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 */
35 #define SCHED_NORMAL 0
39 /* SCHED_ISO: reserved but not implemented yet */
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
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>
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>
98 struct futex_pi_state
;
99 struct robust_list_head
;
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
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) \
133 load += n*(FIXED_1-exp); \
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
);
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
);
158 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
);
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
)
168 print_cfs_rq(struct seq_file
*m
, int cpu
, struct cfs_rq
*cfs_rq
)
174 * Task state bitmask. NOTE! These bits are also
175 * encoded in fs/proc/array.c: get_task_state().
177 * We have two separate sets of flags: task->state
178 * is about runnability, while task->exit_state are
179 * about the task exiting. Confusing, but this way
180 * modifying one set can't modify the other one by
183 #define TASK_RUNNING 0
184 #define TASK_INTERRUPTIBLE 1
185 #define TASK_UNINTERRUPTIBLE 2
186 #define __TASK_STOPPED 4
187 #define __TASK_TRACED 8
188 /* in tsk->exit_state */
189 #define EXIT_ZOMBIE 16
191 /* in tsk->state again */
193 #define TASK_WAKEKILL 128
194 #define TASK_WAKING 256
196 /* Convenience macros for the sake of set_task_state */
197 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
198 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
199 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
201 /* Convenience macros for the sake of wake_up */
202 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
203 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
205 /* get_task_state() */
206 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
207 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
210 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
211 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
212 #define task_is_stopped_or_traced(task) \
213 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
214 #define task_contributes_to_load(task) \
215 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
216 (task->flags & PF_FREEZING) == 0)
218 #define __set_task_state(tsk, state_value) \
219 do { (tsk)->state = (state_value); } while (0)
220 #define set_task_state(tsk, state_value) \
221 set_mb((tsk)->state, (state_value))
224 * set_current_state() includes a barrier so that the write of current->state
225 * is correctly serialised wrt the caller's subsequent test of whether to
228 * set_current_state(TASK_UNINTERRUPTIBLE);
229 * if (do_i_need_to_sleep())
232 * If the caller does not need such serialisation then use __set_current_state()
234 #define __set_current_state(state_value) \
235 do { current->state = (state_value); } while (0)
236 #define set_current_state(state_value) \
237 set_mb(current->state, (state_value))
239 /* Task command name length */
240 #define TASK_COMM_LEN 16
242 #include <linux/spinlock.h>
245 * This serializes "schedule()" and also protects
246 * the run-queue from deletions/modifications (but
247 * _adding_ to the beginning of the run-queue has
250 extern rwlock_t tasklist_lock
;
251 extern spinlock_t mmlist_lock
;
255 extern void sched_init(void);
256 extern void sched_init_smp(void);
257 extern asmlinkage
void schedule_tail(struct task_struct
*prev
);
258 extern void init_idle(struct task_struct
*idle
, int cpu
);
259 extern void init_idle_bootup_task(struct task_struct
*idle
);
261 extern int runqueue_is_locked(int cpu
);
262 extern void task_rq_unlock_wait(struct task_struct
*p
);
264 extern cpumask_var_t nohz_cpu_mask
;
265 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
266 extern int select_nohz_load_balancer(int cpu
);
267 extern int get_nohz_load_balancer(void);
269 static inline int select_nohz_load_balancer(int cpu
)
276 * Only dump TASK_* tasks. (0 for all tasks)
278 extern void show_state_filter(unsigned long state_filter
);
280 static inline void show_state(void)
282 show_state_filter(0);
285 extern void show_regs(struct pt_regs
*);
288 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
289 * task), SP is the stack pointer of the first frame that should be shown in the back
290 * trace (or NULL if the entire call-chain of the task should be shown).
292 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
294 void io_schedule(void);
295 long io_schedule_timeout(long timeout
);
297 extern void cpu_init (void);
298 extern void trap_init(void);
299 extern void update_process_times(int user
);
300 extern void scheduler_tick(void);
302 extern void sched_show_task(struct task_struct
*p
);
304 #ifdef CONFIG_DETECT_SOFTLOCKUP
305 extern void softlockup_tick(void);
306 extern void touch_softlockup_watchdog(void);
307 extern void touch_all_softlockup_watchdogs(void);
308 extern int proc_dosoftlockup_thresh(struct ctl_table
*table
, int write
,
310 size_t *lenp
, loff_t
*ppos
);
311 extern unsigned int softlockup_panic
;
312 extern int softlockup_thresh
;
314 static inline void softlockup_tick(void)
317 static inline void touch_softlockup_watchdog(void)
320 static inline void touch_all_softlockup_watchdogs(void)
325 #ifdef CONFIG_DETECT_HUNG_TASK
326 extern unsigned int sysctl_hung_task_panic
;
327 extern unsigned long sysctl_hung_task_check_count
;
328 extern unsigned long sysctl_hung_task_timeout_secs
;
329 extern unsigned long sysctl_hung_task_warnings
;
330 extern int proc_dohung_task_timeout_secs(struct ctl_table
*table
, int write
,
332 size_t *lenp
, loff_t
*ppos
);
335 /* Attach to any functions which should be ignored in wchan output. */
336 #define __sched __attribute__((__section__(".sched.text")))
338 /* Linker adds these: start and end of __sched functions */
339 extern char __sched_text_start
[], __sched_text_end
[];
341 /* Is this address in the __sched functions? */
342 extern int in_sched_functions(unsigned long addr
);
344 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
345 extern signed long schedule_timeout(signed long timeout
);
346 extern signed long schedule_timeout_interruptible(signed long timeout
);
347 extern signed long schedule_timeout_killable(signed long timeout
);
348 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
349 asmlinkage
void schedule(void);
350 extern int mutex_spin_on_owner(struct mutex
*lock
, struct thread_info
*owner
);
353 struct user_namespace
;
356 * Default maximum number of active map areas, this limits the number of vmas
357 * per mm struct. Users can overwrite this number by sysctl but there is a
360 * When a program's coredump is generated as ELF format, a section is created
361 * per a vma. In ELF, the number of sections is represented in unsigned short.
362 * This means the number of sections should be smaller than 65535 at coredump.
363 * Because the kernel adds some informative sections to a image of program at
364 * generating coredump, we need some margin. The number of extra sections is
365 * 1-3 now and depends on arch. We use "5" as safe margin, here.
367 #define MAPCOUNT_ELF_CORE_MARGIN (5)
368 #define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
370 extern int sysctl_max_map_count
;
372 #include <linux/aio.h>
375 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
376 unsigned long, unsigned long);
378 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
379 unsigned long len
, unsigned long pgoff
,
380 unsigned long flags
);
381 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
382 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
384 #if USE_SPLIT_PTLOCKS
386 * The mm counters are not protected by its page_table_lock,
387 * so must be incremented atomically.
389 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
390 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
391 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
392 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
393 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
395 #else /* !USE_SPLIT_PTLOCKS */
397 * The mm counters are protected by its page_table_lock,
398 * so can be incremented directly.
