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
195 #define TASK_STATE_MAX 512
197 #define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
199 extern char ___assert_task_state
[1 - 2*!!(
200 sizeof(TASK_STATE_TO_CHAR_STR
)-1 != ilog2(TASK_STATE_MAX
)+1)];
202 /* Convenience macros for the sake of set_task_state */
203 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
204 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
205 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
207 /* Convenience macros for the sake of wake_up */
208 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
209 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
211 /* get_task_state() */
212 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
213 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
216 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
217 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
218 #define task_is_stopped_or_traced(task) \
219 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
220 #define task_contributes_to_load(task) \
221 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
222 (task->flags & PF_FREEZING) == 0)
224 #define __set_task_state(tsk, state_value) \
225 do { (tsk)->state = (state_value); } while (0)
226 #define set_task_state(tsk, state_value) \
227 set_mb((tsk)->state, (state_value))
230 * set_current_state() includes a barrier so that the write of current->state
231 * is correctly serialised wrt the caller's subsequent test of whether to
234 * set_current_state(TASK_UNINTERRUPTIBLE);
235 * if (do_i_need_to_sleep())
238 * If the caller does not need such serialisation then use __set_current_state()
240 #define __set_current_state(state_value) \
241 do { current->state = (state_value); } while (0)
242 #define set_current_state(state_value) \
243 set_mb(current->state, (state_value))
245 /* Task command name length */
246 #define TASK_COMM_LEN 16
248 #include <linux/spinlock.h>
251 * This serializes "schedule()" and also protects
252 * the run-queue from deletions/modifications (but
253 * _adding_ to the beginning of the run-queue has
256 extern rwlock_t tasklist_lock
;
257 extern spinlock_t mmlist_lock
;
261 extern void sched_init(void);
262 extern void sched_init_smp(void);
263 extern asmlinkage
void schedule_tail(struct task_struct
*prev
);
264 extern void init_idle(struct task_struct
*idle
, int cpu
);
265 extern void init_idle_bootup_task(struct task_struct
*idle
);
267 extern int runqueue_is_locked(int cpu
);
268 extern void task_rq_unlock_wait(struct task_struct
*p
);
270 extern cpumask_var_t nohz_cpu_mask
;
271 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
272 extern int select_nohz_load_balancer(int cpu
);
273 extern int get_nohz_load_balancer(void);
275 static inline int select_nohz_load_balancer(int cpu
)
282 * Only dump TASK_* tasks. (0 for all tasks)
284 extern void show_state_filter(unsigned long state_filter
);
286 static inline void show_state(void)
288 show_state_filter(0);
291 extern void show_regs(struct pt_regs
*);
294 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
295 * task), SP is the stack pointer of the first frame that should be shown in the back
296 * trace (or NULL if the entire call-chain of the task should be shown).
298 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
300 void io_schedule(void);
301 long io_schedule_timeout(long timeout
);
303 extern void cpu_init (void);
304 extern void trap_init(void);
305 extern void update_process_times(int user
);
306 extern void scheduler_tick(void);
308 extern void sched_show_task(struct task_struct
*p
);
310 #ifdef CONFIG_DETECT_SOFTLOCKUP
311 extern void softlockup_tick(void);
312 extern void touch_softlockup_watchdog(void);
313 extern void touch_all_softlockup_watchdogs(void);
314 extern int proc_dosoftlockup_thresh(struct ctl_table
*table
, int write
,
316 size_t *lenp
, loff_t
*ppos
);
317 extern unsigned int softlockup_panic
;
318 extern int softlockup_thresh
;
320 static inline void softlockup_tick(void)
323 static inline void touch_softlockup_watchdog(void)
326 static inline void touch_all_softlockup_watchdogs(void)
331 #ifdef CONFIG_DETECT_HUNG_TASK
332 extern unsigned int sysctl_hung_task_panic
;
333 extern unsigned long sysctl_hung_task_check_count
;
334 extern unsigned long sysctl_hung_task_timeout_secs
;
335 extern unsigned long sysctl_hung_task_warnings
;
336 extern int proc_dohung_task_timeout_secs(struct ctl_table
*table
, int write
,
338 size_t *lenp
, loff_t
*ppos
);
341 /* Attach to any functions which should be ignored in wchan output. */
342 #define __sched __attribute__((__section__(".sched.text")))
344 /* Linker adds these: start and end of __sched functions */
345 extern char __sched_text_start
[], __sched_text_end
[];
347 /* Is this address in the __sched functions? */
348 extern int in_sched_functions(unsigned long addr
);
350 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
351 extern signed long schedule_timeout(signed long timeout
);
352 extern signed long schedule_timeout_interruptible(signed long timeout
);
353 extern signed long schedule_timeout_killable(signed long timeout
);
354 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
355 asmlinkage
void schedule(void);
356 extern int mutex_spin_on_owner(struct mutex
*lock
, struct thread_info
*owner
);
359 struct user_namespace
;
362 * Default maximum number of active map areas, this limits the number of vmas
363 * per mm struct. Users can overwrite this number by sysctl but there is a
366 * When a program's coredump is generated as ELF format, a section is created
367 * per a vma. In ELF, the number of sections is represented in unsigned short.
368 * This means the number of sections should be smaller than 65535 at coredump.
369 * Because the kernel adds some informative sections to a image of program at
370 * generating coredump, we need some margin. The number of extra sections is
371 * 1-3 now and depends on arch. We use "5" as safe margin, here.
373 #define MAPCOUNT_ELF_CORE_MARGIN (5)
374 #define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
376 extern int sysctl_max_map_count
;
378 #include <linux/aio.h>
381 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
383 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
384 unsigned long, unsigned long);
386 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
387 unsigned long len
, unsigned long pgoff
,
388 unsigned long flags
);
389 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
390 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
392 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
) {}
395 #if USE_SPLIT_PTLOCKS
397 * The mm counters are not protected by its page_table_lock,
398 * so must be incremented atomically.
400 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
401 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
402 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
403 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
404 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
406 #else /* !USE_SPLIT_PTLOCKS */
408 * The mm counters are protected by its page_table_lock,
409 * so can be incremented directly.
411 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
412 #define get_mm_counter(mm, member) ((mm)->_##member)
413 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
414 #define inc_mm_counter(mm, member) (mm)->_##member++
415 #define dec_mm_counter(mm, member) (mm)->_##member--
417 #endif /* !USE_SPLIT_PTLOCKS */
419 #define get_mm_rss(mm) \
420 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
421 #define update_hiwater_rss(mm) do { \
422 unsigned long _rss = get_mm_rss(mm); \
423 if ((mm)->hiwater_rss < _rss) \
424 (mm)->hiwater_rss = _rss; \
426 #define update_hiwater_vm(mm) do { \
427 if ((mm)->hiwater_vm < (mm)->total_vm) \
428 (mm)->hiwater_vm = (mm)->total_vm; \
431 static inline unsigned long get_mm_hiwater_rss(struct mm_struct
*mm
)
433 return max(mm
->hiwater_rss
, get_mm_rss(mm
));
436 static inline void setmax_mm_hiwater_rss(unsigned long *maxrss
,
437 struct mm_struct
*mm
)
439 unsigned long hiwater_rss
= get_mm_hiwater_rss(mm
);
441 if (*maxrss
< hiwater_rss
)
442 *maxrss
= hiwater_rss
;
445 static inline unsigned long get_mm_hiwater_vm(struct mm_struct
*mm
)
447 return max(mm
->hiwater_vm
, mm
->total_vm
);
450 extern void set_dumpable(struct mm_struct
*mm
, int value
);
451 extern int get_dumpable(struct mm_struct
*mm
);
455 #define MMF_DUMPABLE 0 /* core dump is permitted */
456 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
458 #define MMF_DUMPABLE_BITS 2
459 #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
461 /* coredump filter bits */
462 #define MMF_DUMP_ANON_PRIVATE 2
463 #define MMF_DUMP_ANON_SHARED 3
464 #define MMF_DUMP_MAPPED_PRIVATE 4
465 #define MMF_DUMP_MAPPED_SHARED 5
466 #define MMF_DUMP_ELF_HEADERS 6
467 #define MMF_DUMP_HUGETLB_PRIVATE 7
468 #define MMF_DUMP_HUGETLB_SHARED 8
470 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
471 #define MMF_DUMP_FILTER_BITS 7
472 #define MMF_DUMP_FILTER_MASK \
473 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
474 #define MMF_DUMP_FILTER_DEFAULT \
475 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
476 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
478 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
479 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
481 # define MMF_DUMP_MASK_DEFAULT_ELF 0
483 /* leave room for more dump flags */
484 #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
486 #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
488 struct sighand_struct
{
490 struct k_sigaction action
[_NSIG
];
492 wait_queue_head_t signalfd_wqh
;
495 struct pacct_struct
{
498 unsigned long ac_mem
;
499 cputime_t ac_utime
, ac_stime
;
500 unsigned long ac_minflt
, ac_majflt
;
511 * struct task_cputime - collected CPU time counts
512 * @utime: time spent in user mode, in &cputime_t units
513 * @stime: time spent in kernel mode, in &cputime_t units
514 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
516 * This structure groups together three kinds of CPU time that are
517 * tracked for threads and thread groups. Most things considering
518 * CPU time want to group these counts together and treat all three
519 * of them in parallel.
