]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/linux/sched.h
sched: Check for idle task in might_sleep()
[mirror_ubuntu-artful-kernel.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
607ca46e 4#include <uapi/linux/sched.h>
b7b3c76a 5
5c228079
DY
6#include <linux/sched/prio.h>
7
b7b3c76a
DW
8
9struct sched_param {
10 int sched_priority;
11};
12
1da177e4
LT
13#include <asm/param.h> /* for HZ */
14
1da177e4
LT
15#include <linux/capability.h>
16#include <linux/threads.h>
17#include <linux/kernel.h>
18#include <linux/types.h>
19#include <linux/timex.h>
20#include <linux/jiffies.h>
fb00aca4 21#include <linux/plist.h>
1da177e4
LT
22#include <linux/rbtree.h>
23#include <linux/thread_info.h>
24#include <linux/cpumask.h>
25#include <linux/errno.h>
26#include <linux/nodemask.h>
c92ff1bd 27#include <linux/mm_types.h>
00d1a39e 28#include <linux/preempt_mask.h>
1da177e4 29
1da177e4
LT
30#include <asm/page.h>
31#include <asm/ptrace.h>
1da177e4
LT
32#include <asm/cputime.h>
33
34#include <linux/smp.h>
35#include <linux/sem.h>
36#include <linux/signal.h>
1da177e4
LT
37#include <linux/compiler.h>
38#include <linux/completion.h>
39#include <linux/pid.h>
40#include <linux/percpu.h>
41#include <linux/topology.h>
3e26c149 42#include <linux/proportions.h>
1da177e4 43#include <linux/seccomp.h>
e56d0903 44#include <linux/rcupdate.h>
05725f7e 45#include <linux/rculist.h>
23f78d4a 46#include <linux/rtmutex.h>
1da177e4 47
a3b6714e
DW
48#include <linux/time.h>
49#include <linux/param.h>
50#include <linux/resource.h>
51#include <linux/timer.h>
52#include <linux/hrtimer.h>
7c3ab738 53#include <linux/task_io_accounting.h>
9745512c 54#include <linux/latencytop.h>
9e2b2dc4 55#include <linux/cred.h>
fa14ff4a 56#include <linux/llist.h>
7b44ab97 57#include <linux/uidgid.h>
21caf2fc 58#include <linux/gfp.h>
a3b6714e
DW
59
60#include <asm/processor.h>
36d57ac4 61
d50dde5a
DF
62#define SCHED_ATTR_SIZE_VER0 48 /* sizeof first published struct */
63
64/*
65 * Extended scheduling parameters data structure.
66 *
67 * This is needed because the original struct sched_param can not be
68 * altered without introducing ABI issues with legacy applications
69 * (e.g., in sched_getparam()).
70 *
71 * However, the possibility of specifying more than just a priority for
72 * the tasks may be useful for a wide variety of application fields, e.g.,
73 * multimedia, streaming, automation and control, and many others.
74 *
75 * This variant (sched_attr) is meant at describing a so-called
76 * sporadic time-constrained task. In such model a task is specified by:
77 * - the activation period or minimum instance inter-arrival time;
78 * - the maximum (or average, depending on the actual scheduling
79 * discipline) computation time of all instances, a.k.a. runtime;
80 * - the deadline (relative to the actual activation time) of each
81 * instance.
82 * Very briefly, a periodic (sporadic) task asks for the execution of
83 * some specific computation --which is typically called an instance--
84 * (at most) every period. Moreover, each instance typically lasts no more
85 * than the runtime and must be completed by time instant t equal to
86 * the instance activation time + the deadline.
87 *
88 * This is reflected by the actual fields of the sched_attr structure:
89 *
90 * @size size of the structure, for fwd/bwd compat.
91 *
92 * @sched_policy task's scheduling policy
93 * @sched_flags for customizing the scheduler behaviour
94 * @sched_nice task's nice value (SCHED_NORMAL/BATCH)
95 * @sched_priority task's static priority (SCHED_FIFO/RR)
96 * @sched_deadline representative of the task's deadline
97 * @sched_runtime representative of the task's runtime
98 * @sched_period representative of the task's period
99 *
100 * Given this task model, there are a multiplicity of scheduling algorithms
101 * and policies, that can be used to ensure all the tasks will make their
102 * timing constraints.
aab03e05
DF
103 *
104 * As of now, the SCHED_DEADLINE policy (sched_dl scheduling class) is the
105 * only user of this new interface. More information about the algorithm
106 * available in the scheduling class file or in Documentation/.
d50dde5a
DF
107 */
108struct sched_attr {
109 u32 size;
110
111 u32 sched_policy;
112 u64 sched_flags;
113
114 /* SCHED_NORMAL, SCHED_BATCH */
115 s32 sched_nice;
116
117 /* SCHED_FIFO, SCHED_RR */
118 u32 sched_priority;
119
120 /* SCHED_DEADLINE */
121 u64 sched_runtime;
122 u64 sched_deadline;
123 u64 sched_period;
124};
125
1da177e4 126struct exec_domain;
c87e2837 127struct futex_pi_state;
286100a6 128struct robust_list_head;
bddd87c7 129struct bio_list;
5ad4e53b 130struct fs_struct;
cdd6c482 131struct perf_event_context;
73c10101 132struct blk_plug;
c4ad8f98 133struct filename;
1da177e4 134
1da177e4
LT
135/*
136 * List of flags we want to share for kernel threads,
137 * if only because they are not used by them anyway.
138 */
139#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
140
141/*
142 * These are the constant used to fake the fixed-point load-average
143 * counting. Some notes:
144 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
145 * a load-average precision of 10 bits integer + 11 bits fractional
146 * - if you want to count load-averages more often, you need more
147 * precision, or rounding will get you. With 2-second counting freq,
148 * the EXP_n values would be 1981, 2034 and 2043 if still using only
149 * 11 bit fractions.
150 */
151extern unsigned long avenrun[]; /* Load averages */
2d02494f 152extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
153
154#define FSHIFT 11 /* nr of bits of precision */
155#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 156#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
157#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
158#define EXP_5 2014 /* 1/exp(5sec/5min) */
159#define EXP_15 2037 /* 1/exp(5sec/15min) */
160
161#define CALC_LOAD(load,exp,n) \
162 load *= exp; \
163 load += n*(FIXED_1-exp); \
164 load >>= FSHIFT;
165
166extern unsigned long total_forks;
167extern int nr_threads;
1da177e4
LT
168DECLARE_PER_CPU(unsigned long, process_counts);
169extern int nr_processes(void);
170extern unsigned long nr_running(void);
1da177e4 171extern unsigned long nr_iowait(void);
8c215bd3 172extern unsigned long nr_iowait_cpu(int cpu);
69d25870
AV
173extern unsigned long this_cpu_load(void);
174
175
0f004f5a 176extern void calc_global_load(unsigned long ticks);
5aaa0b7a 177extern void update_cpu_load_nohz(void);
1da177e4 178
7e49fcce
SR
179extern unsigned long get_parent_ip(unsigned long addr);
180
b637a328
PM
181extern void dump_cpu_task(int cpu);
182
43ae34cb
IM
183struct seq_file;
184struct cfs_rq;
4cf86d77 185struct task_group;
43ae34cb
IM
186#ifdef CONFIG_SCHED_DEBUG
187extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
188extern void proc_sched_set_task(struct task_struct *p);
189extern void
5cef9eca 190print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb 191#endif
1da177e4 192
4a8342d2
LT
193/*
194 * Task state bitmask. NOTE! These bits are also
195 * encoded in fs/proc/array.c: get_task_state().
196 *
197 * We have two separate sets of flags: task->state
198 * is about runnability, while task->exit_state are
199 * about the task exiting. Confusing, but this way
200 * modifying one set can't modify the other one by
201 * mistake.
202 */
1da177e4
LT
203#define TASK_RUNNING 0
204#define TASK_INTERRUPTIBLE 1
205#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
206#define __TASK_STOPPED 4
207#define __TASK_TRACED 8
4a8342d2
LT
208/* in tsk->exit_state */
209#define EXIT_ZOMBIE 16
210#define EXIT_DEAD 32
211/* in tsk->state again */
af927232 212#define TASK_DEAD 64
f021a3c2 213#define TASK_WAKEKILL 128
e9c84311 214#define TASK_WAKING 256
f2530dc7
TG
215#define TASK_PARKED 512
216#define TASK_STATE_MAX 1024
f021a3c2 217
f2530dc7 218#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKWP"
73342151 219
e1781538
PZ
220extern char ___assert_task_state[1 - 2*!!(
221 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
222
223/* Convenience macros for the sake of set_task_state */
224#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
225#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
226#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 227
92a1f4bc
MW
228/* Convenience macros for the sake of wake_up */
229#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 230#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
231
232/* get_task_state() */
233#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2 234 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
74e37200 235 __TASK_TRACED | EXIT_ZOMBIE | EXIT_DEAD)
92a1f4bc 236
f021a3c2
MW
237#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
238#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 239#define task_is_stopped_or_traced(task) \
f021a3c2 240 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 241#define task_contributes_to_load(task) \
e3c8ca83 242 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
376fede8 243 (task->flags & PF_FROZEN) == 0)
1da177e4
LT
244
245#define __set_task_state(tsk, state_value) \
246 do { (tsk)->state = (state_value); } while (0)
247#define set_task_state(tsk, state_value) \
248 set_mb((tsk)->state, (state_value))
249
498d0c57
AM
250/*
251 * set_current_state() includes a barrier so that the write of current->state
252 * is correctly serialised wrt the caller's subsequent test of whether to
253 * actually sleep:
254 *
255 * set_current_state(TASK_UNINTERRUPTIBLE);
256 * if (do_i_need_to_sleep())
257 * schedule();
258 *
259 * If the caller does not need such serialisation then use __set_current_state()
260 */
1da177e4
LT
261#define __set_current_state(state_value) \
262 do { current->state = (state_value); } while (0)
263#define set_current_state(state_value) \
264 set_mb(current->state, (state_value))
265
266/* Task command name length */
267#define TASK_COMM_LEN 16
268
1da177e4
LT
269#include <linux/spinlock.h>
270
271/*
272 * This serializes "schedule()" and also protects
273 * the run-queue from deletions/modifications (but
274 * _adding_ to the beginning of the run-queue has
275 * a separate lock).
276 */
277extern rwlock_t tasklist_lock;
278extern spinlock_t mmlist_lock;
279
36c8b586 280struct task_struct;
1da177e4 281
db1466b3
PM
282#ifdef CONFIG_PROVE_RCU
283extern int lockdep_tasklist_lock_is_held(void);
284#endif /* #ifdef CONFIG_PROVE_RCU */
285
1da177e4
LT
286extern void sched_init(void);
287extern void sched_init_smp(void);
2d07b255 288extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 289extern void init_idle(struct task_struct *idle, int cpu);
1df21055 290extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 291
89f19f04 292extern int runqueue_is_locked(int cpu);
017730c1 293
3451d024 294#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
c1cc017c 295extern void nohz_balance_enter_idle(int cpu);
69e1e811 296extern void set_cpu_sd_state_idle(void);
83cd4fe2 297extern int get_nohz_timer_target(void);
46cb4b7c 298#else
c1cc017c 299static inline void nohz_balance_enter_idle(int cpu) { }
fdaabd80 300static inline void set_cpu_sd_state_idle(void) { }
46cb4b7c 301#endif
1da177e4 302
e59e2ae2 303/*
39bc89fd 304 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
305 */
306extern void show_state_filter(unsigned long state_filter);
307
308static inline void show_state(void)
309{
39bc89fd 310 show_state_filter(0);
e59e2ae2
IM
311}
312
1da177e4
LT
313extern void show_regs(struct pt_regs *);
314
315/*
316 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
317 * task), SP is the stack pointer of the first frame that should be shown in the back
318 * trace (or NULL if the entire call-chain of the task should be shown).
