sched: Move rq->idle_stamp up to the core
[linux-2.6-block.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
607ca46e 4#include <uapi/linux/sched.h>
b7b3c76a 5
5c228079
DY
6#include <linux/sched/prio.h>
7
b7b3c76a
DW
8
9struct sched_param {
10 int sched_priority;
11};
12
1da177e4
LT
13#include <asm/param.h> /* for HZ */
14
1da177e4
LT
15#include <linux/capability.h>
16#include <linux/threads.h>
17#include <linux/kernel.h>
18#include <linux/types.h>
19#include <linux/timex.h>
20#include <linux/jiffies.h>
fb00aca4 21#include <linux/plist.h>
1da177e4
LT
22#include <linux/rbtree.h>
23#include <linux/thread_info.h>
24#include <linux/cpumask.h>
25#include <linux/errno.h>
26#include <linux/nodemask.h>
c92ff1bd 27#include <linux/mm_types.h>
00d1a39e 28#include <linux/preempt_mask.h>
1da177e4 29
1da177e4
LT
30#include <asm/page.h>
31#include <asm/ptrace.h>
1da177e4
LT
32#include <asm/cputime.h>
33
34#include <linux/smp.h>
35#include <linux/sem.h>
36#include <linux/signal.h>
1da177e4
LT
37#include <linux/compiler.h>
38#include <linux/completion.h>
39#include <linux/pid.h>
40#include <linux/percpu.h>
41#include <linux/topology.h>
3e26c149 42#include <linux/proportions.h>
1da177e4 43#include <linux/seccomp.h>
e56d0903 44#include <linux/rcupdate.h>
05725f7e 45#include <linux/rculist.h>
23f78d4a 46#include <linux/rtmutex.h>
1da177e4 47
a3b6714e
DW
48#include <linux/time.h>
49#include <linux/param.h>
50#include <linux/resource.h>
51#include <linux/timer.h>
52#include <linux/hrtimer.h>
7c3ab738 53#include <linux/task_io_accounting.h>
9745512c 54#include <linux/latencytop.h>
9e2b2dc4 55#include <linux/cred.h>
fa14ff4a 56#include <linux/llist.h>
7b44ab97 57#include <linux/uidgid.h>
21caf2fc 58#include <linux/gfp.h>
a3b6714e
DW
59
60#include <asm/processor.h>
36d57ac4 61
d50dde5a
DF
62#define SCHED_ATTR_SIZE_VER0 48 /* sizeof first published struct */
63
64/*
65 * Extended scheduling parameters data structure.
66 *
67 * This is needed because the original struct sched_param can not be
68 * altered without introducing ABI issues with legacy applications
69 * (e.g., in sched_getparam()).
70 *
71 * However, the possibility of specifying more than just a priority for
72 * the tasks may be useful for a wide variety of application fields, e.g.,
73 * multimedia, streaming, automation and control, and many others.
74 *
75 * This variant (sched_attr) is meant at describing a so-called
76 * sporadic time-constrained task. In such model a task is specified by:
77 * - the activation period or minimum instance inter-arrival time;
78 * - the maximum (or average, depending on the actual scheduling
79 * discipline) computation time of all instances, a.k.a. runtime;
80 * - the deadline (relative to the actual activation time) of each
81 * instance.
82 * Very briefly, a periodic (sporadic) task asks for the execution of
83 * some specific computation --which is typically called an instance--
84 * (at most) every period. Moreover, each instance typically lasts no more
85 * than the runtime and must be completed by time instant t equal to
86 * the instance activation time + the deadline.
87 *
88 * This is reflected by the actual fields of the sched_attr structure:
89 *
90 * @size size of the structure, for fwd/bwd compat.