521 struct task_cputime
{
524 unsigned long long sum_exec_runtime
;
526 /* Alternate field names when used to cache expirations. */
527 #define prof_exp stime
528 #define virt_exp utime
529 #define sched_exp sum_exec_runtime
531 #define INIT_CPUTIME \
532 (struct task_cputime) { \
533 .utime = cputime_zero, \
534 .stime = cputime_zero, \
535 .sum_exec_runtime = 0, \
539 * Disable preemption until the scheduler is running.
540 * Reset by start_kernel()->sched_init()->init_idle().
542 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
543 * before the scheduler is active -- see should_resched().
545 #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
548 * struct thread_group_cputimer - thread group interval timer counts
549 * @cputime: thread group interval timers.
550 * @running: non-zero when there are timers running and
551 * @cputime receives updates.
552 * @lock: lock for fields in this struct.
554 * This structure contains the version of task_cputime, above, that is
555 * used for thread group CPU timer calculations.
557 struct thread_group_cputimer
{
558 struct task_cputime cputime
;
564 * NOTE! "signal_struct" does not have it's own
565 * locking, because a shared signal_struct always
566 * implies a shared sighand_struct, so locking
567 * sighand_struct is always a proper superset of
568 * the locking of signal_struct.
570 struct signal_struct
{
574 wait_queue_head_t wait_chldexit
; /* for wait4() */
576 /* current thread group signal load-balancing target: */
577 struct task_struct
*curr_target
;
579 /* shared signal handling: */
580 struct sigpending shared_pending
;
582 /* thread group exit support */
585 * - notify group_exit_task when ->count is equal to notify_count
586 * - everyone except group_exit_task is stopped during signal delivery
587 * of fatal signals, group_exit_task processes the signal.
590 struct task_struct
*group_exit_task
;
592 /* thread group stop support, overloads group_exit_code too */
593 int group_stop_count
;
594 unsigned int flags
; /* see SIGNAL_* flags below */
596 /* POSIX.1b Interval Timers */
597 struct list_head posix_timers
;
599 /* ITIMER_REAL timer for the process */
600 struct hrtimer real_timer
;
601 struct pid
*leader_pid
;
602 ktime_t it_real_incr
;
605 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
606 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
607 * values are defined to 0 and 1 respectively
609 struct cpu_itimer it
[2];
612 * Thread group totals for process CPU timers.
613 * See thread_group_cputimer(), et al, for details.
615 struct thread_group_cputimer cputimer
;
617 /* Earliest-expiration cache. */
618 struct task_cputime cputime_expires
;
620 struct list_head cpu_timers
[3];
622 struct pid
*tty_old_pgrp
;
624 /* boolean value for session group leader */
627 struct tty_struct
*tty
; /* NULL if no tty */
630 * Cumulative resource counters for dead threads in the group,
631 * and for reaped dead child processes forked by this group.
632 * Live threads maintain their own counters and add to these
633 * in __exit_signal, except for the group leader.
635 cputime_t utime
, stime
, cutime
, cstime
;
638 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
639 cputime_t prev_utime
, prev_stime
;
641 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
642 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
643 unsigned long inblock
, oublock
, cinblock
, coublock
;
644 unsigned long maxrss
, cmaxrss
;
645 struct task_io_accounting ioac
;
648 * Cumulative ns of schedule CPU time fo dead threads in the
649 * group, not including a zombie group leader, (This only differs
650 * from jiffies_to_ns(utime + stime) if sched_clock uses something
651 * other than jiffies.)
653 unsigned long long sum_sched_runtime
;
656 * We don't bother to synchronize most readers of this at all,
657 * because there is no reader checking a limit that actually needs
658 * to get both rlim_cur and rlim_max atomically, and either one
659 * alone is a single word that can safely be read normally.
660 * getrlimit/setrlimit use task_lock(current->group_leader) to
661 * protect this instead of the siglock, because they really
662 * have no need to disable irqs.
664 struct rlimit rlim
[RLIM_NLIMITS
];
666 #ifdef CONFIG_BSD_PROCESS_ACCT
667 struct pacct_struct pacct
; /* per-process accounting information */
669 #ifdef CONFIG_TASKSTATS
670 struct taskstats
*stats
;
674 struct tty_audit_buf
*tty_audit_buf
;
677 int oom_adj
; /* OOM kill score adjustment (bit shift) */
680 /* Context switch must be unlocked if interrupts are to be enabled */
681 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
682 # define __ARCH_WANT_UNLOCKED_CTXSW
686 * Bits in flags field of signal_struct.
688 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
689 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
690 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
691 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
693 * Pending notifications to parent.
695 #define SIGNAL_CLD_STOPPED 0x00000010
696 #define SIGNAL_CLD_CONTINUED 0x00000020
697 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
699 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
701 /* If true, all threads except ->group_exit_task have pending SIGKILL */
702 static inline int signal_group_exit(const struct signal_struct
*sig
)
704 return (sig
->flags
& SIGNAL_GROUP_EXIT
) ||
705 (sig
->group_exit_task
!= NULL
);
709 * Some day this will be a full-fledged user tracking system..
712 atomic_t __count
; /* reference count */
713 atomic_t processes
; /* How many processes does this user have? */
714 atomic_t files
; /* How many open files does this user have? */
715 atomic_t sigpending
; /* How many pending signals does this user have? */
716 #ifdef CONFIG_INOTIFY_USER
717 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
718 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
721 atomic_t epoll_watches
; /* The number of file descriptors currently watched */
723 #ifdef CONFIG_POSIX_MQUEUE
724 /* protected by mq_lock */
725 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
727 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
730 struct key
*uid_keyring
; /* UID specific keyring */
731 struct key
*session_keyring
; /* UID's default session keyring */
734 /* Hash table maintenance information */
735 struct hlist_node uidhash_node
;
737 struct user_namespace
*user_ns
;
739 #ifdef CONFIG_USER_SCHED
740 struct task_group
*tg
;
743 struct delayed_work work
;
747 #ifdef CONFIG_PERF_EVENTS
748 atomic_long_t locked_vm
;
752 extern int uids_sysfs_init(void);
754 extern struct user_struct
*find_user(uid_t
);
756 extern struct user_struct root_user
;
757 #define INIT_USER (&root_user)
760 struct backing_dev_info
;
761 struct reclaim_state
;
763 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
765 /* cumulative counters */
766 unsigned long pcount
; /* # of times run on this cpu */
767 unsigned long long run_delay
; /* time spent waiting on a runqueue */
770 unsigned long long last_arrival
,/* when we last ran on a cpu */
771 last_queued
; /* when we were last queued to run */
772 #ifdef CONFIG_SCHEDSTATS
774 unsigned int bkl_count
;
777 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
779 #ifdef CONFIG_TASK_DELAY_ACCT
780 struct task_delay_info
{
782 unsigned int flags
; /* Private per-task flags */
784 /* For each stat XXX, add following, aligned appropriately
786 * struct timespec XXX_start, XXX_end;
790 * Atomicity of updates to XXX_delay, XXX_count protected by
791 * single lock above (split into XXX_lock if contention is an issue).
795 * XXX_count is incremented on every XXX operation, the delay
796 * associated with the operation is added to XXX_delay.
797 * XXX_delay contains the accumulated delay time in nanoseconds.