319 */
320extern void show_stack(struct task_struct *task, unsigned long *sp);
321
322void io_schedule(void);
323long io_schedule_timeout(long timeout);
324
325extern void cpu_init (void);
326extern void trap_init(void);
327extern void update_process_times(int user);
328extern void scheduler_tick(void);
329
82a1fcb9
IM
330extern void sched_show_task(struct task_struct *p);
331
19cc36c0 332#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 333extern void touch_softlockup_watchdog(void);
d6ad3e28 334extern void touch_softlockup_watchdog_sync(void);
04c9167f 335extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
336extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
337 void __user *buffer,
338 size_t *lenp, loff_t *ppos);
9c44bc03 339extern unsigned int softlockup_panic;
004417a6 340void lockup_detector_init(void);
8446f1d3 341#else
8446f1d3
IM
342static inline void touch_softlockup_watchdog(void)
343{
344}
d6ad3e28
JW
345static inline void touch_softlockup_watchdog_sync(void)
346{
347}
04c9167f
JF
348static inline void touch_all_softlockup_watchdogs(void)
349{
350}
004417a6
PZ
351static inline void lockup_detector_init(void)
352{
353}
8446f1d3
IM
354#endif
355
8b414521
MT
356#ifdef CONFIG_DETECT_HUNG_TASK
357void reset_hung_task_detector(void);
358#else
359static inline void reset_hung_task_detector(void)
360{
361}
362#endif
363
1da177e4
LT
364/* Attach to any functions which should be ignored in wchan output. */
365#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
366
367/* Linker adds these: start and end of __sched functions */
368extern char __sched_text_start[], __sched_text_end[];
369
1da177e4
LT
370/* Is this address in the __sched functions? */
371extern int in_sched_functions(unsigned long addr);
372
373#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 374extern signed long schedule_timeout(signed long timeout);
64ed93a2 375extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 376extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 377extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 378asmlinkage void schedule(void);
c5491ea7 379extern void schedule_preempt_disabled(void);
1da177e4 380
ab516013 381struct nsproxy;
acce292c 382struct user_namespace;
1da177e4 383
efc1a3b1
DH
384#ifdef CONFIG_MMU
385extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
386extern unsigned long
387arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
388 unsigned long, unsigned long);
389extern unsigned long
390arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
391 unsigned long len, unsigned long pgoff,
392 unsigned long flags);
efc1a3b1
DH
393#else
394static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
395#endif
1da177e4 396
d049f74f
KC
397#define SUID_DUMP_DISABLE 0 /* No setuid dumping */
398#define SUID_DUMP_USER 1 /* Dump as user of process */
399#define SUID_DUMP_ROOT 2 /* Dump as root */
400
6c5d5238 401/* mm flags */
f8af4da3 402
7288e118 403/* for SUID_DUMP_* above */
3cb4a0bb 404#define MMF_DUMPABLE_BITS 2
f8af4da3 405#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb 406
942be387
ON
407extern void set_dumpable(struct mm_struct *mm, int value);
408/*
409 * This returns the actual value of the suid_dumpable flag. For things
410 * that are using this for checking for privilege transitions, it must
411 * test against SUID_DUMP_USER rather than treating it as a boolean
412 * value.
413 */
414static inline int __get_dumpable(unsigned long mm_flags)
415{
416 return mm_flags & MMF_DUMPABLE_MASK;
417}
418
419static inline int get_dumpable(struct mm_struct *mm)
420{
421 return __get_dumpable(mm->flags);
422}
423
3cb4a0bb
KH
424/* coredump filter bits */
425#define MMF_DUMP_ANON_PRIVATE 2
426#define MMF_DUMP_ANON_SHARED 3
427#define MMF_DUMP_MAPPED_PRIVATE 4
428#define MMF_DUMP_MAPPED_SHARED 5
82df3973 429#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
430#define MMF_DUMP_HUGETLB_PRIVATE 7
431#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 432
3cb4a0bb 433#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 434#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
435#define MMF_DUMP_FILTER_MASK \
436 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
437#define MMF_DUMP_FILTER_DEFAULT \
e575f111 438 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
439 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
440
441#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
442# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
443#else
444# define MMF_DUMP_MASK_DEFAULT_ELF 0
445#endif
f8af4da3
HD
446 /* leave room for more dump flags */
447#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 448#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
bafb282d 449#define MMF_EXE_FILE_CHANGED 18 /* see prctl_set_mm_exe_file() */
f8af4da3 450
9f68f672
ON
451#define MMF_HAS_UPROBES 19 /* has uprobes */
452#define MMF_RECALC_UPROBES 20 /* MMF_HAS_UPROBES can be wrong */
f8ac4ec9 453
f8af4da3 454#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 455
1da177e4
LT
456struct sighand_struct {
457 atomic_t count;
458 struct k_sigaction action[_NSIG];
459 spinlock_t siglock;
b8fceee1 460 wait_queue_head_t signalfd_wqh;
1da177e4
LT
461};
462
0e464814 463struct pacct_struct {
f6ec29a4
KK
464 int ac_flag;
465 long ac_exitcode;
0e464814 466 unsigned long ac_mem;
77787bfb
KK
467 cputime_t ac_utime, ac_stime;
468 unsigned long ac_minflt, ac_majflt;
0e464814
KK
469};
470
42c4ab41
SG
471struct cpu_itimer {
472 cputime_t expires;
473 cputime_t incr;
8356b5f9
SG
474 u32 error;
475 u32 incr_error;
42c4ab41
SG
476};
477
d37f761d
FW
478/**
479 * struct cputime - snaphsot of system and user cputime
480 * @utime: time spent in user mode
481 * @stime: time spent in system mode
482 *
483 * Gathers a generic snapshot of user and system time.
484 */
485struct cputime {
486 cputime_t utime;
487 cputime_t stime;
488};
489
f06febc9
FM
490/**
491 * struct task_cputime - collected CPU time counts
492 * @utime: time spent in user mode, in &cputime_t units
493 * @stime: time spent in kernel mode, in &cputime_t units
494 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 495 *
d37f761d
FW
496 * This is an extension of struct cputime that includes the total runtime
497 * spent by the task from the scheduler point of view.
498 *
499 * As a result, this structure groups together three kinds of CPU time
500 * that are tracked for threads and thread groups. Most things considering
f06febc9
FM
501 * CPU time want to group these counts together and treat all three
502 * of them in parallel.
503 */
504struct task_cputime {
505 cputime_t utime;
506 cputime_t stime;
507 unsigned long long sum_exec_runtime;
508};
509/* Alternate field names when used to cache expirations. */
510#define prof_exp stime
511#define virt_exp utime
512#define sched_exp sum_exec_runtime
513
4cd4c1b4
PZ
514#define INIT_CPUTIME \
515 (struct task_cputime) { \
64861634
MS
516 .utime = 0, \
517 .stime = 0, \
4cd4c1b4
PZ
518 .sum_exec_runtime = 0, \
519 }
520
a233f112
PZ
521#ifdef CONFIG_PREEMPT_COUNT
522#define PREEMPT_DISABLED (1 + PREEMPT_ENABLED)
523#else
524#define PREEMPT_DISABLED PREEMPT_ENABLED
525#endif
526
c99e6efe
PZ
527/*
528 * Disable preemption until the scheduler is running.
529 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
530 *
531 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
532 * before the scheduler is active -- see should_resched().
c99e6efe 533 */
a233f112 534#define INIT_PREEMPT_COUNT (PREEMPT_DISABLED + PREEMPT_ACTIVE)
c99e6efe 535
f06febc9 536/**
4cd4c1b4
PZ
537 * struct thread_group_cputimer - thread group interval timer counts
538 * @cputime: thread group interval timers.
539 * @running: non-zero when there are timers running and
540 * @cputime receives updates.
541 * @lock: lock for fields in this struct.
f06febc9
FM
542 *
543 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 544 * used for thread group CPU timer calculations.
f06febc9 545 */
4cd4c1b4
PZ
546struct thread_group_cputimer {
547 struct task_cputime cputime;
548 int running;
ee30a7b2 549 raw_spinlock_t lock;
f06febc9 550};
f06febc9 551
4714d1d3 552#include <linux/rwsem.h>
5091faa4
MG
553struct autogroup;
554
1da177e4 555/*
e815f0a8 556 * NOTE! "signal_struct" does not have its own
1da177e4
LT
557 * locking, because a shared signal_struct always
558 * implies a shared sighand_struct, so locking
559 * sighand_struct is always a proper superset of
560 * the locking of signal_struct.
561 */
562struct signal_struct {
ea6d290c 563 atomic_t sigcnt;
1da177e4 564 atomic_t live;
b3ac022c 565 int nr_threads;
0c740d0a 566 struct list_head thread_head;
1da177e4
LT
567
568 wait_queue_head_t wait_chldexit; /* for wait4() */
569
570 /* current thread group signal load-balancing target: */
36c8b586 571 struct task_struct *curr_target;
1da177e4
LT
572
573 /* shared signal handling: */
574 struct sigpending shared_pending;
575
576 /* thread group exit support */
577 int group_exit_code;
578 /* overloaded:
579 * - notify group_exit_task when ->count is equal to notify_count
580 * - everyone except group_exit_task is stopped during signal delivery
581 * of fatal signals, group_exit_task processes the signal.
582 */
1da177e4 583 int notify_count;
07dd20e0 584 struct task_struct *group_exit_task;
1da177e4
LT
585
586 /* thread group stop support, overloads group_exit_code too */
587 int group_stop_count;
588 unsigned int flags; /* see SIGNAL_* flags below */
589
ebec18a6
LP
590 /*
591 * PR_SET_CHILD_SUBREAPER marks a process, like a service
592 * manager, to re-parent orphan (double-forking) child processes
593 * to this process instead of 'init'. The service manager is
594 * able to receive SIGCHLD signals and is able to investigate
595 * the process until it calls wait(). All children of this
596 * process will inherit a flag if they should look for a
597 * child_subreaper process at exit.
598 */
599 unsigned int is_child_subreaper:1;
600 unsigned int has_child_subreaper:1;
601
1da177e4 602 /* POSIX.1b Interval Timers */
5ed67f05
PE
603 int posix_timer_id;
604 struct list_head posix_timers;
1da177e4
LT
605
606 /* ITIMER_REAL timer for the process */
2ff678b8 607 struct hrtimer real_timer;
fea9d175 608 struct pid *leader_pid;
2ff678b8 609 ktime_t it_real_incr;
1da177e4 610
42c4ab41
SG
611 /*
612 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
613 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
614 * values are defined to 0 and 1 respectively
615 */
616 struct cpu_itimer it[2];
1da177e4 617
f06febc9 618 /*
4cd4c1b4
PZ
619 * Thread group totals for process CPU timers.
620 * See thread_group_cputimer(), et al, for details.
f06febc9 621 */
4cd4c1b4 622 struct thread_group_cputimer cputimer;
f06febc9
FM
623
624 /* Earliest-expiration cache. */
625 struct task_cputime cputime_expires;
626
627 struct list_head cpu_timers[3];
628
ab521dc0 629 struct pid *tty_old_pgrp;
1ec320af 630
1da177e4
LT
631 /* boolean value for session group leader */
632 int leader;
633
634 struct tty_struct *tty; /* NULL if no tty */
635
5091faa4
MG
636#ifdef CONFIG_SCHED_AUTOGROUP
637 struct autogroup *autogroup;
638#endif
1da177e4
LT
639 /*
640 * Cumulative resource counters for dead threads in the group,
641 * and for reaped dead child processes forked by this group.
642 * Live threads maintain their own counters and add to these
643 * in __exit_signal, except for the group leader.
644 */
32bd671d 645 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
646 cputime_t gtime;
647 cputime_t cgtime;
9fbc42ea 648#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 649 struct cputime prev_cputime;
0cf55e1e 650#endif
1da177e4
LT
651 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
652 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 653 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 654 unsigned long maxrss, cmaxrss;
940389b8 655 struct task_io_accounting ioac;
1da177e4 656
32bd671d
PZ
657 /*
658 * Cumulative ns of schedule CPU time fo dead threads in the
659 * group, not including a zombie group leader, (This only differs
660 * from jiffies_to_ns(utime + stime) if sched_clock uses something
661 * other than jiffies.)
662 */
663 unsigned long long sum_sched_runtime;
664
1da177e4
LT
665 /*
666 * We don't bother to synchronize most readers of this at all,
667 * because there is no reader checking a limit that actually needs
668 * to get both rlim_cur and rlim_max atomically, and either one
669 * alone is a single word that can safely be read normally.
670 * getrlimit/setrlimit use task_lock(current->group_leader) to
671 * protect this instead of the siglock, because they really
672 * have no need to disable irqs.
673 */
674 struct rlimit rlim[RLIM_NLIMITS];
675
0e464814
KK
676#ifdef CONFIG_BSD_PROCESS_ACCT
677 struct pacct_struct pacct; /* per-process accounting information */
678#endif
ad4ecbcb 679#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
680 struct taskstats *stats;
681#endif
522ed776
MT
682#ifdef CONFIG_AUDIT
683 unsigned audit_tty;
46e959ea 684 unsigned audit_tty_log_passwd;
522ed776
MT
685 struct tty_audit_buf *tty_audit_buf;
686#endif
4714d1d3
BB
687#ifdef CONFIG_CGROUPS
688 /*
77e4ef99
TH
689 * group_rwsem prevents new tasks from entering the threadgroup and
690 * member tasks from exiting,a more specifically, setting of
691 * PF_EXITING. fork and exit paths are protected with this rwsem
692 * using threadgroup_change_begin/end(). Users which require
693 * threadgroup to remain stable should use threadgroup_[un]lock()
694 * which also takes care of exec path. Currently, cgroup is the
695 * only user.
4714d1d3 696 */
257058ae 697 struct rw_semaphore group_rwsem;
4714d1d3 698#endif
28b83c51 699
e1e12d2f 700 oom_flags_t oom_flags;
a9c58b90
DR
701 short oom_score_adj; /* OOM kill score adjustment */
702 short oom_score_adj_min; /* OOM kill score adjustment min value.
703 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
704
705 struct mutex cred_guard_mutex; /* guard against foreign influences on
706 * credential calculations
707 * (notably. ptrace) */
1da177e4
LT
708};
709
710/*
711 * Bits in flags field of signal_struct.
712 */
713#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
ee77f075
ON
714#define SIGNAL_STOP_CONTINUED 0x00000002 /* SIGCONT since WCONTINUED reap */
715#define SIGNAL_GROUP_EXIT 0x00000004 /* group exit in progress */
403bad72 716#define SIGNAL_GROUP_COREDUMP 0x00000008 /* coredump in progress */
e4420551
ON
717/*
718 * Pending notifications to parent.