91 *
92 * @sched_policy task's scheduling policy
93 * @sched_flags for customizing the scheduler behaviour
94 * @sched_nice task's nice value (SCHED_NORMAL/BATCH)
95 * @sched_priority task's static priority (SCHED_FIFO/RR)
96 * @sched_deadline representative of the task's deadline
97 * @sched_runtime representative of the task's runtime
98 * @sched_period representative of the task's period
99 *
100 * Given this task model, there are a multiplicity of scheduling algorithms
101 * and policies, that can be used to ensure all the tasks will make their
102 * timing constraints.
aab03e05
DF
103 *
104 * As of now, the SCHED_DEADLINE policy (sched_dl scheduling class) is the
105 * only user of this new interface. More information about the algorithm
106 * available in the scheduling class file or in Documentation/.
d50dde5a
DF
107 */
108struct sched_attr {
109 u32 size;
110
111 u32 sched_policy;
112 u64 sched_flags;
113
114 /* SCHED_NORMAL, SCHED_BATCH */
115 s32 sched_nice;
116
117 /* SCHED_FIFO, SCHED_RR */
118 u32 sched_priority;
119
120 /* SCHED_DEADLINE */
121 u64 sched_runtime;
122 u64 sched_deadline;
123 u64 sched_period;
124};
125
1da177e4 126struct exec_domain;
c87e2837 127struct futex_pi_state;
286100a6 128struct robust_list_head;
bddd87c7 129struct bio_list;
5ad4e53b 130struct fs_struct;
cdd6c482 131struct perf_event_context;
73c10101 132struct blk_plug;
1da177e4 133
1da177e4
LT
134/*
135 * List of flags we want to share for kernel threads,
136 * if only because they are not used by them anyway.
137 */
138#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
139
140/*
141 * These are the constant used to fake the fixed-point load-average
142 * counting. Some notes:
143 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
144 * a load-average precision of 10 bits integer + 11 bits fractional
145 * - if you want to count load-averages more often, you need more
146 * precision, or rounding will get you. With 2-second counting freq,
147 * the EXP_n values would be 1981, 2034 and 2043 if still using only
148 * 11 bit fractions.
149 */
150extern unsigned long avenrun[]; /* Load averages */
2d02494f 151extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
152
153#define FSHIFT 11 /* nr of bits of precision */
154#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 155#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
156#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
157#define EXP_5 2014 /* 1/exp(5sec/5min) */
158#define EXP_15 2037 /* 1/exp(5sec/15min) */
159
160#define CALC_LOAD(load,exp,n) \
161 load *= exp; \
162 load += n*(FIXED_1-exp); \
163 load >>= FSHIFT;
164
165extern unsigned long total_forks;
166extern int nr_threads;
1da177e4
LT
167DECLARE_PER_CPU(unsigned long, process_counts);
168extern int nr_processes(void);
169extern unsigned long nr_running(void);
1da177e4 170extern unsigned long nr_iowait(void);
8c215bd3 171extern unsigned long nr_iowait_cpu(int cpu);
69d25870
AV
172extern unsigned long this_cpu_load(void);
173
174
0f004f5a 175extern void calc_global_load(unsigned long ticks);
5aaa0b7a 176extern void update_cpu_load_nohz(void);
1da177e4 177
7e49fcce
SR
178extern unsigned long get_parent_ip(unsigned long addr);
179
b637a328
PM
180extern void dump_cpu_task(int cpu);
181
43ae34cb
IM
182struct seq_file;
183struct cfs_rq;
4cf86d77 184struct task_group;
43ae34cb
IM
185#ifdef CONFIG_SCHED_DEBUG
186extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
187extern void proc_sched_set_task(struct task_struct *p);
188extern void
5cef9eca 189print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb 190#endif
1da177e4 191
4a8342d2
LT
192/*
193 * Task state bitmask. NOTE! These bits are also
194 * encoded in fs/proc/array.c: get_task_state().
195 *
196 * We have two separate sets of flags: task->state
197 * is about runnability, while task->exit_state are
198 * about the task exiting. Confusing, but this way
199 * modifying one set can't modify the other one by
200 * mistake.
