);
2148 extern void flush_thread(void);
2149 extern void exit_thread(void);
2151 extern void exit_files(struct task_struct *);
2152 extern void __cleanup_signal(struct signal_struct *);
2153 extern void __cleanup_sighand(struct sighand_struct *);
2155 extern void exit_itimers(struct signal_struct *);
2156 extern void flush_itimer_signals(void);
2158 extern NORET_TYPE void do_group_exit(int);
2160 extern void daemonize(const char *, ...);
2161 extern int allow_signal(int);
2162 extern int disallow_signal(int);
2164 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2165 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
2166 struct task_struct *fork_idle(int);
2168 extern void set_task_comm(struct task_struct *tsk, char *from);
2169 extern char *get_task_comm(char *to, struct task_struct *tsk);
2171 #ifdef CONFIG_SMP
2172 extern void wait_task_context_switch(struct task_struct *p);
2173 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
2174 #else
2175 static inline void wait_task_context_switch(struct task_struct *p) {}
2176 static inline unsigned long wait_task_inactive(struct task_struct *p,
2177 long match_state)
2179 return 1;
2181 #endif
2183 #define next_task(p) \
2184 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2186 #define for_each_process(p) \
2187 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2189 extern bool current_is_single_threaded(void);
2192 * Careful: do_each_thread/while_each_thread is a double loop so
2193 * 'break' will not work as expected - use goto instead.
2195 #define do_each_thread(g, t) \
2196 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2198 #define while_each_thread(g, t) \
2199 while ((t = next_thread(t)) != g)
2201 /* de_thread depends on thread_group_leader not being a pid based check */
2202 #define thread_group_leader(p) (p == p->group_leader)
2204 /* Do to the insanities of de_thread it is possible for a process
2205 * to have the pid of the thread group leader without actually being
2206 * the thread group leader. For iteration through the pids in proc
2207 * all we care about is that we have a task with the appropriate
2208 * pid, we don't actually care if we have the right task.
2210 static inline int has_group_leader_pid(struct task_struct *p)
2212 return p->pid == p->tgid;
2215 static inline
2216 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2218 return p1->tgid == p2->tgid;
2221 static inline struct task_struct *next_thread(const struct task_struct *p)
2223 return list_entry_rcu(p->thread_group.next,
2224 struct task_struct, thread_group);
2227 static inline int thread_group_empty(struct task_struct *p)
2229 return list_empty(&p->thread_group);
2232 #define delay_group_leader(p) \
2233 (thread_group_leader(p) && !thread_group_empty(p))
2235 static inline int task_detached(struct task_struct *p)
2237 return p->exit_signal == -1;
2241 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2242 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2243 * pins the final release of task.io_context. Also protects ->cpuset and
2244 * ->cgroup.subsys[].
2246 * Nests both inside and outside of read_lock(&tasklist_lock).
2247 * It must not be nested with write_lock_irq(&tasklist_lock),
2248 * neither inside nor outside.
2250 static inline void task_lock(struct task_struct *p)
2252 spin_lock(&p->alloc_lock);
2255 static inline void task_unlock(struct task_struct *p)
2257 spin_unlock(&p->alloc_lock);
2260 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2261 unsigned long *flags);
2263 static inline void unlock_task_sighand(struct task_struct *tsk,
2264 unsigned long *flags)
2266 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2269 #ifndef __HAVE_THREAD_FUNCTIONS
2271 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2272 #define task_stack_page(task) ((task)->stack)
2274 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2276 *task_thread_info(p) = *task_thread_info(org);
2277 task_thread_info(p)->task = p;
2280 static inline unsigned long *end_of_stack(struct task_struct *p)
2282 return (unsigned long *)(task_thread_info(p) + 1);
2285 #endif
2287 static inline int object_is_on_stack(void *obj)
2289 void *stack = task_stack_page(current);
2291 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2294 extern void thread_info_cache_init(void);
2296 #ifdef CONFIG_DEBUG_STACK_USAGE
2297 static inline unsigned long stack_not_used(struct task_struct *p)
2299 unsigned long *n = end_of_stack(p);
2301 do { /* Skip over canary */
2302 n++;
2303 } while (!*n);
2305 return (unsigned long)n - (unsigned long)end_of_stack(p);
2307 #endif
2309 /* set thread flags in other task's structures
2310 * - see asm/thread_info.h for TIF_xxxx flags available
2312 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2314 set_ti_thread_flag(task_thread_info(tsk), flag);
2317 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2319 clear_ti_thread_flag(task_thread_info(tsk), flag);
2322 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2324 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2327 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2329 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2332 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2334 return test_ti_thread_flag(task_thread_info(tsk), flag);
2337 static inline void set_tsk_need_resched(struct task_struct *tsk)
2339 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2342 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2344 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2347 static inline int test_tsk_need_resched(struct task_struct *tsk)
2349 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2352 static inline int restart_syscall(void)
2354 set_tsk_thread_flag(current, TIF_SIGPENDING);
2355 return -ERESTARTNOINTR;
2358 static inline int signal_pending(struct task_struct *p)
2360 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2363 static inline int __fatal_signal_pending(struct task_struct *p)
2365 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2368 static inline int fatal_signal_pending(struct task_struct *p)
2370 return signal_pending(p) && __fatal_signal_pending(p);
2373 static inline int signal_pending_state(long state, struct task_struct *p)
2375 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2376 return 0;
2377 if (!signal_pending(p))
2378 return 0;
2380 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2383 static inline int need_resched(void)
2385 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2389 * cond_resched() and cond_resched_lock(): latency reduction via
2390 * explicit rescheduling in places that are safe. The return
2391 * value indicates whether a reschedule was done in fact.
