]> git.proxmox.com Git - mirror_ubuntu-hirsute-kernel.git/blame - include/linux/sched.h
sched/headers: Remove the <linux/topology.h> include from <linux/sched.h>
[mirror_ubuntu-hirsute-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 6#include <linux/sched/prio.h>
ee6a3d19 7#include <linux/nodemask.h>
5c228079 8
b69339ba 9#include <linux/mutex.h>
fb00aca4 10#include <linux/plist.h>
77ba809e 11#include <linux/mm_types_task.h>
1da177e4 12
1da177e4 13#include <linux/sem.h>
ab602f79 14#include <linux/shm.h>
f361bf4a 15#include <linux/signal_types.h>
1da177e4 16#include <linux/pid.h>
1da177e4 17#include <linux/seccomp.h>
b68070e1 18#include <linux/rcupdate.h>
1da177e4 19
a3b6714e 20#include <linux/resource.h>
a3b6714e 21#include <linux/hrtimer.h>
5c9a8750 22#include <linux/kcov.h>
7c3ab738 23#include <linux/task_io_accounting.h>
9745512c 24#include <linux/latencytop.h>
a3b6714e 25
70b8157e
IM
26#include <asm/current.h>
27
c7af7877
IM
28/* task_struct member predeclarations: */
29struct audit_context;
30struct autogroup;
31struct backing_dev_info;
bddd87c7 32struct bio_list;
73c10101 33struct blk_plug;
c7af7877 34struct cfs_rq;
c4ad8f98 35struct filename;
c7af7877
IM
36struct fs_struct;
37struct futex_pi_state;
38struct io_context;
39struct mempolicy;
89076bc3 40struct nameidata;
c7af7877
IM
41struct nsproxy;
42struct perf_event_context;
43struct pid_namespace;
44struct pipe_inode_info;
45struct rcu_node;
46struct reclaim_state;
47struct robust_list_head;
48struct sched_attr;
49struct sched_param;
43ae34cb 50struct seq_file;
c7af7877
IM
51struct sighand_struct;
52struct signal_struct;
53struct task_delay_info;
4cf86d77 54struct task_group;
c7af7877
IM
55struct task_struct;
56struct uts_namespace;
1da177e4 57
4a8342d2
LT
58/*
59 * Task state bitmask. NOTE! These bits are also
60 * encoded in fs/proc/array.c: get_task_state().
61 *
62 * We have two separate sets of flags: task->state
63 * is about runnability, while task->exit_state are
64 * about the task exiting. Confusing, but this way
65 * modifying one set can't modify the other one by
66 * mistake.
67 */
1da177e4
LT
68#define TASK_RUNNING 0
69#define TASK_INTERRUPTIBLE 1
70#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
71#define __TASK_STOPPED 4
72#define __TASK_TRACED 8
4a8342d2 73/* in tsk->exit_state */
ad86622b
ON
74#define EXIT_DEAD 16
75#define EXIT_ZOMBIE 32
abd50b39 76#define EXIT_TRACE (EXIT_ZOMBIE | EXIT_DEAD)
4a8342d2 77/* in tsk->state again */
af927232 78#define TASK_DEAD 64
f021a3c2 79#define TASK_WAKEKILL 128
e9c84311 80#define TASK_WAKING 256
f2530dc7 81#define TASK_PARKED 512
80ed87c8 82#define TASK_NOLOAD 1024
7dc603c9
PZ
83#define TASK_NEW 2048
84#define TASK_STATE_MAX 4096
f021a3c2 85
7dc603c9 86#define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWPNn"
73342151 87
642fa448 88/* Convenience macros for the sake of set_current_state */
f021a3c2
MW
89#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
90#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
91#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 92
80ed87c8
PZ
93#define TASK_IDLE (TASK_UNINTERRUPTIBLE | TASK_NOLOAD)
94
92a1f4bc
MW
95/* Convenience macros for the sake of wake_up */
96#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 97#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
98
99/* get_task_state() */
100#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2 101 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
74e37200 102 __TASK_TRACED | EXIT_ZOMBIE | EXIT_DEAD)
92a1f4bc 103
f021a3c2
MW
104#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
105#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 106#define task_is_stopped_or_traced(task) \
f021a3c2 107 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 108#define task_contributes_to_load(task) \
e3c8ca83 109 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
80ed87c8
PZ
110 (task->flags & PF_FROZEN) == 0 && \
111 (task->state & TASK_NOLOAD) == 0)
1da177e4 112
8eb23b9f
PZ
113#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
114
8eb23b9f
PZ
115#define __set_current_state(state_value) \
116 do { \
117 current->task_state_change = _THIS_IP_; \
118 current->state = (state_value); \
119 } while (0)
120#define set_current_state(state_value) \
121 do { \
122 current->task_state_change = _THIS_IP_; \
a2250238 123 smp_store_mb(current->state, (state_value)); \
8eb23b9f
PZ
124 } while (0)
125
126#else
498d0c57
AM
127/*
128 * set_current_state() includes a barrier so that the write of current->state
129 * is correctly serialised wrt the caller's subsequent test of whether to
130 * actually sleep:
131 *
a2250238 132 * for (;;) {
498d0c57 133 * set_current_state(TASK_UNINTERRUPTIBLE);
a2250238
PZ
134 * if (!need_sleep)
135 * break;
136 *
137 * schedule();
138 * }
139 * __set_current_state(TASK_RUNNING);
140 *
141 * If the caller does not need such serialisation (because, for instance, the
142 * condition test and condition change and wakeup are under the same lock) then
143 * use __set_current_state().
144 *
145 * The above is typically ordered against the wakeup, which does:
146 *
147 * need_sleep = false;
148 * wake_up_state(p, TASK_UNINTERRUPTIBLE);
149 *
150 * Where wake_up_state() (and all other wakeup primitives) imply enough
151 * barriers to order the store of the variable against wakeup.
152 *
153 * Wakeup will do: if (@state & p->state) p->state = TASK_RUNNING, that is,
154 * once it observes the TASK_UNINTERRUPTIBLE store the waking CPU can issue a
155 * TASK_RUNNING store which can collide with __set_current_state(TASK_RUNNING).
498d0c57 156 *
a2250238 157 * This is obviously fine, since they both store the exact same value.
498d0c57 158 *
a2250238 159 * Also see the comments of try_to_wake_up().