201 */
1da177e4
LT
202#define TASK_RUNNING 0
203#define TASK_INTERRUPTIBLE 1
204#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
205#define __TASK_STOPPED 4
206#define __TASK_TRACED 8
4a8342d2
LT
207/* in tsk->exit_state */
208#define EXIT_ZOMBIE 16
209#define EXIT_DEAD 32
210/* in tsk->state again */
af927232 211#define TASK_DEAD 64
f021a3c2 212#define TASK_WAKEKILL 128
e9c84311 213#define TASK_WAKING 256
f2530dc7
TG
214#define TASK_PARKED 512
215#define TASK_STATE_MAX 1024
f021a3c2 216
f2530dc7 217#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKWP"
73342151 218
e1781538
PZ
219extern char ___assert_task_state[1 - 2*!!(
220 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
221
222/* Convenience macros for the sake of set_task_state */
223#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
224#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
225#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 226
92a1f4bc
MW
227/* Convenience macros for the sake of wake_up */
228#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 229#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
230
231/* get_task_state() */
232#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2 233 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
74e37200 234 __TASK_TRACED | EXIT_ZOMBIE | EXIT_DEAD)
92a1f4bc 235
f021a3c2
MW
236#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
237#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 238#define task_is_stopped_or_traced(task) \
f021a3c2 239 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 240#define task_contributes_to_load(task) \
e3c8ca83 241 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
376fede8 242 (task->flags & PF_FROZEN) == 0)
1da177e4
LT
243
244#define __set_task_state(tsk, state_value) \
245 do { (tsk)->state = (state_value); } while (0)
246#define set_task_state(tsk, state_value) \
247 set_mb((tsk)->state, (state_value))
248
498d0c57
AM
249/*
250 * set_current_state() includes a barrier so that the write of current->state
251 * is correctly serialised wrt the caller's subsequent test of whether to
252 * actually sleep:
253 *
254 * set_current_state(TASK_UNINTERRUPTIBLE);
255 * if (do_i_need_to_sleep())
256 * schedule();
257 *
258 * If the caller does not need such serialisation then use __set_current_state()
259 */
1da177e4
LT
260#define __set_current_state(state_value) \
261 do { current->state = (state_value); } while (0)
262#define set_current_state(state_value) \
263 set_mb(current->state, (state_value))
264
265/* Task command name length */
266#define TASK_COMM_LEN 16
267
1da177e4
LT
268#include <linux/spinlock.h>
269
270/*
271 * This serializes "schedule()" and also protects
272 * the run-queue from deletions/modifications (but
273 * _adding_ to the beginning of the run-queue has
274 * a separate lock).
275 */
276extern rwlock_t tasklist_lock;
277extern spinlock_t mmlist_lock;
278
36c8b586 279struct task_struct;
1da177e4 280
db1466b3
PM
281#ifdef CONFIG_PROVE_RCU
282extern int lockdep_tasklist_lock_is_held(void);
283#endif /* #ifdef CONFIG_PROVE_RCU */
284
1da177e4
LT
285extern void sched_init(void);
286extern void sched_init_smp(void);
2d07b255 287extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 288extern void init_idle(struct task_struct *idle, int cpu);
1df21055 289extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 290
89f19f04 291extern int runqueue_is_locked(int cpu);
017730c1 292
3451d024 293#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
c1cc017c 294extern void nohz_balance_enter_idle(int cpu);
69e1e811 295extern void set_cpu_sd_state_idle(void);
83cd4fe2 296extern int get_nohz_timer_target(void);
46cb4b7c 297#else
c1cc017c 298static inline void nohz_balance_enter_idle(int cpu) { }
fdaabd80 299static inline void set_cpu_sd_state_idle(void) { }
46cb4b7c 300#endif
1da177e4 301
e59e2ae2 302/*
39bc89fd 303 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
304 */
305extern void show_state_filter(unsigned long state_filter);
306
307static inline void show_state(void)
308{
39bc89fd 309 show_state_filter(0);
e59e2ae2
IM
310}
311
1da177e4
LT
312extern void show_regs(struct pt_regs *);
313
314/*
315 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
316 * task), SP is the stack pointer of the first frame that should be shown in the back
317 * trace (or NULL if the entire call-chain of the task should be shown).
