]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/linux/sched.h
sched/headers: Remove <linux/cgroup-defs.h> 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#include <asm/ptrace.h>
1da177e4 12
1da177e4 13#include <linux/sem.h>
ab602f79 14#include <linux/shm.h>
1da177e4 15#include <linux/signal.h>
f361bf4a 16#include <linux/signal_types.h>
1da177e4 17#include <linux/pid.h>
1da177e4 18#include <linux/seccomp.h>
05725f7e 19#include <linux/rculist.h>
23f78d4a 20#include <linux/rtmutex.h>
1da177e4 21
a3b6714e 22#include <linux/resource.h>
a3b6714e 23#include <linux/hrtimer.h>
5c9a8750 24#include <linux/kcov.h>
7c3ab738 25#include <linux/task_io_accounting.h>
9745512c 26#include <linux/latencytop.h>
9e2b2dc4 27#include <linux/cred.h>
21caf2fc 28#include <linux/gfp.h>
fd771233 29#include <linux/topology.h>
d4311ff1 30#include <linux/magic.h>
a3b6714e 31
70b8157e
IM
32#include <asm/current.h>
33
c7af7877
IM
34/* task_struct member predeclarations: */
35struct audit_context;
36struct autogroup;
37struct backing_dev_info;
bddd87c7 38struct bio_list;
73c10101 39struct blk_plug;
c7af7877 40struct cfs_rq;
c4ad8f98 41struct filename;
c7af7877
IM
42struct fs_struct;
43struct futex_pi_state;
44struct io_context;
45struct mempolicy;
89076bc3 46struct nameidata;
c7af7877
IM
47struct nsproxy;
48struct perf_event_context;
49struct pid_namespace;
50struct pipe_inode_info;
51struct rcu_node;
52struct reclaim_state;
53struct robust_list_head;
54struct sched_attr;
55struct sched_param;
43ae34cb 56struct seq_file;
c7af7877
IM
57struct sighand_struct;
58struct signal_struct;
59struct task_delay_info;
4cf86d77 60struct task_group;
c7af7877
IM
61struct task_struct;
62struct uts_namespace;
1da177e4 63
4a8342d2
LT
64/*
65 * Task state bitmask. NOTE! These bits are also
66 * encoded in fs/proc/array.c: get_task_state().
67 *
68 * We have two separate sets of flags: task->state
69 * is about runnability, while task->exit_state are
70 * about the task exiting. Confusing, but this way
71 * modifying one set can't modify the other one by
72 * mistake.
73 */
1da177e4
LT
74#define TASK_RUNNING 0
75#define TASK_INTERRUPTIBLE 1
76#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
77#define __TASK_STOPPED 4
78#define __TASK_TRACED 8
4a8342d2 79/* in tsk->exit_state */
ad86622b
ON
80#define EXIT_DEAD 16
81#define EXIT_ZOMBIE 32
abd50b39 82#define EXIT_TRACE (EXIT_ZOMBIE | EXIT_DEAD)
4a8342d2 83/* in tsk->state again */
af927232 84#define TASK_DEAD 64
f021a3c2 85#define TASK_WAKEKILL 128
e9c84311 86#define TASK_WAKING 256
f2530dc7 87#define TASK_PARKED 512
80ed87c8 88#define TASK_NOLOAD 1024
7dc603c9
PZ
89#define TASK_NEW 2048
90#define TASK_STATE_MAX 4096
f021a3c2 91
7dc603c9 92#define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWPNn"
73342151 93
642fa448 94/* Convenience macros for the sake of set_current_state */
f021a3c2
MW
95#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
96#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
97#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 98
80ed87c8
PZ
99#define TASK_IDLE (TASK_UNINTERRUPTIBLE | TASK_NOLOAD)
100
92a1f4bc
MW
101/* Convenience macros for the sake of wake_up */
102#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 103#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
104
105/* get_task_state() */
106#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2 107 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
74e37200 108 __TASK_TRACED | EXIT_ZOMBIE | EXIT_DEAD)
92a1f4bc 109
f021a3c2
MW
110#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
111#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 112#define task_is_stopped_or_traced(task) \
f021a3c2 113 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 114#define task_contributes_to_load(task) \
e3c8ca83 115 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
80ed87c8
PZ
116 (task->flags & PF_FROZEN) == 0 && \
117 (task->state & TASK_NOLOAD) == 0)
1da177e4 118
8eb23b9f
PZ
119#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
120
8eb23b9f
PZ
121#define __set_current_state(state_value) \
122 do { \
123 current->task_state_change = _THIS_IP_; \
124 current->state = (state_value); \
125 } while (0)
126#define set_current_state(state_value) \
127 do { \
128 current->task_state_change = _THIS_IP_; \
a2250238 129 smp_store_mb(current->state, (state_value)); \
8eb23b9f
PZ
130 } while (0)
131
132#else
498d0c57
AM
133/*
134 * set_current_state() includes a barrier so that the write of current->state
135 * is correctly serialised wrt the caller's subsequent test of whether to
136 * actually sleep:
137 *
a2250238 138 * for (;;) {
498d0c57 139 * set_current_state(TASK_UNINTERRUPTIBLE);
a2250238
PZ
140 * if (!need_sleep)
141 * break;
142 *
143 * schedule();
144 * }
145 * __set_current_state(TASK_RUNNING);
146 *
147 * If the caller does not need such serialisation (because, for instance, the
148 * condition test and condition change and wakeup are under the same lock) then
149 * use __set_current_state().
150 *
151 * The above is typically ordered against the wakeup, which does:
152 *
153 * need_sleep = false;
154 * wake_up_state(p, TASK_UNINTERRUPTIBLE);
155 *
156 * Where wake_up_state() (and all other wakeup primitives) imply enough
157 * barriers to order the store of the variable against wakeup.
158 *
159 * Wakeup will do: if (@state & p->state) p->state = TASK_RUNNING, that is,
160 * once it observes the TASK_UNINTERRUPTIBLE store the waking CPU can issue a
161 * TASK_RUNNING store which can collide with __set_current_state(TASK_RUNNING).
498d0c57 162 *
a2250238 163 * This is obviously fine, since they both store the exact same value.
498d0c57 164 *
a2250238 165 * Also see the comments of try_to_wake_up().