498d0c57 160 */
8eb23b9f 161#define __set_current_state(state_value) \
1da177e4 162 do { current->state = (state_value); } while (0)
8eb23b9f 163#define set_current_state(state_value) \
b92b8b35 164 smp_store_mb(current->state, (state_value))
1da177e4 165
8eb23b9f
PZ
166#endif
167
1da177e4
LT
168/* Task command name length */
169#define TASK_COMM_LEN 16
170
3fa0818b
RR
171extern cpumask_var_t cpu_isolated_map;
172
1da177e4
LT
173extern void scheduler_tick(void);
174
1da177e4 175#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 176extern signed long schedule_timeout(signed long timeout);
64ed93a2 177extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 178extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 179extern signed long schedule_timeout_uninterruptible(signed long timeout);
69b27baf 180extern signed long schedule_timeout_idle(signed long timeout);
1da177e4 181asmlinkage void schedule(void);
c5491ea7 182extern void schedule_preempt_disabled(void);
1da177e4 183
10ab5643
TH
184extern int __must_check io_schedule_prepare(void);
185extern void io_schedule_finish(int token);
9cff8ade 186extern long io_schedule_timeout(long timeout);
10ab5643 187extern void io_schedule(void);
9cff8ade 188
d37f761d 189/**
9d7fb042 190 * struct prev_cputime - snaphsot of system and user cputime
d37f761d
FW
191 * @utime: time spent in user mode
192 * @stime: time spent in system mode
9d7fb042 193 * @lock: protects the above two fields
d37f761d 194 *
9d7fb042
PZ
195 * Stores previous user/system time values such that we can guarantee
196 * monotonicity.
d37f761d 197 */
9d7fb042
PZ
198struct prev_cputime {
199#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
5613fda9
FW
200 u64 utime;
201 u64 stime;
9d7fb042
PZ
202 raw_spinlock_t lock;
203#endif
d37f761d
FW
204};
205
f06febc9
FM
206/**
207 * struct task_cputime - collected CPU time counts
5613fda9
FW
208 * @utime: time spent in user mode, in nanoseconds
209 * @stime: time spent in kernel mode, in nanoseconds
f06febc9 210 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 211 *
9d7fb042
PZ
212 * This structure groups together three kinds of CPU time that are tracked for
213 * threads and thread groups. Most things considering CPU time want to group
214 * these counts together and treat all three of them in parallel.
f06febc9
FM
215 */
216struct task_cputime {
5613fda9
FW
217 u64 utime;
218 u64 stime;
f06febc9
FM
219 unsigned long long sum_exec_runtime;
220};
9d7fb042 221
f06febc9 222/* Alternate field names when used to cache expirations. */
f06febc9 223#define virt_exp utime
9d7fb042 224#define prof_exp stime
f06febc9
FM
225#define sched_exp sum_exec_runtime
226
f6db8347 227#ifdef CONFIG_SCHED_INFO
1da177e4
LT
228struct sched_info {
229 /* cumulative counters */
2d72376b 230 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 231 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
232
233 /* timestamps */
172ba844
BS
234 unsigned long long last_arrival,/* when we last ran on a cpu */
235 last_queued; /* when we were last queued to run */
1da177e4 236};
f6db8347 237#endif /* CONFIG_SCHED_INFO */
1da177e4 238
6ecdd749
YD
239/*
240 * Integer metrics need fixed point arithmetic, e.g., sched/fair
241 * has a few: load, load_avg, util_avg, freq, and capacity.
242 *
243 * We define a basic fixed point arithmetic range, and then formalize
244 * all these metrics based on that basic range.
245 */
246# define SCHED_FIXEDPOINT_SHIFT 10
247# define SCHED_FIXEDPOINT_SCALE (1L << SCHED_FIXEDPOINT_SHIFT)
248
20b8a59f 249struct load_weight {
9dbdb155
PZ
250 unsigned long weight;
251 u32 inv_weight;
20b8a59f
IM
252};
253
9d89c257 254/*
7b595334
YD
255 * The load_avg/util_avg accumulates an infinite geometric series
256 * (see __update_load_avg() in kernel/sched/fair.c).
257 *
258 * [load_avg definition]
259 *
260 * load_avg = runnable% * scale_load_down(load)
261 *
262 * where runnable% is the time ratio that a sched_entity is runnable.
263 * For cfs_rq, it is the aggregated load_avg of all runnable and
9d89c257 264 * blocked sched_entities.
7b595334
YD
265 *
266 * load_avg may also take frequency scaling into account:
267 *
268 * load_avg = runnable% * scale_load_down(load) * freq%
269 *
270 * where freq% is the CPU frequency normalized to the highest frequency.
271 *
272 * [util_avg definition]
273 *
274 * util_avg = running% * SCHED_CAPACITY_SCALE
275 *
276 * where running% is the time ratio that a sched_entity is running on
277 * a CPU. For cfs_rq, it is the aggregated util_avg of all runnable
278 * and blocked sched_entities.
279 *
280 * util_avg may also factor frequency scaling and CPU capacity scaling:
281 *
282 * util_avg = running% * SCHED_CAPACITY_SCALE * freq% * capacity%
283 *
284 * where freq% is the same as above, and capacity% is the CPU capacity
285 * normalized to the greatest capacity (due to uarch differences, etc).
286 *
287 * N.B., the above ratios (runnable%, running%, freq%, and capacity%)
288 * themselves are in the range of [0, 1]. To do fixed point arithmetics,
289 * we therefore scale them to as large a range as necessary. This is for
290 * example reflected by util_avg's SCHED_CAPACITY_SCALE.
291 *
292 * [Overflow issue]
293 *
294 * The 64-bit load_sum can have 4353082796 (=2^64/47742/88761) entities
295 * with the highest load (=88761), always runnable on a single cfs_rq,
296 * and should not overflow as the number already hits PID_MAX_LIMIT.
297 *
298 * For all other cases (including 32-bit kernels), struct load_weight's
299 * weight will overflow first before we do, because:
300 *
301 * Max(load_avg) <= Max(load.weight)
302 *
303 * Then it is the load_weight's responsibility to consider overflow
304 * issues.
9d89c257 305 */
9d85f21c 306struct sched_avg {
9d89c257
YD
307 u64 last_update_time, load_sum;
308 u32 util_sum, period_contrib;
309 unsigned long load_avg, util_avg;
9d85f21c
PT
310};
311
94c18227 312#ifdef CONFIG_SCHEDSTATS
41acab88 313struct sched_statistics {
20b8a59f 314 u64 wait_start;
94c18227 315 u64 wait_max;
6d082592
AV
316 u64 wait_count;
317 u64 wait_sum;
8f0dfc34
AV
318 u64 iowait_count;
319 u64 iowait_sum;
94c18227 320
20b8a59f 321 u64 sleep_start;
20b8a59f 322 u64 sleep_max;
94c18227
IM
323 s64 sum_sleep_runtime;
324
325 u64 block_start;
20b8a59f
IM
326 u64 block_max;
327 u64 exec_max;
eba1ed4b 328 u64 slice_max;
cc367732 329
cc367732
IM
330 u64 nr_migrations_cold;
331 u64 nr_failed_migrations_affine;
332 u64 nr_failed_migrations_running;
333 u64 nr_failed_migrations_hot;
334 u64 nr_forced_migrations;
cc367732
IM
335
336 u64 nr_wakeups;
337 u64 nr_wakeups_sync;
338 u64 nr_wakeups_migrate;
339 u64 nr_wakeups_local;
340 u64 nr_wakeups_remote;
341 u64 nr_wakeups_affine;
342 u64 nr_wakeups_affine_attempts;
343 u64 nr_wakeups_passive;
344 u64 nr_wakeups_idle;
41acab88
LDM
345};
346#endif
347
348struct sched_entity {
349 struct load_weight load; /* for load-balancing */
350 struct rb_node run_node;
351 struct list_head group_node;
352 unsigned int on_rq;
353
354 u64 exec_start;
355 u64 sum_exec_runtime;
356 u64 vruntime;
357 u64 prev_sum_exec_runtime;
358
41acab88
LDM
359 u64 nr_migrations;
360
41acab88
LDM
361#ifdef CONFIG_SCHEDSTATS
362 struct sched_statistics statistics;
94c18227
IM
363#endif
364
20b8a59f 365#ifdef CONFIG_FAIR_GROUP_SCHED
fed14d45 366 int depth;
20b8a59f
IM
367 struct sched_entity *parent;
368 /* rq on which this entity is (to be) queued: */
369 struct cfs_rq *cfs_rq;
370 /* rq "owned" by this entity/group: */
371 struct cfs_rq *my_q;
372#endif
8bd75c77 373
141965c7 374#ifdef CONFIG_SMP
5a107804
JO
375 /*
376 * Per entity load average tracking.
