]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/linux/sched.h
sched/headers: Remove the 'init_pid_ns' prototype from <linux/sched.h>
[mirror_ubuntu-artful-kernel.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
607ca46e 4#include <uapi/linux/sched.h>
b7b3c76a 5
5c228079
DY
6#include <linux/sched/prio.h>
7
b69339ba 8#include <linux/mutex.h>
fb00aca4 9#include <linux/plist.h>
77ba809e 10#include <linux/mm_types_task.h>
1da177e4 11
1da177e4 12#include <linux/sem.h>
ab602f79 13#include <linux/shm.h>
f361bf4a 14#include <linux/signal_types.h>
1da177e4 15#include <linux/pid.h>
1da177e4 16#include <linux/seccomp.h>
05725f7e 17#include <linux/rculist.h>
1da177e4 18
a3b6714e 19#include <linux/resource.h>
a3b6714e 20#include <linux/hrtimer.h>
5c9a8750 21#include <linux/kcov.h>
7c3ab738 22#include <linux/task_io_accounting.h>
9745512c 23#include <linux/latencytop.h>
fd771233 24#include <linux/topology.h>
d4311ff1 25#include <linux/magic.h>
a3b6714e 26
70b8157e
IM
27#include <asm/current.h>
28
c7af7877
IM
29/* task_struct member predeclarations: */
30struct audit_context;
31struct autogroup;
32struct backing_dev_info;
bddd87c7 33struct bio_list;
73c10101 34struct blk_plug;
c7af7877 35struct cfs_rq;
c4ad8f98 36struct filename;
c7af7877
IM
37struct fs_struct;
38struct futex_pi_state;
39struct io_context;
40struct mempolicy;
89076bc3 41struct nameidata;
c7af7877
IM
42struct nsproxy;
43struct perf_event_context;
44struct pid_namespace;
45struct pipe_inode_info;
46struct rcu_node;
47struct reclaim_state;
48struct robust_list_head;
49struct sched_attr;
50struct sched_param;
43ae34cb 51struct seq_file;
c7af7877
IM
52struct sighand_struct;
53struct signal_struct;
54struct task_delay_info;
4cf86d77 55struct task_group;
c7af7877
IM
56struct task_struct;
57struct uts_namespace;
1da177e4 58
4a8342d2
LT
59/*
60 * Task state bitmask. NOTE! These bits are also
61 * encoded in fs/proc/array.c: get_task_state().
62 *
63 * We have two separate sets of flags: task->state
64 * is about runnability, while task->exit_state are
65 * about the task exiting. Confusing, but this way
66 * modifying one set can't modify the other one by
67 * mistake.
68 */
1da177e4
LT
69#define TASK_RUNNING 0
70#define TASK_INTERRUPTIBLE 1
71#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
72#define __TASK_STOPPED 4
73#define __TASK_TRACED 8
4a8342d2 74/* in tsk->exit_state */
ad86622b
ON
75#define EXIT_DEAD 16
76#define EXIT_ZOMBIE 32
abd50b39 77#define EXIT_TRACE (EXIT_ZOMBIE | EXIT_DEAD)
4a8342d2 78/* in tsk->state again */
af927232 79#define TASK_DEAD 64
f021a3c2 80#define TASK_WAKEKILL 128
e9c84311 81#define TASK_WAKING 256
f2530dc7 82#define TASK_PARKED 512
80ed87c8 83#define TASK_NOLOAD 1024
7dc603c9
PZ
84#define TASK_NEW 2048
85#define TASK_STATE_MAX 4096
f021a3c2 86
7dc603c9 87#define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWPNn"
73342151 88
642fa448 89/* Convenience macros for the sake of set_current_state */
f021a3c2
MW
90#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
91#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
92#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 93
80ed87c8
PZ
94#define TASK_IDLE (TASK_UNINTERRUPTIBLE | TASK_NOLOAD)
95
92a1f4bc
MW
96/* Convenience macros for the sake of wake_up */
97#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 98#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
99
100/* get_task_state() */
101#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2 102 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
74e37200 103 __TASK_TRACED | EXIT_ZOMBIE | EXIT_DEAD)
92a1f4bc 104
f021a3c2
MW
105#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
106#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 107#define task_is_stopped_or_traced(task) \
f021a3c2 108 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 109#define task_contributes_to_load(task) \
e3c8ca83 110 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
80ed87c8
PZ
111 (task->flags & PF_FROZEN) == 0 && \
112 (task->state & TASK_NOLOAD) == 0)
1da177e4 113
8eb23b9f
PZ
114#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
115
8eb23b9f
PZ
116#define __set_current_state(state_value) \
117 do { \
118 current->task_state_change = _THIS_IP_; \
119 current->state = (state_value); \
120 } while (0)
121#define set_current_state(state_value) \
122 do { \
123 current->task_state_change = _THIS_IP_; \
a2250238 124 smp_store_mb(current->state, (state_value)); \
8eb23b9f
PZ
125 } while (0)
126
127#else
498d0c57
AM
128/*
129 * set_current_state() includes a barrier so that the write of current->state
130 * is correctly serialised wrt the caller's subsequent test of whether to
131 * actually sleep:
132 *
a2250238 133 * for (;;) {
498d0c57 134 * set_current_state(TASK_UNINTERRUPTIBLE);
a2250238
PZ
135 * if (!need_sleep)
136 * break;
137 *
138 * schedule();
139 * }
140 * __set_current_state(TASK_RUNNING);
141 *
142 * If the caller does not need such serialisation (because, for instance, the
143 * condition test and condition change and wakeup are under the same lock) then
144 * use __set_current_state().
145 *
146 * The above is typically ordered against the wakeup, which does:
147 *
148 * need_sleep = false;
149 * wake_up_state(p, TASK_UNINTERRUPTIBLE);
150 *
151 * Where wake_up_state() (and all other wakeup primitives) imply enough
152 * barriers to order the store of the variable against wakeup.
153 *
154 * Wakeup will do: if (@state & p->state) p->state = TASK_RUNNING, that is,
155 * once it observes the TASK_UNINTERRUPTIBLE store the waking CPU can issue a
156 * TASK_RUNNING store which can collide with __set_current_state(TASK_RUNNING).
498d0c57 157 *
a2250238 158 * This is obviously fine, since they both store the exact same value.
498d0c57 159 *
a2250238 160 * Also see the comments of try_to_wake_up().