498d0c57 166 */
8eb23b9f 167#define __set_current_state(state_value) \
1da177e4 168 do { current->state = (state_value); } while (0)
8eb23b9f 169#define set_current_state(state_value) \
b92b8b35 170 smp_store_mb(current->state, (state_value))
1da177e4 171
8eb23b9f
PZ
172#endif
173
1da177e4
LT
174/* Task command name length */
175#define TASK_COMM_LEN 16
176
3fa0818b
RR
177extern cpumask_var_t cpu_isolated_map;
178
89f19f04 179extern int runqueue_is_locked(int cpu);
017730c1 180
1da177e4
LT
181extern void scheduler_tick(void);
182
1da177e4 183#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 184extern signed long schedule_timeout(signed long timeout);
64ed93a2 185extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 186extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 187extern signed long schedule_timeout_uninterruptible(signed long timeout);
69b27baf 188extern signed long schedule_timeout_idle(signed long timeout);
1da177e4 189asmlinkage void schedule(void);
c5491ea7 190extern void schedule_preempt_disabled(void);
1da177e4 191
10ab5643
TH
192extern int __must_check io_schedule_prepare(void);
193extern void io_schedule_finish(int token);
9cff8ade 194extern long io_schedule_timeout(long timeout);
10ab5643 195extern void io_schedule(void);
9cff8ade 196
d37f761d 197/**
9d7fb042 198 * struct prev_cputime - snaphsot of system and user cputime
d37f761d
FW
199 * @utime: time spent in user mode
200 * @stime: time spent in system mode
9d7fb042 201 * @lock: protects the above two fields
d37f761d 202 *
9d7fb042
PZ
203 * Stores previous user/system time values such that we can guarantee
204 * monotonicity.
d37f761d 205 */
9d7fb042
PZ
206struct prev_cputime {
207#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
5613fda9
FW
208 u64 utime;
209 u64 stime;
9d7fb042
PZ
210 raw_spinlock_t lock;
211#endif
d37f761d
FW
212};
213
f06febc9
FM
214/**
215 * struct task_cputime - collected CPU time counts
5613fda9
FW
216 * @utime: time spent in user mode, in nanoseconds
217 * @stime: time spent in kernel mode, in nanoseconds
f06febc9 218 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 219 *
9d7fb042
PZ
220 * This structure groups together three kinds of CPU time that are tracked for
221 * threads and thread groups. Most things considering CPU time want to group
222 * these counts together and treat all three of them in parallel.
f06febc9
FM
223 */
224struct task_cputime {
5613fda9
FW
225 u64 utime;
226 u64 stime;
f06febc9
FM
227 unsigned long long sum_exec_runtime;
228};
9d7fb042 229
f06febc9 230/* Alternate field names when used to cache expirations. */
f06febc9 231#define virt_exp utime
9d7fb042 232#define prof_exp stime
f06febc9
FM
233#define sched_exp sum_exec_runtime
234
4714d1d3 235#include <linux/rwsem.h>
1da177e4 236
f6db8347 237#ifdef CONFIG_SCHED_INFO
1da177e4
LT
238struct sched_info {
239 /* cumulative counters */
2d72376b 240 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 241 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
242
243 /* timestamps */
172ba844
BS
244 unsigned long long last_arrival,/* when we last ran on a cpu */
245 last_queued; /* when we were last queued to run */
1da177e4 246};
f6db8347 247#endif /* CONFIG_SCHED_INFO */
1da177e4 248
6ecdd749
YD
249/*
250 * Integer metrics need fixed point arithmetic, e.g., sched/fair
251 * has a few: load, load_avg, util_avg, freq, and capacity.
252 *
253 * We define a basic fixed point arithmetic range, and then formalize
254 * all these metrics based on that basic range.
255 */
256# define SCHED_FIXEDPOINT_SHIFT 10
257# define SCHED_FIXEDPOINT_SCALE (1L << SCHED_FIXEDPOINT_SHIFT)
258
383f2835 259#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 260extern void prefetch_stack(struct task_struct *t);
383f2835
KC
261#else
262static inline void prefetch_stack(struct task_struct *t) { }
263#endif
1da177e4 264
20b8a59f 265struct load_weight {
9dbdb155
PZ
266 unsigned long weight;
267 u32 inv_weight;
20b8a59f
IM
268};
269
9d89c257 270/*
7b595334
YD
271 * The load_avg/util_avg accumulates an infinite geometric series
272 * (see __update_load_avg() in kernel/sched/fair.c).
273 *
274 * [load_avg definition]
275 *
276 * load_avg = runnable% * scale_load_down(load)
277 *
278 * where runnable% is the time ratio that a sched_entity is runnable.
279 * For cfs_rq, it is the aggregated load_avg of all runnable and
9d89c257 280 * blocked sched_entities.
7b595334
YD
281 *
282 * load_avg may also take frequency scaling into account:
283 *
284 * load_avg = runnable% * scale_load_down(load) * freq%
285 *
286 * where freq% is the CPU frequency normalized to the highest frequency.
287 *
288 * [util_avg definition]
289 *
290 * util_avg = running% * SCHED_CAPACITY_SCALE
291 *
292 * where running% is the time ratio that a sched_entity is running on
293 * a CPU. For cfs_rq, it is the aggregated util_avg of all runnable
294 * and blocked sched_entities.
295 *
296 * util_avg may also factor frequency scaling and CPU capacity scaling:
297 *
298 * util_avg = running% * SCHED_CAPACITY_SCALE * freq% * capacity%
299 *
300 * where freq% is the same as above, and capacity% is the CPU capacity
301 * normalized to the greatest capacity (due to uarch differences, etc).
302 *
303 * N.B., the above ratios (runnable%, running%, freq%, and capacity%)
304 * themselves are in the range of [0, 1]. To do fixed point arithmetics,
305 * we therefore scale them to as large a range as necessary. This is for
306 * example reflected by util_avg's SCHED_CAPACITY_SCALE.
307 *
308 * [Overflow issue]
309 *
310 * The 64-bit load_sum can have 4353082796 (=2^64/47742/88761) entities
311 * with the highest load (=88761), always runnable on a single cfs_rq,
312 * and should not overflow as the number already hits PID_MAX_LIMIT.
313 *
314 * For all other cases (including 32-bit kernels), struct load_weight's
315 * weight will overflow first before we do, because:
316 *
317 * Max(load_avg) <= Max(load.weight)
318 *
319 * Then it is the load_weight's responsibility to consider overflow
320 * issues.
9d89c257 321 */
9d85f21c 322struct sched_avg {
9d89c257
YD
323 u64 last_update_time, load_sum;
324 u32 util_sum, period_contrib;
325 unsigned long load_avg, util_avg;
9d85f21c
PT
326};
327
94c18227 328#ifdef CONFIG_SCHEDSTATS
41acab88 329struct sched_statistics {
20b8a59f 330 u64 wait_start;
94c18227 331 u64 wait_max;
6d082592
AV
332 u64 wait_count;
333 u64 wait_sum;
8f0dfc34
AV
334 u64 iowait_count;
335 u64 iowait_sum;
94c18227 336
20b8a59f 337 u64 sleep_start;
20b8a59f 338 u64 sleep_max;
94c18227
IM
339 s64 sum_sleep_runtime;
340
341 u64 block_start;
20b8a59f
IM
342 u64 block_max;
343 u64 exec_max;
eba1ed4b 344 u64 slice_max;
cc367732 345
cc367732
IM
346 u64 nr_migrations_cold;
347 u64 nr_failed_migrations_affine;
348 u64 nr_failed_migrations_running;
349 u64 nr_failed_migrations_hot;
350 u64 nr_forced_migrations;
cc367732
IM
351
352 u64 nr_wakeups;
353 u64 nr_wakeups_sync;
354 u64 nr_wakeups_migrate;
355 u64 nr_wakeups_local;
356 u64 nr_wakeups_remote;
357 u64 nr_wakeups_affine;
358 u64 nr_wakeups_affine_attempts;
359 u64 nr_wakeups_passive;
360 u64 nr_wakeups_idle;
41acab88
LDM
361};
362#endif
363
364struct sched_entity {
365 struct load_weight load; /* for load-balancing */
366 struct rb_node run_node;
367 struct list_head group_node;
368 unsigned int on_rq;
369
370 u64 exec_start;
371 u64 sum_exec_runtime;
372 u64 vruntime;
373 u64 prev_sum_exec_runtime;
374
41acab88
LDM
375 u64 nr_migrations;
376
41acab88
LDM
377#ifdef CONFIG_SCHEDSTATS
378 struct sched_statistics statistics;
94c18227
IM
379#endif
380
20b8a59f 381#ifdef CONFIG_FAIR_GROUP_SCHED
fed14d45 382 int depth;
20b8a59f
IM
383 struct sched_entity *parent;
384 /* rq on which this entity is (to be) queued: */
385 struct cfs_rq *cfs_rq;
386 /* rq "owned" by this entity/group: */
387 struct cfs_rq *my_q;
388#endif
8bd75c77 389
141965c7 390#ifdef CONFIG_SMP
5a107804
JO
391 /*
392 * Per entity load average tracking.