400 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
401 #define get_mm_counter(mm, member) ((mm)->_##member)
402 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
403 #define inc_mm_counter(mm, member) (mm)->_##member++
404 #define dec_mm_counter(mm, member) (mm)->_##member--
406 #endif /* !USE_SPLIT_PTLOCKS */
408 #define get_mm_rss(mm) \
409 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
410 #define update_hiwater_rss(mm) do { \
411 unsigned long _rss = get_mm_rss(mm); \
412 if ((mm)->hiwater_rss < _rss) \
413 (mm)->hiwater_rss = _rss; \
415 #define update_hiwater_vm(mm) do { \
416 if ((mm)->hiwater_vm < (mm)->total_vm) \
417 (mm)->hiwater_vm = (mm)->total_vm; \
420 static inline unsigned long get_mm_hiwater_rss(struct mm_struct
*mm
)
422 return max(mm
->hiwater_rss
, get_mm_rss(mm
));
425 static inline void setmax_mm_hiwater_rss(unsigned long *maxrss
,
426 struct mm_struct
*mm
)
428 unsigned long hiwater_rss
= get_mm_hiwater_rss(mm
);
430 if (*maxrss
< hiwater_rss
)
431 *maxrss
= hiwater_rss
;
434 static inline unsigned long get_mm_hiwater_vm(struct mm_struct
*mm
)
436 return max(mm
->hiwater_vm
, mm
->total_vm
);
439 extern void set_dumpable(struct mm_struct
*mm
, int value
);
440 extern int get_dumpable(struct mm_struct
*mm
);
444 #define MMF_DUMPABLE 0 /* core dump is permitted */
445 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
447 #define MMF_DUMPABLE_BITS 2
448 #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
450 /* coredump filter bits */
451 #define MMF_DUMP_ANON_PRIVATE 2
452 #define MMF_DUMP_ANON_SHARED 3
453 #define MMF_DUMP_MAPPED_PRIVATE 4
454 #define MMF_DUMP_MAPPED_SHARED 5
455 #define MMF_DUMP_ELF_HEADERS 6
456 #define MMF_DUMP_HUGETLB_PRIVATE 7
457 #define MMF_DUMP_HUGETLB_SHARED 8
459 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
460 #define MMF_DUMP_FILTER_BITS 7
461 #define MMF_DUMP_FILTER_MASK \
462 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
463 #define MMF_DUMP_FILTER_DEFAULT \
464 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
465 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
467 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
468 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
470 # define MMF_DUMP_MASK_DEFAULT_ELF 0
472 /* leave room for more dump flags */
473 #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
475 #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
477 struct sighand_struct
{
479 struct k_sigaction action
[_NSIG
];
481 wait_queue_head_t signalfd_wqh
;
484 struct pacct_struct
{
487 unsigned long ac_mem
;
488 cputime_t ac_utime
, ac_stime
;
489 unsigned long ac_minflt
, ac_majflt
;
500 * struct task_cputime - collected CPU time counts
501 * @utime: time spent in user mode, in &cputime_t units
502 * @stime: time spent in kernel mode, in &cputime_t units
503 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
505 * This structure groups together three kinds of CPU time that are
506 * tracked for threads and thread groups. Most things considering
507 * CPU time want to group these counts together and treat all three
508 * of them in parallel.
510 struct task_cputime
{
513 unsigned long long sum_exec_runtime
;
515 /* Alternate field names when used to cache expirations. */
516 #define prof_exp stime
517 #define virt_exp utime
518 #define sched_exp sum_exec_runtime
520 #define INIT_CPUTIME \
521 (struct task_cputime) { \
522 .utime = cputime_zero, \
523 .stime = cputime_zero, \
524 .sum_exec_runtime = 0, \
528 * Disable preemption until the scheduler is running.
529 * Reset by start_kernel()->sched_init()->init_idle().
531 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
532 * before the scheduler is active -- see should_resched().
534 #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
537 * struct thread_group_cputimer - thread group interval timer counts
538 * @cputime: thread group interval timers.
539 * @running: non-zero when there are timers running and
540 * @cputime receives updates.
541 * @lock: lock for fields in this struct.
543 * This structure contains the version of task_cputime, above, that is
544 * used for thread group CPU timer calculations.
546 struct thread_group_cputimer
{
547 struct task_cputime cputime
;
553 * NOTE! "signal_struct" does not have it's own
554 * locking, because a shared signal_struct always
555 * implies a shared sighand_struct, so locking
556 * sighand_struct is always a proper superset of
557 * the locking of signal_struct.
559 struct signal_struct
{
563 wait_queue_head_t wait_chldexit
; /* for wait4() */
565 /* current thread group signal load-balancing target: */
566 struct task_struct
*curr_target
;
568 /* shared signal handling: */
569 struct sigpending shared_pending
;
571 /* thread group exit support */
574 * - notify group_exit_task when ->count is equal to notify_count
575 * - everyone except group_exit_task is stopped during signal delivery
576 * of fatal signals, group_exit_task processes the signal.
579 struct task_struct
*group_exit_task
;
581 /* thread group stop support, overloads group_exit_code too */
582 int group_stop_count
;
583 unsigned int flags
; /* see SIGNAL_* flags below */
585 /* POSIX.1b Interval Timers */
586 struct list_head posix_timers
;
588 /* ITIMER_REAL timer for the process */
589 struct hrtimer real_timer
;
590 struct pid
*leader_pid
;
591 ktime_t it_real_incr
;
594 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
595 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
596 * values are defined to 0 and 1 respectively
598 struct cpu_itimer it
[2];
601 * Thread group totals for process CPU timers.
602 * See thread_group_cputimer(), et al, for details.
604 struct thread_group_cputimer cputimer
;
606 /* Earliest-expiration cache. */
607 struct task_cputime cputime_expires
;
609 struct list_head cpu_timers
[3];
611 struct pid
*tty_old_pgrp
;
613 /* boolean value for session group leader */
616 struct tty_struct
*tty
; /* NULL if no tty */
619 * Cumulative resource counters for dead threads in the group,
620 * and for reaped dead child processes forked by this group.
621 * Live threads maintain their own counters and add to these
622 * in __exit_signal, except for the group leader.
624 cputime_t utime
, stime
, cutime
, cstime
;
627 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
628 cputime_t prev_utime
, prev_stime
;
630 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
631 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
632 unsigned long inblock
, oublock
, cinblock
, coublock
;
633 unsigned long maxrss
, cmaxrss
;
634 struct task_io_accounting ioac
;
637 * Cumulative ns of schedule CPU time fo dead threads in the
638 * group, not including a zombie group leader, (This only differs
639 * from jiffies_to_ns(utime + stime) if sched_clock uses something
640 * other than jiffies.)
642 unsigned long long sum_sched_runtime
;
645 * We don't bother to synchronize most readers of this at all,
646 * because there is no reader checking a limit that actually needs
647 * to get both rlim_cur and rlim_max atomically, and either one
648 * alone is a single word that can safely be read normally.
649 * getrlimit/setrlimit use task_lock(current->group_leader) to
650 * protect this instead of the siglock, because they really
651 * have no need to disable irqs.
653 struct rlimit rlim
[RLIM_NLIMITS
];
655 #ifdef CONFIG_BSD_PROCESS_ACCT
656 struct pacct_struct pacct
; /* per-process accounting information */
658 #ifdef CONFIG_TASKSTATS
659 struct taskstats
*stats
;
663 struct tty_audit_buf
*tty_audit_buf
;
666 int oom_adj
; /* OOM kill score adjustment (bit shift) */
669 /* Context switch must be unlocked if interrupts are to be enabled */
670 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
671 # define __ARCH_WANT_UNLOCKED_CTXSW
675 * Bits in flags field of signal_struct.
677 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
678 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
679 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
680 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
682 * Pending notifications to parent.
684 #define SIGNAL_CLD_STOPPED 0x00000010
685 #define SIGNAL_CLD_CONTINUED 0x00000020
686 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
688 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
690 /* If true, all threads except ->group_exit_task have pending SIGKILL */
691 static inline int signal_group_exit(const struct signal_struct
*sig
)
693 return (sig
->flags
& SIGNAL_GROUP_EXIT
) ||
694 (sig
->group_exit_task
!= NULL
);
698 * Some day this will be a full-fledged user tracking system..