799 struct timespec blkio_start
, blkio_end
; /* Shared by blkio, swapin */
800 u64 blkio_delay
; /* wait for sync block io completion */
801 u64 swapin_delay
; /* wait for swapin block io completion */
802 u32 blkio_count
; /* total count of the number of sync block */
803 /* io operations performed */
804 u32 swapin_count
; /* total count of the number of swapin block */
805 /* io operations performed */
807 struct timespec freepages_start
, freepages_end
;
808 u64 freepages_delay
; /* wait for memory reclaim */
809 u32 freepages_count
; /* total count of memory reclaim */
811 #endif /* CONFIG_TASK_DELAY_ACCT */
813 static inline int sched_info_on(void)
815 #ifdef CONFIG_SCHEDSTATS
817 #elif defined(CONFIG_TASK_DELAY_ACCT)
818 extern int delayacct_on
;
833 * sched-domains (multiprocessor balancing) declarations:
837 * Increase resolution of nice-level calculations:
839 #define SCHED_LOAD_SHIFT 10
840 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
842 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
845 #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
846 #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
847 #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
848 #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
849 #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
850 #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
851 #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
852 #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
853 #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
854 #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
855 #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
857 #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
859 enum powersavings_balance_level
{
860 POWERSAVINGS_BALANCE_NONE
= 0, /* No power saving load balance */
861 POWERSAVINGS_BALANCE_BASIC
, /* Fill one thread/core/package
862 * first for long running threads
864 POWERSAVINGS_BALANCE_WAKEUP
, /* Also bias task wakeups to semi-idle
865 * cpu package for power savings
867 MAX_POWERSAVINGS_BALANCE_LEVELS
870 extern int sched_mc_power_savings
, sched_smt_power_savings
;
872 static inline int sd_balance_for_mc_power(void)
874 if (sched_smt_power_savings
)
875 return SD_POWERSAVINGS_BALANCE
;
877 return SD_PREFER_SIBLING
;
880 static inline int sd_balance_for_package_power(void)
882 if (sched_mc_power_savings
| sched_smt_power_savings
)
883 return SD_POWERSAVINGS_BALANCE
;
885 return SD_PREFER_SIBLING
;
889 * Optimise SD flags for power savings:
890 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
891 * Keep default SD flags if sched_{smt,mc}_power_saving=0
894 static inline int sd_power_saving_flags(void)
896 if (sched_mc_power_savings
| sched_smt_power_savings
)
897 return SD_BALANCE_NEWIDLE
;
903 struct sched_group
*next
; /* Must be a circular list */
906 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
909 unsigned int cpu_power
;
912 * The CPUs this group covers.
914 * NOTE: this field is variable length. (Allocated dynamically
915 * by attaching extra space to the end of the structure,
916 * depending on how many CPUs the kernel has booted up with)
918 * It is also be embedded into static data structures at build
919 * time. (See 'struct static_sched_group' in kernel/sched.c)
921 unsigned long cpumask
[0];
924 static inline struct cpumask
*sched_group_cpus(struct sched_group
*sg
)
926 return to_cpumask(sg
->cpumask
);
929 enum sched_domain_level
{
939 struct sched_domain_attr
{
940 int relax_domain_level
;
943 #define SD_ATTR_INIT (struct sched_domain_attr) { \
944 .relax_domain_level = -1, \
947 struct sched_domain
{
948 /* These fields must be setup */
949 struct sched_domain
*parent
; /* top domain must be null terminated */
950 struct sched_domain
*child
; /* bottom domain must be null terminated */
951 struct sched_group
*groups
; /* the balancing groups of the domain */
952 unsigned long min_interval
; /* Minimum balance interval ms */
953 unsigned long max_interval
; /* Maximum balance interval ms */
954 unsigned int busy_factor
; /* less balancing by factor if busy */
955 unsigned int imbalance_pct
; /* No balance until over watermark */
956 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
957 unsigned int busy_idx
;
958 unsigned int idle_idx
;
959 unsigned int newidle_idx
;
960 unsigned int wake_idx
;
961 unsigned int forkexec_idx
;
962 unsigned int smt_gain
;
963 int flags
; /* See SD_* */
964 enum sched_domain_level level
;
966 /* Runtime fields. */
967 unsigned long last_balance
; /* init to jiffies. units in jiffies */
968 unsigned int balance_interval
; /* initialise to 1. units in ms. */
969 unsigned int nr_balance_failed
; /* initialise to 0 */
973 #ifdef CONFIG_SCHEDSTATS
974 /* load_balance() stats */
975 unsigned int lb_count
[CPU_MAX_IDLE_TYPES
];
976 unsigned int lb_failed
[CPU_MAX_IDLE_TYPES
];
977 unsigned int lb_balanced
[CPU_MAX_IDLE_TYPES
];
978 unsigned int lb_imbalance
[CPU_MAX_IDLE_TYPES
];
979 unsigned int lb_gained
[CPU_MAX_IDLE_TYPES
];
980 unsigned int lb_hot_gained
[CPU_MAX_IDLE_TYPES
];
981 unsigned int lb_nobusyg
[CPU_MAX_IDLE_TYPES
];
982 unsigned int lb_nobusyq
[CPU_MAX_IDLE_TYPES
];
984 /* Active load balancing */
985 unsigned int alb_count
;
986 unsigned int alb_failed
;
987 unsigned int alb_pushed
;
989 /* SD_BALANCE_EXEC stats */
990 unsigned int sbe_count
;
991 unsigned int sbe_balanced
;
992 unsigned int sbe_pushed
;
994 /* SD_BALANCE_FORK stats */
995 unsigned int sbf_count
;
996 unsigned int sbf_balanced
;
997 unsigned int sbf_pushed
;
999 /* try_to_wake_up() stats */
1000 unsigned int ttwu_wake_remote
;
1001 unsigned int ttwu_move_affine
;
1002 unsigned int ttwu_move_balance
;
1004 #ifdef CONFIG_SCHED_DEBUG
1009 * Span of all CPUs in this domain.
1011 * NOTE: this field is variable length. (Allocated dynamically
1012 * by attaching extra space to the end of the structure,
1013 * depending on how many CPUs the kernel has booted up with)
1015 * It is also be embedded into static data structures at build
1016 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1018 unsigned long span
[0];
1021 static inline struct cpumask
*sched_domain_span(struct sched_domain
*sd
)
1023 return to_cpumask(sd
->span
);
1026 extern void partition_sched_domains(int ndoms_new
, cpumask_var_t doms_new
[],
1027 struct sched_domain_attr
*dattr_new
);
1029 /* Allocate an array of sched domains, for partition_sched_domains(). */
1030 cpumask_var_t
*alloc_sched_domains(unsigned int ndoms
);
1031 void free_sched_domains(cpumask_var_t doms
[], unsigned int ndoms
);
1033 /* Test a flag in parent sched domain */
1034 static inline int test_sd_parent(struct sched_domain
*sd
, int flag
)
1036 if (sd
->parent
&& (sd
->parent
->flags
& flag
))
1042 unsigned long default_scale_freq_power(struct sched_domain
*sd
, int cpu
);
1043 unsigned long default_scale_smt_power(struct sched_domain
*sd
, int cpu
);
1045 #else /* CONFIG_SMP */
1047 struct sched_domain_attr
;
1050 partition_sched_domains(int ndoms_new
, cpumask_var_t doms_new
[],
1051 struct sched_domain_attr
*dattr_new
)
1054 #endif /* !CONFIG_SMP */
1057 struct io_context
; /* See blkdev.h */
1060 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
1061 extern void prefetch_stack(struct task_struct
*t
);
1063 static inline void prefetch_stack(struct task_struct
*t
) { }
1066 struct audit_context
; /* See audit.c */
1068 struct pipe_inode_info
;
1069 struct uts_namespace
;
1072 struct sched_domain
;
1077 #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
1078 #define WF_FORK 0x02 /* child wakeup after fork */
1080 struct sched_class
{
1081 const struct sched_class
*next
;
1083 void (*enqueue_task
) (struct rq
*rq
, struct task_struct
*p
, int wakeup
);
1084 void (*dequeue_task
) (struct rq
*rq
, struct task_struct
*p
, int sleep
);
1085 void (*yield_task
) (struct rq
*rq
);
1087 void (*check_preempt_curr
) (struct rq
*rq
, struct task_struct
*p
, int flags
);
1089 struct task_struct
* (*pick_next_task
) (struct rq
*rq
);
1090 void (*put_prev_task
) (struct rq
*rq
, struct task_struct
*p
);
1093 int (*select_task_rq
)(struct task_struct
*p
, int sd_flag
, int flags
);
1095 unsigned long (*load_balance
) (struct rq
*this_rq
, int this_cpu
,
1096 struct rq
*busiest
