Merge tag 'pci-v3.15-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaa...
[deliverable/linux.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>
bfc3f028 32#include <linux/cputime.h>
1da177e4
LT
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);
6201b4d6 297extern int get_nohz_timer_target(int pinned);
46cb4b7c 298#else
c1cc017c 299static inline void nohz_balance_enter_idle(int cpu) { }
fdaabd80 300static inline void set_cpu_sd_state_idle(void) { }
6201b4d6
VK
301static inline int get_nohz_timer_target(int pinned)
302{
303 return smp_processor_id();
304}
46cb4b7c 305#endif
1da177e4 306
e59e2ae2 307/*
39bc89fd 308 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
309 */
310extern void show_state_filter(unsigned long state_filter);
311
312static inline void show_state(void)
313{
39bc89fd 314 show_state_filter(0);
e59e2ae2
IM
315}
316
1da177e4
LT
317extern void show_regs(struct pt_regs *);
318
319/*
320 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
321 * task), SP is the stack pointer of the first frame that should be shown in the back
322 * trace (or NULL if the entire call-chain of the task should be shown).
323 */
324extern void show_stack(struct task_struct *task, unsigned long *sp);
325
326void io_schedule(void);
327long io_schedule_timeout(long timeout);
328
329extern void cpu_init (void);
330extern void trap_init(void);
331extern void update_process_times(int user);
332extern void scheduler_tick(void);
333
82a1fcb9
IM
334extern void sched_show_task(struct task_struct *p);
335
19cc36c0 336#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 337extern void touch_softlockup_watchdog(void);
d6ad3e28 338extern void touch_softlockup_watchdog_sync(void);
04c9167f 339extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
340extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
341 void __user *buffer,
342 size_t *lenp, loff_t *ppos);
9c44bc03 343extern unsigned int softlockup_panic;
004417a6 344void lockup_detector_init(void);
8446f1d3 345#else
8446f1d3
IM
346static inline void touch_softlockup_watchdog(void)
347{
348}
d6ad3e28
JW
349static inline void touch_softlockup_watchdog_sync(void)
350{
351}
04c9167f
JF
352static inline void touch_all_softlockup_watchdogs(void)
353{
354}
004417a6
PZ
355static inline void lockup_detector_init(void)
356{
357}
8446f1d3
IM
358#endif
359
8b414521
MT
360#ifdef CONFIG_DETECT_HUNG_TASK
361void reset_hung_task_detector(void);
362#else
363static inline void reset_hung_task_detector(void)
364{
365}
366#endif
367
1da177e4
LT
368/* Attach to any functions which should be ignored in wchan output. */
369#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
370
371/* Linker adds these: start and end of __sched functions */
372extern char __sched_text_start[], __sched_text_end[];
373
1da177e4
LT
374/* Is this address in the __sched functions? */
375extern int in_sched_functions(unsigned long addr);
376
377#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 378extern signed long schedule_timeout(signed long timeout);
64ed93a2 379extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 380extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 381extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 382asmlinkage void schedule(void);
c5491ea7 383extern void schedule_preempt_disabled(void);
1da177e4 384
ab516013 385struct nsproxy;
acce292c 386struct user_namespace;
1da177e4 387
efc1a3b1
DH
388#ifdef CONFIG_MMU
389extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
390extern unsigned long
391arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
392 unsigned long, unsigned long);
393extern unsigned long
394arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
395 unsigned long len, unsigned long pgoff,
396 unsigned long flags);
efc1a3b1
DH
397#else
398static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
399#endif
1da177e4 400
d049f74f
KC
401#define SUID_DUMP_DISABLE 0 /* No setuid dumping */
402#define SUID_DUMP_USER 1 /* Dump as user of process */
403#define SUID_DUMP_ROOT 2 /* Dump as root */
404
6c5d5238 405/* mm flags */
f8af4da3 406
7288e118 407/* for SUID_DUMP_* above */
3cb4a0bb 408#define MMF_DUMPABLE_BITS 2
f8af4da3 409#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb 410
942be387
ON
411extern void set_dumpable(struct mm_struct *mm, int value);
412/*
413 * This returns the actual value of the suid_dumpable flag. For things
414 * that are using this for checking for privilege transitions, it must
415 * test against SUID_DUMP_USER rather than treating it as a boolean
416 * value.
417 */
418static inline int __get_dumpable(unsigned long mm_flags)
419{
420 return mm_flags & MMF_DUMPABLE_MASK;
421}
422
423static inline int get_dumpable(struct mm_struct *mm)
424{
425 return __get_dumpable(mm->flags);
426}
427
3cb4a0bb
KH
428/* coredump filter bits */
429#define MMF_DUMP_ANON_PRIVATE 2
430#define MMF_DUMP_ANON_SHARED 3
431#define MMF_DUMP_MAPPED_PRIVATE 4
432#define MMF_DUMP_MAPPED_SHARED 5
82df3973 433#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
434#define MMF_DUMP_HUGETLB_PRIVATE 7
435#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 436
3cb4a0bb 437#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 438#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
439#define MMF_DUMP_FILTER_MASK \
440 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
441#define MMF_DUMP_FILTER_DEFAULT \
e575f111 442 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
443 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
444
445#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
446# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
447#else
448# define MMF_DUMP_MASK_DEFAULT_ELF 0
449#endif
f8af4da3
HD
450 /* leave room for more dump flags */
451#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 452#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
bafb282d 453#define MMF_EXE_FILE_CHANGED 18 /* see prctl_set_mm_exe_file() */
f8af4da3 454
9f68f672
ON
455#define MMF_HAS_UPROBES 19 /* has uprobes */
456#define MMF_RECALC_UPROBES 20 /* MMF_HAS_UPROBES can be wrong */
f8ac4ec9 457
f8af4da3 458#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 459
1da177e4
LT
460struct sighand_struct {
461 atomic_t count;
462 struct k_sigaction action[_NSIG];
463 spinlock_t siglock;
b8fceee1 464 wait_queue_head_t signalfd_wqh;
1da177e4
LT
465};
466
0e464814 467struct pacct_struct {
f6ec29a4
KK
468 int ac_flag;
469 long ac_exitcode;
0e464814 470 unsigned long ac_mem;
77787bfb
KK
471 cputime_t ac_utime, ac_stime;
472 unsigned long ac_minflt, ac_majflt;
0e464814
KK
473};
474
42c4ab41
SG
475struct cpu_itimer {
476 cputime_t expires;
477 cputime_t incr;
8356b5f9
SG
478 u32 error;
479 u32 incr_error;
42c4ab41
SG
480};
481
d37f761d
FW
482/**
483 * struct cputime - snaphsot of system and user cputime
484 * @utime: time spent in user mode
485 * @stime: time spent in system mode
486 *
487 * Gathers a generic snapshot of user and system time.
488 */
489struct cputime {
490 cputime_t utime;
491 cputime_t stime;
492};
493
f06febc9
FM
494/**
495 * struct task_cputime - collected CPU time counts
496 * @utime: time spent in user mode, in &cputime_t units
497 * @stime: time spent in kernel mode, in &cputime_t units
498 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 499 *
d37f761d
FW
500 * This is an extension of struct cputime that includes the total runtime
501 * spent by the task from the scheduler point of view.
502 *
503 * As a result, this structure groups together three kinds of CPU time
504 * that are tracked for threads and thread groups. Most things considering
f06febc9
FM
505 * CPU time want to group these counts together and treat all three
506 * of them in parallel.
507 */
508struct task_cputime {
509 cputime_t utime;
510 cputime_t stime;
511 unsigned long long sum_exec_runtime;
512};
513/* Alternate field names when used to cache expirations. */
514#define prof_exp stime
515#define virt_exp utime
516#define sched_exp sum_exec_runtime
517
4cd4c1b4
PZ
518#define INIT_CPUTIME \
519 (struct task_cputime) { \
64861634
MS
520 .utime = 0, \
521 .stime = 0, \
4cd4c1b4
PZ
522 .sum_exec_runtime = 0, \
523 }
524
a233f112
PZ
525#ifdef CONFIG_PREEMPT_COUNT
526#define PREEMPT_DISABLED (1 + PREEMPT_ENABLED)
527#else
528#define PREEMPT_DISABLED PREEMPT_ENABLED
529#endif
530
c99e6efe
PZ
531/*
532 * Disable preemption until the scheduler is running.
533 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
534 *
535 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
536 * before the scheduler is active -- see should_resched().
c99e6efe 537 */
a233f112 538#define INIT_PREEMPT_COUNT (PREEMPT_DISABLED + PREEMPT_ACTIVE)
c99e6efe 539
f06febc9 540/**
4cd4c1b4
PZ
541 * struct thread_group_cputimer - thread group interval timer counts
542 * @cputime: thread group interval timers.
543 * @running: non-zero when there are timers running and
544 * @cputime receives updates.
545 * @lock: lock for fields in this struct.
f06febc9
FM
546 *
547 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 548 * used for thread group CPU timer calculations.
f06febc9 549 */
4cd4c1b4
PZ
550struct thread_group_cputimer {
551 struct task_cputime cputime;
552 int running;
ee30a7b2 553 raw_spinlock_t lock;
f06febc9 554};
f06febc9 555
4714d1d3 556#include <linux/rwsem.h>
5091faa4
MG
557struct autogroup;
558
1da177e4 559/*
e815f0a8 560 * NOTE! "signal_struct" does not have its own
1da177e4
LT
561 * locking, because a shared signal_struct always
562 * implies a shared sighand_struct, so locking
563 * sighand_struct is always a proper superset of
564 * the locking of signal_struct.
565 */
566struct signal_struct {
ea6d290c 567 atomic_t sigcnt;
1da177e4 568 atomic_t live;
b3ac022c 569 int nr_threads;
0c740d0a 570 struct list_head thread_head;
1da177e4
LT
571
572 wait_queue_head_t wait_chldexit; /* for wait4() */
573
574 /* current thread group signal load-balancing target: */
36c8b586 575 struct task_struct *curr_target;
1da177e4
LT
576
577 /* shared signal handling: */
578 struct sigpending shared_pending;
579
580 /* thread group exit support */
581 int group_exit_code;
582 /* overloaded:
583 * - notify group_exit_task when ->count is equal to notify_count
584 * - everyone except group_exit_task is stopped during signal delivery
585 * of fatal signals, group_exit_task processes the signal.
586 */
1da177e4 587 int notify_count;
07dd20e0 588 struct task_struct *group_exit_task;
1da177e4
LT
589
590 /* thread group stop support, overloads group_exit_code too */
591 int group_stop_count;
592 unsigned int flags; /* see SIGNAL_* flags below */
593
ebec18a6
LP
594 /*
595 * PR_SET_CHILD_SUBREAPER marks a process, like a service
596 * manager, to re-parent orphan (double-forking) child processes
597 * to this process instead of 'init'. The service manager is
598 * able to receive SIGCHLD signals and is able to investigate
599 * the process until it calls wait(). All children of this
600 * process will inherit a flag if they should look for a
601 * child_subreaper process at exit.
602 */
603 unsigned int is_child_subreaper:1;
604 unsigned int has_child_subreaper:1;
605
1da177e4 606 /* POSIX.1b Interval Timers */
5ed67f05
PE
607 int posix_timer_id;
608 struct list_head posix_timers;
1da177e4
LT
609
610 /* ITIMER_REAL timer for the process */
2ff678b8 611 struct hrtimer real_timer;
fea9d175 612 struct pid *leader_pid;
2ff678b8 613 ktime_t it_real_incr;
1da177e4 614
42c4ab41
SG
615 /*
616 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
617 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
618 * values are defined to 0 and 1 respectively
619 */
620 struct cpu_itimer it[2];
1da177e4 621
f06febc9 622 /*
4cd4c1b4
PZ
623 * Thread group totals for process CPU timers.