701 atomic_t __count
; /* reference count */
702 atomic_t processes
; /* How many processes does this user have? */
703 atomic_t files
; /* How many open files does this user have? */
704 atomic_t sigpending
; /* How many pending signals does this user have? */
705 #ifdef CONFIG_INOTIFY_USER
706 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
707 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
710 atomic_t epoll_watches
; /* The number of file descriptors currently watched */
712 #ifdef CONFIG_POSIX_MQUEUE
713 /* protected by mq_lock */
714 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
716 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
719 struct key
*uid_keyring
; /* UID specific keyring */
720 struct key
*session_keyring
; /* UID's default session keyring */
723 /* Hash table maintenance information */
724 struct hlist_node uidhash_node
;
726 struct user_namespace
*user_ns
;
728 #ifdef CONFIG_USER_SCHED
729 struct task_group
*tg
;
732 struct delayed_work work
;
736 #ifdef CONFIG_PERF_EVENTS
737 atomic_long_t locked_vm
;
741 extern int uids_sysfs_init(void);
743 extern struct user_struct
*find_user(uid_t
);
745 extern struct user_struct root_user
;
746 #define INIT_USER (&root_user)
749 struct backing_dev_info
;
750 struct reclaim_state
;
752 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
754 /* cumulative counters */
755 unsigned long pcount
; /* # of times run on this cpu */
756 unsigned long long run_delay
; /* time spent waiting on a runqueue */
759 unsigned long long last_arrival
,/* when we last ran on a cpu */
760 last_queued
; /* when we were last queued to run */
761 #ifdef CONFIG_SCHEDSTATS
763 unsigned int bkl_count
;
766 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
768 #ifdef CONFIG_TASK_DELAY_ACCT
769 struct task_delay_info
{
771 unsigned int flags
; /* Private per-task flags */
773 /* For each stat XXX, add following, aligned appropriately
775 * struct timespec XXX_start, XXX_end;
779 * Atomicity of updates to XXX_delay, XXX_count protected by
780 * single lock above (split into XXX_lock if contention is an issue).
784 * XXX_count is incremented on every XXX operation, the delay
785 * associated with the operation is added to XXX_delay.
786 * XXX_delay contains the accumulated delay time in nanoseconds.
788 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
789 u64 blkio_delay
; /* wait for sync block io completion */
790 u64 swapin_delay
; /* wait for swapin block io completion */
791 u32 blkio_count
; /* total count of the number of sync block */
792 /* io operations performed */
793 u32 swapin_count
; /* total count of the number of swapin block */
794 /* io operations performed */
796 struct timespec freepages_start
, freepages_end
;
797 u64 freepages_delay
; /* wait for memory reclaim */
798 u32 freepages_count
; /* total count of memory reclaim */
800 #endif /* CONFIG_TASK_DELAY_ACCT */
802 static inline int sched_info_on(void)
804 #ifdef CONFIG_SCHEDSTATS
806 #elif defined(CONFIG_TASK_DELAY_ACCT)
807 extern int delayacct_on
;
822 * sched-domains (multiprocessor balancing) declarations:
826 * Increase resolution of nice-level calculations:
828 #define SCHED_LOAD_SHIFT 10
829 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
831 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
834 #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
835 #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
836 #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
837 #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
838 #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
839 #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
840 #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
841 #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
842 #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
843 #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
844 #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
846 #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
848 enum powersavings_balance_level
{
849 POWERSAVINGS_BALANCE_NONE
= 0, /* No power saving load balance */
850 POWERSAVINGS_BALANCE_BASIC
, /* Fill one thread/core/package
851 * first for long running threads
853 POWERSAVINGS_BALANCE_WAKEUP
, /* Also bias task wakeups to semi-idle
854 * cpu package for power savings
856 MAX_POWERSAVINGS_BALANCE_LEVELS
859 extern int sched_mc_power_savings
, sched_smt_power_savings
;
861 static inline int sd_balance_for_mc_power(void)
863 if (sched_smt_power_savings
)
864 return SD_POWERSAVINGS_BALANCE
;
866 return SD_PREFER_SIBLING
;
869 static inline int sd_balance_for_package_power(void)
871 if (sched_mc_power_savings
| sched_smt_power_savings
)
872 return SD_POWERSAVINGS_BALANCE
;
874 return SD_PREFER_SIBLING
;
878 * Optimise SD flags for power savings:
879 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
880 * Keep default SD flags if sched_{smt,mc}_power_saving=0
883 static inline int sd_power_saving_flags(void)
885 if (sched_mc_power_savings
| sched_smt_power_savings
)
886 return SD_BALANCE_NEWIDLE
;
892 struct sched_group
*next
; /* Must be a circular list */
895 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
898 unsigned int cpu_power
;
901 * The CPUs this group covers.
903 * NOTE: this field is variable length. (Allocated dynamically
904 * by attaching extra space to the end of the structure,
905 * depending on how many CPUs the kernel has booted up with)
907 * It is also be embedded into static data structures at build
908 * time. (See 'struct static_sched_group' in kernel/sched.c)
910 unsigned long cpumask
[0];
913 static inline struct cpumask
*sched_group_cpus(struct sched_group
*sg
)
915 return to_cpumask(sg
->cpumask
);
918 enum sched_domain_level
{
928 struct sched_domain_attr
{
929 int relax_domain_level
;
932 #define SD_ATTR_INIT (struct sched_domain_attr) { \
933 .relax_domain_level = -1, \
936 struct sched_domain
{
937 /* These fields must be setup */
938 struct sched_domain
*parent
; /* top domain must be null terminated */
939 struct sched_domain
*child
; /* bottom domain must be null terminated */
940 struct sched_group
*groups
; /* the balancing groups of the domain */
941 unsigned long min_interval
; /* Minimum balance interval ms */
942 unsigned long max_interval
; /* Maximum balance interval ms */
943 unsigned int busy_factor
; /* less balancing by factor if busy */
944 unsigned int imbalance_pct
; /* No balance until over watermark */
945 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
946 unsigned int busy_idx
;
947 unsigned int idle_idx
;
948 unsigned int newidle_idx
;
949 unsigned int wake_idx
;
950 unsigned int forkexec_idx
;
951 unsigned int smt_gain
;
952 int flags
; /* See SD_* */
953 enum sched_domain_level level
;
955 /* Runtime fields. */
956 unsigned long last_balance
; /* init to jiffies. units in jiffies */
957 unsigned int balance_interval
; /* initialise to 1. units in ms. */
958 unsigned int nr_balance_failed
; /* initialise to 0 */
962 #ifdef CONFIG_SCHEDSTATS
963 /* load_balance() stats */
964 unsigned int lb_count
[CPU_MAX_IDLE_TYPES
];
965 unsigned int lb_failed
[CPU_MAX_IDLE_TYPES
];
966 unsigned int lb_balanced
[CPU_MAX_IDLE_TYPES
];
967 unsigned int lb_imbalance
[CPU_MAX_IDLE_TYPES
];
968 unsigned int lb_gained
[CPU_MAX_IDLE_TYPES
];
969 unsigned int lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
970 unsigned int lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
971 unsigned int lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
973 /* Active load balancing */
974 unsigned int alb_count
;
975 unsigned int alb_failed
;
976 unsigned int alb_pushed
;
978 /* SD_BALANCE_EXEC stats */
979 unsigned int sbe_count
;