, unsigned long max_load_move
,
1097 struct sched_domain
*sd
, enum cpu_idle_type idle
,
1098 int *all_pinned
, int *this_best_prio
);
1100 int (*move_one_task
) (struct rq
*this_rq
, int this_cpu
,
1101 struct rq
*busiest
, struct sched_domain
*sd
,
1102 enum cpu_idle_type idle
);
1103 void (*pre_schedule
) (struct rq
*this_rq
, struct task_struct
*task
);
1104 void (*post_schedule
) (struct rq
*this_rq
);
1105 void (*task_waking
) (struct rq
*this_rq
, struct task_struct
*task
);
1106 void (*task_woken
) (struct rq
*this_rq
, struct task_struct
*task
);
1108 void (*set_cpus_allowed
)(struct task_struct
*p
,
1109 const struct cpumask
*newmask
);
1111 void (*rq_online
)(struct rq
*rq
);
1112 void (*rq_offline
)(struct rq
*rq
);
1115 void (*set_curr_task
) (struct rq
*rq
);
1116 void (*task_tick
) (struct rq
*rq
, struct task_struct
*p
, int queued
);
1117 void (*task_fork
) (struct task_struct
*p
);
1119 void (*switched_from
) (struct rq
*this_rq
, struct task_struct
*task
,
1121 void (*switched_to
) (struct rq
*this_rq
, struct task_struct
*task
,
1123 void (*prio_changed
) (struct rq
*this_rq
, struct task_struct
*task
,
1124 int oldprio
, int running
);
1126 unsigned int (*get_rr_interval
) (struct rq
*rq
,
1127 struct task_struct
*task
);
1129 #ifdef CONFIG_FAIR_GROUP_SCHED
1130 void (*moved_group
) (struct task_struct
*p
, int on_rq
);
1134 struct load_weight
{
1135 unsigned long weight
, inv_weight
;
1139 * CFS stats for a schedulable entity (task, task-group etc)
1141 * Current field usage histogram:
1148 struct sched_entity
{
1149 struct load_weight load
; /* for load-balancing */
1150 struct rb_node run_node
;
1151 struct list_head group_node
;
1155 u64 sum_exec_runtime
;
1157 u64 prev_sum_exec_runtime
;
1167 #ifdef CONFIG_SCHEDSTATS
1177 s64 sum_sleep_runtime
;
1184 u64 nr_migrations_cold
;
1185 u64 nr_failed_migrations_affine
;
1186 u64 nr_failed_migrations_running
;
1187 u64 nr_failed_migrations_hot
;
1188 u64 nr_forced_migrations
;
1191 u64 nr_wakeups_sync
;
1192 u64 nr_wakeups_migrate
;
1193 u64 nr_wakeups_local
;
1194 u64 nr_wakeups_remote
;
1195 u64 nr_wakeups_affine
;
1196 u64 nr_wakeups_affine_attempts
;
1197 u64 nr_wakeups_passive
;
1198 u64 nr_wakeups_idle
;
1201 #ifdef CONFIG_FAIR_GROUP_SCHED
1202 struct sched_entity
*parent
;
1203 /* rq on which this entity is (to be) queued: */
1204 struct cfs_rq
*cfs_rq
;
1205 /* rq "owned" by this entity/group: */
1206 struct cfs_rq
*my_q
;
1210 struct sched_rt_entity
{
1211 struct list_head run_list
;
1212 unsigned long timeout
;
1213 unsigned int time_slice
;
1214 int nr_cpus_allowed
;
1216 struct sched_rt_entity
*back
;
1217 #ifdef CONFIG_RT_GROUP_SCHED
1218 struct sched_rt_entity
*parent
;
1219 /* rq on which this entity is (to be) queued: */
1220 struct rt_rq
*rt_rq
;
1221 /* rq "owned" by this entity/group: */
1228 struct task_struct
{
1229 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
1232 unsigned int flags
; /* per process flags, defined below */
1233 unsigned int ptrace
;
1235 int lock_depth
; /* BKL lock depth */
1238 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1243 int prio
, static_prio
, normal_prio
;
1244 unsigned int rt_priority
;
1245 const struct sched_class
*sched_class
;
1246 struct sched_entity se
;
1247 struct sched_rt_entity rt
;
1249 #ifdef CONFIG_PREEMPT_NOTIFIERS
1250 /* list of struct preempt_notifier: */
1251 struct hlist_head preempt_notifiers
;
1255 * fpu_counter contains the number of consecutive context switches
1256 * that the FPU is used. If this is over a threshold, the lazy fpu
1257 * saving becomes unlazy to save the trap. This is an unsigned char
1258 * so that after 256 times the counter wraps and the behavior turns
1259 * lazy again; this to deal with bursty apps that only use FPU for
1262 unsigned char fpu_counter
;
1263 #ifdef CONFIG_BLK_DEV_IO_TRACE
1264 unsigned int btrace_seq
;
1267 unsigned int policy
;
1268 cpumask_t cpus_allowed
;
1270 #ifdef CONFIG_TREE_PREEMPT_RCU
1271 int rcu_read_lock_nesting
;
1272 char rcu_read_unlock_special
;
1273 struct rcu_node
*rcu_blocked_node
;
1274 struct list_head rcu_node_entry
;
1275 #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1277 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1278 struct sched_info sched_info
;
1281 struct list_head tasks
;
1282 struct plist_node pushable_tasks
;
1284 struct mm_struct
*mm
, *active_mm
;
1288 int exit_code
, exit_signal
;
1289 int pdeath_signal
; /* The signal sent when the parent dies */
1291 unsigned int personality
;
1292 unsigned did_exec
:1;
1293 unsigned in_execve
:1; /* Tell the LSMs that the process is doing an
1295 unsigned in_iowait
:1;
1298 /* Revert to default priority/policy when forking */
1299 unsigned sched_reset_on_fork
:1;
1304 #ifdef CONFIG_CC_STACKPROTECTOR
1305 /* Canary value for the -fstack-protector gcc feature */
1306 unsigned long stack_canary
;
1310 * pointers to (original) parent process, youngest child, younger sibling,
1311 * older sibling, respectively. (p->father can be replaced with
1312 * p->real_parent->pid)
1314 struct task_struct
*real_parent
; /* real parent process */
1315 struct task_struct
*parent
; /* recipient of SIGCHLD, wait4() reports */
1317 * children/sibling forms the list of my natural children
1319 struct list_head children
; /* list of my children */
1320 struct list_head sibling
; /* linkage in my parent's children list */
1321 struct task_struct
*group_leader
; /* threadgroup leader */
1324 * ptraced is the list of tasks this task is using ptrace on.
1325 * This includes both natural children and PTRACE_ATTACH targets.
1326 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1328 struct list_head ptraced
;
1329 struct list_head ptrace_entry
;
1332 * This is the tracer handle for the ptrace BTS extension.
1333 * This field actually belongs to the ptracer task.
1335 struct bts_context
*bts
;
1337 /* PID/PID hash table linkage. */
1338 struct pid_link pids
[PIDTYPE_MAX
];
1339 struct list_head thread_group
;
1341 struct completion
*vfork_done
; /* for vfork() */
1342 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
1343 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
1345 cputime_t utime
, stime
, utimescaled
, stimescaled
;
1347 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
1348 cputime_t prev_utime
, prev_stime
;
1350 unsigned long nvcsw
, nivcsw
; /* context switch counts */
1351 struct timespec start_time
; /* monotonic time */
1352 struct timespec real_start_time
; /* boot based time */
1353 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1354 unsigned long min_flt
, maj_flt
;
1356 struct task_cputime cputime_expires
;
1357 struct list_head cpu_timers
[3];
1359 /* process credentials */
1360 const struct cred
*real_cred
; /* objective and real subjective task
1361 * credentials (COW) */
1362 const struct cred
*cred
; /* effective (overridable) subjective task
1363 * credentials (COW) */
1364 struct mutex cred_guard_mutex
; /* guard against foreign influences on
1365 * credential calculations
1366 * (notably. ptrace) */
1367 struct cred
*replacement_session_keyring
; /* for KEYCTL_SESSION_TO_PARENT */
1369 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
1370 - access with [gs]et_task_comm (which lock
1371 it with task_lock())
1372 - initialized normally by setup_new_exec */
1373 /* file system info */
1374 int link_count
, total_link_count
;
1375 #ifdef CONFIG_SYSVIPC
1377 struct sysv_sem sysvsem
;
1379 #ifdef CONFIG_DETECT_HUNG_TASK
1380 /* hung task detection */
1381 unsigned long last_switch_count
;
1383 /* CPU-specific state of this task */
1384 struct thread_struct thread
;
1385 /* filesystem information */
1386 struct fs_struct
*fs
;
1387 /* open file information */
1388 struct files_struct
*files
;
1390 struct nsproxy
*nsproxy
;
1391 /* signal handlers */
1392 struct signal_struct
*signal
;
1393 struct sighand_struct
*sighand
;
1395 sigset_t blocked
, real_blocked
;
1396 sigset_t saved_sigmask
; /* restored if set_restore_sigmask() was used */
1397 struct sigpending pending
;
1399 unsigned long sas_ss_sp
;
1401 int (*notifier
)(void *priv
);
1402 void *notifier_data
;
1403 sigset_t
*notifier_mask
;
1404 struct audit_context
*audit_context
;
1405 #ifdef CONFIG_AUDITSYSCALL
1407 unsigned int sessionid
;
1411 /* Thread group tracking */