393 *
394 * Put into separate cache line so it does not
395 * collide with read-mostly values above.
396 */
397 struct sched_avg avg ____cacheline_aligned_in_smp;
9d85f21c 398#endif
20b8a59f 399};
70b97a7f 400
fa717060
PZ
401struct sched_rt_entity {
402 struct list_head run_list;
78f2c7db 403 unsigned long timeout;
57d2aa00 404 unsigned long watchdog_stamp;
bee367ed 405 unsigned int time_slice;
ff77e468
PZ
406 unsigned short on_rq;
407 unsigned short on_list;
6f505b16 408
58d6c2d7 409 struct sched_rt_entity *back;
052f1dc7 410#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
411 struct sched_rt_entity *parent;
412 /* rq on which this entity is (to be) queued: */
413 struct rt_rq *rt_rq;
414 /* rq "owned" by this entity/group: */
415 struct rt_rq *my_q;
416#endif
fa717060
PZ
417};
418
aab03e05
DF
419struct sched_dl_entity {
420 struct rb_node rb_node;
421
422 /*
423 * Original scheduling parameters. Copied here from sched_attr
4027d080 424 * during sched_setattr(), they will remain the same until
425 * the next sched_setattr().
aab03e05
DF
426 */
427 u64 dl_runtime; /* maximum runtime for each instance */
428 u64 dl_deadline; /* relative deadline of each instance */
755378a4 429 u64 dl_period; /* separation of two instances (period) */
332ac17e 430 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
431
432 /*
433 * Actual scheduling parameters. Initialized with the values above,
434 * they are continously updated during task execution. Note that
435 * the remaining runtime could be < 0 in case we are in overrun.
436 */
437 s64 runtime; /* remaining runtime for this instance */
438 u64 deadline; /* absolute deadline for this instance */
439 unsigned int flags; /* specifying the scheduler behaviour */
440
441 /*
442 * Some bool flags:
443 *
444 * @dl_throttled tells if we exhausted the runtime. If so, the
445 * task has to wait for a replenishment to be performed at the
446 * next firing of dl_timer.
447 *
2d3d891d
DF
448 * @dl_boosted tells if we are boosted due to DI. If so we are
449 * outside bandwidth enforcement mechanism (but only until we
5bfd126e
JL
450 * exit the critical section);
451 *
452 * @dl_yielded tells if task gave up the cpu before consuming
453 * all its available runtime during the last job.
aab03e05 454 */
72f9f3fd 455 int dl_throttled, dl_boosted, dl_yielded;
aab03e05
DF
456
457 /*
458 * Bandwidth enforcement timer. Each -deadline task has its
459 * own bandwidth to be enforced, thus we need one timer per task.
460 */
461 struct hrtimer dl_timer;
462};
8bd75c77 463
1d082fd0
PM
464union rcu_special {
465 struct {
8203d6d0
PM
466 u8 blocked;
467 u8 need_qs;
468 u8 exp_need_qs;
469 u8 pad; /* Otherwise the compiler can store garbage here. */
470 } b; /* Bits. */
471 u32 s; /* Set of bits. */
1d082fd0 472};
86848966 473
8dc85d54
PZ
474enum perf_event_task_context {
475 perf_invalid_context = -1,
476 perf_hw_context = 0,
89a1e187 477 perf_sw_context,
8dc85d54
PZ
478 perf_nr_task_contexts,
479};
480
eb61baf6
IM
481struct wake_q_node {
482 struct wake_q_node *next;
483};
484
1da177e4 485struct task_struct {
c65eacbe
AL
486#ifdef CONFIG_THREAD_INFO_IN_TASK
487 /*
488 * For reasons of header soup (see current_thread_info()), this
489 * must be the first element of task_struct.
490 */
491 struct thread_info thread_info;
492#endif
1da177e4 493 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 494 void *stack;
1da177e4 495 atomic_t usage;
97dc32cd
WC
496 unsigned int flags; /* per process flags, defined below */
497 unsigned int ptrace;
1da177e4 498
2dd73a4f 499#ifdef CONFIG_SMP
fa14ff4a 500 struct llist_node wake_entry;
3ca7a440 501 int on_cpu;
c65eacbe
AL
502#ifdef CONFIG_THREAD_INFO_IN_TASK
503 unsigned int cpu; /* current CPU */
504#endif
63b0e9ed 505 unsigned int wakee_flips;
62470419 506 unsigned long wakee_flip_decay_ts;
63b0e9ed 507 struct task_struct *last_wakee;
ac66f547
PZ
508
509 int wake_cpu;
2dd73a4f 510#endif
fd2f4419 511 int on_rq;
50e645a8 512
b29739f9 513 int prio, static_prio, normal_prio;
c7aceaba 514 unsigned int rt_priority;
5522d5d5 515 const struct sched_class *sched_class;
20b8a59f 516 struct sched_entity se;
fa717060 517 struct sched_rt_entity rt;
8323f26c
PZ
518#ifdef CONFIG_CGROUP_SCHED
519 struct task_group *sched_task_group;
520#endif
aab03e05 521 struct sched_dl_entity dl;
1da177e4 522
e107be36
AK
523#ifdef CONFIG_PREEMPT_NOTIFIERS
524 /* list of struct preempt_notifier: */
525 struct hlist_head preempt_notifiers;
526#endif
527
6c5c9341 528#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 529 unsigned int btrace_seq;
6c5c9341 530#endif
1da177e4 531
97dc32cd 532 unsigned int policy;
29baa747 533 int nr_cpus_allowed;
1da177e4 534 cpumask_t cpus_allowed;
1da177e4 535
a57eb940 536#ifdef CONFIG_PREEMPT_RCU
e260be67 537 int rcu_read_lock_nesting;
1d082fd0 538 union rcu_special rcu_read_unlock_special;
f41d911f 539 struct list_head rcu_node_entry;
a57eb940 540 struct rcu_node *rcu_blocked_node;
28f6569a 541#endif /* #ifdef CONFIG_PREEMPT_RCU */
8315f422
PM
542#ifdef CONFIG_TASKS_RCU
543 unsigned long rcu_tasks_nvcsw;
544 bool rcu_tasks_holdout;
545 struct list_head rcu_tasks_holdout_list;
176f8f7a 546 int rcu_tasks_idle_cpu;
8315f422 547#endif /* #ifdef CONFIG_TASKS_RCU */
e260be67 548
f6db8347 549#ifdef CONFIG_SCHED_INFO
1da177e4
LT
550 struct sched_info sched_info;
551#endif
552
553 struct list_head tasks;
806c09a7 554#ifdef CONFIG_SMP
917b627d 555 struct plist_node pushable_tasks;
1baca4ce 556 struct rb_node pushable_dl_tasks;
806c09a7 557#endif
1da177e4
LT
558
559 struct mm_struct *mm, *active_mm;
314ff785
IM
560
561 /* Per-thread vma caching: */
562 struct vmacache vmacache;
563
34e55232
KH
564#if defined(SPLIT_RSS_COUNTING)
565 struct task_rss_stat rss_stat;
566#endif
1da177e4 567/* task state */
97dc32cd 568 int exit_state;
1da177e4
LT
569 int exit_code, exit_signal;
570 int pdeath_signal; /* The signal sent when the parent dies */
e7cc4173 571 unsigned long jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
572
573 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 574 unsigned int personality;
9b89f6ba 575
be958bdc 576 /* scheduler bits, serialized by scheduler locks */