980 unsigned int sbe_balanced
;
981 unsigned int sbe_pushed
;
983 /* SD_BALANCE_FORK stats */
984 unsigned int sbf_count
;
985 unsigned int sbf_balanced
;
986 unsigned int sbf_pushed
;
988 /* try_to_wake_up() stats */
989 unsigned int ttwu_wake_remote
;
990 unsigned int ttwu_move_affine
;
991 unsigned int ttwu_move_balance
;
993 #ifdef CONFIG_SCHED_DEBUG
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
, cpumask_var_t doms_new
[],
1016 struct sched_domain_attr
*dattr_new
);
1018 /* Allocate an array of sched domains, for partition_sched_domains(). */
1019 cpumask_var_t
*alloc_sched_domains(unsigned int ndoms
);
1020 void free_sched_domains(cpumask_var_t doms
[], unsigned int ndoms
);
1022 /* Test a flag in parent sched domain */
1023 static inline int test_sd_parent(struct sched_domain
*sd
, int flag
)
1025 if (sd
->parent
&& (sd
->parent
->flags
& flag
))
1031 unsigned long default_scale_freq_power(struct sched_domain
*sd
, int cpu
);
1032 unsigned long default_scale_smt_power(struct sched_domain
*sd
, int cpu
);
1034 #else /* CONFIG_SMP */
1036 struct sched_domain_attr
;
1039 partition_sched_domains(int ndoms_new
, cpumask_var_t doms_new
[],
1040 struct sched_domain_attr
*dattr_new
)
1043 #endif /* !CONFIG_SMP */
1046 struct io_context
; /* See blkdev.h */
1049 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
1050 extern void prefetch_stack(struct task_struct
*t
);
1052 static inline void prefetch_stack(struct task_struct
*t
) { }
1055 struct audit_context
; /* See audit.c */
1057 struct pipe_inode_info
;
1058 struct uts_namespace
;
1061 struct sched_domain
;
1066 #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
1067 #define WF_FORK 0x02 /* child wakeup after fork */
1069 struct sched_class
{
1070 const struct sched_class
*next
;
1072 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
, int wakeup
);
1073 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
, int sleep
);
1074 void (*yield_task
) (struct rq
*rq
);
1076 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
, int flags
);
1078 struct task_struct
* (*pick_next_task
) (struct rq
*rq
);
1079 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
);
1082 int (*select_task_rq
)(struct task_struct
*p
, int sd_flag
, int flags
);
1084 unsigned long (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
1085 struct rq
*busiest
, unsigned long max_load_move
,
1086 struct sched_domain
*sd
, enum cpu_idle_type idle
,
1087 int *all_pinned
, int *this_best_prio
);
1089 int (*move_one_task
) (struct rq
*this_rq
, int this_cpu
,
1090 struct rq
*busiest
, struct sched_domain
*sd
,
1091 enum cpu_idle_type idle
);
1092 void (*pre_schedule
) (struct rq
*this_rq
, struct task_struct
*task
);
1093 void (*post_schedule
) (struct rq
*this_rq
);
1094 void (*task_wake_up
) (struct rq
*this_rq
, struct task_struct
*task
);
1096 void (*set_cpus_allowed
)(struct task_struct
*p
,
1097 const struct cpumask
*newmask
);
1099 void (*rq_online
)(struct rq
*rq
);
1100 void (*rq_offline
)(struct rq
*rq
);
1103 void (*set_curr_task
) (struct rq
*rq
);
1104 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
, int queued
);
1105 void (*task_fork
) (struct task_struct
*p
);
1107 void (*switched_from
) (struct rq
*this_rq
, struct task_struct
*task
,
1109 void (*switched_to
) (struct rq
*this_rq
, struct task_struct
*task
,
1111 void (*prio_changed
) (struct rq
*this_rq
, struct task_struct
*task
,
1112 int oldprio
, int running
);
1114 unsigned int (*get_rr_interval
) (struct rq
*rq
,
1115 struct task_struct
*task
);
1117 #ifdef CONFIG_FAIR_GROUP_SCHED
1118 void (*moved_group
) (struct task_struct
*p
);
1122 struct load_weight
{
1123 unsigned long weight
, inv_weight
;
1127 * CFS stats for a schedulable entity (task, task-group etc)
1129 * Current field usage histogram:
1136 struct sched_entity
{
1137 struct load_weight load
; /* for load-balancing */
1138 struct rb_node run_node
;
1139 struct list_head group_node
;
1143 u64 sum_exec_runtime
;
1145 u64 prev_sum_exec_runtime
;
1155 #ifdef CONFIG_SCHEDSTATS
1165 s64 sum_sleep_runtime
;
1172 u64 nr_migrations_cold
;
1173 u64 nr_failed_migrations_affine
;
1174 u64 nr_failed_migrations_running
;
1175 u64 nr_failed_migrations_hot
;
1176 u64 nr_forced_migrations
;
1177 u64 nr_forced2_migrations
;
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
;
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
;
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: */
1217 struct task_struct
{
1218 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
1221 unsigned int flags
; /* per process flags, defined below */
1222 unsigned int ptrace
;
1224 int lock_depth
; /* BKL lock depth */
1227 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
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
;
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
1251 unsigned char fpu_counter
;
1252 #ifdef CONFIG_BLK_DEV_IO_TRACE
1253 unsigned int btrace_seq
;
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
;
1270 struct list_head tasks
;
1271 struct plist_node pushable_tasks
;
1273 struct mm_struct
*mm
, *active_mm
;
1277 int exit_code
, exit_signal
;
1278 int pdeath_signal
; /* The signal sent when the parent dies */
1280 unsigned int personality
;
1281 unsigned did_exec
:1;
1282 unsigned in_execve
:1; /* Tell the LSMs that the process is doing an
1284 unsigned in_iowait
:1;
1287 /* Revert to default priority/policy when forking */
1288 unsigned sched_reset_on_fork
:1;
1293 #ifdef CONFIG_CC_STACKPROTECTOR
1294 /* Canary value for the -fstack-protector gcc feature */
1295 unsigned long stack_canary
;
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
;
1336 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
1337 cputime_t prev_utime
, prev_stime
;
1339 unsigned long nvcsw
, nivcsw
; /* context switch counts */
1340 struct timespec start_time
; /* monotonic time */
1341 struct timespec real_start_time
; /* boot based time */
1342 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1343 unsigned long min_flt
, maj_flt
;
1345 struct task_cputime cputime_expires
;
1346 struct list_head cpu_timers
[3];
1348 /* process credentials */
1349 const struct cred
*real_cred
; /* objective and real subjective task
1350 * credentials (COW) */
1351 const struct cred
*cred
; /* effective (overridable) subjective task
1352 * credentials (COW) */
1353 struct mutex cred_guard_mutex
; /* guard against foreign influences on
1354 * credential calculations
1355 * (notably. ptrace) */
1356 struct cred
*replacement_session_keyring
; /* for KEYCTL_SESSION_TO_PARENT */
1358 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1359 - access with [gs]et_task_comm (which lock
1360 it with task_lock())
1361 - initialized normally by flush_old_exec */
1362 /* file system info */
1363 int link_count
, total_link_count
;
1364 #ifdef CONFIG_SYSVIPC
1366 struct sysv_sem sysvsem
;
1368 #ifdef CONFIG_DETECT_HUNG_TASK
1369 /* hung task detection */
1370 unsigned long last_switch_count
;
1372 /* CPU-specific state of this task */
1373 struct thread_struct thread
;
1374 /* filesystem information */
1375 struct fs_struct
*fs
;
1376 /* open file information */
1377 struct files_struct
*files
;
1379 struct nsproxy
*nsproxy
;
1380 /* signal handlers */
1381 struct signal_struct
*signal
;
1382 struct sighand_struct
*sighand
;
1384 sigset_t blocked
, real_blocked
;
1385 sigset_t saved_sigmask
; /* restored if set_restore_sigmask() was used */
1386 struct sigpending pending
;
1388 unsigned long sas_ss_sp
;
1390 int (*notifier
)(void *priv
);
1391 void *notifier_data
;
1392 sigset_t
*notifier_mask
;
1393 struct audit_context
*audit_context
;
1394 #ifdef CONFIG_AUDITSYSCALL
1396 unsigned int sessionid
;
1400 /* Thread group tracking */
