usb: musb: gadget: fix kernel panic if using out ep with FIFO_TXRX style
[linux-2.6/linux-acpi-2.6/ibm-acpi-2.6.git] / include / linux / sched.h
blob957a25fff8fc10920eee128f058f37a0697352d3
1 #ifndef _LINUX_SCHED_H
2 #define _LINUX_SCHED_H
4 /*
5 * cloning flags:
6 */
7 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
9 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15 #define CLONE_THREAD 0x00010000 /* Same thread group? */
16 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
17 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
24 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
25 #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
26 #define CLONE_NEWIPC 0x08000000 /* New ipcs */
27 #define CLONE_NEWUSER 0x10000000 /* New user namespace */
28 #define CLONE_NEWPID 0x20000000 /* New pid namespace */
29 #define CLONE_NEWNET 0x40000000 /* New network namespace */
30 #define CLONE_IO 0x80000000 /* Clone io context */
33 * Scheduling policies
35 #define SCHED_NORMAL 0
36 #define SCHED_FIFO 1
37 #define SCHED_RR 2
38 #define SCHED_BATCH 3
39 /* SCHED_ISO: reserved but not implemented yet */
40 #define SCHED_IDLE 5
41 /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
42 #define SCHED_RESET_ON_FORK 0x40000000
44 #ifdef __KERNEL__
46 struct sched_param {
47 int sched_priority;
50 #include <asm/param.h> /* for HZ */
52 #include <linux/capability.h>
53 #include <linux/threads.h>
54 #include <linux/kernel.h>
55 #include <linux/types.h>
56 #include <linux/timex.h>
57 #include <linux/jiffies.h>
58 #include <linux/rbtree.h>
59 #include <linux/thread_info.h>
60 #include <linux/cpumask.h>
61 #include <linux/errno.h>
62 #include <linux/nodemask.h>
63 #include <linux/mm_types.h>
65 #include <asm/system.h>
66 #include <asm/page.h>
67 #include <asm/ptrace.h>
68 #include <asm/cputime.h>
70 #include <linux/smp.h>
71 #include <linux/sem.h>
72 #include <linux/signal.h>
73 #include <linux/path.h>
74 #include <linux/compiler.h>
75 #include <linux/completion.h>
76 #include <linux/pid.h>
77 #include <linux/percpu.h>
78 #include <linux/topology.h>
79 #include <linux/proportions.h>
80 #include <linux/seccomp.h>
81 #include <linux/rcupdate.h>
82 #include <linux/rculist.h>
83 #include <linux/rtmutex.h>
85 #include <linux/time.h>
86 #include <linux/param.h>
87 #include <linux/resource.h>
88 #include <linux/timer.h>
89 #include <linux/hrtimer.h>
90 #include <linux/task_io_accounting.h>
91 #include <linux/kobject.h>
92 #include <linux/latencytop.h>
93 #include <linux/cred.h>
95 #include <asm/processor.h>
97 struct exec_domain;
98 struct futex_pi_state;
99 struct robust_list_head;
100 struct bio;
101 struct fs_struct;
102 struct bts_context;
103 struct perf_event_context;
106 * List of flags we want to share for kernel threads,
107 * if only because they are not used by them anyway.
109 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
112 * These are the constant used to fake the fixed-point load-average
113 * counting. Some notes:
114 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
115 * a load-average precision of 10 bits integer + 11 bits fractional
116 * - if you want to count load-averages more often, you need more
117 * precision, or rounding will get you. With 2-second counting freq,
118 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 * 11 bit fractions.
121 extern unsigned long avenrun[]; /* Load averages */
122 extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
124 #define FSHIFT 11 /* nr of bits of precision */
125 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
126 #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
127 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
128 #define EXP_5 2014 /* 1/exp(5sec/5min) */
129 #define EXP_15 2037 /* 1/exp(5sec/15min) */
131 #define CALC_LOAD(load,exp,n) \
132 load *= exp; \
133 load += n*(FIXED_1-exp); \
134 load >>= FSHIFT;
136 extern unsigned long total_forks;
137 extern int nr_threads;
138 DECLARE_PER_CPU(unsigned long, process_counts);
139 extern int nr_processes(void);
140 extern unsigned long nr_running(void);
141 extern unsigned long nr_uninterruptible(void);
142 extern unsigned long nr_iowait(void);
143 extern unsigned long nr_iowait_cpu(void);
144 extern unsigned long this_cpu_load(void);
147 extern void calc_global_load(void);
149 extern unsigned long get_parent_ip(unsigned long addr);
151 struct seq_file;
152 struct cfs_rq;
153 struct task_group;
154 #ifdef CONFIG_SCHED_DEBUG
155 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
156 extern void proc_sched_set_task(struct task_struct *p);
157 extern void
158 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
159 #else
160 static inline void
161 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
164 static inline void proc_sched_set_task(struct task_struct *p)
167 static inline void
168 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
171 #endif
173 extern unsigned long long time_sync_thresh;
176 * Task state bitmask. NOTE! These bits are also
177 * encoded in fs/proc/array.c: get_task_state().
179 * We have two separate sets of flags: task->state
180 * is about runnability, while task->exit_state are
181 * about the task exiting. Confusing, but this way
182 * modifying one set can't modify the other one by
183 * mistake.
185 #define TASK_RUNNING 0
186 #define TASK_INTERRUPTIBLE 1
187 #define TASK_UNINTERRUPTIBLE 2
188 #define __TASK_STOPPED 4
189 #define __TASK_TRACED 8
190 /* in tsk->exit_state */
191 #define EXIT_ZOMBIE 16
192 #define EXIT_DEAD 32
193 /* in tsk->state again */
194 #define TASK_DEAD 64
195 #define TASK_WAKEKILL 128
196 #define TASK_WAKING 256
198 /* Convenience macros for the sake of set_task_state */
199 #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
200 #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
201 #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
203 /* Convenience macros for the sake of wake_up */
204 #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
205 #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
207 /* get_task_state() */
208 #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
209 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
210 __TASK_TRACED)
212 #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
213 #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
214 #define task_is_stopped_or_traced(task) \
215 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
216 #define task_contributes_to_load(task) \
217 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
218 (task->flags & PF_FREEZING) == 0)
220 #define __set_task_state(tsk, state_value) \
221 do { (tsk)->state = (state_value); } while (0)
222 #define set_task_state(tsk, state_value) \
223 set_mb((tsk)->state, (state_value))
226 * set_current_state() includes a barrier so that the write of current->state
227 * is correctly serialised wrt the caller's subsequent test of whether to
228 * actually sleep:
230 * set_current_state(TASK_UNINTERRUPTIBLE);
231 * if (do_i_need_to_sleep())
232 * schedule();
234 * If the caller does not need such serialisation then use __set_current_state()
236 #define __set_current_state(state_value) \
237 do { current->state = (state_value); } while (0)
238 #define set_current_state(state_value) \
239 set_mb(current->state, (state_value))
241 /* Task command name length */
242 #define TASK_COMM_LEN 16
244 #include <linux/spinlock.h>
247 * This serializes "schedule()" and also protects
248 * the run-queue from deletions/modifications (but
249 * _adding_ to the beginning of the run-queue has
250 * a separate lock).
252 extern rwlock_t tasklist_lock;
253 extern spinlock_t mmlist_lock;
255 struct task_struct;
257 extern void sched_init(void);
258 extern void sched_init_smp(void);
259 extern asmlinkage void schedule_tail(struct task_struct *prev);
260 extern void init_idle(struct task_struct *idle, int cpu);
261 extern void init_idle_bootup_task(struct task_struct *idle);
263 extern int runqueue_is_locked(int cpu);
264 extern void task_rq_unlock_wait(struct task_struct *p);
266 extern cpumask_var_t nohz_cpu_mask;
267 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
268 extern int select_nohz_load_balancer(int cpu);
269 extern int get_nohz_load_balancer(void);
270 #else
271 static inline int select_nohz_load_balancer(int cpu)
273 return 0;
275 #endif
278 * Only dump TASK_* tasks. (0 for all tasks)
280 extern void show_state_filter(unsigned long state_filter);
282 static inline void show_state(void)
284 show_state_filter(0);
287 extern void show_regs(struct pt_regs *);
290 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
291 * task), SP is the stack pointer of the first frame that should be shown in the back
292 * trace (or NULL if the entire call-chain of the task should be shown).