ca94c442 577 unsigned sched_reset_on_fork:1;
a8e4f2ea 578 unsigned sched_contributes_to_load:1;
ff303e66 579 unsigned sched_migrated:1;
b7e7ade3 580 unsigned sched_remote_wakeup:1;
be958bdc
PZ
581 unsigned :0; /* force alignment to the next boundary */
582
583 /* unserialized, strictly 'current' */
584 unsigned in_execve:1; /* bit to tell LSMs we're in execve */
585 unsigned in_iowait:1;
7e781418
AL
586#if !defined(TIF_RESTORE_SIGMASK)
587 unsigned restore_sigmask:1;
588#endif
626ebc41
TH
589#ifdef CONFIG_MEMCG
590 unsigned memcg_may_oom:1;
127424c8 591#ifndef CONFIG_SLOB
6f185c29
VD
592 unsigned memcg_kmem_skip_account:1;
593#endif
127424c8 594#endif
ff303e66
PZ
595#ifdef CONFIG_COMPAT_BRK
596 unsigned brk_randomized:1;
597#endif
6f185c29 598
1d4457f9
KC
599 unsigned long atomic_flags; /* Flags needing atomic access. */
600
f56141e3
AL
601 struct restart_block restart_block;
602
1da177e4
LT
603 pid_t pid;
604 pid_t tgid;
0a425405 605
1314562a 606#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
607 /* Canary value for the -fstack-protector gcc feature */
608 unsigned long stack_canary;
1314562a 609#endif
4d1d61a6 610 /*
1da177e4 611 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 612 * older sibling, respectively. (p->father can be replaced with
f470021a 613 * p->real_parent->pid)
1da177e4 614 */
abd63bc3
KC
615 struct task_struct __rcu *real_parent; /* real parent process */
616 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 617 /*
f470021a 618 * children/sibling forms the list of my natural children
1da177e4
LT
619 */
620 struct list_head children; /* list of my children */
621 struct list_head sibling; /* linkage in my parent's children list */
622 struct task_struct *group_leader; /* threadgroup leader */
623
f470021a
RM
624 /*
625 * ptraced is the list of tasks this task is using ptrace on.
626 * This includes both natural children and PTRACE_ATTACH targets.
627 * p->ptrace_entry is p's link on the p->parent->ptraced list.
628 */
629 struct list_head ptraced;
630 struct list_head ptrace_entry;
631
1da177e4 632 /* PID/PID hash table linkage. */
92476d7f 633 struct pid_link pids[PIDTYPE_MAX];
47e65328 634 struct list_head thread_group;
0c740d0a 635 struct list_head thread_node;
1da177e4
LT
636
637 struct completion *vfork_done; /* for vfork() */
638 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
639 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
640
5613fda9 641 u64 utime, stime;
40565b5a 642#ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME
5613fda9 643 u64 utimescaled, stimescaled;
40565b5a 644#endif
16a6d9be 645 u64 gtime;
9d7fb042 646 struct prev_cputime prev_cputime;
6a61671b 647#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
b7ce2277 648 seqcount_t vtime_seqcount;
6a61671b
FW
649 unsigned long long vtime_snap;
650 enum {
7098c1ea
FW
651 /* Task is sleeping or running in a CPU with VTIME inactive */
652 VTIME_INACTIVE = 0,
653 /* Task runs in userspace in a CPU with VTIME active */
6a61671b 654 VTIME_USER,
7098c1ea 655 /* Task runs in kernelspace in a CPU with VTIME active */
6a61671b
FW
656 VTIME_SYS,
657 } vtime_snap_whence;
d99ca3b9 658#endif
d027d45d
FW
659
660#ifdef CONFIG_NO_HZ_FULL
f009a7a7 661 atomic_t tick_dep_mask;
d027d45d 662#endif
1da177e4 663 unsigned long nvcsw, nivcsw; /* context switch counts */
ccbf62d8 664 u64 start_time; /* monotonic time in nsec */
57e0be04 665 u64 real_start_time; /* boot based time in nsec */
1da177e4
LT
666/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
667 unsigned long min_flt, maj_flt;
668
b18b6a9c 669#ifdef CONFIG_POSIX_TIMERS
f06febc9 670 struct task_cputime cputime_expires;
1da177e4 671 struct list_head cpu_timers[3];
b18b6a9c 672#endif
1da177e4
LT
673
674/* process credentials */
64b875f7 675 const struct cred __rcu *ptracer_cred; /* Tracer's credentials at attach */
1b0ba1c9 676 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 677 * credentials (COW) */
1b0ba1c9 678 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 679 * credentials (COW) */
36772092
PBG
680 char comm[TASK_COMM_LEN]; /* executable name excluding path
681 - access with [gs]et_task_comm (which lock
682 it with task_lock())
221af7f8 683 - initialized normally by setup_new_exec */
1da177e4 684/* file system info */
756daf26 685 struct nameidata *nameidata;
3d5b6fcc 686#ifdef CONFIG_SYSVIPC
1da177e4
LT
687/* ipc stuff */
688 struct sysv_sem sysvsem;
ab602f79 689 struct sysv_shm sysvshm;
3d5b6fcc 690#endif
e162b39a 691#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 692/* hung task detection */
82a1fcb9
IM
693 unsigned long last_switch_count;
694#endif
1da177e4
LT
695/* filesystem information */
696 struct fs_struct *fs;
697/* open file information */
698 struct files_struct *files;
1651e14e 699/* namespaces */
ab516013 700 struct nsproxy *nsproxy;
1da177e4
LT
701/* signal handlers */
702 struct signal_struct *signal;
703 struct sighand_struct *sighand;
704
705 sigset_t blocked, real_blocked;
f3de272b 706 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
707 struct sigpending pending;
708
709 unsigned long sas_ss_sp;
710 size_t sas_ss_size;
2a742138 711 unsigned sas_ss_flags;
2e01fabe 712
67d12145 713 struct callback_head *task_works;
e73f8959 714
1da177e4 715 struct audit_context *audit_context;
bfef93a5 716#ifdef CONFIG_AUDITSYSCALL
e1760bd5 717 kuid_t loginuid;
4746ec5b 718 unsigned int sessionid;
bfef93a5 719#endif
932ecebb 720 struct seccomp seccomp;
1da177e4
LT
721
722/* Thread group tracking */
723 u32 parent_exec_id;
724 u32 self_exec_id;
58568d2a
MX
725/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
726 * mempolicy */
1da177e4 727 spinlock_t alloc_lock;
1da177e4 728
b29739f9 729 /* Protection of the PI data structures: */
1d615482 730 raw_spinlock_t pi_lock;
b29739f9 731
76751049
PZ
732 struct wake_q_node wake_q;
733
23f78d4a
IM
734#ifdef CONFIG_RT_MUTEXES
735 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
736 struct rb_root pi_waiters;
737 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
738 /* Deadlock detection and priority inheritance handling */
739 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
740#endif
741
408894ee
IM
