Commit | Line | Data |
---|---|---|
b2441318 | 1 | /* SPDX-License-Identifier: GPL-2.0 */ |
1da177e4 LT |
2 | #ifndef _LINUX_SCHED_H |
3 | #define _LINUX_SCHED_H | |
4 | ||
5eca1c10 IM |
5 | /* |
6 | * Define 'struct task_struct' and provide the main scheduler | |
7 | * APIs (schedule(), wakeup variants, etc.) | |
8 | */ | |
b7b3c76a | 9 | |
5eca1c10 | 10 | #include <uapi/linux/sched.h> |
5c228079 | 11 | |
5eca1c10 | 12 | #include <asm/current.h> |
1da177e4 | 13 | |
5eca1c10 | 14 | #include <linux/pid.h> |
1da177e4 | 15 | #include <linux/sem.h> |
ab602f79 | 16 | #include <linux/shm.h> |
5eca1c10 IM |
17 | #include <linux/kcov.h> |
18 | #include <linux/mutex.h> | |
19 | #include <linux/plist.h> | |
20 | #include <linux/hrtimer.h> | |
1da177e4 | 21 | #include <linux/seccomp.h> |
5eca1c10 | 22 | #include <linux/nodemask.h> |
b68070e1 | 23 | #include <linux/rcupdate.h> |
ec1d2819 | 24 | #include <linux/refcount.h> |
a3b6714e | 25 | #include <linux/resource.h> |
9745512c | 26 | #include <linux/latencytop.h> |
5eca1c10 | 27 | #include <linux/sched/prio.h> |
9eacb5c7 | 28 | #include <linux/sched/types.h> |
5eca1c10 IM |
29 | #include <linux/signal_types.h> |
30 | #include <linux/mm_types_task.h> | |
31 | #include <linux/task_io_accounting.h> | |
2b69942f | 32 | #include <linux/posix-timers.h> |
d7822b1e | 33 | #include <linux/rseq.h> |
a3b6714e | 34 | |
5eca1c10 | 35 | /* task_struct member predeclarations (sorted alphabetically): */ |
c7af7877 | 36 | struct audit_context; |
c7af7877 | 37 | struct backing_dev_info; |
bddd87c7 | 38 | struct bio_list; |
73c10101 | 39 | struct blk_plug; |
3c93a0c0 | 40 | struct capture_control; |
c7af7877 | 41 | struct cfs_rq; |
c7af7877 IM |
42 | struct fs_struct; |
43 | struct futex_pi_state; | |
44 | struct io_context; | |
45 | struct mempolicy; | |
89076bc3 | 46 | struct nameidata; |
c7af7877 IM |
47 | struct nsproxy; |
48 | struct perf_event_context; | |
49 | struct pid_namespace; | |
50 | struct pipe_inode_info; | |
51 | struct rcu_node; | |
52 | struct reclaim_state; | |
53 | struct robust_list_head; | |
3c93a0c0 QY |
54 | struct root_domain; |
55 | struct rq; | |
c7af7877 IM |
56 | struct sched_attr; |
57 | struct sched_param; | |
43ae34cb | 58 | struct seq_file; |
c7af7877 IM |
59 | struct sighand_struct; |
60 | struct signal_struct; | |
61 | struct task_delay_info; | |
4cf86d77 | 62 | struct task_group; |
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 | */ | |
5eca1c10 IM |
74 | |
75 | /* Used in tsk->state: */ | |
92c4bc9f PZ |
76 | #define TASK_RUNNING 0x0000 |
77 | #define TASK_INTERRUPTIBLE 0x0001 | |
78 | #define TASK_UNINTERRUPTIBLE 0x0002 | |
79 | #define __TASK_STOPPED 0x0004 | |
80 | #define __TASK_TRACED 0x0008 | |
5eca1c10 | 81 | /* Used in tsk->exit_state: */ |
92c4bc9f PZ |
82 | #define EXIT_DEAD 0x0010 |
83 | #define EXIT_ZOMBIE 0x0020 | |
5eca1c10 IM |
84 | #define EXIT_TRACE (EXIT_ZOMBIE | EXIT_DEAD) |
85 | /* Used in tsk->state again: */ | |
8ef9925b PZ |
86 | #define TASK_PARKED 0x0040 |
87 | #define TASK_DEAD 0x0080 | |
88 | #define TASK_WAKEKILL 0x0100 | |
89 | #define TASK_WAKING 0x0200 | |
92c4bc9f PZ |
90 | #define TASK_NOLOAD 0x0400 |
91 | #define TASK_NEW 0x0800 | |
92 | #define TASK_STATE_MAX 0x1000 | |
5eca1c10 | 93 | |
5eca1c10 IM |
94 | /* Convenience macros for the sake of set_current_state: */ |
95 | #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE) | |
96 | #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED) | |
97 | #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED) | |
98 | ||
99 | #define TASK_IDLE (TASK_UNINTERRUPTIBLE | TASK_NOLOAD) | |
100 | ||
101 | /* Convenience macros for the sake of wake_up(): */ | |
102 | #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE) | |
5eca1c10 IM |
103 | |
104 | /* get_task_state(): */ | |
105 | #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \ | |
106 | TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \ | |
8ef9925b PZ |
107 | __TASK_TRACED | EXIT_DEAD | EXIT_ZOMBIE | \ |
108 | TASK_PARKED) | |
5eca1c10 IM |
109 | |
110 | #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0) | |
111 | ||
112 | #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0) | |
113 | ||
114 | #define task_is_stopped_or_traced(task) ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0) | |
115 | ||
116 | #define task_contributes_to_load(task) ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \ | |
117 | (task->flags & PF_FROZEN) == 0 && \ | |
118 | (task->state & TASK_NOLOAD) == 0) | |
1da177e4 | 119 | |
8eb23b9f PZ |
120 | #ifdef CONFIG_DEBUG_ATOMIC_SLEEP |
121 | ||
b5bf9a90 PZ |
122 | /* |
123 | * Special states are those that do not use the normal wait-loop pattern. See | |
124 | * the comment with set_special_state(). | |
125 | */ | |
126 | #define is_special_task_state(state) \ | |
1cef1150 | 127 | ((state) & (__TASK_STOPPED | __TASK_TRACED | TASK_PARKED | TASK_DEAD)) |
b5bf9a90 | 128 | |
8eb23b9f PZ |
129 | #define __set_current_state(state_value) \ |
130 | do { \ | |
b5bf9a90 | 131 | WARN_ON_ONCE(is_special_task_state(state_value));\ |
8eb23b9f PZ |
132 | current->task_state_change = _THIS_IP_; \ |
133 | current->state = (state_value); \ | |
134 | } while (0) | |
b5bf9a90 | 135 | |
8eb23b9f PZ |
136 | #define set_current_state(state_value) \ |
137 | do { \ | |
b5bf9a90 | 138 | WARN_ON_ONCE(is_special_task_state(state_value));\ |
8eb23b9f | 139 | current->task_state_change = _THIS_IP_; \ |
a2250238 | 140 | smp_store_mb(current->state, (state_value)); \ |
8eb23b9f PZ |
141 | } while (0) |
142 | ||
b5bf9a90 PZ |
143 | #define set_special_state(state_value) \ |
144 | do { \ | |
145 | unsigned long flags; /* may shadow */ \ | |
146 | WARN_ON_ONCE(!is_special_task_state(state_value)); \ | |
147 | raw_spin_lock_irqsave(¤t->pi_lock, flags); \ | |
148 | current->task_state_change = _THIS_IP_; \ | |
149 | current->state = (state_value); \ | |
150 | raw_spin_unlock_irqrestore(¤t->pi_lock, flags); \ | |
151 | } while (0) | |
8eb23b9f | 152 | #else |
498d0c57 AM |
153 | /* |
154 | * set_current_state() includes a barrier so that the write of current->state | |
155 | * is correctly serialised wrt the caller's subsequent test of whether to | |
156 | * actually sleep: | |
157 | * | |
a2250238 | 158 | * for (;;) { |
498d0c57 | 159 | * set_current_state(TASK_UNINTERRUPTIBLE); |
a2250238 PZ |
160 | * if (!need_sleep) |
161 | * break; | |
162 | * | |
163 | * schedule(); | |
164 | * } | |
165 | * __set_current_state(TASK_RUNNING); | |
166 | * | |
167 | * If the caller does not need such serialisation (because, for instance, the | |
168 | * condition test and condition change and wakeup are under the same lock) then | |
169 | * use __set_current_state(). | |
170 | * | |
171 | * The above is typically ordered against the wakeup, which does: | |
172 | * | |
b5bf9a90 PZ |
173 | * need_sleep = false; |
174 | * wake_up_state(p, TASK_UNINTERRUPTIBLE); | |
a2250238 | 175 | * |
7696f991 AP |
176 | * where wake_up_state() executes a full memory barrier before accessing the |
177 | * task state. | |
a2250238 PZ |
178 | * |
179 | * Wakeup will do: if (@state & p->state) p->state = TASK_RUNNING, that is, | |
180 | * once it observes the TASK_UNINTERRUPTIBLE store the waking CPU can issue a | |
181 | * TASK_RUNNING store which can collide with __set_current_state(TASK_RUNNING). | |
498d0c57 | 182 | * |
b5bf9a90 | 183 | * However, with slightly different timing the wakeup TASK_RUNNING store can |
dfcb245e | 184 | * also collide with the TASK_UNINTERRUPTIBLE store. Losing that store is not |
b5bf9a90 PZ |
185 | * a problem either because that will result in one extra go around the loop |
186 | * and our @cond test will save the day. | |
498d0c57 | 187 | * |
a2250238 | 188 | * Also see the comments of try_to_wake_up(). |
498d0c57 | 189 | */ |
b5bf9a90 PZ |
190 | #define __set_current_state(state_value) \ |
191 | current->state = (state_value) | |
192 | ||
193 | #define set_current_state(state_value) \ | |
194 | smp_store_mb(current->state, (state_value)) | |
195 | ||
196 | /* | |
197 | * set_special_state() should be used for those states when the blocking task | |
198 | * can not use the regular condition based wait-loop. In that case we must | |
199 | * serialize against wakeups such that any possible in-flight TASK_RUNNING stores | |
200 | * will not collide with our state change. | |
201 | */ | |
202 | #define set_special_state(state_value) \ | |
203 | do { \ | |
204 | unsigned long flags; /* may shadow */ \ | |
205 | raw_spin_lock_irqsave(¤t->pi_lock, flags); \ | |
206 | current->state = (state_value); \ | |
207 | raw_spin_unlock_irqrestore(¤t->pi_lock, flags); \ | |
208 | } while (0) | |
209 | ||
8eb23b9f PZ |
210 | #endif |
211 | ||
5eca1c10 IM |
212 | /* Task command name length: */ |
213 | #define TASK_COMM_LEN 16 | |
1da177e4 | 214 | |
1da177e4 LT |
215 | extern void scheduler_tick(void); |
216 | ||
5eca1c10 IM |
217 | #define MAX_SCHEDULE_TIMEOUT LONG_MAX |
218 | ||
219 | extern long schedule_timeout(long timeout); | |
220 | extern long schedule_timeout_interruptible(long timeout); | |
221 | extern long schedule_timeout_killable(long timeout); | |
222 | extern long schedule_timeout_uninterruptible(long timeout); | |
223 | extern long schedule_timeout_idle(long timeout); | |
1da177e4 | 224 | asmlinkage void schedule(void); |
c5491ea7 | 225 | extern void schedule_preempt_disabled(void); |
19c95f26 | 226 | asmlinkage void preempt_schedule_irq(void); |
1da177e4 | 227 | |
10ab5643 TH |
228 | extern int __must_check io_schedule_prepare(void); |
229 | extern void io_schedule_finish(int token); | |
9cff8ade | 230 | extern long io_schedule_timeout(long timeout); |
10ab5643 | 231 | extern void io_schedule(void); |
9cff8ade | 232 | |
d37f761d | 233 | /** |
0ba42a59 | 234 | * struct prev_cputime - snapshot of system and user cputime |
d37f761d FW |
235 | * @utime: time spent in user mode |
236 | * @stime: time spent in system mode | |
9d7fb042 | 237 | * @lock: protects the above two fields |
d37f761d | 238 | * |
9d7fb042 PZ |
239 | * Stores previous user/system time values such that we can guarantee |
240 | * monotonicity. | |
d37f761d | 241 | */ |
9d7fb042 PZ |
242 | struct prev_cputime { |
243 | #ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE | |
5eca1c10 IM |
244 | u64 utime; |
245 | u64 stime; | |
246 | raw_spinlock_t lock; | |
9d7fb042 | 247 | #endif |
d37f761d FW |
248 | }; |