294 extern void show_stack(struct task_struct *task, unsigned long *sp);
296 void io_schedule(void);
297 long io_schedule_timeout(long timeout);
299 extern void cpu_init (void);
300 extern void trap_init(void);
301 extern void update_process_times(int user);
302 extern void scheduler_tick(void);
304 extern void sched_show_task(struct task_struct *p);
306 #ifdef CONFIG_DETECT_SOFTLOCKUP
307 extern void softlockup_tick(void);
308 extern void touch_softlockup_watchdog(void);
309 extern void touch_all_softlockup_watchdogs(void);
310 extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
311 void __user *buffer,
312 size_t *lenp, loff_t *ppos);
313 extern unsigned int softlockup_panic;
314 extern int softlockup_thresh;
315 #else
316 static inline void softlockup_tick(void)
319 static inline void touch_softlockup_watchdog(void)
322 static inline void touch_all_softlockup_watchdogs(void)
325 #endif
327 #ifdef CONFIG_DETECT_HUNG_TASK
328 extern unsigned int sysctl_hung_task_panic;
329 extern unsigned long sysctl_hung_task_check_count;
330 extern unsigned long sysctl_hung_task_timeout_secs;
331 extern unsigned long sysctl_hung_task_warnings;
332 extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
333 void __user *buffer,
334 size_t *lenp, loff_t *ppos);
335 #endif
337 /* Attach to any functions which should be ignored in wchan output. */
338 #define __sched __attribute__((__section__(".sched.text")))
340 /* Linker adds these: start and end of __sched functions */
341 extern char __sched_text_start[], __sched_text_end[];
343 /* Is this address in the __sched functions? */
344 extern int in_sched_functions(unsigned long addr);
346 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
347 extern signed long schedule_timeout(signed long timeout);
348 extern signed long schedule_timeout_interruptible(signed long timeout);
349 extern signed long schedule_timeout_killable(signed long timeout);
350 extern signed long schedule_timeout_uninterruptible(signed long timeout);
351 asmlinkage void __schedule(void);
352 asmlinkage void schedule(void);
353 extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
355 struct nsproxy;
356 struct user_namespace;
359 * Default maximum number of active map areas, this limits the number of vmas
360 * per mm struct. Users can overwrite this number by sysctl but there is a
361 * problem.
363 * When a program's coredump is generated as ELF format, a section is created
364 * per a vma. In ELF, the number of sections is represented in unsigned short.
365 * This means the number of sections should be smaller than 65535 at coredump.
366 * Because the kernel adds some informative sections to a image of program at
367 * generating coredump, we need some margin. The number of extra sections is
368 * 1-3 now and depends on arch. We use "5" as safe margin, here.
370 #define MAPCOUNT_ELF_CORE_MARGIN (5)
371 #define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
373 extern int sysctl_max_map_count;
375 #include <linux/aio.h>
377 extern unsigned long
378 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
379 unsigned long, unsigned long);
380 extern unsigned long
381 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
382 unsigned long len, unsigned long pgoff,
383 unsigned long flags);
384 extern void arch_unmap_area(struct mm_struct *, unsigned long);
385 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
387 #if USE_SPLIT_PTLOCKS
389 * The mm counters are not protected by its page_table_lock,
390 * so must be incremented atomically.
392 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
393 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
394 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
395 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
396 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
398 #else /* !USE_SPLIT_PTLOCKS */
400 * The mm counters are protected by its page_table_lock,
401 * so can be incremented directly.
403 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
404 #define get_mm_counter(mm, member) ((mm)->_##member)
405 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
406 #define inc_mm_counter(mm, member) (mm)->_##member++
407 #define dec_mm_counter(mm, member) (mm)->_##member--
409 #endif /* !USE_SPLIT_PTLOCKS */
411 #define get_mm_rss(mm) \
412 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
413 #define update_hiwater_rss(mm) do { \
414 unsigned long _rss = get_mm_rss(mm); \
415 if ((mm)->hiwater_rss < _rss) \
416 (mm)->hiwater_rss = _rss; \
417 } while (0)
418 #define update_hiwater_vm(mm) do { \
419 if ((mm)->hiwater_vm < (mm)->total_vm) \
420 (mm)->hiwater_vm = (mm)->total_vm; \
421 } while (0)
423 static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
425 return max(mm->hiwater_rss, get_mm_rss(mm));
428 static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
429 struct mm_struct *mm)
431 unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
433 if (*maxrss < hiwater_rss)
434 *maxrss = hiwater_rss;
437 static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
439 return max(mm->hiwater_vm, mm->total_vm);
442 extern void set_dumpable(struct mm_struct *mm, int value);
443 extern int get_dumpable(struct mm_struct *mm);
445 /* mm flags */
446 /* dumpable bits */
447 #define MMF_DUMPABLE 0 /* core dump is permitted */
448 #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
450 #define MMF_DUMPABLE_BITS 2
451 #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
453 /* coredump filter bits */
454 #define MMF_DUMP_ANON_PRIVATE 2
455 #define MMF_DUMP_ANON_SHARED 3
456 #define MMF_DUMP_MAPPED_PRIVATE 4
457 #define MMF_DUMP_MAPPED_SHARED 5
458 #define MMF_DUMP_ELF_HEADERS 6
459 #define MMF_DUMP_HUGETLB_PRIVATE 7
460 #define MMF_DUMP_HUGETLB_SHARED 8
462 #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
463 #define MMF_DUMP_FILTER_BITS 7
464 #define MMF_DUMP_FILTER_MASK \
465 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
466 #define MMF_DUMP_FILTER_DEFAULT \
467 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
468 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
470 #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
471 # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
472 #else
473 # define MMF_DUMP_MASK_DEFAULT_ELF 0
474 #endif
475 /* leave room for more dump flags */
476 #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
478 #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
480 struct sighand_struct {
481 atomic_t count;
482 struct k_sigaction action[_NSIG];
483 spinlock_t siglock;
484 wait_queue_head_t signalfd_wqh;
487 struct pacct_struct {
488 int ac_flag;
489 long ac_exitcode;
490 unsigned long ac_mem;
491 cputime_t ac_utime, ac_stime;
492 unsigned long ac_minflt, ac_majflt;
495 struct cpu_itimer {
496 cputime_t expires;
497 cputime_t incr;
498 u32 error;
499 u32 incr_error;
503 * struct task_cputime - collected CPU time counts
504 * @utime: time spent in user mode, in &cputime_t units
505 * @stime: time spent in kernel mode, in &cputime_t units
506 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
508 * This structure groups together three kinds of CPU time that are
509 * tracked for threads and thread groups. Most things considering
510 * CPU time want to group these counts together and treat all three
511 * of them in parallel.
513 struct task_cputime {
514 cputime_t utime;
515 cputime_t stime;
516 unsigned long long sum_exec_runtime;
518 /* Alternate field names when used to cache expirations. */
519 #define prof_exp stime
520 #define virt_exp utime
521 #define sched_exp sum_exec_runtime
523 #define INIT_CPUTIME \
524 (struct task_cputime) { \
525 .utime = cputime_zero, \
526 .stime = cputime_zero, \
527 .sum_exec_runtime = 0, \
531 * Disable preemption until the scheduler is running.
532 * Reset by start_kernel()->sched_init()->init_idle().
534 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
535 * before the scheduler is active -- see should_resched().
537 #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
540 * struct thread_group_cputimer - thread group interval timer counts
541 * @cputime: thread group interval timers.
542 * @running: non-zero when there are timers running and
543 * @cputime receives updates.
544 * @lock: lock for fields in this struct.
546 * This structure contains the version of task_cputime, above, that is
547 * used for thread group CPU timer calculations.
549 struct thread_group_cputimer {
550 struct task_cputime cputime;
551 int running;
552 spinlock_t lock;
556 * NOTE! "signal_struct" does not have it's own
557 * locking, because a shared signal_struct always
558 * implies a shared sighand_struct, so locking
559 * sighand_struct is always a proper superset of
560 * the locking of signal_struct.
562 struct signal_struct {
563 atomic_t count;
564 atomic_t live;
566 wait_queue_head_t wait_chldexit; /* for wait4() */
568 /* current thread group signal load-balancing target: */
569 struct task_struct *curr_target;
571 /* shared signal handling: */
572 struct sigpending shared_pending;
574 /* thread group exit support */
575 int group_exit_code;
576 /* overloaded:
577 * - notify group_exit_task when ->count is equal to notify_count
578 * - everyone except group_exit_task is stopped during signal delivery
579 * of fatal signals, group_exit_task processes the signal.
581 int notify_count;
582 struct task_struct *group_exit_task;
584 /* thread group stop support, overloads group_exit_code too */
585 int group_stop_count;
586 unsigned int flags; /* see SIGNAL_* flags below */
588 /* POSIX.1b Interval Timers */
589 struct list_head posix_timers;
591 /* ITIMER_REAL timer for the process */
592 struct hrtimer real_timer;
593 struct pid *leader_pid;
594 ktime_t it_real_incr;
597 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
598 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
599 * values are defined to 0 and 1 respectively
601 struct cpu_itimer it[2];
604 * Thread group totals for process CPU timers.
605 * See thread_group_cputimer(), et al, for details.
607 struct thread_group_cputimer cputimer;
609 /* Earliest-expiration cache. */
610 struct task_cputime cputime_expires;
612 struct list_head cpu_timers[3];
614 struct pid *tty_old_pgrp;
616 /* boolean value for session group leader */
617 int leader;
619 struct tty_struct *tty; /* NULL if no tty */
622 * Cumulative resource counters for dead threads in the group,
623 * and for reaped dead child processes forked by this group.
624 * Live threads maintain their own counters and add to these
625 * in __exit_signal, except for the group leader.