624 * See thread_group_cputimer(), et al, for details.
f06febc9 625 */
4cd4c1b4 626 struct thread_group_cputimer cputimer;
f06febc9
FM
627
628 /* Earliest-expiration cache. */
629 struct task_cputime cputime_expires;
630
631 struct list_head cpu_timers[3];
632
ab521dc0 633 struct pid *tty_old_pgrp;
1ec320af 634
1da177e4
LT
635 /* boolean value for session group leader */
636 int leader;
637
638 struct tty_struct *tty; /* NULL if no tty */
639
5091faa4
MG
640#ifdef CONFIG_SCHED_AUTOGROUP
641 struct autogroup *autogroup;
642#endif
1da177e4
LT
643 /*
644 * Cumulative resource counters for dead threads in the group,
645 * and for reaped dead child processes forked by this group.
646 * Live threads maintain their own counters and add to these
647 * in __exit_signal, except for the group leader.
648 */
32bd671d 649 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
650 cputime_t gtime;
651 cputime_t cgtime;
9fbc42ea 652#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 653 struct cputime prev_cputime;
0cf55e1e 654#endif
1da177e4
LT
655 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
656 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 657 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 658 unsigned long maxrss, cmaxrss;
940389b8 659 struct task_io_accounting ioac;
1da177e4 660
32bd671d
PZ
661 /*
662 * Cumulative ns of schedule CPU time fo dead threads in the
663 * group, not including a zombie group leader, (This only differs
664 * from jiffies_to_ns(utime + stime) if sched_clock uses something
665 * other than jiffies.)
666 */
667 unsigned long long sum_sched_runtime;
668
1da177e4
LT
669 /*
670 * We don't bother to synchronize most readers of this at all,
671 * because there is no reader checking a limit that actually needs
672 * to get both rlim_cur and rlim_max atomically, and either one
673 * alone is a single word that can safely be read normally.
674 * getrlimit/setrlimit use task_lock(current->group_leader) to
675 * protect this instead of the siglock, because they really
676 * have no need to disable irqs.
677 */
678 struct rlimit rlim[RLIM_NLIMITS];
679
0e464814
KK
680#ifdef CONFIG_BSD_PROCESS_ACCT
681 struct pacct_struct pacct; /* per-process accounting information */
682#endif
ad4ecbcb 683#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
684 struct taskstats *stats;
685#endif
522ed776
MT
686#ifdef CONFIG_AUDIT
687 unsigned audit_tty;
46e959ea 688 unsigned audit_tty_log_passwd;
522ed776
MT
689 struct tty_audit_buf *tty_audit_buf;
690#endif
4714d1d3
BB
691#ifdef CONFIG_CGROUPS
692 /*
77e4ef99
TH
693 * group_rwsem prevents new tasks from entering the threadgroup and
694 * member tasks from exiting,a more specifically, setting of
695 * PF_EXITING. fork and exit paths are protected with this rwsem
696 * using threadgroup_change_begin/end(). Users which require
697 * threadgroup to remain stable should use threadgroup_[un]lock()
698 * which also takes care of exec path. Currently, cgroup is the
699 * only user.
4714d1d3 700 */
257058ae 701 struct rw_semaphore group_rwsem;
4714d1d3 702#endif
28b83c51 703
e1e12d2f 704 oom_flags_t oom_flags;
a9c58b90
DR
705 short oom_score_adj; /* OOM kill score adjustment */
706 short oom_score_adj_min; /* OOM kill score adjustment min value.
707 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
708
709 struct mutex cred_guard_mutex; /* guard against foreign influences on
710 * credential calculations
711 * (notably. ptrace) */
1da177e4
LT
712};
713
714/*
715 * Bits in flags field of signal_struct.
716 */
717#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
ee77f075
ON
718#define SIGNAL_STOP_CONTINUED 0x00000002 /* SIGCONT since WCONTINUED reap */
719#define SIGNAL_GROUP_EXIT 0x00000004 /* group exit in progress */
403bad72 720#define SIGNAL_GROUP_COREDUMP 0x00000008 /* coredump in progress */
e4420551
ON
721/*
722 * Pending notifications to parent.
723 */
724#define SIGNAL_CLD_STOPPED 0x00000010
725#define SIGNAL_CLD_CONTINUED 0x00000020
726#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 727
fae5fa44
ON
728#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
729
ed5d2cac
ON
730/* If true, all threads except ->group_exit_task have pending SIGKILL */
731static inline int signal_group_exit(const struct signal_struct *sig)
732{
733 return (sig->flags & SIGNAL_GROUP_EXIT) ||
734 (sig->group_exit_task != NULL);
735}
736
1da177e4
LT
737/*
738 * Some day this will be a full-fledged user tracking system..
739 */
740struct user_struct {
741 atomic_t __count; /* reference count */
742 atomic_t processes; /* How many processes does this user have? */
743 atomic_t files; /* How many open files does this user have? */
744 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 745#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
746 atomic_t inotify_watches; /* How many inotify watches does this user have? */
747 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
748#endif
4afeff85
EP
749#ifdef CONFIG_FANOTIFY
750 atomic_t fanotify_listeners;
751#endif
7ef9964e 752#ifdef CONFIG_EPOLL
52bd19f7 753 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 754#endif
970a8645 755#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
756 /* protected by mq_lock */
757 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 758#endif
1da177e4
LT
759 unsigned long locked_shm; /* How many pages of mlocked shm ? */
760
761#ifdef CONFIG_KEYS
762 struct key *uid_keyring; /* UID specific keyring */
763 struct key *session_keyring; /* UID's default session keyring */
764#endif
765
766 /* Hash table maintenance information */
735de223 767 struct hlist_node uidhash_node;
7b44ab97 768 kuid_t uid;
24e377a8 769
cdd6c482 770#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
771 atomic_long_t locked_vm;
772#endif
1da177e4
LT
773};
774
eb41d946 775extern int uids_sysfs_init(void);
5cb350ba 776
7b44ab97 777extern struct user_struct *find_user(kuid_t);
1da177e4
LT
778
779extern struct user_struct root_user;
780#define INIT_USER (&root_user)
781
b6dff3ec 782
1da177e4
LT
783struct backing_dev_info;
784struct reclaim_state;
785
52f17b6c 786#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
787struct sched_info {
788 /* cumulative counters */
2d72376b 789 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 790 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
791
792 /* timestamps */
172ba844
BS
793 unsigned long long last_arrival,/* when we last ran on a cpu */
794 last_queued; /* when we were last queued to run */
1da177e4 795};
52f17b6c 796#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 797
ca74e92b
SN
798#ifdef CONFIG_TASK_DELAY_ACCT
799struct task_delay_info {
800 spinlock_t lock;
801 unsigned int flags; /* Private per-task flags */
802
803 /* For each stat XXX, add following, aligned appropriately
804 *
805 * struct timespec XXX_start, XXX_end;
806 * u64 XXX_delay;
807 * u32 XXX_count;
808 *
809 * Atomicity of updates to XXX_delay, XXX_count protected by
810 * single lock above (split into XXX_lock if contention is an issue).
811 */
0ff92245
SN
812
813 /*
814 * XXX_count is incremented on every XXX operation, the delay
815 * associated with the operation is added to XXX_delay.
816 * XXX_delay contains the accumulated delay time in nanoseconds.
817 */
818 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
819 u64 blkio_delay; /* wait for sync block io completion */
820 u64 swapin_delay; /* wait for swapin block io completion */
821 u32 blkio_count; /* total count of the number of sync block */
822 /* io operations performed */
823 u32 swapin_count; /* total count of the number of swapin block */
824 /* io operations performed */
873b4771
KK
825
826 struct timespec freepages_start, freepages_end;
827 u64 freepages_delay; /* wait for memory reclaim */
828 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 829};
52f17b6c
CS
830#endif /* CONFIG_TASK_DELAY_ACCT */
831
832static inline int sched_info_on(void)
833{
834#ifdef CONFIG_SCHEDSTATS
835 return 1;
836#elif defined(CONFIG_TASK_DELAY_ACCT)
837 extern int delayacct_on;
838 return delayacct_on;
839#else
840 return 0;
ca74e92b 841#endif
52f17b6c 842}
ca74e92b 843
d15bcfdb
IM
844enum cpu_idle_type {
845 CPU_IDLE,
846 CPU_NOT_IDLE,
847 CPU_NEWLY_IDLE,
848 CPU_MAX_IDLE_TYPES
1da177e4
LT
849};
850
1399fa78
NR
851/*
852 * Increase resolution of cpu_power calculations
853 */
854#define SCHED_POWER_SHIFT 10
855#define SCHED_POWER_SCALE (1L << SCHED_POWER_SHIFT)
1da177e4 856
1399fa78
NR
857/*
858 * sched-domains (multiprocessor balancing) declarations:
859 */
2dd73a4f 860#ifdef CONFIG_SMP
b5d978e0
PZ
861#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
862#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
863#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
864#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 865#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 866#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
b5d978e0 867#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
b5d978e0
PZ
868#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
869#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 870#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 871#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
e3589f6c 872#define SD_OVERLAP 0x2000 /* sched_domains of this level overlap */
3a7053b3 873#define SD_NUMA 0x4000 /* cross-node balancing */
5c45bf27 874
532cb4c4
MN
875extern int __weak arch_sd_sibiling_asym_packing(void);
876
1d3504fc
HS
877struct sched_domain_attr {
878 int relax_domain_level;
879};
880
881#define SD_ATTR_INIT (struct sched_domain_attr) { \
882 .relax_domain_level = -1, \
883}
884
60495e77
PZ
885extern int sched_domain_level_max;
886
5e6521ea
LZ
887struct sched_group;
888
1da177e4
LT
889struct sched_domain {
890 /* These fields must be setup */
891 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 892 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 893 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
894 unsigned long min_interval; /* Minimum balance interval ms */
895 unsigned long max_interval; /* Maximum balance interval ms */
896 unsigned int busy_factor; /* less balancing by factor if busy */
897 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 898 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
899 unsigned int busy_idx;
900 unsigned int idle_idx;
901 unsigned int newidle_idx;
902 unsigned int wake_idx;
147cbb4b 903 unsigned int forkexec_idx;
a52bfd73 904 unsigned int smt_gain;
25f55d9d
VG
905
906 int nohz_idle; /* NOHZ IDLE status */
1da177e4 907 int flags; /* See SD_* */
60495e77 908 int level;
1da177e4
LT
909
910 /* Runtime fields. */
911 unsigned long last_balance; /* init to jiffies. units in jiffies */
912 unsigned int balance_interval; /* initialise to 1. units in ms. */
913 unsigned int nr_balance_failed; /* initialise to 0 */
914
f48627e6 915 /* idle_balance() stats */
9bd721c5 916 u64 max_newidle_lb_cost;
f48627e6 917 unsigned long next_decay_max_lb_cost;
2398f2c6 918
1da177e4
LT
919#ifdef CONFIG_SCHEDSTATS
920 /* load_balance() stats */
480b9434
KC
921 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
922 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
923 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
924 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
925 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
926 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
927 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
928 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
929
930 /* Active load balancing */
480b9434
KC
931 unsigned int alb_count;
932 unsigned int alb_failed;
933 unsigned int alb_pushed;
1da177e4 934
68767a0a 935 /* SD_BALANCE_EXEC stats */
480b9434
KC
936 unsigned int sbe_count;
937 unsigned int sbe_balanced;
938 unsigned int sbe_pushed;
1da177e4 939
68767a0a 940 /* SD_BALANCE_FORK stats */
480b9434
KC
941 unsigned int sbf_count;
942 unsigned int sbf_balanced;
943 unsigned int sbf_pushed;
68767a0a 944
1da177e4 945 /* try_to_wake_up() stats */
480b9434
KC
946 unsigned int ttwu_wake_remote;
947 unsigned int ttwu_move_affine;
948 unsigned int ttwu_move_balance;
1da177e4 949#endif
a5d8c348
IM
950#ifdef CONFIG_SCHED_DEBUG
951 char *name;
952#endif
dce840a0
PZ
953 union {
954 void *private; /* used during construction */
955 struct rcu_head rcu; /* used during destruction */
956 };
6c99e9ad 957
669c55e9 958 unsigned int span_weight;
4200efd9
IM
959 /*
960 * Span of all CPUs in this domain.