249 | ||
bac5b6b6 FW |
250 | enum vtime_state { |
251 | /* Task is sleeping or running in a CPU with VTIME inactive: */ | |
252 | VTIME_INACTIVE = 0, | |
14faf6fc FW |
253 | /* Task is idle */ |
254 | VTIME_IDLE, | |
bac5b6b6 FW |
255 | /* Task runs in kernelspace in a CPU with VTIME active: */ |
256 | VTIME_SYS, | |
14faf6fc FW |
257 | /* Task runs in userspace in a CPU with VTIME active: */ |
258 | VTIME_USER, | |
e6d5bf3e FW |
259 | /* Task runs as guests in a CPU with VTIME active: */ |
260 | VTIME_GUEST, | |
bac5b6b6 FW |
261 | }; |
262 | ||
263 | struct vtime { | |
264 | seqcount_t seqcount; | |
265 | unsigned long long starttime; | |
266 | enum vtime_state state; | |
802f4a82 | 267 | unsigned int cpu; |
2a42eb95 WL |
268 | u64 utime; |
269 | u64 stime; | |
270 | u64 gtime; | |
bac5b6b6 FW |
271 | }; |
272 | ||
69842cba PB |
273 | /* |
274 | * Utilization clamp constraints. | |
275 | * @UCLAMP_MIN: Minimum utilization | |
276 | * @UCLAMP_MAX: Maximum utilization | |
277 | * @UCLAMP_CNT: Utilization clamp constraints count | |
278 | */ | |
279 | enum uclamp_id { | |
280 | UCLAMP_MIN = 0, | |
281 | UCLAMP_MAX, | |
282 | UCLAMP_CNT | |
283 | }; | |
284 | ||
f9a25f77 MP |
285 | #ifdef CONFIG_SMP |
286 | extern struct root_domain def_root_domain; | |
287 | extern struct mutex sched_domains_mutex; | |
288 | #endif | |
289 | ||
1da177e4 | 290 | struct sched_info { |
7f5f8e8d | 291 | #ifdef CONFIG_SCHED_INFO |
5eca1c10 IM |
292 | /* Cumulative counters: */ |
293 | ||
294 | /* # of times we have run on this CPU: */ | |
295 | unsigned long pcount; | |
296 | ||
297 | /* Time spent waiting on a runqueue: */ | |
298 | unsigned long long run_delay; | |
299 | ||
300 | /* Timestamps: */ | |
301 | ||
302 | /* When did we last run on a CPU? */ | |
303 | unsigned long long last_arrival; | |
304 | ||
305 | /* When were we last queued to run? */ | |
306 | unsigned long long last_queued; | |
1da177e4 | 307 | |
f6db8347 | 308 | #endif /* CONFIG_SCHED_INFO */ |
7f5f8e8d | 309 | }; |
1da177e4 | 310 | |
6ecdd749 YD |
311 | /* |
312 | * Integer metrics need fixed point arithmetic, e.g., sched/fair | |
313 | * has a few: load, load_avg, util_avg, freq, and capacity. | |
314 | * | |
315 | * We define a basic fixed point arithmetic range, and then formalize | |
316 | * all these metrics based on that basic range. | |
317 | */ | |
5eca1c10 IM |
318 | # define SCHED_FIXEDPOINT_SHIFT 10 |
319 | # define SCHED_FIXEDPOINT_SCALE (1L << SCHED_FIXEDPOINT_SHIFT) | |
6ecdd749 | 320 | |
69842cba PB |
321 | /* Increase resolution of cpu_capacity calculations */ |
322 | # define SCHED_CAPACITY_SHIFT SCHED_FIXEDPOINT_SHIFT | |
323 | # define SCHED_CAPACITY_SCALE (1L << SCHED_CAPACITY_SHIFT) | |
324 | ||
20b8a59f | 325 | struct load_weight { |
5eca1c10 IM |
326 | unsigned long weight; |
327 | u32 inv_weight; | |
20b8a59f IM |
328 | }; |
329 | ||
7f65ea42 PB |
330 | /** |
331 | * struct util_est - Estimation utilization of FAIR tasks | |
332 | * @enqueued: instantaneous estimated utilization of a task/cpu | |
333 | * @ewma: the Exponential Weighted Moving Average (EWMA) | |
334 | * utilization of a task | |
335 | * | |
336 | * Support data structure to track an Exponential Weighted Moving Average | |
337 | * (EWMA) of a FAIR task's utilization. New samples are added to the moving | |
338 | * average each time a task completes an activation. Sample's weight is chosen | |
339 | * so that the EWMA will be relatively insensitive to transient changes to the | |
340 | * task's workload. | |
341 | * | |
342 | * The enqueued attribute has a slightly different meaning for tasks and cpus: | |
343 | * - task: the task's util_avg at last task dequeue time | |
344 | * - cfs_rq: the sum of util_est.enqueued for each RUNNABLE task on that CPU | |
345 | * Thus, the util_est.enqueued of a task represents the contribution on the | |
346 | * estimated utilization of the CPU where that task is currently enqueued. | |
347 | * | |
348 | * Only for tasks we track a moving average of the past instantaneous | |
349 | * estimated utilization. This allows to absorb sporadic drops in utilization | |
350 | * of an otherwise almost periodic task. | |
351 | */ | |
352 | struct util_est { | |
353 | unsigned int enqueued; | |
354 | unsigned int ewma; | |
355 | #define UTIL_EST_WEIGHT_SHIFT 2 | |
317d359d | 356 | } __attribute__((__aligned__(sizeof(u64)))); |
7f65ea42 | 357 | |
9d89c257 | 358 | /* |
7b595334 | 359 | * The load_avg/util_avg accumulates an infinite geometric series |
0dacee1b | 360 | * (see __update_load_avg_cfs_rq() in kernel/sched/pelt.c). |
7b595334 YD |
361 | * |
362 | * [load_avg definition] | |
363 | * | |
364 | * load_avg = runnable% * scale_load_down(load) | |
365 | * | |
366 | * where runnable% is the time ratio that a sched_entity is runnable. | |
367 | * For cfs_rq, it is the aggregated load_avg of all runnable and | |
9d89c257 | 368 | * blocked sched_entities. |
7b595334 | 369 | * |
7b595334 YD |
370 | * [util_avg definition] |
371 | * | |
372 | * util_avg = running% * SCHED_CAPACITY_SCALE | |
373 | * | |
374 | * where running% is the time ratio that a sched_entity is running on | |
375 | * a CPU. For cfs_rq, it is the aggregated util_avg of all runnable | |
376 | * and blocked sched_entities. | |
377 | * | |
23127296 VG |
378 | * load_avg and util_avg don't direcly factor frequency scaling and CPU |
379 | * capacity scaling. The scaling is done through the rq_clock_pelt that | |
380 | * is used for computing those signals (see update_rq_clock_pelt()) | |
7b595334 | 381 | * |
23127296 VG |
382 | * N.B., the above ratios (runnable% and running%) themselves are in the |
383 | * range of [0, 1]. To do fixed point arithmetics, we therefore scale them | |
384 | * to as large a range as necessary. This is for example reflected by | |
385 | * util_avg's SCHED_CAPACITY_SCALE. | |
7b595334 YD |
386 | * |
387 | * [Overflow issue] | |
388 | * | |
389 | * The 64-bit load_sum can have 4353082796 (=2^64/47742/88761) entities | |
390 | * with the highest load (=88761), always runnable on a single cfs_rq, | |
391 | * and should not overflow as the number already hits PID_MAX_LIMIT. | |
392 | * | |
393 | * For all other cases (including 32-bit kernels), struct load_weight's | |
394 | * weight will overflow first before we do, because: | |
395 | * | |
396 | * Max(load_avg) <= Max(load.weight) | |
397 | * | |
398 | * Then it is the load_weight's responsibility to consider overflow | |
399 | * issues. | |
9d89c257 | 400 | */ |
9d85f21c | 401 | struct sched_avg { |
5eca1c10 IM |
402 | u64 last_update_time; |
403 | u64 load_sum; | |
404 | u32 util_sum; | |
405 | u32 period_contrib; | |
406 | unsigned long load_avg; | |
407 | unsigned long util_avg; | |
7f65ea42 | 408 | struct util_est util_est; |
317d359d | 409 | } ____cacheline_aligned; |
9d85f21c | 410 | |
41acab88 | 411 | struct sched_statistics { |
7f5f8e8d | 412 | #ifdef CONFIG_SCHEDSTATS |
5eca1c10 IM |
413 | u64 wait_start; |
414 | u64 wait_max; | |
415 | u64 wait_count; | |
416 | u64 wait_sum; | |
417 | u64 iowait_count; | |
418 | u64 iowait_sum; | |
419 | ||
420 | u64 sleep_start; | |
421 | u64 sleep_max; | |
422 | s64 sum_sleep_runtime; | |
423 | ||
424 | u64 block_start; | |
425 | u64 block_max; | |
426 | u64 exec_max; | |
427 | u64 slice_max; | |
428 | ||
429 | u64 nr_migrations_cold; | |
430 | u64 nr_failed_migrations_affine; | |
431 | u64 nr_failed_migrations_running; | |
432 | u64 nr_failed_migrations_hot; | |
433 | u64 nr_forced_migrations; | |
434 | ||
435 | u64 nr_wakeups; | |
436 | u64 nr_wakeups_sync; | |
437 | u64 nr_wakeups_migrate; | |
438 | u64 nr_wakeups_local; | |
439 | u64 nr_wakeups_remote; | |
440 | u64 nr_wakeups_affine; | |
441 | u64 nr_wakeups_affine_attempts; | |
442 | u64 nr_wakeups_passive; | |
443 | u64 nr_wakeups_idle; | |
41acab88 | 444 | #endif |
7f5f8e8d | 445 | }; |
41acab88 LDM |
446 | |
447 | struct sched_entity { | |
5eca1c10 IM |
448 | /* For load-balancing: */ |
449 | struct load_weight load; | |
450 | struct rb_node run_node; | |
451 | struct list_head group_node; | |
452 | unsigned int on_rq; | |
41acab88 | 453 | |
5eca1c10 IM |
454 | u64 exec_start; |
455 | u64 sum_exec_runtime; | |
456 | u64 vruntime; | |
457 | u64 prev_sum_exec_runtime; | |
41acab88 | 458 | |
5eca1c10 | 459 | u64 nr_migrations; |
41acab88 | 460 | |
5eca1c10 | 461 | struct sched_statistics statistics; |
94c18227 | 462 | |
20b8a59f | 463 | #ifdef CONFIG_FAIR_GROUP_SCHED |
5eca1c10 IM |
464 | int depth; |
465 | struct sched_entity *parent; | |
20b8a59f | 466 | /* rq on which this entity is (to be) queued: */ |
5eca1c10 | 467 | struct cfs_rq *cfs_rq; |
20b8a59f | 468 | /* rq "owned" by this entity/group: */ |
5eca1c10 | 469 | struct cfs_rq *my_q; |
20b8a59f | 470 | #endif |
8bd75c77 | 471 | |
141965c7 | 472 | #ifdef CONFIG_SMP |
5a107804 JO |
473 | /* |
474 | * Per entity load average tracking. | |
475 | * | |
476 | * Put into separate cache line so it does not | |
477 | * collide with read-mostly values above. | |
478 | */ | |
317d359d | 479 | struct sched_avg avg; |
9d85f21c | 480 | #endif |
20b8a59f | 481 | }; |
70b97a7f | 482 | |
fa717060 | 483 | struct sched_rt_entity { |
5eca1c10 IM |
484 | struct list_head run_list; |
485 | unsigned long timeout; | |
486 | unsigned long watchdog_stamp; | |
487 | unsigned int time_slice; | |
488 | unsigned short on_rq; | |
489 | unsigned short on_list; | |
490 | ||
491 | struct sched_rt_entity *back; | |
052f1dc7 | 492 | #ifdef CONFIG_RT_GROUP_SCHED |
5eca1c10 | 493 | struct sched_rt_entity *parent; |
6f505b16 | 494 | /* rq on which this entity is (to be) queued: */ |
5eca1c10 | 495 | struct rt_rq *rt_rq; |
6f505b16 | 496 | /* rq "owned" by this entity/group: */ |
5eca1c10 | 497 | struct rt_rq *my_q; |
6f505b16 | 498 | #endif |
3859a271 | 499 | } __randomize_layout; |
fa717060 | 500 | |
aab03e05 | 501 | struct sched_dl_entity { |
5eca1c10 | 502 | struct rb_node rb_node; |
aab03e05 DF |
503 | |
504 | /* | |
505 | * Original scheduling parameters. Copied here from sched_attr | |
4027d080 | 506 | * during sched_setattr(), they will remain the same until |
507 | * the next sched_setattr(). | |
aab03e05 | 508 | */ |
5eca1c10 IM |
509 | u64 dl_runtime; /* Maximum runtime for each instance */ |
510 | u64 dl_deadline; /* Relative deadline of each instance */ | |
511 | u64 dl_period; /* Separation of two instances (period) */ | |
54d6d303 | 512 | u64 dl_bw; /* dl_runtime / dl_period */ |