627 cputime_t utime, stime, cutime, cstime;
628 cputime_t gtime;
629 cputime_t cgtime;
630 #ifndef CONFIG_VIRT_CPU_ACCOUNTING
631 cputime_t prev_utime, prev_stime;
632 #endif
633 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
634 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
635 unsigned long inblock, oublock, cinblock, coublock;
636 unsigned long maxrss, cmaxrss;
637 struct task_io_accounting ioac;
640 * Cumulative ns of schedule CPU time fo dead threads in the
641 * group, not including a zombie group leader, (This only differs
642 * from jiffies_to_ns(utime + stime) if sched_clock uses something
643 * other than jiffies.)
645 unsigned long long sum_sched_runtime;
648 * We don't bother to synchronize most readers of this at all,
649 * because there is no reader checking a limit that actually needs
650 * to get both rlim_cur and rlim_max atomically, and either one
651 * alone is a single word that can safely be read normally.
652 * getrlimit/setrlimit use task_lock(current->group_leader) to
653 * protect this instead of the siglock, because they really
654 * have no need to disable irqs.
656 struct rlimit rlim[RLIM_NLIMITS];
658 #ifdef CONFIG_BSD_PROCESS_ACCT
659 struct pacct_struct pacct; /* per-process accounting information */
660 #endif
661 #ifdef CONFIG_TASKSTATS
662 struct taskstats *stats;
663 #endif
664 #ifdef CONFIG_AUDIT
665 unsigned audit_tty;
666 struct tty_audit_buf *tty_audit_buf;
667 #endif
669 int oom_adj; /* OOM kill score adjustment (bit shift) */
672 /* Context switch must be unlocked if interrupts are to be enabled */
673 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
674 # define __ARCH_WANT_UNLOCKED_CTXSW
675 #endif
678 * Bits in flags field of signal_struct.
680 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
681 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
682 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
683 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
685 * Pending notifications to parent.
687 #define SIGNAL_CLD_STOPPED 0x00000010
688 #define SIGNAL_CLD_CONTINUED 0x00000020
689 #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
691 #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
693 /* If true, all threads except ->group_exit_task have pending SIGKILL */
694 static inline int signal_group_exit(const struct signal_struct *sig)
696 return (sig->flags & SIGNAL_GROUP_EXIT) ||
697 (sig->group_exit_task != NULL);
701 * Some day this will be a full-fledged user tracking system..
703 struct user_struct {
704 atomic_t __count; /* reference count */
705 atomic_t processes; /* How many processes does this user have? */
706 atomic_t files; /* How many open files does this user have? */
707 atomic_t sigpending; /* How many pending signals does this user have? */
708 #ifdef CONFIG_INOTIFY_USER
709 atomic_t inotify_watches; /* How many inotify watches does this user have? */
710 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
711 #endif
712 #ifdef CONFIG_EPOLL
713 atomic_t epoll_watches; /* The number of file descriptors currently watched */
714 #endif
715 #ifdef CONFIG_POSIX_MQUEUE
716 /* protected by mq_lock */
717 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
718 #endif
719 unsigned long locked_shm; /* How many pages of mlocked shm ? */
721 #ifdef CONFIG_KEYS
722 struct key *uid_keyring; /* UID specific keyring */
723 struct key *session_keyring; /* UID's default session keyring */
724 #endif
726 /* Hash table maintenance information */
727 struct hlist_node uidhash_node;
728 uid_t uid;
729 struct user_namespace *user_ns;
731 #ifdef CONFIG_USER_SCHED
732 struct task_group *tg;
733 #ifdef CONFIG_SYSFS
734 struct kobject kobj;
735 struct delayed_work work;
736 #endif
737 #endif
739 #ifdef CONFIG_PERF_EVENTS
740 atomic_long_t locked_vm;
741 #endif
744 extern int uids_sysfs_init(void);
746 extern struct user_struct *find_user(uid_t);
748 extern struct user_struct root_user;
749 #define INIT_USER (&root_user)
752 struct backing_dev_info;
753 struct reclaim_state;
755 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
756 struct sched_info {
757 /* cumulative counters */
758 unsigned long pcount; /* # of times run on this cpu */
759 unsigned long long run_delay; /* time spent waiting on a runqueue */
761 /* timestamps */
762 unsigned long long last_arrival,/* when we last ran on a cpu */
763 last_queued; /* when we were last queued to run */
764 #ifdef CONFIG_SCHEDSTATS
765 /* BKL stats */
766 unsigned int bkl_count;
767 #endif
769 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
771 #ifdef CONFIG_TASK_DELAY_ACCT
772 struct task_delay_info {
773 spinlock_t lock;
774 unsigned int flags; /* Private per-task flags */
776 /* For each stat XXX, add following, aligned appropriately
778 * struct timespec XXX_start, XXX_end;
779 * u64 XXX_delay;
780 * u32 XXX_count;
782 * Atomicity of updates to XXX_delay, XXX_count protected by
783 * single lock above (split into XXX_lock if contention is an issue).
787 * XXX_count is incremented on every XXX operation, the delay
788 * associated with the operation is added to XXX_delay.
789 * XXX_delay contains the accumulated delay time in nanoseconds.
791 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
792 u64 blkio_delay; /* wait for sync block io completion */
793 u64 swapin_delay; /* wait for swapin block io completion */
794 u32 blkio_count; /* total count of the number of sync block */
795 /* io operations performed */
796 u32 swapin_count; /* total count of the number of swapin block */
797 /* io operations performed */
799 struct timespec freepages_start, freepages_end;
800 u64 freepages_delay; /* wait for memory reclaim */
801 u32 freepages_count; /* total count of memory reclaim */
803 #endif /* CONFIG_TASK_DELAY_ACCT */
805 static inline int sched_info_on(void)
807 #ifdef CONFIG_SCHEDSTATS
808 return 1;
809 #elif defined(CONFIG_TASK_DELAY_ACCT)
810 extern int delayacct_on;
811 return delayacct_on;
812 #else
813 return 0;
814 #endif
817 enum cpu_idle_type {
818 CPU_IDLE,
819 CPU_NOT_IDLE,
820 CPU_NEWLY_IDLE,
821 CPU_MAX_IDLE_TYPES
825 * sched-domains (multiprocessor balancing) declarations:
829 * Increase resolution of nice-level calculations:
831 #define SCHED_LOAD_SHIFT 10
832 #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
834 #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
836 #ifdef CONFIG_SMP
837 #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
838 #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
839 #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
840 #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
841 #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
842 #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
843 #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
844 #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
845 #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
846 #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
847 #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
849 #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
851 enum powersavings_balance_level {
852 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
853 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
854 * first for long running threads
856 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
857 * cpu package for power savings
859 MAX_POWERSAVINGS_BALANCE_LEVELS
862 extern int sched_mc_power_savings, sched_smt_power_savings;
864 static inline int sd_balance_for_mc_power(void)
866 if (sched_smt_power_savings)
867 return SD_POWERSAVINGS_BALANCE;
869 if (!sched_mc_power_savings)
870 return SD_PREFER_SIBLING;
872 return 0;
875 static inline int sd_balance_for_package_power(void)
877 if (sched_mc_power_savings | sched_smt_power_savings)
878 return SD_POWERSAVINGS_BALANCE;
880 return SD_PREFER_SIBLING;
884 * Optimise SD flags for power savings:
885 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
886 * Keep default SD flags if sched_{smt,mc}_power_saving=0
889 static inline int sd_power_saving_flags(void)
891 if (sched_mc_power_savings | sched_smt_power_savings)
892 return SD_BALANCE_NEWIDLE;
894 return 0;
897 struct sched_group {
898 struct sched_group *next; /* Must be a circular list */
901 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
902 * single CPU.
904 unsigned int cpu_power;
907 * The CPUs this group covers.
909 * NOTE: this field is variable length. (Allocated dynamically
910 * by attaching extra space to the end of the structure,
911 * depending on how many CPUs the kernel has booted up with)
913 * It is also be embedded into static data structures at build
914 * time. (See 'struct static_sched_group' in kernel/sched.c)
916 unsigned long cpumask[0];
919 static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
921 return to_cpumask(sg->cpumask);
924 enum sched_domain_level {
925 SD_LV_NONE = 0,
926 SD_LV_SIBLING,
927 SD_LV_MC,
928 SD_LV_CPU,
929 SD_LV_NODE,
930 SD_LV_ALLNODES,
931 SD_LV_MAX
934 struct sched_domain_attr {
935 int relax_domain_level;
938 #define SD_ATTR_INIT (struct sched_domain_attr) { \
939 .relax_domain_level = -1, \
942 struct sched_domain {
943 /* These fields must be setup */
944 struct sched_domain *parent; /* top domain must be null terminated */
945 struct sched_domain *child; /* bottom domain must be null terminated */
946 struct sched_group *groups; /* the balancing groups of the domain */
947 unsigned long min_interval; /* Minimum balance interval ms */
948 unsigned long max_interval; /* Maximum balance interval ms */
949 unsigned int busy_factor; /* less balancing by factor if busy */
950 unsigned int imbalance_pct; /* No balance until over watermark */
951 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
952 unsigned int busy_idx;
953 unsigned int idle_idx;
954 unsigned int newidle_idx;
955 unsigned int wake_idx;
956 unsigned int forkexec_idx;
957 unsigned int smt_gain;
958 int flags; /* See SD_* */
959 enum sched_domain_level level;
961 /* Runtime fields. */
962 unsigned long last_balance; /* init to jiffies. units in jiffies */
963 unsigned int balance_interval; /* initialise to 1. units in ms. */
964 unsigned int nr_balance_failed; /* initialise to 0 */
966 u64 last_update;
968 #ifdef CONFIG_SCHEDSTATS
969 /* load_balance() stats */
970 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
971 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
972 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
973 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
974 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
975 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
976 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
977 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
979 /* Active load balancing */
980 unsigned int alb_count;
981 unsigned int alb_failed;
982 unsigned int alb_pushed;
984 /* SD_BALANCE_EXEC stats */
985 unsigned int sbe_count;
986 unsigned int sbe_balanced;
987 unsigned int sbe_pushed;
989 /* SD_BALANCE_FORK stats */
990 unsigned int sbf_count;
991 unsigned int sbf_balanced;
992 unsigned int sbf_pushed;
994 /* try_to_wake_up() stats */
995 unsigned int ttwu_wake_remote;
996 unsigned int ttwu_move_affine;
997 unsigned int ttwu_move_balance;
998 #endif
999 #ifdef CONFIG_SCHED_DEBUG
1000 char *name;
1001 #endif
1003 unsigned int span_weight;
1005 * Span of all CPUs in this domain.