498d0c57 161 */
8eb23b9f 162#define __set_current_state(state_value) \
1da177e4 163 do { current->state = (state_value); } while (0)
8eb23b9f 164#define set_current_state(state_value) \
b92b8b35 165 smp_store_mb(current->state, (state_value))
1da177e4 166
8eb23b9f
PZ
167#endif
168
1da177e4
LT
169/* Task command name length */
170#define TASK_COMM_LEN 16
171
3fa0818b
RR
172extern cpumask_var_t cpu_isolated_map;
173
1da177e4
LT
174extern void scheduler_tick(void);
175
1da177e4 176#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 177extern signed long schedule_timeout(signed long timeout);
64ed93a2 178extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 179extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 180extern signed long schedule_timeout_uninterruptible(signed long timeout);
69b27baf 181extern signed long schedule_timeout_idle(signed long timeout);
1da177e4 182asmlinkage void schedule(void);
c5491ea7 183extern void schedule_preempt_disabled(void);
1da177e4 184
10ab5643
TH
185extern int __must_check io_schedule_prepare(void);
186extern void io_schedule_finish(int token);
9cff8ade 187extern long io_schedule_timeout(long timeout);
10ab5643 188extern void io_schedule(void);
9cff8ade 189
d37f761d 190/**
9d7fb042 191 * struct prev_cputime - snaphsot of system and user cputime
d37f761d
FW
192 * @utime: time spent in user mode
193 * @stime: time spent in system mode
9d7fb042 194 * @lock: protects the above two fields
d37f761d 195 *
9d7fb042
PZ
196 * Stores previous user/system time values such that we can guarantee
197 * monotonicity.
d37f761d 198 */
9d7fb042
PZ
199struct prev_cputime {
200#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
5613fda9
FW
201 u64 utime;
202 u64 stime;
9d7fb042
PZ
203 raw_spinlock_t lock;
204#endif
d37f761d
FW
205};
206
f06febc9
FM
207/**
208 * struct task_cputime - collected CPU time counts
5613fda9
FW
209 * @utime: time spent in user mode, in nanoseconds
210 * @stime: time spent in kernel mode, in nanoseconds
f06febc9 211 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 212 *
9d7fb042
PZ
213 * This structure groups together three kinds of CPU time that are tracked for
214 * threads and thread groups. Most things considering CPU time want to group
215 * these counts together and treat all three of them in parallel.
f06febc9
FM
216 */
217struct task_cputime {
5613fda9
FW
218 u64 utime;
219 u64 stime;
f06febc9
FM
220 unsigned long long sum_exec_runtime;
221};
9d7fb042 222
f06febc9 223/* Alternate field names when used to cache expirations. */
f06febc9 224#define virt_exp utime
9d7fb042 225#define prof_exp stime
f06febc9
FM
226#define sched_exp sum_exec_runtime
227
f6db8347 228#ifdef CONFIG_SCHED_INFO
1da177e4
LT
229struct sched_info {
230 /* cumulative counters */
2d72376b 231 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 232 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
233
234 /* timestamps */
172ba844
BS
235 unsigned long long last_arrival,/* when we last ran on a cpu */
236 last_queued; /* when we were last queued to run */
1da177e4 237};
f6db8347 238#endif /* CONFIG_SCHED_INFO */
1da177e4 239
6ecdd749
YD
240/*
241 * Integer metrics need fixed point arithmetic, e.g., sched/fair
242 * has a few: load, load_avg, util_avg, freq, and capacity.
243 *
244 * We define a basic fixed point arithmetic range, and then formalize
245 * all these metrics based on that basic range.
246 */
247# define SCHED_FIXEDPOINT_SHIFT 10
248# define SCHED_FIXEDPOINT_SCALE (1L << SCHED_FIXEDPOINT_SHIFT)
249
383f2835 250#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 251extern void prefetch_stack(struct task_struct *t);
383f2835
KC
252#else
253static inline void prefetch_stack(struct task_struct *t) { }
254#endif
1da177e4 255
20b8a59f 256struct load_weight {
9dbdb155
PZ
257 unsigned long weight;
258 u32 inv_weight;
20b8a59f
IM
259};
260
9d89c257 261/*
7b595334
YD
262 * The load_avg/util_avg accumulates an infinite geometric series
263 * (see __update_load_avg() in kernel/sched/fair.c).
264 *
265 * [load_avg definition]
266 *
267 * load_avg = runnable% * scale_load_down(load)
268 *
269 * where runnable% is the time ratio that a sched_entity is runnable.
270 * For cfs_rq, it is the aggregated load_avg of all runnable and
9d89c257 271 * blocked sched_entities.
7b595334
YD
272 *
273 * load_avg may also take frequency scaling into account:
274 *
275 * load_avg = runnable% * scale_load_down(load) * freq%
276 *
277 * where freq% is the CPU frequency normalized to the highest frequency.
278 *
279 * [util_avg definition]
280 *
281 * util_avg = running% * SCHED_CAPACITY_SCALE
282 *
283 * where running% is the time ratio that a sched_entity is running on
284 * a CPU. For cfs_rq, it is the aggregated util_avg of all runnable
285 * and blocked sched_entities.
286 *
287 * util_avg may also factor frequency scaling and CPU capacity scaling:
288 *
289 * util_avg = running% * SCHED_CAPACITY_SCALE * freq% * capacity%
290 *
291 * where freq% is the same as above, and capacity% is the CPU capacity
292 * normalized to the greatest capacity (due to uarch differences, etc).
293 *
294 * N.B., the above ratios (runnable%, running%, freq%, and capacity%)
295 * themselves are in the range of [0, 1]. To do fixed point arithmetics,
296 * we therefore scale them to as large a range as necessary. This is for
297 * example reflected by util_avg's SCHED_CAPACITY_SCALE.
298 *
299 * [Overflow issue]
300 *
301 * The 64-bit load_sum can have 4353082796 (=2^64/47742/88761) entities
302 * with the highest load (=88761), always runnable on a single cfs_rq,
303 * and should not overflow as the number already hits PID_MAX_LIMIT.
304 *
305 * For all other cases (including 32-bit kernels), struct load_weight's
306 * weight will overflow first before we do, because:
307 *
308 * Max(load_avg) <= Max(load.weight)
309 *
310 * Then it is the load_weight's responsibility to consider overflow
311 * issues.
9d89c257 312 */
9d85f21c 313struct sched_avg {
9d89c257
YD
314 u64 last_update_time, load_sum;
315 u32 util_sum, period_contrib;
316 unsigned long load_avg, util_avg;
9d85f21c
PT
317};
318
94c18227 319#ifdef CONFIG_SCHEDSTATS
41acab88 320struct sched_statistics {
20b8a59f 321 u64 wait_start;
94c18227 322 u64 wait_max;
6d082592
AV
323 u64 wait_count;
324 u64 wait_sum;
8f0dfc34
AV
325 u64 iowait_count;
326 u64 iowait_sum;
94c18227 327
20b8a59f 328 u64 sleep_start;
20b8a59f 329 u64 sleep_max;
94c18227
IM
330 s64 sum_sleep_runtime;
331
332 u64 block_start;
20b8a59f
IM
333 u64 block_max;
334 u64 exec_max;
eba1ed4b 335 u64 slice_max;
cc367732 336
cc367732
IM
337 u64 nr_migrations_cold;
338 u64 nr_failed_migrations_affine;
339 u64 nr_failed_migrations_running;
340 u64 nr_failed_migrations_hot;
341 u64 nr_forced_migrations;
cc367732
IM
342
343 u64 nr_wakeups;
344 u64 nr_wakeups_sync;
345 u64 nr_wakeups_migrate;
346 u64 nr_wakeups_local;
347 u64 nr_wakeups_remote;
348 u64 nr_wakeups_affine;
349 u64 nr_wakeups_affine_attempts;
350 u64 nr_wakeups_passive;
351 u64 nr_wakeups_idle;
41acab88
LDM
352};
353#endif
354
355struct sched_entity {
356 struct load_weight load; /* for load-balancing */
357 struct rb_node run_node;
358 struct list_head group_node;
359 unsigned int on_rq;
360
361 u64 exec_start;
362 u64 sum_exec_runtime;
363 u64 vruntime;
364 u64 prev_sum_exec_runtime;
365
41acab88
LDM
366 u64 nr_migrations;
367
41acab88
LDM
368#ifdef CONFIG_SCHEDSTATS
369 struct sched_statistics statistics;
94c18227
IM
370#endif
371
20b8a59f 372#ifdef CONFIG_FAIR_GROUP_SCHED
fed14d45 373 int depth;
20b8a59f
IM
374 struct sched_entity *parent;
375 /* rq on which this entity is (to be) queued: */
376 struct cfs_rq *cfs_rq;
377 /* rq "owned" by this entity/group: */
378 struct cfs_rq *my_q;
379#endif
8bd75c77 380
141965c7 381#ifdef CONFIG_SMP
5a107804
JO
382 /*
383 * Per entity load average tracking.