377 *
378 * Put into separate cache line so it does not
379 * collide with read-mostly values above.
380 */
381 struct sched_avg avg ____cacheline_aligned_in_smp;
9d85f21c 382#endif
20b8a59f 383};
70b97a7f 384
fa717060
PZ
385struct sched_rt_entity {
386 struct list_head run_list;
78f2c7db 387 unsigned long timeout;
57d2aa00 388 unsigned long watchdog_stamp;
bee367ed 389 unsigned int time_slice;
ff77e468
PZ
390 unsigned short on_rq;
391 unsigned short on_list;
6f505b16 392
58d6c2d7 393 struct sched_rt_entity *back;
052f1dc7 394#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
395 struct sched_rt_entity *parent;
396 /* rq on which this entity is (to be) queued: */
397 struct rt_rq *rt_rq;
398 /* rq "owned" by this entity/group: */
399 struct rt_rq *my_q;
400#endif
fa717060
PZ
401};
402
aab03e05
DF
403struct sched_dl_entity {
404 struct rb_node rb_node;
405
406 /*
407 * Original scheduling parameters. Copied here from sched_attr
4027d080 408 * during sched_setattr(), they will remain the same until
409 * the next sched_setattr().
aab03e05
DF
410 */
411 u64 dl_runtime; /* maximum runtime for each instance */
412 u64 dl_deadline; /* relative deadline of each instance */
755378a4 413 u64 dl_period; /* separation of two instances (period) */
332ac17e 414 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
415
416 /*
417 * Actual scheduling parameters. Initialized with the values above,
418 * they are continously updated during task execution. Note that
419 * the remaining runtime could be < 0 in case we are in overrun.
420 */
421 s64 runtime; /* remaining runtime for this instance */
422 u64 deadline; /* absolute deadline for this instance */
423 unsigned int flags; /* specifying the scheduler behaviour */
424
425 /*
426 * Some bool flags:
427 *
428 * @dl_throttled tells if we exhausted the runtime. If so, the
429 * task has to wait for a replenishment to be performed at the
430 * next firing of dl_timer.
431 *
2d3d891d
DF
432 * @dl_boosted tells if we are boosted due to DI. If so we are
433 * outside bandwidth enforcement mechanism (but only until we
5bfd126e
JL
434 * exit the critical section);
435 *
436 * @dl_yielded tells if task gave up the cpu before consuming
437 * all its available runtime during the last job.
aab03e05 438 */
72f9f3fd 439 int dl_throttled, dl_boosted, dl_yielded;
aab03e05
DF
440
441 /*
442 * Bandwidth enforcement timer. Each -deadline task has its
443 * own bandwidth to be enforced, thus we need one timer per task.
444 */
445 struct hrtimer dl_timer;
446};
8bd75c77 447
1d082fd0
PM
448union rcu_special {
449 struct {
8203d6d0
PM
450 u8 blocked;
451 u8 need_qs;
452 u8 exp_need_qs;
453 u8 pad; /* Otherwise the compiler can store garbage here. */
454 } b; /* Bits. */
455 u32 s; /* Set of bits. */
1d082fd0 456};
86848966 457
8dc85d54
PZ
458enum perf_event_task_context {
459 perf_invalid_context = -1,
460 perf_hw_context = 0,
89a1e187 461 perf_sw_context,
8dc85d54
PZ
462 perf_nr_task_contexts,
463};
464
eb61baf6
IM
465struct wake_q_node {
466 struct wake_q_node *next;
467};
468
1da177e4 469struct task_struct {
c65eacbe
AL
470#ifdef CONFIG_THREAD_INFO_IN_TASK
471 /*
472 * For reasons of header soup (see current_thread_info()), this
473 * must be the first element of task_struct.
474 */
475 struct thread_info thread_info;
476#endif
1da177e4 477 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 478 void *stack;
1da177e4 479 atomic_t usage;
97dc32cd
WC
480 unsigned int flags; /* per process flags, defined below */
481 unsigned int ptrace;
1da177e4 482
2dd73a4f 483#ifdef CONFIG_SMP
fa14ff4a 484 struct llist_node wake_entry;
3ca7a440 485 int on_cpu;
c65eacbe
AL
486#ifdef CONFIG_THREAD_INFO_IN_TASK
487 unsigned int cpu; /* current CPU */
488#endif
63b0e9ed 489 unsigned int wakee_flips;
62470419 490 unsigned long wakee_flip_decay_ts;
63b0e9ed 491 struct task_struct *last_wakee;
ac66f547
PZ
492
493 int wake_cpu;
2dd73a4f 494#endif
fd2f4419 495 int on_rq;
50e645a8 496
b29739f9 497 int prio, static_prio, normal_prio;
c7aceaba 498 unsigned int rt_priority;
5522d5d5 499 const struct sched_class *sched_class;
20b8a59f 500 struct sched_entity se;
fa717060 501 struct sched_rt_entity rt;
8323f26c
PZ
502#ifdef CONFIG_CGROUP_SCHED
503 struct task_group *sched_task_group;
504#endif
aab03e05 505 struct sched_dl_entity dl;
1da177e4 506
e107be36
AK
507#ifdef CONFIG_PREEMPT_NOTIFIERS
508 /* list of struct preempt_notifier: */
509 struct hlist_head preempt_notifiers;
510#endif
511
6c5c9341 512#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 513 unsigned int btrace_seq;
6c5c9341 514#endif
1da177e4 515
97dc32cd 516 unsigned int policy;
29baa747 517 int nr_cpus_allowed;
1da177e4 518 cpumask_t cpus_allowed;
1da177e4 519
a57eb940 520#ifdef CONFIG_PREEMPT_RCU
e260be67 521 int rcu_read_lock_nesting;
1d082fd0 522 union rcu_special rcu_read_unlock_special;
f41d911f 523 struct list_head rcu_node_entry;
a57eb940 524 struct rcu_node *rcu_blocked_node;
28f6569a 525#endif /* #ifdef CONFIG_PREEMPT_RCU */
8315f422
PM
526#ifdef CONFIG_TASKS_RCU
527 unsigned long rcu_tasks_nvcsw;
528 bool rcu_tasks_holdout;
529 struct list_head rcu_tasks_holdout_list;
176f8f7a 530 int rcu_tasks_idle_cpu;
8315f422 531#endif /* #ifdef CONFIG_TASKS_RCU */
e260be67 532
f6db8347 533#ifdef CONFIG_SCHED_INFO
1da177e4
LT
534 struct sched_info sched_info;
535#endif
536
537 struct list_head tasks;
806c09a7 538#ifdef CONFIG_SMP
917b627d 539 struct plist_node pushable_tasks;
1baca4ce 540 struct rb_node pushable_dl_tasks;
806c09a7 541#endif
1da177e4
LT
542
543 struct mm_struct *mm, *active_mm;
314ff785
IM
544
545 /* Per-thread vma caching: */
546 struct vmacache vmacache;