1007 * NOTE: this field is variable length. (Allocated dynamically
1008 * by attaching extra space to the end of the structure,
1009 * depending on how many CPUs the kernel has booted up with)
1011 * It is also be embedded into static data structures at build
1012 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1014 unsigned long span[0];
1017 static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1019 return to_cpumask(sd->span);
1022 extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1023 struct sched_domain_attr *dattr_new);
1025 /* Test a flag in parent sched domain */
1026 static inline int test_sd_parent(struct sched_domain *sd, int flag)
1028 if (sd->parent && (sd->parent->flags & flag))
1029 return 1;
1031 return 0;
1034 unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1035 unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1037 #else /* CONFIG_SMP */
1039 struct sched_domain_attr;
1041 static inline void
1042 partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1043 struct sched_domain_attr *dattr_new)
1046 #endif /* !CONFIG_SMP */
1049 struct io_context; /* See blkdev.h */
1052 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
1053 extern void prefetch_stack(struct task_struct *t);
1054 #else
1055 static inline void prefetch_stack(struct task_struct *t) { }
1056 #endif
1058 struct audit_context; /* See audit.c */
1059 struct mempolicy;
1060 struct pipe_inode_info;
1061 struct uts_namespace;
1063 struct rq;
1064 struct sched_domain;
1067 * wake flags
1069 #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
1070 #define WF_FORK 0x02 /* child wakeup after fork */
1072 struct sched_class {
1073 const struct sched_class *next;
1075 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup,
1076 bool head);
1077 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
1078 void (*yield_task) (struct rq *rq);
1080 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
1082 struct task_struct * (*pick_next_task) (struct rq *rq);
1083 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
1085 #ifdef CONFIG_SMP
1086 int (*select_task_rq)(struct rq *rq, struct task_struct *p,
1087 int sd_flag, int flags);
1089 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
1090 struct rq *busiest, unsigned long max_load_move,
1091 struct sched_domain *sd, enum cpu_idle_type idle,
1092 int *all_pinned, int *this_best_prio);
1094 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1095 struct rq *busiest, struct sched_domain *sd,
1096 enum cpu_idle_type idle);
1097 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1098 void (*post_schedule) (struct rq *this_rq);
1099 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1100 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
1102 void (*set_cpus_allowed)(struct task_struct *p,
1103 const struct cpumask *newmask);
1105 void (*rq_online)(struct rq *rq);
1106 void (*rq_offline)(struct rq *rq);
1107 #endif
1109 void (*set_curr_task) (struct rq *rq);
1110 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1111 void (*task_fork) (struct task_struct *p);
1113 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1114 int running);
1115 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1116 int running);
1117 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1118 int oldprio, int running);
1120 unsigned int (*get_rr_interval) (struct rq *rq,
1121 struct task_struct *task);
1123 #ifdef CONFIG_FAIR_GROUP_SCHED
1124 void (*moved_group) (struct task_struct *p, int on_rq);
1125 #endif
1128 struct load_weight {
1129 unsigned long weight, inv_weight;
1133 * CFS stats for a schedulable entity (task, task-group etc)
1135 * Current field usage histogram:
1137 * 4 se->block_start
1138 * 4 se->run_node
1139 * 4 se->sleep_start
1140 * 6 se->load.weight
1142 struct sched_entity {
1143 struct load_weight load; /* for load-balancing */
1144 struct rb_node run_node;
1145 struct list_head group_node;
1146 unsigned int on_rq;
1148 u64 exec_start;
1149 u64 sum_exec_runtime;
1150 u64 vruntime;
1151 u64 prev_sum_exec_runtime;
1153 u64 last_wakeup;
1154 u64 avg_overlap;
1156 u64 nr_migrations;
1158 u64 start_runtime;
1159 u64 avg_wakeup;
1161 u64 avg_running;
1163 #ifdef CONFIG_SCHEDSTATS
1164 u64 wait_start;
1165 u64 wait_max;
1166 u64 wait_count;
1167 u64 wait_sum;
1168 u64 iowait_count;
1169 u64 iowait_sum;
1171 u64 sleep_start;
1172 u64 sleep_max;
1173 s64 sum_sleep_runtime;
1175 u64 block_start;
1176 u64 block_max;
1177 u64 exec_max;
1178 u64 slice_max;
1180 u64 nr_migrations_cold;
1181 u64 nr_failed_migrations_affine;
1182 u64 nr_failed_migrations_running;
1183 u64 nr_failed_migrations_hot;
1184 u64 nr_forced_migrations;
1186 u64 nr_wakeups;
1187 u64 nr_wakeups_sync;
1188 u64 nr_wakeups_migrate;
1189 u64 nr_wakeups_local;
1190 u64 nr_wakeups_remote;
1191 u64 nr_wakeups_affine;
1192 u64 nr_wakeups_affine_attempts;
1193 u64 nr_wakeups_passive;
1194 u64 nr_wakeups_idle;
1195 #endif
1197 #ifdef CONFIG_FAIR_GROUP_SCHED
1198 struct sched_entity *parent;
1199 /* rq on which this entity is (to be) queued: */
1200 struct cfs_rq *cfs_rq;
1201 /* rq "owned" by this entity/group: */
1202 struct cfs_rq *my_q;
1203 #endif
1206 struct sched_rt_entity {
1207 struct list_head run_list;
1208 unsigned long timeout;
1209 unsigned int time_slice;
1210 int nr_cpus_allowed;
1212 struct sched_rt_entity *back;
1213 #ifdef CONFIG_RT_GROUP_SCHED
1214 struct sched_rt_entity *parent;
1215 /* rq on which this entity is (to be) queued: */
1216 struct rt_rq *rt_rq;
1217 /* rq "owned" by this entity/group: */
1218 struct rt_rq *my_q;
1219 #endif
1222 struct rcu_node;
1224 struct task_struct {
1225 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
1226 void *stack;
1227 atomic_t usage;
1228 unsigned int flags; /* per process flags, defined below */
1229 unsigned int ptrace;
1231 int lock_depth; /* BKL lock depth */
1233 #ifdef CONFIG_SMP
1234 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1235 int oncpu;
1236 #endif
1237 #endif
1239 int prio, static_prio, normal_prio;
1240 unsigned int rt_priority;
1241 const struct sched_class *sched_class;
1242 struct sched_entity se;
1243 struct sched_rt_entity rt;
1245 #ifdef CONFIG_PREEMPT_NOTIFIERS
1246 /* list of struct preempt_notifier: */
1247 struct hlist_head preempt_notifiers;
1248 #endif
1251 * fpu_counter contains the number of consecutive context switches
1252 * that the FPU is used. If this is over a threshold, the lazy fpu
1253 * saving becomes unlazy to save the trap. This is an unsigned char
1254 * so that after 256 times the counter wraps and the behavior turns
1255 * lazy again; this to deal with bursty apps that only use FPU for
1256 * a short time
1258 unsigned char fpu_counter;
1259 #ifdef CONFIG_BLK_DEV_IO_TRACE
1260 unsigned int btrace_seq;
1261 #endif
1263 unsigned int policy;
1264 cpumask_t cpus_allowed;
1266 #ifdef CONFIG_TREE_PREEMPT_RCU
1267 int rcu_read_lock_nesting;
1268 char rcu_read_unlock_special;
1269 struct rcu_node *rcu_blocked_node;
1270 struct list_head rcu_node_entry;
1271 #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1273 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1274 struct sched_info sched_info;
1275 #endif
1277 struct list_head tasks;
1278 struct plist_node pushable_tasks;
1280 struct mm_struct *mm, *active_mm;
1282 /* task state */
1283 int exit_state;
1284 int exit_code, exit_signal;
1285 int pdeath_signal; /* The signal sent when the parent dies */
1286 /* ??? */
1287 unsigned int personality;
1288 unsigned did_exec:1;
1289 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1290 * execve */
1291 unsigned in_iowait:1;
1294 /* Revert to default priority/policy when forking */
1295 unsigned sched_reset_on_fork:1;
1297 pid_t pid;
1298 pid_t tgid;
1300 #ifdef CONFIG_CC_STACKPROTECTOR
1301 /* Canary value for the -fstack-protector gcc feature */
1302 unsigned long stack_canary;
1303 #endif
1306 * pointers to (original) parent process, youngest child, younger sibling,
1307 * older sibling, respectively. (p->father can be replaced with
1308 * p->real_parent->pid)
1310 struct task_struct *real_parent; /* real parent process */
1311 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1313 * children/sibling forms the list of my natural children
1315 struct list_head children; /* list of my children */
1316 struct list_head sibling; /* linkage in my parent's children list */
1317 struct task_struct *group_leader; /* threadgroup leader */
1320 * ptraced is the list of tasks this task is using ptrace on.