719 */
720#define SIGNAL_CLD_STOPPED 0x00000010
721#define SIGNAL_CLD_CONTINUED 0x00000020
722#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 723
fae5fa44
ON
724#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
725
ed5d2cac
ON
726/* If true, all threads except ->group_exit_task have pending SIGKILL */
727static inline int signal_group_exit(const struct signal_struct *sig)
728{
729 return (sig->flags & SIGNAL_GROUP_EXIT) ||
730 (sig->group_exit_task != NULL);
731}
732
1da177e4
LT
733/*
734 * Some day this will be a full-fledged user tracking system..
735 */
736struct user_struct {
737 atomic_t __count; /* reference count */
738 atomic_t processes; /* How many processes does this user have? */
739 atomic_t files; /* How many open files does this user have? */
740 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 741#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
742 atomic_t inotify_watches; /* How many inotify watches does this user have? */
743 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
744#endif
4afeff85
EP
745#ifdef CONFIG_FANOTIFY
746 atomic_t fanotify_listeners;
747#endif
7ef9964e 748#ifdef CONFIG_EPOLL
52bd19f7 749 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 750#endif
970a8645 751#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
752 /* protected by mq_lock */
753 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 754#endif
1da177e4
LT
755 unsigned long locked_shm; /* How many pages of mlocked shm ? */
756
757#ifdef CONFIG_KEYS
758 struct key *uid_keyring; /* UID specific keyring */
759 struct key *session_keyring; /* UID's default session keyring */
760#endif
761
762 /* Hash table maintenance information */
735de223 763 struct hlist_node uidhash_node;
7b44ab97 764 kuid_t uid;
24e377a8 765
cdd6c482 766#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
767 atomic_long_t locked_vm;
768#endif
1da177e4
LT
769};
770
eb41d946 771extern int uids_sysfs_init(void);
5cb350ba 772
7b44ab97 773extern struct user_struct *find_user(kuid_t);
1da177e4
LT
774
775extern struct user_struct root_user;
776#define INIT_USER (&root_user)
777
b6dff3ec 778
1da177e4
LT
779struct backing_dev_info;
780struct reclaim_state;
781
52f17b6c 782#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
783struct sched_info {
784 /* cumulative counters */
2d72376b 785 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 786 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
787
788 /* timestamps */
172ba844
BS
789 unsigned long long last_arrival,/* when we last ran on a cpu */
790 last_queued; /* when we were last queued to run */
1da177e4 791};
52f17b6c 792#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 793
ca74e92b
SN
794#ifdef CONFIG_TASK_DELAY_ACCT
795struct task_delay_info {
796 spinlock_t lock;
797 unsigned int flags; /* Private per-task flags */
798
799 /* For each stat XXX, add following, aligned appropriately
800 *
801 * struct timespec XXX_start, XXX_end;
802 * u64 XXX_delay;
803 * u32 XXX_count;
804 *
805 * Atomicity of updates to XXX_delay, XXX_count protected by
806 * single lock above (split into XXX_lock if contention is an issue).
807 */
0ff92245
SN
808
809 /*
810 * XXX_count is incremented on every XXX operation, the delay
811 * associated with the operation is added to XXX_delay.
812 * XXX_delay contains the accumulated delay time in nanoseconds.
813 */
814 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
815 u64 blkio_delay; /* wait for sync block io completion */
816 u64 swapin_delay; /* wait for swapin block io completion */
817 u32 blkio_count; /* total count of the number of sync block */
818 /* io operations performed */
819 u32 swapin_count; /* total count of the number of swapin block */
820 /* io operations performed */
873b4771
KK
821
822 struct timespec freepages_start, freepages_end;
823 u64 freepages_delay; /* wait for memory reclaim */
824 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 825};
52f17b6c
CS
826#endif /* CONFIG_TASK_DELAY_ACCT */
827
828static inline int sched_info_on(void)
829{
830#ifdef CONFIG_SCHEDSTATS
831 return 1;
832#elif defined(CONFIG_TASK_DELAY_ACCT)
833 extern int delayacct_on;
834 return delayacct_on;
835#else
836 return 0;
ca74e92b 837#endif
52f17b6c 838}
ca74e92b 839
d15bcfdb
IM
840enum cpu_idle_type {
841 CPU_IDLE,
842 CPU_NOT_IDLE,
843 CPU_NEWLY_IDLE,
844 CPU_MAX_IDLE_TYPES
1da177e4
LT
845};
846
1399fa78
NR
847/*
848 * Increase resolution of cpu_power calculations
849 */
850#define SCHED_POWER_SHIFT 10
851#define SCHED_POWER_SCALE (1L << SCHED_POWER_SHIFT)
1da177e4 852
1399fa78
NR
853/*
854 * sched-domains (multiprocessor balancing) declarations:
855 */
2dd73a4f 856#ifdef CONFIG_SMP
b5d978e0
PZ
857#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
858#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
859#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
860#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 861#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 862#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
b5d978e0 863#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
b5d978e0
PZ
864#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
865#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 866#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 867#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
e3589f6c 868#define SD_OVERLAP 0x2000 /* sched_domains of this level overlap */
3a7053b3 869#define SD_NUMA 0x4000 /* cross-node balancing */
5c45bf27 870
532cb4c4
MN
871extern int __weak arch_sd_sibiling_asym_packing(void);
872
1d3504fc
HS
873struct sched_domain_attr {
874 int relax_domain_level;
875};
876
877#define SD_ATTR_INIT (struct sched_domain_attr) { \
878 .relax_domain_level = -1, \
879}
880
60495e77
PZ
881extern int sched_domain_level_max;
882
5e6521ea
LZ
883struct sched_group;
884
1da177e4
LT
885struct sched_domain {
886 /* These fields must be setup */
887 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 888 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 889 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
890 unsigned long min_interval; /* Minimum balance interval ms */
891 unsigned long max_interval; /* Maximum balance interval ms */
892 unsigned int busy_factor; /* less balancing by factor if busy */
893 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 894 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
895 unsigned int busy_idx;
896 unsigned int idle_idx;
897 unsigned int newidle_idx;
898 unsigned int wake_idx;
147cbb4b 899 unsigned int forkexec_idx;
a52bfd73 900 unsigned int smt_gain;
25f55d9d
VG
901
902 int nohz_idle; /* NOHZ IDLE status */
1da177e4 903 int flags; /* See SD_* */
60495e77 904 int level;
1da177e4
LT
905
906 /* Runtime fields. */
907 unsigned long last_balance; /* init to jiffies. units in jiffies */
908 unsigned int balance_interval; /* initialise to 1. units in ms. */
909 unsigned int nr_balance_failed; /* initialise to 0 */
910
f48627e6 911 /* idle_balance() stats */
9bd721c5 912 u64 max_newidle_lb_cost;
f48627e6 913 unsigned long next_decay_max_lb_cost;
2398f2c6 914
1da177e4
LT
915#ifdef CONFIG_SCHEDSTATS
916 /* load_balance() stats */
480b9434
KC
917 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
918 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
919 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
920 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
921 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
922 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
923 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
924 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
925
926 /* Active load balancing */
480b9434
KC
927 unsigned int alb_count;
928 unsigned int alb_failed;
929 unsigned int alb_pushed;
1da177e4 930
68767a0a 931 /* SD_BALANCE_EXEC stats */
480b9434
KC
932 unsigned int sbe_count;
933 unsigned int sbe_balanced;
934 unsigned int sbe_pushed;
1da177e4 935
68767a0a 936 /* SD_BALANCE_FORK stats */
480b9434
KC
937 unsigned int sbf_count;
938 unsigned int sbf_balanced;
939 unsigned int sbf_pushed;
68767a0a 940
1da177e4 941 /* try_to_wake_up() stats */
480b9434
KC
942 unsigned int ttwu_wake_remote;
943 unsigned int ttwu_move_affine;
944 unsigned int ttwu_move_balance;
1da177e4 945#endif
a5d8c348
IM
946#ifdef CONFIG_SCHED_DEBUG
947 char *name;
948#endif
dce840a0
PZ
949 union {
950 void *private; /* used during construction */
951 struct rcu_head rcu; /* used during destruction */
952 };
6c99e9ad 953
669c55e9 954 unsigned int span_weight;
4200efd9
IM
955 /*
956 * Span of all CPUs in this domain.
957 *
958 * NOTE: this field is variable length. (Allocated dynamically
959 * by attaching extra space to the end of the structure,
960 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
961 */
962 unsigned long span[0];
1da177e4
LT
963};
964
758b2cdc
RR
965static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
966{
6c99e9ad 967 return to_cpumask(sd->span);
758b2cdc
RR
968}
969
acc3f5d7 970extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 971 struct sched_domain_attr *dattr_new);
029190c5 972
acc3f5d7
RR
973/* Allocate an array of sched domains, for partition_sched_domains(). */
974cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
975void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
976
39be3501
PZ
977bool cpus_share_cache(int this_cpu, int that_cpu);
978
1b427c15 979#else /* CONFIG_SMP */
1da177e4 980
1b427c15 981struct sched_domain_attr;
d02c7a8c 982
1b427c15 983static inline void
acc3f5d7 984partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
985 struct sched_domain_attr *dattr_new)
986{
d02c7a8c 987}
39be3501
PZ
988
989static inline bool cpus_share_cache(int this_cpu, int that_cpu)
990{
991 return true;
992}
993
1b427c15 994#endif /* !CONFIG_SMP */
1da177e4 995
47fe38fc 996
1da177e4 997struct io_context; /* See blkdev.h */
1da177e4 998
1da177e4 999
383f2835 1000#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1001extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1002#else
1003static inline void prefetch_stack(struct task_struct *t) { }
1004#endif
1da177e4
LT
1005
1006struct audit_context; /* See audit.c */
1007struct mempolicy;
b92ce558 1008struct pipe_inode_info;
4865ecf1 1009struct uts_namespace;
1da177e4 1010
20b8a59f 1011struct load_weight {
9dbdb155
PZ
1012 unsigned long weight;
1013 u32 inv_weight;
20b8a59f
IM
1014};
1015
9d85f21c
PT
1016struct sched_avg {
1017 /*
1018 * These sums represent an infinite geometric series and so are bound
239003ea 1019 * above by 1024/(1-y). Thus we only need a u32 to store them for all
9d85f21c
PT
1020 * choices of y < 1-2^(-32)*1024.
1021 */
1022 u32 runnable_avg_sum, runnable_avg_period;
1023 u64 last_runnable_update;
9ee474f5 1024 s64 decay_count;
2dac754e 1025 unsigned long load_avg_contrib;
9d85f21c
PT
1026};
1027
94c18227 1028#ifdef CONFIG_SCHEDSTATS
41acab88 1029struct sched_statistics {
20b8a59f 1030 u64 wait_start;
94c18227 1031 u64 wait_max;
6d082592
AV
1032 u64 wait_count;
1033 u64 wait_sum;
8f0dfc34
AV
1034 u64 iowait_count;
1035 u64 iowait_sum;
94c18227 1036
20b8a59f 1037 u64 sleep_start;
20b8a59f 1038 u64 sleep_max;
94c18227
IM
1039 s64 sum_sleep_runtime;
1040
1041 u64 block_start;
20b8a59f
IM
1042 u64 block_max;
1043 u64 exec_max;
eba1ed4b 1044 u64 slice_max;
cc367732 1045
cc367732
IM
1046 u64 nr_migrations_cold;
1047 u64 nr_failed_migrations_affine;
1048 u64 nr_failed_migrations_running;
1049 u64 nr_failed_migrations_hot;
1050 u64 nr_forced_migrations;
cc367732
IM
1051
1052 u64 nr_wakeups;
1053 u64 nr_wakeups_sync;
1054 u64 nr_wakeups_migrate;
1055 u64 nr_wakeups_local;
1056 u64 nr_wakeups_remote;
1057 u64 nr_wakeups_affine;
1058 u64 nr_wakeups_affine_attempts;
1059 u64 nr_wakeups_passive;
1060 u64 nr_wakeups_idle;
41acab88
LDM
1061};
1062#endif
1063
1064struct sched_entity {
1065 struct load_weight load; /* for load-balancing */
1066 struct rb_node run_node;
1067 struct list_head group_node;
1068 unsigned int on_rq;
1069
1070 u64 exec_start;
1071 u64 sum_exec_runtime;
1072 u64 vruntime;
1073 u64 prev_sum_exec_runtime;
1074
41acab88
LDM
1075 u64 nr_migrations;
1076
41acab88
LDM
1077#ifdef CONFIG_SCHEDSTATS
1078 struct sched_statistics statistics;
94c18227
IM
1079#endif
1080
20b8a59f 1081#ifdef CONFIG_FAIR_GROUP_SCHED
fed14d45 1082 int depth;
20b8a59f
IM
1083 struct sched_entity *parent;
1084 /* rq on which this entity is (to be) queued: */
1085 struct cfs_rq *cfs_rq;
1086 /* rq "owned" by this entity/group: */
1087 struct cfs_rq *my_q;
1088#endif
8bd75c77 1089
141965c7 1090#ifdef CONFIG_SMP
f4e26b12 1091 /* Per-entity load-tracking */
9d85f21c
PT
1092 struct sched_avg avg;
1093#endif
20b8a59f 1094};
70b97a7f 1095
fa717060
PZ
1096struct sched_rt_entity {
1097 struct list_head run_list;
78f2c7db 1098 unsigned long timeout;
57d2aa00 1099 unsigned long watchdog_stamp;
bee367ed 1100 unsigned int time_slice;
6f505b16 1101
58d6c2d7 1102 struct sched_rt_entity *back;
052f1dc7 1103#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1104 struct sched_rt_entity *parent;
1105 /* rq on which this entity is (to be) queued: */
1106 struct rt_rq *rt_rq;
1107 /* rq "owned" by this entity/group: */
1108 struct rt_rq *my_q;
1109#endif
fa717060
PZ
1110};
1111
aab03e05
DF
1112struct sched_dl_entity {
1113 struct rb_node rb_node;
1114
1115 /*
1116 * Original scheduling parameters. Copied here from sched_attr
1117 * during sched_setscheduler2(), they will remain the same until
1118 * the next sched_setscheduler2().