3effcb42 | 513 | u64 dl_density; /* dl_runtime / dl_deadline */ |
aab03e05 DF |
514 | |
515 | /* | |
516 | * Actual scheduling parameters. Initialized with the values above, | |
dfcb245e | 517 | * they are continuously updated during task execution. Note that |
aab03e05 DF |
518 | * the remaining runtime could be < 0 in case we are in overrun. |
519 | */ | |
5eca1c10 IM |
520 | s64 runtime; /* Remaining runtime for this instance */ |
521 | u64 deadline; /* Absolute deadline for this instance */ | |
522 | unsigned int flags; /* Specifying the scheduler behaviour */ | |
aab03e05 DF |
523 | |
524 | /* | |
525 | * Some bool flags: | |
526 | * | |
527 | * @dl_throttled tells if we exhausted the runtime. If so, the | |
528 | * task has to wait for a replenishment to be performed at the | |
529 | * next firing of dl_timer. | |
530 | * | |
2d3d891d DF |
531 | * @dl_boosted tells if we are boosted due to DI. If so we are |
532 | * outside bandwidth enforcement mechanism (but only until we | |
5bfd126e JL |
533 | * exit the critical section); |
534 | * | |
5eca1c10 | 535 | * @dl_yielded tells if task gave up the CPU before consuming |
5bfd126e | 536 | * all its available runtime during the last job. |
209a0cbd LA |
537 | * |
538 | * @dl_non_contending tells if the task is inactive while still | |
539 | * contributing to the active utilization. In other words, it | |
540 | * indicates if the inactive timer has been armed and its handler | |
541 | * has not been executed yet. This flag is useful to avoid race | |
542 | * conditions between the inactive timer handler and the wakeup | |
543 | * code. | |
34be3930 JL |
544 | * |
545 | * @dl_overrun tells if the task asked to be informed about runtime | |
546 | * overruns. | |
aab03e05 | 547 | */ |
aa5222e9 DC |
548 | unsigned int dl_throttled : 1; |
549 | unsigned int dl_boosted : 1; | |
550 | unsigned int dl_yielded : 1; | |
551 | unsigned int dl_non_contending : 1; | |
34be3930 | 552 | unsigned int dl_overrun : 1; |
aab03e05 DF |
553 | |
554 | /* | |
555 | * Bandwidth enforcement timer. Each -deadline task has its | |
556 | * own bandwidth to be enforced, thus we need one timer per task. | |
557 | */ | |
5eca1c10 | 558 | struct hrtimer dl_timer; |
209a0cbd LA |
559 | |
560 | /* | |
561 | * Inactive timer, responsible for decreasing the active utilization | |
562 | * at the "0-lag time". When a -deadline task blocks, it contributes | |
563 | * to GRUB's active utilization until the "0-lag time", hence a | |
564 | * timer is needed to decrease the active utilization at the correct | |
565 | * time. | |
566 | */ | |
567 | struct hrtimer inactive_timer; | |
aab03e05 | 568 | }; |
8bd75c77 | 569 | |
69842cba PB |
570 | #ifdef CONFIG_UCLAMP_TASK |
571 | /* Number of utilization clamp buckets (shorter alias) */ | |
572 | #define UCLAMP_BUCKETS CONFIG_UCLAMP_BUCKETS_COUNT | |
573 | ||
574 | /* | |
575 | * Utilization clamp for a scheduling entity | |
576 | * @value: clamp value "assigned" to a se | |
577 | * @bucket_id: bucket index corresponding to the "assigned" value | |
e8f14172 | 578 | * @active: the se is currently refcounted in a rq's bucket |
a509a7cd | 579 | * @user_defined: the requested clamp value comes from user-space |
69842cba PB |
580 | * |
581 | * The bucket_id is the index of the clamp bucket matching the clamp value | |
582 | * which is pre-computed and stored to avoid expensive integer divisions from | |
583 | * the fast path. | |
e8f14172 PB |
584 | * |
585 | * The active bit is set whenever a task has got an "effective" value assigned, | |
586 | * which can be different from the clamp value "requested" from user-space. | |
587 | * This allows to know a task is refcounted in the rq's bucket corresponding | |
588 | * to the "effective" bucket_id. | |
a509a7cd PB |
589 | * |
590 | * The user_defined bit is set whenever a task has got a task-specific clamp | |
591 | * value requested from userspace, i.e. the system defaults apply to this task | |
592 | * just as a restriction. This allows to relax default clamps when a less | |
593 | * restrictive task-specific value has been requested, thus allowing to | |
594 | * implement a "nice" semantic. For example, a task running with a 20% | |
595 | * default boost can still drop its own boosting to 0%. | |
69842cba PB |
596 | */ |
597 | struct uclamp_se { | |
598 | unsigned int value : bits_per(SCHED_CAPACITY_SCALE); | |
599 | unsigned int bucket_id : bits_per(UCLAMP_BUCKETS); | |
e8f14172 | 600 | unsigned int active : 1; |
a509a7cd | 601 | unsigned int user_defined : 1; |
69842cba PB |
602 | }; |
603 | #endif /* CONFIG_UCLAMP_TASK */ | |
604 | ||
1d082fd0 PM |
605 | union rcu_special { |
606 | struct { | |
5eca1c10 IM |
607 | u8 blocked; |
608 | u8 need_qs; | |
05f41571 | 609 | u8 exp_hint; /* Hint for performance. */ |
23634ebc | 610 | u8 deferred_qs; |
8203d6d0 | 611 | } b; /* Bits. */ |
05f41571 | 612 | u32 s; /* Set of bits. */ |
1d082fd0 | 613 | }; |
86848966 | 614 | |
8dc85d54 PZ |
615 | enum perf_event_task_context { |
616 | perf_invalid_context = -1, | |
617 | perf_hw_context = 0, | |
89a1e187 | 618 | perf_sw_context, |
8dc85d54 PZ |
619 | perf_nr_task_contexts, |
620 | }; | |
621 | ||
eb61baf6 IM |
622 | struct wake_q_node { |
623 | struct wake_q_node *next; | |
624 | }; | |
625 | ||
1da177e4 | 626 | struct task_struct { |
c65eacbe AL |
627 | #ifdef CONFIG_THREAD_INFO_IN_TASK |
628 | /* | |
629 | * For reasons of header soup (see current_thread_info()), this | |
630 | * must be the first element of task_struct. | |
631 | */ | |
5eca1c10 | 632 | struct thread_info thread_info; |
c65eacbe | 633 | #endif |
5eca1c10 IM |
634 | /* -1 unrunnable, 0 runnable, >0 stopped: */ |
635 | volatile long state; | |
29e48ce8 KC |
636 | |
637 | /* | |
638 | * This begins the randomizable portion of task_struct. Only | |
639 | * scheduling-critical items should be added above here. | |
640 | */ | |
641 | randomized_struct_fields_start | |
642 | ||
5eca1c10 | 643 | void *stack; |
ec1d2819 | 644 | refcount_t usage; |
5eca1c10 IM |
645 | /* Per task flags (PF_*), defined further below: */ |
646 | unsigned int flags; | |
647 | unsigned int ptrace; | |
1da177e4 | 648 | |
2dd73a4f | 649 | #ifdef CONFIG_SMP |
5eca1c10 IM |
650 | struct llist_node wake_entry; |
651 | int on_cpu; | |
c65eacbe | 652 | #ifdef CONFIG_THREAD_INFO_IN_TASK |
5eca1c10 IM |
653 | /* Current CPU: */ |
654 | unsigned int cpu; | |
c65eacbe | 655 | #endif |
5eca1c10 IM |
656 | unsigned int wakee_flips; |
657 | unsigned long wakee_flip_decay_ts; | |
658 | struct task_struct *last_wakee; | |
ac66f547 | 659 | |
32e839dd MG |
660 | /* |
661 | * recent_used_cpu is initially set as the last CPU used by a task | |
662 | * that wakes affine another task. Waker/wakee relationships can | |
663 | * push tasks around a CPU where each wakeup moves to the next one. | |
664 | * Tracking a recently used CPU allows a quick search for a recently | |
665 | * used CPU that may be idle. | |
666 | */ | |
667 | int recent_used_cpu; | |
5eca1c10 | 668 | int wake_cpu; |
2dd73a4f | 669 | #endif |
5eca1c10 IM |
670 | int on_rq; |
671 | ||
672 | int prio; | |
673 | int static_prio; | |
674 | int normal_prio; | |
675 | unsigned int rt_priority; | |
50e645a8 | 676 | |
5eca1c10 IM |
677 | const struct sched_class *sched_class; |
678 | struct sched_entity se; | |
679 | struct sched_rt_entity rt; | |
8323f26c | 680 | #ifdef CONFIG_CGROUP_SCHED |
5eca1c10 | 681 | struct task_group *sched_task_group; |
8323f26c | 682 | #endif |
5eca1c10 | 683 | struct sched_dl_entity dl; |
1da177e4 | 684 | |
69842cba | 685 | #ifdef CONFIG_UCLAMP_TASK |
e8f14172 PB |
686 | /* Clamp values requested for a scheduling entity */ |
687 | struct uclamp_se uclamp_req[UCLAMP_CNT]; | |
688 | /* Effective clamp values used for a scheduling entity */ | |
69842cba PB |
689 | struct uclamp_se uclamp[UCLAMP_CNT]; |
690 | #endif | |
691 | ||
e107be36 | 692 | #ifdef CONFIG_PREEMPT_NOTIFIERS |
5eca1c10 IM |
693 | /* List of struct preempt_notifier: */ |
694 | struct hlist_head preempt_notifiers; | |
e107be36 AK |
695 | #endif |
696 | ||
6c5c9341 | 697 | #ifdef CONFIG_BLK_DEV_IO_TRACE |
5eca1c10 | 698 | unsigned int btrace_seq; |
6c5c9341 | 699 | #endif |
1da177e4 | 700 | |
5eca1c10 IM |
701 | unsigned int policy; |
702 | int nr_cpus_allowed; | |
3bd37062 SAS |
703 | const cpumask_t *cpus_ptr; |
704 | cpumask_t cpus_mask; | |
1da177e4 | 705 | |
a57eb940 | 706 | #ifdef CONFIG_PREEMPT_RCU |
5eca1c10 IM |
707 | int rcu_read_lock_nesting; |
708 | union rcu_special rcu_read_unlock_special; | |
709 | struct list_head rcu_node_entry; | |
710 | struct rcu_node *rcu_blocked_node; | |
28f6569a | 711 | #endif /* #ifdef CONFIG_PREEMPT_RCU */ |
5eca1c10 | 712 | |
8315f422 | 713 | #ifdef CONFIG_TASKS_RCU |
5eca1c10 | 714 | unsigned long rcu_tasks_nvcsw; |
ccdd29ff PM |
715 | u8 rcu_tasks_holdout; |
716 | u8 rcu_tasks_idx; | |
5eca1c10 | 717 | int rcu_tasks_idle_cpu; |
ccdd29ff | 718 | struct list_head rcu_tasks_holdout_list; |
8315f422 | 719 | #endif /* #ifdef CONFIG_TASKS_RCU */ |
e260be67 | 720 | |
5eca1c10 | 721 | struct sched_info sched_info; |
1da177e4 | 722 | |
5eca1c10 | 723 | struct list_head tasks; |
806c09a7 | 724 | #ifdef CONFIG_SMP |
5eca1c10 IM |
725 | struct plist_node pushable_tasks; |
726 | struct rb_node pushable_dl_tasks; | |
806c09a7 | 727 | #endif |
1da177e4 | 728 | |
5eca1c10 IM |
729 | struct mm_struct *mm; |
730 | struct mm_struct *active_mm; | |
314ff785 IM |
731 | |
732 | /* Per-thread vma caching: */ | |
5eca1c10 | 733 | struct vmacache vmacache; |
314ff785 | 734 | |
5eca1c10 IM |
735 | #ifdef SPLIT_RSS_COUNTING |
736 | struct task_rss_stat rss_stat; | |
34e55232 | 737 | #endif |
5eca1c10 IM |
738 | int exit_state; |
739 | int exit_code; | |
740 | int exit_signal; | |
741 | /* The signal sent when the parent dies: */ | |
742 | int pdeath_signal; | |
743 | /* JOBCTL_*, siglock protected: */ | |
744 | unsigned long jobctl; | |
745 | ||
746 | /* Used for emulating ABI behavior of previous Linux versions: */ | |
747 | unsigned int personality; | |
748 | ||
749 | /* Scheduler bits, serialized by scheduler locks: */ | |
750 | unsigned sched_reset_on_fork:1; | |
751 | unsigned sched_contributes_to_load:1; | |
752 | unsigned sched_migrated:1; | |
753 | unsigned sched_remote_wakeup:1; | |
eb414681 JW |
754 | #ifdef CONFIG_PSI |
755 | unsigned sched_psi_wake_requeue:1; | |
756 | #endif | |