547
34e55232
KH
548#if defined(SPLIT_RSS_COUNTING)
549 struct task_rss_stat rss_stat;
550#endif
1da177e4 551/* task state */
97dc32cd 552 int exit_state;
1da177e4
LT
553 int exit_code, exit_signal;
554 int pdeath_signal; /* The signal sent when the parent dies */
e7cc4173 555 unsigned long jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
556
557 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 558 unsigned int personality;
9b89f6ba 559
be958bdc 560 /* scheduler bits, serialized by scheduler locks */
ca94c442 561 unsigned sched_reset_on_fork:1;
a8e4f2ea 562 unsigned sched_contributes_to_load:1;
ff303e66 563 unsigned sched_migrated:1;
b7e7ade3 564 unsigned sched_remote_wakeup:1;
be958bdc
PZ
565 unsigned :0; /* force alignment to the next boundary */
566
567 /* unserialized, strictly 'current' */
568 unsigned in_execve:1; /* bit to tell LSMs we're in execve */
569 unsigned in_iowait:1;
7e781418
AL
570#if !defined(TIF_RESTORE_SIGMASK)
571 unsigned restore_sigmask:1;
572#endif
626ebc41
TH
573#ifdef CONFIG_MEMCG
574 unsigned memcg_may_oom:1;
127424c8 575#ifndef CONFIG_SLOB
6f185c29
VD
576 unsigned memcg_kmem_skip_account:1;
577#endif
127424c8 578#endif
ff303e66
PZ
579#ifdef CONFIG_COMPAT_BRK
580 unsigned brk_randomized:1;
581#endif
6f185c29 582
1d4457f9
KC
583 unsigned long atomic_flags; /* Flags needing atomic access. */
584
f56141e3
AL
585 struct restart_block restart_block;
586
1da177e4
LT
587 pid_t pid;
588 pid_t tgid;
0a425405 589
1314562a 590#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
591 /* Canary value for the -fstack-protector gcc feature */
592 unsigned long stack_canary;
1314562a 593#endif
4d1d61a6 594 /*
1da177e4 595 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 596 * older sibling, respectively. (p->father can be replaced with
f470021a 597 * p->real_parent->pid)
1da177e4 598 */
abd63bc3
KC
599 struct task_struct __rcu *real_parent; /* real parent process */
600 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 601 /*
f470021a 602 * children/sibling forms the list of my natural children
1da177e4
LT
603 */
604 struct list_head children; /* list of my children */
605 struct list_head sibling; /* linkage in my parent's children list */
606 struct task_struct *group_leader; /* threadgroup leader */
607
f470021a
RM
608 /*
609 * ptraced is the list of tasks this task is using ptrace on.
610 * This includes both natural children and PTRACE_ATTACH targets.
611 * p->ptrace_entry is p's link on the p->parent->ptraced list.
612 */
613 struct list_head ptraced;
614 struct list_head ptrace_entry;
615
1da177e4 616 /* PID/PID hash table linkage. */
92476d7f 617 struct pid_link pids[PIDTYPE_MAX];
47e65328 618 struct list_head thread_group;
0c740d0a 619 struct list_head thread_node;
1da177e4
LT
620
621 struct completion *vfork_done; /* for vfork() */
622 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
623 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
624
5613fda9 625 u64 utime, stime;
40565b5a 626#ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME
5613fda9 627 u64 utimescaled, stimescaled;
40565b5a 628#endif
16a6d9be 629 u64 gtime;
9d7fb042 630 struct prev_cputime prev_cputime;
6a61671b 631#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
b7ce2277 632 seqcount_t vtime_seqcount;
6a61671b
FW
633 unsigned long long vtime_snap;
634 enum {
7098c1ea
FW
635 /* Task is sleeping or running in a CPU with VTIME inactive */
636 VTIME_INACTIVE = 0,
637 /* Task runs in userspace in a CPU with VTIME active */
6a61671b 638 VTIME_USER,
7098c1ea 639 /* Task runs in kernelspace in a CPU with VTIME active */
6a61671b
FW
640 VTIME_SYS,
641 } vtime_snap_whence;
d99ca3b9 642#endif
d027d45d
FW
643
644#ifdef CONFIG_NO_HZ_FULL
f009a7a7 645 atomic_t tick_dep_mask;
d027d45d 646#endif
1da177e4 647 unsigned long nvcsw, nivcsw; /* context switch counts */
ccbf62d8 648 u64 start_time; /* monotonic time in nsec */
57e0be04 649 u64 real_start_time; /* boot based time in nsec */
1da177e4
LT
650/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
651 unsigned long min_flt, maj_flt;
652
b18b6a9c 653#ifdef CONFIG_POSIX_TIMERS
f06febc9 654 struct task_cputime cputime_expires;
1da177e4 655 struct list_head cpu_timers[3];
b18b6a9c 656#endif
1da177e4
LT
657
658/* process credentials */
64b875f7 659 const struct cred __rcu *ptracer_cred; /* Tracer's credentials at attach */
1b0ba1c9 660 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 661 * credentials (COW) */
1b0ba1c9 662 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 663 * credentials (COW) */
36772092
PBG
664 char comm[TASK_COMM_LEN]; /* executable name excluding path
665 - access with [gs]et_task_comm (which lock
666 it with task_lock())
221af7f8 667 - initialized normally by setup_new_exec */
1da177e4 668/* file system info */
756daf26 669 struct nameidata *nameidata;
3d5b6fcc 670#ifdef CONFIG_SYSVIPC
1da177e4
LT
671/* ipc stuff */
672 struct sysv_sem sysvsem;
ab602f79 673 struct sysv_shm sysvshm;
3d5b6fcc 674#endif
e162b39a 675#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 676/* hung task detection */
82a1fcb9
IM
677 unsigned long last_switch_count;
678#endif
1da177e4
LT
679/* filesystem information */
680 struct fs_struct *fs;
681/* open file information */
682 struct files_struct *files;
1651e14e 683/* namespaces */
ab516013 684 struct nsproxy *nsproxy;
1da177e4
LT
685/* signal handlers */
686 struct signal_struct *signal;
687 struct sighand_struct *sighand;
688
689 sigset_t blocked, real_blocked;
f3de272b 690 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
691 struct sigpending pending;
692
693 unsigned long sas_ss_sp;