1119 */
1120 u64 dl_runtime; /* maximum runtime for each instance */
1121 u64 dl_deadline; /* relative deadline of each instance */
755378a4 1122 u64 dl_period; /* separation of two instances (period) */
332ac17e 1123 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
1124
1125 /*
1126 * Actual scheduling parameters. Initialized with the values above,
1127 * they are continously updated during task execution. Note that
1128 * the remaining runtime could be < 0 in case we are in overrun.
1129 */
1130 s64 runtime; /* remaining runtime for this instance */
1131 u64 deadline; /* absolute deadline for this instance */
1132 unsigned int flags; /* specifying the scheduler behaviour */
1133
1134 /*
1135 * Some bool flags:
1136 *
1137 * @dl_throttled tells if we exhausted the runtime. If so, the
1138 * task has to wait for a replenishment to be performed at the
1139 * next firing of dl_timer.
1140 *
1141 * @dl_new tells if a new instance arrived. If so we must
1142 * start executing it with full runtime and reset its absolute
1143 * deadline;
2d3d891d
DF
1144 *
1145 * @dl_boosted tells if we are boosted due to DI. If so we are
1146 * outside bandwidth enforcement mechanism (but only until we
1147 * exit the critical section).
aab03e05 1148 */
2d3d891d 1149 int dl_throttled, dl_new, dl_boosted;
aab03e05
DF
1150
1151 /*
1152 * Bandwidth enforcement timer. Each -deadline task has its
1153 * own bandwidth to be enforced, thus we need one timer per task.
1154 */
1155 struct hrtimer dl_timer;
1156};
8bd75c77 1157
86848966
PM
1158struct rcu_node;
1159
8dc85d54
PZ
1160enum perf_event_task_context {
1161 perf_invalid_context = -1,
1162 perf_hw_context = 0,
89a1e187 1163 perf_sw_context,
8dc85d54
PZ
1164 perf_nr_task_contexts,
1165};
1166
1da177e4
LT
1167struct task_struct {
1168 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1169 void *stack;
1da177e4 1170 atomic_t usage;
97dc32cd
WC
1171 unsigned int flags; /* per process flags, defined below */
1172 unsigned int ptrace;
1da177e4 1173
2dd73a4f 1174#ifdef CONFIG_SMP
fa14ff4a 1175 struct llist_node wake_entry;
3ca7a440 1176 int on_cpu;
62470419
MW
1177 struct task_struct *last_wakee;
1178 unsigned long wakee_flips;
1179 unsigned long wakee_flip_decay_ts;
ac66f547
PZ
1180
1181 int wake_cpu;
2dd73a4f 1182#endif
fd2f4419 1183 int on_rq;
50e645a8 1184
b29739f9 1185 int prio, static_prio, normal_prio;
c7aceaba 1186 unsigned int rt_priority;
5522d5d5 1187 const struct sched_class *sched_class;
20b8a59f 1188 struct sched_entity se;
fa717060 1189 struct sched_rt_entity rt;
8323f26c
PZ
1190#ifdef CONFIG_CGROUP_SCHED
1191 struct task_group *sched_task_group;
1192#endif
aab03e05 1193 struct sched_dl_entity dl;
1da177e4 1194
e107be36
AK
1195#ifdef CONFIG_PREEMPT_NOTIFIERS
1196 /* list of struct preempt_notifier: */
1197 struct hlist_head preempt_notifiers;
1198#endif
1199
6c5c9341 1200#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1201 unsigned int btrace_seq;
6c5c9341 1202#endif
1da177e4 1203
97dc32cd 1204 unsigned int policy;
29baa747 1205 int nr_cpus_allowed;
1da177e4 1206 cpumask_t cpus_allowed;
1da177e4 1207
a57eb940 1208#ifdef CONFIG_PREEMPT_RCU
e260be67 1209 int rcu_read_lock_nesting;
f41d911f 1210 char rcu_read_unlock_special;
f41d911f 1211 struct list_head rcu_node_entry;
a57eb940
PM
1212#endif /* #ifdef CONFIG_PREEMPT_RCU */
1213#ifdef CONFIG_TREE_PREEMPT_RCU
1214 struct rcu_node *rcu_blocked_node;
f41d911f 1215#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1216#ifdef CONFIG_RCU_BOOST
1217 struct rt_mutex *rcu_boost_mutex;
1218#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1219
52f17b6c 1220#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1221 struct sched_info sched_info;
1222#endif
1223
1224 struct list_head tasks;
806c09a7 1225#ifdef CONFIG_SMP
917b627d 1226 struct plist_node pushable_tasks;
1baca4ce 1227 struct rb_node pushable_dl_tasks;
806c09a7 1228#endif
1da177e4
LT
1229
1230 struct mm_struct *mm, *active_mm;
4471a675
JK
1231#ifdef CONFIG_COMPAT_BRK
1232 unsigned brk_randomized:1;
1233#endif
34e55232
KH
1234#if defined(SPLIT_RSS_COUNTING)
1235 struct task_rss_stat rss_stat;
1236#endif
1da177e4 1237/* task state */
97dc32cd 1238 int exit_state;
1da177e4
LT
1239 int exit_code, exit_signal;
1240 int pdeath_signal; /* The signal sent when the parent dies */
a8f072c1 1241 unsigned int jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
1242
1243 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 1244 unsigned int personality;
9b89f6ba 1245
f9ce1f1c
KT
1246 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1247 * execve */
8f0dfc34
AV
1248 unsigned in_iowait:1;
1249
259e5e6c
AL
1250 /* task may not gain privileges */
1251 unsigned no_new_privs:1;
ca94c442
LP
1252
1253 /* Revert to default priority/policy when forking */
1254 unsigned sched_reset_on_fork:1;
a8e4f2ea 1255 unsigned sched_contributes_to_load:1;
ca94c442 1256
1da177e4
LT
1257 pid_t pid;
1258 pid_t tgid;
0a425405 1259
1314562a 1260#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1261 /* Canary value for the -fstack-protector gcc feature */
1262 unsigned long stack_canary;
1314562a 1263#endif
4d1d61a6 1264 /*
1da177e4 1265 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 1266 * older sibling, respectively. (p->father can be replaced with
f470021a 1267 * p->real_parent->pid)
1da177e4 1268 */
abd63bc3
KC
1269 struct task_struct __rcu *real_parent; /* real parent process */
1270 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1271 /*
f470021a 1272 * children/sibling forms the list of my natural children
1da177e4
LT
1273 */
1274 struct list_head children; /* list of my children */
1275 struct list_head sibling; /* linkage in my parent's children list */
1276 struct task_struct *group_leader; /* threadgroup leader */
1277
f470021a
RM
1278 /*
1279 * ptraced is the list of tasks this task is using ptrace on.
1280 * This includes both natural children and PTRACE_ATTACH targets.
1281 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1282 */
1283 struct list_head ptraced;
1284 struct list_head ptrace_entry;
1285
1da177e4 1286 /* PID/PID hash table linkage. */
92476d7f 1287 struct pid_link pids[PIDTYPE_MAX];
47e65328 1288 struct list_head thread_group;
0c740d0a 1289 struct list_head thread_node;
1da177e4
LT
1290
1291 struct completion *vfork_done; /* for vfork() */
1292 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1293 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1294
c66f08be 1295 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1296 cputime_t gtime;
9fbc42ea 1297#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 1298 struct cputime prev_cputime;
6a61671b
FW
1299#endif
1300#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1301 seqlock_t vtime_seqlock;
1302 unsigned long long vtime_snap;
1303 enum {
1304 VTIME_SLEEPING = 0,
1305 VTIME_USER,
1306 VTIME_SYS,
1307 } vtime_snap_whence;
d99ca3b9 1308#endif
1da177e4 1309 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1310 struct timespec start_time; /* monotonic time */
1311 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1312/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1313 unsigned long min_flt, maj_flt;
1314
f06febc9 1315 struct task_cputime cputime_expires;
1da177e4
LT
1316 struct list_head cpu_timers[3];
1317
1318/* process credentials */
1b0ba1c9 1319 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1320 * credentials (COW) */
1b0ba1c9 1321 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1322 * credentials (COW) */
36772092
PBG
1323 char comm[TASK_COMM_LEN]; /* executable name excluding path
1324 - access with [gs]et_task_comm (which lock
1325 it with task_lock())
221af7f8 1326 - initialized normally by setup_new_exec */
1da177e4
LT
1327/* file system info */
1328 int link_count, total_link_count;
3d5b6fcc 1329#ifdef CONFIG_SYSVIPC
1da177e4
LT
1330/* ipc stuff */
1331 struct sysv_sem sysvsem;
3d5b6fcc 1332#endif
e162b39a 1333#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1334/* hung task detection */
82a1fcb9
IM
1335 unsigned long last_switch_count;
1336#endif
1da177e4
LT
1337/* CPU-specific state of this task */
1338 struct thread_struct thread;
1339/* filesystem information */
1340 struct fs_struct *fs;
1341/* open file information */
1342 struct files_struct *files;
1651e14e 1343/* namespaces */
ab516013 1344 struct nsproxy *nsproxy;
1da177e4
LT
1345/* signal handlers */
1346 struct signal_struct *signal;
1347 struct sighand_struct *sighand;
1348
1349 sigset_t blocked, real_blocked;
f3de272b 1350 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1351 struct sigpending pending;
1352
1353 unsigned long sas_ss_sp;
1354 size_t sas_ss_size;
1355 int (*notifier)(void *priv);
1356 void *notifier_data;
1357 sigset_t *notifier_mask;
67d12145 1358 struct callback_head *task_works;
e73f8959 1359
1da177e4 1360 struct audit_context *audit_context;
bfef93a5 1361#ifdef CONFIG_AUDITSYSCALL
e1760bd5 1362 kuid_t loginuid;
4746ec5b 1363 unsigned int sessionid;
bfef93a5 1364#endif
932ecebb 1365 struct seccomp seccomp;
1da177e4
LT
1366
1367/* Thread group tracking */
1368 u32 parent_exec_id;
1369 u32 self_exec_id;
58568d2a
MX
1370/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1371 * mempolicy */
1da177e4 1372 spinlock_t alloc_lock;
1da177e4 1373
b29739f9 1374 /* Protection of the PI data structures: */
1d615482 1375 raw_spinlock_t pi_lock;
b29739f9 1376
23f78d4a
IM
1377#ifdef CONFIG_RT_MUTEXES
1378 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
1379 struct rb_root pi_waiters;
1380 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
1381 /* Deadlock detection and priority inheritance handling */
1382 struct rt_mutex_waiter *pi_blocked_on;
2d3d891d
DF
1383 /* Top pi_waiters task */
1384 struct task_struct *pi_top_task;
23f78d4a
IM
1385#endif
1386
408894ee
IM
1387#ifdef CONFIG_DEBUG_MUTEXES
1388 /* mutex deadlock detection */
1389 struct mutex_waiter *blocked_on;
1390#endif
de30a2b3
IM
1391#ifdef CONFIG_TRACE_IRQFLAGS
1392 unsigned int irq_events;
de30a2b3 1393 unsigned long hardirq_enable_ip;
de30a2b3 1394 unsigned long hardirq_disable_ip;
fa1452e8 1395 unsigned int hardirq_enable_event;
de30a2b3 1396 unsigned int hardirq_disable_event;
fa1452e8
HS
1397 int hardirqs_enabled;
1398 int hardirq_context;
de30a2b3 1399 unsigned long softirq_disable_ip;
de30a2b3 1400 unsigned long softirq_enable_ip;
fa1452e8 1401 unsigned int softirq_disable_event;
de30a2b3 1402 unsigned int softirq_enable_event;
fa1452e8 1403 int softirqs_enabled;
de30a2b3
IM
1404 int softirq_context;
1405#endif
fbb9ce95 1406#ifdef CONFIG_LOCKDEP
bdb9441e 1407# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1408 u64 curr_chain_key;
1409 int lockdep_depth;
fbb9ce95 1410 unsigned int lockdep_recursion;
c7aceaba 1411 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1412 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1413#endif
408894ee 1414
1da177e4
LT
1415/* journalling filesystem info */
1416 void *journal_info;
1417
d89d8796 1418/* stacked block device info */
bddd87c7 1419 struct bio_list *bio_list;
d89d8796 1420
73c10101
JA
1421#ifdef CONFIG_BLOCK
1422/* stack plugging */
1423 struct blk_plug *plug;
1424#endif
1425
1da177e4
LT
1426/* VM state */
1427 struct reclaim_state *reclaim_state;
1428
1da177e4
LT
1429 struct backing_dev_info *backing_dev_info;
1430
1431 struct io_context *io_context;
1432
1433 unsigned long ptrace_message;
1434 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1435 struct task_io_accounting ioac;
8f0ab514 1436#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1437 u64 acct_rss_mem1; /* accumulated rss usage */
1438 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1439 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1440#endif
1441#ifdef CONFIG_CPUSETS
58568d2a 1442 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 1443 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 1444 int cpuset_mem_spread_rotor;
6adef3eb 1445 int cpuset_slab_spread_rotor;
1da177e4 1446#endif
ddbcc7e8 1447#ifdef CONFIG_CGROUPS
817929ec 1448 /* Control Group info protected by css_set_lock */
2c392b8c 1449 struct css_set __rcu *cgroups;
817929ec
PM
1450 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1451 struct list_head cg_list;
ddbcc7e8 1452#endif
42b2dd0a 1453#ifdef CONFIG_FUTEX
0771dfef 1454 struct robust_list_head __user *robust_list;
34f192c6
IM
1455#ifdef CONFIG_COMPAT
1456 struct compat_robust_list_head __user *compat_robust_list;
1457#endif
c87e2837
IM
1458 struct list_head pi_state_list;
1459 struct futex_pi_state *pi_state_cache;
c7aceaba 1460#endif
cdd6c482 1461#ifdef CONFIG_PERF_EVENTS
8dc85d54 1462 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1463 struct mutex perf_event_mutex;
1464 struct list_head perf_event_list;
a63eaf34 1465#endif
c7aceaba 1466#ifdef CONFIG_NUMA
58568d2a 1467 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1468 short il_next;
207205a2 1469 short pref_node_fork;
42b2dd0a 1470#endif
cbee9f88
PZ
1471#ifdef CONFIG_NUMA_BALANCING
1472 int numa_scan_seq;
cbee9f88 1473 unsigned int numa_scan_period;
598f0ec0 1474 unsigned int numa_scan_period_max;
de1c9ce6 1475 int numa_preferred_nid;
6b9a7460 1476 unsigned long numa_migrate_retry;
cbee9f88 1477 u64 node_stamp; /* migration stamp */
7e2703e6
RR
1478 u64 last_task_numa_placement;
1479 u64 last_sum_exec_runtime;
cbee9f88 1480 struct callback_head numa_work;
f809ca9a 1481
8c8a743c
PZ
1482 struct list_head numa_entry;
1483 struct numa_group *numa_group;
1484
745d6147
MG
1485 /*
1486 * Exponential decaying average of faults on a per-node basis.