318 */
319extern void show_stack(struct task_struct *task, unsigned long *sp);
320
321void io_schedule(void);
322long io_schedule_timeout(long timeout);
323
324extern void cpu_init (void);
325extern void trap_init(void);
326extern void update_process_times(int user);
327extern void scheduler_tick(void);
328
82a1fcb9
IM
329extern void sched_show_task(struct task_struct *p);
330
19cc36c0 331#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 332extern void touch_softlockup_watchdog(void);
d6ad3e28 333extern void touch_softlockup_watchdog_sync(void);
04c9167f 334extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
335extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
336 void __user *buffer,
337 size_t *lenp, loff_t *ppos);
9c44bc03 338extern unsigned int softlockup_panic;
004417a6 339void lockup_detector_init(void);
8446f1d3 340#else
8446f1d3
IM
341static inline void touch_softlockup_watchdog(void)
342{
343}
d6ad3e28
JW
344static inline void touch_softlockup_watchdog_sync(void)
345{
346}
04c9167f
JF
347static inline void touch_all_softlockup_watchdogs(void)
348{
349}
004417a6
PZ
350static inline void lockup_detector_init(void)
351{
352}
8446f1d3
IM
353#endif
354
8b414521
MT
355#ifdef CONFIG_DETECT_HUNG_TASK
356void reset_hung_task_detector(void);
357#else
358static inline void reset_hung_task_detector(void)
359{
360}
361#endif
362
1da177e4
LT
363/* Attach to any functions which should be ignored in wchan output. */
364#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
365
366/* Linker adds these: start and end of __sched functions */
367extern char __sched_text_start[], __sched_text_end[];
368
1da177e4
LT
369/* Is this address in the __sched functions? */
370extern int in_sched_functions(unsigned long addr);
371
372#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 373extern signed long schedule_timeout(signed long timeout);
64ed93a2 374extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 375extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 376extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 377asmlinkage void schedule(void);
c5491ea7 378extern void schedule_preempt_disabled(void);
1da177e4 379
ab516013 380struct nsproxy;
acce292c 381struct user_namespace;
1da177e4 382
efc1a3b1
DH
383#ifdef CONFIG_MMU
384extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
385extern unsigned long
386arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
387 unsigned long, unsigned long);
388extern unsigned long
389arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
390 unsigned long len, unsigned long pgoff,
391 unsigned long flags);
efc1a3b1
DH
392#else
393static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
394#endif
1da177e4 395
d049f74f
KC
396#define SUID_DUMP_DISABLE 0 /* No setuid dumping */
397#define SUID_DUMP_USER 1 /* Dump as user of process */
398#define SUID_DUMP_ROOT 2 /* Dump as root */
399
6c5d5238 400/* mm flags */
f8af4da3 401
7288e118 402/* for SUID_DUMP_* above */
3cb4a0bb 403#define MMF_DUMPABLE_BITS 2
f8af4da3 404#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb 405
942be387
ON
406extern void set_dumpable(struct mm_struct *mm, int value);
407/*
408 * This returns the actual value of the suid_dumpable flag. For things
409 * that are using this for checking for privilege transitions, it must
410 * test against SUID_DUMP_USER rather than treating it as a boolean
411 * value.