384 *
385 * Put into separate cache line so it does not
386 * collide with read-mostly values above.
387 */
388 struct sched_avg avg ____cacheline_aligned_in_smp;
9d85f21c 389#endif
20b8a59f 390};
70b97a7f 391
fa717060
PZ
392struct sched_rt_entity {
393 struct list_head run_list;
78f2c7db 394 unsigned long timeout;
57d2aa00 395 unsigned long watchdog_stamp;
bee367ed 396 unsigned int time_slice;
ff77e468
PZ
397 unsigned short on_rq;
398 unsigned short on_list;
6f505b16 399
58d6c2d7 400 struct sched_rt_entity *back;
052f1dc7 401#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
402 struct sched_rt_entity *parent;
403 /* rq on which this entity is (to be) queued: */
404 struct rt_rq *rt_rq;
405 /* rq "owned" by this entity/group: */
406 struct rt_rq *my_q;
407#endif
fa717060
PZ
408};
409
aab03e05
DF
410struct sched_dl_entity {
411 struct rb_node rb_node;
412
413 /*
414 * Original scheduling parameters. Copied here from sched_attr
4027d080 415 * during sched_setattr(), they will remain the same until
416 * the next sched_setattr().
aab03e05
DF
417 */
418 u64 dl_runtime; /* maximum runtime for each instance */
419 u64 dl_deadline; /* relative deadline of each instance */
755378a4 420 u64 dl_period; /* separation of two instances (period) */
332ac17e 421 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
422
423 /*
424 * Actual scheduling parameters. Initialized with the values above,
425 * they are continously updated during task execution. Note that
426 * the remaining runtime could be < 0 in case we are in overrun.
427 */
428 s64 runtime; /* remaining runtime for this instance */
429 u64 deadline; /* absolute deadline for this instance */
430 unsigned int flags; /* specifying the scheduler behaviour */
431
432 /*
433 * Some bool flags:
434 *
435 * @dl_throttled tells if we exhausted the runtime. If so, the
436 * task has to wait for a replenishment to be performed at the
437 * next firing of dl_timer.
438 *
2d3d891d
DF
439 * @dl_boosted tells if we are boosted due to DI. If so we are
440 * outside bandwidth enforcement mechanism (but only until we
5bfd126e
JL
441 * exit the critical section);
442 *
443 * @dl_yielded tells if task gave up the cpu before consuming
444 * all its available runtime during the last job.
aab03e05 445 */
72f9f3fd 446 int dl_throttled, dl_boosted, dl_yielded;
aab03e05
DF
447
448 /*
449 * Bandwidth enforcement timer. Each -deadline task has its
450 * own bandwidth to be enforced, thus we need one timer per task.
451 */
452 struct hrtimer dl_timer;
453};
8bd75c77 454
1d082fd0
PM
455union rcu_special {
456 struct {
8203d6d0
PM
457 u8 blocked;
458 u8 need_qs;
459 u8 exp_need_qs;
460 u8 pad; /* Otherwise the compiler can store garbage here. */
461 } b; /* Bits. */
462 u32 s; /* Set of bits. */
1d082fd0 463};
86848966 464
8dc85d54
PZ
465enum perf_event_task_context {
466 perf_invalid_context = -1,
467 perf_hw_context = 0,
89a1e187 468 perf_sw_context,
8dc85d54
PZ
469 perf_nr_task_contexts,
470};
471
eb61baf6
IM
472struct wake_q_node {
473 struct wake_q_node *next;
474};
475
1da177e4 476struct task_struct {
c65eacbe
AL
477#ifdef CONFIG_THREAD_INFO_IN_TASK
478 /*
479 * For reasons of header soup (see current_thread_info()), this
480 * must be the first element of task_struct.
481 */
482 struct thread_info thread_info;
483#endif
1da177e4 484 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 485 void *stack;
1da177e4 486 atomic_t usage;
97dc32cd
WC
487 unsigned int flags; /* per process flags, defined below */
488 unsigned int ptrace;
1da177e4 489
2dd73a4f 490#ifdef CONFIG_SMP
fa14ff4a 491 struct llist_node wake_entry;
3ca7a440 492 int on_cpu;
c65eacbe
AL
493#ifdef CONFIG_THREAD_INFO_IN_TASK
494 unsigned int cpu; /* current CPU */
495#endif
63b0e9ed 496 unsigned int wakee_flips;
62470419 497 unsigned long wakee_flip_decay_ts;
63b0e9ed 498 struct task_struct *last_wakee;
ac66f547
PZ
499
500 int wake_cpu;
2dd73a4f 501#endif
fd2f4419 502 int on_rq;
50e645a8 503
b29739f9 504 int prio, static_prio, normal_prio;
c7aceaba 505 unsigned int rt_priority;
5522d5d5 506 const struct sched_class *sched_class;
20b8a59f 507 struct sched_entity se;
fa717060 508 struct sched_rt_entity rt;
8323f26c
PZ
509#ifdef CONFIG_CGROUP_SCHED
510 struct task_group *sched_task_group;
511#endif
aab03e05 512 struct sched_dl_entity dl;
1da177e4 513
e107be36
AK
514#ifdef CONFIG_PREEMPT_NOTIFIERS
515 /* list of struct preempt_notifier: */
516 struct hlist_head preempt_notifiers;
517#endif
518
6c5c9341 519#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 520 unsigned int btrace_seq;
6c5c9341 521#endif
1da177e4 522
97dc32cd 523 unsigned int policy;
29baa747 524 int nr_cpus_allowed;
1da177e4 525 cpumask_t cpus_allowed;
1da177e4 526
a57eb940 527#ifdef CONFIG_PREEMPT_RCU
e260be67 528 int rcu_read_lock_nesting;
1d082fd0 529 union rcu_special rcu_read_unlock_special;
f41d911f 530 struct list_head rcu_node_entry;
a57eb940 531 struct rcu_node *rcu_blocked_node;
28f6569a 532#endif /* #ifdef CONFIG_PREEMPT_RCU */
8315f422
PM
533#ifdef CONFIG_TASKS_RCU
534 unsigned long rcu_tasks_nvcsw;
535 bool rcu_tasks_holdout;
536 struct list_head rcu_tasks_holdout_list;
176f8f7a 537 int rcu_tasks_idle_cpu;
8315f422 538#endif /* #ifdef CONFIG_TASKS_RCU */
e260be67 539
f6db8347 540#ifdef CONFIG_SCHED_INFO
1da177e4
LT
541 struct sched_info sched_info;
542#endif
543
544 struct list_head tasks;
806c09a7 545#ifdef CONFIG_SMP
917b627d 546 struct plist_node pushable_tasks;
1baca4ce 547 struct rb_node pushable_dl_tasks;
806c09a7 548#endif
1da177e4
LT
549
550 struct mm_struct *mm, *active_mm;
314ff785
IM
551
552 /* Per-thread vma caching: */
553 struct vmacache vmacache;