1321 * This includes both natural children and PTRACE_ATTACH targets.
1322 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1324 struct list_head ptraced;
1325 struct list_head ptrace_entry;
1328 * This is the tracer handle for the ptrace BTS extension.
1329 * This field actually belongs to the ptracer task.
1331 struct bts_context *bts;
1333 /* PID/PID hash table linkage. */
1334 struct pid_link pids[PIDTYPE_MAX];
1335 struct list_head thread_group;
1337 struct completion *vfork_done; /* for vfork() */
1338 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1339 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1341 cputime_t utime, stime, utimescaled, stimescaled;
1342 cputime_t gtime;
1343 cputime_t prev_utime, prev_stime;
1344 unsigned long nvcsw, nivcsw; /* context switch counts */
1345 struct timespec start_time; /* monotonic time */
1346 struct timespec real_start_time; /* boot based time */
1347 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1348 unsigned long min_flt, maj_flt;
1350 struct task_cputime cputime_expires;
1351 struct list_head cpu_timers[3];
1353 /* process credentials */
1354 const struct cred *real_cred; /* objective and real subjective task
1355 * credentials (COW) */
1356 const struct cred *cred; /* effective (overridable) subjective task
1357 * credentials (COW) */
1358 struct mutex cred_guard_mutex; /* guard against foreign influences on
1359 * credential calculations
1360 * (notably. ptrace) */
1361 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
1363 char comm[TASK_COMM_LEN]; /* executable name excluding path
1364 - access with [gs]et_task_comm (which lock
1365 it with task_lock())
1366 - initialized normally by setup_new_exec */
1367 /* file system info */
1368 int link_count, total_link_count;
1369 #ifdef CONFIG_SYSVIPC
1370 /* ipc stuff */
1371 struct sysv_sem sysvsem;
1372 #endif
1373 #ifdef CONFIG_DETECT_HUNG_TASK
1374 /* hung task detection */
1375 unsigned long last_switch_count;
1376 #endif
1377 /* CPU-specific state of this task */
1378 struct thread_struct thread;
1379 /* filesystem information */
1380 struct fs_struct *fs;
1381 /* open file information */
1382 struct files_struct *files;
1383 /* namespaces */
1384 struct nsproxy *nsproxy;
1385 /* signal handlers */
1386 struct signal_struct *signal;
1387 struct sighand_struct *sighand;
1389 sigset_t blocked, real_blocked;
1390 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1391 struct sigpending pending;
1393 unsigned long sas_ss_sp;
1394 size_t sas_ss_size;
1395 int (*notifier)(void *priv);
1396 void *notifier_data;
1397 sigset_t *notifier_mask;
1398 struct audit_context *audit_context;
1399 #ifdef CONFIG_AUDITSYSCALL
1400 uid_t loginuid;
1401 unsigned int sessionid;
1402 #endif
1403 seccomp_t seccomp;
1405 /* Thread group tracking */
1406 u32 parent_exec_id;
1407 u32 self_exec_id;
1408 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1409 * mempolicy */
1410 spinlock_t alloc_lock;
1412 #ifdef CONFIG_GENERIC_HARDIRQS
1413 /* IRQ handler threads */
1414 struct irqaction *irqaction;
1415 #endif
1417 /* Protection of the PI data structures: */
1418 spinlock_t pi_lock;
1420 #ifdef CONFIG_RT_MUTEXES
1421 /* PI waiters blocked on a rt_mutex held by this task */
1422 struct plist_head pi_waiters;
1423 /* Deadlock detection and priority inheritance handling */
1424 struct rt_mutex_waiter *pi_blocked_on;
1425 #endif
1427 #ifdef CONFIG_DEBUG_MUTEXES
1428 /* mutex deadlock detection */
1429 struct mutex_waiter *blocked_on;
1430 #endif
1431 #ifdef CONFIG_TRACE_IRQFLAGS
1432 unsigned int irq_events;
1433 int hardirqs_enabled;
1434 unsigned long hardirq_enable_ip;
1435 unsigned int hardirq_enable_event;
1436 unsigned long hardirq_disable_ip;
1437 unsigned int hardirq_disable_event;
1438 int softirqs_enabled;
1439 unsigned long softirq_disable_ip;
1440 unsigned int softirq_disable_event;
1441 unsigned long softirq_enable_ip;
1442 unsigned int softirq_enable_event;
1443 int hardirq_context;
1444 int softirq_context;
1445 #endif
1446 #ifdef CONFIG_LOCKDEP
1447 # define MAX_LOCK_DEPTH 48UL
1448 u64 curr_chain_key;
1449 int lockdep_depth;
1450 unsigned int lockdep_recursion;
1451 struct held_lock held_locks[MAX_LOCK_DEPTH];
1452 gfp_t lockdep_reclaim_gfp;
1453 #endif
1455 /* journalling filesystem info */
1456 void *journal_info;
1458 /* stacked block device info */
1459 struct bio *bio_list, **bio_tail;
1461 /* VM state */
1462 struct reclaim_state *reclaim_state;
1464 struct backing_dev_info *backing_dev_info;
1466 struct io_context *io_context;
1468 unsigned long ptrace_message;
1469 siginfo_t *last_siginfo; /* For ptrace use. */
1470 struct task_io_accounting ioac;
1471 #if defined(CONFIG_TASK_XACCT)
1472 u64 acct_rss_mem1; /* accumulated rss usage */
1473 u64 acct_vm_mem1; /* accumulated virtual memory usage */
1474 cputime_t acct_timexpd; /* stime + utime since last update */
1475 #endif
1476 #ifdef CONFIG_CPUSETS
1477 nodemask_t mems_allowed; /* Protected by alloc_lock */
1478 int cpuset_mem_spread_rotor;
1479 #endif
1480 #ifdef CONFIG_CGROUPS
1481 /* Control Group info protected by css_set_lock */
1482 struct css_set *cgroups;
1483 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1484 struct list_head cg_list;
1485 #endif
1486 #ifdef CONFIG_FUTEX
1487 struct robust_list_head __user *robust_list;
1488 #ifdef CONFIG_COMPAT
1489 struct compat_robust_list_head __user *compat_robust_list;
1490 #endif
1491 struct list_head pi_state_list;
1492 struct futex_pi_state *pi_state_cache;
1493 #endif
1494 #ifdef CONFIG_PERF_EVENTS
1495 struct perf_event_context *perf_event_ctxp;
1496 struct mutex perf_event_mutex;
1497 struct list_head perf_event_list;
1498 #endif
1499 #ifdef CONFIG_NUMA
1500 struct mempolicy *mempolicy; /* Protected by alloc_lock */
1501 short il_next;
1502 #endif
1503 atomic_t fs_excl; /* holding fs exclusive resources */
1504 struct rcu_head rcu;
1507 * cache last used pipe for splice
1509 struct pipe_inode_info *splice_pipe;
1510 #ifdef CONFIG_TASK_DELAY_ACCT
1511 struct task_delay_info *delays;
1512 #endif
1513 #ifdef CONFIG_FAULT_INJECTION
1514 int make_it_fail;
1515 #endif
1516 struct prop_local_single dirties;
1517 #ifdef CONFIG_LATENCYTOP
1518 int latency_record_count;
1519 struct latency_record latency_record[LT_SAVECOUNT];
1520 #endif
1522 * time slack values; these are used to round up poll() and
1523 * select() etc timeout values. These are in nanoseconds.
1525 unsigned long timer_slack_ns;
1526 unsigned long default_timer_slack_ns;
1528 struct list_head *scm_work_list;
1529 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
1530 /* Index of current stored adress in ret_stack */
1531 int curr_ret_stack;
1532 /* Stack of return addresses for return function tracing */
1533 struct ftrace_ret_stack *ret_stack;
1534 /* time stamp for last schedule */
1535 unsigned long long ftrace_timestamp;
1537 * Number of functions that haven't been traced
1538 * because of depth overrun.
1540 atomic_t trace_overrun;
1541 /* Pause for the tracing */
1542 atomic_t tracing_graph_pause;
1543 #endif
1544 #ifdef CONFIG_TRACING
1545 /* state flags for use by tracers */
1546 unsigned long trace;
1547 /* bitmask of trace recursion */
1548 unsigned long trace_recursion;
1549 #endif /* CONFIG_TRACING */
1552 /* Future-safe accessor for struct task_struct's cpus_allowed. */
1553 #define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1556 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1557 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1558 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1559 * values are inverted: lower p->prio value means higher priority.