961 *
962 * NOTE: this field is variable length. (Allocated dynamically
963 * by attaching extra space to the end of the structure,
964 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
965 */
966 unsigned long span[0];
1da177e4
LT
967};
968
758b2cdc
RR
969static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
970{
6c99e9ad 971 return to_cpumask(sd->span);
758b2cdc
RR
972}
973
acc3f5d7 974extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 975 struct sched_domain_attr *dattr_new);
029190c5 976
acc3f5d7
RR
977/* Allocate an array of sched domains, for partition_sched_domains(). */
978cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
979void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
980
39be3501
PZ
981bool cpus_share_cache(int this_cpu, int that_cpu);
982
1b427c15 983#else /* CONFIG_SMP */
1da177e4 984
1b427c15 985struct sched_domain_attr;
d02c7a8c 986
1b427c15 987static inline void
acc3f5d7 988partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
989 struct sched_domain_attr *dattr_new)
990{
d02c7a8c 991}
39be3501
PZ
992
993static inline bool cpus_share_cache(int this_cpu, int that_cpu)
994{
995 return true;
996}
997
1b427c15 998#endif /* !CONFIG_SMP */
1da177e4 999
47fe38fc 1000
1da177e4 1001struct io_context; /* See blkdev.h */
1da177e4 1002
1da177e4 1003
383f2835 1004#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1005extern void prefetch_stack(struct task_struct *t);
383f2835
CK
1006#else
1007static inline void prefetch_stack(struct task_struct *t) { }
1008#endif
1da177e4
LT
1009
1010struct audit_context; /* See audit.c */
1011struct mempolicy;
b92ce558 1012struct pipe_inode_info;
4865ecf1 1013struct uts_namespace;
1da177e4 1014
20b8a59f 1015struct load_weight {
9dbdb155
PZ
1016 unsigned long weight;
1017 u32 inv_weight;
20b8a59f
IM
1018};
1019
9d85f21c
PT
1020struct sched_avg {
1021 /*
1022 * These sums represent an infinite geometric series and so are bound
239003ea 1023 * above by 1024/(1-y). Thus we only need a u32 to store them for all
9d85f21c
PT
1024 * choices of y < 1-2^(-32)*1024.
1025 */
1026 u32 runnable_avg_sum, runnable_avg_period;
1027 u64 last_runnable_update;
9ee474f5 1028 s64 decay_count;
2dac754e 1029 unsigned long load_avg_contrib;
9d85f21c
PT
1030};
1031
94c18227 1032#ifdef CONFIG_SCHEDSTATS
41acab88 1033struct sched_statistics {
20b8a59f 1034 u64 wait_start;
94c18227 1035 u64 wait_max;
6d082592
AV
1036 u64 wait_count;
1037 u64 wait_sum;
8f0dfc34
AV
1038 u64 iowait_count;
1039 u64 iowait_sum;
94c18227 1040
20b8a59f 1041 u64 sleep_start;
20b8a59f 1042 u64 sleep_max;
94c18227
IM
1043 s64 sum_sleep_runtime;
1044
1045 u64 block_start;
20b8a59f
IM
1046 u64 block_max;
1047 u64 exec_max;
eba1ed4b 1048 u64 slice_max;
cc367732 1049
cc367732
IM
1050 u64 nr_migrations_cold;
1051 u64 nr_failed_migrations_affine;
1052 u64 nr_failed_migrations_running;
1053 u64 nr_failed_migrations_hot;
1054 u64 nr_forced_migrations;
cc367732
IM
1055
1056 u64 nr_wakeups;
1057 u64 nr_wakeups_sync;
1058 u64 nr_wakeups_migrate;
1059 u64 nr_wakeups_local;
1060 u64 nr_wakeups_remote;
1061 u64 nr_wakeups_affine;
1062 u64 nr_wakeups_affine_attempts;
1063 u64 nr_wakeups_passive;
1064 u64 nr_wakeups_idle;
41acab88
LDM
1065};
1066#endif
1067
1068struct sched_entity {
1069 struct load_weight load; /* for load-balancing */
1070 struct rb_node run_node;
1071 struct list_head group_node;
1072 unsigned int on_rq;
1073
1074 u64 exec_start;
1075 u64 sum_exec_runtime;
1076 u64 vruntime;
1077 u64 prev_sum_exec_runtime;
1078
41acab88
LDM
1079 u64 nr_migrations;
1080
41acab88
LDM
1081#ifdef CONFIG_SCHEDSTATS
1082 struct sched_statistics statistics;
94c18227
IM
1083#endif
1084
20b8a59f 1085#ifdef CONFIG_FAIR_GROUP_SCHED
fed14d45 1086 int depth;
20b8a59f
IM
1087 struct sched_entity *parent;
1088 /* rq on which this entity is (to be) queued: */
1089 struct cfs_rq *cfs_rq;
1090 /* rq "owned" by this entity/group: */
1091 struct cfs_rq *my_q;
1092#endif
8bd75c77 1093
141965c7 1094#ifdef CONFIG_SMP
f4e26b12 1095 /* Per-entity load-tracking */
9d85f21c
PT
1096 struct sched_avg avg;
1097#endif
20b8a59f 1098};
70b97a7f 1099
fa717060
PZ
1100struct sched_rt_entity {
1101 struct list_head run_list;
78f2c7db 1102 unsigned long timeout;
57d2aa00 1103 unsigned long watchdog_stamp;
bee367ed 1104 unsigned int time_slice;
6f505b16 1105
58d6c2d7 1106 struct sched_rt_entity *back;
052f1dc7 1107#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1108 struct sched_rt_entity *parent;
1109 /* rq on which this entity is (to be) queued: */
1110 struct rt_rq *rt_rq;
1111 /* rq "owned" by this entity/group: */
1112 struct rt_rq *my_q;
1113#endif
fa717060
PZ
1114};
1115
aab03e05
DF
1116struct sched_dl_entity {
1117 struct rb_node rb_node;
1118
1119 /*
1120 * Original scheduling parameters. Copied here from sched_attr
1121 * during sched_setscheduler2(), they will remain the same until
1122 * the next sched_setscheduler2().
1123 */
1124 u64 dl_runtime; /* maximum runtime for each instance */
1125 u64 dl_deadline; /* relative deadline of each instance */
755378a4 1126 u64 dl_period; /* separation of two instances (period) */
332ac17e 1127 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
1128
1129 /*
1130 * Actual scheduling parameters. Initialized with the values above,
1131 * they are continously updated during task execution. Note that
1132 * the remaining runtime could be < 0 in case we are in overrun.
1133 */
1134 s64 runtime; /* remaining runtime for this instance */
1135 u64 deadline; /* absolute deadline for this instance */
1136 unsigned int flags; /* specifying the scheduler behaviour */
1137
1138 /*
1139 * Some bool flags:
1140 *
1141 * @dl_throttled tells if we exhausted the runtime. If so, the
1142 * task has to wait for a replenishment to be performed at the
1143 * next firing of dl_timer.
1144 *
1145 * @dl_new tells if a new instance arrived. If so we must
1146 * start executing it with full runtime and reset its absolute
1147 * deadline;
2d3d891d
DF
1148 *
1149 * @dl_boosted tells if we are boosted due to DI. If so we are
1150 * outside bandwidth enforcement mechanism (but only until we
1151 * exit the critical section).
aab03e05 1152 */
2d3d891d 1153 int dl_throttled, dl_new, dl_boosted;
aab03e05
DF
1154
1155 /*
1156 * Bandwidth enforcement timer. Each -deadline task has its
1157 * own bandwidth to be enforced, thus we need one timer per task.
1158 */
1159 struct hrtimer dl_timer;
1160};
8bd75c77 1161
86848966
PM
1162struct rcu_node;
1163
8dc85d54
PZ
1164enum perf_event_task_context {
1165 perf_invalid_context = -1,
1166 perf_hw_context = 0,
89a1e187 1167 perf_sw_context,
8dc85d54
PZ
1168 perf_nr_task_contexts,
1169};
1170
1da177e4
LT
1171struct task_struct {
1172 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1173 void *stack;
1da177e4 1174 atomic_t usage;
97dc32cd
WC
1175 unsigned int flags; /* per process flags, defined below */
1176 unsigned int ptrace;
1da177e4 1177
2dd73a4f 1178#ifdef CONFIG_SMP
fa14ff4a 1179 struct llist_node wake_entry;
3ca7a440 1180 int on_cpu;
62470419
MW
1181 struct task_struct *last_wakee;
1182 unsigned long wakee_flips;
1183 unsigned long wakee_flip_decay_ts;
ac66f547
PZ
1184
1185 int wake_cpu;
2dd73a4f 1186#endif
fd2f4419 1187 int on_rq;
50e645a8 1188
b29739f9 1189 int prio, static_prio, normal_prio;
c7aceaba 1190 unsigned int rt_priority;
5522d5d5 1191 const struct sched_class *sched_class;
20b8a59f 1192 struct sched_entity se;
fa717060 1193 struct sched_rt_entity rt;
8323f26c
PZ
1194#ifdef CONFIG_CGROUP_SCHED
1195 struct task_group *sched_task_group;
1196#endif
aab03e05 1197 struct sched_dl_entity dl;
1da177e4 1198
e107be36
AK
1199#ifdef CONFIG_PREEMPT_NOTIFIERS
1200 /* list of struct preempt_notifier: */
1201 struct hlist_head preempt_notifiers;
1202#endif
1203
6c5c9341 1204#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1205 unsigned int btrace_seq;
6c5c9341 1206#endif
1da177e4 1207
97dc32cd 1208 unsigned int policy;
29baa747 1209 int nr_cpus_allowed;
1da177e4 1210 cpumask_t cpus_allowed;
1da177e4 1211
a57eb940 1212#ifdef CONFIG_PREEMPT_RCU
e260be67 1213 int rcu_read_lock_nesting;
f41d911f 1214 char rcu_read_unlock_special;
f41d911f 1215 struct list_head rcu_node_entry;
a57eb940
PM
1216#endif /* #ifdef CONFIG_PREEMPT_RCU */
1217#ifdef CONFIG_TREE_PREEMPT_RCU
1218 struct rcu_node *rcu_blocked_node;
f41d911f 1219#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1220#ifdef CONFIG_RCU_BOOST
1221 struct rt_mutex *rcu_boost_mutex;
1222#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1223
52f17b6c 1224#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1225 struct sched_info sched_info;
1226#endif
1227
1228 struct list_head tasks;
806c09a7 1229#ifdef CONFIG_SMP
917b627d 1230 struct plist_node pushable_tasks;
1baca4ce 1231 struct rb_node pushable_dl_tasks;
806c09a7 1232#endif
1da177e4
LT
1233
1234 struct mm_struct *mm, *active_mm;
4471a675
JK
1235#ifdef CONFIG_COMPAT_BRK
1236 unsigned brk_randomized:1;
1237#endif
34e55232
KH
1238#if defined(SPLIT_RSS_COUNTING)
1239 struct task_rss_stat rss_stat;
1240#endif
1da177e4 1241/* task state */
97dc32cd 1242 int exit_state;
1da177e4
LT
1243 int exit_code, exit_signal;
1244 int pdeath_signal; /* The signal sent when the parent dies */
a8f072c1 1245 unsigned int jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
1246
1247 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 1248 unsigned int personality;
9b89f6ba 1249
f9ce1f1c
KT
1250 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1251 * execve */
8f0dfc34
AV
1252 unsigned in_iowait:1;
1253
259e5e6c
AL
1254 /* task may not gain privileges */
1255 unsigned no_new_privs:1;
ca94c442
LP
1256
1257 /* Revert to default priority/policy when forking */
1258 unsigned sched_reset_on_fork:1;
a8e4f2ea 1259 unsigned sched_contributes_to_load:1;
ca94c442 1260
1da177e4
LT
1261 pid_t pid;
1262 pid_t tgid;
0a425405 1263
1314562a 1264#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1265 /* Canary value for the -fstack-protector gcc feature */
1266 unsigned long stack_canary;
1314562a 1267#endif
4d1d61a6 1268 /*
1da177e4 1269 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 1270 * older sibling, respectively. (p->father can be replaced with
f470021a 1271 * p->real_parent->pid)
1da177e4 1272 */
abd63bc3
KC
1273 struct task_struct __rcu *real_parent; /* real parent process */
1274 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1275 /*
f470021a 1276 * children/sibling forms the list of my natural children
1da177e4
LT
1277 */
1278 struct list_head children; /* list of my children */
1279 struct list_head sibling; /* linkage in my parent's children list */
1280 struct task_struct *group_leader; /* threadgroup leader */
1281
f470021a
RM
1282 /*
1283 * ptraced is the list of tasks this task is using ptrace on.