757 | ||
5eca1c10 IM |
758 | /* Force alignment to the next boundary: */ |
759 | unsigned :0; | |
760 | ||
761 | /* Unserialized, strictly 'current' */ | |
762 | ||
763 | /* Bit to tell LSMs we're in execve(): */ | |
764 | unsigned in_execve:1; | |
765 | unsigned in_iowait:1; | |
766 | #ifndef TIF_RESTORE_SIGMASK | |
767 | unsigned restore_sigmask:1; | |
7e781418 | 768 | #endif |
626ebc41 | 769 | #ifdef CONFIG_MEMCG |
29ef680a | 770 | unsigned in_user_fault:1; |
127424c8 | 771 | #endif |
ff303e66 | 772 | #ifdef CONFIG_COMPAT_BRK |
5eca1c10 | 773 | unsigned brk_randomized:1; |
ff303e66 | 774 | #endif |
77f88796 TH |
775 | #ifdef CONFIG_CGROUPS |
776 | /* disallow userland-initiated cgroup migration */ | |
777 | unsigned no_cgroup_migration:1; | |
76f969e8 RG |
778 | /* task is frozen/stopped (used by the cgroup freezer) */ |
779 | unsigned frozen:1; | |
77f88796 | 780 | #endif |
d09d8df3 JB |
781 | #ifdef CONFIG_BLK_CGROUP |
782 | /* to be used once the psi infrastructure lands upstream. */ | |
783 | unsigned use_memdelay:1; | |
784 | #endif | |
6f185c29 | 785 | |
5eca1c10 | 786 | unsigned long atomic_flags; /* Flags requiring atomic access. */ |
1d4457f9 | 787 | |
5eca1c10 | 788 | struct restart_block restart_block; |
f56141e3 | 789 | |
5eca1c10 IM |
790 | pid_t pid; |
791 | pid_t tgid; | |
0a425405 | 792 | |
050e9baa | 793 | #ifdef CONFIG_STACKPROTECTOR |
5eca1c10 IM |
794 | /* Canary value for the -fstack-protector GCC feature: */ |
795 | unsigned long stack_canary; | |
1314562a | 796 | #endif |
4d1d61a6 | 797 | /* |
5eca1c10 | 798 | * Pointers to the (original) parent process, youngest child, younger sibling, |
4d1d61a6 | 799 | * older sibling, respectively. (p->father can be replaced with |
f470021a | 800 | * p->real_parent->pid) |
1da177e4 | 801 | */ |
5eca1c10 IM |
802 | |
803 | /* Real parent process: */ | |
804 | struct task_struct __rcu *real_parent; | |
805 | ||
806 | /* Recipient of SIGCHLD, wait4() reports: */ | |
807 | struct task_struct __rcu *parent; | |
808 | ||
1da177e4 | 809 | /* |
5eca1c10 | 810 | * Children/sibling form the list of natural children: |
1da177e4 | 811 | */ |
5eca1c10 IM |
812 | struct list_head children; |
813 | struct list_head sibling; | |
814 | struct task_struct *group_leader; | |
1da177e4 | 815 | |
f470021a | 816 | /* |
5eca1c10 IM |
817 | * 'ptraced' is the list of tasks this task is using ptrace() on. |
818 | * | |
f470021a | 819 | * This includes both natural children and PTRACE_ATTACH targets. |
5eca1c10 | 820 | * 'ptrace_entry' is this task's link on the p->parent->ptraced list. |
f470021a | 821 | */ |
5eca1c10 IM |
822 | struct list_head ptraced; |
823 | struct list_head ptrace_entry; | |
f470021a | 824 | |
1da177e4 | 825 | /* PID/PID hash table linkage. */ |
2c470475 EB |
826 | struct pid *thread_pid; |
827 | struct hlist_node pid_links[PIDTYPE_MAX]; | |
5eca1c10 IM |
828 | struct list_head thread_group; |
829 | struct list_head thread_node; | |
830 | ||
831 | struct completion *vfork_done; | |
1da177e4 | 832 | |
5eca1c10 IM |
833 | /* CLONE_CHILD_SETTID: */ |
834 | int __user *set_child_tid; | |
1da177e4 | 835 | |
5eca1c10 IM |
836 | /* CLONE_CHILD_CLEARTID: */ |
837 | int __user *clear_child_tid; | |
838 | ||
839 | u64 utime; | |
840 | u64 stime; | |
40565b5a | 841 | #ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME |
5eca1c10 IM |
842 | u64 utimescaled; |
843 | u64 stimescaled; | |
40565b5a | 844 | #endif |
5eca1c10 IM |
845 | u64 gtime; |
846 | struct prev_cputime prev_cputime; | |
6a61671b | 847 | #ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN |
bac5b6b6 | 848 | struct vtime vtime; |
d99ca3b9 | 849 | #endif |
d027d45d FW |
850 | |
851 | #ifdef CONFIG_NO_HZ_FULL | |
5eca1c10 | 852 | atomic_t tick_dep_mask; |
d027d45d | 853 | #endif |
5eca1c10 IM |
854 | /* Context switch counts: */ |
855 | unsigned long nvcsw; | |
856 | unsigned long nivcsw; | |
857 | ||
858 | /* Monotonic time in nsecs: */ | |
859 | u64 start_time; | |
860 | ||
861 | /* Boot based time in nsecs: */ | |
cf25e24d | 862 | u64 start_boottime; |
5eca1c10 IM |
863 | |
864 | /* MM fault and swap info: this can arguably be seen as either mm-specific or thread-specific: */ | |
865 | unsigned long min_flt; | |
866 | unsigned long maj_flt; | |
1da177e4 | 867 | |
2b69942f TG |
868 | /* Empty if CONFIG_POSIX_CPUTIMERS=n */ |
869 | struct posix_cputimers posix_cputimers; | |
1da177e4 | 870 | |
5eca1c10 IM |
871 | /* Process credentials: */ |
872 | ||
873 | /* Tracer's credentials at attach: */ | |
874 | const struct cred __rcu *ptracer_cred; | |
875 | ||
876 | /* Objective and real subjective task credentials (COW): */ | |
877 | const struct cred __rcu *real_cred; | |
878 | ||
879 | /* Effective (overridable) subjective task credentials (COW): */ | |
880 | const struct cred __rcu *cred; | |
881 | ||
7743c48e DH |
882 | #ifdef CONFIG_KEYS |
883 | /* Cached requested key. */ | |
884 | struct key *cached_requested_key; | |
885 | #endif | |
886 | ||
5eca1c10 IM |
887 | /* |
888 | * executable name, excluding path. | |
889 | * | |
890 | * - normally initialized setup_new_exec() | |
891 | * - access it with [gs]et_task_comm() | |
892 | * - lock it with task_lock() | |
893 | */ | |
894 | char comm[TASK_COMM_LEN]; | |
895 | ||
896 | struct nameidata *nameidata; | |
897 | ||
3d5b6fcc | 898 | #ifdef CONFIG_SYSVIPC |
5eca1c10 IM |
899 | struct sysv_sem sysvsem; |
900 | struct sysv_shm sysvshm; | |
3d5b6fcc | 901 | #endif |
e162b39a | 902 | #ifdef CONFIG_DETECT_HUNG_TASK |
5eca1c10 | 903 | unsigned long last_switch_count; |
a2e51445 | 904 | unsigned long last_switch_time; |
82a1fcb9 | 905 | #endif |
5eca1c10 IM |
906 | /* Filesystem information: */ |
907 | struct fs_struct *fs; | |
908 | ||
909 | /* Open file information: */ | |
910 | struct files_struct *files; | |
911 | ||
912 | /* Namespaces: */ | |
913 | struct nsproxy *nsproxy; | |
914 | ||
915 | /* Signal handlers: */ | |
916 | struct signal_struct *signal; | |
913292c9 | 917 | struct sighand_struct __rcu *sighand; |
5eca1c10 IM |
918 | sigset_t blocked; |
919 | sigset_t real_blocked; | |
920 | /* Restored if set_restore_sigmask() was used: */ | |
921 | sigset_t saved_sigmask; | |
922 | struct sigpending pending; | |
923 | unsigned long sas_ss_sp; | |
924 | size_t sas_ss_size; | |
925 | unsigned int sas_ss_flags; | |
926 | ||
927 | struct callback_head *task_works; | |
928 | ||
4b7d248b | 929 | #ifdef CONFIG_AUDIT |
bfef93a5 | 930 | #ifdef CONFIG_AUDITSYSCALL |
5f3d544f RGB |
931 | struct audit_context *audit_context; |
932 | #endif | |
5eca1c10 IM |
933 | kuid_t loginuid; |
934 | unsigned int sessionid; | |
bfef93a5 | 935 | #endif |
5eca1c10 IM |
936 | struct seccomp seccomp; |
937 | ||
938 | /* Thread group tracking: */ | |
939 | u32 parent_exec_id; | |
940 | u32 self_exec_id; | |
1da177e4 | 941 | |
5eca1c10 IM |
942 | /* Protection against (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed, mempolicy: */ |
943 | spinlock_t alloc_lock; | |
1da177e4 | 944 | |
b29739f9 | 945 | /* Protection of the PI data structures: */ |
5eca1c10 | 946 | raw_spinlock_t pi_lock; |
b29739f9 | 947 | |
5eca1c10 | 948 | struct wake_q_node wake_q; |
76751049 | 949 | |
23f78d4a | 950 | #ifdef CONFIG_RT_MUTEXES |
5eca1c10 | 951 | /* PI waiters blocked on a rt_mutex held by this task: */ |
a23ba907 | 952 | struct rb_root_cached pi_waiters; |
e96a7705 XP |
953 | /* Updated under owner's pi_lock and rq lock */ |
954 | struct task_struct *pi_top_task; | |
5eca1c10 IM |
955 | /* Deadlock detection and priority inheritance handling: */ |
956 | struct rt_mutex_waiter *pi_blocked_on; | |
23f78d4a IM |
957 | #endif |
958 | ||
408894ee | 959 | #ifdef CONFIG_DEBUG_MUTEXES |
5eca1c10 IM |
960 | /* Mutex deadlock detection: */ |
961 | struct mutex_waiter *blocked_on; | |
408894ee | 962 | #endif |
5eca1c10 | 963 | |
312364f3 DV |
964 | #ifdef CONFIG_DEBUG_ATOMIC_SLEEP |
965 | int non_block_count; | |
966 | #endif | |
967 | ||
de30a2b3 | 968 | #ifdef CONFIG_TRACE_IRQFLAGS |
5eca1c10 IM |
969 | unsigned int irq_events; |
970 | unsigned long hardirq_enable_ip; | |
971 | unsigned long hardirq_disable_ip; | |
972 | unsigned int hardirq_enable_event; | |
973 | unsigned int hardirq_disable_event; | |
974 | int hardirqs_enabled; | |
975 | int hardirq_context; | |
976 | unsigned long softirq_disable_ip; | |
977 | unsigned long softirq_enable_ip; | |
978 | unsigned int softirq_disable_event; | |
979 | unsigned int softirq_enable_event; | |
980 | int softirqs_enabled; | |
981 | int softirq_context; | |
de30a2b3 | 982 | #endif |
5eca1c10 | 983 | |
fbb9ce95 | 984 | #ifdef CONFIG_LOCKDEP |
5eca1c10 IM |
985 | # define MAX_LOCK_DEPTH 48UL |
986 | u64 curr_chain_key; | |
987 | int lockdep_depth; | |
988 | unsigned int lockdep_recursion; | |
989 | struct held_lock held_locks[MAX_LOCK_DEPTH]; | |
fbb9ce95 | 990 | #endif |
5eca1c10 | 991 | |
c6d30853 | 992 | #ifdef CONFIG_UBSAN |
5eca1c10 | 993 | unsigned int in_ubsan; |
c6d30853 | 994 | #endif |
408894ee | 995 | |
5eca1c10 IM |
996 | /* Journalling filesystem info: */ |
997 | void *journal_info; | |
1da177e4 | 998 | |
5eca1c10 IM |
999 | /* Stacked block device info: */ |
1000 | struct bio_list *bio_list; | |
d89d8796 | 1001 | |
73c10101 | 1002 | #ifdef CONFIG_BLOCK |
5eca1c10 IM |
1003 | /* Stack plugging: */ |
1004 | struct blk_plug *plug; | |
73c10101 JA |
1005 | #endif |
1006 | ||
5eca1c10 IM |
1007 | /* VM state: */ |
1008 | struct reclaim_state *reclaim_state; | |
1009 | ||
1010 | struct backing_dev_info *backing_dev_info; | |
1da177e4 | 1011 | |
5eca1c10 | 1012 | struct io_context *io_context; |
1da177e4 | 1013 | |
5e1f0f09 MG |
1014 | #ifdef CONFIG_COMPACTION |
1015 | struct capture_control *capture_control; | |
1016 | #endif | |
5eca1c10 IM |
1017 | /* Ptrace state: */ |
1018 | unsigned long ptrace_message; | |
ae7795bc | 1019 | kernel_siginfo_t *last_siginfo; |
1da177e4 | 1020 | |
5eca1c10 | 1021 | struct task_io_accounting ioac; |
eb414681 JW |
1022 | #ifdef CONFIG_PSI |
1023 | /* Pressure stall state */ | |
1024 | unsigned int psi_flags; | |
1025 | #endif | |
5eca1c10 IM |
1026 | #ifdef CONFIG_TASK_XACCT |
1027 | /* Accumulated RSS usage: */ | |
1028 | u64 acct_rss_mem1; | |
1029 | /* Accumulated virtual memory usage: */ | |
1030 | u64 acct_vm_mem1; | |
1031 | /* stime + utime since last update: */ | |
1032 | u64 acct_timexpd; | |
1da177e4 LT |
1033 | #endif |
1034 | #ifdef CONFIG_CPUSETS | |
5eca1c10 IM |