1414 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1416 spinlock_t alloc_lock
;
1418 #ifdef CONFIG_GENERIC_HARDIRQS
1419 /* IRQ handler threads */
1420 struct irqaction
*irqaction
;
1423 /* Protection of the PI data structures: */
1424 raw_spinlock_t pi_lock
;
1426 #ifdef CONFIG_RT_MUTEXES
1427 /* PI waiters blocked on a rt_mutex held by this task */
1428 struct plist_head pi_waiters
;
1429 /* Deadlock detection and priority inheritance handling */
1430 struct rt_mutex_waiter
*pi_blocked_on
;
1433 #ifdef CONFIG_DEBUG_MUTEXES
1434 /* mutex deadlock detection */
1435 struct mutex_waiter
*blocked_on
;
1437 #ifdef CONFIG_TRACE_IRQFLAGS
1438 unsigned int irq_events
;
1439 unsigned long hardirq_enable_ip
;
1440 unsigned long hardirq_disable_ip
;
1441 unsigned int hardirq_enable_event
;
1442 unsigned int hardirq_disable_event
;
1443 int hardirqs_enabled
;
1444 int hardirq_context
;
1445 unsigned long softirq_disable_ip
;
1446 unsigned long softirq_enable_ip
;
1447 unsigned int softirq_disable_event
;
1448 unsigned int softirq_enable_event
;
1449 int softirqs_enabled
;
1450 int softirq_context
;
1452 #ifdef CONFIG_LOCKDEP
1453 # define MAX_LOCK_DEPTH 48UL
1456 unsigned int lockdep_recursion
;
1457 struct held_lock held_locks
[MAX_LOCK_DEPTH
];
1458 gfp_t lockdep_reclaim_gfp
;
1461 /* journalling filesystem info */
1464 /* stacked block device info */
1465 struct bio
*bio_list
, **bio_tail
;
1468 struct reclaim_state
*reclaim_state
;
1470 struct backing_dev_info
*backing_dev_info
;
1472 struct io_context
*io_context
;
1474 unsigned long ptrace_message
;
1475 siginfo_t
*last_siginfo
; /* For ptrace use. */
1476 struct task_io_accounting ioac
;
1477 #if defined(CONFIG_TASK_XACCT)
1478 u64 acct_rss_mem1
; /* accumulated rss usage */
1479 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
1480 cputime_t acct_timexpd
; /* stime + utime since last update */
1482 #ifdef CONFIG_CPUSETS
1483 nodemask_t mems_allowed
; /* Protected by alloc_lock */
1484 int cpuset_mem_spread_rotor
;
1486 #ifdef CONFIG_CGROUPS
1487 /* Control Group info protected by css_set_lock */
1488 struct css_set
*cgroups
;
1489 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1490 struct list_head cg_list
;
1493 struct robust_list_head __user
*robust_list
;
1494 #ifdef CONFIG_COMPAT
1495 struct compat_robust_list_head __user
*compat_robust_list
;
1497 struct list_head pi_state_list
;
1498 struct futex_pi_state
*pi_state_cache
;
1500 #ifdef CONFIG_PERF_EVENTS
1501 struct perf_event_context
*perf_event_ctxp
;
1502 struct mutex perf_event_mutex
;
1503 struct list_head perf_event_list
;
1506 struct mempolicy
*mempolicy
; /* Protected by alloc_lock */
1509 atomic_t fs_excl
; /* holding fs exclusive resources */
1510 struct rcu_head rcu
;
1513 * cache last used pipe for splice
1515 struct pipe_inode_info
*splice_pipe
;
1516 #ifdef CONFIG_TASK_DELAY_ACCT
1517 struct task_delay_info
*delays
;
1519 #ifdef CONFIG_FAULT_INJECTION
1522 struct prop_local_single dirties
;
1523 #ifdef CONFIG_LATENCYTOP
1524 int latency_record_count
;
1525 struct latency_record latency_record
[LT_SAVECOUNT
];
1528 * time slack values; these are used to round up poll() and
1529 * select() etc timeout values. These are in nanoseconds.
1531 unsigned long timer_slack_ns
;
1532 unsigned long default_timer_slack_ns
;
1534 struct list_head
*scm_work_list
;
1535 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1536 /* Index of current stored address in ret_stack */
1538 /* Stack of return addresses for return function tracing */
1539 struct ftrace_ret_stack
*ret_stack
;
1540 /* time stamp for last schedule */
1541 unsigned long long ftrace_timestamp
;
1543 * Number of functions that haven't been traced
1544 * because of depth overrun.
1546 atomic_t trace_overrun
;
1547 /* Pause for the tracing */
1548 atomic_t tracing_graph_pause
;
1550 #ifdef CONFIG_TRACING
1551 /* state flags for use by tracers */
1552 unsigned long trace
;
1553 /* bitmask of trace recursion */
1554 unsigned long trace_recursion
;
1555 #endif /* CONFIG_TRACING */
1556 unsigned long stack_start
;
1557 #ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1558 struct memcg_batch_info
{
1559 int do_batch
; /* incremented when batch uncharge started */
1560 struct mem_cgroup
*memcg
; /* target memcg of uncharge */
1561 unsigned long bytes
; /* uncharged usage */
1562 unsigned long memsw_bytes
; /* uncharged mem+swap usage */
1567 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1568 #define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
1571 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1572 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1573 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1574 * values are inverted: lower p->prio value means higher priority.
1576 * The MAX_USER_RT_PRIO value allows the actual maximum
1577 * RT priority to be separate from the value exported to
1578 * user-space. This allows kernel threads to set their
1579 * priority to a value higher than any user task. Note:
1580 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1583 #define MAX_USER_RT_PRIO 100
1584 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1586 #define MAX_PRIO (MAX_RT_PRIO + 40)
1587 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1589 static inline int rt_prio(int prio
)
1591 if (unlikely(prio
< MAX_RT_PRIO
))
1596 static inline int rt_task(struct task_struct
*p
)
1598 return rt_prio(p
->prio
);
1601 static inline struct pid
*task_pid(struct task_struct
*task
)
1603 return task
->pids
[PIDTYPE_PID
].pid
;
1606 static inline struct pid
*task_tgid(struct task_struct
*task
)
1608 return task
->group_leader
->pids
[PIDTYPE_PID
].pid
;
1612 * Without tasklist or rcu lock it is not safe to dereference
1613 * the result of task_pgrp/task_session even if task == current,
1614 * we can race with another thread doing sys_setsid/sys_setpgid.
1616 static inline struct pid
*task_pgrp(struct task_struct
*task
)
1618 return task
->group_leader
->pids
[PIDTYPE_PGID
].pid
;
1621 static inline struct pid
*task_session(struct task_struct
*task
)
1623 return task
->group_leader
->pids
[PIDTYPE_SID
].pid
;
1626 struct pid_namespace
;
1629 * the helpers to get the task's different pids as they are seen
1630 * from various namespaces
1632 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1633 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1635 * task_xid_nr_ns() : id seen from the ns specified;
1637 * set_task_vxid() : assigns a virtual id to a task;
1639 * see also pid_nr() etc in include/linux/pid.h
1641 pid_t
__task_pid_nr_ns(struct task_struct
*task
, enum pid_type type
,
1642 struct pid_namespace
*ns
);
1644 static inline pid_t
task_pid_nr(struct task_struct
*tsk
)
1649 static inline pid_t
task_pid_nr_ns(struct task_struct
*tsk
,
1650 struct pid_namespace
*ns
)
1652 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, ns
);
1655 static inline pid_t
task_pid_vnr(struct task_struct
*tsk
)
1657 return __task_pid_nr_ns(tsk
, PIDTYPE_PID
, NULL
);
1661 static inline pid_t
task_tgid_nr(struct task_struct
*tsk
)
1666 pid_t
task_tgid_nr_ns(struct task_struct
*tsk
, struct pid_namespace
*ns
);
1668 static inline pid_t
task_tgid_vnr(struct task_struct
*tsk
)
1670 return pid_vnr(task_tgid(tsk
));
1674 static inline pid_t
task_pgrp_nr_ns(struct task_struct
*tsk
,
1675 struct pid_namespace
*ns
)
1677 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, ns
);
1680 static inline pid_t
task_pgrp_vnr(struct task_struct
*tsk
)
1682 return __task_pid_nr_ns(tsk
, PIDTYPE_PGID
, NULL
);
1686 static inline pid_t
task_session_nr_ns(struct task_struct
*tsk
,
1687 struct pid_namespace
*ns
)
1689 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, ns
);
1692 static inline pid_t
task_session_vnr(struct task_struct
*tsk
)
1694 return __task_pid_nr_ns(tsk
, PIDTYPE_SID
, NULL
);
1697 /* obsolete, do not use */
1698 static inline pid_t
task_pgrp_nr(struct task_struct
*tsk
)
1700 return task_pgrp_nr_ns(tsk
, &init_pid_ns
);
1704 * pid_alive - check that a task structure is not stale
1705 * @p: Task structure to be checked.
1707 * Test if a process is not yet dead (at most zombie state)
1708 * If pid_alive fails, then pointers within the task structure
1709 * can be stale and must not be dereferenced.