1284 * This includes both natural children and PTRACE_ATTACH targets.
1285 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1286 */
1287 struct list_head ptraced;
1288 struct list_head ptrace_entry;
1289
1da177e4 1290 /* PID/PID hash table linkage. */
92476d7f 1291 struct pid_link pids[PIDTYPE_MAX];
47e65328 1292 struct list_head thread_group;
0c740d0a 1293 struct list_head thread_node;
1da177e4
LT
1294
1295 struct completion *vfork_done; /* for vfork() */
1296 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1297 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1298
c66f08be 1299 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1300 cputime_t gtime;
9fbc42ea 1301#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 1302 struct cputime prev_cputime;
6a61671b
FW
1303#endif
1304#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1305 seqlock_t vtime_seqlock;
1306 unsigned long long vtime_snap;
1307 enum {
1308 VTIME_SLEEPING = 0,
1309 VTIME_USER,
1310 VTIME_SYS,
1311 } vtime_snap_whence;
d99ca3b9 1312#endif
1da177e4 1313 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1314 struct timespec start_time; /* monotonic time */
1315 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1316/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1317 unsigned long min_flt, maj_flt;
1318
f06febc9 1319 struct task_cputime cputime_expires;
1da177e4
LT
1320 struct list_head cpu_timers[3];
1321
1322/* process credentials */
1b0ba1c9 1323 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1324 * credentials (COW) */
1b0ba1c9 1325 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1326 * credentials (COW) */
36772092
PBG
1327 char comm[TASK_COMM_LEN]; /* executable name excluding path
1328 - access with [gs]et_task_comm (which lock
1329 it with task_lock())
221af7f8 1330 - initialized normally by setup_new_exec */
1da177e4
LT
1331/* file system info */
1332 int link_count, total_link_count;
3d5b6fcc 1333#ifdef CONFIG_SYSVIPC
1da177e4
LT
1334/* ipc stuff */
1335 struct sysv_sem sysvsem;
3d5b6fcc 1336#endif
e162b39a 1337#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1338/* hung task detection */
82a1fcb9
IM
1339 unsigned long last_switch_count;
1340#endif
1da177e4
LT
1341/* CPU-specific state of this task */
1342 struct thread_struct thread;
1343/* filesystem information */
1344 struct fs_struct *fs;
1345/* open file information */
1346 struct files_struct *files;
1651e14e 1347/* namespaces */
ab516013 1348 struct nsproxy *nsproxy;
1da177e4
LT
1349/* signal handlers */
1350 struct signal_struct *signal;
1351 struct sighand_struct *sighand;
1352
1353 sigset_t blocked, real_blocked;
f3de272b 1354 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1355 struct sigpending pending;
1356
1357 unsigned long sas_ss_sp;
1358 size_t sas_ss_size;
1359 int (*notifier)(void *priv);
1360 void *notifier_data;
1361 sigset_t *notifier_mask;
67d12145 1362 struct callback_head *task_works;
e73f8959 1363
1da177e4 1364 struct audit_context *audit_context;
bfef93a5 1365#ifdef CONFIG_AUDITSYSCALL
e1760bd5 1366 kuid_t loginuid;
4746ec5b 1367 unsigned int sessionid;
bfef93a5 1368#endif
932ecebb 1369 struct seccomp seccomp;
1da177e4
LT
1370
1371/* Thread group tracking */
1372 u32 parent_exec_id;
1373 u32 self_exec_id;
58568d2a
MX
1374/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1375 * mempolicy */
1da177e4 1376 spinlock_t alloc_lock;
1da177e4 1377
b29739f9 1378 /* Protection of the PI data structures: */
1d615482 1379 raw_spinlock_t pi_lock;
b29739f9 1380
23f78d4a
IM
1381#ifdef CONFIG_RT_MUTEXES
1382 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
1383 struct rb_root pi_waiters;
1384 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
1385 /* Deadlock detection and priority inheritance handling */
1386 struct rt_mutex_waiter *pi_blocked_on;
2d3d891d
DF
1387 /* Top pi_waiters task */
1388 struct task_struct *pi_top_task;
23f78d4a
IM
1389#endif
1390
408894ee
IM
1391#ifdef CONFIG_DEBUG_MUTEXES
1392 /* mutex deadlock detection */
1393 struct mutex_waiter *blocked_on;
1394#endif
de30a2b3
IM
1395#ifdef CONFIG_TRACE_IRQFLAGS
1396 unsigned int irq_events;
de30a2b3 1397 unsigned long hardirq_enable_ip;
de30a2b3 1398 unsigned long hardirq_disable_ip;
fa1452e8 1399 unsigned int hardirq_enable_event;
de30a2b3 1400 unsigned int hardirq_disable_event;
fa1452e8
HS
1401 int hardirqs_enabled;
1402 int hardirq_context;
de30a2b3 1403 unsigned long softirq_disable_ip;
de30a2b3 1404 unsigned long softirq_enable_ip;
fa1452e8 1405 unsigned int softirq_disable_event;
de30a2b3 1406 unsigned int softirq_enable_event;
fa1452e8 1407 int softirqs_enabled;
de30a2b3
IM
1408 int softirq_context;
1409#endif
fbb9ce95 1410#ifdef CONFIG_LOCKDEP
bdb9441e 1411# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1412 u64 curr_chain_key;
1413 int lockdep_depth;
fbb9ce95 1414 unsigned int lockdep_recursion;
c7aceaba 1415 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1416 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1417#endif
408894ee 1418
1da177e4
LT
1419/* journalling filesystem info */
1420 void *journal_info;
1421
d89d8796 1422/* stacked block device info */
bddd87c7 1423 struct bio_list *bio_list;
d89d8796 1424
73c10101
JA
1425#ifdef CONFIG_BLOCK
1426/* stack plugging */
1427 struct blk_plug *plug;
1428#endif
1429
1da177e4
LT
1430/* VM state */
1431 struct reclaim_state *reclaim_state;
1432
1da177e4
LT
1433 struct backing_dev_info *backing_dev_info;
1434
1435 struct io_context *io_context;
1436
1437 unsigned long ptrace_message;
1438 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1439 struct task_io_accounting ioac;
8f0ab514 1440#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1441 u64 acct_rss_mem1; /* accumulated rss usage */
1442 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1443 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1444#endif
1445#ifdef CONFIG_CPUSETS
58568d2a 1446 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 1447 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 1448 int cpuset_mem_spread_rotor;
6adef3eb 1449 int cpuset_slab_spread_rotor;
1da177e4 1450#endif
ddbcc7e8 1451#ifdef CONFIG_CGROUPS
817929ec 1452 /* Control Group info protected by css_set_lock */
2c392b8c 1453 struct css_set __rcu *cgroups;
817929ec
PM
1454 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1455 struct list_head cg_list;
ddbcc7e8 1456#endif
42b2dd0a 1457#ifdef CONFIG_FUTEX
0771dfef 1458 struct robust_list_head __user *robust_list;
34f192c6
IM
1459#ifdef CONFIG_COMPAT
1460 struct compat_robust_list_head __user *compat_robust_list;
1461#endif
c87e2837
IM
1462 struct list_head pi_state_list;
1463 struct futex_pi_state *pi_state_cache;
c7aceaba 1464#endif
cdd6c482 1465#ifdef CONFIG_PERF_EVENTS
8dc85d54 1466 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1467 struct mutex perf_event_mutex;
1468 struct list_head perf_event_list;
a63eaf34 1469#endif
8f47b187
TG
1470#ifdef CONFIG_DEBUG_PREEMPT
1471 unsigned long preempt_disable_ip;
1472#endif
c7aceaba 1473#ifdef CONFIG_NUMA
58568d2a 1474 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1475 short il_next;
207205a2 1476 short pref_node_fork;
42b2dd0a 1477#endif
cbee9f88
PZ
1478#ifdef CONFIG_NUMA_BALANCING
1479 int numa_scan_seq;
cbee9f88 1480 unsigned int numa_scan_period;
598f0ec0 1481 unsigned int numa_scan_period_max;
de1c9ce6 1482 int numa_preferred_nid;
6b9a7460 1483 unsigned long numa_migrate_retry;
cbee9f88 1484 u64 node_stamp; /* migration stamp */
7e2703e6
RR
1485 u64 last_task_numa_placement;
1486 u64 last_sum_exec_runtime;
cbee9f88 1487 struct callback_head numa_work;
f809ca9a 1488
8c8a743c
PZ
1489 struct list_head numa_entry;
1490 struct numa_group *numa_group;
1491
745d6147
MG
1492 /*
1493 * Exponential decaying average of faults on a per-node basis.
1494 * Scheduling placement decisions are made based on the these counts.
1495 * The values remain static for the duration of a PTE scan
1496 */
ff1df896 1497 unsigned long *numa_faults_memory;
83e1d2cd 1498 unsigned long total_numa_faults;
745d6147
MG
1499
1500 /*
1501 * numa_faults_buffer records faults per node during the current
ff1df896
RR
1502 * scan window. When the scan completes, the counts in
1503 * numa_faults_memory decay and these values are copied.
745d6147 1504 */
ff1df896 1505 unsigned long *numa_faults_buffer_memory;
745d6147 1506
50ec8a40
RR
1507 /*
1508 * Track the nodes the process was running on when a NUMA hinting
1509 * fault was incurred.
1510 */
1511 unsigned long *numa_faults_cpu;
1512 unsigned long *numa_faults_buffer_cpu;
1513
04bb2f94
RR
1514 /*
1515 * numa_faults_locality tracks if faults recorded during the last
1516 * scan window were remote/local. The task scan period is adapted
1517 * based on the locality of the faults with different weights
1518 * depending on whether they were shared or private faults
1519 */
1520 unsigned long numa_faults_locality[2];
1521
b32e86b4 1522 unsigned long numa_pages_migrated;
cbee9f88
PZ
1523#endif /* CONFIG_NUMA_BALANCING */
1524
e56d0903 1525 struct rcu_head rcu;
b92ce558
JA
1526
1527 /*
1528 * cache last used pipe for splice
1529 */
1530 struct pipe_inode_info *splice_pipe;
5640f768
ED
1531
1532 struct page_frag task_frag;
1533
ca74e92b
SN
1534#ifdef CONFIG_TASK_DELAY_ACCT
1535 struct task_delay_info *delays;
f4f154fd
AM
1536#endif
1537#ifdef CONFIG_FAULT_INJECTION
1538 int make_it_fail;
ca74e92b 1539#endif
9d823e8f
WF
1540 /*
1541 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
1542 * balance_dirty_pages() for some dirty throttling pause
1543 */
1544 int nr_dirtied;
1545 int nr_dirtied_pause;
83712358 1546 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 1547
9745512c
AV
1548#ifdef CONFIG_LATENCYTOP
1549 int latency_record_count;
1550 struct latency_record latency_record[LT_SAVECOUNT];
1551#endif
6976675d
AV
1552 /*
1553 * time slack values; these are used to round up poll() and
1554 * select() etc timeout values. These are in nanoseconds.
1555 */
1556 unsigned long timer_slack_ns;
1557 unsigned long default_timer_slack_ns;
f8d570a4 1558
fb52607a 1559#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1560 /* Index of current stored address in ret_stack */
f201ae23
FW
1561 int curr_ret_stack;
1562 /* Stack of return addresses for return function tracing */
1563 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1564 /* time stamp for last schedule */
1565 unsigned long long ftrace_timestamp;
f201ae23
FW
1566 /*
1567 * Number of functions that haven't been traced
1568 * because of depth overrun.