1487 * Scheduling placement decisions are made based on the these counts.
1488 * The values remain static for the duration of a PTE scan
1489 */
ff1df896 1490 unsigned long *numa_faults_memory;
83e1d2cd 1491 unsigned long total_numa_faults;
745d6147
MG
1492
1493 /*
1494 * numa_faults_buffer records faults per node during the current
ff1df896
RR
1495 * scan window. When the scan completes, the counts in
1496 * numa_faults_memory decay and these values are copied.
745d6147 1497 */
ff1df896 1498 unsigned long *numa_faults_buffer_memory;
745d6147 1499
50ec8a40
RR
1500 /*
1501 * Track the nodes the process was running on when a NUMA hinting
1502 * fault was incurred.
1503 */
1504 unsigned long *numa_faults_cpu;
1505 unsigned long *numa_faults_buffer_cpu;
1506
04bb2f94
RR
1507 /*
1508 * numa_faults_locality tracks if faults recorded during the last
1509 * scan window were remote/local. The task scan period is adapted
1510 * based on the locality of the faults with different weights
1511 * depending on whether they were shared or private faults
1512 */
1513 unsigned long numa_faults_locality[2];
1514
b32e86b4 1515 unsigned long numa_pages_migrated;
cbee9f88
PZ
1516#endif /* CONFIG_NUMA_BALANCING */
1517
e56d0903 1518 struct rcu_head rcu;
b92ce558
JA
1519
1520 /*
1521 * cache last used pipe for splice
1522 */
1523 struct pipe_inode_info *splice_pipe;
5640f768
ED
1524
1525 struct page_frag task_frag;
1526
ca74e92b
SN
1527#ifdef CONFIG_TASK_DELAY_ACCT
1528 struct task_delay_info *delays;
f4f154fd
AM
1529#endif
1530#ifdef CONFIG_FAULT_INJECTION
1531 int make_it_fail;
ca74e92b 1532#endif
9d823e8f
WF
1533 /*
1534 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
1535 * balance_dirty_pages() for some dirty throttling pause
1536 */
1537 int nr_dirtied;
1538 int nr_dirtied_pause;
83712358 1539 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 1540
9745512c
AV
1541#ifdef CONFIG_LATENCYTOP
1542 int latency_record_count;
1543 struct latency_record latency_record[LT_SAVECOUNT];
1544#endif
6976675d
AV
1545 /*
1546 * time slack values; these are used to round up poll() and
1547 * select() etc timeout values. These are in nanoseconds.
1548 */
1549 unsigned long timer_slack_ns;
1550 unsigned long default_timer_slack_ns;
f8d570a4 1551
fb52607a 1552#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1553 /* Index of current stored address in ret_stack */
f201ae23
FW
1554 int curr_ret_stack;
1555 /* Stack of return addresses for return function tracing */
1556 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1557 /* time stamp for last schedule */
1558 unsigned long long ftrace_timestamp;
f201ae23
FW
1559 /*
1560 * Number of functions that haven't been traced
1561 * because of depth overrun.
1562 */
1563 atomic_t trace_overrun;
380c4b14
FW
1564 /* Pause for the tracing */
1565 atomic_t tracing_graph_pause;
f201ae23 1566#endif
ea4e2bc4
SR
1567#ifdef CONFIG_TRACING
1568 /* state flags for use by tracers */
1569 unsigned long trace;
b1cff0ad 1570 /* bitmask and counter of trace recursion */
261842b7
SR
1571 unsigned long trace_recursion;
1572#endif /* CONFIG_TRACING */
c255a458 1573#ifdef CONFIG_MEMCG /* memcg uses this to do batch job */
569b846d
KH
1574 struct memcg_batch_info {
1575 int do_batch; /* incremented when batch uncharge started */
1576 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1577 unsigned long nr_pages; /* uncharged usage */
1578 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d 1579 } memcg_batch;
0e9d92f2 1580 unsigned int memcg_kmem_skip_account;
519e5247 1581 struct memcg_oom_info {
49426420
JW
1582 struct mem_cgroup *memcg;
1583 gfp_t gfp_mask;
1584 int order;
519e5247
JW
1585 unsigned int may_oom:1;
1586 } memcg_oom;
569b846d 1587#endif
0326f5a9
SD
1588#ifdef CONFIG_UPROBES
1589 struct uprobe_task *utask;
0326f5a9 1590#endif
cafe5635
KO
1591#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
1592 unsigned int sequential_io;
1593 unsigned int sequential_io_avg;
1594#endif
1da177e4
LT
1595};
1596
76e6eee0 1597/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1598#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1599
6688cc05
PZ
1600#define TNF_MIGRATED 0x01
1601#define TNF_NO_GROUP 0x02
dabe1d99 1602#define TNF_SHARED 0x04
04bb2f94 1603#define TNF_FAULT_LOCAL 0x08
6688cc05 1604
cbee9f88 1605#ifdef CONFIG_NUMA_BALANCING
6688cc05 1606extern void task_numa_fault(int last_node, int node, int pages, int flags);
e29cf08b 1607extern pid_t task_numa_group_id(struct task_struct *p);
1a687c2e 1608extern void set_numabalancing_state(bool enabled);
82727018 1609extern void task_numa_free(struct task_struct *p);
10f39042
RR
1610extern bool should_numa_migrate_memory(struct task_struct *p, struct page *page,
1611 int src_nid, int dst_cpu);
cbee9f88 1612#else
ac8e895b 1613static inline void task_numa_fault(int last_node, int node, int pages,
6688cc05 1614 int flags)
cbee9f88
PZ
1615{
1616}
e29cf08b
MG
1617static inline pid_t task_numa_group_id(struct task_struct *p)
1618{
1619 return 0;
1620}
1a687c2e
MG
1621static inline void set_numabalancing_state(bool enabled)
1622{
1623}
82727018
RR
1624static inline void task_numa_free(struct task_struct *p)
1625{
1626}
10f39042
RR
1627static inline bool should_numa_migrate_memory(struct task_struct *p,
1628 struct page *page, int src_nid, int dst_cpu)
1629{
1630 return true;
1631}
cbee9f88
PZ
1632#endif
1633
e868171a 1634static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1635{
1636 return task->pids[PIDTYPE_PID].pid;
1637}
1638
e868171a 1639static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1640{
1641 return task->group_leader->pids[PIDTYPE_PID].pid;
1642}
1643
6dda81f4
ON
1644/*
1645 * Without tasklist or rcu lock it is not safe to dereference
1646 * the result of task_pgrp/task_session even if task == current,
1647 * we can race with another thread doing sys_setsid/sys_setpgid.
1648 */
e868171a 1649static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1650{
1651 return task->group_leader->pids[PIDTYPE_PGID].pid;
1652}
1653
e868171a 1654static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1655{
1656 return task->group_leader->pids[PIDTYPE_SID].pid;
1657}
1658
7af57294
PE
1659struct pid_namespace;
1660
1661/*
1662 * the helpers to get the task's different pids as they are seen
1663 * from various namespaces
1664 *
1665 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1666 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1667 * current.
7af57294
PE
1668 * task_xid_nr_ns() : id seen from the ns specified;
1669 *
1670 * set_task_vxid() : assigns a virtual id to a task;
1671 *
7af57294
PE
1672 * see also pid_nr() etc in include/linux/pid.h
1673 */
52ee2dfd
ON
1674pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1675 struct pid_namespace *ns);
7af57294 1676
e868171a 1677static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1678{
1679 return tsk->pid;
1680}
1681
52ee2dfd
ON
1682static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1683 struct pid_namespace *ns)
1684{
1685 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1686}
7af57294
PE
1687
1688static inline pid_t task_pid_vnr(struct task_struct *tsk)
1689{
52ee2dfd 1690 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1691}
1692
1693
e868171a 1694static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1695{
1696 return tsk->tgid;
1697}
1698
2f2a3a46 1699pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1700
1701static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1702{
1703 return pid_vnr(task_tgid(tsk));
1704}
1705
1706
52ee2dfd
ON
1707static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1708 struct pid_namespace *ns)
7af57294 1709{
52ee2dfd 1710 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1711}
1712
7af57294
PE
1713static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1714{
52ee2dfd 1715 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1716}
1717
1718
52ee2dfd
ON
1719static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1720 struct pid_namespace *ns)
7af57294 1721{
52ee2dfd 1722 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1723}
1724
7af57294
PE
1725static inline pid_t task_session_vnr(struct task_struct *tsk)
1726{
52ee2dfd 1727 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1728}
1729
1b0f7ffd
ON
1730/* obsolete, do not use */
1731static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1732{
1733 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1734}
7af57294 1735
1da177e4
LT
1736/**
1737 * pid_alive - check that a task structure is not stale
1738 * @p: Task structure to be checked.
1739 *
1740 * Test if a process is not yet dead (at most zombie state)
1741 * If pid_alive fails, then pointers within the task structure
1742 * can be stale and must not be dereferenced.
e69f6186
YB
1743 *
1744 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1745 */
e868171a 1746static inline int pid_alive(struct task_struct *p)
1da177e4 1747{
92476d7f 1748 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1749}
1750
f400e198 1751/**
b460cbc5 1752 * is_global_init - check if a task structure is init
3260259f
HK
1753 * @tsk: Task structure to be checked.
1754 *
1755 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1756 *
1757 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1758 */
e868171a 1759static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1760{
1761 return tsk->pid == 1;
1762}
b460cbc5 1763
9ec52099
CLG
1764extern struct pid *cad_pid;
1765
1da177e4 1766extern void free_task(struct task_struct *tsk);
1da177e4 1767#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1768
158d9ebd 1769extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1770
1771static inline void put_task_struct(struct task_struct *t)
1772{
1773 if (atomic_dec_and_test(&t->usage))
8c7904a0 1774 __put_task_struct(t);
e56d0903 1775}
1da177e4 1776
6a61671b
FW
1777#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1778extern void task_cputime(struct task_struct *t,
1779 cputime_t *utime, cputime_t *stime);
1780extern void task_cputime_scaled(struct task_struct *t,
1781 cputime_t *utimescaled, cputime_t *stimescaled);
1782extern cputime_t task_gtime(struct task_struct *t);
1783#else
6fac4829
FW
1784static inline void task_cputime(struct task_struct *t,
1785 cputime_t *utime, cputime_t *stime)
1786{
1787 if (utime)
1788 *utime = t->utime;
1789 if (stime)
1790 *stime = t->stime;
1791}
1792
1793static inline void task_cputime_scaled(struct task_struct *t,
1794 cputime_t *utimescaled,
1795 cputime_t *stimescaled)
1796{
1797 if (utimescaled)
1798 *utimescaled = t->utimescaled;
1799 if (stimescaled)
1800 *stimescaled = t->stimescaled;
1801}
6a61671b
FW
1802
1803static inline cputime_t task_gtime(struct task_struct *t)
1804{
1805 return t->gtime;
1806}
1807#endif
e80d0a1a
FW
1808extern void task_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
1809extern void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1810
1da177e4
LT
1811/*
1812 * Per process flags
1813 */
1da177e4 1814#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1815#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1816#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1817#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1818#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1819#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1820#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1821#define PF_DUMPCORE 0x00000200 /* dumped core */
1822#define PF_SIGNALED 0x00000400 /* killed by a signal */
1823#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1824#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1825#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1826#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1827#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1828#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1829#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1830#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1831#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1832#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1833#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1834#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1835#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1836#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1837#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
14a40ffc 1838#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1839#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1840#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1841#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1842#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1843#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1844
1845/*
1846 * Only the _current_ task can read/write to tsk->flags, but other
1847 * tasks can access tsk->flags in readonly mode for example
1848 * with tsk_used_math (like during threaded core dumping).