554
34e55232
KH
555#if defined(SPLIT_RSS_COUNTING)
556 struct task_rss_stat rss_stat;
557#endif
1da177e4 558/* task state */
97dc32cd 559 int exit_state;
1da177e4
LT
560 int exit_code, exit_signal;
561 int pdeath_signal; /* The signal sent when the parent dies */
e7cc4173 562 unsigned long jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
563
564 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 565 unsigned int personality;
9b89f6ba 566
be958bdc 567 /* scheduler bits, serialized by scheduler locks */
ca94c442 568 unsigned sched_reset_on_fork:1;
a8e4f2ea 569 unsigned sched_contributes_to_load:1;
ff303e66 570 unsigned sched_migrated:1;
b7e7ade3 571 unsigned sched_remote_wakeup:1;
be958bdc
PZ
572 unsigned :0; /* force alignment to the next boundary */
573
574 /* unserialized, strictly 'current' */
575 unsigned in_execve:1; /* bit to tell LSMs we're in execve */
576 unsigned in_iowait:1;
7e781418
AL
577#if !defined(TIF_RESTORE_SIGMASK)
578 unsigned restore_sigmask:1;
579#endif
626ebc41
TH
580#ifdef CONFIG_MEMCG
581 unsigned memcg_may_oom:1;
127424c8 582#ifndef CONFIG_SLOB
6f185c29
VD
583 unsigned memcg_kmem_skip_account:1;
584#endif
127424c8 585#endif
ff303e66
PZ
586#ifdef CONFIG_COMPAT_BRK
587 unsigned brk_randomized:1;
588#endif
6f185c29 589
1d4457f9
KC
590 unsigned long atomic_flags; /* Flags needing atomic access. */
591
f56141e3
AL
592 struct restart_block restart_block;
593
1da177e4
LT
594 pid_t pid;
595 pid_t tgid;
0a425405 596
1314562a 597#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
598 /* Canary value for the -fstack-protector gcc feature */
599 unsigned long stack_canary;
1314562a 600#endif
4d1d61a6 601 /*
1da177e4 602 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 603 * older sibling, respectively. (p->father can be replaced with
f470021a 604 * p->real_parent->pid)
1da177e4 605 */
abd63bc3
KC
606 struct task_struct __rcu *real_parent; /* real parent process */
607 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 608 /*
f470021a 609 * children/sibling forms the list of my natural children
1da177e4
LT
610 */
611 struct list_head children; /* list of my children */
612 struct list_head sibling; /* linkage in my parent's children list */
613 struct task_struct *group_leader; /* threadgroup leader */
614
f470021a
RM
615 /*
616 * ptraced is the list of tasks this task is using ptrace on.
617 * This includes both natural children and PTRACE_ATTACH targets.
618 * p->ptrace_entry is p's link on the p->parent->ptraced list.
619 */
620 struct list_head ptraced;
621 struct list_head ptrace_entry;
622
1da177e4 623 /* PID/PID hash table linkage. */
92476d7f 624 struct pid_link pids[PIDTYPE_MAX];
47e65328 625 struct list_head thread_group;
0c740d0a 626 struct list_head thread_node;
1da177e4
LT
627
628 struct completion *vfork_done; /* for vfork() */
629 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
630 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
631
5613fda9 632 u64 utime, stime;
40565b5a 633#ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME
5613fda9 634 u64 utimescaled, stimescaled;
40565b5a 635#endif
16a6d9be 636 u64 gtime;
9d7fb042 637 struct prev_cputime prev_cputime;
6a61671b 638#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
b7ce2277 639 seqcount_t vtime_seqcount;
6a61671b
FW
640 unsigned long long vtime_snap;
641 enum {
7098c1ea
FW
642 /* Task is sleeping or running in a CPU with VTIME inactive */
643 VTIME_INACTIVE = 0,
644 /* Task runs in userspace in a CPU with VTIME active */
6a61671b 645 VTIME_USER,
7098c1ea 646 /* Task runs in kernelspace in a CPU with VTIME active */
6a61671b
FW
647 VTIME_SYS,
648 } vtime_snap_whence;
d99ca3b9 649#endif
d027d45d
FW
650
651#ifdef CONFIG_NO_HZ_FULL
f009a7a7 652 atomic_t tick_dep_mask;
d027d45d 653#endif
1da177e4 654 unsigned long nvcsw, nivcsw; /* context switch counts */
ccbf62d8 655 u64 start_time; /* monotonic time in nsec */
57e0be04 656 u64 real_start_time; /* boot based time in nsec */
1da177e4
LT
657/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
658 unsigned long min_flt, maj_flt;
659
b18b6a9c 660#ifdef CONFIG_POSIX_TIMERS
f06febc9 661 struct task_cputime cputime_expires;
1da177e4 662 struct list_head cpu_timers[3];
b18b6a9c 663#endif
1da177e4
LT
664
665/* process credentials */
64b875f7 666 const struct cred __rcu *ptracer_cred; /* Tracer's credentials at attach */
1b0ba1c9 667 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 668 * credentials (COW) */
1b0ba1c9 669 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 670 * credentials (COW) */
36772092
PBG
671 char comm[TASK_COMM_LEN]; /* executable name excluding path
672 - access with [gs]et_task_comm (which lock
673 it with task_lock())
221af7f8 674 - initialized normally by setup_new_exec */
1da177e4 675/* file system info */
756daf26 676 struct nameidata *nameidata;
3d5b6fcc 677#ifdef CONFIG_SYSVIPC
1da177e4
LT
678/* ipc stuff */
679 struct sysv_sem sysvsem;
ab602f79 680 struct sysv_shm sysvshm;
3d5b6fcc 681#endif
e162b39a 682#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 683/* hung task detection */
82a1fcb9
IM
684 unsigned long last_switch_count;
685#endif
1da177e4
LT
686/* filesystem information */
687 struct fs_struct *fs;
688/* open file information */
689 struct files_struct *files;
1651e14e 690/* namespaces */
ab516013 691 struct nsproxy *nsproxy;
1da177e4
LT
692/* signal handlers */
693 struct signal_struct *signal;
694 struct sighand_struct *sighand;
695
696 sigset_t blocked, real_blocked;
f3de272b 697 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
698 struct sigpending pending;
699
700 unsigned long sas_ss_sp;