694 size_t sas_ss_size;
2a742138 695 unsigned sas_ss_flags;
2e01fabe 696
67d12145 697 struct callback_head *task_works;
e73f8959 698
1da177e4 699 struct audit_context *audit_context;
bfef93a5 700#ifdef CONFIG_AUDITSYSCALL
e1760bd5 701 kuid_t loginuid;
4746ec5b 702 unsigned int sessionid;
bfef93a5 703#endif
932ecebb 704 struct seccomp seccomp;
1da177e4
LT
705
706/* Thread group tracking */
707 u32 parent_exec_id;
708 u32 self_exec_id;
58568d2a
MX
709/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
710 * mempolicy */
1da177e4 711 spinlock_t alloc_lock;
1da177e4 712
b29739f9 713 /* Protection of the PI data structures: */
1d615482 714 raw_spinlock_t pi_lock;
b29739f9 715
76751049
PZ
716 struct wake_q_node wake_q;
717
23f78d4a
IM
718#ifdef CONFIG_RT_MUTEXES
719 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
720 struct rb_root pi_waiters;
721 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
722 /* Deadlock detection and priority inheritance handling */
723 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
724#endif
725
408894ee
IM
726#ifdef CONFIG_DEBUG_MUTEXES
727 /* mutex deadlock detection */
728 struct mutex_waiter *blocked_on;
729#endif
de30a2b3
IM
730#ifdef CONFIG_TRACE_IRQFLAGS
731 unsigned int irq_events;
de30a2b3 732 unsigned long hardirq_enable_ip;
de30a2b3 733 unsigned long hardirq_disable_ip;
fa1452e8 734 unsigned int hardirq_enable_event;
de30a2b3 735 unsigned int hardirq_disable_event;
fa1452e8
HS
736 int hardirqs_enabled;
737 int hardirq_context;
de30a2b3 738 unsigned long softirq_disable_ip;
de30a2b3 739 unsigned long softirq_enable_ip;
fa1452e8 740 unsigned int softirq_disable_event;
de30a2b3 741 unsigned int softirq_enable_event;
fa1452e8 742 int softirqs_enabled;
de30a2b3
IM
743 int softirq_context;
744#endif
fbb9ce95 745#ifdef CONFIG_LOCKDEP
bdb9441e 746# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
747 u64 curr_chain_key;
748 int lockdep_depth;
fbb9ce95 749 unsigned int lockdep_recursion;
c7aceaba 750 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 751 gfp_t lockdep_reclaim_gfp;
fbb9ce95 752#endif
c6d30853
AR
753#ifdef CONFIG_UBSAN
754 unsigned int in_ubsan;
755#endif
408894ee 756
1da177e4
LT
757/* journalling filesystem info */
758 void *journal_info;
759
d89d8796 760/* stacked block device info */
bddd87c7 761 struct bio_list *bio_list;
d89d8796 762
73c10101
JA
763#ifdef CONFIG_BLOCK
764/* stack plugging */
765 struct blk_plug *plug;
766#endif
767
1da177e4
LT
768/* VM state */
769 struct reclaim_state *reclaim_state;
770
1da177e4
LT
771 struct backing_dev_info *backing_dev_info;
772
773 struct io_context *io_context;
774
775 unsigned long ptrace_message;
776 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 777 struct task_io_accounting ioac;
8f0ab514 778#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
779 u64 acct_rss_mem1; /* accumulated rss usage */
780 u64 acct_vm_mem1; /* accumulated virtual memory usage */
605dc2b3 781 u64 acct_timexpd; /* stime + utime since last update */
1da177e4
LT
782#endif
783#ifdef CONFIG_CPUSETS
58568d2a 784 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 785 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 786 int cpuset_mem_spread_rotor;
6adef3eb 787 int cpuset_slab_spread_rotor;
1da177e4 788#endif
ddbcc7e8 789#ifdef CONFIG_CGROUPS
817929ec 790 /* Control Group info protected by css_set_lock */
2c392b8c 791 struct css_set __rcu *cgroups;
817929ec
PM
792 /* cg_list protected by css_set_lock and tsk->alloc_lock */
793 struct list_head cg_list;
ddbcc7e8 794#endif
e02737d5
FY
795#ifdef CONFIG_INTEL_RDT_A
796 int closid;
797#endif
42b2dd0a 798#ifdef CONFIG_FUTEX
0771dfef 799 struct robust_list_head __user *robust_list;
34f192c6
IM
800#ifdef CONFIG_COMPAT
801 struct compat_robust_list_head __user *compat_robust_list;
802#endif
c87e2837
IM
803 struct list_head pi_state_list;
804 struct futex_pi_state *pi_state_cache;
c7aceaba 805#endif
cdd6c482 806#ifdef CONFIG_PERF_EVENTS
8dc85d54 807 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
808 struct mutex perf_event_mutex;
809 struct list_head perf_event_list;
a63eaf34 810#endif
8f47b187
TG
811#ifdef CONFIG_DEBUG_PREEMPT
812 unsigned long preempt_disable_ip;
813#endif
c7aceaba 814#ifdef CONFIG_NUMA
58568d2a 815 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 816 short il_next;
207205a2 817 short pref_node_fork;
42b2dd0a 818#endif
cbee9f88
PZ
819#ifdef CONFIG_NUMA_BALANCING
820 int numa_scan_seq;
cbee9f88 821 unsigned int numa_scan_period;
598f0ec0 822 unsigned int numa_scan_period_max;
de1c9ce6 823 int numa_preferred_nid;
6b9a7460 824 unsigned long numa_migrate_retry;
cbee9f88 825 u64 node_stamp; /* migration stamp */
7e2703e6
RR
826 u64 last_task_numa_placement;
827 u64 last_sum_exec_runtime;
cbee9f88 828 struct callback_head numa_work;
f809ca9a 829
8c8a743c
PZ
830 struct list_head numa_entry;
831 struct numa_group *numa_group;
832
745d6147 833 /*
44dba3d5
IM
834 * numa_faults is an array split into four regions:
835 * faults_memory, faults_cpu, faults_memory_buffer, faults_cpu_buffer
836 * in this precise order.
837 *
838 * faults_memory: Exponential decaying average of faults on a per-node
839 * basis. Scheduling placement decisions are made based on these
840 * counts. The values remain static for the duration of a PTE scan.
841 * faults_cpu: Track the nodes the process was running on when a NUMA
842 * hinting fault was incurred.
843 * faults_memory_buffer and faults_cpu_buffer: Record faults per node
844 * during the current scan window. When the scan completes, the counts
845 * in faults_memory and faults_cpu decay and these values are copied.