412 */
413static inline int __get_dumpable(unsigned long mm_flags)
414{
415 return mm_flags & MMF_DUMPABLE_MASK;
416}
417
418static inline int get_dumpable(struct mm_struct *mm)
419{
420 return __get_dumpable(mm->flags);
421}
422
3cb4a0bb
KH
423/* coredump filter bits */
424#define MMF_DUMP_ANON_PRIVATE 2
425#define MMF_DUMP_ANON_SHARED 3
426#define MMF_DUMP_MAPPED_PRIVATE 4
427#define MMF_DUMP_MAPPED_SHARED 5
82df3973 428#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
429#define MMF_DUMP_HUGETLB_PRIVATE 7
430#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 431
3cb4a0bb 432#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 433#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
434#define MMF_DUMP_FILTER_MASK \
435 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
436#define MMF_DUMP_FILTER_DEFAULT \
e575f111 437 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
438 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
439
440#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
441# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
442#else
443# define MMF_DUMP_MASK_DEFAULT_ELF 0
444#endif
f8af4da3
HD
445 /* leave room for more dump flags */
446#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 447#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
bafb282d 448#define MMF_EXE_FILE_CHANGED 18 /* see prctl_set_mm_exe_file() */
f8af4da3 449
9f68f672
ON
450#define MMF_HAS_UPROBES 19 /* has uprobes */
451#define MMF_RECALC_UPROBES 20 /* MMF_HAS_UPROBES can be wrong */
f8ac4ec9 452
f8af4da3 453#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 454
1da177e4
LT
455struct sighand_struct {
456 atomic_t count;
457 struct k_sigaction action[_NSIG];
458 spinlock_t siglock;
b8fceee1 459 wait_queue_head_t signalfd_wqh;
1da177e4
LT
460};
461
0e464814 462struct pacct_struct {
f6ec29a4
KK
463 int ac_flag;
464 long ac_exitcode;
0e464814 465 unsigned long ac_mem;
77787bfb
KK
466 cputime_t ac_utime, ac_stime;
467 unsigned long ac_minflt, ac_majflt;
0e464814
KK
468};
469
42c4ab41
SG
470struct cpu_itimer {
471 cputime_t expires;
472 cputime_t incr;
8356b5f9
SG
473 u32 error;
474 u32 incr_error;
42c4ab41
SG
475};
476
d37f761d
FW
477/**
478 * struct cputime - snaphsot of system and user cputime
479 * @utime: time spent in user mode
480 * @stime: time spent in system mode
481 *
482 * Gathers a generic snapshot of user and system time.
483 */
484struct cputime {
485 cputime_t utime;
486 cputime_t stime;
487};
488
f06febc9
FM
489/**
490 * struct task_cputime - collected CPU time counts
491 * @utime: time spent in user mode, in &cputime_t units
492 * @stime: time spent in kernel mode, in &cputime_t units
493 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 494 *
d37f761d
FW
495 * This is an extension of struct cputime that includes the total runtime
496 * spent by the task from the scheduler point of view.
497 *
498 * As a result, this structure groups together three kinds of CPU time
499 * that are tracked for threads and thread groups. Most things considering
f06febc9
FM
500 * CPU time want to group these counts together and treat all three
501 * of them in parallel.
502 */
503struct task_cputime {
504 cputime_t utime;
505 cputime_t stime;
506 unsigned long long sum_exec_runtime;
507};
508/* Alternate field names when used to cache expirations. */
509#define prof_exp stime
510#define virt_exp utime
511#define sched_exp sum_exec_runtime
512
4cd4c1b4
PZ
513#define INIT_CPUTIME \
514 (struct task_cputime) { \
64861634
MS
515 .utime = 0, \
516 .stime = 0, \
4cd4c1b4
PZ
517 .sum_exec_runtime = 0, \
518 }
519
a233f112
PZ
520#ifdef CONFIG_PREEMPT_COUNT
521#define PREEMPT_DISABLED (1 + PREEMPT_ENABLED)
522#else
523#define PREEMPT_DISABLED PREEMPT_ENABLED
524#endif
525
c99e6efe
PZ
526/*
527 * Disable preemption until the scheduler is running.
528 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
529 *
530 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
531 * before the scheduler is active -- see should_resched().
c99e6efe 532 */
a233f112 533#define INIT_PREEMPT_COUNT (PREEMPT_DISABLED + PREEMPT_ACTIVE)
c99e6efe 534
f06febc9 535/**
4cd4c1b4
PZ
536 * struct thread_group_cputimer - thread group interval timer counts
537 * @cputime: thread group interval timers.
538 * @running: non-zero when there are timers running and
539 * @cputime receives updates.