742#ifdef CONFIG_DEBUG_MUTEXES
743 /* mutex deadlock detection */
744 struct mutex_waiter *blocked_on;
745#endif
de30a2b3
IM
746#ifdef CONFIG_TRACE_IRQFLAGS
747 unsigned int irq_events;
de30a2b3 748 unsigned long hardirq_enable_ip;
de30a2b3 749 unsigned long hardirq_disable_ip;
fa1452e8 750 unsigned int hardirq_enable_event;
de30a2b3 751 unsigned int hardirq_disable_event;
fa1452e8
HS
752 int hardirqs_enabled;
753 int hardirq_context;
de30a2b3 754 unsigned long softirq_disable_ip;
de30a2b3 755 unsigned long softirq_enable_ip;
fa1452e8 756 unsigned int softirq_disable_event;
de30a2b3 757 unsigned int softirq_enable_event;
fa1452e8 758 int softirqs_enabled;
de30a2b3
IM
759 int softirq_context;
760#endif
fbb9ce95 761#ifdef CONFIG_LOCKDEP
bdb9441e 762# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
763 u64 curr_chain_key;
764 int lockdep_depth;
fbb9ce95 765 unsigned int lockdep_recursion;
c7aceaba 766 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 767 gfp_t lockdep_reclaim_gfp;
fbb9ce95 768#endif
c6d30853
AR
769#ifdef CONFIG_UBSAN
770 unsigned int in_ubsan;
771#endif
408894ee 772
1da177e4
LT
773/* journalling filesystem info */
774 void *journal_info;
775
d89d8796 776/* stacked block device info */
bddd87c7 777 struct bio_list *bio_list;
d89d8796 778
73c10101
JA
779#ifdef CONFIG_BLOCK
780/* stack plugging */
781 struct blk_plug *plug;
782#endif
783
1da177e4
LT
784/* VM state */
785 struct reclaim_state *reclaim_state;
786
1da177e4
LT
787 struct backing_dev_info *backing_dev_info;
788
789 struct io_context *io_context;
790
791 unsigned long ptrace_message;
792 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 793 struct task_io_accounting ioac;
8f0ab514 794#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
795 u64 acct_rss_mem1; /* accumulated rss usage */
796 u64 acct_vm_mem1; /* accumulated virtual memory usage */
605dc2b3 797 u64 acct_timexpd; /* stime + utime since last update */
1da177e4
LT
798#endif
799#ifdef CONFIG_CPUSETS
58568d2a 800 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 801 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 802 int cpuset_mem_spread_rotor;
6adef3eb 803 int cpuset_slab_spread_rotor;
1da177e4 804#endif
ddbcc7e8 805#ifdef CONFIG_CGROUPS
817929ec 806 /* Control Group info protected by css_set_lock */
2c392b8c 807 struct css_set __rcu *cgroups;
817929ec
PM
808 /* cg_list protected by css_set_lock and tsk->alloc_lock */
809 struct list_head cg_list;
ddbcc7e8 810#endif
e02737d5
FY
811#ifdef CONFIG_INTEL_RDT_A
812 int closid;
813#endif
42b2dd0a 814#ifdef CONFIG_FUTEX
0771dfef 815 struct robust_list_head __user *robust_list;
34f192c6
IM
816#ifdef CONFIG_COMPAT
817 struct compat_robust_list_head __user *compat_robust_list;
818#endif
c87e2837
IM
819 struct list_head pi_state_list;
820 struct futex_pi_state *pi_state_cache;
c7aceaba 821#endif
cdd6c482 822#ifdef CONFIG_PERF_EVENTS
8dc85d54 823 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
824 struct mutex perf_event_mutex;
825 struct list_head perf_event_list;
a63eaf34 826#endif
8f47b187
TG
827#ifdef CONFIG_DEBUG_PREEMPT
828 unsigned long preempt_disable_ip;
829#endif
c7aceaba 830#ifdef CONFIG_NUMA
58568d2a 831 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 832 short il_next;
207205a2 833 short pref_node_fork;
42b2dd0a 834#endif
cbee9f88
PZ
835#ifdef CONFIG_NUMA_BALANCING
836 int numa_scan_seq;
cbee9f88 837 unsigned int numa_scan_period;
598f0ec0 838 unsigned int numa_scan_period_max;
de1c9ce6 839 int numa_preferred_nid;
6b9a7460 840 unsigned long numa_migrate_retry;
cbee9f88 841 u64 node_stamp; /* migration stamp */
7e2703e6
RR
842 u64 last_task_numa_placement;
843 u64 last_sum_exec_runtime;
cbee9f88 844 struct callback_head numa_work;
f809ca9a 845
8c8a743c
PZ
846 struct list_head numa_entry;
847 struct numa_group *numa_group;
848
745d6147 849 /*
44dba3d5
IM
850 * numa_faults is an array split into four regions:
851 * faults_memory, faults_cpu, faults_memory_buffer, faults_cpu_buffer
852 * in this precise order.
853 *
854 * faults_memory: Exponential decaying average of faults on a per-node
855 * basis. Scheduling placement decisions are made based on these
856 * counts. The values remain static for the duration of a PTE scan.
857 * faults_cpu: Track the nodes the process was running on when a NUMA
858 * hinting fault was incurred.
859 * faults_memory_buffer and faults_cpu_buffer: Record faults per node
860 * during the current scan window. When the scan completes, the counts
861 * in faults_memory and faults_cpu decay and these values are copied.
745d6147 862 */
44dba3d5 863 unsigned long *numa_faults;
83e1d2cd 864 unsigned long total_numa_faults;
745d6147 865
04bb2f94
RR
866 /*
867 * numa_faults_locality tracks if faults recorded during the last
074c2381
MG
868 * scan window were remote/local or failed to migrate. The task scan
869 * period is adapted based on the locality of the faults with different
870 * weights depending on whether they were shared or private faults
04bb2f94 871 */
074c2381 872 unsigned long numa_faults_locality[3];
04bb2f94 873
b32e86b4 874 unsigned long numa_pages_migrated;
cbee9f88
PZ
875#endif /* CONFIG_NUMA_BALANCING */
876
72b252ae 877 struct tlbflush_unmap_batch tlb_ubc;
72b252ae 878
e56d0903 879 struct rcu_head rcu;
b92ce558
JA
880
881 /*
882 * cache last used pipe for splice
883 */
884 struct pipe_inode_info *splice_pipe;
5640f768
ED
885
886 struct page_frag task_frag;
887
47913d4e
IM
888#ifdef CONFIG_TASK_DELAY_ACCT
889 struct task_delay_info *delays;
f4f154fd 890#endif
47913d4e 891
f4f154fd
AM
892#ifdef CONFIG_FAULT_INJECTION
893 int make_it_fail;
ca74e92b 894#endif
9d823e8f
WF
895 /*
896 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
897 * balance_dirty_pages() for some dirty throttling pause
898 */
899 int nr_dirtied;
900 int nr_dirtied_pause;
83712358 901 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 902
9745512c
AV
903#ifdef CONFIG_LATENCYTOP
904 int latency_record_count;
905 struct latency_record latency_record[LT_SAVECOUNT];
906#endif
6976675d
AV
907 /*
908 * time slack values; these are used to round up poll() and
909 * select() etc timeout values. These are in nanoseconds.