745d6147 846 */
44dba3d5 847 unsigned long *numa_faults;
83e1d2cd 848 unsigned long total_numa_faults;
745d6147 849
04bb2f94
RR
850 /*
851 * numa_faults_locality tracks if faults recorded during the last
074c2381
MG
852 * scan window were remote/local or failed to migrate. The task scan
853 * period is adapted based on the locality of the faults with different
854 * weights depending on whether they were shared or private faults
04bb2f94 855 */
074c2381 856 unsigned long numa_faults_locality[3];
04bb2f94 857
b32e86b4 858 unsigned long numa_pages_migrated;
cbee9f88
PZ
859#endif /* CONFIG_NUMA_BALANCING */
860
72b252ae 861 struct tlbflush_unmap_batch tlb_ubc;
72b252ae 862
e56d0903 863 struct rcu_head rcu;
b92ce558
JA
864
865 /*
866 * cache last used pipe for splice
867 */
868 struct pipe_inode_info *splice_pipe;
5640f768
ED
869
870 struct page_frag task_frag;
871
47913d4e
IM
872#ifdef CONFIG_TASK_DELAY_ACCT
873 struct task_delay_info *delays;
f4f154fd 874#endif
47913d4e 875
f4f154fd
AM
876#ifdef CONFIG_FAULT_INJECTION
877 int make_it_fail;
ca74e92b 878#endif
9d823e8f
WF
879 /*
880 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
881 * balance_dirty_pages() for some dirty throttling pause
882 */
883 int nr_dirtied;
884 int nr_dirtied_pause;
83712358 885 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 886
9745512c
AV
887#ifdef CONFIG_LATENCYTOP
888 int latency_record_count;
889 struct latency_record latency_record[LT_SAVECOUNT];
890#endif
6976675d
AV
891 /*
892 * time slack values; these are used to round up poll() and
893 * select() etc timeout values. These are in nanoseconds.
894 */
da8b44d5
JS
895 u64 timer_slack_ns;
896 u64 default_timer_slack_ns;
f8d570a4 897
0b24becc
AR
898#ifdef CONFIG_KASAN
899 unsigned int kasan_depth;
900#endif
fb52607a 901#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 902 /* Index of current stored address in ret_stack */
f201ae23
FW
903 int curr_ret_stack;
904 /* Stack of return addresses for return function tracing */
905 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
906 /* time stamp for last schedule */
907 unsigned long long ftrace_timestamp;
f201ae23
FW
908 /*
909 * Number of functions that haven't been traced
910 * because of depth overrun.
911 */
912 atomic_t trace_overrun;
380c4b14
FW
913 /* Pause for the tracing */
914 atomic_t tracing_graph_pause;
f201ae23 915#endif
ea4e2bc4
SR
916#ifdef CONFIG_TRACING
917 /* state flags for use by tracers */
918 unsigned long trace;
b1cff0ad 919 /* bitmask and counter of trace recursion */
261842b7
SR
920 unsigned long trace_recursion;
921#endif /* CONFIG_TRACING */
5c9a8750
DV
922#ifdef CONFIG_KCOV
923 /* Coverage collection mode enabled for this task (0 if disabled). */
924 enum kcov_mode kcov_mode;
925 /* Size of the kcov_area. */
926 unsigned kcov_size;
927 /* Buffer for coverage collection. */
928 void *kcov_area;
929 /* kcov desciptor wired with this task or NULL. */
930 struct kcov *kcov;
931#endif
6f185c29 932#ifdef CONFIG_MEMCG
626ebc41
TH
933 struct mem_cgroup *memcg_in_oom;
934 gfp_t memcg_oom_gfp_mask;
935 int memcg_oom_order;
b23afb93
TH
936
937 /* number of pages to reclaim on returning to userland */
938 unsigned int memcg_nr_pages_over_high;
569b846d 939#endif
0326f5a9
SD
940#ifdef CONFIG_UPROBES
941 struct uprobe_task *utask;
0326f5a9 942#endif
cafe5635
KO
943#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
944 unsigned int sequential_io;
945 unsigned int sequential_io_avg;
946#endif
8eb23b9f
PZ
947#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
948 unsigned long task_state_change;
949#endif
8bcbde54 950 int pagefault_disabled;
03049269 951#ifdef CONFIG_MMU
29c696e1 952 struct task_struct *oom_reaper_list;
03049269 953#endif
ba14a194
AL
954#ifdef CONFIG_VMAP_STACK
955 struct vm_struct *stack_vm_area;
956#endif
68f24b08
AL
957#ifdef CONFIG_THREAD_INFO_IN_TASK
958 /* A live task holds one reference. */
959 atomic_t stack_refcount;
960#endif
0c8c0f03
DH
961/* CPU-specific state of this task */
962 struct thread_struct thread;
963/*
964 * WARNING: on x86, 'thread_struct' contains a variable-sized
965 * structure. It *MUST* be at the end of 'task_struct'.
966 *
967 * Do not put anything below here!
968 */
1da177e4
LT
969};
970
e868171a 971static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
972{
973 return task->pids[PIDTYPE_PID].pid;
974}
975
e868171a 976static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
977{
978 return task->group_leader->pids[PIDTYPE_PID].pid;
979}
980
6dda81f4
ON
981/*
982 * Without tasklist or rcu lock it is not safe to dereference
983 * the result of task_pgrp/task_session even if task == current,
984 * we can race with another thread doing sys_setsid/sys_setpgid.
985 */
e868171a 986static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
987{
988 return task->group_leader->pids[PIDTYPE_PGID].pid;
989}
990
e868171a 991static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
992{
993 return task->group_leader->pids[PIDTYPE_SID].pid;
994}
995
7af57294
PE
996/*
997 * the helpers to get the task's different pids as they are seen
998 * from various namespaces
999 *
1000 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1001 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1002 * current.
7af57294
PE
1003 * task_xid_nr_ns() : id seen from the ns specified;
1004 *
1005 * set_task_vxid() : assigns a virtual id to a task;
1006 *
7af57294
PE
1007 * see also pid_nr() etc in include/linux/pid.h
1008 */
52ee2dfd
ON
1009pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1010 struct pid_namespace *ns);
7af57294 1011
e868171a 1012static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1013{
1014 return tsk->pid;
1015}
1016
52ee2dfd
ON
1017static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1018 struct pid_namespace *ns)
1019{
1020 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1021}
7af57294
PE
1022
1023static inline pid_t task_pid_vnr(struct task_struct *tsk)
1024{
52ee2dfd 1025 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1026}
1027
1028
e868171a 1029static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1030{
1031 return tsk->tgid;
1032}
1033
2f2a3a46 1034pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1035
1036static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1037{
1038 return pid_vnr(task_tgid(tsk));
1039}
1040
1041
80e0b6e8 1042static inline int pid_alive(const struct task_struct *p);
ad36d282
RGB
1043static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns)
1044{
1045 pid_t pid = 0;
1046
1047 rcu_read_lock();
1048 if (pid_alive(tsk))
1049 pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns);
1050 rcu_read_unlock();
1051
1052 return pid;
1053}
1054
1055static inline pid_t task_ppid_nr(const struct task_struct *tsk)
1056{
1057 return task_ppid_nr_ns(tsk, &init_pid_ns);
1058}
1059
52ee2dfd
ON
1060static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1061 struct pid_namespace *ns)
7af57294 1062{
52ee2dfd 1063 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1064}
1065
7af57294
PE
1066static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1067{
52ee2dfd 1068 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1069}
1070
1071
52ee2dfd
ON
1072static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1073 struct pid_namespace *ns)
7af57294 1074{
52ee2dfd 1075 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1076}
1077
7af57294
PE
1078static inline pid_t task_session_vnr(struct task_struct *tsk)
1079{
52ee2dfd 1080 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1081}
1082
1b0f7ffd
ON
1083/* obsolete, do not use */
1084static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1085{
1086 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1087}
7af57294 1088
1da177e4
LT
1089/**
1090 * pid_alive - check that a task structure is not stale
1091 * @p: Task structure to be checked.