540 * @lock: lock for fields in this struct.
f06febc9
FM
541 *
542 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 543 * used for thread group CPU timer calculations.
f06febc9 544 */
4cd4c1b4
PZ
545struct thread_group_cputimer {
546 struct task_cputime cputime;
547 int running;
ee30a7b2 548 raw_spinlock_t lock;
f06febc9 549};
f06febc9 550
4714d1d3 551#include <linux/rwsem.h>
5091faa4
MG
552struct autogroup;
553
1da177e4 554/*
e815f0a8 555 * NOTE! "signal_struct" does not have its own
1da177e4
LT
556 * locking, because a shared signal_struct always
557 * implies a shared sighand_struct, so locking
558 * sighand_struct is always a proper superset of
559 * the locking of signal_struct.
560 */
561struct signal_struct {
ea6d290c 562 atomic_t sigcnt;
1da177e4 563 atomic_t live;
b3ac022c 564 int nr_threads;
0c740d0a 565 struct list_head thread_head;
1da177e4
LT
566
567 wait_queue_head_t wait_chldexit; /* for wait4() */
568
569 /* current thread group signal load-balancing target: */
36c8b586 570 struct task_struct *curr_target;
1da177e4
LT
571
572 /* shared signal handling: */
573 struct sigpending shared_pending;
574
575 /* thread group exit support */
576 int group_exit_code;
577 /* overloaded:
578 * - notify group_exit_task when ->count is equal to notify_count
579 * - everyone except group_exit_task is stopped during signal delivery
580 * of fatal signals, group_exit_task processes the signal.
581 */
1da177e4 582 int notify_count;
07dd20e0 583 struct task_struct *group_exit_task;
1da177e4
LT
584
585 /* thread group stop support, overloads group_exit_code too */
586 int group_stop_count;
587 unsigned int flags; /* see SIGNAL_* flags below */
588
ebec18a6
LP
589 /*
590 * PR_SET_CHILD_SUBREAPER marks a process, like a service
591 * manager, to re-parent orphan (double-forking) child processes
592 * to this process instead of 'init'. The service manager is
593 * able to receive SIGCHLD signals and is able to investigate
594 * the process until it calls wait(). All children of this
595 * process will inherit a flag if they should look for a
596 * child_subreaper process at exit.
597 */
598 unsigned int is_child_subreaper:1;
599 unsigned int has_child_subreaper:1;
600
1da177e4 601 /* POSIX.1b Interval Timers */
5ed67f05
PE
602 int posix_timer_id;
603 struct list_head posix_timers;
1da177e4
LT
604
605 /* ITIMER_REAL timer for the process */
2ff678b8 606 struct hrtimer real_timer;
fea9d175 607 struct pid *leader_pid;
2ff678b8 608 ktime_t it_real_incr;
1da177e4 609
42c4ab41
SG
610 /*
611 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
612 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
613 * values are defined to 0 and 1 respectively
614 */
615 struct cpu_itimer it[2];
1da177e4 616
f06febc9 617 /*
4cd4c1b4
PZ
618 * Thread group totals for process CPU timers.
619 * See thread_group_cputimer(), et al, for details.
f06febc9 620 */
4cd4c1b4 621 struct thread_group_cputimer cputimer;
f06febc9
FM
622
623 /* Earliest-expiration cache. */
624 struct task_cputime cputime_expires;
625
626 struct list_head cpu_timers[3];
627
ab521dc0 628 struct pid *tty_old_pgrp;
1ec320af 629
1da177e4
LT
630 /* boolean value for session group leader */
631 int leader;
632
633 struct tty_struct *tty; /* NULL if no tty */
634
5091faa4
MG
635#ifdef CONFIG_SCHED_AUTOGROUP
636 struct autogroup *autogroup;
637#endif
1da177e4
LT
638 /*
639 * Cumulative resource counters for dead threads in the group,
640 * and for reaped dead child processes forked by this group.
641 * Live threads maintain their own counters and add to these
642 * in __exit_signal, except for the group leader.