701 size_t sas_ss_size;
2a742138 702 unsigned sas_ss_flags;
2e01fabe 703
67d12145 704 struct callback_head *task_works;
e73f8959 705
1da177e4 706 struct audit_context *audit_context;
bfef93a5 707#ifdef CONFIG_AUDITSYSCALL
e1760bd5 708 kuid_t loginuid;
4746ec5b 709 unsigned int sessionid;
bfef93a5 710#endif
932ecebb 711 struct seccomp seccomp;
1da177e4
LT
712
713/* Thread group tracking */
714 u32 parent_exec_id;
715 u32 self_exec_id;
58568d2a
MX
716/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
717 * mempolicy */
1da177e4 718 spinlock_t alloc_lock;
1da177e4 719
b29739f9 720 /* Protection of the PI data structures: */
1d615482 721 raw_spinlock_t pi_lock;
b29739f9 722
76751049
PZ
723 struct wake_q_node wake_q;
724
23f78d4a
IM
725#ifdef CONFIG_RT_MUTEXES
726 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
727 struct rb_root pi_waiters;
728 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
729 /* Deadlock detection and priority inheritance handling */
730 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
731#endif
732
408894ee
IM
733#ifdef CONFIG_DEBUG_MUTEXES
734 /* mutex deadlock detection */
735 struct mutex_waiter *blocked_on;
736#endif
de30a2b3
IM
737#ifdef CONFIG_TRACE_IRQFLAGS
738 unsigned int irq_events;
de30a2b3 739 unsigned long hardirq_enable_ip;
de30a2b3 740 unsigned long hardirq_disable_ip;
fa1452e8 741 unsigned int hardirq_enable_event;
de30a2b3 742 unsigned int hardirq_disable_event;
fa1452e8
HS
743 int hardirqs_enabled;
744 int hardirq_context;
de30a2b3 745 unsigned long softirq_disable_ip;
de30a2b3 746 unsigned long softirq_enable_ip;
fa1452e8 747 unsigned int softirq_disable_event;
de30a2b3 748 unsigned int softirq_enable_event;
fa1452e8 749 int softirqs_enabled;
de30a2b3
IM
750 int softirq_context;
751#endif
fbb9ce95 752#ifdef CONFIG_LOCKDEP
bdb9441e 753# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
754 u64 curr_chain_key;
755 int lockdep_depth;
fbb9ce95 756 unsigned int lockdep_recursion;
c7aceaba 757 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 758 gfp_t lockdep_reclaim_gfp;
fbb9ce95 759#endif
c6d30853
AR
760#ifdef CONFIG_UBSAN
761 unsigned int in_ubsan;
762#endif
408894ee 763
1da177e4
LT
764/* journalling filesystem info */
765 void *journal_info;
766
d89d8796 767/* stacked block device info */
bddd87c7 768 struct bio_list *bio_list;
d89d8796 769
73c10101
JA
770#ifdef CONFIG_BLOCK
771/* stack plugging */
772 struct blk_plug *plug;
773#endif
774
1da177e4
LT
775/* VM state */
776 struct reclaim_state *reclaim_state;
777
1da177e4
LT
778 struct backing_dev_info *backing_dev_info;
779
780 struct io_context *io_context;
781
782 unsigned long ptrace_message;
783 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 784 struct task_io_accounting ioac;
8f0ab514 785#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
786 u64 acct_rss_mem1; /* accumulated rss usage */
787 u64 acct_vm_mem1; /* accumulated virtual memory usage */
605dc2b3 788 u64 acct_timexpd; /* stime + utime since last update */
1da177e4
LT
789#endif
790#ifdef CONFIG_CPUSETS
58568d2a 791 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 792 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 793 int cpuset_mem_spread_rotor;
6adef3eb 794 int cpuset_slab_spread_rotor;
1da177e4 795#endif
ddbcc7e8 796#ifdef CONFIG_CGROUPS
817929ec 797 /* Control Group info protected by css_set_lock */
2c392b8c 798 struct css_set __rcu *cgroups;
817929ec
PM
799 /* cg_list protected by css_set_lock and tsk->alloc_lock */
800 struct list_head cg_list;
ddbcc7e8 801#endif
e02737d5
FY
802#ifdef CONFIG_INTEL_RDT_A
803 int closid;
804#endif
42b2dd0a 805#ifdef CONFIG_FUTEX
0771dfef 806 struct robust_list_head __user *robust_list;
34f192c6
IM
807#ifdef CONFIG_COMPAT
808 struct compat_robust_list_head __user *compat_robust_list;
809#endif
c87e2837
IM
810 struct list_head pi_state_list;
811 struct futex_pi_state *pi_state_cache;
c7aceaba 812#endif
cdd6c482 813#ifdef CONFIG_PERF_EVENTS
8dc85d54 814 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
815 struct mutex perf_event_mutex;
816 struct list_head perf_event_list;
a63eaf34 817#endif
8f47b187
TG
818#ifdef CONFIG_DEBUG_PREEMPT
819 unsigned long preempt_disable_ip;
820#endif
c7aceaba 821#ifdef CONFIG_NUMA
58568d2a 822 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 823 short il_next;
207205a2 824 short pref_node_fork;
42b2dd0a 825#endif
cbee9f88
PZ
826#ifdef CONFIG_NUMA_BALANCING
827 int numa_scan_seq;
cbee9f88 828 unsigned int numa_scan_period;
598f0ec0 829 unsigned int numa_scan_period_max;
de1c9ce6 830 int numa_preferred_nid;
6b9a7460 831 unsigned long numa_migrate_retry;
cbee9f88 832 u64 node_stamp; /* migration stamp */
7e2703e6
RR
833 u64 last_task_numa_placement;
834 u64 last_sum_exec_runtime;
cbee9f88 835 struct callback_head numa_work;
f809ca9a 836
8c8a743c
PZ
837 struct list_head numa_entry;
838 struct numa_group *numa_group;
839
745d6147 840 /*
44dba3d5
IM
841 * numa_faults is an array split into four regions:
842 * faults_memory, faults_cpu, faults_memory_buffer, faults_cpu_buffer
843 * in this precise order.
844 *
845 * faults_memory: Exponential decaying average of faults on a per-node
846 * basis. Scheduling placement decisions are made based on these
847 * counts. The values remain static for the duration of a PTE scan.
848 * faults_cpu: Track the nodes the process was running on when a NUMA
849 * hinting fault was incurred.
850 * faults_memory_buffer and faults_cpu_buffer: Record faults per node
851 * during the current scan window. When the scan completes, the counts
852 * in faults_memory and faults_cpu decay and these values are copied.