1569 */
1570 atomic_t trace_overrun;
380c4b14
FW
1571 /* Pause for the tracing */
1572 atomic_t tracing_graph_pause;
f201ae23 1573#endif
ea4e2bc4
SR
1574#ifdef CONFIG_TRACING
1575 /* state flags for use by tracers */
1576 unsigned long trace;
b1cff0ad 1577 /* bitmask and counter of trace recursion */
261842b7
SR
1578 unsigned long trace_recursion;
1579#endif /* CONFIG_TRACING */
c255a458 1580#ifdef CONFIG_MEMCG /* memcg uses this to do batch job */
569b846d
KH
1581 struct memcg_batch_info {
1582 int do_batch; /* incremented when batch uncharge started */
1583 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1584 unsigned long nr_pages; /* uncharged usage */
1585 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d 1586 } memcg_batch;
0e9d92f2 1587 unsigned int memcg_kmem_skip_account;
519e5247 1588 struct memcg_oom_info {
49426420
JW
1589 struct mem_cgroup *memcg;
1590 gfp_t gfp_mask;
1591 int order;
519e5247
JW
1592 unsigned int may_oom:1;
1593 } memcg_oom;
569b846d 1594#endif
0326f5a9
SD
1595#ifdef CONFIG_UPROBES
1596 struct uprobe_task *utask;
0326f5a9 1597#endif
cafe5635
KO
1598#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
1599 unsigned int sequential_io;
1600 unsigned int sequential_io_avg;
1601#endif
1da177e4
LT
1602};
1603
76e6eee0 1604/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1605#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1606
6688cc05
PZ
1607#define TNF_MIGRATED 0x01
1608#define TNF_NO_GROUP 0x02
dabe1d99 1609#define TNF_SHARED 0x04
04bb2f94 1610#define TNF_FAULT_LOCAL 0x08
6688cc05 1611
cbee9f88 1612#ifdef CONFIG_NUMA_BALANCING
6688cc05 1613extern void task_numa_fault(int last_node, int node, int pages, int flags);
e29cf08b 1614extern pid_t task_numa_group_id(struct task_struct *p);
1a687c2e 1615extern void set_numabalancing_state(bool enabled);
82727018 1616extern void task_numa_free(struct task_struct *p);
10f39042
RR
1617extern bool should_numa_migrate_memory(struct task_struct *p, struct page *page,
1618 int src_nid, int dst_cpu);
cbee9f88 1619#else
ac8e895b 1620static inline void task_numa_fault(int last_node, int node, int pages,
6688cc05 1621 int flags)
cbee9f88
PZ
1622{
1623}
e29cf08b
MG
1624static inline pid_t task_numa_group_id(struct task_struct *p)
1625{
1626 return 0;
1627}
1a687c2e
MG
1628static inline void set_numabalancing_state(bool enabled)
1629{
1630}
82727018
RR
1631static inline void task_numa_free(struct task_struct *p)
1632{
1633}
10f39042
RR
1634static inline bool should_numa_migrate_memory(struct task_struct *p,
1635 struct page *page, int src_nid, int dst_cpu)
1636{
1637 return true;
1638}
cbee9f88
PZ
1639#endif
1640
e868171a 1641static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1642{
1643 return task->pids[PIDTYPE_PID].pid;
1644}
1645
e868171a 1646static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1647{
1648 return task->group_leader->pids[PIDTYPE_PID].pid;
1649}
1650
6dda81f4
ON
1651/*
1652 * Without tasklist or rcu lock it is not safe to dereference
1653 * the result of task_pgrp/task_session even if task == current,
1654 * we can race with another thread doing sys_setsid/sys_setpgid.
1655 */
e868171a 1656static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1657{
1658 return task->group_leader->pids[PIDTYPE_PGID].pid;
1659}
1660
e868171a 1661static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1662{
1663 return task->group_leader->pids[PIDTYPE_SID].pid;
1664}
1665
7af57294
PE
1666struct pid_namespace;
1667
1668/*
1669 * the helpers to get the task's different pids as they are seen
1670 * from various namespaces
1671 *
1672 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1673 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1674 * current.
7af57294
PE
1675 * task_xid_nr_ns() : id seen from the ns specified;
1676 *
1677 * set_task_vxid() : assigns a virtual id to a task;
1678 *
7af57294
PE
1679 * see also pid_nr() etc in include/linux/pid.h
1680 */
52ee2dfd
ON
1681pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1682 struct pid_namespace *ns);
7af57294 1683
e868171a 1684static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1685{
1686 return tsk->pid;
1687}
1688
52ee2dfd
ON
1689static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1690 struct pid_namespace *ns)
1691{
1692 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1693}
7af57294
PE
1694
1695static inline pid_t task_pid_vnr(struct task_struct *tsk)
1696{
52ee2dfd 1697 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1698}
1699
1700
e868171a 1701static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1702{
1703 return tsk->tgid;
1704}
1705
2f2a3a46 1706pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1707
1708static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1709{
1710 return pid_vnr(task_tgid(tsk));
1711}
1712
1713
52ee2dfd
ON
1714static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1715 struct pid_namespace *ns)
7af57294 1716{
52ee2dfd 1717 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1718}
1719
7af57294
PE
1720static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1721{
52ee2dfd 1722 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1723}
1724
1725
52ee2dfd
ON
1726static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1727 struct pid_namespace *ns)
7af57294 1728{
52ee2dfd 1729 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1730}
1731
7af57294
PE
1732static inline pid_t task_session_vnr(struct task_struct *tsk)
1733{
52ee2dfd 1734 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1735}
1736
1b0f7ffd
ON
1737/* obsolete, do not use */
1738static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1739{
1740 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1741}
7af57294 1742
1da177e4
LT
1743/**
1744 * pid_alive - check that a task structure is not stale
1745 * @p: Task structure to be checked.
1746 *
1747 * Test if a process is not yet dead (at most zombie state)
1748 * If pid_alive fails, then pointers within the task structure
1749 * can be stale and must not be dereferenced.
e69f6186
YB
1750 *
1751 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1752 */
e868171a 1753static inline int pid_alive(struct task_struct *p)
1da177e4 1754{
92476d7f 1755 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1756}
1757
f400e198 1758/**
b460cbc5 1759 * is_global_init - check if a task structure is init
3260259f
H
1760 * @tsk: Task structure to be checked.
1761 *
1762 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1763 *
1764 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1765 */
e868171a 1766static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1767{
1768 return tsk->pid == 1;
1769}
b460cbc5 1770
9ec52099
CLG
1771extern struct pid *cad_pid;
1772
1da177e4 1773extern void free_task(struct task_struct *tsk);
1da177e4 1774#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1775
158d9ebd 1776extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1777
1778static inline void put_task_struct(struct task_struct *t)
1779{
1780 if (atomic_dec_and_test(&t->usage))
8c7904a0 1781 __put_task_struct(t);
e56d0903 1782}
1da177e4 1783
6a61671b
FW
1784#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1785extern void task_cputime(struct task_struct *t,
1786 cputime_t *utime, cputime_t *stime);
1787extern void task_cputime_scaled(struct task_struct *t,
1788 cputime_t *utimescaled, cputime_t *stimescaled);
1789extern cputime_t task_gtime(struct task_struct *t);
1790#else
6fac4829
FW
1791static inline void task_cputime(struct task_struct *t,
1792 cputime_t *utime, cputime_t *stime)
1793{
1794 if (utime)
1795 *utime = t->utime;
1796 if (stime)
1797 *stime = t->stime;
1798}
1799
1800static inline void task_cputime_scaled(struct task_struct *t,
1801 cputime_t *utimescaled,
1802 cputime_t *stimescaled)
1803{
1804 if (utimescaled)
1805 *utimescaled = t->utimescaled;
1806 if (stimescaled)
1807 *stimescaled = t->stimescaled;
1808}
6a61671b
FW
1809
1810static inline cputime_t task_gtime(struct task_struct *t)
1811{
1812 return t->gtime;
1813}
1814#endif
e80d0a1a
FW
1815extern void task_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
1816extern void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1817
1da177e4
LT
1818/*
1819 * Per process flags
1820 */
1da177e4 1821#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1822#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1823#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1824#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1825#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1826#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1827#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1828#define PF_DUMPCORE 0x00000200 /* dumped core */
1829#define PF_SIGNALED 0x00000400 /* killed by a signal */
1830#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1831#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1832#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1833#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1834#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1835#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1836#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1837#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1838#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1839#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1840#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1841#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1842#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1843#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1844#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
14a40ffc 1845#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1846#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1847#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1848#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1849#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1850#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1851
1852/*
1853 * Only the _current_ task can read/write to tsk->flags, but other
1854 * tasks can access tsk->flags in readonly mode for example
1855 * with tsk_used_math (like during threaded core dumping).
1856 * There is however an exception to this rule during ptrace
1857 * or during fork: the ptracer task is allowed to write to the
1858 * child->flags of its traced child (same goes for fork, the parent
1859 * can write to the child->flags), because we're guaranteed the
1860 * child is not running and in turn not changing child->flags
1861 * at the same time the parent does it.
1862 */
1863#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1864#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1865#define clear_used_math() clear_stopped_child_used_math(current)
1866#define set_used_math() set_stopped_child_used_math(current)
1867#define conditional_stopped_child_used_math(condition, child) \
1868 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1869#define conditional_used_math(condition) \
1870 conditional_stopped_child_used_math(condition, current)
1871#define copy_to_stopped_child_used_math(child) \
1872 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1873/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1874#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1875#define used_math() tsk_used_math(current)
1876
21caf2fc
ML
1877/* __GFP_IO isn't allowed if PF_MEMALLOC_NOIO is set in current->flags */
1878static inline gfp_t memalloc_noio_flags(gfp_t flags)
1879{
1880 if (unlikely(current->flags & PF_MEMALLOC_NOIO))
1881 flags &= ~__GFP_IO;
1882 return flags;
1883}
1884
1885static inline unsigned int memalloc_noio_save(void)
1886{
1887 unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
1888 current->flags |= PF_MEMALLOC_NOIO;
1889 return flags;
1890}
1891
1892static inline void memalloc_noio_restore(unsigned int flags)
1893{
1894 current->flags = (current->flags & ~PF_MEMALLOC_NOIO) | flags;
1895}
1896
e5c1902e 1897/*
a8f072c1 1898 * task->jobctl flags
e5c1902e 1899 */
a8f072c1 1900#define JOBCTL_STOP_SIGMASK 0xffff /* signr of the last group stop */
e5c1902e 1901
a8f072c1
TH
1902#define JOBCTL_STOP_DEQUEUED_BIT 16 /* stop signal dequeued */
1903#define JOBCTL_STOP_PENDING_BIT 17 /* task should stop for group stop */
1904#define JOBCTL_STOP_CONSUME_BIT 18 /* consume group stop count */
73ddff2b 1905#define JOBCTL_TRAP_STOP_BIT 19 /* trap for STOP */
fb1d910c 1906#define JOBCTL_TRAP_NOTIFY_BIT 20 /* trap for NOTIFY */
a8f072c1 1907#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */
544b2c91 1908#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */
a8f072c1
TH
1909
1910#define JOBCTL_STOP_DEQUEUED (1 << JOBCTL_STOP_DEQUEUED_BIT)
1911#define JOBCTL_STOP_PENDING (1 << JOBCTL_STOP_PENDING_BIT)
1912#define JOBCTL_STOP_CONSUME (1 << JOBCTL_STOP_CONSUME_BIT)
73ddff2b 1913#define JOBCTL_TRAP_STOP (1 << JOBCTL_TRAP_STOP_BIT)
fb1d910c 1914#define JOBCTL_TRAP_NOTIFY (1 << JOBCTL_TRAP_NOTIFY_BIT)
a8f072c1 1915#define JOBCTL_TRAPPING (1 << JOBCTL_TRAPPING_BIT)
544b2c91 1916#define JOBCTL_LISTENING (1 << JOBCTL_LISTENING_BIT)
a8f072c1 1917
fb1d910c 1918#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
73ddff2b 1919#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK)
3759a0d9 1920
7dd3db54
TH
1921extern bool task_set_jobctl_pending(struct task_struct *task,
1922 unsigned int mask);
73ddff2b 1923extern void task_clear_jobctl_trapping(struct task_struct *task);
3759a0d9
TH
1924extern void task_clear_jobctl_pending(struct task_struct *task,
1925 unsigned int mask);
39efa3ef 1926
a57eb940 1927#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1928
1929#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1aa03f11 1930#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1931
1932static inline void rcu_copy_process(struct task_struct *p)
1933{
1934 p->rcu_read_lock_nesting = 0;
1935 p->rcu_read_unlock_special = 0;
a57eb940 1936#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1937 p->rcu_blocked_node = NULL;
24278d14
PM
1938#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1939#ifdef CONFIG_RCU_BOOST
1940 p->rcu_boost_mutex = NULL;
1941#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1942 INIT_LIST_HEAD(&p->rcu_node_entry);
1943}
1944
f41d911f
PM
1945#else
1946
1947static inline void rcu_copy_process(struct task_struct *p)
1948{
1949}
1950
1951#endif
1952
907aed48
MG
1953static inline void tsk_restore_flags(struct task_struct *task,
1954 unsigned long orig_flags, unsigned long flags)
1955{
1956 task->flags &= ~flags;
1957 task->flags |= orig_flags & flags;
1958}
1959
1da177e4 1960#ifdef CONFIG_SMP
1e1b6c51
KM
1961extern void do_set_cpus_allowed(struct task_struct *p,
1962 const struct cpumask *new_mask);
1963
cd8ba7cd 1964extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1965 const struct cpumask *new_mask);
1da177e4 1966#else
1e1b6c51
KM
1967static inline void do_set_cpus_allowed(struct task_struct *p,
1968 const struct cpumask *new_mask)
1969{
1970}
cd8ba7cd 1971static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1972 const struct cpumask *new_mask)
1da177e4 1973{
96f874e2 1974 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1975 return -EINVAL;
1976 return 0;
1977}
1978#endif
e0ad9556 1979
3451d024 1980#ifdef CONFIG_NO_HZ_COMMON
5167e8d5
PZ
1981void calc_load_enter_idle(void);
1982void calc_load_exit_idle(void);
1983#else
1984static inline void calc_load_enter_idle(void) { }
1985static inline void calc_load_exit_idle(void) { }
3451d024 1986#endif /* CONFIG_NO_HZ_COMMON */
5167e8d5 1987
e0ad9556 1988#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1989static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1990{
1991 return set_cpus_allowed_ptr(p, &new_mask);
1992}
e0ad9556 1993#endif
1da177e4 1994
b342501c 1995/*
c676329a
PZ
1996 * Do not use outside of architecture code which knows its limitations.