1711 static inline int pid_alive(struct task_struct
*p
)
1713 return p
->pids
[PIDTYPE_PID
].pid
!= NULL
;
1717 * is_global_init - check if a task structure is init
1718 * @tsk: Task structure to be checked.
1720 * Check if a task structure is the first user space task the kernel created.
1722 static inline int is_global_init(struct task_struct
*tsk
)
1724 return tsk
->pid
== 1;
1728 * is_container_init:
1729 * check whether in the task is init in its own pid namespace.
1731 extern int is_container_init(struct task_struct
*tsk
);
1733 extern struct pid
*cad_pid
;
1735 extern void free_task(struct task_struct
*tsk
);
1736 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1738 extern void __put_task_struct(struct task_struct
*t
);
1740 static inline void put_task_struct(struct task_struct
*t
)
1742 if (atomic_dec_and_test(&t
->usage
))
1743 __put_task_struct(t
);
1746 extern void task_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1747 extern void thread_group_times(struct task_struct
*p
, cputime_t
*ut
, cputime_t
*st
);
1752 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1753 /* Not implemented yet, only for 486*/
1754 #define PF_STARTING 0x00000002 /* being created */
1755 #define PF_EXITING 0x00000004 /* getting shut down */
1756 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1757 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1758 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1759 #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1760 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1761 #define PF_DUMPCORE 0x00000200 /* dumped core */
1762 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1763 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1764 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1765 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1766 #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1767 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1768 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1769 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1770 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1771 #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1772 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1773 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1774 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1775 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1776 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1777 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1778 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1779 #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
1780 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1781 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1782 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1783 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1786 * Only the _current_ task can read/write to tsk->flags, but other
1787 * tasks can access tsk->flags in readonly mode for example
1788 * with tsk_used_math (like during threaded core dumping).
1789 * There is however an exception to this rule during ptrace
1790 * or during fork: the ptracer task is allowed to write to the
1791 * child->flags of its traced child (same goes for fork, the parent
1792 * can write to the child->flags), because we're guaranteed the
1793 * child is not running and in turn not changing child->flags
1794 * at the same time the parent does it.
1796 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1797 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1798 #define clear_used_math() clear_stopped_child_used_math(current)
1799 #define set_used_math() set_stopped_child_used_math(current)
1800 #define conditional_stopped_child_used_math(condition, child) \
1801 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1802 #define conditional_used_math(condition) \
1803 conditional_stopped_child_used_math(condition, current)
1804 #define copy_to_stopped_child_used_math(child) \
1805 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1806 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1807 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1808 #define used_math() tsk_used_math(current)
1810 #ifdef CONFIG_TREE_PREEMPT_RCU
1812 #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1813 #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1815 static inline void rcu_copy_process(struct task_struct
*p
)
1817 p
->rcu_read_lock_nesting
= 0;
1818 p
->rcu_read_unlock_special
= 0;
1819 p
->rcu_blocked_node
= NULL
;
1820 INIT_LIST_HEAD(&p
->rcu_node_entry
);
1825 static inline void rcu_copy_process(struct task_struct
*p
)
1832 extern int set_cpus_allowed_ptr(struct task_struct
*p
,
1833 const struct cpumask
*new_mask
);
1835 static inline int set_cpus_allowed_ptr(struct task_struct
*p
,
1836 const struct cpumask
*new_mask
)
1838 if (!cpumask_test_cpu(0, new_mask
))
1844 #ifndef CONFIG_CPUMASK_OFFSTACK
1845 static inline int set_cpus_allowed(struct task_struct
*p
, cpumask_t new_mask
)
1847 return set_cpus_allowed_ptr(p
, &new_mask
);
1852 * Architectures can set this to 1 if they have specified
1853 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1854 * but then during bootup it turns out that sched_clock()
1855 * is reliable after all:
1857 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1858 extern int sched_clock_stable
;
1861 /* ftrace calls sched_clock() directly */
1862 extern unsigned long long notrace
sched_clock(void);
1864 extern void sched_clock_init(void);
1865 extern u64
sched_clock_cpu(int cpu
);
1867 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1868 static inline void sched_clock_tick(void)
1872 static inline void sched_clock_idle_sleep_event(void)
1876 static inline void sched_clock_idle_wakeup_event(u64 delta_ns
)
1880 extern void sched_clock_tick(void);
1881 extern void sched_clock_idle_sleep_event(void);
1882 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1886 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1887 * clock constructed from sched_clock():
1889 extern unsigned long long cpu_clock(int cpu
);
1891 extern unsigned long long
1892 task_sched_runtime(struct task_struct
*task
);
1893 extern unsigned long long thread_group_sched_runtime(struct task_struct
*task
);
1895 /* sched_exec is called by processes performing an exec */
1897 extern void sched_exec(void);
1899 #define sched_exec() {}
1902 extern void sched_clock_idle_sleep_event(void);
1903 extern void sched_clock_idle_wakeup_event(u64 delta_ns
);
1905 #ifdef CONFIG_HOTPLUG_CPU
1906 extern void idle_task_exit(void);
1908 static inline void idle_task_exit(void) {}
1911 extern void sched_idle_next(void);
1913 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1914 extern void wake_up_idle_cpu(int cpu
);
1916 static inline void wake_up_idle_cpu(int cpu
) { }
1919 extern unsigned int sysctl_sched_latency
;
1920 extern unsigned int sysctl_sched_min_granularity
;
1921 extern unsigned int sysctl_sched_wakeup_granularity
;
1922 extern unsigned int sysctl_sched_shares_ratelimit
;
1923 extern unsigned int sysctl_sched_shares_thresh
;
1924 extern unsigned int sysctl_sched_child_runs_first
;
1926 enum sched_tunable_scaling
{
1927 SCHED_TUNABLESCALING_NONE
,
1928 SCHED_TUNABLESCALING_LOG
,
1929 SCHED_TUNABLESCALING_LINEAR
,
1930 SCHED_TUNABLESCALING_END
,
1932 extern enum sched_tunable_scaling sysctl_sched_tunable_scaling
;
1934 #ifdef CONFIG_SCHED_DEBUG
1935 extern unsigned int sysctl_sched_migration_cost
;
1936 extern unsigned int sysctl_sched_nr_migrate
;
1937 extern unsigned int sysctl_sched_time_avg
;
1938 extern unsigned int sysctl_timer_migration
;
1940 int sched_proc_update_handler(struct ctl_table
*table
, int write
,
1941 void __user
*buffer
, size_t *length
,
1944 #ifdef CONFIG_SCHED_DEBUG
1945 static inline unsigned int get_sysctl_timer_migration(void)
1947 return sysctl_timer_migration
;
1950 static inline unsigned int get_sysctl_timer_migration(void)
1955 extern unsigned int sysctl_sched_rt_period
;
1956 extern int sysctl_sched_rt_runtime
;
1958 int sched_rt_handler(struct ctl_table
*table
, int write
,
1959 void __user
*buffer
, size_t *lenp
,
1962 extern unsigned int sysctl_sched_compat_yield
;
1964 #ifdef CONFIG_RT_MUTEXES
1965 extern int rt_mutex_getprio(struct task_struct
*p
);
1966 extern void rt_mutex_setprio(struct task_struct
*p
, int prio
);
1967 extern void rt_mutex_adjust_pi(struct task_struct
*p
);
1969 static inline int rt_mutex_getprio(struct task_struct
*p
)
1971 return p
->normal_prio
;
1973 # define rt_mutex_adjust_pi(p) do { } while (0)
1976 extern void set_user_nice(struct task_struct
*p
, long nice
);
1977 extern int task_prio(const struct task_struct
*p
);
1978 extern int task_nice(const struct task_struct
*p
);
1979 extern int can_nice(const struct task_struct
*p
, const int nice
);
1980 extern int task_curr(const struct task_struct
*p
);
1981 extern int idle_cpu(int cpu
);
1982 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
1983 extern int sched_setscheduler_nocheck(struct task_struct
*, int,
1984 struct sched_param
*);
1985 extern struct task_struct
*idle_task(int cpu
);
1986 extern struct task_struct
*curr_task(int cpu
);
1987 extern void set_curr_task(int cpu
, struct task_struct
*p
);
1992 * The default (Linux) execution domain.