1849 * There is however an exception to this rule during ptrace
1850 * or during fork: the ptracer task is allowed to write to the
1851 * child->flags of its traced child (same goes for fork, the parent
1852 * can write to the child->flags), because we're guaranteed the
1853 * child is not running and in turn not changing child->flags
1854 * at the same time the parent does it.
1855 */
1856#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1857#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1858#define clear_used_math() clear_stopped_child_used_math(current)
1859#define set_used_math() set_stopped_child_used_math(current)
1860#define conditional_stopped_child_used_math(condition, child) \
1861 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1862#define conditional_used_math(condition) \
1863 conditional_stopped_child_used_math(condition, current)
1864#define copy_to_stopped_child_used_math(child) \
1865 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1866/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1867#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1868#define used_math() tsk_used_math(current)
1869
21caf2fc
ML
1870/* __GFP_IO isn't allowed if PF_MEMALLOC_NOIO is set in current->flags */
1871static inline gfp_t memalloc_noio_flags(gfp_t flags)
1872{
1873 if (unlikely(current->flags & PF_MEMALLOC_NOIO))
1874 flags &= ~__GFP_IO;
1875 return flags;
1876}
1877
1878static inline unsigned int memalloc_noio_save(void)
1879{
1880 unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
1881 current->flags |= PF_MEMALLOC_NOIO;
1882 return flags;
1883}
1884
1885static inline void memalloc_noio_restore(unsigned int flags)
1886{
1887 current->flags = (current->flags & ~PF_MEMALLOC_NOIO) | flags;
1888}
1889
e5c1902e 1890/*
a8f072c1 1891 * task->jobctl flags
e5c1902e 1892 */
a8f072c1 1893#define JOBCTL_STOP_SIGMASK 0xffff /* signr of the last group stop */
e5c1902e 1894
a8f072c1
TH
1895#define JOBCTL_STOP_DEQUEUED_BIT 16 /* stop signal dequeued */
1896#define JOBCTL_STOP_PENDING_BIT 17 /* task should stop for group stop */
1897#define JOBCTL_STOP_CONSUME_BIT 18 /* consume group stop count */
73ddff2b 1898#define JOBCTL_TRAP_STOP_BIT 19 /* trap for STOP */
fb1d910c 1899#define JOBCTL_TRAP_NOTIFY_BIT 20 /* trap for NOTIFY */
a8f072c1 1900#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */
544b2c91 1901#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */
a8f072c1
TH
1902
1903#define JOBCTL_STOP_DEQUEUED (1 << JOBCTL_STOP_DEQUEUED_BIT)
1904#define JOBCTL_STOP_PENDING (1 << JOBCTL_STOP_PENDING_BIT)
1905#define JOBCTL_STOP_CONSUME (1 << JOBCTL_STOP_CONSUME_BIT)
73ddff2b 1906#define JOBCTL_TRAP_STOP (1 << JOBCTL_TRAP_STOP_BIT)
fb1d910c 1907#define JOBCTL_TRAP_NOTIFY (1 << JOBCTL_TRAP_NOTIFY_BIT)
a8f072c1 1908#define JOBCTL_TRAPPING (1 << JOBCTL_TRAPPING_BIT)
544b2c91 1909#define JOBCTL_LISTENING (1 << JOBCTL_LISTENING_BIT)
a8f072c1 1910
fb1d910c 1911#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
73ddff2b 1912#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK)
3759a0d9 1913
7dd3db54
TH
1914extern bool task_set_jobctl_pending(struct task_struct *task,
1915 unsigned int mask);
73ddff2b 1916extern void task_clear_jobctl_trapping(struct task_struct *task);
3759a0d9
TH
1917extern void task_clear_jobctl_pending(struct task_struct *task,
1918 unsigned int mask);
39efa3ef 1919
a57eb940 1920#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1921
1922#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1aa03f11 1923#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1924
1925static inline void rcu_copy_process(struct task_struct *p)
1926{
1927 p->rcu_read_lock_nesting = 0;
1928 p->rcu_read_unlock_special = 0;
a57eb940 1929#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1930 p->rcu_blocked_node = NULL;
24278d14
PM
1931#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1932#ifdef CONFIG_RCU_BOOST
1933 p->rcu_boost_mutex = NULL;
1934#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1935 INIT_LIST_HEAD(&p->rcu_node_entry);
1936}
1937
f41d911f
PM
1938#else
1939
1940static inline void rcu_copy_process(struct task_struct *p)
1941{
1942}
1943
1944#endif
1945
907aed48
MG
1946static inline void tsk_restore_flags(struct task_struct *task,
1947 unsigned long orig_flags, unsigned long flags)
1948{
1949 task->flags &= ~flags;
1950 task->flags |= orig_flags & flags;
1951}
1952
1da177e4 1953#ifdef CONFIG_SMP
1e1b6c51
KM
1954extern void do_set_cpus_allowed(struct task_struct *p,
1955 const struct cpumask *new_mask);
1956
cd8ba7cd 1957extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1958 const struct cpumask *new_mask);
1da177e4 1959#else
1e1b6c51
KM
1960static inline void do_set_cpus_allowed(struct task_struct *p,
1961 const struct cpumask *new_mask)
1962{
1963}
cd8ba7cd 1964static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1965 const struct cpumask *new_mask)
1da177e4 1966{
96f874e2 1967 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1968 return -EINVAL;
1969 return 0;
1970}
1971#endif
e0ad9556 1972
3451d024 1973#ifdef CONFIG_NO_HZ_COMMON
5167e8d5
PZ
1974void calc_load_enter_idle(void);
1975void calc_load_exit_idle(void);
1976#else
1977static inline void calc_load_enter_idle(void) { }
1978static inline void calc_load_exit_idle(void) { }
3451d024 1979#endif /* CONFIG_NO_HZ_COMMON */
5167e8d5 1980
e0ad9556 1981#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1982static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1983{
1984 return set_cpus_allowed_ptr(p, &new_mask);
1985}
e0ad9556 1986#endif
1da177e4 1987
b342501c 1988/*
c676329a
PZ
1989 * Do not use outside of architecture code which knows its limitations.
1990 *
1991 * sched_clock() has no promise of monotonicity or bounded drift between
1992 * CPUs, use (which you should not) requires disabling IRQs.
1993 *
1994 * Please use one of the three interfaces below.
b342501c 1995 */
1bbfa6f2 1996extern unsigned long long notrace sched_clock(void);
c676329a 1997/*
489a71b0 1998 * See the comment in kernel/sched/clock.c
c676329a
PZ
1999 */
2000extern u64 cpu_clock(int cpu);
2001extern u64 local_clock(void);
2002extern u64 sched_clock_cpu(int cpu);
2003
e436d800 2004
c1955a3d 2005extern void sched_clock_init(void);
3e51f33f 2006
c1955a3d 2007#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
2008static inline void sched_clock_tick(void)
2009{
2010}
2011
2012static inline void sched_clock_idle_sleep_event(void)
2013{
2014}
2015
2016static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
2017{
2018}
2019#else
c676329a
PZ
2020/*
2021 * Architectures can set this to 1 if they have specified
2022 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
2023 * but then during bootup it turns out that sched_clock()
2024 * is reliable after all:
2025 */
35af99e6
PZ
2026extern int sched_clock_stable(void);
2027extern void set_sched_clock_stable(void);
2028extern void clear_sched_clock_stable(void);
c676329a 2029
3e51f33f
PZ
2030extern void sched_clock_tick(void);
2031extern void sched_clock_idle_sleep_event(void);
2032extern void sched_clock_idle_wakeup_event(u64 delta_ns);
2033#endif
2034
b52bfee4
VP
2035#ifdef CONFIG_IRQ_TIME_ACCOUNTING
2036/*
2037 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
2038 * The reason for this explicit opt-in is not to have perf penalty with
2039 * slow sched_clocks.
2040 */
2041extern void enable_sched_clock_irqtime(void);
2042extern void disable_sched_clock_irqtime(void);
2043#else
2044static inline void enable_sched_clock_irqtime(void) {}
2045static inline void disable_sched_clock_irqtime(void) {}
2046#endif
2047
36c8b586 2048extern unsigned long long
41b86e9c 2049task_sched_runtime(struct task_struct *task);
1da177e4
LT
2050
2051/* sched_exec is called by processes performing an exec */
2052#ifdef CONFIG_SMP
2053extern void sched_exec(void);
2054#else
2055#define sched_exec() {}
2056#endif
2057
2aa44d05
IM
2058extern void sched_clock_idle_sleep_event(void);
2059extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 2060
1da177e4
LT
2061#ifdef CONFIG_HOTPLUG_CPU
2062extern void idle_task_exit(void);
2063#else
2064static inline void idle_task_exit(void) {}
2065#endif
2066
3451d024 2067#if defined(CONFIG_NO_HZ_COMMON) && defined(CONFIG_SMP)
1c20091e 2068extern void wake_up_nohz_cpu(int cpu);
06d8308c 2069#else
1c20091e 2070static inline void wake_up_nohz_cpu(int cpu) { }
06d8308c
TG
2071#endif
2072
ce831b38
FW
2073#ifdef CONFIG_NO_HZ_FULL
2074extern bool sched_can_stop_tick(void);
265f22a9 2075extern u64 scheduler_tick_max_deferment(void);
ce831b38
FW
2076#else
2077static inline bool sched_can_stop_tick(void) { return false; }
06d8308c
TG
2078#endif
2079
5091faa4 2080#ifdef CONFIG_SCHED_AUTOGROUP
5091faa4
MG
2081extern void sched_autogroup_create_attach(struct task_struct *p);
2082extern void sched_autogroup_detach(struct task_struct *p);
2083extern void sched_autogroup_fork(struct signal_struct *sig);
2084extern void sched_autogroup_exit(struct signal_struct *sig);
2085#ifdef CONFIG_PROC_FS
2086extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
2e5b5b3a 2087extern int proc_sched_autogroup_set_nice(struct task_struct *p, int nice);
5091faa4
MG
2088#endif
2089#else
2090static inline void sched_autogroup_create_attach(struct task_struct *p) { }
2091static inline void sched_autogroup_detach(struct task_struct *p) { }
2092static inline void sched_autogroup_fork(struct signal_struct *sig) { }
2093static inline void sched_autogroup_exit(struct signal_struct *sig) { }
2094#endif
2095
d95f4122 2096extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
2097extern void set_user_nice(struct task_struct *p, long nice);
2098extern int task_prio(const struct task_struct *p);
d0ea0268
DY
2099/**
2100 * task_nice - return the nice value of a given task.
2101 * @p: the task in question.
2102 *
2103 * Return: The nice value [ -20 ... 0 ... 19 ].
2104 */
2105static inline int task_nice(const struct task_struct *p)
2106{
2107 return PRIO_TO_NICE((p)->static_prio);
2108}
36c8b586
IM
2109extern int can_nice(const struct task_struct *p, const int nice);
2110extern int task_curr(const struct task_struct *p);
1da177e4 2111extern int idle_cpu(int cpu);
fe7de49f
KM
2112extern int sched_setscheduler(struct task_struct *, int,
2113 const struct sched_param *);
961ccddd 2114extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 2115 const struct sched_param *);
d50dde5a
DF
2116extern int sched_setattr(struct task_struct *,
2117 const struct sched_attr *);
36c8b586 2118extern struct task_struct *idle_task(int cpu);
c4f30608
PM
2119/**
2120 * is_idle_task - is the specified task an idle task?
fa757281 2121 * @p: the task in question.
e69f6186
YB
2122 *
2123 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 2124 */
7061ca3b 2125static inline bool is_idle_task(const struct task_struct *p)
c4f30608
PM
2126{
2127 return p->pid == 0;
2128}
36c8b586
IM
2129extern struct task_struct *curr_task(int cpu);
2130extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
2131
2132void yield(void);
2133
2134/*
2135 * The default (Linux) execution domain.