1092 *
1093 * Test if a process is not yet dead (at most zombie state)
1094 * If pid_alive fails, then pointers within the task structure
1095 * can be stale and must not be dereferenced.
e69f6186
YB
1096 *
1097 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1098 */
ad36d282 1099static inline int pid_alive(const struct task_struct *p)
1da177e4 1100{
92476d7f 1101 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1102}
1103
f400e198 1104/**
570f5241
SS
1105 * is_global_init - check if a task structure is init. Since init
1106 * is free to have sub-threads we need to check tgid.
3260259f
HK
1107 * @tsk: Task structure to be checked.
1108 *
1109 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1110 *
1111 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1112 */
e868171a 1113static inline int is_global_init(struct task_struct *tsk)
b461cc03 1114{
570f5241 1115 return task_tgid_nr(tsk) == 1;
b461cc03 1116}
b460cbc5 1117
9ec52099
CLG
1118extern struct pid *cad_pid;
1119
1da177e4
LT
1120/*
1121 * Per process flags
1122 */
c1de45ca 1123#define PF_IDLE 0x00000002 /* I am an IDLE thread */
1da177e4 1124#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1125#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1126#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1127#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1128#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1129#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1130#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1131#define PF_DUMPCORE 0x00000200 /* dumped core */
1132#define PF_SIGNALED 0x00000400 /* killed by a signal */
1133#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1134#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1135#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1136#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1137#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1138#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1139#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1140#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1141#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1142#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1143#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1144#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1145#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
14a40ffc 1146#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1147#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
61a87122 1148#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1149#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1150#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1151
1152/*
1153 * Only the _current_ task can read/write to tsk->flags, but other
1154 * tasks can access tsk->flags in readonly mode for example
1155 * with tsk_used_math (like during threaded core dumping).
1156 * There is however an exception to this rule during ptrace
1157 * or during fork: the ptracer task is allowed to write to the
1158 * child->flags of its traced child (same goes for fork, the parent
1159 * can write to the child->flags), because we're guaranteed the
1160 * child is not running and in turn not changing child->flags
1161 * at the same time the parent does it.
1162 */
1163#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1164#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1165#define clear_used_math() clear_stopped_child_used_math(current)
1166#define set_used_math() set_stopped_child_used_math(current)
1167#define conditional_stopped_child_used_math(condition, child) \
1168 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1169#define conditional_used_math(condition) \
1170 conditional_stopped_child_used_math(condition, current)
1171#define copy_to_stopped_child_used_math(child) \
1172 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1173/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1174#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1175#define used_math() tsk_used_math(current)
1176
1d4457f9 1177/* Per-process atomic flags. */
a2b86f77 1178#define PFA_NO_NEW_PRIVS 0 /* May not gain new privileges. */
2ad654bc
ZL
1179#define PFA_SPREAD_PAGE 1 /* Spread page cache over cpuset */
1180#define PFA_SPREAD_SLAB 2 /* Spread some slab caches over cpuset */
77ed2c57 1181#define PFA_LMK_WAITING 3 /* Lowmemorykiller is waiting */
2ad654bc 1182
1d4457f9 1183
e0e5070b
ZL
1184#define TASK_PFA_TEST(name, func) \
1185 static inline bool task_##func(struct task_struct *p) \
1186 { return test_bit(PFA_##name, &p->atomic_flags); }
1187#define TASK_PFA_SET(name, func) \
1188 static inline void task_set_##func(struct task_struct *p) \
1189 { set_bit(PFA_##name, &p->atomic_flags); }
1190#define TASK_PFA_CLEAR(name, func) \
1191 static inline void task_clear_##func(struct task_struct *p) \
1192 { clear_bit(PFA_##name, &p->atomic_flags); }
1193
1194TASK_PFA_TEST(NO_NEW_PRIVS, no_new_privs)
1195TASK_PFA_SET(NO_NEW_PRIVS, no_new_privs)
1d4457f9 1196
2ad654bc
ZL
1197TASK_PFA_TEST(SPREAD_PAGE, spread_page)
1198TASK_PFA_SET(SPREAD_PAGE, spread_page)
1199TASK_PFA_CLEAR(SPREAD_PAGE, spread_page)
1200
1201TASK_PFA_TEST(SPREAD_SLAB, spread_slab)
1202TASK_PFA_SET(SPREAD_SLAB, spread_slab)
1203TASK_PFA_CLEAR(SPREAD_SLAB, spread_slab)
1d4457f9 1204
77ed2c57
TH
1205TASK_PFA_TEST(LMK_WAITING, lmk_waiting)
1206TASK_PFA_SET(LMK_WAITING, lmk_waiting)
1207
907aed48
MG
1208static inline void tsk_restore_flags(struct task_struct *task,
1209 unsigned long orig_flags, unsigned long flags)
1210{
1211 task->flags &= ~flags;
1212 task->flags |= orig_flags & flags;
1213}
1214
f82f8042
JL
1215extern int cpuset_cpumask_can_shrink(const struct cpumask *cur,
1216 const struct cpumask *trial);
7f51412a
JL
1217extern int task_can_attach(struct task_struct *p,
1218 const struct cpumask *cs_cpus_allowed);
1da177e4 1219#ifdef CONFIG_SMP
1e1b6c51
KM
1220extern void do_set_cpus_allowed(struct task_struct *p,
1221 const struct cpumask *new_mask);
1222
cd8ba7cd 1223extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1224 const struct cpumask *new_mask);
1da177e4 1225#else
1e1b6c51
KM
1226static inline void do_set_cpus_allowed(struct task_struct *p,
1227 const struct cpumask *new_mask)
1228{
1229}
cd8ba7cd 1230static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1231 const struct cpumask *new_mask)
1da177e4 1232{
96f874e2 1233 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1234 return -EINVAL;
1235 return 0;
1236}
1237#endif
e0ad9556 1238
6d0d2878
CB
1239#ifndef cpu_relax_yield
1240#define cpu_relax_yield() cpu_relax()
1241#endif
1242
fa93384f 1243extern int yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
1244extern void set_user_nice(struct task_struct *p, long nice);
1245extern int task_prio(const struct task_struct *p);
d0ea0268
DY
1246/**
1247 * task_nice - return the nice value of a given task.
1248 * @p: the task in question.
1249 *
1250 * Return: The nice value [ -20 ... 0 ... 19 ].