1035 | /* Protected by ->alloc_lock: */ |
1036 | nodemask_t mems_allowed; | |
1037 | /* Seqence number to catch updates: */ | |
1038 | seqcount_t mems_allowed_seq; | |
1039 | int cpuset_mem_spread_rotor; | |
1040 | int cpuset_slab_spread_rotor; | |
1da177e4 | 1041 | #endif |
ddbcc7e8 | 1042 | #ifdef CONFIG_CGROUPS |
5eca1c10 IM |
1043 | /* Control Group info protected by css_set_lock: */ |
1044 | struct css_set __rcu *cgroups; | |
1045 | /* cg_list protected by css_set_lock and tsk->alloc_lock: */ | |
1046 | struct list_head cg_list; | |
ddbcc7e8 | 1047 | #endif |
e6d42931 | 1048 | #ifdef CONFIG_X86_CPU_RESCTRL |
0734ded1 | 1049 | u32 closid; |
d6aaba61 | 1050 | u32 rmid; |
e02737d5 | 1051 | #endif |
42b2dd0a | 1052 | #ifdef CONFIG_FUTEX |
5eca1c10 | 1053 | struct robust_list_head __user *robust_list; |
34f192c6 IM |
1054 | #ifdef CONFIG_COMPAT |
1055 | struct compat_robust_list_head __user *compat_robust_list; | |
1056 | #endif | |
5eca1c10 IM |
1057 | struct list_head pi_state_list; |
1058 | struct futex_pi_state *pi_state_cache; | |
3f186d97 | 1059 | struct mutex futex_exit_mutex; |
3d4775df | 1060 | unsigned int futex_state; |
c7aceaba | 1061 | #endif |
cdd6c482 | 1062 | #ifdef CONFIG_PERF_EVENTS |
5eca1c10 IM |
1063 | struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts]; |
1064 | struct mutex perf_event_mutex; | |
1065 | struct list_head perf_event_list; | |
a63eaf34 | 1066 | #endif |
8f47b187 | 1067 | #ifdef CONFIG_DEBUG_PREEMPT |
5eca1c10 | 1068 | unsigned long preempt_disable_ip; |
8f47b187 | 1069 | #endif |
c7aceaba | 1070 | #ifdef CONFIG_NUMA |
5eca1c10 IM |
1071 | /* Protected by alloc_lock: */ |
1072 | struct mempolicy *mempolicy; | |
45816682 | 1073 | short il_prev; |
5eca1c10 | 1074 | short pref_node_fork; |
42b2dd0a | 1075 | #endif |
cbee9f88 | 1076 | #ifdef CONFIG_NUMA_BALANCING |
5eca1c10 IM |
1077 | int numa_scan_seq; |
1078 | unsigned int numa_scan_period; | |
1079 | unsigned int numa_scan_period_max; | |
1080 | int numa_preferred_nid; | |
1081 | unsigned long numa_migrate_retry; | |
1082 | /* Migration stamp: */ | |
1083 | u64 node_stamp; | |
1084 | u64 last_task_numa_placement; | |
1085 | u64 last_sum_exec_runtime; | |
1086 | struct callback_head numa_work; | |
1087 | ||
cb361d8c JH |
1088 | /* |
1089 | * This pointer is only modified for current in syscall and | |
1090 | * pagefault context (and for tasks being destroyed), so it can be read | |
1091 | * from any of the following contexts: | |
1092 | * - RCU read-side critical section | |
1093 | * - current->numa_group from everywhere | |
1094 | * - task's runqueue locked, task not running | |
1095 | */ | |
1096 | struct numa_group __rcu *numa_group; | |
8c8a743c | 1097 | |
745d6147 | 1098 | /* |
44dba3d5 IM |
1099 | * numa_faults is an array split into four regions: |
1100 | * faults_memory, faults_cpu, faults_memory_buffer, faults_cpu_buffer | |
1101 | * in this precise order. | |
1102 | * | |
1103 | * faults_memory: Exponential decaying average of faults on a per-node | |
1104 | * basis. Scheduling placement decisions are made based on these | |
1105 | * counts. The values remain static for the duration of a PTE scan. | |
1106 | * faults_cpu: Track the nodes the process was running on when a NUMA | |
1107 | * hinting fault was incurred. | |
1108 | * faults_memory_buffer and faults_cpu_buffer: Record faults per node | |
1109 | * during the current scan window. When the scan completes, the counts | |
1110 | * in faults_memory and faults_cpu decay and these values are copied. | |
745d6147 | 1111 | */ |
5eca1c10 IM |
1112 | unsigned long *numa_faults; |
1113 | unsigned long total_numa_faults; | |
745d6147 | 1114 | |
04bb2f94 RR |
1115 | /* |
1116 | * numa_faults_locality tracks if faults recorded during the last | |
074c2381 MG |
1117 | * scan window were remote/local or failed to migrate. The task scan |
1118 | * period is adapted based on the locality of the faults with different | |
1119 | * weights depending on whether they were shared or private faults | |
04bb2f94 | 1120 | */ |
5eca1c10 | 1121 | unsigned long numa_faults_locality[3]; |
04bb2f94 | 1122 | |
5eca1c10 | 1123 | unsigned long numa_pages_migrated; |
cbee9f88 PZ |
1124 | #endif /* CONFIG_NUMA_BALANCING */ |
1125 | ||
d7822b1e MD |
1126 | #ifdef CONFIG_RSEQ |
1127 | struct rseq __user *rseq; | |
d7822b1e MD |
1128 | u32 rseq_sig; |
1129 | /* | |
1130 | * RmW on rseq_event_mask must be performed atomically | |
1131 | * with respect to preemption. | |
1132 | */ | |
1133 | unsigned long rseq_event_mask; | |
1134 | #endif | |
1135 | ||
5eca1c10 | 1136 | struct tlbflush_unmap_batch tlb_ubc; |
72b252ae | 1137 | |
3fbd7ee2 EB |
1138 | union { |
1139 | refcount_t rcu_users; | |
1140 | struct rcu_head rcu; | |
1141 | }; | |
b92ce558 | 1142 | |
5eca1c10 IM |
1143 | /* Cache last used pipe for splice(): */ |
1144 | struct pipe_inode_info *splice_pipe; | |
5640f768 | 1145 | |
5eca1c10 | 1146 | struct page_frag task_frag; |
5640f768 | 1147 | |
47913d4e IM |
1148 | #ifdef CONFIG_TASK_DELAY_ACCT |
1149 | struct task_delay_info *delays; | |
f4f154fd | 1150 | #endif |
47913d4e | 1151 | |
f4f154fd | 1152 | #ifdef CONFIG_FAULT_INJECTION |
5eca1c10 | 1153 | int make_it_fail; |
9049f2f6 | 1154 | unsigned int fail_nth; |
ca74e92b | 1155 | #endif |
9d823e8f | 1156 | /* |
5eca1c10 IM |
1157 | * When (nr_dirtied >= nr_dirtied_pause), it's time to call |
1158 | * balance_dirty_pages() for a dirty throttling pause: | |
9d823e8f | 1159 | */ |
5eca1c10 IM |
1160 | int nr_dirtied; |
1161 | int nr_dirtied_pause; | |
1162 | /* Start of a write-and-pause period: */ | |
1163 | unsigned long dirty_paused_when; | |
9d823e8f | 1164 | |
9745512c | 1165 | #ifdef CONFIG_LATENCYTOP |
5eca1c10 IM |
1166 | int latency_record_count; |
1167 | struct latency_record latency_record[LT_SAVECOUNT]; | |
9745512c | 1168 | #endif |
6976675d | 1169 | /* |
5eca1c10 | 1170 | * Time slack values; these are used to round up poll() and |
6976675d AV |
1171 | * select() etc timeout values. These are in nanoseconds. |
1172 | */ | |
5eca1c10 IM |
1173 | u64 timer_slack_ns; |
1174 | u64 default_timer_slack_ns; | |
f8d570a4 | 1175 | |
0b24becc | 1176 | #ifdef CONFIG_KASAN |
5eca1c10 | 1177 | unsigned int kasan_depth; |
0b24becc | 1178 | #endif |
5eca1c10 | 1179 | |
fb52607a | 1180 | #ifdef CONFIG_FUNCTION_GRAPH_TRACER |
5eca1c10 IM |
1181 | /* Index of current stored address in ret_stack: */ |
1182 | int curr_ret_stack; | |
39eb456d | 1183 | int curr_ret_depth; |
5eca1c10 IM |
1184 | |
1185 | /* Stack of return addresses for return function tracing: */ | |
1186 | struct ftrace_ret_stack *ret_stack; | |
1187 | ||
1188 | /* Timestamp for last schedule: */ | |
1189 | unsigned long long ftrace_timestamp; | |
1190 | ||
f201ae23 FW |
1191 | /* |
1192 | * Number of functions that haven't been traced | |
5eca1c10 | 1193 | * because of depth overrun: |
f201ae23 | 1194 | */ |
5eca1c10 IM |
1195 | atomic_t trace_overrun; |
1196 | ||
1197 | /* Pause tracing: */ | |
1198 | atomic_t tracing_graph_pause; | |
f201ae23 | 1199 | #endif |
5eca1c10 | 1200 | |
ea4e2bc4 | 1201 | #ifdef CONFIG_TRACING |
5eca1c10 IM |
1202 | /* State flags for use by tracers: */ |
1203 | unsigned long trace; | |
1204 | ||
1205 | /* Bitmask and counter of trace recursion: */ | |
1206 | unsigned long trace_recursion; | |
261842b7 | 1207 | #endif /* CONFIG_TRACING */ |
5eca1c10 | 1208 | |
5c9a8750 | 1209 | #ifdef CONFIG_KCOV |
eec028c9 AK |
1210 | /* See kernel/kcov.c for more details. */ |
1211 | ||
5eca1c10 | 1212 | /* Coverage collection mode enabled for this task (0 if disabled): */ |
0ed557aa | 1213 | unsigned int kcov_mode; |
5eca1c10 IM |
1214 | |
1215 | /* Size of the kcov_area: */ | |
1216 | unsigned int kcov_size; | |
1217 | ||
1218 | /* Buffer for coverage collection: */ | |
1219 | void *kcov_area; | |
1220 | ||
1221 | /* KCOV descriptor wired with this task or NULL: */ | |
1222 | struct kcov *kcov; | |
eec028c9 AK |
1223 | |
1224 | /* KCOV common handle for remote coverage collection: */ | |
1225 | u64 kcov_handle; | |
1226 | ||
1227 | /* KCOV sequence number: */ | |
1228 | int kcov_sequence; | |
5c9a8750 | 1229 | #endif |
5eca1c10 | 1230 | |
6f185c29 | 1231 | #ifdef CONFIG_MEMCG |
5eca1c10 IM |
1232 | struct mem_cgroup *memcg_in_oom; |
1233 | gfp_t memcg_oom_gfp_mask; | |
1234 | int memcg_oom_order; | |
b23afb93 | 1235 | |
5eca1c10 IM |
1236 | /* Number of pages to reclaim on returning to userland: */ |
1237 | unsigned int memcg_nr_pages_over_high; | |
d46eb14b SB |
1238 | |
1239 | /* Used by memcontrol for targeted memcg charge: */ | |
1240 | struct mem_cgroup *active_memcg; | |
569b846d | 1241 | #endif |
5eca1c10 | 1242 | |
d09d8df3 JB |
1243 | #ifdef CONFIG_BLK_CGROUP |
1244 | struct request_queue *throttle_queue; | |
1245 | #endif | |
1246 | ||
0326f5a9 | 1247 | #ifdef CONFIG_UPROBES |
5eca1c10 | 1248 | struct uprobe_task *utask; |
0326f5a9 | 1249 | #endif |
cafe5635 | 1250 | #if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE) |
5eca1c10 IM |
1251 | unsigned int sequential_io; |
1252 | unsigned int sequential_io_avg; | |
cafe5635 | 1253 | #endif |
8eb23b9f | 1254 | #ifdef CONFIG_DEBUG_ATOMIC_SLEEP |
5eca1c10 | 1255 | unsigned long task_state_change; |
8eb23b9f | 1256 | #endif |
5eca1c10 | 1257 | int pagefault_disabled; |
03049269 | 1258 | #ifdef CONFIG_MMU |
5eca1c10 | 1259 | struct task_struct *oom_reaper_list; |
03049269 | 1260 | #endif |
ba14a194 | 1261 | #ifdef CONFIG_VMAP_STACK |
5eca1c10 | 1262 | struct vm_struct *stack_vm_area; |
ba14a194 | 1263 | #endif |
68f24b08 | 1264 | #ifdef CONFIG_THREAD_INFO_IN_TASK |
5eca1c10 | 1265 | /* A live task holds one reference: */ |
f0b89d39 | 1266 | refcount_t stack_refcount; |
d83a7cb3 JP |
1267 | #endif |
1268 | #ifdef CONFIG_LIVEPATCH | |
1269 | int patch_state; | |
0302e28d | 1270 | #endif |
e4e55b47 TH |
1271 | #ifdef CONFIG_SECURITY |
1272 | /* Used by LSM modules for access restriction: */ | |
1273 | void *security; | |
68f24b08 | 1274 | #endif |
29e48ce8 | 1275 | |
afaef01c AP |
1276 | #ifdef CONFIG_GCC_PLUGIN_STACKLEAK |
1277 | unsigned long lowest_stack; | |
c8d12627 | 1278 | unsigned long prev_lowest_stack; |
afaef01c AP |
1279 | #endif |
1280 | ||
29e48ce8 KC |
1281 | /* |
1282 | * New fields for task_struct should be added above here, so that | |
1283 | * they are included in the randomized portion of task_struct. | |
1284 | */ | |
1285 | randomized_struct_fields_end | |