1994 extern struct exec_domain default_exec_domain
;
1996 union thread_union
{
1997 struct thread_info thread_info
;
1998 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
2001 #ifndef __HAVE_ARCH_KSTACK_END
2002 static inline int kstack_end(void *addr
)
2004 /* Reliable end of stack detection:
2005 * Some APM bios versions misalign the stack
2007 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
2011 extern union thread_union init_thread_union
;
2012 extern struct task_struct init_task
;
2014 extern struct mm_struct init_mm
;
2016 extern struct pid_namespace init_pid_ns
;
2019 * find a task by one of its numerical ids
2021 * find_task_by_pid_ns():
2022 * finds a task by its pid in the specified namespace
2023 * find_task_by_vpid():
2024 * finds a task by its virtual pid
2026 * see also find_vpid() etc in include/linux/pid.h
2029 extern struct task_struct
*find_task_by_vpid(pid_t nr
);
2030 extern struct task_struct
*find_task_by_pid_ns(pid_t nr
,
2031 struct pid_namespace
*ns
);
2033 extern void __set_special_pids(struct pid
*pid
);
2035 /* per-UID process charging. */
2036 extern struct user_struct
* alloc_uid(struct user_namespace
*, uid_t
);
2037 static inline struct user_struct
*get_uid(struct user_struct
*u
)
2039 atomic_inc(&u
->__count
);
2042 extern void free_uid(struct user_struct
*);
2043 extern void release_uids(struct user_namespace
*ns
);
2045 #include <asm/current.h>
2047 extern void do_timer(unsigned long ticks
);
2049 extern int wake_up_state(struct task_struct
*tsk
, unsigned int state
);
2050 extern int wake_up_process(struct task_struct
*tsk
);
2051 extern void wake_up_new_task(struct task_struct
*tsk
,
2052 unsigned long clone_flags
);
2054 extern void kick_process(struct task_struct
*tsk
);
2056 static inline void kick_process(struct task_struct
*tsk
) { }
2058 extern void sched_fork(struct task_struct
*p
, int clone_flags
);
2059 extern void sched_dead(struct task_struct
*p
);
2061 extern void proc_caches_init(void);
2062 extern void flush_signals(struct task_struct
*);
2063 extern void __flush_signals(struct task_struct
*);
2064 extern void ignore_signals(struct task_struct
*);
2065 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
2066 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
2068 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
2070 unsigned long flags
;
2073 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
2074 ret
= dequeue_signal(tsk
, mask
, info
);
2075 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
2080 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
2082 extern void unblock_all_signals(void);
2083 extern void release_task(struct task_struct
* p
);
2084 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
2085 extern int force_sigsegv(int, struct task_struct
*);
2086 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
2087 extern int __kill_pgrp_info(int sig
, struct siginfo
*info
, struct pid
*pgrp
);
2088 extern int kill_pid_info(int sig
, struct siginfo
*info
, struct pid
*pid
);
2089 extern int kill_pid_info_as_uid(int, struct siginfo
*, struct pid
*, uid_t
, uid_t
, u32
);
2090 extern int kill_pgrp(struct pid
*pid
, int sig
, int priv
);
2091 extern int kill_pid(struct pid
*pid
, int sig
, int priv
);
2092 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
2093 extern int do_notify_parent(struct task_struct
*, int);
2094 extern void __wake_up_parent(struct task_struct
*p
, struct task_struct
*parent
);
2095 extern void force_sig(int, struct task_struct
*);
2096 extern int send_sig(int, struct task_struct
*, int);
2097 extern void zap_other_threads(struct task_struct
*p
);
2098 extern struct sigqueue
*sigqueue_alloc(void);
2099 extern void sigqueue_free(struct sigqueue
*);
2100 extern int send_sigqueue(struct sigqueue
*, struct task_struct
*, int group
);
2101 extern int do_sigaction(int, struct k_sigaction
*, struct k_sigaction
*);
2102 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
2104 static inline int kill_cad_pid(int sig
, int priv
)
2106 return kill_pid(cad_pid
, sig
, priv
);
2109 /* These can be the second arg to send_sig_info/send_group_sig_info. */
2110 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
2111 #define SEND_SIG_PRIV ((struct siginfo *) 1)
2112 #define SEND_SIG_FORCED ((struct siginfo *) 2)
2115 * True if we are on the alternate signal stack.
2117 static inline int on_sig_stack(unsigned long sp
)
2119 #ifdef CONFIG_STACK_GROWSUP
2120 return sp
>= current
->sas_ss_sp
&&
2121 sp
- current
->sas_ss_sp
< current
->sas_ss_size
;
2123 return sp
> current
->sas_ss_sp
&&
2124 sp
- current
->sas_ss_sp
<= current
->sas_ss_size
;
2128 static inline int sas_ss_flags(unsigned long sp
)
2130 return (current
->sas_ss_size
== 0 ? SS_DISABLE
2131 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
2135 * Routines for handling mm_structs
2137 extern struct mm_struct
* mm_alloc(void);
2139 /* mmdrop drops the mm and the page tables */
2140 extern void __mmdrop(struct mm_struct
*);
2141 static inline void mmdrop(struct mm_struct
* mm
)
2143 if (unlikely(atomic_dec_and_test(&mm
->mm_count
)))
2147 /* mmput gets rid of the mappings and all user-space */
2148 extern void mmput(struct mm_struct
*);
2149 /* Grab a reference to a task's mm, if it is not already going away */
2150 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
2151 /* Remove the current tasks stale references to the old mm_struct */
2152 extern void mm_release(struct task_struct
*, struct mm_struct
*);
2153 /* Allocate a new mm structure and copy contents from tsk->mm */
2154 extern struct mm_struct
*dup_mm(struct task_struct
*tsk
);
2156 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2157 struct task_struct
*, struct pt_regs
*);
2158 extern void flush_thread(void);
2159 extern void exit_thread(void);
2161 extern void exit_files(struct task_struct
*);
2162 extern void __cleanup_signal(struct signal_struct
*);
2163 extern void __cleanup_sighand(struct sighand_struct
*);
2165 extern void exit_itimers(struct signal_struct
*);
2166 extern void flush_itimer_signals(void);
2168 extern NORET_TYPE
void do_group_exit(int);
2170 extern void daemonize(const char *, ...);
2171 extern int allow_signal(int);
2172 extern int disallow_signal(int);
2174 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
2175 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
2176 struct task_struct
*fork_idle(int);
2178 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
2179 extern char *get_task_comm(char *to
, struct task_struct
*tsk
);
2182 extern void wait_task_context_switch(struct task_struct
*p
);
2183 extern unsigned long wait_task_inactive(struct task_struct
*, long match_state
);
2185 static inline void wait_task_context_switch(struct task_struct
*p
) {}
2186 static inline unsigned long wait_task_inactive(struct task_struct
*p
,
2193 #define next_task(p) \
2194 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2196 #define for_each_process(p) \
2197 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2199 extern bool current_is_single_threaded(void);
2202 * Careful: do_each_thread/while_each_thread is a double loop so
2203 * 'break' will not work as expected - use goto instead.
2205 #define do_each_thread(g, t) \
2206 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2208 #define while_each_thread(g, t) \
2209 while ((t = next_thread(t)) != g)
2211 /* de_thread depends on thread_group_leader not being a pid based check */
2212 #define thread_group_leader(p) (p == p->group_leader)
2214 /* Do to the insanities of de_thread it is possible for a process
2215 * to have the pid of the thread group leader without actually being
2216 * the thread group leader. For iteration through the pids in proc
2217 * all we care about is that we have a task with the appropriate
2218 * pid, we don't actually care if we have the right task.
2220 static inline int has_group_leader_pid(struct task_struct
*p
)
2222 return p
->pid
== p
->tgid
;
2226 int same_thread_group(struct task_struct
*p1
, struct task_struct
*p2
)
2228 return p1
->tgid
== p2
->tgid
;
2231 static inline struct task_struct
*next_thread(const struct task_struct
*p
)
2233 return list_entry_rcu(p
->thread_group
.next
,
2234 struct task_struct
, thread_group
);
2237 static inline int thread_group_empty(struct task_struct
*p
)
2239 return list_empty(&p
->thread_group
);
2242 #define delay_group_leader(p) \
2243 (thread_group_leader(p) && !thread_group_empty(p))
2245 static inline int task_detached(struct task_struct
*p
)
2247 return p
->exit_signal
== -1;
2251 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2252 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2253 * pins the final release of task.io_context. Also protects ->cpuset and
2254 * ->cgroup.subsys[].
2256 * Nests both inside and outside of read_lock(&tasklist_lock).
2257 * It must not be nested with write_lock_irq(&tasklist_lock),
2258 * neither inside nor outside.