1561 * The MAX_USER_RT_PRIO value allows the actual maximum
1562 * RT priority to be separate from the value exported to
1563 * user-space. This allows kernel threads to set their
1564 * priority to a value higher than any user task. Note:
1565 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1568 #define MAX_USER_RT_PRIO 100
1569 #define MAX_RT_PRIO MAX_USER_RT_PRIO
1571 #define MAX_PRIO (MAX_RT_PRIO + 40)
1572 #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1574 static inline int rt_prio(int prio)
1576 if (unlikely(prio < MAX_RT_PRIO))
1577 return 1;
1578 return 0;
1581 static inline int rt_task(struct task_struct *p)
1583 return rt_prio(p->prio);
1586 static inline struct pid *task_pid(struct task_struct *task)
1588 return task->pids[PIDTYPE_PID].pid;
1591 static inline struct pid *task_tgid(struct task_struct *task)
1593 return task->group_leader->pids[PIDTYPE_PID].pid;
1597 * Without tasklist or rcu lock it is not safe to dereference
1598 * the result of task_pgrp/task_session even if task == current,
1599 * we can race with another thread doing sys_setsid/sys_setpgid.
1601 static inline struct pid *task_pgrp(struct task_struct *task)
1603 return task->group_leader->pids[PIDTYPE_PGID].pid;
1606 static inline struct pid *task_session(struct task_struct *task)
1608 return task->group_leader->pids[PIDTYPE_SID].pid;
1611 struct pid_namespace;
1614 * the helpers to get the task's different pids as they are seen
1615 * from various namespaces
1617 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
1618 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1619 * current.
1620 * task_xid_nr_ns() : id seen from the ns specified;
1622 * set_task_vxid() : assigns a virtual id to a task;
1624 * see also pid_nr() etc in include/linux/pid.h
1626 pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1627 struct pid_namespace *ns);
1629 static inline pid_t task_pid_nr(struct task_struct *tsk)
1631 return tsk->pid;
1634 static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1635 struct pid_namespace *ns)
1637 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1640 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1642 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
1646 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1648 return tsk->tgid;
1651 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1653 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1655 return pid_vnr(task_tgid(tsk));
1659 static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1660 struct pid_namespace *ns)
1662 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
1665 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1667 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
1671 static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1672 struct pid_namespace *ns)
1674 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
1677 static inline pid_t task_session_vnr(struct task_struct *tsk)
1679 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
1682 /* obsolete, do not use */
1683 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1685 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1689 * pid_alive - check that a task structure is not stale
1690 * @p: Task structure to be checked.
1692 * Test if a process is not yet dead (at most zombie state)
1693 * If pid_alive fails, then pointers within the task structure
1694 * can be stale and must not be dereferenced.
1696 static inline int pid_alive(struct task_struct *p)
1698 return p->pids[PIDTYPE_PID].pid != NULL;
1702 * is_global_init - check if a task structure is init
1703 * @tsk: Task structure to be checked.
1705 * Check if a task structure is the first user space task the kernel created.
1707 static inline int is_global_init(struct task_struct *tsk)
1709 return tsk->pid == 1;
1713 * is_container_init:
1714 * check whether in the task is init in its own pid namespace.
1716 extern int is_container_init(struct task_struct *tsk);
1718 extern struct pid *cad_pid;
1720 extern void free_task(struct task_struct *tsk);
1721 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1723 extern void __put_task_struct(struct task_struct *t);
1725 static inline void put_task_struct(struct task_struct *t)
1727 if (atomic_dec_and_test(&t->usage))
1728 __put_task_struct(t);
1731 extern cputime_t task_utime(struct task_struct *p);
1732 extern cputime_t task_stime(struct task_struct *p);
1733 extern cputime_t task_gtime(struct task_struct *p);
1734 extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
1737 * Per process flags
1739 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1740 /* Not implemented yet, only for 486*/
1741 #define PF_STARTING 0x00000002 /* being created */
1742 #define PF_EXITING 0x00000004 /* getting shut down */
1743 #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
1744 #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1745 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1746 #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1747 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1748 #define PF_DUMPCORE 0x00000200 /* dumped core */
1749 #define PF_SIGNALED 0x00000400 /* killed by a signal */
1750 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
1751 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1752 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1753 #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1754 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1755 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
1756 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1757 #define PF_KSWAPD 0x00040000 /* I am kswapd */
1758 #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1759 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
1760 #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
1761 #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1762 #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1763 #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1764 #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
1765 #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
1766 #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
1767 #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1768 #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
1769 #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
1770 #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1773 * Only the _current_ task can read/write to tsk->flags, but other
1774 * tasks can access tsk->flags in readonly mode for example
1775 * with tsk_used_math (like during threaded core dumping).
1776 * There is however an exception to this rule during ptrace
1777 * or during fork: the ptracer task is allowed to write to the
1778 * child->flags of its traced child (same goes for fork, the parent
1779 * can write to the child->flags), because we're guaranteed the
1780 * child is not running and in turn not changing child->flags
1781 * at the same time the parent does it.
1783 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1784 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1785 #define clear_used_math() clear_stopped_child_used_math(current)
1786 #define set_used_math() set_stopped_child_used_math(current)
1787 #define conditional_stopped_child_used_math(condition, child) \
1788 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1789 #define conditional_used_math(condition) \
1790 conditional_stopped_child_used_math(condition, current)
1791 #define copy_to_stopped_child_used_math(child) \
1792 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1793 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1794 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1795 #define used_math() tsk_used_math(current)
1797 #ifdef CONFIG_TREE_PREEMPT_RCU
1799 #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1800 #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1802 static inline void rcu_copy_process(struct task_struct *p)
1804 p->rcu_read_lock_nesting = 0;
1805 p->rcu_read_unlock_special = 0;
1806 p->rcu_blocked_node = NULL;
1807 INIT_LIST_HEAD(&p->rcu_node_entry);
1810 #else
1812 static inline void rcu_copy_process(struct task_struct *p)
1816 #endif
1818 #ifdef CONFIG_SMP
1819 extern int set_cpus_allowed_ptr(struct task_struct *p,
1820 const struct cpumask *new_mask);
1821 #else
1822 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1823 const struct cpumask *new_mask)
1825 if (!cpumask_test_cpu(0, new_mask))
1826 return -EINVAL;
1827 return 0;
1829 #endif
1831 #ifndef CONFIG_CPUMASK_OFFSTACK
1832 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1834 return set_cpus_allowed_ptr(p, &new_mask);
1836 #endif
1839 * Architectures can set this to 1 if they have specified
1840 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1841 * but then during bootup it turns out that sched_clock()
1842 * is reliable after all:
1844 #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1845 extern int sched_clock_stable;
1846 #endif
1848 extern unsigned long long sched_clock(void);
1850 extern void sched_clock_init(void);
1851 extern u64 sched_clock_cpu(int cpu);
1853 #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1854 static inline void sched_clock_tick(void)
1858 static inline void sched_clock_idle_sleep_event(void)
1862 static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1865 #else
1866 extern void sched_clock_tick(void);
1867 extern void sched_clock_idle_sleep_event(void);
1868 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1869 #endif
1872 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1873 * clock constructed from sched_clock():
1875 extern unsigned long long cpu_clock(int cpu);
1877 extern unsigned long long
1878 task_sched_runtime(struct task_struct *task);
1879 extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1881 /* sched_exec is called by processes performing an exec */
1882 #ifdef CONFIG_SMP
1883 extern void sched_exec(void);
1884 #else
1885 #define sched_exec() {}
1886 #endif
1888 extern void sched_clock_idle_sleep_event(void);
1889 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1891 #ifdef CONFIG_HOTPLUG_CPU
1892 extern void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p);
1893 extern void idle_task_exit(void);
1894 #else
1895 static inline void idle_task_exit(void) {}
1896 #endif
1898 extern void sched_idle_next(void);
1900 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1901 extern void wake_up_idle_cpu(int cpu);
1902 #else
1903 static inline void wake_up_idle_cpu(int cpu) { }
1904 #endif
1906 extern unsigned int sysctl_sched_latency;
1907 extern unsigned int sysctl_sched_min_granularity;
1908 extern unsigned int sysctl_sched_wakeup_granularity;
1909 extern unsigned int sysctl_sched_shares_ratelimit;
1910 extern unsigned int sysctl_sched_shares_thresh;
1911 extern unsigned int sysctl_sched_child_runs_first;
1912 #ifdef CONFIG_SCHED_DEBUG
1913 extern unsigned int sysctl_sched_features;
1914 extern unsigned int sysctl_sched_migration_cost;
1915 extern unsigned int sysctl_sched_nr_migrate;
1916 extern unsigned int sysctl_sched_time_avg;
1917 extern unsigned int sysctl_timer_migration;
1919 int sched_nr_latency_handler(struct ctl_table *table, int write,
1920 void __user *buffer, size_t *length,
1921 loff_t *ppos);
1922 #endif
1923 #ifdef CONFIG_SCHED_DEBUG
1924 static inline unsigned int get_sysctl_timer_migration(void)
1926 return sysctl_timer_migration;
1928 #else
1929 static inline unsigned int get_sysctl_timer_migration(void)
1931 return 1;
1933 #endif
1934 extern unsigned int sysctl_sched_rt_period;
1935 extern int sysctl_sched_rt_runtime;
1937 int sched_rt_handler(struct ctl_table *table, int write,
1938 void __user *buffer, size_t *lenp,
1939 loff_t *ppos);
1941 extern unsigned int sysctl_sched_compat_yield;
1943 #ifdef CONFIG_RT_MUTEXES
1944 extern int rt_mutex_getprio(struct task_struct *p);
1945 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1946 extern void rt_mutex_adjust_pi(struct task_struct *p);
1947 #else
1948 static inline int rt_mutex_getprio(struct task_struct *p)
1950 return p->normal_prio;
1952 # define rt_mutex_adjust_pi(p) do { } while (0)
1953 #endif
1955 extern void set_user_nice(struct task_struct *p, long nice);
1956 extern int task_prio(const struct task_struct *p);
1957 extern int task_nice(const struct task_struct *p);
1958 extern int can_nice(const struct task_struct *p, const int nice);
1959 extern int task_curr(const struct task_struct *p);
1960 extern int idle_cpu(int cpu);
1961 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1962 extern int sched_setscheduler_nocheck(struct task_struct *, int,
1963 struct sched_param *);
1964 extern struct task_struct *idle_task(int cpu);
1965 extern struct task_struct *curr_task(int cpu);
1966 extern void set_curr_task(int cpu, struct task_struct *p);
1968 void yield(void);
1971 * The default (Linux) execution domain.