1286 | ||
5eca1c10 IM |
1287 | /* CPU-specific state of this task: */ |
1288 | struct thread_struct thread; | |
1289 | ||
1290 | /* | |
1291 | * WARNING: on x86, 'thread_struct' contains a variable-sized | |
1292 | * structure. It *MUST* be at the end of 'task_struct'. | |
1293 | * | |
1294 | * Do not put anything below here! | |
1295 | */ | |
1da177e4 LT |
1296 | }; |
1297 | ||
e868171a | 1298 | static inline struct pid *task_pid(struct task_struct *task) |
22c935f4 | 1299 | { |
2c470475 | 1300 | return task->thread_pid; |
22c935f4 EB |
1301 | } |
1302 | ||
7af57294 PE |
1303 | /* |
1304 | * the helpers to get the task's different pids as they are seen | |
1305 | * from various namespaces | |
1306 | * | |
1307 | * task_xid_nr() : global id, i.e. the id seen from the init namespace; | |
44c4e1b2 EB |
1308 | * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of |
1309 | * current. | |
7af57294 PE |
1310 | * task_xid_nr_ns() : id seen from the ns specified; |
1311 | * | |
7af57294 PE |
1312 | * see also pid_nr() etc in include/linux/pid.h |
1313 | */ | |
5eca1c10 | 1314 | pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type, struct pid_namespace *ns); |
7af57294 | 1315 | |
e868171a | 1316 | static inline pid_t task_pid_nr(struct task_struct *tsk) |
7af57294 PE |
1317 | { |
1318 | return tsk->pid; | |
1319 | } | |
1320 | ||
5eca1c10 | 1321 | static inline pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) |
52ee2dfd ON |
1322 | { |
1323 | return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns); | |
1324 | } | |
7af57294 PE |
1325 | |
1326 | static inline pid_t task_pid_vnr(struct task_struct *tsk) | |
1327 | { | |
52ee2dfd | 1328 | return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL); |
7af57294 PE |
1329 | } |
1330 | ||
1331 | ||
e868171a | 1332 | static inline pid_t task_tgid_nr(struct task_struct *tsk) |
7af57294 PE |
1333 | { |
1334 | return tsk->tgid; | |
1335 | } | |
1336 | ||
5eca1c10 IM |
1337 | /** |
1338 | * pid_alive - check that a task structure is not stale | |
1339 | * @p: Task structure to be checked. | |
1340 | * | |
1341 | * Test if a process is not yet dead (at most zombie state) | |
1342 | * If pid_alive fails, then pointers within the task structure | |
1343 | * can be stale and must not be dereferenced. | |
1344 | * | |
1345 | * Return: 1 if the process is alive. 0 otherwise. | |
1346 | */ | |
1347 | static inline int pid_alive(const struct task_struct *p) | |
1348 | { | |
2c470475 | 1349 | return p->thread_pid != NULL; |
5eca1c10 | 1350 | } |
7af57294 | 1351 | |
5eca1c10 | 1352 | static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) |
7af57294 | 1353 | { |
52ee2dfd | 1354 | return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns); |
7af57294 PE |
1355 | } |
1356 | ||
7af57294 PE |
1357 | static inline pid_t task_pgrp_vnr(struct task_struct *tsk) |
1358 | { | |
52ee2dfd | 1359 | return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL); |
7af57294 PE |
1360 | } |
1361 | ||
1362 | ||
5eca1c10 | 1363 | static inline pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) |
7af57294 | 1364 | { |
52ee2dfd | 1365 | return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns); |
7af57294 PE |
1366 | } |
1367 | ||
7af57294 PE |
1368 | static inline pid_t task_session_vnr(struct task_struct *tsk) |
1369 | { | |
52ee2dfd | 1370 | return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL); |
7af57294 PE |
1371 | } |
1372 | ||
dd1c1f2f ON |
1373 | static inline pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) |
1374 | { | |
6883f81a | 1375 | return __task_pid_nr_ns(tsk, PIDTYPE_TGID, ns); |
dd1c1f2f ON |
1376 | } |
1377 | ||
1378 | static inline pid_t task_tgid_vnr(struct task_struct *tsk) | |
1379 | { | |
6883f81a | 1380 | return __task_pid_nr_ns(tsk, PIDTYPE_TGID, NULL); |
dd1c1f2f ON |
1381 | } |
1382 | ||
1383 | static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns) | |
1384 | { | |
1385 | pid_t pid = 0; | |
1386 | ||
1387 | rcu_read_lock(); | |
1388 | if (pid_alive(tsk)) | |
1389 | pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns); | |
1390 | rcu_read_unlock(); | |
1391 | ||
1392 | return pid; | |
1393 | } | |
1394 | ||
1395 | static inline pid_t task_ppid_nr(const struct task_struct *tsk) | |
1396 | { | |
1397 | return task_ppid_nr_ns(tsk, &init_pid_ns); | |
1398 | } | |
1399 | ||
5eca1c10 | 1400 | /* Obsolete, do not use: */ |
1b0f7ffd ON |
1401 | static inline pid_t task_pgrp_nr(struct task_struct *tsk) |
1402 | { | |
1403 | return task_pgrp_nr_ns(tsk, &init_pid_ns); | |
1404 | } | |
7af57294 | 1405 | |
06eb6184 PZ |
1406 | #define TASK_REPORT_IDLE (TASK_REPORT + 1) |
1407 | #define TASK_REPORT_MAX (TASK_REPORT_IDLE << 1) | |
1408 | ||
1d48b080 | 1409 | static inline unsigned int task_state_index(struct task_struct *tsk) |
20435d84 | 1410 | { |
1593baab PZ |
1411 | unsigned int tsk_state = READ_ONCE(tsk->state); |
1412 | unsigned int state = (tsk_state | tsk->exit_state) & TASK_REPORT; | |
20435d84 | 1413 | |
06eb6184 PZ |
1414 | BUILD_BUG_ON_NOT_POWER_OF_2(TASK_REPORT_MAX); |
1415 | ||
06eb6184 PZ |
1416 | if (tsk_state == TASK_IDLE) |
1417 | state = TASK_REPORT_IDLE; | |
1418 | ||
1593baab PZ |
1419 | return fls(state); |
1420 | } | |
1421 | ||
1d48b080 | 1422 | static inline char task_index_to_char(unsigned int state) |
1593baab | 1423 | { |
8ef9925b | 1424 | static const char state_char[] = "RSDTtXZPI"; |
1593baab | 1425 | |
06eb6184 | 1426 | BUILD_BUG_ON(1 + ilog2(TASK_REPORT_MAX) != sizeof(state_char) - 1); |
20435d84 | 1427 | |
1593baab PZ |
1428 | return state_char[state]; |
1429 | } | |
1430 | ||
1431 | static inline char task_state_to_char(struct task_struct *tsk) | |
1432 | { | |
1d48b080 | 1433 | return task_index_to_char(task_state_index(tsk)); |
20435d84 XX |
1434 | } |
1435 | ||
f400e198 | 1436 | /** |
570f5241 SS |
1437 | * is_global_init - check if a task structure is init. Since init |
1438 | * is free to have sub-threads we need to check tgid. | |
3260259f H |
1439 | * @tsk: Task structure to be checked. |
1440 | * | |
1441 | * Check if a task structure is the first user space task the kernel created. | |
e69f6186 YB |
1442 | * |
1443 | * Return: 1 if the task structure is init. 0 otherwise. | |
b460cbc5 | 1444 | */ |
e868171a | 1445 | static inline int is_global_init(struct task_struct *tsk) |
b461cc03 | 1446 | { |
570f5241 | 1447 | return task_tgid_nr(tsk) == 1; |
b461cc03 | 1448 | } |
b460cbc5 | 1449 | |
9ec52099 CLG |
1450 | extern struct pid *cad_pid; |
1451 | ||
1da177e4 LT |
1452 | /* |
1453 | * Per process flags | |
1454 | */ | |
5eca1c10 IM |
1455 | #define PF_IDLE 0x00000002 /* I am an IDLE thread */ |
1456 | #define PF_EXITING 0x00000004 /* Getting shut down */ | |
5eca1c10 IM |
1457 | #define PF_VCPU 0x00000010 /* I'm a virtual CPU */ |
1458 | #define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */ | |
1459 | #define PF_FORKNOEXEC 0x00000040 /* Forked but didn't exec */ | |
1460 | #define PF_MCE_PROCESS 0x00000080 /* Process policy on mce errors */ | |
1461 | #define PF_SUPERPRIV 0x00000100 /* Used super-user privileges */ | |
1462 | #define PF_DUMPCORE 0x00000200 /* Dumped core */ | |
1463 | #define PF_SIGNALED 0x00000400 /* Killed by a signal */ | |
1464 | #define PF_MEMALLOC 0x00000800 /* Allocating memory */ | |
1465 | #define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */ | |
1466 | #define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */ | |
1467 | #define PF_USED_ASYNC 0x00004000 /* Used async_schedule*(), used by module init */ | |
1468 | #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */ | |
1469 | #define PF_FROZEN 0x00010000 /* Frozen for system suspend */ | |
7dea19f9 MH |
1470 | #define PF_KSWAPD 0x00020000 /* I am kswapd */ |
1471 | #define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */ | |
1472 | #define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */ | |
5eca1c10 IM |
1473 | #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */ |
1474 | #define PF_KTHREAD 0x00200000 /* I am a kernel thread */ | |
1475 | #define PF_RANDOMIZE 0x00400000 /* Randomize virtual address space */ | |
1476 | #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */ | |
eb414681 | 1477 | #define PF_MEMSTALL 0x01000000 /* Stalled due to lack of memory */ |
73ab1cb2 | 1478 | #define PF_UMH 0x02000000 /* I'm an Usermodehelper process */ |
3bd37062 | 1479 | #define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_mask */ |
5eca1c10 | 1480 | #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */ |
d7fefcc8 | 1481 | #define PF_MEMALLOC_NOCMA 0x10000000 /* All allocation request will have _GFP_MOVABLE cleared */ |
771b53d0 | 1482 | #define PF_IO_WORKER 0x20000000 /* Task is an IO worker */ |
5eca1c10 IM |
1483 | #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */ |
1484 | #define PF_SUSPEND_TASK 0x80000000 /* This thread called freeze_processes() and should not be frozen */ | |
1da177e4 LT |
1485 | |
1486 | /* | |
1487 | * Only the _current_ task can read/write to tsk->flags, but other | |
1488 | * tasks can access tsk->flags in readonly mode for example | |
1489 | * with tsk_used_math (like during threaded core dumping). | |
1490 | * There is however an exception to this rule during ptrace | |
1491 | * or during fork: the ptracer task is allowed to write to the | |
1492 | * child->flags of its traced child (same goes for fork, the parent | |
1493 | * can write to the child->flags), because we're guaranteed the | |
1494 | * child is not running and in turn not changing child->flags | |
1495 | * at the same time the parent does it. | |
1496 | */ | |
5eca1c10 IM |
1497 | #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0) |
1498 | #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0) | |
1499 | #define clear_used_math() clear_stopped_child_used_math(current) | |
1500 | #define set_used_math() set_stopped_child_used_math(current) | |
1501 | ||
1da177e4 LT |
1502 | #define conditional_stopped_child_used_math(condition, child) \ |
1503 | do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0) | |
5eca1c10 IM |
1504 | |
1505 | #define conditional_used_math(condition) conditional_stopped_child_used_math(condition, current) | |
1506 | ||
1da177e4 LT |
1507 | #define copy_to_stopped_child_used_math(child) \ |
1508 | do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0) | |
5eca1c10 | 1509 | |