643 */
32bd671d 644 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
645 cputime_t gtime;
646 cputime_t cgtime;
9fbc42ea 647#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 648 struct cputime prev_cputime;
0cf55e1e 649#endif
1da177e4
LT
650 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
651 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 652 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 653 unsigned long maxrss, cmaxrss;
940389b8 654 struct task_io_accounting ioac;
1da177e4 655
32bd671d
PZ
656 /*
657 * Cumulative ns of schedule CPU time fo dead threads in the
658 * group, not including a zombie group leader, (This only differs
659 * from jiffies_to_ns(utime + stime) if sched_clock uses something
660 * other than jiffies.)
661 */
662 unsigned long long sum_sched_runtime;
663
1da177e4
LT
664 /*
665 * We don't bother to synchronize most readers of this at all,
666 * because there is no reader checking a limit that actually needs
667 * to get both rlim_cur and rlim_max atomically, and either one
668 * alone is a single word that can safely be read normally.
669 * getrlimit/setrlimit use task_lock(current->group_leader) to
670 * protect this instead of the siglock, because they really
671 * have no need to disable irqs.
672 */
673 struct rlimit rlim[RLIM_NLIMITS];
674
0e464814
KK
675#ifdef CONFIG_BSD_PROCESS_ACCT
676 struct pacct_struct pacct; /* per-process accounting information */
677#endif
ad4ecbcb 678#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
679 struct taskstats *stats;
680#endif
522ed776
MT
681#ifdef CONFIG_AUDIT
682 unsigned audit_tty;
46e959ea 683 unsigned audit_tty_log_passwd;
522ed776
MT
684 struct tty_audit_buf *tty_audit_buf;
685#endif
4714d1d3
BB
686#ifdef CONFIG_CGROUPS
687 /*
77e4ef99
TH
688 * group_rwsem prevents new tasks from entering the threadgroup and
689 * member tasks from exiting,a more specifically, setting of
690 * PF_EXITING. fork and exit paths are protected with this rwsem
691 * using threadgroup_change_begin/end(). Users which require
692 * threadgroup to remain stable should use threadgroup_[un]lock()
693 * which also takes care of exec path. Currently, cgroup is the
694 * only user.
4714d1d3 695 */
257058ae 696 struct rw_semaphore group_rwsem;
4714d1d3 697#endif
28b83c51 698
e1e12d2f 699 oom_flags_t oom_flags;
a9c58b90
DR
700 short oom_score_adj; /* OOM kill score adjustment */
701 short oom_score_adj_min; /* OOM kill score adjustment min value.
702 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
703
704 struct mutex cred_guard_mutex; /* guard against foreign influences on
705 * credential calculations
706 * (notably. ptrace) */
1da177e4
LT
707};
708
709/*
710 * Bits in flags field of signal_struct.
711 */
712#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
ee77f075
ON
713#define SIGNAL_STOP_CONTINUED 0x00000002 /* SIGCONT since WCONTINUED reap */
714#define SIGNAL_GROUP_EXIT 0x00000004 /* group exit in progress */
403bad72 715#define SIGNAL_GROUP_COREDUMP 0x00000008 /* coredump in progress */
e4420551
ON
716/*
717 * Pending notifications to parent.
718 */
719#define SIGNAL_CLD_STOPPED 0x00000010
720#define SIGNAL_CLD_CONTINUED 0x00000020
721#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 722
fae5fa44
ON
723#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
724
ed5d2cac
ON
725/* If true, all threads except ->group_exit_task have pending SIGKILL */
726static inline int signal_group_exit(const struct signal_struct *sig)
727{
728 return (sig->flags & SIGNAL_GROUP_EXIT) ||
729 (sig->group_exit_task != NULL);
730}
731
1da177e4
LT
732/*
733 * Some day this will be a full-fledged user tracking system..