745d6147 853 */
44dba3d5 854 unsigned long *numa_faults;
83e1d2cd 855 unsigned long total_numa_faults;
745d6147 856
04bb2f94
RR
857 /*
858 * numa_faults_locality tracks if faults recorded during the last
074c2381
MG
859 * scan window were remote/local or failed to migrate. The task scan
860 * period is adapted based on the locality of the faults with different
861 * weights depending on whether they were shared or private faults
04bb2f94 862 */
074c2381 863 unsigned long numa_faults_locality[3];
04bb2f94 864
b32e86b4 865 unsigned long numa_pages_migrated;
cbee9f88
PZ
866#endif /* CONFIG_NUMA_BALANCING */
867
72b252ae 868 struct tlbflush_unmap_batch tlb_ubc;
72b252ae 869
e56d0903 870 struct rcu_head rcu;
b92ce558
JA
871
872 /*
873 * cache last used pipe for splice
874 */
875 struct pipe_inode_info *splice_pipe;
5640f768
ED
876
877 struct page_frag task_frag;
878
47913d4e
IM
879#ifdef CONFIG_TASK_DELAY_ACCT
880 struct task_delay_info *delays;
f4f154fd 881#endif
47913d4e 882
f4f154fd
AM
883#ifdef CONFIG_FAULT_INJECTION
884 int make_it_fail;
ca74e92b 885#endif
9d823e8f
WF
886 /*
887 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
888 * balance_dirty_pages() for some dirty throttling pause
889 */
890 int nr_dirtied;
891 int nr_dirtied_pause;
83712358 892 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 893
9745512c
AV
894#ifdef CONFIG_LATENCYTOP
895 int latency_record_count;
896 struct latency_record latency_record[LT_SAVECOUNT];
897#endif
6976675d
AV
898 /*
899 * time slack values; these are used to round up poll() and
900 * select() etc timeout values. These are in nanoseconds.
901 */
da8b44d5
JS
902 u64 timer_slack_ns;
903 u64 default_timer_slack_ns;
f8d570a4 904
0b24becc
AR
905#ifdef CONFIG_KASAN
906 unsigned int kasan_depth;
907#endif
fb52607a 908#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 909 /* Index of current stored address in ret_stack */
f201ae23
FW
910 int curr_ret_stack;
911 /* Stack of return addresses for return function tracing */
912 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
913 /* time stamp for last schedule */
914 unsigned long long ftrace_timestamp;
f201ae23
FW
915 /*
916 * Number of functions that haven't been traced
917 * because of depth overrun.
918 */
919 atomic_t trace_overrun;
380c4b14
FW
920 /* Pause for the tracing */
921 atomic_t tracing_graph_pause;
f201ae23 922#endif
ea4e2bc4
SR
923#ifdef CONFIG_TRACING
924 /* state flags for use by tracers */
925 unsigned long trace;
b1cff0ad 926 /* bitmask and counter of trace recursion */
261842b7
SR
927 unsigned long trace_recursion;
928#endif /* CONFIG_TRACING */
5c9a8750
DV
929#ifdef CONFIG_KCOV
930 /* Coverage collection mode enabled for this task (0 if disabled). */
931 enum kcov_mode kcov_mode;
932 /* Size of the kcov_area. */
933 unsigned kcov_size;
934 /* Buffer for coverage collection. */
935 void *kcov_area;
936 /* kcov desciptor wired with this task or NULL. */
937 struct kcov *kcov;
938#endif
6f185c29 939#ifdef CONFIG_MEMCG
626ebc41
TH
940 struct mem_cgroup *memcg_in_oom;
941 gfp_t memcg_oom_gfp_mask;
942 int memcg_oom_order;
b23afb93
TH
943
944 /* number of pages to reclaim on returning to userland */
945 unsigned int memcg_nr_pages_over_high;
569b846d 946#endif
0326f5a9
SD
947#ifdef CONFIG_UPROBES
948 struct uprobe_task *utask;
0326f5a9 949#endif
cafe5635
KO
950#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
951 unsigned int sequential_io;
952 unsigned int sequential_io_avg;
953#endif
8eb23b9f
PZ
954#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
955 unsigned long task_state_change;
956#endif
8bcbde54 957 int pagefault_disabled;
03049269 958#ifdef CONFIG_MMU
29c696e1 959 struct task_struct *oom_reaper_list;
03049269 960#endif
ba14a194
AL
961#ifdef CONFIG_VMAP_STACK
962 struct vm_struct *stack_vm_area;
963#endif
68f24b08
AL
964#ifdef CONFIG_THREAD_INFO_IN_TASK
965 /* A live task holds one reference. */
966 atomic_t stack_refcount;
967#endif
0c8c0f03
DH
968/* CPU-specific state of this task */
969 struct thread_struct thread;
970/*
971 * WARNING: on x86, 'thread_struct' contains a variable-sized
972 * structure. It *MUST* be at the end of 'task_struct'.
973 *
974 * Do not put anything below here!
975 */
1da177e4
LT
976};
977
e868171a 978static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
979{
980 return task->pids[PIDTYPE_PID].pid;
981}
982
e868171a 983static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
984{
985 return task->group_leader->pids[PIDTYPE_PID].pid;
986}
987
6dda81f4
ON
988/*
989 * Without tasklist or rcu lock it is not safe to dereference
990 * the result of task_pgrp/task_session even if task == current,
991 * we can race with another thread doing sys_setsid/sys_setpgid.
992 */
e868171a 993static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
994{
995 return task->group_leader->pids[PIDTYPE_PGID].pid;
996}
997
e868171a 998static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
999{
1000 return task->group_leader->pids[PIDTYPE_SID].pid;
1001}
1002
7af57294
PE
1003/*
1004 * the helpers to get the task's different pids as they are seen
1005 * from various namespaces
1006 *
1007 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1008 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1009 * current.
7af57294
PE
1010 * task_xid_nr_ns() : id seen from the ns specified;
1011 *
1012 * set_task_vxid() : assigns a virtual id to a task;
1013 *
7af57294
PE
1014 * see also pid_nr() etc in include/linux/pid.h
1015 */
52ee2dfd
ON
1016pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1017 struct pid_namespace *ns);
7af57294 1018
e868171a 1019static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1020{
1021 return tsk->pid;
1022}
1023
52ee2dfd
ON
1024static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1025 struct pid_namespace *ns)
1026{
1027 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1028}
7af57294
PE
1029
1030static inline pid_t task_pid_vnr(struct task_struct *tsk)
1031{
52ee2dfd 1032 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1033}
1034
1035
e868171a 1036static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1037{
1038 return tsk->tgid;
1039}
1040
2f2a3a46 1041pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1042
1043static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1044{
1045 return pid_vnr(task_tgid(tsk));
1046}
1047
1048
80e0b6e8 1049static inline int pid_alive(const struct task_struct *p);
ad36d282
RGB
1050static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns)
1051{
1052 pid_t pid = 0;
1053
1054 rcu_read_lock();
1055 if (pid_alive(tsk))
1056 pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns);
1057 rcu_read_unlock();
1058
1059 return pid;
1060}
1061
1062static inline pid_t task_ppid_nr(const struct task_struct *tsk)
1063{
1064 return task_ppid_nr_ns(tsk, &init_pid_ns);
1065}
1066
52ee2dfd
ON
1067static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1068 struct pid_namespace *ns)
7af57294 1069{
52ee2dfd 1070 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1071}
1072
7af57294
PE
1073static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1074{
52ee2dfd 1075 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1076}
1077
1078
52ee2dfd
ON
1079static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1080 struct pid_namespace *ns)
7af57294 1081{
52ee2dfd 1082 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1083}
1084
7af57294
PE
1085static inline pid_t task_session_vnr(struct task_struct *tsk)
1086{
52ee2dfd 1087 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1088}
1089
1b0f7ffd
ON
1090/* obsolete, do not use */
1091static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1092{
1093 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1094}
7af57294 1095
1da177e4
LT
1096/**
1097 * pid_alive - check that a task structure is not stale
1098 * @p: Task structure to be checked.