2136 */
2137extern struct exec_domain default_exec_domain;
2138
2139union thread_union {
2140 struct thread_info thread_info;
2141 unsigned long stack[THREAD_SIZE/sizeof(long)];
2142};
2143
2144#ifndef __HAVE_ARCH_KSTACK_END
2145static inline int kstack_end(void *addr)
2146{
2147 /* Reliable end of stack detection:
2148 * Some APM bios versions misalign the stack
2149 */
2150 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2151}
2152#endif
2153
2154extern union thread_union init_thread_union;
2155extern struct task_struct init_task;
2156
2157extern struct mm_struct init_mm;
2158
198fe21b
PE
2159extern struct pid_namespace init_pid_ns;
2160
2161/*
2162 * find a task by one of its numerical ids
2163 *
198fe21b
PE
2164 * find_task_by_pid_ns():
2165 * finds a task by its pid in the specified namespace
228ebcbe
PE
2166 * find_task_by_vpid():
2167 * finds a task by its virtual pid
198fe21b 2168 *
e49859e7 2169 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2170 */
2171
228ebcbe
PE
2172extern struct task_struct *find_task_by_vpid(pid_t nr);
2173extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2174 struct pid_namespace *ns);
198fe21b 2175
1da177e4 2176/* per-UID process charging. */
7b44ab97 2177extern struct user_struct * alloc_uid(kuid_t);
1da177e4
LT
2178static inline struct user_struct *get_uid(struct user_struct *u)
2179{
2180 atomic_inc(&u->__count);
2181 return u;
2182}
2183extern void free_uid(struct user_struct *);
1da177e4
LT
2184
2185#include <asm/current.h>
2186
f0af911a 2187extern void xtime_update(unsigned long ticks);
1da177e4 2188
b3c97528
HH
2189extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2190extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 2191extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
2192#ifdef CONFIG_SMP
2193 extern void kick_process(struct task_struct *tsk);
2194#else
2195 static inline void kick_process(struct task_struct *tsk) { }
2196#endif
aab03e05 2197extern int sched_fork(unsigned long clone_flags, struct task_struct *p);
ad46c2c4 2198extern void sched_dead(struct task_struct *p);
1da177e4 2199
1da177e4
LT
2200extern void proc_caches_init(void);
2201extern void flush_signals(struct task_struct *);
3bcac026 2202extern void __flush_signals(struct task_struct *);
10ab825b 2203extern void ignore_signals(struct task_struct *);
1da177e4
LT
2204extern void flush_signal_handlers(struct task_struct *, int force_default);
2205extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2206
2207static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2208{
2209 unsigned long flags;
2210 int ret;
2211
2212 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2213 ret = dequeue_signal(tsk, mask, info);
2214 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2215
2216 return ret;
53c8f9f1 2217}
1da177e4
LT
2218
2219extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2220 sigset_t *mask);
2221extern void unblock_all_signals(void);
2222extern void release_task(struct task_struct * p);
2223extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2224extern int force_sigsegv(int, struct task_struct *);
2225extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2226extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2227extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
d178bc3a
SH
2228extern int kill_pid_info_as_cred(int, struct siginfo *, struct pid *,
2229 const struct cred *, u32);
c4b92fc1
EB
2230extern int kill_pgrp(struct pid *pid, int sig, int priv);
2231extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2232extern int kill_proc_info(int, struct siginfo *, pid_t);
86773473 2233extern __must_check bool do_notify_parent(struct task_struct *, int);
a7f0765e 2234extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2235extern void force_sig(int, struct task_struct *);
1da177e4 2236extern int send_sig(int, struct task_struct *, int);
09faef11 2237extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2238extern struct sigqueue *sigqueue_alloc(void);
2239extern void sigqueue_free(struct sigqueue *);
ac5c2153 2240extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2241extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4 2242
51a7b448
AV
2243static inline void restore_saved_sigmask(void)
2244{
2245 if (test_and_clear_restore_sigmask())
77097ae5 2246 __set_current_blocked(&current->saved_sigmask);
51a7b448
AV
2247}
2248
b7f9a11a
AV
2249static inline sigset_t *sigmask_to_save(void)
2250{
2251 sigset_t *res = &current->blocked;
2252 if (unlikely(test_restore_sigmask()))
2253 res = &current->saved_sigmask;
2254 return res;
2255}
2256
9ec52099
CLG
2257static inline int kill_cad_pid(int sig, int priv)
2258{
2259 return kill_pid(cad_pid, sig, priv);
2260}
2261
1da177e4
LT
2262/* These can be the second arg to send_sig_info/send_group_sig_info. */
2263#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2264#define SEND_SIG_PRIV ((struct siginfo *) 1)
2265#define SEND_SIG_FORCED ((struct siginfo *) 2)
2266
2a855dd0
SAS
2267/*
2268 * True if we are on the alternate signal stack.
2269 */
1da177e4
LT
2270static inline int on_sig_stack(unsigned long sp)
2271{
2a855dd0
SAS
2272#ifdef CONFIG_STACK_GROWSUP
2273 return sp >= current->sas_ss_sp &&
2274 sp - current->sas_ss_sp < current->sas_ss_size;
2275#else
2276 return sp > current->sas_ss_sp &&
2277 sp - current->sas_ss_sp <= current->sas_ss_size;
2278#endif
1da177e4
LT
2279}
2280
2281static inline int sas_ss_flags(unsigned long sp)
2282{
2283 return (current->sas_ss_size == 0 ? SS_DISABLE
2284 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2285}
2286
5a1b98d3
AV
2287static inline unsigned long sigsp(unsigned long sp, struct ksignal *ksig)
2288{
2289 if (unlikely((ksig->ka.sa.sa_flags & SA_ONSTACK)) && ! sas_ss_flags(sp))
2290#ifdef CONFIG_STACK_GROWSUP
2291 return current->sas_ss_sp;
2292#else
2293 return current->sas_ss_sp + current->sas_ss_size;
2294#endif
2295 return sp;
2296}
2297
1da177e4
LT
2298/*
2299 * Routines for handling mm_structs
2300 */
2301extern struct mm_struct * mm_alloc(void);
2302
2303/* mmdrop drops the mm and the page tables */
b3c97528 2304extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2305static inline void mmdrop(struct mm_struct * mm)
2306{
6fb43d7b 2307 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2308 __mmdrop(mm);
2309}
2310
2311/* mmput gets rid of the mappings and all user-space */
2312extern void mmput(struct mm_struct *);
2313/* Grab a reference to a task's mm, if it is not already going away */
2314extern struct mm_struct *get_task_mm(struct task_struct *task);
8cdb878d
CY
2315/*
2316 * Grab a reference to a task's mm, if it is not already going away
2317 * and ptrace_may_access with the mode parameter passed to it
2318 * succeeds.
2319 */
2320extern struct mm_struct *mm_access(struct task_struct *task, unsigned int mode);
1da177e4
LT
2321/* Remove the current tasks stale references to the old mm_struct */
2322extern void mm_release(struct task_struct *, struct mm_struct *);
2323
6f2c55b8 2324extern int copy_thread(unsigned long, unsigned long, unsigned long,
afa86fc4 2325 struct task_struct *);
1da177e4
LT
2326extern void flush_thread(void);
2327extern void exit_thread(void);
2328
1da177e4 2329extern void exit_files(struct task_struct *);
a7e5328a 2330extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2331
1da177e4 2332extern void exit_itimers(struct signal_struct *);
cbaffba1 2333extern void flush_itimer_signals(void);
1da177e4 2334
9402c95f 2335extern void do_group_exit(int);
1da177e4 2336
1da177e4
LT
2337extern int allow_signal(int);
2338extern int disallow_signal(int);
1da177e4 2339
c4ad8f98 2340extern int do_execve(struct filename *,
d7627467 2341 const char __user * const __user *,
da3d4c5f 2342 const char __user * const __user *);
e80d6661 2343extern long do_fork(unsigned long, unsigned long, unsigned long, int __user *, int __user *);
36c8b586 2344struct task_struct *fork_idle(int);
2aa3a7f8 2345extern pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
1da177e4
LT
2346
2347extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2348extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2349
2350#ifdef CONFIG_SMP
317f3941 2351void scheduler_ipi(void);
85ba2d86 2352extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2353#else
184748cc 2354static inline void scheduler_ipi(void) { }
85ba2d86
RM
2355static inline unsigned long wait_task_inactive(struct task_struct *p,
2356 long match_state)
2357{
2358 return 1;
2359}
1da177e4
LT
2360#endif
2361
05725f7e
JP
2362#define next_task(p) \
2363 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2364
2365#define for_each_process(p) \
2366 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2367
5bb459bb 2368extern bool current_is_single_threaded(void);
d84f4f99 2369
1da177e4
LT
2370/*
2371 * Careful: do_each_thread/while_each_thread is a double loop so
2372 * 'break' will not work as expected - use goto instead.
2373 */
2374#define do_each_thread(g, t) \
2375 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2376
2377#define while_each_thread(g, t) \
2378 while ((t = next_thread(t)) != g)
2379
0c740d0a
ON
2380#define __for_each_thread(signal, t) \
2381 list_for_each_entry_rcu(t, &(signal)->thread_head, thread_node)
2382
2383#define for_each_thread(p, t) \
2384 __for_each_thread((p)->signal, t)
2385
2386/* Careful: this is a double loop, 'break' won't work as expected. */
2387#define for_each_process_thread(p, t) \
2388 for_each_process(p) for_each_thread(p, t)
2389
7e49827c
ON
2390static inline int get_nr_threads(struct task_struct *tsk)
2391{
b3ac022c 2392 return tsk->signal->nr_threads;
7e49827c
ON
2393}
2394
087806b1
ON
2395static inline bool thread_group_leader(struct task_struct *p)
2396{
2397 return p->exit_signal >= 0;
2398}
1da177e4 2399
0804ef4b
EB
2400/* Do to the insanities of de_thread it is possible for a process
2401 * to have the pid of the thread group leader without actually being
2402 * the thread group leader. For iteration through the pids in proc
2403 * all we care about is that we have a task with the appropriate
2404 * pid, we don't actually care if we have the right task.
2405 */
e1403b8e 2406static inline bool has_group_leader_pid(struct task_struct *p)
0804ef4b 2407{
e1403b8e 2408 return task_pid(p) == p->signal->leader_pid;
0804ef4b
EB
2409}
2410
bac0abd6 2411static inline
e1403b8e 2412bool same_thread_group(struct task_struct *p1, struct task_struct *p2)
bac0abd6 2413{
e1403b8e 2414 return p1->signal == p2->signal;
bac0abd6
PE
2415}
2416
36c8b586 2417static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2418{
05725f7e
JP
2419 return list_entry_rcu(p->thread_group.next,
2420 struct task_struct, thread_group);
47e65328
ON
2421}
2422
e868171a 2423static inline int thread_group_empty(struct task_struct *p)
1da177e4 2424{
47e65328 2425 return list_empty(&p->thread_group);
1da177e4
LT
2426}
2427
2428#define delay_group_leader(p) \
2429 (thread_group_leader(p) && !thread_group_empty(p))
2430
1da177e4 2431/*
260ea101 2432 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2433 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8 2434 * pins the final release of task.io_context. Also protects ->cpuset and
d68b46fe 2435 * ->cgroup.subsys[]. And ->vfork_done.
1da177e4
LT
2436 *
2437 * Nests both inside and outside of read_lock(&tasklist_lock).
2438 * It must not be nested with write_lock_irq(&tasklist_lock),
2439 * neither inside nor outside.
2440 */
2441static inline void task_lock(struct task_struct *p)
2442{
2443 spin_lock(&p->alloc_lock);
2444}
2445
2446static inline void task_unlock(struct task_struct *p)
2447{
2448 spin_unlock(&p->alloc_lock);
2449}
2450
b8ed374e 2451extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2452 unsigned long *flags);
2453
9388dc30
AV
2454static inline struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2455 unsigned long *flags)
2456{
2457 struct sighand_struct *ret;
2458
2459 ret = __lock_task_sighand(tsk, flags);
2460 (void)__cond_lock(&tsk->sighand->siglock, ret);
2461 return ret;
2462}
b8ed374e 2463
f63ee72e
ON
2464static inline void unlock_task_sighand(struct task_struct *tsk,
2465 unsigned long *flags)
2466{
2467 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2468}
2469
4714d1d3 2470#ifdef CONFIG_CGROUPS
257058ae 2471static inline void threadgroup_change_begin(struct task_struct *tsk)
4714d1d3 2472{
257058ae 2473 down_read(&tsk->signal->group_rwsem);
4714d1d3 2474}
257058ae 2475static inline void threadgroup_change_end(struct task_struct *tsk)
4714d1d3 2476{
257058ae 2477 up_read(&tsk->signal->group_rwsem);
4714d1d3 2478}
77e4ef99
TH
2479
2480/**
2481 * threadgroup_lock - lock threadgroup
2482 * @tsk: member task of the threadgroup to lock
2483 *
2484 * Lock the threadgroup @tsk belongs to. No new task is allowed to enter
2485 * and member tasks aren't allowed to exit (as indicated by PF_EXITING) or
e56fb287
ON
2486 * change ->group_leader/pid. This is useful for cases where the threadgroup
2487 * needs to stay stable across blockable operations.
77e4ef99
TH
2488 *
2489 * fork and exit paths explicitly call threadgroup_change_{begin|end}() for
2490 * synchronization. While held, no new task will be added to threadgroup
2491 * and no existing live task will have its PF_EXITING set.
2492 *
e56fb287
ON
2493 * de_thread() does threadgroup_change_{begin|end}() when a non-leader
2494 * sub-thread becomes a new leader.
77e4ef99 2495 */
257058ae 2496static inline void threadgroup_lock(struct task_struct *tsk)
4714d1d3 2497{
257058ae 2498 down_write(&tsk->signal->group_rwsem);
4714d1d3 2499}
77e4ef99
TH
2500
2501/**
2502 * threadgroup_unlock - unlock threadgroup
2503 * @tsk: member task of the threadgroup to unlock
2504 *
2505 * Reverse threadgroup_lock().