1997 *
1998 * sched_clock() has no promise of monotonicity or bounded drift between
1999 * CPUs, use (which you should not) requires disabling IRQs.
2000 *
2001 * Please use one of the three interfaces below.
b342501c 2002 */
1bbfa6f2 2003extern unsigned long long notrace sched_clock(void);
c676329a 2004/*
489a71b0 2005 * See the comment in kernel/sched/clock.c
c676329a
PZ
2006 */
2007extern u64 cpu_clock(int cpu);
2008extern u64 local_clock(void);
2009extern u64 sched_clock_cpu(int cpu);
2010
e436d800 2011
c1955a3d 2012extern void sched_clock_init(void);
3e51f33f 2013
c1955a3d 2014#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
2015static inline void sched_clock_tick(void)
2016{
2017}
2018
2019static inline void sched_clock_idle_sleep_event(void)
2020{
2021}
2022
2023static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
2024{
2025}
2026#else
c676329a
PZ
2027/*
2028 * Architectures can set this to 1 if they have specified
2029 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
2030 * but then during bootup it turns out that sched_clock()
2031 * is reliable after all:
2032 */
35af99e6
PZ
2033extern int sched_clock_stable(void);
2034extern void set_sched_clock_stable(void);
2035extern void clear_sched_clock_stable(void);
c676329a 2036
3e51f33f
PZ
2037extern void sched_clock_tick(void);
2038extern void sched_clock_idle_sleep_event(void);
2039extern void sched_clock_idle_wakeup_event(u64 delta_ns);
2040#endif
2041
b52bfee4
VP
2042#ifdef CONFIG_IRQ_TIME_ACCOUNTING
2043/*
2044 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
2045 * The reason for this explicit opt-in is not to have perf penalty with
2046 * slow sched_clocks.
2047 */
2048extern void enable_sched_clock_irqtime(void);
2049extern void disable_sched_clock_irqtime(void);
2050#else
2051static inline void enable_sched_clock_irqtime(void) {}
2052static inline void disable_sched_clock_irqtime(void) {}
2053#endif
2054
36c8b586 2055extern unsigned long long
41b86e9c 2056task_sched_runtime(struct task_struct *task);
1da177e4
LT
2057
2058/* sched_exec is called by processes performing an exec */
2059#ifdef CONFIG_SMP
2060extern void sched_exec(void);
2061#else
2062#define sched_exec() {}
2063#endif
2064
2aa44d05
IM
2065extern void sched_clock_idle_sleep_event(void);
2066extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 2067
1da177e4
LT
2068#ifdef CONFIG_HOTPLUG_CPU
2069extern void idle_task_exit(void);
2070#else
2071static inline void idle_task_exit(void) {}
2072#endif
2073
3451d024 2074#if defined(CONFIG_NO_HZ_COMMON) && defined(CONFIG_SMP)
1c20091e 2075extern void wake_up_nohz_cpu(int cpu);
06d8308c 2076#else
1c20091e 2077static inline void wake_up_nohz_cpu(int cpu) { }
06d8308c
TG
2078#endif
2079
ce831b38
FW
2080#ifdef CONFIG_NO_HZ_FULL
2081extern bool sched_can_stop_tick(void);
265f22a9 2082extern u64 scheduler_tick_max_deferment(void);
ce831b38
FW
2083#else
2084static inline bool sched_can_stop_tick(void) { return false; }
06d8308c
TG
2085#endif
2086
5091faa4 2087#ifdef CONFIG_SCHED_AUTOGROUP
5091faa4
MG
2088extern void sched_autogroup_create_attach(struct task_struct *p);
2089extern void sched_autogroup_detach(struct task_struct *p);
2090extern void sched_autogroup_fork(struct signal_struct *sig);
2091extern void sched_autogroup_exit(struct signal_struct *sig);
2092#ifdef CONFIG_PROC_FS
2093extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
2e5b5b3a 2094extern int proc_sched_autogroup_set_nice(struct task_struct *p, int nice);
5091faa4
MG
2095#endif
2096#else
2097static inline void sched_autogroup_create_attach(struct task_struct *p) { }
2098static inline void sched_autogroup_detach(struct task_struct *p) { }
2099static inline void sched_autogroup_fork(struct signal_struct *sig) { }
2100static inline void sched_autogroup_exit(struct signal_struct *sig) { }
2101#endif
2102
d95f4122 2103extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
2104extern void set_user_nice(struct task_struct *p, long nice);
2105extern int task_prio(const struct task_struct *p);
d0ea0268
DY
2106/**
2107 * task_nice - return the nice value of a given task.
2108 * @p: the task in question.
2109 *
2110 * Return: The nice value [ -20 ... 0 ... 19 ].
2111 */
2112static inline int task_nice(const struct task_struct *p)
2113{
2114 return PRIO_TO_NICE((p)->static_prio);
2115}
36c8b586
IM
2116extern int can_nice(const struct task_struct *p, const int nice);
2117extern int task_curr(const struct task_struct *p);
1da177e4 2118extern int idle_cpu(int cpu);
fe7de49f
KM
2119extern int sched_setscheduler(struct task_struct *, int,
2120 const struct sched_param *);
961ccddd 2121extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 2122 const struct sched_param *);
d50dde5a
DF
2123extern int sched_setattr(struct task_struct *,
2124 const struct sched_attr *);
36c8b586 2125extern struct task_struct *idle_task(int cpu);
c4f30608
PM
2126/**
2127 * is_idle_task - is the specified task an idle task?
fa757281 2128 * @p: the task in question.
e69f6186
YB
2129 *
2130 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 2131 */
7061ca3b 2132static inline bool is_idle_task(const struct task_struct *p)
c4f30608
PM
2133{
2134 return p->pid == 0;
2135}
36c8b586
IM
2136extern struct task_struct *curr_task(int cpu);
2137extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
2138
2139void yield(void);
2140
2141/*
2142 * The default (Linux) execution domain.
2143 */
2144extern struct exec_domain default_exec_domain;
2145
2146union thread_union {
2147 struct thread_info thread_info;
2148 unsigned long stack[THREAD_SIZE/sizeof(long)];
2149};
2150
2151#ifndef __HAVE_ARCH_KSTACK_END
2152static inline int kstack_end(void *addr)
2153{
2154 /* Reliable end of stack detection:
2155 * Some APM bios versions misalign the stack
2156 */
2157 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2158}
2159#endif
2160
2161extern union thread_union init_thread_union;
2162extern struct task_struct init_task;
2163
2164extern struct mm_struct init_mm;
2165
198fe21b
PE
2166extern struct pid_namespace init_pid_ns;
2167
2168/*
2169 * find a task by one of its numerical ids
2170 *
198fe21b
PE
2171 * find_task_by_pid_ns():
2172 * finds a task by its pid in the specified namespace
228ebcbe
PE
2173 * find_task_by_vpid():
2174 * finds a task by its virtual pid
198fe21b 2175 *
e49859e7 2176 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2177 */
2178
228ebcbe
PE
2179extern struct task_struct *find_task_by_vpid(pid_t nr);
2180extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2181 struct pid_namespace *ns);
198fe21b 2182
1da177e4 2183/* per-UID process charging. */
7b44ab97 2184extern struct user_struct * alloc_uid(kuid_t);
1da177e4
LT
2185static inline struct user_struct *get_uid(struct user_struct *u)
2186{
2187 atomic_inc(&u->__count);
2188 return u;
2189}
2190extern void free_uid(struct user_struct *);
1da177e4
LT
2191
2192#include <asm/current.h>
2193
f0af911a 2194extern void xtime_update(unsigned long ticks);
1da177e4 2195
b3c97528
HH
2196extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2197extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 2198extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
2199#ifdef CONFIG_SMP
2200 extern void kick_process(struct task_struct *tsk);
2201#else
2202 static inline void kick_process(struct task_struct *tsk) { }
2203#endif
aab03e05 2204extern int sched_fork(unsigned long clone_flags, struct task_struct *p);
ad46c2c4 2205extern void sched_dead(struct task_struct *p);
1da177e4 2206
1da177e4
LT
2207extern void proc_caches_init(void);
2208extern void flush_signals(struct task_struct *);
3bcac026 2209extern void __flush_signals(struct task_struct *);
10ab825b 2210extern void ignore_signals(struct task_struct *);
1da177e4
LT
2211extern void flush_signal_handlers(struct task_struct *, int force_default);
2212extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2213
2214static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2215{
2216 unsigned long flags;
2217 int ret;
2218
2219 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2220 ret = dequeue_signal(tsk, mask, info);
2221 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2222
2223 return ret;
53c8f9f1 2224}
1da177e4
LT
2225
2226extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2227 sigset_t *mask);
2228extern void unblock_all_signals(void);
2229extern void release_task(struct task_struct * p);
2230extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2231extern int force_sigsegv(int, struct task_struct *);
2232extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2233extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2234extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
d178bc3a
SH
2235extern int kill_pid_info_as_cred(int, struct siginfo *, struct pid *,
2236 const struct cred *, u32);
c4b92fc1
EB
2237extern int kill_pgrp(struct pid *pid, int sig, int priv);
2238extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2239extern int kill_proc_info(int, struct siginfo *, pid_t);
86773473 2240extern __must_check bool do_notify_parent(struct task_struct *, int);
a7f0765e 2241extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2242extern void force_sig(int, struct task_struct *);
1da177e4 2243extern int send_sig(int, struct task_struct *, int);
09faef11 2244extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2245extern struct sigqueue *sigqueue_alloc(void);
2246extern void sigqueue_free(struct sigqueue *);
ac5c2153 2247extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2248extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4 2249
51a7b448
AV
2250static inline void restore_saved_sigmask(void)
2251{
2252 if (test_and_clear_restore_sigmask())
77097ae5 2253 __set_current_blocked(&current->saved_sigmask);
51a7b448
AV
2254}
2255
b7f9a11a
AV
2256static inline sigset_t *sigmask_to_save(void)
2257{
2258 sigset_t *res = &current->blocked;
2259 if (unlikely(test_restore_sigmask()))
2260 res = &current->saved_sigmask;
2261 return res;
2262}
2263
9ec52099
CLG
2264static inline int kill_cad_pid(int sig, int priv)
2265{
2266 return kill_pid(cad_pid, sig, priv);
2267}
2268
1da177e4
LT
2269/* These can be the second arg to send_sig_info/send_group_sig_info. */
2270#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2271#define SEND_SIG_PRIV ((struct siginfo *) 1)
2272#define SEND_SIG_FORCED ((struct siginfo *) 2)
2273
2a855dd0
SAS
2274/*
2275 * True if we are on the alternate signal stack.