1403 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1405 spinlock_t alloc_lock
;
1407 #ifdef CONFIG_GENERIC_HARDIRQS
1408 /* IRQ handler threads */
1409 struct irqaction
*irqaction
;
1412 /* Protection of the PI data structures: */
1415 #ifdef CONFIG_RT_MUTEXES
1416 /* PI waiters blocked on a rt_mutex held by this task */
1417 struct plist_head pi_waiters
;
1418 /* Deadlock detection and priority inheritance handling */
1419 struct rt_mutex_waiter
*pi_blocked_on
;
1422 #ifdef CONFIG_DEBUG_MUTEXES
1423 /* mutex deadlock detection */
1424 struct mutex_waiter
*blocked_on
;
1426 #ifdef CONFIG_TRACE_IRQFLAGS
1427 unsigned int irq_events
;
1428 unsigned long hardirq_enable_ip
;
1429 unsigned long hardirq_disable_ip
;
1430 unsigned int hardirq_enable_event
;
1431 unsigned int hardirq_disable_event
;
1432 int hardirqs_enabled
;
1433 int hardirq_context
;
1434 unsigned long softirq_disable_ip
;
1435 unsigned long softirq_enable_ip
;
1436 unsigned int softirq_disable_event
;
1437 unsigned int softirq_enable_event
;
1438 int softirqs_enabled
;
1439 int softirq_context
;
1441 #ifdef CONFIG_LOCKDEP
1442 # define MAX_LOCK_DEPTH 48UL
1445 unsigned int lockdep_recursion
;
1446 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1447 gfp_t lockdep_reclaim_gfp
;
1450 /* journalling filesystem info */
1453 /* stacked block device info */
1454 struct bio
*bio_list
, **bio_tail
;
1457 struct reclaim_state
*reclaim_state
;
1459 struct backing_dev_info
*backing_dev_info
;
1461 struct io_context
*io_context
;
1463 unsigned long ptrace_message
;
1464 siginfo_t
*last_siginfo
; /* For ptrace use. */
1465 struct task_io_accounting ioac
;
1466 #if defined(CONFIG_TASK_XACCT)
1467 u64 acct_rss_mem1
; /* accumulated rss usage */
1468 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1469 cputime_t acct_timexpd
; /* stime + utime since last update */
1471 #ifdef CONFIG_CPUSETS
1472 nodemask_t mems_allowed
; /* Protected by alloc_lock */
1473 int cpuset_mem_spread_rotor
;
1475 #ifdef CONFIG_CGROUPS
1476 /* Control Group info protected by css_set_lock */
1477 struct css_set
*cgroups
;
1478 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1479 struct list_head cg_list
;
1482 struct robust_list_head __user
*robust_list
;
1483 #ifdef CONFIG_COMPAT
1484 struct compat_robust_list_head __user
*compat_robust_list
;
1486 struct list_head pi_state_list
;
1487 struct futex_pi_state
*pi_state_cache
;
1489 #ifdef CONFIG_PERF_EVENTS
1490 struct perf_event_context
*perf_event_ctxp
;
1491 struct mutex perf_event_mutex
;
1492 struct list_head perf_event_list
;
1495 struct mempolicy
*mempolicy
; /* Protected by alloc_lock */
1498 atomic_t fs_excl
; /* holding fs exclusive resources */
1499 struct rcu_head rcu
;
1502 * cache last used pipe for splice
1504 struct pipe_inode_info
*splice_pipe
;
1505 #ifdef CONFIG_TASK_DELAY_ACCT
1506 struct task_delay_info
*delays
;
1508 #ifdef CONFIG_FAULT_INJECTION
1511 struct prop_local_single dirties
;
1512 #ifdef CONFIG_LATENCYTOP
1513 int latency_record_count
;
1514 struct latency_record latency_record
[LT_SAVECOUNT
];
1517 * time slack values; these are used to round up poll() and
1518 * select() etc timeout values. These are in nanoseconds.
1520 unsigned long timer_slack_ns
;
1521 unsigned long default_timer_slack_ns
;
1523 struct list_head
*scm_work_list
;
1524 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1525 /* Index of current stored adress in ret_stack */
1527 /* Stack of return addresses for return function tracing */
1528 struct ftrace_ret_stack
*ret_stack
;
1529 /* time stamp for last schedule */
1530 unsigned long long ftrace_timestamp
;
1532 * Number of functions that haven't been traced
1533 * because of depth overrun.
1535 atomic_t trace_overrun
;
1536 /* Pause for the tracing */
1537 atomic_t tracing_graph_pause
;
1539 #ifdef CONFIG_TRACING
1540 /* state flags for use by tracers */
1541 unsigned long trace
;
1542 /* bitmask of trace recursion */
1543 unsigned long trace_recursion
;
1544 #endif /* CONFIG_TRACING */
1545 unsigned long stack_start
;
1548 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1549 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1552 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1553 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1554 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1555 * values are inverted: lower p->prio value means higher priority.
1557 * The MAX_USER_RT_PRIO value allows the actual maximum
1558 * RT priority to be separate from the value exported to
1559 * user-space. This allows kernel threads to set their
1560 * priority to a value higher than any user task. Note:
1561 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1564 #define MAX_USER_RT_PRIO 100
1565 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1567 #define MAX_PRIO (MAX_RT_PRIO + 40)
1568 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1570 static inline int rt_prio(int prio
)
1572 if (unlikely(prio
< MAX_RT_PRIO
))
1577 static inline int rt_task(struct task_struct
*p
)
1579 return rt_prio(p
->prio
);
1582 static inline struct pid
*task_pid(struct task_struct
*task
)
1584 return task
->pids
[PIDTYPE_PID
].pid
;
1587 static inline struct pid
*task_tgid(struct task_struct
*task
)
1589 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1593 * Without tasklist or rcu lock it is not safe to dereference
1594 * the result of task_pgrp/task_session even if task == current,
1595 * we can race with another thread doing sys_setsid/sys_setpgid.
1597 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1599 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1602 static inline struct pid
*task_session(struct task_struct
*task
)
1604 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1607 struct pid_namespace
;
1610 * the helpers to get the task's different pids as they are seen
1611 * from various namespaces
1613 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1614 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1616 * task_xid_nr_ns() : id seen from the ns specified;
1618 * set_task_vxid() : assigns a virtual id to a task;
1620 * see also pid_nr() etc in include/linux/pid.h
1622 pid_t
__task_pid_nr_ns(struct task_struct
*task
, enum pid_type type
,
1623 struct pid_namespace
*ns
);
1625 static inline pid_t
task_pid_nr(struct task_struct
*tsk
)
1630 static inline pid_t
task_pid_nr_ns(struct task_struct
*tsk
,
1631 struct pid_namespace
*ns
)
1633 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, ns
);
1636 static inline pid_t
task_pid_vnr(struct task_struct
*tsk
)
1638 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, NULL
);
1642 static inline pid_t
task_tgid_nr(struct task_struct
*tsk
)
1647 pid_t
task_tgid_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1649 static inline pid_t
task_tgid_vnr(struct task_struct
*tsk
)
1651 return pid_vnr(task_tgid(tsk
));
1655 static inline pid_t
task_pgrp_nr_ns(struct task_struct
*tsk
,
1656 struct pid_namespace
*ns
)
1658 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, ns
);
1661 static inline pid_t
task_pgrp_vnr(struct task_struct
*tsk
)
1663 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, NULL
);
1667 static inline pid_t
task_session_nr_ns(struct task_struct
*tsk
,
1668 struct pid_namespace
*ns
)
1670 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, ns
);
1673 static inline pid_t
task_session_vnr(struct task_struct
*tsk
)
1675 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, NULL
);
1678 /* obsolete, do not use */
1679 static inline pid_t
task_pgrp_nr(struct task_struct
*tsk
)
1681 return task_pgrp_nr_ns(tsk
, &init_pid_ns
);
1685 * pid_alive - check that a task structure is not stale
1686 * @p: Task structure to be checked.