1251 */
1252static inline int task_nice(const struct task_struct *p)
1253{
1254 return PRIO_TO_NICE((p)->static_prio);
1255}
36c8b586
IM
1256extern int can_nice(const struct task_struct *p, const int nice);
1257extern int task_curr(const struct task_struct *p);
1da177e4 1258extern int idle_cpu(int cpu);
fe7de49f
KM
1259extern int sched_setscheduler(struct task_struct *, int,
1260 const struct sched_param *);
961ccddd 1261extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 1262 const struct sched_param *);
d50dde5a
DF
1263extern int sched_setattr(struct task_struct *,
1264 const struct sched_attr *);
36c8b586 1265extern struct task_struct *idle_task(int cpu);
c4f30608
PM
1266/**
1267 * is_idle_task - is the specified task an idle task?
fa757281 1268 * @p: the task in question.
e69f6186
YB
1269 *
1270 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 1271 */
7061ca3b 1272static inline bool is_idle_task(const struct task_struct *p)
c4f30608 1273{
c1de45ca 1274 return !!(p->flags & PF_IDLE);
c4f30608 1275}
36c8b586 1276extern struct task_struct *curr_task(int cpu);
a458ae2e 1277extern void ia64_set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1278
1279void yield(void);
1280
1da177e4 1281union thread_union {
c65eacbe 1282#ifndef CONFIG_THREAD_INFO_IN_TASK
1da177e4 1283 struct thread_info thread_info;
c65eacbe 1284#endif
1da177e4
LT
1285 unsigned long stack[THREAD_SIZE/sizeof(long)];
1286};
1287
f3ac6067
IM
1288#ifdef CONFIG_THREAD_INFO_IN_TASK
1289static inline struct thread_info *task_thread_info(struct task_struct *task)
1290{
1291 return &task->thread_info;
1292}
1293#elif !defined(__HAVE_THREAD_FUNCTIONS)
1294# define task_thread_info(task) ((struct thread_info *)(task)->stack)
1295#endif
1296
198fe21b
PE
1297/*
1298 * find a task by one of its numerical ids
1299 *
198fe21b
PE
1300 * find_task_by_pid_ns():
1301 * finds a task by its pid in the specified namespace
228ebcbe
PE
1302 * find_task_by_vpid():
1303 * finds a task by its virtual pid
198fe21b 1304 *
e49859e7 1305 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
1306 */
1307
228ebcbe
PE
1308extern struct task_struct *find_task_by_vpid(pid_t nr);
1309extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1310 struct pid_namespace *ns);
198fe21b 1311
b3c97528
HH
1312extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1313extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 1314extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
1315#ifdef CONFIG_SMP
1316 extern void kick_process(struct task_struct *tsk);
1317#else
1318 static inline void kick_process(struct task_struct *tsk) { }
1319#endif
1da177e4 1320
82b89778
AH
1321extern void __set_task_comm(struct task_struct *tsk, const char *from, bool exec);
1322static inline void set_task_comm(struct task_struct *tsk, const char *from)
1323{
1324 __set_task_comm(tsk, from, false);
1325}
59714d65 1326extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
1327
1328#ifdef CONFIG_SMP
317f3941 1329void scheduler_ipi(void);
85ba2d86 1330extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 1331#else
184748cc 1332static inline void scheduler_ipi(void) { }
85ba2d86
RM
1333static inline unsigned long wait_task_inactive(struct task_struct *p,
1334 long match_state)
1335{
1336 return 1;
1337}
1da177e4
LT
1338#endif
1339
1da177e4
LT
1340/* set thread flags in other task's structures
1341 * - see asm/thread_info.h for TIF_xxxx flags available
1342 */
1343static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1344{
a1261f54 1345 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1346}
1347
1348static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1349{
a1261f54 1350 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1351}
1352
1353static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1354{
a1261f54 1355 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1356}
1357
1358static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1359{
a1261f54 1360 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1361}
1362
1363static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1364{
a1261f54 1365 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1366}
1367
1368static inline void set_tsk_need_resched(struct task_struct *tsk)
1369{
1370 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1371}
1372
1373static inline void clear_tsk_need_resched(struct task_struct *tsk)
1374{
1375 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1376}
1377
8ae121ac
GH
1378static inline int test_tsk_need_resched(struct task_struct *tsk)
1379{
1380 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
1381}
1382
1da177e4
LT
1383/*
1384 * cond_resched() and cond_resched_lock(): latency reduction via
1385 * explicit rescheduling in places that are safe. The return
1386 * value indicates whether a reschedule was done in fact.
1387 * cond_resched_lock() will drop the spinlock before scheduling,
1388 * cond_resched_softirq() will enable bhs before scheduling.
1389 */
35a773a0 1390#ifndef CONFIG_PREEMPT
c3921ab7 1391extern int _cond_resched(void);
35a773a0
PZ
1392#else
1393static inline int _cond_resched(void) { return 0; }
1394#endif
6f80bd98 1395
613afbf8 1396#define cond_resched() ({ \
3427445a 1397 ___might_sleep(__FILE__, __LINE__, 0); \
613afbf8
FW
1398 _cond_resched(); \
1399})
6f80bd98 1400
613afbf8
FW
1401extern int __cond_resched_lock(spinlock_t *lock);
1402
1403#define cond_resched_lock(lock) ({ \
3427445a 1404 ___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\
613afbf8
FW
1405 __cond_resched_lock(lock); \
1406})
1407
1408extern int __cond_resched_softirq(void);
1409
75e1056f 1410#define cond_resched_softirq() ({ \
3427445a 1411 ___might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
75e1056f 1412 __cond_resched_softirq(); \
613afbf8 1413})
1da177e4 1414
f6f3c437
SH
1415static inline void cond_resched_rcu(void)
1416{
1417#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
1418 rcu_read_unlock();
1419 cond_resched();
1420 rcu_read_lock();
1421#endif
1422}
1423
1da177e4
LT
1424/*
1425 * Does a critical section need to be broken due to another
95c354fe
NP
1426 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
1427 * but a general need for low latency)
1da177e4 1428 */
95c354fe 1429static inline int spin_needbreak(spinlock_t *lock)
1da177e4 1430{
95c354fe
NP
1431#ifdef CONFIG_PREEMPT
1432 return spin_is_contended(lock);
1433#else
1da177e4 1434 return 0;
95c354fe 1435#endif
1da177e4
LT
1436}
1437
75f93fed
PZ
1438static __always_inline bool need_resched(void)
1439{
1440 return unlikely(tif_need_resched());
1441}
1442
1da177e4
LT
1443/*
1444 * Wrappers for p->thread_info->cpu access. No-op on UP.
1445 */
1446#ifdef CONFIG_SMP
1447
1448static inline unsigned int task_cpu(const struct task_struct *p)
1449{
c65eacbe
AL
1450#ifdef CONFIG_THREAD_INFO_IN_TASK
1451 return p->cpu;
1452#else
a1261f54 1453 return task_thread_info(p)->cpu;
c65eacbe 1454#endif
1da177e4
LT
1455}
1456
c65cc870 1457extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
1458
1459#else
1460
1461static inline unsigned int task_cpu(const struct task_struct *p)
1462{
1463 return 0;
1464}
1465
1466static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1467{
1468}
1469
1470#endif /* CONFIG_SMP */
1471
d9345c65
PX
1472/*
1473 * In order to reduce various lock holder preemption latencies provide an
1474 * interface to see if a vCPU is currently running or not.
1475 *
1476 * This allows us to terminate optimistic spin loops and block, analogous to
1477 * the native optimistic spin heuristic of testing if the lock owner task is
1478 * running or not.
1479 */
1480#ifndef vcpu_is_preempted
1481# define vcpu_is_preempted(cpu) false
1482#endif
1483
96f874e2
RR
1484extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
1485extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 1486
82455257
DH
1487#ifndef TASK_SIZE_OF
1488#define TASK_SIZE_OF(tsk) TASK_SIZE
1489#endif
1490
1da177e4 1491#endif