2260 static inline void task_lock(struct task_struct
*p
)
2262 spin_lock(&p
->alloc_lock
);
2265 static inline void task_unlock(struct task_struct
*p
)
2267 spin_unlock(&p
->alloc_lock
);
2270 extern struct sighand_struct
*lock_task_sighand(struct task_struct
*tsk
,
2271 unsigned long *flags
);
2273 static inline void unlock_task_sighand(struct task_struct
*tsk
,
2274 unsigned long *flags
)
2276 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, *flags
);
2279 #ifndef __HAVE_THREAD_FUNCTIONS
2281 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2282 #define task_stack_page(task) ((task)->stack)
2284 static inline void setup_thread_stack(struct task_struct
*p
, struct task_struct
*org
)
2286 *task_thread_info(p
) = *task_thread_info(org
);
2287 task_thread_info(p
)->task
= p
;
2290 static inline unsigned long *end_of_stack(struct task_struct
*p
)
2292 return (unsigned long *)(task_thread_info(p
) + 1);
2297 static inline int object_is_on_stack(void *obj
)
2299 void *stack
= task_stack_page(current
);
2301 return (obj
>= stack
) && (obj
< (stack
+ THREAD_SIZE
));
2304 extern void thread_info_cache_init(void);
2306 #ifdef CONFIG_DEBUG_STACK_USAGE
2307 static inline unsigned long stack_not_used(struct task_struct
*p
)
2309 unsigned long *n
= end_of_stack(p
);
2311 do { /* Skip over canary */
2315 return (unsigned long)n
- (unsigned long)end_of_stack(p
);
2319 /* set thread flags in other task's structures
2320 * - see asm/thread_info.h for TIF_xxxx flags available
2322 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2324 set_ti_thread_flag(task_thread_info(tsk
), flag
);
2327 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2329 clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2332 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2334 return test_and_set_ti_thread_flag(task_thread_info(tsk
), flag
);
2337 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2339 return test_and_clear_ti_thread_flag(task_thread_info(tsk
), flag
);
2342 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
2344 return test_ti_thread_flag(task_thread_info(tsk
), flag
);
2347 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
2349 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2352 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
2354 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
2357 static inline int test_tsk_need_resched(struct task_struct
*tsk
)
2359 return unlikely(test_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
));
2362 static inline int restart_syscall(void)
2364 set_tsk_thread_flag(current
, TIF_SIGPENDING
);
2365 return -ERESTARTNOINTR
;
2368 static inline int signal_pending(struct task_struct
*p
)
2370 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
2373 static inline int __fatal_signal_pending(struct task_struct
*p
)
2375 return unlikely(sigismember(&p
->pending
.signal
, SIGKILL
));
2378 static inline int fatal_signal_pending(struct task_struct
*p
)
2380 return signal_pending(p
) && __fatal_signal_pending(p
);
2383 static inline int signal_pending_state(long state
, struct task_struct
*p
)
2385 if (!(state
& (TASK_INTERRUPTIBLE
| TASK_WAKEKILL
)))
2387 if (!signal_pending(p
))
2390 return (state
& TASK_INTERRUPTIBLE
) || __fatal_signal_pending(p
);
2393 static inline int need_resched(void)
2395 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
2399 * cond_resched() and cond_resched_lock(): latency reduction via
2400 * explicit rescheduling in places that are safe. The return
2401 * value indicates whether a reschedule was done in fact.
2402 * cond_resched_lock() will drop the spinlock before scheduling,
2403 * cond_resched_softirq() will enable bhs before scheduling.
2405 extern int _cond_resched(void);
2407 #define cond_resched() ({ \
2408 __might_sleep(__FILE__, __LINE__, 0); \
2412 extern int __cond_resched_lock(spinlock_t
*lock
);
2414 #ifdef CONFIG_PREEMPT
2415 #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
2417 #define PREEMPT_LOCK_OFFSET 0
2420 #define cond_resched_lock(lock) ({ \
2421 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
2422 __cond_resched_lock(lock); \
2425 extern int __cond_resched_softirq(void);
2427 #define cond_resched_softirq() ({ \
2428 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2429 __cond_resched_softirq(); \
2433 * Does a critical section need to be broken due to another
2434 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2435 * but a general need for low latency)
2437 static inline int spin_needbreak(spinlock_t
*lock
)
2439 #ifdef CONFIG_PREEMPT
2440 return spin_is_contended(lock
);
2447 * Thread group CPU time accounting.
2449 void thread_group_cputime(struct task_struct
*tsk
, struct task_cputime
*times
);
2450 void thread_group_cputimer(struct task_struct
*tsk
, struct task_cputime
*times
);
2452 static inline void thread_group_cputime_init(struct signal_struct
*sig
)
2454 sig
->cputimer
.cputime
= INIT_CPUTIME
;
2455 spin_lock_init(&sig
->cputimer
.lock
);
2456 sig
->cputimer
.running
= 0;
2459 static inline void thread_group_cputime_free(struct signal_struct
*sig
)
2464 * Reevaluate whether the task has signals pending delivery.
2465 * Wake the task if so.
2466 * This is required every time the blocked sigset_t changes.
2467 * callers must hold sighand->siglock.
2469 extern void recalc_sigpending_and_wake(struct task_struct
*t
);
2470 extern void recalc_sigpending(void);
2472 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
2475 * Wrappers for p->thread_info->cpu access. No-op on UP.
2479 static inline unsigned int task_cpu(const struct task_struct
*p
)
2481 return task_thread_info(p
)->cpu
;
2484 extern void set_task_cpu(struct task_struct
*p
, unsigned int cpu
);
2488 static inline unsigned int task_cpu(const struct task_struct
*p
)
2493 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
2497 #endif /* CONFIG_SMP */
2499 #ifdef CONFIG_TRACING
2501 __trace_special(void *__tr
, void *__data
,
2502 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
);
2505 __trace_special(void *__tr
, void *__data
,
2506 unsigned long arg1
, unsigned long arg2
, unsigned long arg3
)
2511 extern long sched_setaffinity(pid_t pid
, const struct cpumask
*new_mask
);
2512 extern long sched_getaffinity(pid_t pid
, struct cpumask
*mask
);
2514 extern void normalize_rt_tasks(void);
2516 #ifdef CONFIG_GROUP_SCHED
2518 extern struct task_group init_task_group
;
2519 #ifdef CONFIG_USER_SCHED
2520 extern struct task_group root_task_group
;
2521 extern void set_tg_uid(struct user_struct
*user
);
2524 extern struct task_group
*sched_create_group(struct task_group
*parent
);
2525 extern void sched_destroy_group(struct task_group
*tg
);
2526 extern void sched_move_task(struct task_struct
*tsk
);
2527 #ifdef CONFIG_FAIR_GROUP_SCHED
2528 extern int sched_group_set_shares(struct task_group
*tg
, unsigned long shares
);
2529 extern unsigned long sched_group_shares(struct task_group
*tg
);
2531 #ifdef CONFIG_RT_GROUP_SCHED
2532 extern int sched_group_set_rt_runtime(struct task_group
*tg
,
2533 long rt_runtime_us
);
2534 extern long sched_group_rt_runtime(struct task_group
*tg
);
2535 extern int sched_group_set_rt_period(struct task_group
*tg
,
2537 extern long sched_group_rt_period(struct task_group
*tg
);
2538 extern int sched_rt_can_attach(struct task_group
*tg
, struct task_struct
*tsk
);
2542 extern int task_can_switch_user(struct user_struct
*up
,
2543 struct task_struct
*tsk
);
2545 #ifdef CONFIG_TASK_XACCT
2546 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2548 tsk
->ioac
.rchar
+= amt
;
2551 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2553 tsk
->ioac
.wchar
+= amt
;
2556 static inline void inc_syscr(struct task_struct
*tsk
)
2561 static inline void inc_syscw(struct task_struct
*tsk
)
2566 static inline void add_rchar(struct task_struct
*tsk
, ssize_t amt
)
2570 static inline void add_wchar(struct task_struct
*tsk
, ssize_t amt
)
2574 static inline void inc_syscr(struct task_struct
*tsk
)
2578 static inline void inc_syscw(struct task_struct
*tsk
)
2583 #ifndef TASK_SIZE_OF
2584 #define TASK_SIZE_OF(tsk) TASK_SIZE
2588 * Call the function if the target task is executing on a CPU right now:
2590 extern void task_oncpu_function_call(struct task_struct
*p
,
2591 void (*func
) (void *info
), void *info
);
2594 #ifdef CONFIG_MM_OWNER
2595 extern void mm_update_next_owner(struct mm_struct
*mm
);
2596 extern void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
);
2598 static inline void mm_update_next_owner(struct mm_struct
*mm
)
2602 static inline void mm_init_owner(struct mm_struct
*mm
, struct task_struct
*p
)
2605 #endif /* CONFIG_MM_OWNER */
2607 static inline unsigned long task_rlimit(const struct task_struct
*tsk
,
2610 return ACCESS_ONCE(tsk
->signal
->rlim
[limit
].rlim_cur
);
2613 static inline unsigned long task_rlimit_max(const struct task_struct
*tsk
,
2616 return ACCESS_ONCE(tsk
->signal
->rlim
[limit
].rlim_max
);
2619 static inline unsigned long rlimit(unsigned int limit
)
2621 return task_rlimit(current
, limit
);
2624 static inline unsigned long rlimit_max(unsigned int limit
)
2626 return task_rlimit_max(current
, limit
);
2629 #endif /* __KERNEL__ */