910 */
da8b44d5
JS
911 u64 timer_slack_ns;
912 u64 default_timer_slack_ns;
f8d570a4 913
0b24becc
AR
914#ifdef CONFIG_KASAN
915 unsigned int kasan_depth;
916#endif
fb52607a 917#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 918 /* Index of current stored address in ret_stack */
f201ae23
FW
919 int curr_ret_stack;
920 /* Stack of return addresses for return function tracing */
921 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
922 /* time stamp for last schedule */
923 unsigned long long ftrace_timestamp;
f201ae23
FW
924 /*
925 * Number of functions that haven't been traced
926 * because of depth overrun.
927 */
928 atomic_t trace_overrun;
380c4b14
FW
929 /* Pause for the tracing */
930 atomic_t tracing_graph_pause;
f201ae23 931#endif
ea4e2bc4
SR
932#ifdef CONFIG_TRACING
933 /* state flags for use by tracers */
934 unsigned long trace;
b1cff0ad 935 /* bitmask and counter of trace recursion */
261842b7
SR
936 unsigned long trace_recursion;
937#endif /* CONFIG_TRACING */
5c9a8750
DV
938#ifdef CONFIG_KCOV
939 /* Coverage collection mode enabled for this task (0 if disabled). */
940 enum kcov_mode kcov_mode;
941 /* Size of the kcov_area. */
942 unsigned kcov_size;
943 /* Buffer for coverage collection. */
944 void *kcov_area;
945 /* kcov desciptor wired with this task or NULL. */
946 struct kcov *kcov;
947#endif
6f185c29 948#ifdef CONFIG_MEMCG
626ebc41
TH
949 struct mem_cgroup *memcg_in_oom;
950 gfp_t memcg_oom_gfp_mask;
951 int memcg_oom_order;
b23afb93
TH
952
953 /* number of pages to reclaim on returning to userland */
954 unsigned int memcg_nr_pages_over_high;
569b846d 955#endif
0326f5a9
SD
956#ifdef CONFIG_UPROBES
957 struct uprobe_task *utask;
0326f5a9 958#endif
cafe5635
KO
959#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
960 unsigned int sequential_io;
961 unsigned int sequential_io_avg;
962#endif
8eb23b9f
PZ
963#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
964 unsigned long task_state_change;
965#endif
8bcbde54 966 int pagefault_disabled;
03049269 967#ifdef CONFIG_MMU
29c696e1 968 struct task_struct *oom_reaper_list;
03049269 969#endif
ba14a194
AL
970#ifdef CONFIG_VMAP_STACK
971 struct vm_struct *stack_vm_area;
972#endif
68f24b08
AL
973#ifdef CONFIG_THREAD_INFO_IN_TASK
974 /* A live task holds one reference. */
975 atomic_t stack_refcount;
976#endif
0c8c0f03
DH
977/* CPU-specific state of this task */
978 struct thread_struct thread;
979/*
980 * WARNING: on x86, 'thread_struct' contains a variable-sized
981 * structure. It *MUST* be at the end of 'task_struct'.
982 *
983 * Do not put anything below here!
984 */
1da177e4
LT
985};
986
e868171a 987static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
988{
989 return task->pids[PIDTYPE_PID].pid;
990}
991
e868171a 992static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
993{
994 return task->group_leader->pids[PIDTYPE_PID].pid;
995}
996
6dda81f4
ON
997/*
998 * Without tasklist or rcu lock it is not safe to dereference
999 * the result of task_pgrp/task_session even if task == current,
1000 * we can race with another thread doing sys_setsid/sys_setpgid.
1001 */
e868171a 1002static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1003{
1004 return task->group_leader->pids[PIDTYPE_PGID].pid;
1005}
1006
e868171a 1007static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1008{
1009 return task->group_leader->pids[PIDTYPE_SID].pid;
1010}
1011
7af57294
PE
1012/*
1013 * the helpers to get the task's different pids as they are seen
1014 * from various namespaces
1015 *
1016 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1017 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1018 * current.
7af57294
PE
1019 * task_xid_nr_ns() : id seen from the ns specified;
1020 *
1021 * set_task_vxid() : assigns a virtual id to a task;
1022 *
7af57294
PE
1023 * see also pid_nr() etc in include/linux/pid.h
1024 */
52ee2dfd
ON
1025pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1026 struct pid_namespace *ns);
7af57294 1027
e868171a 1028static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1029{
1030 return tsk->pid;
1031}
1032
52ee2dfd
ON
1033static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1034 struct pid_namespace *ns)
1035{
1036 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1037}
7af57294
PE
1038
1039static inline pid_t task_pid_vnr(struct task_struct *tsk)
1040{
52ee2dfd 1041 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1042}
1043
1044
e868171a 1045static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1046{
1047 return tsk->tgid;
1048}
1049
2f2a3a46 1050pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1051
1052static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1053{
1054 return pid_vnr(task_tgid(tsk));
1055}
1056
1057
80e0b6e8 1058static inline int pid_alive(const struct task_struct *p);
ad36d282
RGB
1059static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns)
1060{
1061 pid_t pid = 0;
1062
1063 rcu_read_lock();
1064 if (pid_alive(tsk))
1065 pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns);
1066 rcu_read_unlock();
1067
1068 return pid;
1069}
1070
1071static inline pid_t task_ppid_nr(const struct task_struct *tsk)
1072{
1073 return task_ppid_nr_ns(tsk, &init_pid_ns);
1074}
1075
52ee2dfd
ON
1076static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1077 struct pid_namespace *ns)
7af57294 1078{
52ee2dfd 1079 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1080}
1081
7af57294
PE
1082static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1083{
52ee2dfd 1084 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1085}
1086
1087
52ee2dfd
ON
1088static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1089 struct pid_namespace *ns)
7af57294 1090{
52ee2dfd 1091 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1092}
1093
7af57294
PE
1094static inline pid_t task_session_vnr(struct task_struct *tsk)
1095{
52ee2dfd 1096 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1097}
1098
1b0f7ffd
ON
1099/* obsolete, do not use */
1100static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1101{
1102 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1103}
7af57294 1104
1da177e4
LT
1105/**
1106 * pid_alive - check that a task structure is not stale
1107 * @p: Task structure to be checked.