1da177e4 | 1510 | /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */ |
5eca1c10 IM |
1511 | #define tsk_used_math(p) ((p)->flags & PF_USED_MATH) |
1512 | #define used_math() tsk_used_math(current) | |
1da177e4 | 1513 | |
62ec05dd TG |
1514 | static inline bool is_percpu_thread(void) |
1515 | { | |
1516 | #ifdef CONFIG_SMP | |
1517 | return (current->flags & PF_NO_SETAFFINITY) && | |
1518 | (current->nr_cpus_allowed == 1); | |
1519 | #else | |
1520 | return true; | |
1521 | #endif | |
1522 | } | |
1523 | ||
1d4457f9 | 1524 | /* Per-process atomic flags. */ |
5eca1c10 IM |
1525 | #define PFA_NO_NEW_PRIVS 0 /* May not gain new privileges. */ |
1526 | #define PFA_SPREAD_PAGE 1 /* Spread page cache over cpuset */ | |
1527 | #define PFA_SPREAD_SLAB 2 /* Spread some slab caches over cpuset */ | |
356e4bff TG |
1528 | #define PFA_SPEC_SSB_DISABLE 3 /* Speculative Store Bypass disabled */ |
1529 | #define PFA_SPEC_SSB_FORCE_DISABLE 4 /* Speculative Store Bypass force disabled*/ | |
9137bb27 TG |
1530 | #define PFA_SPEC_IB_DISABLE 5 /* Indirect branch speculation restricted */ |
1531 | #define PFA_SPEC_IB_FORCE_DISABLE 6 /* Indirect branch speculation permanently restricted */ | |
71368af9 | 1532 | #define PFA_SPEC_SSB_NOEXEC 7 /* Speculative Store Bypass clear on execve() */ |
1d4457f9 | 1533 | |
e0e5070b ZL |
1534 | #define TASK_PFA_TEST(name, func) \ |
1535 | static inline bool task_##func(struct task_struct *p) \ | |
1536 | { return test_bit(PFA_##name, &p->atomic_flags); } | |
5eca1c10 | 1537 | |
e0e5070b ZL |
1538 | #define TASK_PFA_SET(name, func) \ |
1539 | static inline void task_set_##func(struct task_struct *p) \ | |
1540 | { set_bit(PFA_##name, &p->atomic_flags); } | |
5eca1c10 | 1541 | |
e0e5070b ZL |
1542 | #define TASK_PFA_CLEAR(name, func) \ |
1543 | static inline void task_clear_##func(struct task_struct *p) \ | |
1544 | { clear_bit(PFA_##name, &p->atomic_flags); } | |
1545 | ||
1546 | TASK_PFA_TEST(NO_NEW_PRIVS, no_new_privs) | |
1547 | TASK_PFA_SET(NO_NEW_PRIVS, no_new_privs) | |
1d4457f9 | 1548 | |
2ad654bc ZL |
1549 | TASK_PFA_TEST(SPREAD_PAGE, spread_page) |
1550 | TASK_PFA_SET(SPREAD_PAGE, spread_page) | |
1551 | TASK_PFA_CLEAR(SPREAD_PAGE, spread_page) | |
1552 | ||
1553 | TASK_PFA_TEST(SPREAD_SLAB, spread_slab) | |
1554 | TASK_PFA_SET(SPREAD_SLAB, spread_slab) | |
1555 | TASK_PFA_CLEAR(SPREAD_SLAB, spread_slab) | |
1d4457f9 | 1556 | |
356e4bff TG |
1557 | TASK_PFA_TEST(SPEC_SSB_DISABLE, spec_ssb_disable) |
1558 | TASK_PFA_SET(SPEC_SSB_DISABLE, spec_ssb_disable) | |
1559 | TASK_PFA_CLEAR(SPEC_SSB_DISABLE, spec_ssb_disable) | |
1560 | ||
71368af9 WL |
1561 | TASK_PFA_TEST(SPEC_SSB_NOEXEC, spec_ssb_noexec) |
1562 | TASK_PFA_SET(SPEC_SSB_NOEXEC, spec_ssb_noexec) | |
1563 | TASK_PFA_CLEAR(SPEC_SSB_NOEXEC, spec_ssb_noexec) | |
1564 | ||
356e4bff TG |
1565 | TASK_PFA_TEST(SPEC_SSB_FORCE_DISABLE, spec_ssb_force_disable) |
1566 | TASK_PFA_SET(SPEC_SSB_FORCE_DISABLE, spec_ssb_force_disable) | |
1567 | ||
9137bb27 TG |
1568 | TASK_PFA_TEST(SPEC_IB_DISABLE, spec_ib_disable) |
1569 | TASK_PFA_SET(SPEC_IB_DISABLE, spec_ib_disable) | |
1570 | TASK_PFA_CLEAR(SPEC_IB_DISABLE, spec_ib_disable) | |
1571 | ||
1572 | TASK_PFA_TEST(SPEC_IB_FORCE_DISABLE, spec_ib_force_disable) | |
1573 | TASK_PFA_SET(SPEC_IB_FORCE_DISABLE, spec_ib_force_disable) | |
1574 | ||
5eca1c10 | 1575 | static inline void |
717a94b5 | 1576 | current_restore_flags(unsigned long orig_flags, unsigned long flags) |
907aed48 | 1577 | { |
717a94b5 N |
1578 | current->flags &= ~flags; |
1579 | current->flags |= orig_flags & flags; | |
907aed48 MG |
1580 | } |
1581 | ||
5eca1c10 IM |
1582 | extern int cpuset_cpumask_can_shrink(const struct cpumask *cur, const struct cpumask *trial); |
1583 | extern int task_can_attach(struct task_struct *p, const struct cpumask *cs_cpus_allowed); | |
1da177e4 | 1584 | #ifdef CONFIG_SMP |
5eca1c10 IM |
1585 | extern void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask); |
1586 | extern int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask); | |
1da177e4 | 1587 | #else |
5eca1c10 | 1588 | static inline void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) |
1e1b6c51 KM |
1589 | { |
1590 | } | |
5eca1c10 | 1591 | static inline int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) |
1da177e4 | 1592 | { |
96f874e2 | 1593 | if (!cpumask_test_cpu(0, new_mask)) |
1da177e4 LT |
1594 | return -EINVAL; |
1595 | return 0; | |
1596 | } | |
1597 | #endif | |
e0ad9556 | 1598 | |
fa93384f | 1599 | extern int yield_to(struct task_struct *p, bool preempt); |
36c8b586 IM |
1600 | extern void set_user_nice(struct task_struct *p, long nice); |
1601 | extern int task_prio(const struct task_struct *p); | |
5eca1c10 | 1602 | |
d0ea0268 DY |
1603 | /** |
1604 | * task_nice - return the nice value of a given task. | |
1605 | * @p: the task in question. | |
1606 | * | |
1607 | * Return: The nice value [ -20 ... 0 ... 19 ]. | |
1608 | */ | |
1609 | static inline int task_nice(const struct task_struct *p) | |
1610 | { | |
1611 | return PRIO_TO_NICE((p)->static_prio); | |
1612 | } | |
5eca1c10 | 1613 | |
36c8b586 IM |
1614 | extern int can_nice(const struct task_struct *p, const int nice); |
1615 | extern int task_curr(const struct task_struct *p); | |
1da177e4 | 1616 | extern int idle_cpu(int cpu); |
943d355d | 1617 | extern int available_idle_cpu(int cpu); |
5eca1c10 IM |
1618 | extern int sched_setscheduler(struct task_struct *, int, const struct sched_param *); |
1619 | extern int sched_setscheduler_nocheck(struct task_struct *, int, const struct sched_param *); | |
1620 | extern int sched_setattr(struct task_struct *, const struct sched_attr *); | |
794a56eb | 1621 | extern int sched_setattr_nocheck(struct task_struct *, const struct sched_attr *); |
36c8b586 | 1622 | extern struct task_struct *idle_task(int cpu); |
5eca1c10 | 1623 | |
c4f30608 PM |
1624 | /** |
1625 | * is_idle_task - is the specified task an idle task? | |
fa757281 | 1626 | * @p: the task in question. |
e69f6186 YB |
1627 | * |
1628 | * Return: 1 if @p is an idle task. 0 otherwise. | |
c4f30608 | 1629 | */ |
7061ca3b | 1630 | static inline bool is_idle_task(const struct task_struct *p) |
c4f30608 | 1631 | { |
c1de45ca | 1632 | return !!(p->flags & PF_IDLE); |
c4f30608 | 1633 | } |
5eca1c10 | 1634 | |
36c8b586 | 1635 | extern struct task_struct *curr_task(int cpu); |
a458ae2e | 1636 | extern void ia64_set_curr_task(int cpu, struct task_struct *p); |
1da177e4 LT |
1637 | |
1638 | void yield(void); | |
1639 | ||
1da177e4 | 1640 | union thread_union { |
0500871f DH |
1641 | #ifndef CONFIG_ARCH_TASK_STRUCT_ON_STACK |
1642 | struct task_struct task; | |
1643 | #endif | |
c65eacbe | 1644 | #ifndef CONFIG_THREAD_INFO_IN_TASK |
1da177e4 | 1645 | struct thread_info thread_info; |
c65eacbe | 1646 | #endif |
1da177e4 LT |
1647 | unsigned long stack[THREAD_SIZE/sizeof(long)]; |
1648 | }; | |
1649 | ||
0500871f DH |
1650 | #ifndef CONFIG_THREAD_INFO_IN_TASK |
1651 | extern struct thread_info init_thread_info; | |
1652 | #endif | |
1653 | ||
1654 | extern unsigned long init_stack[THREAD_SIZE / sizeof(unsigned long)]; | |
1655 | ||
f3ac6067 IM |
1656 | #ifdef CONFIG_THREAD_INFO_IN_TASK |
1657 | static inline struct thread_info *task_thread_info(struct task_struct *task) | |
1658 | { | |
1659 | return &task->thread_info; | |
1660 | } | |
1661 | #elif !defined(__HAVE_THREAD_FUNCTIONS) | |
1662 | # define task_thread_info(task) ((struct thread_info *)(task)->stack) | |
1663 | #endif | |
1664 | ||
198fe21b PE |
1665 | /* |
1666 | * find a task by one of its numerical ids | |
1667 | * | |
198fe21b PE |
1668 | * find_task_by_pid_ns(): |
1669 | * finds a task by its pid in the specified namespace | |
228ebcbe PE |
1670 | * find_task_by_vpid(): |
1671 | * finds a task by its virtual pid | |
198fe21b | 1672 | * |
e49859e7 | 1673 | * see also find_vpid() etc in include/linux/pid.h |
198fe21b PE |
1674 | */ |
1675 | ||
228ebcbe | 1676 | extern struct task_struct *find_task_by_vpid(pid_t nr); |
5eca1c10 | 1677 | extern struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns); |
198fe21b | 1678 | |
2ee08260 MR |
1679 | /* |
1680 | * find a task by its virtual pid and get the task struct | |
1681 | */ | |
1682 | extern struct task_struct *find_get_task_by_vpid(pid_t nr); | |
1683 | ||
b3c97528 HH |
1684 | extern int wake_up_state(struct task_struct *tsk, unsigned int state); |
1685 | extern int wake_up_process(struct task_struct *tsk); | |
3e51e3ed | 1686 | extern void wake_up_new_task(struct task_struct *tsk); |
5eca1c10 | 1687 | |
1da177e4 | 1688 | #ifdef CONFIG_SMP |
5eca1c10 | 1689 | extern void kick_process(struct task_struct *tsk); |
1da177e4 | 1690 | #else |
5eca1c10 | 1691 | static inline void kick_process(struct task_struct *tsk) { } |
1da177e4 | 1692 | #endif |
1da177e4 | 1693 | |
82b89778 | 1694 | extern void __set_task_comm(struct task_struct *tsk, const char *from, bool exec); |
5eca1c10 | 1695 | |
82b89778 AH |
1696 | static inline void set_task_comm(struct task_struct *tsk, const char *from) |
1697 | { | |
1698 | __set_task_comm(tsk, from, false); | |
1699 | } | |
5eca1c10 | 1700 | |
3756f640 AB |
1701 | extern char *__get_task_comm(char *to, size_t len, struct task_struct *tsk); |
1702 | #define get_task_comm(buf, tsk) ({ \ | |
1703 | BUILD_BUG_ON(sizeof(buf) != TASK_COMM_LEN); \ | |
1704 | __get_task_comm(buf, sizeof(buf), tsk); \ | |
1705 | }) | |
1da177e4 LT |
1706 | |
1707 | #ifdef CONFIG_SMP | |
317f3941 | 1708 | void scheduler_ipi(void); |
85ba2d86 | 1709 | extern unsigned long wait_task_inactive(struct task_struct *, long match_state); |
1da177e4 | 1710 | #else |
184748cc | 1711 | static inline void scheduler_ipi(void) { } |
5eca1c10 | 1712 | static inline unsigned long wait_task_inactive(struct task_struct *p, long match_state) |
85ba2d86 RM |
1713 | { |
1714 | return 1; | |
1715 | } | |
1da177e4 LT |
1716 | #endif |
1717 | ||
5eca1c10 IM |
1718 | /* |
1719 | * Set thread flags in other task's structures. | |
1720 | * See asm/thread_info.h for TIF_xxxx flags available: | |
1da177e4 LT |
1721 | */ |
1722 | static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1723 | { | |
a1261f54 | 1724 | set_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1725 | } |
1726 | ||
1727 | static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1728 | { | |
a1261f54 | 1729 | clear_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1730 | } |
1731 | ||
93ee37c2 DM |