1973 extern struct exec_domain default_exec_domain;
1975 union thread_union {
1976 struct thread_info thread_info;
1977 unsigned long stack[THREAD_SIZE/sizeof(long)];
1980 #ifndef __HAVE_ARCH_KSTACK_END
1981 static inline int kstack_end(void *addr)
1983 /* Reliable end of stack detection:
1984 * Some APM bios versions misalign the stack
1986 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1988 #endif
1990 extern union thread_union init_thread_union;
1991 extern struct task_struct init_task;
1993 extern struct mm_struct init_mm;
1995 extern struct pid_namespace init_pid_ns;
1998 * find a task by one of its numerical ids
2000 * find_task_by_pid_ns():
2001 * finds a task by its pid in the specified namespace
2002 * find_task_by_vpid():
2003 * finds a task by its virtual pid
2005 * see also find_vpid() etc in include/linux/pid.h
2008 extern struct task_struct *find_task_by_vpid(pid_t nr);
2009 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2010 struct pid_namespace *ns);
2012 extern void __set_special_pids(struct pid *pid);
2014 /* per-UID process charging. */
2015 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
2016 static inline struct user_struct *get_uid(struct user_struct *u)
2018 atomic_inc(&u->__count);
2019 return u;
2021 extern void free_uid(struct user_struct *);
2022 extern void release_uids(struct user_namespace *ns);
2024 #include <asm/current.h>
2026 extern void do_timer(unsigned long ticks);
2028 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2029 extern int wake_up_process(struct task_struct *tsk);
2030 extern void wake_up_new_task(struct task_struct *tsk,
2031 unsigned long clone_flags);
2032 #ifdef CONFIG_SMP
2033 extern void kick_process(struct task_struct *tsk);
2034 #else
2035 static inline void kick_process(struct task_struct *tsk) { }
2036 #endif
2037 extern void sched_fork(struct task_struct *p, int clone_flags);
2038 extern void sched_dead(struct task_struct *p);
2040 extern void proc_caches_init(void);
2041 extern void flush_signals(struct task_struct *);
2042 extern void __flush_signals(struct task_struct *);
2043 extern void ignore_signals(struct task_struct *);
2044 extern void flush_signal_handlers(struct task_struct *, int force_default);
2045 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2047 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2049 unsigned long flags;
2050 int ret;
2052 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2053 ret = dequeue_signal(tsk, mask, info);
2054 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2056 return ret;
2059 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2060 sigset_t *mask);
2061 extern void unblock_all_signals(void);
2062 extern void release_task(struct task_struct * p);
2063 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
2064 extern int force_sigsegv(int, struct task_struct *);
2065 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
2066 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
2067 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2068 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
2069 extern int kill_pgrp(struct pid *pid, int sig, int priv);
2070 extern int kill_pid(struct pid *pid, int sig, int priv);
2071 extern int kill_proc_info(int, struct siginfo *, pid_t);
2072 extern int do_notify_parent(struct task_struct *, int);
2073 extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
2074 extern void force_sig(int, struct task_struct *);
2075 extern void force_sig_specific(int, struct task_struct *);
2076 extern int send_sig(int, struct task_struct *, int);
2077 extern void zap_other_threads(struct task_struct *p);
2078 extern struct sigqueue *sigqueue_alloc(void);
2079 extern void sigqueue_free(struct sigqueue *);
2080 extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
2081 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
2082 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2084 static inline int kill_cad_pid(int sig, int priv)
2086 return kill_pid(cad_pid, sig, priv);
2089 /* These can be the second arg to send_sig_info/send_group_sig_info. */
2090 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
2091 #define SEND_SIG_PRIV ((struct siginfo *) 1)
2092 #define SEND_SIG_FORCED ((struct siginfo *) 2)
2094 static inline int is_si_special(const struct siginfo *info)
2096 return info <= SEND_SIG_FORCED;
2100 * True if we are on the alternate signal stack.
2102 static inline int on_sig_stack(unsigned long sp)
2104 #ifdef CONFIG_STACK_GROWSUP
2105 return sp >= current->sas_ss_sp &&
2106 sp - current->sas_ss_sp < current->sas_ss_size;
2107 #else
2108 return sp > current->sas_ss_sp &&
2109 sp - current->sas_ss_sp <= current->sas_ss_size;
2110 #endif
2113 static inline int sas_ss_flags(unsigned long sp)
2115 return (current->sas_ss_size == 0 ? SS_DISABLE
2116 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2120 * Routines for handling mm_structs
2122 extern struct mm_struct * mm_alloc(void);
2124 /* mmdrop drops the mm and the page tables */
2125 extern void __mmdrop(struct mm_struct *);
2126 static inline void mmdrop(struct mm_struct * mm)
2128 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
2129 __mmdrop(mm);
2132 /* mmput gets rid of the mappings and all user-space */
2133 extern void mmput(struct mm_struct *);
2134 /* Grab a reference to a task's mm, if it is not already going away */
2135 extern struct mm_struct *get_task_mm(struct task_struct *task);
2136 /* Remove the current tasks stale references to the old mm_struct */
2137 extern void mm_release(struct task_struct *, struct mm_struct *);
2138 /* Allocate a new mm structure and copy contents from tsk->mm */
2139 extern struct mm_struct *dup_mm(struct task_struct *tsk);
2141 extern int copy_thread(unsigned long, unsigned long, unsigned long,
2142 struct task_struct *, struct pt_regs *);
2143 extern void flush_thread(void);
2144 extern void exit_thread(void);
2146 extern void exit_files(struct task_struct *);
2147 extern void __cleanup_signal(struct signal_struct *);
2148 extern void __cleanup_sighand(struct sighand_struct *);
2150 extern void exit_itimers(struct signal_struct *);
2151 extern void flush_itimer_signals(void);
2153 extern NORET_TYPE void do_group_exit(int);
2155 extern void daemonize(const char *, ...);
2156 extern int allow_signal(int);
2157 extern int disallow_signal(int);
2159 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2160 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
2161 struct task_struct *fork_idle(int);
2163 extern void set_task_comm(struct task_struct *tsk, char *from);
2164 extern char *get_task_comm(char *to, struct task_struct *tsk);
2166 #ifdef CONFIG_SMP
2167 extern void wait_task_context_switch(struct task_struct *p);
2168 extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
2169 #else
2170 static inline void wait_task_context_switch(struct task_struct *p) {}
2171 static inline unsigned long wait_task_inactive(struct task_struct *p,
2172 long match_state)
2174 return 1;
2176 #endif
2178 #define next_task(p) \
2179 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
2181 #define for_each_process(p) \
2182 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2184 extern bool current_is_single_threaded(void);
2187 * Careful: do_each_thread/while_each_thread is a double loop so
2188 * 'break' will not work as expected - use goto instead.
2190 #define do_each_thread(g, t) \
2191 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2193 #define while_each_thread(g, t) \
2194 while ((t = next_thread(t)) != g)
2196 /* de_thread depends on thread_group_leader not being a pid based check */
2197 #define thread_group_leader(p) (p == p->group_leader)
2199 /* Do to the insanities of de_thread it is possible for a process
2200 * to have the pid of the thread group leader without actually being
2201 * the thread group leader. For iteration through the pids in proc
2202 * all we care about is that we have a task with the appropriate
2203 * pid, we don't actually care if we have the right task.