1688 * Test if a process is not yet dead (at most zombie state)
1689 * If pid_alive fails, then pointers within the task structure
1690 * can be stale and must not be dereferenced.
1692 static inline int pid_alive(struct task_struct
*p
)
1694 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1698 * is_global_init - check if a task structure is init
1699 * @tsk: Task structure to be checked.
1701 * Check if a task structure is the first user space task the kernel created.
1703 static inline int is_global_init(struct task_struct
*tsk
)
1705 return tsk
->pid
== 1;
1709 * is_container_init:
1710 * check whether in the task is init in its own pid namespace.
1712 extern int is_container_init(struct task_struct
*tsk
);
1714 extern struct pid
*cad_pid
;
1716 extern void free_task(struct task_struct
*tsk
);
1717 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1719 extern void __put_task_struct(struct task_struct
*t
);
1721 static inline void put_task_struct(struct task_struct
*t
)
1723 if (atomic_dec_and_test(&t
->usage
))
1724 __put_task_struct(t
);
1727 extern void task_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1728 extern void thread_group_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1733 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1734 /* Not implemented yet, only for 486*/
1735 #define PF_STARTING 0x00000002 /* being created */
1736 #define PF_EXITING 0x00000004 /* getting shut down */
1737 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1738 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1739 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1740 #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1741 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1742 #define PF_DUMPCORE 0x00000200 /* dumped core */
1743 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1744 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1745 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1746 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1747 #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1748 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1749 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1750 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1751 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1752 #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1753 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1754 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1755 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1756 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1757 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1758 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1759 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1760 #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
1761 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1762 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1763 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1764 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1767 * Only the _current_ task can read/write to tsk->flags, but other
1768 * tasks can access tsk->flags in readonly mode for example
1769 * with tsk_used_math (like during threaded core dumping).
1770 * There is however an exception to this rule during ptrace
1771 * or during fork: the ptracer task is allowed to write to the
1772 * child->flags of its traced child (same goes for fork, the parent
1773 * can write to the child->flags), because we're guaranteed the
1774 * child is not running and in turn not changing child->flags
1775 * at the same time the parent does it.
1777 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1778 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1779 #define clear_used_math() clear_stopped_child_used_math(current)
1780 #define set_used_math() set_stopped_child_used_math(current)
1781 #define conditional_stopped_child_used_math(condition, child) \
1782 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1783 #define conditional_used_math(condition) \
1784 conditional_stopped_child_used_math(condition, current)
1785 #define copy_to_stopped_child_used_math(child) \
1786 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1787 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1788 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1789 #define used_math() tsk_used_math(current)
1791 #ifdef CONFIG_TREE_PREEMPT_RCU
1793 #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1794 #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1796 static inline void rcu_copy_process(struct task_struct
*p
)
1798 p
->rcu_read_lock_nesting
= 0;
1799 p
->rcu_read_unlock_special
= 0;
1800 p
->rcu_blocked_node
= NULL
;
1801 INIT_LIST_HEAD(&p
->rcu_node_entry
);
1806 static inline void rcu_copy_process(struct task_struct
*p
)
1813 extern int set_cpus_allowed_ptr(struct task_struct
*p
,
1814 const struct cpumask
*new_mask
);
1816 static inline int set_cpus_allowed_ptr(struct task_struct
*p
,
1817 const struct cpumask
*new_mask
)
1819 if (!cpumask_test_cpu(0, new_mask
))
1825 #ifndef CONFIG_CPUMASK_OFFSTACK
1826 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1828 return set_cpus_allowed_ptr(p
, &new_mask
);
1833 * Architectures can set this to 1 if they have specified
1834 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1835 * but then during bootup it turns out that sched_clock()
1836 * is reliable after all:
1838 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1839 extern int sched_clock_stable
;
1842 extern unsigned long long sched_clock(void);
1844 extern void sched_clock_init(void);
1845 extern u64
sched_clock_cpu(int cpu
);
1847 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1848 static inline void sched_clock_tick(void)
1852 static inline void sched_clock_idle_sleep_event(void)
1856 static inline void sched_clock_idle_wakeup_event(u64 delta_ns
)
1860 extern void sched_clock_tick(void);
1861 extern void sched_clock_idle_sleep_event(void);
1862 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1866 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1867 * clock constructed from sched_clock():
1869 extern unsigned long long cpu_clock(int cpu
);
1871 extern unsigned long long
1872 task_sched_runtime(struct task_struct
*task
);
1873 extern unsigned long long thread_group_sched_runtime(struct task_struct
*task
);
1875 /* sched_exec is called by processes performing an exec */
1877 extern void sched_exec(void);
1879 #define sched_exec() {}
1882 extern void sched_clock_idle_sleep_event(void);
1883 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1885 #ifdef CONFIG_HOTPLUG_CPU
1886 extern void idle_task_exit(void);
1888 static inline void idle_task_exit(void) {}
1891 extern void sched_idle_next(void);
1893 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1894 extern void wake_up_idle_cpu(int cpu
);
1896 static inline void wake_up_idle_cpu(int cpu
) { }
1899 extern unsigned int sysctl_sched_latency
;
1900 extern unsigned int sysctl_sched_min_granularity
;
1901 extern unsigned int sysctl_sched_wakeup_granularity
;
1902 extern unsigned int sysctl_sched_shares_ratelimit
;
1903 extern unsigned int sysctl_sched_shares_thresh
;
1904 extern unsigned int sysctl_sched_child_runs_first
;
1906 enum sched_tunable_scaling
{
1907 SCHED_TUNABLESCALING_NONE
,
1908 SCHED_TUNABLESCALING_LOG
,
1909 SCHED_TUNABLESCALING_LINEAR
,
1910 SCHED_TUNABLESCALING_END
,
1912 extern enum sched_tunable_scaling sysctl_sched_tunable_scaling
;
1914 #ifdef CONFIG_SCHED_DEBUG
1915 extern unsigned int sysctl_sched_migration_cost
;
1916 extern unsigned int sysctl_sched_nr_migrate
;
1917 extern unsigned int sysctl_sched_time_avg
;
1918 extern unsigned int sysctl_timer_migration
;
1920 int sched_proc_update_handler(struct ctl_table
*table
, int write
,
1921 void __user
*buffer
, size_t *length
,
1924 #ifdef CONFIG_SCHED_DEBUG
1925 static inline unsigned int get_sysctl_timer_migration(void)
1927 return sysctl_timer_migration
;
1930 static inline unsigned int get_sysctl_timer_migration(void)
1935 extern unsigned int sysctl_sched_rt_period
;
1936 extern int sysctl_sched_rt_runtime
;
1938 int sched_rt_handler(struct ctl_table
*table
, int write
,
1939 void __user
*buffer
, size_t *lenp
,
1942 extern unsigned int sysctl_sched_compat_yield
;
1944 #ifdef CONFIG_RT_MUTEXES
1945 extern int rt_mutex_getprio(struct task_struct
*p
);
1946 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1947 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1949 static inline int rt_mutex_getprio(struct task_struct
*p
)
1951 return p
->normal_prio
;
1953 # define rt_mutex_adjust_pi(p) do { } while (0)
1956 extern void set_user_nice(struct task_struct
*p
, long nice
);
1957 extern int task_prio(const struct task_struct
*p
);
1958 extern int task_nice(const struct task_struct
*p
);
1959 extern int can_nice(const struct task_struct
*p
, const int nice
);
1960 extern int task_curr(const struct task_struct
*p
);
1961 extern int idle_cpu(int cpu
);
1962 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1963 extern int sched_setscheduler_nocheck(struct task_struct
*, int,
1964 struct sched_param
*);
1965 extern struct task_struct
*idle_task(int cpu
);
1966 extern struct task_struct
*curr_task(int cpu
);
1967 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1972 * The default (Linux) execution domain.