1108 *
1109 * Test if a process is not yet dead (at most zombie state)
1110 * If pid_alive fails, then pointers within the task structure
1111 * can be stale and must not be dereferenced.
e69f6186
YB
1112 *
1113 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1114 */
ad36d282 1115static inline int pid_alive(const struct task_struct *p)
1da177e4 1116{
92476d7f 1117 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1118}
1119
f400e198 1120/**
570f5241
SS
1121 * is_global_init - check if a task structure is init. Since init
1122 * is free to have sub-threads we need to check tgid.
3260259f
HK
1123 * @tsk: Task structure to be checked.
1124 *
1125 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1126 *
1127 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1128 */
e868171a 1129static inline int is_global_init(struct task_struct *tsk)
b461cc03 1130{
570f5241 1131 return task_tgid_nr(tsk) == 1;
b461cc03 1132}
b460cbc5 1133
9ec52099
CLG
1134extern struct pid *cad_pid;
1135
1da177e4
LT
1136/*
1137 * Per process flags
1138 */
c1de45ca 1139#define PF_IDLE 0x00000002 /* I am an IDLE thread */
1da177e4 1140#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1141#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1142#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1143#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1144#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1145#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1146#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1147#define PF_DUMPCORE 0x00000200 /* dumped core */
1148#define PF_SIGNALED 0x00000400 /* killed by a signal */
1149#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1150#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1151#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1152#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1153#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1154#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1155#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1156#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1157#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1158#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1159#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1160#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1161#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
14a40ffc 1162#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1163#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
61a87122 1164#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1165#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1166#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1167
1168/*
1169 * Only the _current_ task can read/write to tsk->flags, but other
1170 * tasks can access tsk->flags in readonly mode for example
1171 * with tsk_used_math (like during threaded core dumping).
1172 * There is however an exception to this rule during ptrace
1173 * or during fork: the ptracer task is allowed to write to the
1174 * child->flags of its traced child (same goes for fork, the parent
1175 * can write to the child->flags), because we're guaranteed the
1176 * child is not running and in turn not changing child->flags
1177 * at the same time the parent does it.
1178 */
1179#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1180#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1181#define clear_used_math() clear_stopped_child_used_math(current)
1182#define set_used_math() set_stopped_child_used_math(current)
1183#define conditional_stopped_child_used_math(condition, child) \
1184 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1185#define conditional_used_math(condition) \
1186 conditional_stopped_child_used_math(condition, current)
1187#define copy_to_stopped_child_used_math(child) \
1188 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1189/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1190#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1191#define used_math() tsk_used_math(current)
1192
1d4457f9 1193/* Per-process atomic flags. */
a2b86f77 1194#define PFA_NO_NEW_PRIVS 0 /* May not gain new privileges. */
2ad654bc
ZL
1195#define PFA_SPREAD_PAGE 1 /* Spread page cache over cpuset */
1196#define PFA_SPREAD_SLAB 2 /* Spread some slab caches over cpuset */
77ed2c57 1197#define PFA_LMK_WAITING 3 /* Lowmemorykiller is waiting */
2ad654bc 1198
1d4457f9 1199
e0e5070b
ZL
1200#define TASK_PFA_TEST(name, func) \
1201 static inline bool task_##func(struct task_struct *p) \
1202 { return test_bit(PFA_##name, &p->atomic_flags); }
1203#define TASK_PFA_SET(name, func) \
1204 static inline void task_set_##func(struct task_struct *p) \
1205 { set_bit(PFA_##name, &p->atomic_flags); }
1206#define TASK_PFA_CLEAR(name, func) \
1207 static inline void task_clear_##func(struct task_struct *p) \
1208 { clear_bit(PFA_##name, &p->atomic_flags); }
1209
1210TASK_PFA_TEST(NO_NEW_PRIVS, no_new_privs)
1211TASK_PFA_SET(NO_NEW_PRIVS, no_new_privs)
1d4457f9 1212
2ad654bc
ZL
1213TASK_PFA_TEST(SPREAD_PAGE, spread_page)
1214TASK_PFA_SET(SPREAD_PAGE, spread_page)
1215TASK_PFA_CLEAR(SPREAD_PAGE, spread_page)
1216
1217TASK_PFA_TEST(SPREAD_SLAB, spread_slab)
1218TASK_PFA_SET(SPREAD_SLAB, spread_slab)
1219TASK_PFA_CLEAR(SPREAD_SLAB, spread_slab)
1d4457f9 1220
77ed2c57
TH
1221TASK_PFA_TEST(LMK_WAITING, lmk_waiting)
1222TASK_PFA_SET(LMK_WAITING, lmk_waiting)
1223
907aed48
MG
1224static inline void tsk_restore_flags(struct task_struct *task,
1225 unsigned long orig_flags, unsigned long flags)
1226{
1227 task->flags &= ~flags;
1228 task->flags |= orig_flags & flags;
1229}
1230
f82f8042
JL
1231extern int cpuset_cpumask_can_shrink(const struct cpumask *cur,
1232 const struct cpumask *trial);
7f51412a
JL
1233extern int task_can_attach(struct task_struct *p,
1234 const struct cpumask *cs_cpus_allowed);
1da177e4 1235#ifdef CONFIG_SMP
1e1b6c51
KM
1236extern void do_set_cpus_allowed(struct task_struct *p,
1237 const struct cpumask *new_mask);
1238
cd8ba7cd 1239extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1240 const struct cpumask *new_mask);
1da177e4 1241#else
1e1b6c51
KM
1242static inline void do_set_cpus_allowed(struct task_struct *p,
1243 const struct cpumask *new_mask)
1244{
1245}
cd8ba7cd 1246static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1247 const struct cpumask *new_mask)
1da177e4 1248{
96f874e2 1249 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1250 return -EINVAL;
1251 return 0;
1252}
1253#endif
e0ad9556 1254
6d0d2878
CB
1255#ifndef cpu_relax_yield
1256#define cpu_relax_yield() cpu_relax()
1257#endif
1258
fa93384f 1259extern int yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
1260extern void set_user_nice(struct task_struct *p, long nice);
1261extern int task_prio(const struct task_struct *p);
d0ea0268
DY
1262/**
1263 * task_nice - return the nice value of a given task.
1264 * @p: the task in question.
1265 *
1266 * Return: The nice value [ -20 ... 0 ... 19 ].