2205 static inline int has_group_leader_pid(struct task_struct *p)
2207 return p->pid == p->tgid;
2210 static inline
2211 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2213 return p1->tgid == p2->tgid;
2216 static inline struct task_struct *next_thread(const struct task_struct *p)
2218 return list_entry_rcu(p->thread_group.next,
2219 struct task_struct, thread_group);
2222 static inline int thread_group_empty(struct task_struct *p)
2224 return list_empty(&p->thread_group);
2227 #define delay_group_leader(p) \
2228 (thread_group_leader(p) && !thread_group_empty(p))
2230 static inline int task_detached(struct task_struct *p)
2232 return p->exit_signal == -1;
2236 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
2237 * subscriptions and synchronises with wait4(). Also used in procfs. Also
2238 * pins the final release of task.io_context. Also protects ->cpuset and
2239 * ->cgroup.subsys[].
2241 * Nests both inside and outside of read_lock(&tasklist_lock).
2242 * It must not be nested with write_lock_irq(&tasklist_lock),
2243 * neither inside nor outside.
2245 static inline void task_lock(struct task_struct *p)
2247 spin_lock(&p->alloc_lock);
2250 static inline void task_unlock(struct task_struct *p)
2252 spin_unlock(&p->alloc_lock);
2255 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2256 unsigned long *flags);
2258 static inline void unlock_task_sighand(struct task_struct *tsk,
2259 unsigned long *flags)
2261 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2264 #ifndef __HAVE_THREAD_FUNCTIONS
2266 #define task_thread_info(task) ((struct thread_info *)(task)->stack)
2267 #define task_stack_page(task) ((task)->stack)
2269 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2271 *task_thread_info(p) = *task_thread_info(org);
2272 task_thread_info(p)->task = p;
2275 static inline unsigned long *end_of_stack(struct task_struct *p)
2277 return (unsigned long *)(task_thread_info(p) + 1);
2280 #endif
2282 static inline int object_is_on_stack(void *obj)
2284 void *stack = task_stack_page(current);
2286 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2289 extern void thread_info_cache_init(void);
2291 #ifdef CONFIG_DEBUG_STACK_USAGE
2292 static inline unsigned long stack_not_used(struct task_struct *p)
2294 unsigned long *n = end_of_stack(p);
2296 do { /* Skip over canary */
2297 n++;
2298 } while (!*n);
2300 return (unsigned long)n - (unsigned long)end_of_stack(p);
2302 #endif
2304 /* set thread flags in other task's structures
2305 * - see asm/thread_info.h for TIF_xxxx flags available
2307 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2309 set_ti_thread_flag(task_thread_info(tsk), flag);
2312 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2314 clear_ti_thread_flag(task_thread_info(tsk), flag);
2317 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2319 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
2322 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2324 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
2327 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2329 return test_ti_thread_flag(task_thread_info(tsk), flag);
2332 static inline void set_tsk_need_resched(struct task_struct *tsk)
2334 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2337 static inline void clear_tsk_need_resched(struct task_struct *tsk)
2339 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2342 static inline int test_tsk_need_resched(struct task_struct *tsk)
2344 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2347 static inline int restart_syscall(void)
2349 set_tsk_thread_flag(current, TIF_SIGPENDING);
2350 return -ERESTARTNOINTR;
2353 static inline int signal_pending(struct task_struct *p)
2355 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2358 static inline int __fatal_signal_pending(struct task_struct *p)
2360 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2363 static inline int fatal_signal_pending(struct task_struct *p)
2365 return signal_pending(p) && __fatal_signal_pending(p);
2368 static inline int signal_pending_state(long state, struct task_struct *p)
2370 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2371 return 0;
2372 if (!signal_pending(p))
2373 return 0;
2375 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2378 static inline int need_resched(void)
2380 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
2384 * cond_resched() and cond_resched_lock(): latency reduction via
2385 * explicit rescheduling in places that are safe. The return
2386 * value indicates whether a reschedule was done in fact.
2387 * cond_resched_lock() will drop the spinlock before scheduling,
2388 * cond_resched_softirq() will enable bhs before scheduling.
2390 extern int _cond_resched(void);
2392 #define cond_resched() ({ \
2393 __might_sleep(__FILE__, __LINE__, 0); \
2394 _cond_resched(); \
2397 extern int __cond_resched_lock(spinlock_t *lock);
2399 #ifdef CONFIG_PREEMPT
2400 #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
2401 #else
2402 #define PREEMPT_LOCK_OFFSET 0
2403 #endif
2405 #define cond_resched_lock(lock) ({ \
2406 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
2407 __cond_resched_lock(lock); \
2410 extern int __cond_resched_softirq(void);
2412 #define cond_resched_softirq() ({ \
2413 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2414 __cond_resched_softirq(); \
2418 * Does a critical section need to be broken due to another
2419 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2420 * but a general need for low latency)
2422 static inline int spin_needbreak(spinlock_t *lock)
2424 #ifdef CONFIG_PREEMPT
2425 return spin_is_contended(lock);
2426 #else
2427 return 0;
2428 #endif
2432 * Thread group CPU time accounting.
2434 void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
2435 void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
2437 static inline void thread_group_cputime_init(struct signal_struct *sig)
2439 sig->cputimer.cputime = INIT_CPUTIME;
2440 spin_lock_init(&sig->cputimer.lock);
2441 sig->cputimer.running = 0;
2444 static inline void thread_group_cputime_free(struct signal_struct *sig)
2449 * Reevaluate whether the task has signals pending delivery.
2450 * Wake the task if so.
2451 * This is required every time the blocked sigset_t changes.
2452 * callers must hold sighand->siglock.
2454 extern void recalc_sigpending_and_wake(struct task_struct *t);
2455 extern void recalc_sigpending(void);
2457 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2460 * Wrappers for p->thread_info->cpu access. No-op on UP.
2462 #ifdef CONFIG_SMP
2464 static inline unsigned int task_cpu(const struct task_struct *p)
2466 return task_thread_info(p)->cpu;
2469 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2471 #else
2473 static inline unsigned int task_cpu(const struct task_struct *p)
2475 return 0;
2478 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2482 #endif /* CONFIG_SMP */
2484 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2486 #ifdef CONFIG_TRACING
2487 extern void
2488 __trace_special(void *__tr, void *__data,
2489 unsigned long arg1, unsigned long arg2, unsigned long arg3);
2490 #else
2491 static inline void
2492 __trace_special(void *__tr, void *__data,
2493 unsigned long arg1, unsigned long arg2, unsigned long arg3)
2496 #endif
2498 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2499 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
2501 extern void normalize_rt_tasks(void);
2503 #ifdef CONFIG_GROUP_SCHED
2505 extern struct task_group init_task_group;
2506 #ifdef CONFIG_USER_SCHED
2507 extern struct task_group root_task_group;
2508 extern void set_tg_uid(struct user_struct *user);
2509 #endif
2511 extern struct task_group *sched_create_group(struct task_group *parent);
2512 extern void sched_destroy_group(struct task_group *tg);
2513 extern void sched_move_task(struct task_struct *tsk);
2514 #ifdef CONFIG_FAIR_GROUP_SCHED
2515 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2516 extern unsigned long sched_group_shares(struct task_group *tg);
2517 #endif
2518 #ifdef CONFIG_RT_GROUP_SCHED
2519 extern int sched_group_set_rt_runtime(struct task_group *tg,
2520 long rt_runtime_us);
2521 extern long sched_group_rt_runtime(struct task_group *tg);
2522 extern int sched_group_set_rt_period(struct task_group *tg,
2523 long rt_period_us);
2524 extern long sched_group_rt_period(struct task_group *tg);
2525 extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
2526 #endif
2527 #endif
2529 extern int task_can_switch_user(struct user_struct *up,
2530 struct task_struct *tsk);
2532 #ifdef CONFIG_TASK_XACCT
2533 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2535 tsk->ioac.rchar += amt;
2538 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2540 tsk->ioac.wchar += amt;
2543 static inline void inc_syscr(struct task_struct *tsk)
2545 tsk->ioac.syscr++;
2548 static inline void inc_syscw(struct task_struct *tsk)
2550 tsk->ioac.syscw++;
2552 #else
2553 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2557 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2561 static inline void inc_syscr(struct task_struct *tsk)
2565 static inline void inc_syscw(struct task_struct *tsk)
2568 #endif
2570 #ifndef TASK_SIZE_OF
2571 #define TASK_SIZE_OF(tsk) TASK_SIZE
2572 #endif
2575 * Call the function if the target task is executing on a CPU right now:
2577 extern void task_oncpu_function_call(struct task_struct *p,
2578 void (*func) (void *info), void *info);
2581 #ifdef CONFIG_MM_OWNER
2582 extern void mm_update_next_owner(struct mm_struct *mm);
2583 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2584 #else
2585 static inline void mm_update_next_owner(struct mm_struct *mm)
2589 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2592 #endif /* CONFIG_MM_OWNER */
2594 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2596 static inline unsigned long task_rlimit(const struct task_struct *tsk,
2597 unsigned int limit)
2599 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2602 static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2603 unsigned int limit)
2605 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2608 static inline unsigned long rlimit(unsigned int limit)
2610 return task_rlimit(current, limit);
2613 static inline unsigned long rlimit_max(unsigned int limit)
2615 return task_rlimit_max(current, limit);
2618 #endif /* __KERNEL__ */
2620 #endif