1732 | static inline void update_tsk_thread_flag(struct task_struct *tsk, int flag, |
1733 | bool value) | |
1734 | { | |
1735 | update_ti_thread_flag(task_thread_info(tsk), flag, value); | |
1736 | } | |
1737 | ||
1da177e4 LT |
1738 | static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag) |
1739 | { | |
a1261f54 | 1740 | return test_and_set_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1741 | } |
1742 | ||
1743 | static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1744 | { | |
a1261f54 | 1745 | return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1746 | } |
1747 | ||
1748 | static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag) | |
1749 | { | |
a1261f54 | 1750 | return test_ti_thread_flag(task_thread_info(tsk), flag); |
1da177e4 LT |
1751 | } |
1752 | ||
1753 | static inline void set_tsk_need_resched(struct task_struct *tsk) | |
1754 | { | |
1755 | set_tsk_thread_flag(tsk,TIF_NEED_RESCHED); | |
1756 | } | |
1757 | ||
1758 | static inline void clear_tsk_need_resched(struct task_struct *tsk) | |
1759 | { | |
1760 | clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED); | |
1761 | } | |
1762 | ||
8ae121ac GH |
1763 | static inline int test_tsk_need_resched(struct task_struct *tsk) |
1764 | { | |
1765 | return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED)); | |
1766 | } | |
1767 | ||
1da177e4 LT |
1768 | /* |
1769 | * cond_resched() and cond_resched_lock(): latency reduction via | |
1770 | * explicit rescheduling in places that are safe. The return | |
1771 | * value indicates whether a reschedule was done in fact. | |
1772 | * cond_resched_lock() will drop the spinlock before scheduling, | |
1da177e4 | 1773 | */ |
c1a280b6 | 1774 | #ifndef CONFIG_PREEMPTION |
c3921ab7 | 1775 | extern int _cond_resched(void); |
35a773a0 PZ |
1776 | #else |
1777 | static inline int _cond_resched(void) { return 0; } | |
1778 | #endif | |
6f80bd98 | 1779 | |
613afbf8 | 1780 | #define cond_resched() ({ \ |
3427445a | 1781 | ___might_sleep(__FILE__, __LINE__, 0); \ |
613afbf8 FW |
1782 | _cond_resched(); \ |
1783 | }) | |
6f80bd98 | 1784 | |
613afbf8 FW |
1785 | extern int __cond_resched_lock(spinlock_t *lock); |
1786 | ||
1787 | #define cond_resched_lock(lock) ({ \ | |
3427445a | 1788 | ___might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET);\ |
613afbf8 FW |
1789 | __cond_resched_lock(lock); \ |
1790 | }) | |
1791 | ||
f6f3c437 SH |
1792 | static inline void cond_resched_rcu(void) |
1793 | { | |
1794 | #if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU) | |
1795 | rcu_read_unlock(); | |
1796 | cond_resched(); | |
1797 | rcu_read_lock(); | |
1798 | #endif | |
1799 | } | |
1800 | ||
1da177e4 LT |
1801 | /* |
1802 | * Does a critical section need to be broken due to another | |
c1a280b6 | 1803 | * task waiting?: (technically does not depend on CONFIG_PREEMPTION, |
95c354fe | 1804 | * but a general need for low latency) |
1da177e4 | 1805 | */ |
95c354fe | 1806 | static inline int spin_needbreak(spinlock_t *lock) |
1da177e4 | 1807 | { |
c1a280b6 | 1808 | #ifdef CONFIG_PREEMPTION |
95c354fe NP |
1809 | return spin_is_contended(lock); |
1810 | #else | |
1da177e4 | 1811 | return 0; |
95c354fe | 1812 | #endif |
1da177e4 LT |
1813 | } |
1814 | ||
75f93fed PZ |
1815 | static __always_inline bool need_resched(void) |
1816 | { | |
1817 | return unlikely(tif_need_resched()); | |
1818 | } | |
1819 | ||
1da177e4 LT |
1820 | /* |
1821 | * Wrappers for p->thread_info->cpu access. No-op on UP. | |
1822 | */ | |
1823 | #ifdef CONFIG_SMP | |
1824 | ||
1825 | static inline unsigned int task_cpu(const struct task_struct *p) | |
1826 | { | |
c65eacbe | 1827 | #ifdef CONFIG_THREAD_INFO_IN_TASK |
c546951d | 1828 | return READ_ONCE(p->cpu); |
c65eacbe | 1829 | #else |
c546951d | 1830 | return READ_ONCE(task_thread_info(p)->cpu); |
c65eacbe | 1831 | #endif |
1da177e4 LT |
1832 | } |
1833 | ||
c65cc870 | 1834 | extern void set_task_cpu(struct task_struct *p, unsigned int cpu); |
1da177e4 LT |
1835 | |
1836 | #else | |
1837 | ||
1838 | static inline unsigned int task_cpu(const struct task_struct *p) | |
1839 | { | |
1840 | return 0; | |
1841 | } | |
1842 | ||
1843 | static inline void set_task_cpu(struct task_struct *p, unsigned int cpu) | |
1844 | { | |
1845 | } | |
1846 | ||
1847 | #endif /* CONFIG_SMP */ | |
1848 | ||
d9345c65 PX |
1849 | /* |
1850 | * In order to reduce various lock holder preemption latencies provide an | |
1851 | * interface to see if a vCPU is currently running or not. | |
1852 | * | |
1853 | * This allows us to terminate optimistic spin loops and block, analogous to | |
1854 | * the native optimistic spin heuristic of testing if the lock owner task is | |
1855 | * running or not. | |
1856 | */ | |
1857 | #ifndef vcpu_is_preempted | |
42fd8baa QC |
1858 | static inline bool vcpu_is_preempted(int cpu) |
1859 | { | |
1860 | return false; | |
1861 | } | |
d9345c65 PX |
1862 | #endif |
1863 | ||
96f874e2 RR |
1864 | extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask); |
1865 | extern long sched_getaffinity(pid_t pid, struct cpumask *mask); | |
5c45bf27 | 1866 | |
82455257 DH |
1867 | #ifndef TASK_SIZE_OF |
1868 | #define TASK_SIZE_OF(tsk) TASK_SIZE | |
1869 | #endif | |
1870 | ||
d7822b1e MD |
1871 | #ifdef CONFIG_RSEQ |
1872 | ||
1873 | /* | |
1874 | * Map the event mask on the user-space ABI enum rseq_cs_flags | |
1875 | * for direct mask checks. | |
1876 | */ | |
1877 | enum rseq_event_mask_bits { | |
1878 | RSEQ_EVENT_PREEMPT_BIT = RSEQ_CS_FLAG_NO_RESTART_ON_PREEMPT_BIT, | |
1879 | RSEQ_EVENT_SIGNAL_BIT = RSEQ_CS_FLAG_NO_RESTART_ON_SIGNAL_BIT, | |
1880 | RSEQ_EVENT_MIGRATE_BIT = RSEQ_CS_FLAG_NO_RESTART_ON_MIGRATE_BIT, | |
1881 | }; | |
1882 | ||
1883 | enum rseq_event_mask { | |
1884 | RSEQ_EVENT_PREEMPT = (1U << RSEQ_EVENT_PREEMPT_BIT), | |
1885 | RSEQ_EVENT_SIGNAL = (1U << RSEQ_EVENT_SIGNAL_BIT), | |
1886 | RSEQ_EVENT_MIGRATE = (1U << RSEQ_EVENT_MIGRATE_BIT), | |
1887 | }; | |
1888 | ||
1889 | static inline void rseq_set_notify_resume(struct task_struct *t) | |
1890 | { | |
1891 | if (t->rseq) | |
1892 | set_tsk_thread_flag(t, TIF_NOTIFY_RESUME); | |
1893 | } | |
1894 | ||
784e0300 | 1895 | void __rseq_handle_notify_resume(struct ksignal *sig, struct pt_regs *regs); |
d7822b1e | 1896 | |
784e0300 WD |
1897 | static inline void rseq_handle_notify_resume(struct ksignal *ksig, |
1898 | struct pt_regs *regs) | |
d7822b1e MD |
1899 | { |
1900 | if (current->rseq) | |
784e0300 | 1901 | __rseq_handle_notify_resume(ksig, regs); |
d7822b1e MD |
1902 | } |
1903 | ||
784e0300 WD |
1904 | static inline void rseq_signal_deliver(struct ksignal *ksig, |
1905 | struct pt_regs *regs) | |
d7822b1e MD |
1906 | { |
1907 | preempt_disable(); | |
1908 | __set_bit(RSEQ_EVENT_SIGNAL_BIT, ¤t->rseq_event_mask); | |
1909 | preempt_enable(); | |
784e0300 | 1910 | rseq_handle_notify_resume(ksig, regs); |
d7822b1e MD |
1911 | } |
1912 | ||
1913 | /* rseq_preempt() requires preemption to be disabled. */ | |
1914 | static inline void rseq_preempt(struct task_struct *t) | |
1915 | { | |
1916 | __set_bit(RSEQ_EVENT_PREEMPT_BIT, &t->rseq_event_mask); | |
1917 | rseq_set_notify_resume(t); | |
1918 | } | |
1919 | ||
1920 | /* rseq_migrate() requires preemption to be disabled. */ | |
1921 | static inline void rseq_migrate(struct task_struct *t) | |
1922 | { | |
1923 | __set_bit(RSEQ_EVENT_MIGRATE_BIT, &t->rseq_event_mask); | |
1924 | rseq_set_notify_resume(t); | |
1925 | } | |
1926 | ||
1927 | /* | |
1928 | * If parent process has a registered restartable sequences area, the | |
463f550f | 1929 | * child inherits. Unregister rseq for a clone with CLONE_VM set. |
d7822b1e MD |
1930 | */ |
1931 | static inline void rseq_fork(struct task_struct *t, unsigned long clone_flags) | |
1932 | { | |
463f550f | 1933 | if (clone_flags & CLONE_VM) { |
d7822b1e | 1934 | t->rseq = NULL; |
d7822b1e MD |
1935 | t->rseq_sig = 0; |
1936 | t->rseq_event_mask = 0; | |
1937 | } else { | |
1938 | t->rseq = current->rseq; | |
d7822b1e MD |
1939 | t->rseq_sig = current->rseq_sig; |
1940 | t->rseq_event_mask = current->rseq_event_mask; | |
d7822b1e MD |
1941 | } |
1942 | } | |
1943 | ||
1944 | static inline void rseq_execve(struct task_struct *t) | |
1945 | { | |
1946 | t->rseq = NULL; | |
d7822b1e MD |
1947 | t->rseq_sig = 0; |
1948 | t->rseq_event_mask = 0; | |
1949 | } | |
1950 | ||
1951 | #else | |
1952 | ||
1953 | static inline void rseq_set_notify_resume(struct task_struct *t) | |
1954 | { | |
1955 | } | |
784e0300 WD |
1956 | static inline void rseq_handle_notify_resume(struct ksignal *ksig, |
1957 | struct pt_regs *regs) | |
d7822b1e MD |
1958 | { |
1959 | } | |
784e0300 WD |
1960 | static inline void rseq_signal_deliver(struct ksignal *ksig, |
1961 | struct pt_regs *regs) | |
d7822b1e MD |
1962 | { |
1963 | } | |
1964 | static inline void rseq_preempt(struct task_struct *t) | |
1965 | { | |
1966 | } | |
1967 | static inline void rseq_migrate(struct task_struct *t) | |
1968 | { | |
1969 | } | |
1970 | static inline void rseq_fork(struct task_struct *t, unsigned long clone_flags) | |
1971 | { | |
1972 | } | |
1973 | static inline void rseq_execve(struct task_struct *t) | |
1974 | { | |
1975 | } | |
1976 | ||
1977 | #endif | |
1978 | ||
73ab1cb2 TY |
1979 | void __exit_umh(struct task_struct *tsk); |
1980 | ||
1981 | static inline void exit_umh(struct task_struct *tsk) | |
1982 | { | |
1983 | if (unlikely(tsk->flags & PF_UMH)) | |
1984 | __exit_umh(tsk); | |
1985 | } | |
1986 | ||
d7822b1e MD |
1987 | #ifdef CONFIG_DEBUG_RSEQ |
1988 | ||
1989 | void rseq_syscall(struct pt_regs *regs); | |
1990 | ||
1991 | #else | |
1992 | ||
1993 | static inline void rseq_syscall(struct pt_regs *regs) | |
1994 | { | |
1995 | } | |
1996 | ||
1997 | #endif | |
1998 | ||
3c93a0c0 QY |
1999 | const struct sched_avg *sched_trace_cfs_rq_avg(struct cfs_rq *cfs_rq); |
2000 | char *sched_trace_cfs_rq_path(struct cfs_rq *cfs_rq, char *str, int len); | |
2001 | int sched_trace_cfs_rq_cpu(struct cfs_rq *cfs_rq); | |
2002 | ||
2003 | const struct sched_avg *sched_trace_rq_avg_rt(struct rq *rq); | |
2004 | const struct sched_avg *sched_trace_rq_avg_dl(struct rq *rq); | |
2005 | const struct sched_avg *sched_trace_rq_avg_irq(struct rq *rq); | |
2006 | ||
2007 | int sched_trace_rq_cpu(struct rq *rq); | |
2008 | ||
2009 | const struct cpumask *sched_trace_rd_span(struct root_domain *rd); | |
2010 | ||
1da177e4 | 2011 | #endif |