1099 *
1100 * Test if a process is not yet dead (at most zombie state)
1101 * If pid_alive fails, then pointers within the task structure
1102 * can be stale and must not be dereferenced.
e69f6186
YB
1103 *
1104 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1105 */
ad36d282 1106static inline int pid_alive(const struct task_struct *p)
1da177e4 1107{
92476d7f 1108 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1109}
1110
f400e198 1111/**
570f5241
SS
1112 * is_global_init - check if a task structure is init. Since init
1113 * is free to have sub-threads we need to check tgid.
3260259f
HK
1114 * @tsk: Task structure to be checked.
1115 *
1116 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1117 *
1118 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1119 */
e868171a 1120static inline int is_global_init(struct task_struct *tsk)
b461cc03 1121{
570f5241 1122 return task_tgid_nr(tsk) == 1;
b461cc03 1123}
b460cbc5 1124
9ec52099
CLG
1125extern struct pid *cad_pid;
1126
1da177e4
LT
1127/*
1128 * Per process flags
1129 */
c1de45ca 1130#define PF_IDLE 0x00000002 /* I am an IDLE thread */
1da177e4 1131#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1132#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1133#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1134#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1135#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1136#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1137#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1138#define PF_DUMPCORE 0x00000200 /* dumped core */
1139#define PF_SIGNALED 0x00000400 /* killed by a signal */
1140#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1141#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1142#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1143#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1144#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1145#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1146#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1147#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1148#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1149#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1150#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1151#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1152#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
14a40ffc 1153#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1154#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
61a87122 1155#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1156#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1157#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1158
1159/*
1160 * Only the _current_ task can read/write to tsk->flags, but other
1161 * tasks can access tsk->flags in readonly mode for example
1162 * with tsk_used_math (like during threaded core dumping).
1163 * There is however an exception to this rule during ptrace
1164 * or during fork: the ptracer task is allowed to write to the
1165 * child->flags of its traced child (same goes for fork, the parent
1166 * can write to the child->flags), because we're guaranteed the
1167 * child is not running and in turn not changing child->flags
1168 * at the same time the parent does it.
1169 */
1170#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1171#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1172#define clear_used_math() clear_stopped_child_used_math(current)
1173#define set_used_math() set_stopped_child_used_math(current)
1174#define conditional_stopped_child_used_math(condition, child) \
1175 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1176#define conditional_used_math(condition) \
1177 conditional_stopped_child_used_math(condition, current)
1178#define copy_to_stopped_child_used_math(child) \
1179 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1180/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1181#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1182#define used_math() tsk_used_math(current)
1183
1d4457f9 1184/* Per-process atomic flags. */
a2b86f77 1185#define PFA_NO_NEW_PRIVS 0 /* May not gain new privileges. */
2ad654bc
ZL
1186#define PFA_SPREAD_PAGE 1 /* Spread page cache over cpuset */
1187#define PFA_SPREAD_SLAB 2 /* Spread some slab caches over cpuset */
77ed2c57 1188#define PFA_LMK_WAITING 3 /* Lowmemorykiller is waiting */
2ad654bc 1189
1d4457f9 1190
e0e5070b
ZL
1191#define TASK_PFA_TEST(name, func) \
1192 static inline bool task_##func(struct task_struct *p) \
1193 { return test_bit(PFA_##name, &p->atomic_flags); }
1194#define TASK_PFA_SET(name, func) \
1195 static inline void task_set_##func(struct task_struct *p) \
1196 { set_bit(PFA_##name, &p->atomic_flags); }
1197#define TASK_PFA_CLEAR(name, func) \
1198 static inline void task_clear_##func(struct task_struct *p) \
1199 { clear_bit(PFA_##name, &p->atomic_flags); }
1200
1201TASK_PFA_TEST(NO_NEW_PRIVS, no_new_privs)
1202TASK_PFA_SET(NO_NEW_PRIVS, no_new_privs)
1d4457f9 1203
2ad654bc
ZL
1204TASK_PFA_TEST(SPREAD_PAGE, spread_page)
1205TASK_PFA_SET(SPREAD_PAGE, spread_page)
1206TASK_PFA_CLEAR(SPREAD_PAGE, spread_page)
1207
1208TASK_PFA_TEST(SPREAD_SLAB, spread_slab)
1209TASK_PFA_SET(SPREAD_SLAB, spread_slab)
1210TASK_PFA_CLEAR(SPREAD_SLAB, spread_slab)
1d4457f9 1211
77ed2c57
TH
1212TASK_PFA_TEST(LMK_WAITING, lmk_waiting)
1213TASK_PFA_SET(LMK_WAITING, lmk_waiting)
1214
907aed48
MG
1215static inline void tsk_restore_flags(struct task_struct *task,
1216 unsigned long orig_flags, unsigned long flags)
1217{
1218 task->flags &= ~flags;
1219 task->flags |= orig_flags & flags;
1220}
1221
f82f8042
JL
1222extern int cpuset_cpumask_can_shrink(const struct cpumask *cur,
1223 const struct cpumask *trial);
7f51412a
JL
1224extern int task_can_attach(struct task_struct *p,
1225 const struct cpumask *cs_cpus_allowed);
1da177e4 1226#ifdef CONFIG_SMP
1e1b6c51
KM
1227extern void do_set_cpus_allowed(struct task_struct *p,
1228 const struct cpumask *new_mask);
1229
cd8ba7cd 1230extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1231 const struct cpumask *new_mask);
1da177e4 1232#else
1e1b6c51
KM
1233static inline void do_set_cpus_allowed(struct task_struct *p,
1234 const struct cpumask *new_mask)
1235{
1236}
cd8ba7cd 1237static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1238 const struct cpumask *new_mask)
1da177e4 1239{
96f874e2 1240 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1241 return -EINVAL;
1242 return 0;
1243}
1244#endif
e0ad9556 1245
6d0d2878
CB
1246#ifndef cpu_relax_yield
1247#define cpu_relax_yield() cpu_relax()
1248#endif
1249
fa93384f 1250extern int yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
1251extern void set_user_nice(struct task_struct *p, long nice);
1252extern int task_prio(const struct task_struct *p);
d0ea0268
DY
1253/**
1254 * task_nice - return the nice value of a given task.
1255 * @p: the task in question.
1256 *
1257 * Return: The nice value [ -20 ... 0 ... 19 ].