2276 */
1da177e4
LT
2277static inline int on_sig_stack(unsigned long sp)
2278{
2a855dd0
SAS
2279#ifdef CONFIG_STACK_GROWSUP
2280 return sp >= current->sas_ss_sp &&
2281 sp - current->sas_ss_sp < current->sas_ss_size;
2282#else
2283 return sp > current->sas_ss_sp &&
2284 sp - current->sas_ss_sp <= current->sas_ss_size;
2285#endif
1da177e4
LT
2286}
2287
2288static inline int sas_ss_flags(unsigned long sp)
2289{
2290 return (current->sas_ss_size == 0 ? SS_DISABLE
2291 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2292}
2293
5a1b98d3
AV
2294static inline unsigned long sigsp(unsigned long sp, struct ksignal *ksig)
2295{
2296 if (unlikely((ksig->ka.sa.sa_flags & SA_ONSTACK)) && ! sas_ss_flags(sp))
2297#ifdef CONFIG_STACK_GROWSUP
2298 return current->sas_ss_sp;
2299#else
2300 return current->sas_ss_sp + current->sas_ss_size;
2301#endif
2302 return sp;
2303}
2304
1da177e4
LT
2305/*
2306 * Routines for handling mm_structs
2307 */
2308extern struct mm_struct * mm_alloc(void);
2309
2310/* mmdrop drops the mm and the page tables */
b3c97528 2311extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2312static inline void mmdrop(struct mm_struct * mm)
2313{
6fb43d7b 2314 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2315 __mmdrop(mm);
2316}
2317
2318/* mmput gets rid of the mappings and all user-space */
2319extern void mmput(struct mm_struct *);
2320/* Grab a reference to a task's mm, if it is not already going away */
2321extern struct mm_struct *get_task_mm(struct task_struct *task);
8cdb878d
CY
2322/*
2323 * Grab a reference to a task's mm, if it is not already going away
2324 * and ptrace_may_access with the mode parameter passed to it
2325 * succeeds.
2326 */
2327extern struct mm_struct *mm_access(struct task_struct *task, unsigned int mode);
1da177e4
LT
2328/* Remove the current tasks stale references to the old mm_struct */
2329extern void mm_release(struct task_struct *, struct mm_struct *);
2330
6f2c55b8 2331extern int copy_thread(unsigned long, unsigned long, unsigned long,
afa86fc4 2332 struct task_struct *);
1da177e4
LT
2333extern void flush_thread(void);
2334extern void exit_thread(void);
2335
1da177e4 2336extern void exit_files(struct task_struct *);
a7e5328a 2337extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2338
1da177e4 2339extern void exit_itimers(struct signal_struct *);
cbaffba1 2340extern void flush_itimer_signals(void);
1da177e4 2341
9402c95f 2342extern void do_group_exit(int);
1da177e4 2343
1da177e4
LT
2344extern int allow_signal(int);
2345extern int disallow_signal(int);
1da177e4 2346
c4ad8f98 2347extern int do_execve(struct filename *,
d7627467 2348 const char __user * const __user *,
da3d4c5f 2349 const char __user * const __user *);
e80d6661 2350extern long do_fork(unsigned long, unsigned long, unsigned long, int __user *, int __user *);
36c8b586 2351struct task_struct *fork_idle(int);
2aa3a7f8 2352extern pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
1da177e4
LT
2353
2354extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2355extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2356
2357#ifdef CONFIG_SMP
317f3941 2358void scheduler_ipi(void);
85ba2d86 2359extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2360#else
184748cc 2361static inline void scheduler_ipi(void) { }
85ba2d86
RM
2362static inline unsigned long wait_task_inactive(struct task_struct *p,
2363 long match_state)
2364{
2365 return 1;
2366}
1da177e4
LT
2367#endif
2368
05725f7e
JP
2369#define next_task(p) \
2370 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2371
2372#define for_each_process(p) \
2373 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2374
5bb459bb 2375extern bool current_is_single_threaded(void);
d84f4f99 2376
1da177e4
LT
2377/*
2378 * Careful: do_each_thread/while_each_thread is a double loop so
2379 * 'break' will not work as expected - use goto instead.
2380 */
2381#define do_each_thread(g, t) \
2382 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2383
2384#define while_each_thread(g, t) \
2385 while ((t = next_thread(t)) != g)
2386
0c740d0a
ON
2387#define __for_each_thread(signal, t) \
2388 list_for_each_entry_rcu(t, &(signal)->thread_head, thread_node)
2389
2390#define for_each_thread(p, t) \
2391 __for_each_thread((p)->signal, t)
2392
2393/* Careful: this is a double loop, 'break' won't work as expected. */
2394#define for_each_process_thread(p, t) \
2395 for_each_process(p) for_each_thread(p, t)
2396
7e49827c
ON
2397static inline int get_nr_threads(struct task_struct *tsk)
2398{
b3ac022c 2399 return tsk->signal->nr_threads;
7e49827c
ON
2400}
2401
087806b1
ON
2402static inline bool thread_group_leader(struct task_struct *p)
2403{
2404 return p->exit_signal >= 0;
2405}
1da177e4 2406
0804ef4b
EB
2407/* Do to the insanities of de_thread it is possible for a process
2408 * to have the pid of the thread group leader without actually being
2409 * the thread group leader. For iteration through the pids in proc
2410 * all we care about is that we have a task with the appropriate
2411 * pid, we don't actually care if we have the right task.
2412 */
e1403b8e 2413static inline bool has_group_leader_pid(struct task_struct *p)
0804ef4b 2414{
e1403b8e 2415 return task_pid(p) == p->signal->leader_pid;
0804ef4b
EB
2416}
2417
bac0abd6 2418static inline
e1403b8e 2419bool same_thread_group(struct task_struct *p1, struct task_struct *p2)
bac0abd6 2420{
e1403b8e 2421 return p1->signal == p2->signal;
bac0abd6
PE
2422}
2423
36c8b586 2424static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2425{
05725f7e
JP
2426 return list_entry_rcu(p->thread_group.next,
2427 struct task_struct, thread_group);
47e65328
ON
2428}
2429
e868171a 2430static inline int thread_group_empty(struct task_struct *p)
1da177e4 2431{
47e65328 2432 return list_empty(&p->thread_group);
1da177e4
LT
2433}
2434
2435#define delay_group_leader(p) \
2436 (thread_group_leader(p) && !thread_group_empty(p))
2437
1da177e4 2438/*
260ea101 2439 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2440 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8 2441 * pins the final release of task.io_context. Also protects ->cpuset and
d68b46fe 2442 * ->cgroup.subsys[]. And ->vfork_done.
1da177e4
LT
2443 *
2444 * Nests both inside and outside of read_lock(&tasklist_lock).
2445 * It must not be nested with write_lock_irq(&tasklist_lock),
2446 * neither inside nor outside.
2447 */
2448static inline void task_lock(struct task_struct *p)
2449{
2450 spin_lock(&p->alloc_lock);
2451}
2452
2453static inline void task_unlock(struct task_struct *p)
2454{
2455 spin_unlock(&p->alloc_lock);
2456}
2457
b8ed374e 2458extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2459 unsigned long *flags);
2460
9388dc30
AV
2461static inline struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2462 unsigned long *flags)
2463{
2464 struct sighand_struct *ret;
2465
2466 ret = __lock_task_sighand(tsk, flags);
2467 (void)__cond_lock(&tsk->sighand->siglock, ret);
2468 return ret;
2469}
b8ed374e 2470
f63ee72e
ON
2471static inline void unlock_task_sighand(struct task_struct *tsk,
2472 unsigned long *flags)
2473{
2474 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2475}
2476
4714d1d3 2477#ifdef CONFIG_CGROUPS
257058ae 2478static inline void threadgroup_change_begin(struct task_struct *tsk)
4714d1d3 2479{
257058ae 2480 down_read(&tsk->signal->group_rwsem);
4714d1d3 2481}
257058ae 2482static inline void threadgroup_change_end(struct task_struct *tsk)
4714d1d3 2483{
257058ae 2484 up_read(&tsk->signal->group_rwsem);
4714d1d3 2485}
77e4ef99
TH
2486
2487/**
2488 * threadgroup_lock - lock threadgroup
2489 * @tsk: member task of the threadgroup to lock
2490 *
2491 * Lock the threadgroup @tsk belongs to. No new task is allowed to enter
2492 * and member tasks aren't allowed to exit (as indicated by PF_EXITING) or
e56fb287
ON
2493 * change ->group_leader/pid. This is useful for cases where the threadgroup
2494 * needs to stay stable across blockable operations.
77e4ef99
TH
2495 *
2496 * fork and exit paths explicitly call threadgroup_change_{begin|end}() for
2497 * synchronization. While held, no new task will be added to threadgroup
2498 * and no existing live task will have its PF_EXITING set.
2499 *
e56fb287
ON
2500 * de_thread() does threadgroup_change_{begin|end}() when a non-leader
2501 * sub-thread becomes a new leader.
77e4ef99 2502 */
257058ae 2503static inline void threadgroup_lock(struct task_struct *tsk)
4714d1d3 2504{
257058ae 2505 down_write(&tsk->signal->group_rwsem);
4714d1d3 2506}
77e4ef99
TH
2507
2508/**
2509 * threadgroup_unlock - unlock threadgroup
2510 * @tsk: member task of the threadgroup to unlock
2511 *
2512 * Reverse threadgroup_lock().