1974 extern struct exec_domain default_exec_domain
;
1976 union thread_union
{
1977 struct thread_info thread_info
;
1978 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
1981 #ifndef __HAVE_ARCH_KSTACK_END
1982 static inline int kstack_end(void *addr
)
1984 /* Reliable end of stack detection:
1985 * Some APM bios versions misalign the stack
1987 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
1991 extern union thread_union init_thread_union
;
1992 extern struct task_struct init_task
;
1994 extern struct mm_struct init_mm
;
1996 extern struct pid_namespace init_pid_ns
;
1999 * find a task by one of its numerical ids
2001 * find_task_by_pid_ns():
2002 * finds a task by its pid in the specified namespace
2003 * find_task_by_vpid():
2004 * finds a task by its virtual pid
2006 * see also find_vpid() etc in include/linux/pid.h
2009 extern struct task_struct
*find_task_by_vpid(pid_t nr
);
2010 extern struct task_struct
*find_task_by_pid_ns(pid_t nr
,
2011 struct pid_namespace
*ns
);
2013 extern void __set_special_pids(struct pid
*pid
);
2015 /* per-UID process charging. */
2016 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
2017 static inline struct user_struct
*get_uid(struct user_struct
*u
)
2019 atomic_inc(&u
->__count
);
2022 extern void free_uid(struct user_struct
*);
2023 extern void release_uids(struct user_namespace
*ns
);
2025 #include <asm/current.h>
2027 extern void do_timer(unsigned long ticks
);
2029 extern int wake_up_state(struct task_struct
*tsk
, unsigned int state
);
2030 extern int wake_up_process(struct task_struct
*tsk
);
2031 extern void wake_up_new_task(struct task_struct
*tsk
,
2032 unsigned long clone_flags
);
2034 extern void kick_process(struct task_struct
*tsk
);
2036 static inline void kick_process(struct task_struct
*tsk
) { }
2038 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
2039 extern void sched_dead(struct task_struct
*p
);
2041 extern void proc_caches_init(void);
2042 extern void flush_signals(struct task_struct
*);
2043 extern void __flush_signals(struct task_struct
*);
2044 extern void ignore_signals(struct task_struct
*);
2045 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
2046 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
2048 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
2050 unsigned long flags
;
2053 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
2054 ret
= dequeue_signal(tsk
, mask
, info
);
2055 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
2060 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
2062 extern void unblock_all_signals(void);
2063 extern void release_task(struct task_struct
* p
);
2064 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
2065 extern int force_sigsegv(int, struct task_struct
*);
2066 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
2067 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
2068 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
2069 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
2070 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
2071 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
2072 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
2073 extern int do_notify_parent(struct task_struct
*, int);
2074 extern void __wake_up_parent(struct task_struct
*p
, struct task_struct
*parent
);
2075 extern void force_sig(int, struct task_struct
*);
2076 extern void force_sig_specific(int, struct task_struct
*);
2077 extern int send_sig(int, struct task_struct
*, int);
2078 extern void zap_other_threads(struct task_struct
*p
);
2079 extern struct sigqueue
*sigqueue_alloc(void);
2080 extern void sigqueue_free(struct sigqueue
*);
2081 extern int send_sigqueue(struct sigqueue
*, struct task_struct
*, int group
);
2082 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
2083 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
2085 static inline int kill_cad_pid(int sig
, int priv
)
2087 return kill_pid(cad_pid
, sig
, priv
);
2090 /* These can be the second arg to send_sig_info/send_group_sig_info. */
2091 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
2092 #define SEND_SIG_PRIV ((struct siginfo *) 1)
2093 #define SEND_SIG_FORCED ((struct siginfo *) 2)
2095 static inline int is_si_special(const struct siginfo
*info
)
2097 return info
<= SEND_SIG_FORCED
;
2101 * True if we are on the alternate signal stack.
2103 static inline int on_sig_stack(unsigned long sp
)
2105 #ifdef CONFIG_STACK_GROWSUP
2106 return sp
>= current
->sas_ss_sp
&&
2107 sp
- current
->sas_ss_sp
< current
->sas_ss_size
;
2109 return sp
> current
->sas_ss_sp
&&
2110 sp
- current
->sas_ss_sp
<= current
->sas_ss_size
;
2114 static inline int sas_ss_flags(unsigned long sp
)
2116 return (current
->sas_ss_size
== 0 ? SS_DISABLE
2117 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
2121 * Routines for handling mm_structs
2123 extern struct mm_struct
* mm_alloc(void);
2125 /* mmdrop drops the mm and the page tables */
2126 extern void __mmdrop(struct mm_struct
*);
2127 static inline void mmdrop(struct mm_struct
* mm
)
2129 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
2133 /* mmput gets rid of the mappings and all user-space */
2134 extern void mmput(struct mm_struct
*);
2135 /* Grab a reference to a task's mm, if it is not already going away */
2136 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
2137 /* Remove the current tasks stale references to the old mm_struct */
2138 extern void mm_release(struct task_struct
*, struct mm_struct
*);
2139 /* Allocate a new mm structure and copy contents from tsk->mm */
2140 extern struct mm_struct
*dup_mm(struct task_struct
*tsk
);
2142 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2143 struct task_struct
*, struct pt_regs
*);
2144 extern void flush_thread(void);
2145 extern void exit_thread(void);
2147 extern void exit_files(struct task_struct
*);
2148 extern void __cleanup_signal(struct signal_struct
*);
2149 extern void __cleanup_sighand(struct sighand_struct
*);
2151 extern void exit_itimers(struct signal_struct
*);
2152 extern void flush_itimer_signals(void);
2154 extern NORET_TYPE
void do_group_exit(int);
2156 extern void daemonize(const char *, ...);
2157 extern int allow_signal(int);
2158 extern int disallow_signal(int);
2160 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
2161 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
2162 struct task_struct
*fork_idle(int);
2164 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
2165 extern char *get_task_comm(char *to
, struct task_struct
*tsk
);
2168 extern void wait_task_context_switch(struct task_struct
*p
);
2169 extern unsigned long wait_task_inactive(struct task_struct
*, long match_state
);
2171 static inline void wait_task_context_switch(struct task_struct
*p
) {}
2172 static inline unsigned long wait_task_inactive(struct task_struct
*p
,
2179 #define next_task(p) \
2180 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2182 #define for_each_process(p) \
2183 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2185 extern bool current_is_single_threaded(void);
2188 * Careful: do_each_thread/while_each_thread is a double loop so
2189 * 'break' will not work as expected - use goto instead.