2506 */
257058ae 2507static inline void threadgroup_unlock(struct task_struct *tsk)
4714d1d3 2508{
257058ae 2509 up_write(&tsk->signal->group_rwsem);
4714d1d3
BB
2510}
2511#else
257058ae
TH
2512static inline void threadgroup_change_begin(struct task_struct *tsk) {}
2513static inline void threadgroup_change_end(struct task_struct *tsk) {}
2514static inline void threadgroup_lock(struct task_struct *tsk) {}
2515static inline void threadgroup_unlock(struct task_struct *tsk) {}
4714d1d3
BB
2516#endif
2517
f037360f
AV
2518#ifndef __HAVE_THREAD_FUNCTIONS
2519
f7e4217b
RZ
2520#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2521#define task_stack_page(task) ((task)->stack)
a1261f54 2522
10ebffde
AV
2523static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2524{
2525 *task_thread_info(p) = *task_thread_info(org);
2526 task_thread_info(p)->task = p;
2527}
2528
2529static inline unsigned long *end_of_stack(struct task_struct *p)
2530{
f7e4217b 2531 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2532}
2533
f037360f
AV
2534#endif
2535
8b05c7e6
FT
2536static inline int object_is_on_stack(void *obj)
2537{
2538 void *stack = task_stack_page(current);
2539
2540 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2541}
2542
8c9843e5
BH
2543extern void thread_info_cache_init(void);
2544
7c9f8861
ES
2545#ifdef CONFIG_DEBUG_STACK_USAGE
2546static inline unsigned long stack_not_used(struct task_struct *p)
2547{
2548 unsigned long *n = end_of_stack(p);
2549
2550 do { /* Skip over canary */
2551 n++;
2552 } while (!*n);
2553
2554 return (unsigned long)n - (unsigned long)end_of_stack(p);
2555}
2556#endif
2557
1da177e4
LT
2558/* set thread flags in other task's structures
2559 * - see asm/thread_info.h for TIF_xxxx flags available
2560 */
2561static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2562{
a1261f54 2563 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2564}
2565
2566static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2567{
a1261f54 2568 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2569}
2570
2571static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2572{
a1261f54 2573 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2574}
2575
2576static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2577{
a1261f54 2578 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2579}
2580
2581static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2582{
a1261f54 2583 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2584}
2585
2586static inline void set_tsk_need_resched(struct task_struct *tsk)
2587{
2588 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2589}
2590
2591static inline void clear_tsk_need_resched(struct task_struct *tsk)
2592{
2593 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2594}
2595
8ae121ac
GH
2596static inline int test_tsk_need_resched(struct task_struct *tsk)
2597{
2598 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2599}
2600
690cc3ff
EB
2601static inline int restart_syscall(void)
2602{
2603 set_tsk_thread_flag(current, TIF_SIGPENDING);
2604 return -ERESTARTNOINTR;
2605}
2606
1da177e4
LT
2607static inline int signal_pending(struct task_struct *p)
2608{
2609 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2610}
f776d12d 2611
d9588725
RM
2612static inline int __fatal_signal_pending(struct task_struct *p)
2613{
2614 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2615}
f776d12d
MW
2616
2617static inline int fatal_signal_pending(struct task_struct *p)
2618{
2619 return signal_pending(p) && __fatal_signal_pending(p);
2620}
2621
16882c1e
ON
2622static inline int signal_pending_state(long state, struct task_struct *p)
2623{
2624 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2625 return 0;
2626 if (!signal_pending(p))
2627 return 0;
2628
16882c1e
ON
2629 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2630}
2631
1da177e4
LT
2632/*
2633 * cond_resched() and cond_resched_lock(): latency reduction via
2634 * explicit rescheduling in places that are safe. The return
2635 * value indicates whether a reschedule was done in fact.
2636 * cond_resched_lock() will drop the spinlock before scheduling,
2637 * cond_resched_softirq() will enable bhs before scheduling.
2638 */
c3921ab7 2639extern int _cond_resched(void);
6f80bd98 2640
613afbf8
FW
2641#define cond_resched() ({ \
2642 __might_sleep(__FILE__, __LINE__, 0); \
2643 _cond_resched(); \
2644})
6f80bd98 2645
613afbf8
FW
2646extern int __cond_resched_lock(spinlock_t *lock);
2647
bdd4e85d 2648#ifdef CONFIG_PREEMPT_COUNT
716a4234 2649#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2650#else
716a4234 2651#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2652#endif
716a4234 2653
613afbf8 2654#define cond_resched_lock(lock) ({ \
716a4234 2655 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2656 __cond_resched_lock(lock); \
2657})
2658
2659extern int __cond_resched_softirq(void);
2660
75e1056f
VP
2661#define cond_resched_softirq() ({ \
2662 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2663 __cond_resched_softirq(); \
613afbf8 2664})
1da177e4 2665
f6f3c437
SH
2666static inline void cond_resched_rcu(void)
2667{
2668#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
2669 rcu_read_unlock();
2670 cond_resched();
2671 rcu_read_lock();
2672#endif
2673}
2674
1da177e4
LT
2675/*
2676 * Does a critical section need to be broken due to another
95c354fe
NP
2677 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2678 * but a general need for low latency)
1da177e4 2679 */
95c354fe 2680static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2681{
95c354fe
NP
2682#ifdef CONFIG_PREEMPT
2683 return spin_is_contended(lock);
2684#else
1da177e4 2685 return 0;
95c354fe 2686#endif
1da177e4
LT
2687}
2688
ee761f62
TG
2689/*
2690 * Idle thread specific functions to determine the need_resched
2691 * polling state. We have two versions, one based on TS_POLLING in
2692 * thread_info.status and one based on TIF_POLLING_NRFLAG in
2693 * thread_info.flags
2694 */
2695#ifdef TS_POLLING
2696static inline int tsk_is_polling(struct task_struct *p)
2697{
2698 return task_thread_info(p)->status & TS_POLLING;
2699}
ea811747 2700static inline void __current_set_polling(void)
3a98f871
TG
2701{
2702 current_thread_info()->status |= TS_POLLING;
2703}
2704
ea811747
PZ
2705static inline bool __must_check current_set_polling_and_test(void)
2706{
2707 __current_set_polling();
2708
2709 /*
2710 * Polling state must be visible before we test NEED_RESCHED,
2711 * paired by resched_task()
2712 */
2713 smp_mb();
2714
2715 return unlikely(tif_need_resched());
2716}
2717
2718static inline void __current_clr_polling(void)
3a98f871
TG
2719{
2720 current_thread_info()->status &= ~TS_POLLING;
ea811747
PZ
2721}
2722
2723static inline bool __must_check current_clr_polling_and_test(void)
2724{
2725 __current_clr_polling();
2726
2727 /*
2728 * Polling state must be visible before we test NEED_RESCHED,
2729 * paired by resched_task()
2730 */
2731 smp_mb();
2732
2733 return unlikely(tif_need_resched());
3a98f871 2734}
ee761f62
TG
2735#elif defined(TIF_POLLING_NRFLAG)
2736static inline int tsk_is_polling(struct task_struct *p)
2737{
2738 return test_tsk_thread_flag(p, TIF_POLLING_NRFLAG);
2739}
ea811747
PZ
2740
2741static inline void __current_set_polling(void)
3a98f871
TG
2742{
2743 set_thread_flag(TIF_POLLING_NRFLAG);
2744}
2745
ea811747
PZ
2746static inline bool __must_check current_set_polling_and_test(void)
2747{
2748 __current_set_polling();
2749
2750 /*
2751 * Polling state must be visible before we test NEED_RESCHED,
2752 * paired by resched_task()
2753 *
2754 * XXX: assumes set/clear bit are identical barrier wise.
2755 */
2756 smp_mb__after_clear_bit();
2757
2758 return unlikely(tif_need_resched());
2759}
2760
2761static inline void __current_clr_polling(void)
3a98f871
TG
2762{
2763 clear_thread_flag(TIF_POLLING_NRFLAG);
2764}
ea811747
PZ
2765
2766static inline bool __must_check current_clr_polling_and_test(void)
2767{
2768 __current_clr_polling();
2769
2770 /*
2771 * Polling state must be visible before we test NEED_RESCHED,
2772 * paired by resched_task()
2773 */
2774 smp_mb__after_clear_bit();
2775
2776 return unlikely(tif_need_resched());
2777}
2778
ee761f62
TG
2779#else
2780static inline int tsk_is_polling(struct task_struct *p) { return 0; }
ea811747
PZ
2781static inline void __current_set_polling(void) { }
2782static inline void __current_clr_polling(void) { }
2783
2784static inline bool __must_check current_set_polling_and_test(void)
2785{
2786 return unlikely(tif_need_resched());
2787}
2788static inline bool __must_check current_clr_polling_and_test(void)
2789{
2790 return unlikely(tif_need_resched());
2791}
ee761f62
TG
2792#endif
2793
8cb75e0c
PZ
2794static inline void current_clr_polling(void)
2795{
2796 __current_clr_polling();
2797
2798 /*
2799 * Ensure we check TIF_NEED_RESCHED after we clear the polling bit.
2800 * Once the bit is cleared, we'll get IPIs with every new
2801 * TIF_NEED_RESCHED and the IPI handler, scheduler_ipi(), will also
2802 * fold.
2803 */
2804 smp_mb(); /* paired with resched_task() */
2805
2806 preempt_fold_need_resched();
2807}
2808
75f93fed
PZ
2809static __always_inline bool need_resched(void)
2810{
2811 return unlikely(tif_need_resched());
2812}
2813
f06febc9
FM
2814/*
2815 * Thread group CPU time accounting.
2816 */
4cd4c1b4 2817void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2818void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2819
490dea45 2820static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2821{
ee30a7b2 2822 raw_spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2823}
2824
7bb44ade
RM
2825/*
2826 * Reevaluate whether the task has signals pending delivery.
2827 * Wake the task if so.
2828 * This is required every time the blocked sigset_t changes.
2829 * callers must hold sighand->siglock.
2830 */
2831extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2832extern void recalc_sigpending(void);
2833
910ffdb1
ON
2834extern void signal_wake_up_state(struct task_struct *t, unsigned int state);
2835
2836static inline void signal_wake_up(struct task_struct *t, bool resume)
2837{
2838 signal_wake_up_state(t, resume ? TASK_WAKEKILL : 0);
2839}
2840static inline void ptrace_signal_wake_up(struct task_struct *t, bool resume)
2841{
2842 signal_wake_up_state(t, resume ? __TASK_TRACED : 0);
2843}
1da177e4
LT
2844
2845/*
2846 * Wrappers for p->thread_info->cpu access. No-op on UP.
2847 */
2848#ifdef CONFIG_SMP
2849
2850static inline unsigned int task_cpu(const struct task_struct *p)
2851{
a1261f54 2852 return task_thread_info(p)->cpu;
1da177e4
LT
2853}
2854
b32e86b4
IM
2855static inline int task_node(const struct task_struct *p)
2856{
2857 return cpu_to_node(task_cpu(p));
2858}
2859
c65cc870 2860extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2861
2862#else
2863
2864static inline unsigned int task_cpu(const struct task_struct *p)
2865{
2866 return 0;
2867}
2868
2869static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2870{
2871}
2872
2873#endif /* CONFIG_SMP */
2874
96f874e2
RR
2875extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2876extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2877
7c941438 2878#ifdef CONFIG_CGROUP_SCHED
07e06b01 2879extern struct task_group root_task_group;
8323f26c 2880#endif /* CONFIG_CGROUP_SCHED */
9b5b7751 2881
54e99124
DG
2882extern int task_can_switch_user(struct user_struct *up,
2883 struct task_struct *tsk);
2884
4b98d11b
AD
2885#ifdef CONFIG_TASK_XACCT
2886static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2887{
940389b8 2888 tsk->ioac.rchar += amt;
4b98d11b
AD
2889}
2890
2891static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2892{
940389b8 2893 tsk->ioac.wchar += amt;
4b98d11b
AD
2894}
2895
2896static inline void inc_syscr(struct task_struct *tsk)
2897{
940389b8 2898 tsk->ioac.syscr++;
4b98d11b
AD
2899}
2900
2901static inline void inc_syscw(struct task_struct *tsk)
2902{
940389b8 2903 tsk->ioac.syscw++;
4b98d11b
AD
2904}
2905#else
2906static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2907{
2908}
2909
2910static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2911{
2912}
2913
2914static inline void inc_syscr(struct task_struct *tsk)
2915{
2916}
2917
2918static inline void inc_syscw(struct task_struct *tsk)
2919{
2920}
2921#endif
2922
82455257
DH
2923#ifndef TASK_SIZE_OF
2924#define TASK_SIZE_OF(tsk) TASK_SIZE
2925#endif
2926
cf475ad2
BS
2927#ifdef CONFIG_MM_OWNER
2928extern void mm_update_next_owner(struct mm_struct *mm);
2929extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2930#else
2931static inline void mm_update_next_owner(struct mm_struct *mm)
2932{
2933}
2934
2935static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2936{
2937}
2938#endif /* CONFIG_MM_OWNER */
2939
3e10e716
JS
2940static inline unsigned long task_rlimit(const struct task_struct *tsk,
2941 unsigned int limit)
2942{
2943 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2944}
2945
2946static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2947 unsigned int limit)
2948{
2949 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2950}
2951
2952static inline unsigned long rlimit(unsigned int limit)
2953{
2954 return task_rlimit(current, limit);
2955}
2956
2957static inline unsigned long rlimit_max(unsigned int limit)
2958{
2959 return task_rlimit_max(current, limit);
2960}
2961
1da177e4 2962#endif