2392 * cond_resched_lock() will drop the spinlock before scheduling,
2393 * cond_resched_softirq() will enable bhs before scheduling.
2395 extern int _cond_resched(void);
2397 #define cond_resched() ({ \
2398 __might_sleep(__FILE__, __LINE__, 0); \
2399 _cond_resched(); \
2402 extern int __cond_resched_lock(spinlock_t *lock);
2404 #ifdef CONFIG_PREEMPT
2405 #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
2406 #else
2407 #define PREEMPT_LOCK_OFFSET 0
2408 #endif
2410 #define cond_resched_lock(lock) ({ \
2411 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
2412 __cond_resched_lock(lock); \
2415 extern int __cond_resched_softirq(void);
2417 #define cond_resched_softirq() ({ \
2418 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2419 __cond_resched_softirq(); \
2423 * Does a critical section need to be broken due to another
2424 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2425 * but a general need for low latency)
2427 static inline int spin_needbreak(spinlock_t *lock)
2429 #ifdef CONFIG_PREEMPT
2430 return spin_is_contended(lock);
2431 #else
2432 return 0;
2433 #endif
2437 * Thread group CPU time accounting.
2439 void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2440 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2442 static inline void thread_group_cputime_init(struct signal_struct *sig)
2444 sig->cputimer.cputime = INIT_CPUTIME;
2445 spin_lock_init(&sig->cputimer.lock);
2446 sig->cputimer.running = 0;
2449 static inline void thread_group_cputime_free(struct signal_struct *sig)
2454 * Reevaluate whether the task has signals pending delivery.
2455 * Wake the task if so.
2456 * This is required every time the blocked sigset_t changes.
2457 * callers must hold sighand->siglock.
2459 extern void recalc_sigpending_and_wake(struct task_struct *t);
2460 extern void recalc_sigpending(void);
2462 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2465 * Wrappers for p->thread_info->cpu access. No-op on UP.
2467 #ifdef CONFIG_SMP
2469 static inline unsigned int task_cpu(const struct task_struct *p)
2471 return task_thread_info(p)->cpu;
2474 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2476 #else
2478 static inline unsigned int task_cpu(const struct task_struct *p)
2480 return 0;
2483 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2487 #endif /* CONFIG_SMP */
2489 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2491 #ifdef CONFIG_TRACING
2492 extern void
2493 __trace_special(void *__tr, void *__data,
2494 unsigned long arg1, unsigned long arg2, unsigned long arg3);
2495 #else
2496 static inline void
2497 __trace_special(void *__tr, void *__data,
2498 unsigned long arg1, unsigned long arg2, unsigned long arg3)
2501 #endif
2503 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2504 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2506 extern void normalize_rt_tasks(void);
2508 #ifdef CONFIG_CGROUP_SCHED
2510 extern struct task_group init_task_group;
2512 extern struct task_group *sched_create_group(struct task_group *parent);
2513 extern void sched_destroy_group(struct task_group *tg);
2514 extern void sched_move_task(struct task_struct *tsk);
2515 #ifdef CONFIG_FAIR_GROUP_SCHED
2516 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2517 extern unsigned long sched_group_shares(struct task_group *tg);
2518 #endif
2519 #ifdef CONFIG_RT_GROUP_SCHED
2520 extern int sched_group_set_rt_runtime(struct task_group *tg,
2521 long rt_runtime_us);
2522 extern long sched_group_rt_runtime(struct task_group *tg);
2523 extern int sched_group_set_rt_period(struct task_group *tg,
2524 long rt_period_us);
2525 extern long sched_group_rt_period(struct task_group *tg);
2526 extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
2527 #endif
2528 #endif
2530 extern int task_can_switch_user(struct user_struct *up,
2531 struct task_struct *tsk);
2533 #ifdef CONFIG_TASK_XACCT
2534 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2536 tsk->ioac.rchar += amt;
2539 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2541 tsk->ioac.wchar += amt;
2544 static inline void inc_syscr(struct task_struct *tsk)
2546 tsk->ioac.syscr++;
2549 static inline void inc_syscw(struct task_struct *tsk)
2551 tsk->ioac.syscw++;
2553 #else
2554 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2558 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2562 static inline void inc_syscr(struct task_struct *tsk)
2566 static inline void inc_syscw(struct task_struct *tsk)
2569 #endif
2571 #ifndef TASK_SIZE_OF
2572 #define TASK_SIZE_OF(tsk) TASK_SIZE
2573 #endif
2576 * Call the function if the target task is executing on a CPU right now:
2578 extern void task_oncpu_function_call(struct task_struct *p,
2579 void (*func) (void *info), void *info);
2582 #ifdef CONFIG_MM_OWNER
2583 extern void mm_update_next_owner(struct mm_struct *mm);
2584 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2585 #else
2586 static inline void mm_update_next_owner(struct mm_struct *mm)
2590 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2593 #endif /* CONFIG_MM_OWNER */
2595 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2597 static inline unsigned long task_rlimit(const struct task_struct *tsk,
2598 unsigned int limit)
2600 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2603 static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2604 unsigned int limit)
2606 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2609 static inline unsigned long rlimit(unsigned int limit)
2611 return task_rlimit(current, limit);
2614 static inline unsigned long rlimit_max(unsigned int limit)
2616 return task_rlimit_max(current, limit);
2619 #endif /* __KERNEL__ */
2621 #endif