2191 #define do_each_thread(g, t) \
2192 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2194 #define while_each_thread(g, t) \
2195 while ((t = next_thread(t)) != g)
2197 /* de_thread depends on thread_group_leader not being a pid based check */
2198 #define thread_group_leader(p) (p == p->group_leader)
2200 /* Do to the insanities of de_thread it is possible for a process
2201 * to have the pid of the thread group leader without actually being
2202 * the thread group leader. For iteration through the pids in proc
2203 * all we care about is that we have a task with the appropriate
2204 * pid, we don't actually care if we have the right task.
2206 static inline int has_group_leader_pid(struct task_struct
*p
)
2208 return p
->pid
== p
->tgid
;
2212 int same_thread_group(struct task_struct
*p1
, struct task_struct
*p2
)
2214 return p1
->tgid
== p2
->tgid
;
2217 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
2219 return list_entry_rcu(p
->thread_group
.next
,
2220 struct task_struct
, thread_group
);
2223 static inline int thread_group_empty(struct task_struct
*p
)
2225 return list_empty(&p
->thread_group
);
2228 #define delay_group_leader(p) \
2229 (thread_group_leader(p) && !thread_group_empty(p))
2231 static inline int task_detached(struct task_struct
*p
)
2233 return p
->exit_signal
== -1;
2237 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2238 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2239 * pins the final release of task.io_context. Also protects ->cpuset and
2240 * ->cgroup.subsys[].
2242 * Nests both inside and outside of read_lock(&tasklist_lock).
2243 * It must not be nested with write_lock_irq(&tasklist_lock),
2244 * neither inside nor outside.
2246 static inline void task_lock(struct task_struct
*p
)
2248 spin_lock(&p
->alloc_lock
);
2251 static inline void task_unlock(struct task_struct
*p
)
2253 spin_unlock(&p
->alloc_lock
);
2256 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
2257 unsigned long *flags
);
2259 static inline void unlock_task_sighand(struct task_struct
*tsk
,
2260 unsigned long *flags
)
2262 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
2265 #ifndef __HAVE_THREAD_FUNCTIONS
2267 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2268 #define task_stack_page(task) ((task)->stack)
2270 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
2272 *task_thread_info(p
) = *task_thread_info(org
);
2273 task_thread_info(p
)->task
= p
;
2276 static inline unsigned long *end_of_stack(struct task_struct
*p
)
2278 return (unsigned long *)(task_thread_info(p
) + 1);
2283 static inline int object_is_on_stack(void *obj
)
2285 void *stack
= task_stack_page(current
);
2287 return (obj
>= stack
) && (obj
< (stack
+ THREAD_SIZE
));
2290 extern void thread_info_cache_init(void);
2292 #ifdef CONFIG_DEBUG_STACK_USAGE
2293 static inline unsigned long stack_not_used(struct task_struct
*p
)
2295 unsigned long *n
= end_of_stack(p
);
2297 do { /* Skip over canary */
2301 return (unsigned long)n
- (unsigned long)end_of_stack(p
);
2305 /* set thread flags in other task's structures
2306 * - see asm/thread_info.h for TIF_xxxx flags available
2308 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2310 set_ti_thread_flag(task_thread_info(tsk
), flag
);
2313 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2315 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2318 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2320 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
2323 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2325 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2328 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2330 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
2333 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
2335 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2338 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
2340 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2343 static inline int test_tsk_need_resched(struct task_struct
*tsk
)
2345 return unlikely(test_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
));
2348 static inline int restart_syscall(void)
2350 set_tsk_thread_flag(current
, TIF_SIGPENDING
);
2351 return -ERESTARTNOINTR
;
2354 static inline int signal_pending(struct task_struct
*p
)
2356 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
2359 static inline int __fatal_signal_pending(struct task_struct
*p
)
2361 return unlikely(sigismember(&p
->pending
.signal
, SIGKILL
));
2364 static inline int fatal_signal_pending(struct task_struct
*p
)
2366 return signal_pending(p
) && __fatal_signal_pending(p
);
2369 static inline int signal_pending_state(long state
, struct task_struct
*p
)
2371 if (!(state
& (TASK_INTERRUPTIBLE
| TASK_WAKEKILL
)))
2373 if (!signal_pending(p
))
2376 return (state
& TASK_INTERRUPTIBLE
) || __fatal_signal_pending(p
);
2379 static inline int need_resched(void)
2381 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
2385 * cond_resched() and cond_resched_lock(): latency reduction via
2386 * explicit rescheduling in places that are safe. The return
2387 * value indicates whether a reschedule was done in fact.
2388 * cond_resched_lock() will drop the spinlock before scheduling,
2389 * cond_resched_softirq() will enable bhs before scheduling.
2391 extern int _cond_resched(void);
2393 #define cond_resched() ({ \
2394 __might_sleep(__FILE__, __LINE__, 0); \
2398 extern int __cond_resched_lock(spinlock_t
*lock
);
2400 #ifdef CONFIG_PREEMPT
2401 #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
2403 #define PREEMPT_LOCK_OFFSET 0
2406 #define cond_resched_lock(lock) ({ \
2407 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
2408 __cond_resched_lock(lock); \
2411 extern int __cond_resched_softirq(void);
2413 #define cond_resched_softirq() ({ \
2414 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2415 __cond_resched_softirq(); \
2419 * Does a critical section need to be broken due to another
2420 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2421 * but a general need for low latency)
2423 static inline int spin_needbreak(spinlock_t
*lock
)
2425 #ifdef CONFIG_PREEMPT
2426 return spin_is_contended(lock
);
2433 * Thread group CPU time accounting.
2435 void thread_group_cputime(struct task_struct
*tsk
, struct task_cputime
*times
);
2436 void thread_group_cputimer(struct task_struct
*tsk
, struct task_cputime
*times
);
2438 static inline void thread_group_cputime_init(struct signal_struct
*sig
)
2440 sig
->cputimer
.cputime
= INIT_CPUTIME
;
2441 spin_lock_init(&sig
->cputimer
.lock
);
2442 sig
->cputimer
.running
= 0;
2445 static inline void thread_group_cputime_free(struct signal_struct
*sig
)
2450 * Reevaluate whether the task has signals pending delivery.
2451 * Wake the task if so.
2452 * This is required every time the blocked sigset_t changes.
2453 * callers must hold sighand->siglock.
2455 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
2456 extern void recalc_sigpending(void);
2458 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
2461 * Wrappers for p->thread_info->cpu access. No-op on UP.
2465 static inline unsigned int task_cpu(const struct task_struct
*p
)
2467 return task_thread_info(p
)->cpu
;
2470 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
2474 static inline unsigned int task_cpu(const struct task_struct
*p
)
2479 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
2483 #endif /* CONFIG_SMP */
2485 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
2487 #ifdef CONFIG_TRACING
2489 __trace_special(void *__tr
, void *__data
,
2490 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
);
2493 __trace_special(void *__tr
, void *__data
,
2494 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
)
2499 extern long sched_setaffinity(pid_t pid
, const struct cpumask
*new_mask
);
2500 extern long sched_getaffinity(pid_t pid
, struct cpumask
*mask
);
2502 extern void normalize_rt_tasks(void);
2504 #ifdef CONFIG_GROUP_SCHED
2506 extern struct task_group init_task_group
;
2507 #ifdef CONFIG_USER_SCHED
2508 extern struct task_group root_task_group
;
2509 extern void set_tg_uid(struct user_struct
*user
);
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
);
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
,
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
);
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
)
2549 static inline void inc_syscw(struct task_struct
*tsk
)
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
)
2571 #ifndef TASK_SIZE_OF
2572 #define TASK_SIZE_OF(tsk) TASK_SIZE
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
);
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 #endif /* __KERNEL__ */