2513 */
257058ae 2514static inline void threadgroup_unlock(struct task_struct *tsk)
4714d1d3 2515{
257058ae 2516 up_write(&tsk->signal->group_rwsem);
4714d1d3
BB
2517}
2518#else
257058ae
TH
2519static inline void threadgroup_change_begin(struct task_struct *tsk) {}
2520static inline void threadgroup_change_end(struct task_struct *tsk) {}
2521static inline void threadgroup_lock(struct task_struct *tsk) {}
2522static inline void threadgroup_unlock(struct task_struct *tsk) {}
4714d1d3
BB
2523#endif
2524
f037360f
AV
2525#ifndef __HAVE_THREAD_FUNCTIONS
2526
f7e4217b
RZ
2527#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2528#define task_stack_page(task) ((task)->stack)
a1261f54 2529
10ebffde
AV
2530static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2531{
2532 *task_thread_info(p) = *task_thread_info(org);
2533 task_thread_info(p)->task = p;
2534}
2535
2536static inline unsigned long *end_of_stack(struct task_struct *p)
2537{
f7e4217b 2538 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2539}
2540
f037360f
AV
2541#endif
2542
8b05c7e6
FT
2543static inline int object_is_on_stack(void *obj)
2544{
2545 void *stack = task_stack_page(current);
2546
2547 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2548}
2549
8c9843e5
BH
2550extern void thread_info_cache_init(void);
2551
7c9f8861
ES
2552#ifdef CONFIG_DEBUG_STACK_USAGE
2553static inline unsigned long stack_not_used(struct task_struct *p)
2554{
2555 unsigned long *n = end_of_stack(p);
2556
2557 do { /* Skip over canary */
2558 n++;
2559 } while (!*n);
2560
2561 return (unsigned long)n - (unsigned long)end_of_stack(p);
2562}
2563#endif
2564
1da177e4
LT
2565/* set thread flags in other task's structures
2566 * - see asm/thread_info.h for TIF_xxxx flags available
2567 */
2568static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2569{
a1261f54 2570 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2571}
2572
2573static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2574{
a1261f54 2575 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2576}
2577
2578static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2579{
a1261f54 2580 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2581}
2582
2583static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2584{
a1261f54 2585 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2586}
2587
2588static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2589{
a1261f54 2590 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2591}
2592
2593static inline void set_tsk_need_resched(struct task_struct *tsk)
2594{
2595 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2596}
2597
2598static inline void clear_tsk_need_resched(struct task_struct *tsk)
2599{
2600 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2601}
2602
8ae121ac
GH
2603static inline int test_tsk_need_resched(struct task_struct *tsk)
2604{
2605 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2606}
2607
690cc3ff
EB
2608static inline int restart_syscall(void)
2609{
2610 set_tsk_thread_flag(current, TIF_SIGPENDING);
2611 return -ERESTARTNOINTR;
2612}
2613
1da177e4
LT
2614static inline int signal_pending(struct task_struct *p)
2615{
2616 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2617}
f776d12d 2618
d9588725
RM
2619static inline int __fatal_signal_pending(struct task_struct *p)
2620{
2621 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2622}
f776d12d
MW
2623
2624static inline int fatal_signal_pending(struct task_struct *p)
2625{
2626 return signal_pending(p) && __fatal_signal_pending(p);
2627}
2628
16882c1e
ON
2629static inline int signal_pending_state(long state, struct task_struct *p)
2630{
2631 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2632 return 0;
2633 if (!signal_pending(p))
2634 return 0;
2635
16882c1e
ON
2636 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2637}
2638
1da177e4
LT
2639/*
2640 * cond_resched() and cond_resched_lock(): latency reduction via
2641 * explicit rescheduling in places that are safe. The return
2642 * value indicates whether a reschedule was done in fact.
2643 * cond_resched_lock() will drop the spinlock before scheduling,
2644 * cond_resched_softirq() will enable bhs before scheduling.
2645 */
c3921ab7 2646extern int _cond_resched(void);
6f80bd98 2647
613afbf8
FW
2648#define cond_resched() ({ \
2649 __might_sleep(__FILE__, __LINE__, 0); \
2650 _cond_resched(); \
2651})
6f80bd98 2652
613afbf8
FW
2653extern int __cond_resched_lock(spinlock_t *lock);
2654
bdd4e85d 2655#ifdef CONFIG_PREEMPT_COUNT
716a4234 2656#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2657#else
716a4234 2658#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2659#endif
716a4234 2660
613afbf8 2661#define cond_resched_lock(lock) ({ \
716a4234 2662 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2663 __cond_resched_lock(lock); \
2664})
2665
2666extern int __cond_resched_softirq(void);
2667
75e1056f
VP
2668#define cond_resched_softirq() ({ \
2669 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2670 __cond_resched_softirq(); \
613afbf8 2671})
1da177e4 2672
f6f3c437
SH
2673static inline void cond_resched_rcu(void)
2674{
2675#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
2676 rcu_read_unlock();
2677 cond_resched();
2678 rcu_read_lock();
2679#endif
2680}
2681
1da177e4
LT
2682/*
2683 * Does a critical section need to be broken due to another
95c354fe
NP
2684 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2685 * but a general need for low latency)
1da177e4 2686 */
95c354fe 2687static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2688{
95c354fe
NP
2689#ifdef CONFIG_PREEMPT
2690 return spin_is_contended(lock);
2691#else
1da177e4 2692 return 0;
95c354fe 2693#endif
1da177e4
LT
2694}
2695
ee761f62
TG
2696/*
2697 * Idle thread specific functions to determine the need_resched
2698 * polling state. We have two versions, one based on TS_POLLING in
2699 * thread_info.status and one based on TIF_POLLING_NRFLAG in
2700 * thread_info.flags
2701 */
2702#ifdef TS_POLLING
2703static inline int tsk_is_polling(struct task_struct *p)
2704{
2705 return task_thread_info(p)->status & TS_POLLING;
2706}
ea811747 2707static inline void __current_set_polling(void)
3a98f871
TG
2708{
2709 current_thread_info()->status |= TS_POLLING;
2710}
2711
ea811747
PZ
2712static inline bool __must_check current_set_polling_and_test(void)
2713{
2714 __current_set_polling();
2715
2716 /*
2717 * Polling state must be visible before we test NEED_RESCHED,
2718 * paired by resched_task()
2719 */
2720 smp_mb();
2721
2722 return unlikely(tif_need_resched());
2723}
2724
2725static inline void __current_clr_polling(void)
3a98f871
TG
2726{
2727 current_thread_info()->status &= ~TS_POLLING;
ea811747
PZ
2728}
2729
2730static inline bool __must_check current_clr_polling_and_test(void)
2731{
2732 __current_clr_polling();
2733
2734 /*
2735 * Polling state must be visible before we test NEED_RESCHED,
2736 * paired by resched_task()
2737 */
2738 smp_mb();
2739
2740 return unlikely(tif_need_resched());
3a98f871 2741}
ee761f62
TG
2742#elif defined(TIF_POLLING_NRFLAG)
2743static inline int tsk_is_polling(struct task_struct *p)
2744{
2745 return test_tsk_thread_flag(p, TIF_POLLING_NRFLAG);
2746}
ea811747
PZ
2747
2748static inline void __current_set_polling(void)
3a98f871
TG
2749{
2750 set_thread_flag(TIF_POLLING_NRFLAG);
2751}
2752
ea811747
PZ
2753static inline bool __must_check current_set_polling_and_test(void)
2754{
2755 __current_set_polling();
2756
2757 /*
2758 * Polling state must be visible before we test NEED_RESCHED,
2759 * paired by resched_task()
2760 *
2761 * XXX: assumes set/clear bit are identical barrier wise.
2762 */
2763 smp_mb__after_clear_bit();
2764
2765 return unlikely(tif_need_resched());
2766}
2767
2768static inline void __current_clr_polling(void)
3a98f871
TG
2769{
2770 clear_thread_flag(TIF_POLLING_NRFLAG);
2771}
ea811747
PZ
2772
2773static inline bool __must_check current_clr_polling_and_test(void)
2774{
2775 __current_clr_polling();
2776
2777 /*
2778 * Polling state must be visible before we test NEED_RESCHED,
2779 * paired by resched_task()
2780 */
2781 smp_mb__after_clear_bit();
2782
2783 return unlikely(tif_need_resched());
2784}
2785
ee761f62
TG
2786#else
2787static inline int tsk_is_polling(struct task_struct *p) { return 0; }
ea811747
PZ
2788static inline void __current_set_polling(void) { }
2789static inline void __current_clr_polling(void) { }
2790
2791static inline bool __must_check current_set_polling_and_test(void)
2792{
2793 return unlikely(tif_need_resched());
2794}
2795static inline bool __must_check current_clr_polling_and_test(void)
2796{
2797 return unlikely(tif_need_resched());
2798}
ee761f62
TG
2799#endif
2800
8cb75e0c
PZ
2801static inline void current_clr_polling(void)
2802{
2803 __current_clr_polling();
2804
2805 /*
2806 * Ensure we check TIF_NEED_RESCHED after we clear the polling bit.
2807 * Once the bit is cleared, we'll get IPIs with every new
2808 * TIF_NEED_RESCHED and the IPI handler, scheduler_ipi(), will also
2809 * fold.
2810 */
2811 smp_mb(); /* paired with resched_task() */
2812
2813 preempt_fold_need_resched();
2814}
2815
75f93fed
PZ
2816static __always_inline bool need_resched(void)
2817{
2818 return unlikely(tif_need_resched());
2819}
2820
f06febc9
FM
2821/*
2822 * Thread group CPU time accounting.
2823 */
4cd4c1b4 2824void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2825void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2826
490dea45 2827static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2828{
ee30a7b2 2829 raw_spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2830}
2831
7bb44ade
RM
2832/*
2833 * Reevaluate whether the task has signals pending delivery.
2834 * Wake the task if so.
2835 * This is required every time the blocked sigset_t changes.
2836 * callers must hold sighand->siglock.
2837 */
2838extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2839extern void recalc_sigpending(void);
2840
910ffdb1
ON
2841extern void signal_wake_up_state(struct task_struct *t, unsigned int state);
2842
2843static inline void signal_wake_up(struct task_struct *t, bool resume)
2844{
2845 signal_wake_up_state(t, resume ? TASK_WAKEKILL : 0);
2846}
2847static inline void ptrace_signal_wake_up(struct task_struct *t, bool resume)
2848{
2849 signal_wake_up_state(t, resume ? __TASK_TRACED : 0);
2850}
1da177e4
LT
2851
2852/*
2853 * Wrappers for p->thread_info->cpu access. No-op on UP.
2854 */
2855#ifdef CONFIG_SMP
2856
2857static inline unsigned int task_cpu(const struct task_struct *p)
2858{
a1261f54 2859 return task_thread_info(p)->cpu;
1da177e4
LT
2860}
2861
b32e86b4
IM
2862static inline int task_node(const struct task_struct *p)
2863{
2864 return cpu_to_node(task_cpu(p));
2865}
2866
c65cc870 2867extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2868
2869#else
2870
2871static inline unsigned int task_cpu(const struct task_struct *p)
2872{
2873 return 0;
2874}
2875
2876static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2877{
2878}
2879
2880#endif /* CONFIG_SMP */
2881
96f874e2
RR
2882extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2883extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2884
7c941438 2885#ifdef CONFIG_CGROUP_SCHED
07e06b01 2886extern struct task_group root_task_group;
8323f26c 2887#endif /* CONFIG_CGROUP_SCHED */
9b5b7751 2888
54e99124
DG
2889extern int task_can_switch_user(struct user_struct *up,
2890 struct task_struct *tsk);
2891
4b98d11b
AD
2892#ifdef CONFIG_TASK_XACCT
2893static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2894{
940389b8 2895 tsk->ioac.rchar += amt;
4b98d11b
AD
2896}
2897
2898static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2899{
940389b8 2900 tsk->ioac.wchar += amt;
4b98d11b
AD
2901}
2902
2903static inline void inc_syscr(struct task_struct *tsk)
2904{
940389b8 2905 tsk->ioac.syscr++;
4b98d11b
AD
2906}
2907
2908static inline void inc_syscw(struct task_struct *tsk)
2909{
940389b8 2910 tsk->ioac.syscw++;
4b98d11b
AD
2911}
2912#else
2913static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2914{
2915}
2916
2917static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2918{
2919}
2920
2921static inline void inc_syscr(struct task_struct *tsk)
2922{
2923}
2924
2925static inline void inc_syscw(struct task_struct *tsk)
2926{
2927}
2928#endif
2929
82455257
DH
2930#ifndef TASK_SIZE_OF
2931#define TASK_SIZE_OF(tsk) TASK_SIZE
2932#endif
2933
cf475ad2
BS
2934#ifdef CONFIG_MM_OWNER
2935extern void mm_update_next_owner(struct mm_struct *mm);
2936extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2937#else
2938static inline void mm_update_next_owner(struct mm_struct *mm)
2939{
2940}
2941
2942static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2943{
2944}
2945#endif /* CONFIG_MM_OWNER */
2946
3e10e716
JS
2947static inline unsigned long task_rlimit(const struct task_struct *tsk,
2948 unsigned int limit)
2949{
2950 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2951}
2952
2953static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2954 unsigned int limit)
2955{
2956 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2957}
2958
2959static inline unsigned long rlimit(unsigned int limit)
2960{
2961 return task_rlimit(current, limit);
2962}
2963
2964static inline unsigned long rlimit_max(unsigned int limit)
2965{
2966 return task_rlimit_max(current, limit);
2967}
2968
1da177e4 2969#endif
This page took 2.184162 seconds and 5 git commands to generate.