1258 */
1259static inline int task_nice(const struct task_struct *p)
1260{
1261 return PRIO_TO_NICE((p)->static_prio);
1262}
36c8b586
IM
1263extern int can_nice(const struct task_struct *p, const int nice);
1264extern int task_curr(const struct task_struct *p);
1da177e4 1265extern int idle_cpu(int cpu);
fe7de49f
KM
1266extern int sched_setscheduler(struct task_struct *, int,
1267 const struct sched_param *);
961ccddd 1268extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 1269 const struct sched_param *);
d50dde5a
DF
1270extern int sched_setattr(struct task_struct *,
1271 const struct sched_attr *);
36c8b586 1272extern struct task_struct *idle_task(int cpu);
c4f30608
PM
1273/**
1274 * is_idle_task - is the specified task an idle task?
fa757281 1275 * @p: the task in question.
e69f6186
YB
1276 *
1277 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 1278 */
7061ca3b 1279static inline bool is_idle_task(const struct task_struct *p)
c4f30608 1280{
c1de45ca 1281 return !!(p->flags & PF_IDLE);
c4f30608 1282}
36c8b586 1283extern struct task_struct *curr_task(int cpu);
a458ae2e 1284extern void ia64_set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1285
1286void yield(void);
1287
1da177e4 1288union thread_union {
c65eacbe 1289#ifndef CONFIG_THREAD_INFO_IN_TASK
1da177e4 1290 struct thread_info thread_info;
c65eacbe 1291#endif
1da177e4
LT
1292 unsigned long stack[THREAD_SIZE/sizeof(long)];
1293};
1294
f3ac6067
IM
1295#ifdef CONFIG_THREAD_INFO_IN_TASK
1296static inline struct thread_info *task_thread_info(struct task_struct *task)
1297{
1298 return &task->thread_info;
1299}
1300#elif !defined(__HAVE_THREAD_FUNCTIONS)
1301# define task_thread_info(task) ((struct thread_info *)(task)->stack)
1302#endif
1303
198fe21b
PE
1304/*
1305 * find a task by one of its numerical ids
1306 *
198fe21b
PE
1307 * find_task_by_pid_ns():
1308 * finds a task by its pid in the specified namespace
228ebcbe
PE
1309 * find_task_by_vpid():
1310 * finds a task by its virtual pid
198fe21b 1311 *
e49859e7 1312 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
1313 */
1314
228ebcbe
PE
1315extern struct task_struct *find_task_by_vpid(pid_t nr);
1316extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1317 struct pid_namespace *ns);
198fe21b 1318
b3c97528
HH
1319extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1320extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 1321extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
1322#ifdef CONFIG_SMP
1323 extern void kick_process(struct task_struct *tsk);
1324#else
1325 static inline void kick_process(struct task_struct *tsk) { }
1326#endif
1da177e4 1327
82b89778
AH
1328extern void __set_task_comm(struct task_struct *tsk, const char *from, bool exec);
1329static inline void set_task_comm(struct task_struct *tsk, const char *from)
1330{
1331 __set_task_comm(tsk, from, false);
1332}
59714d65 1333extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
1334
1335#ifdef CONFIG_SMP
317f3941 1336void scheduler_ipi(void);
85ba2d86 1337extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 1338#else
184748cc 1339static inline void scheduler_ipi(void) { }
85ba2d86
RM
1340static inline unsigned long wait_task_inactive(struct task_struct *p,
1341 long match_state)
1342{
1343 return 1;
1344}
1da177e4
LT
1345#endif
1346
1da177e4
LT
1347/* set thread flags in other task's structures
1348 * - see asm/thread_info.h for TIF_xxxx flags available
1349 */
1350static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1351{
a1261f54 1352 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1353}
1354
1355static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1356{
a1261f54 1357 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1358}
1359
1360static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1361{
a1261f54 1362 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1363}
1364
1365static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1366{
a1261f54 1367 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1368}
1369
1370static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1371{
a1261f54 1372 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1373}
1374
1375static inline void set_tsk_need_resched(struct task_struct *tsk)
1376{
1377 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1378}
1379
1380static inline void clear_tsk_need_resched(struct task_struct *tsk)
1381{
1382 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1383}
1384
8ae121ac
GH
1385static inline int test_tsk_need_resched(struct task_struct *tsk)
1386{
1387 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
1388}
1389
1da177e4
LT
1390/*
1391 * cond_resched() and cond_resched_lock(): latency reduction via
1392 * explicit rescheduling in places that are safe. The return
1393 * value indicates whether a reschedule was done in fact.
1394 * cond_resched_lock() will drop the spinlock before scheduling,
1395 * cond_resched_softirq() will enable bhs before scheduling.
1396 */
35a773a0 1397#ifndef CONFIG_PREEMPT
c3921ab7 1398extern int _cond_resched(void);
35a773a0
PZ
1399#else
1400static inline int _cond_resched(void) { return 0; }
1401#endif
6f80bd98 1402
613afbf8 1403#define cond_resched() ({ \
3427445a 1404 ___might_sleep(__FILE__, __LINE__, 0); \
613afbf8
FW
1405 _cond_resched(); \
1406})
6f80bd98 1407
613afbf8
FW
1408extern int __cond_resched_lock(spinlock_t *lock);
1409
1410#define cond_resched_lock(lock) ({ \
3427445a 1411 ___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\
613afbf8
FW
1412 __cond_resched_lock(lock); \
1413})
1414
1415extern int __cond_resched_softirq(void);
1416
75e1056f 1417#define cond_resched_softirq() ({ \
3427445a 1418 ___might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
75e1056f 1419 __cond_resched_softirq(); \
613afbf8 1420})
1da177e4 1421
f6f3c437
SH
1422static inline void cond_resched_rcu(void)
1423{
1424#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
1425 rcu_read_unlock();
1426 cond_resched();
1427 rcu_read_lock();
1428#endif
1429}
1430
1da177e4
LT
1431/*
1432 * Does a critical section need to be broken due to another
95c354fe
NP
1433 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
1434 * but a general need for low latency)
1da177e4 1435 */
95c354fe 1436static inline int spin_needbreak(spinlock_t *lock)
1da177e4 1437{
95c354fe
NP
1438#ifdef CONFIG_PREEMPT
1439 return spin_is_contended(lock);
1440#else
1da177e4 1441 return 0;
95c354fe 1442#endif
1da177e4
LT
1443}
1444
75f93fed
PZ
1445static __always_inline bool need_resched(void)
1446{
1447 return unlikely(tif_need_resched());
1448}
1449
1da177e4
LT
1450/*
1451 * Wrappers for p->thread_info->cpu access. No-op on UP.
1452 */
1453#ifdef CONFIG_SMP
1454
1455static inline unsigned int task_cpu(const struct task_struct *p)
1456{
c65eacbe
AL
1457#ifdef CONFIG_THREAD_INFO_IN_TASK
1458 return p->cpu;
1459#else
a1261f54 1460 return task_thread_info(p)->cpu;
c65eacbe 1461#endif
1da177e4
LT
1462}
1463
b32e86b4
IM
1464static inline int task_node(const struct task_struct *p)
1465{
1466 return cpu_to_node(task_cpu(p));
1467}
1468
c65cc870 1469extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
1470
1471#else
1472
1473static inline unsigned int task_cpu(const struct task_struct *p)
1474{
1475 return 0;
1476}
1477
1478static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1479{
1480}
1481
1482#endif /* CONFIG_SMP */
1483
d9345c65
PX
1484/*
1485 * In order to reduce various lock holder preemption latencies provide an
1486 * interface to see if a vCPU is currently running or not.
1487 *
1488 * This allows us to terminate optimistic spin loops and block, analogous to
1489 * the native optimistic spin heuristic of testing if the lock owner task is
1490 * running or not.
1491 */
1492#ifndef vcpu_is_preempted
1493# define vcpu_is_preempted(cpu) false
1494#endif
1495
96f874e2
RR
1496extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
1497extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 1498
82455257
DH
1499#ifndef TASK_SIZE_OF
1500#define TASK_SIZE_OF(tsk) TASK_SIZE
1501#endif
1502
1da177e4 1503#endif