1267 */
1268static inline int task_nice(const struct task_struct *p)
1269{
1270 return PRIO_TO_NICE((p)->static_prio);
1271}
36c8b586
IM
1272extern int can_nice(const struct task_struct *p, const int nice);
1273extern int task_curr(const struct task_struct *p);
1da177e4 1274extern int idle_cpu(int cpu);
fe7de49f
KM
1275extern int sched_setscheduler(struct task_struct *, int,
1276 const struct sched_param *);
961ccddd 1277extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 1278 const struct sched_param *);
d50dde5a
DF
1279extern int sched_setattr(struct task_struct *,
1280 const struct sched_attr *);
36c8b586 1281extern struct task_struct *idle_task(int cpu);
c4f30608
PM
1282/**
1283 * is_idle_task - is the specified task an idle task?
fa757281 1284 * @p: the task in question.
e69f6186
YB
1285 *
1286 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 1287 */
7061ca3b 1288static inline bool is_idle_task(const struct task_struct *p)
c4f30608 1289{
c1de45ca 1290 return !!(p->flags & PF_IDLE);
c4f30608 1291}
36c8b586 1292extern struct task_struct *curr_task(int cpu);
a458ae2e 1293extern void ia64_set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1294
1295void yield(void);
1296
1da177e4 1297union thread_union {
c65eacbe 1298#ifndef CONFIG_THREAD_INFO_IN_TASK
1da177e4 1299 struct thread_info thread_info;
c65eacbe 1300#endif
1da177e4
LT
1301 unsigned long stack[THREAD_SIZE/sizeof(long)];
1302};
1303
f3ac6067
IM
1304#ifdef CONFIG_THREAD_INFO_IN_TASK
1305static inline struct thread_info *task_thread_info(struct task_struct *task)
1306{
1307 return &task->thread_info;
1308}
1309#elif !defined(__HAVE_THREAD_FUNCTIONS)
1310# define task_thread_info(task) ((struct thread_info *)(task)->stack)
1311#endif
1312
198fe21b
PE
1313extern struct pid_namespace init_pid_ns;
1314
1315/*
1316 * find a task by one of its numerical ids
1317 *
198fe21b
PE
1318 * find_task_by_pid_ns():
1319 * finds a task by its pid in the specified namespace
228ebcbe
PE
1320 * find_task_by_vpid():
1321 * finds a task by its virtual pid
198fe21b 1322 *
e49859e7 1323 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
1324 */
1325
228ebcbe
PE
1326extern struct task_struct *find_task_by_vpid(pid_t nr);
1327extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1328 struct pid_namespace *ns);
198fe21b 1329
b3c97528
HH
1330extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1331extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 1332extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
1333#ifdef CONFIG_SMP
1334 extern void kick_process(struct task_struct *tsk);
1335#else
1336 static inline void kick_process(struct task_struct *tsk) { }
1337#endif
1da177e4 1338
82b89778
AH
1339extern void __set_task_comm(struct task_struct *tsk, const char *from, bool exec);
1340static inline void set_task_comm(struct task_struct *tsk, const char *from)
1341{
1342 __set_task_comm(tsk, from, false);
1343}
59714d65 1344extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
1345
1346#ifdef CONFIG_SMP
317f3941 1347void scheduler_ipi(void);
85ba2d86 1348extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 1349#else
184748cc 1350static inline void scheduler_ipi(void) { }
85ba2d86
RM
1351static inline unsigned long wait_task_inactive(struct task_struct *p,
1352 long match_state)
1353{
1354 return 1;
1355}
1da177e4
LT
1356#endif
1357
1da177e4
LT
1358/* set thread flags in other task's structures
1359 * - see asm/thread_info.h for TIF_xxxx flags available
1360 */
1361static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1362{
a1261f54 1363 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1364}
1365
1366static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1367{
a1261f54 1368 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1369}
1370
1371static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1372{
a1261f54 1373 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1374}
1375
1376static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1377{
a1261f54 1378 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1379}
1380
1381static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1382{
a1261f54 1383 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1384}
1385
1386static inline void set_tsk_need_resched(struct task_struct *tsk)
1387{
1388 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1389}
1390
1391static inline void clear_tsk_need_resched(struct task_struct *tsk)
1392{
1393 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1394}
1395
8ae121ac
GH
1396static inline int test_tsk_need_resched(struct task_struct *tsk)
1397{
1398 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
1399}
1400
1da177e4
LT
1401/*
1402 * cond_resched() and cond_resched_lock(): latency reduction via
1403 * explicit rescheduling in places that are safe. The return
1404 * value indicates whether a reschedule was done in fact.
1405 * cond_resched_lock() will drop the spinlock before scheduling,
1406 * cond_resched_softirq() will enable bhs before scheduling.
1407 */
35a773a0 1408#ifndef CONFIG_PREEMPT
c3921ab7 1409extern int _cond_resched(void);
35a773a0
PZ
1410#else
1411static inline int _cond_resched(void) { return 0; }
1412#endif
6f80bd98 1413
613afbf8 1414#define cond_resched() ({ \
3427445a 1415 ___might_sleep(__FILE__, __LINE__, 0); \
613afbf8
FW
1416 _cond_resched(); \
1417})
6f80bd98 1418
613afbf8
FW
1419extern int __cond_resched_lock(spinlock_t *lock);
1420
1421#define cond_resched_lock(lock) ({ \
3427445a 1422 ___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\
613afbf8
FW
1423 __cond_resched_lock(lock); \
1424})
1425
1426extern int __cond_resched_softirq(void);
1427
75e1056f 1428#define cond_resched_softirq() ({ \
3427445a 1429 ___might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
75e1056f 1430 __cond_resched_softirq(); \
613afbf8 1431})
1da177e4 1432
f6f3c437
SH
1433static inline void cond_resched_rcu(void)
1434{
1435#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
1436 rcu_read_unlock();
1437 cond_resched();
1438 rcu_read_lock();
1439#endif
1440}
1441
1da177e4
LT
1442/*
1443 * Does a critical section need to be broken due to another
95c354fe
NP
1444 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
1445 * but a general need for low latency)
1da177e4 1446 */
95c354fe 1447static inline int spin_needbreak(spinlock_t *lock)
1da177e4 1448{
95c354fe
NP
1449#ifdef CONFIG_PREEMPT
1450 return spin_is_contended(lock);
1451#else
1da177e4 1452 return 0;
95c354fe 1453#endif
1da177e4
LT
1454}
1455
75f93fed
PZ
1456static __always_inline bool need_resched(void)
1457{
1458 return unlikely(tif_need_resched());
1459}
1460
1da177e4
LT
1461/*
1462 * Wrappers for p->thread_info->cpu access. No-op on UP.
1463 */
1464#ifdef CONFIG_SMP
1465
1466static inline unsigned int task_cpu(const struct task_struct *p)
1467{
c65eacbe
AL
1468#ifdef CONFIG_THREAD_INFO_IN_TASK
1469 return p->cpu;
1470#else
a1261f54 1471 return task_thread_info(p)->cpu;
c65eacbe 1472#endif
1da177e4
LT
1473}
1474
b32e86b4
IM
1475static inline int task_node(const struct task_struct *p)
1476{
1477 return cpu_to_node(task_cpu(p));
1478}
1479
c65cc870 1480extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
1481
1482#else
1483
1484static inline unsigned int task_cpu(const struct task_struct *p)
1485{
1486 return 0;
1487}
1488
1489static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1490{
1491}
1492
1493#endif /* CONFIG_SMP */
1494
d9345c65
PX
1495/*
1496 * In order to reduce various lock holder preemption latencies provide an
1497 * interface to see if a vCPU is currently running or not.
1498 *
1499 * This allows us to terminate optimistic spin loops and block, analogous to
1500 * the native optimistic spin heuristic of testing if the lock owner task is
1501 * running or not.
1502 */
1503#ifndef vcpu_is_preempted
1504# define vcpu_is_preempted(cpu) false
1505#endif
1506
96f874e2
RR
1507extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
1508extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 1509
82455257
DH
1510#ifndef TASK_SIZE_OF
1511#define TASK_SIZE_OF(tsk) TASK_SIZE
1512#endif
1513
1da177e4 1514#endif