734 */
735struct user_struct {
736 atomic_t __count; /* reference count */
737 atomic_t processes; /* How many processes does this user have? */
738 atomic_t files; /* How many open files does this user have? */
739 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 740#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
741 atomic_t inotify_watches; /* How many inotify watches does this user have? */
742 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
743#endif
4afeff85
EP
744#ifdef CONFIG_FANOTIFY
745 atomic_t fanotify_listeners;
746#endif
7ef9964e 747#ifdef CONFIG_EPOLL
52bd19f7 748 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 749#endif
970a8645 750#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
751 /* protected by mq_lock */
752 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 753#endif
1da177e4
LT
754 unsigned long locked_shm; /* How many pages of mlocked shm ? */
755
756#ifdef CONFIG_KEYS
757 struct key *uid_keyring; /* UID specific keyring */
758 struct key *session_keyring; /* UID's default session keyring */
759#endif
760
761 /* Hash table maintenance information */
735de223 762 struct hlist_node uidhash_node;
7b44ab97 763 kuid_t uid;
24e377a8 764
cdd6c482 765#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
766 atomic_long_t locked_vm;
767#endif
1da177e4
LT
768};
769
eb41d946 770extern int uids_sysfs_init(void);
5cb350ba 771
7b44ab97 772extern struct user_struct *find_user(kuid_t);
1da177e4
LT
773
774extern struct user_struct root_user;
775#define INIT_USER (&root_user)
776
b6dff3ec 777
1da177e4
LT
778struct backing_dev_info;
779struct reclaim_state;
780
52f17b6c 781#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
782struct sched_info {
783 /* cumulative counters */
2d72376b 784 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 785 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
786
787 /* timestamps */
172ba844
BS
788 unsigned long long last_arrival,/* when we last ran on a cpu */
789 last_queued; /* when we were last queued to run */
1da177e4 790};
52f17b6c 791#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 792
ca74e92b
SN
793#ifdef CONFIG_TASK_DELAY_ACCT
794struct task_delay_info {
795 spinlock_t lock;
796 unsigned int flags; /* Private per-task flags */
797
798 /* For each stat XXX, add following, aligned appropriately
799 *
800 * struct timespec XXX_start, XXX_end;
801 * u64 XXX_delay;
802 * u32 XXX_count;
803 *
804 * Atomicity of updates to XXX_delay, XXX_count protected by
805 * single lock above (split into XXX_lock if contention is an issue).
806 */
0ff92245
SN
807
808 /*
809 * XXX_count is incremented on every XXX operation, the delay
810 * associated with the operation is added to XXX_delay.
811 * XXX_delay contains the accumulated delay time in nanoseconds.
812 */
813 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
814 u64 blkio_delay; /* wait for sync block io completion */
815 u64 swapin_delay; /* wait for swapin block io completion */
816 u32 blkio_count; /* total count of the number of sync block */
817 /* io operations performed */
818 u32 swapin_count; /* total count of the number of swapin block */
819 /* io operations performed */
873b4771
KK
820
821 struct timespec freepages_start, freepages_end;
822 u64 freepages_delay; /* wait for memory reclaim */
823 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 824};
52f17b6c
CS
825#endif /* CONFIG_TASK_DELAY_ACCT */
826
827static inline int sched_info_on(void)
828{
829#ifdef CONFIG_SCHEDSTATS
830 return 1;
831#elif defined(CONFIG_TASK_DELAY_ACCT)
832 extern int delayacct_on;
833 return delayacct_on;
834#else
835 return 0;
ca74e92b 836#endif
52f17b6c 837}
ca74e92b 838
d15bcfdb
IM
839enum cpu_idle_type {
840 CPU_IDLE,
841 CPU_NOT_IDLE,
842 CPU_NEWLY_IDLE,
843 CPU_MAX_IDLE_TYPES
1da177e4
LT
844};
845
1399fa78
NR
846/*
847 * Increase resolution of cpu_power calculations
848 */
849#define SCHED_POWER_SHIFT 10
850#define SCHED_POWER_SCALE (1L << SCHED_POWER_SHIFT)
1da177e4 851
1399fa78
NR
852/*
853 * sched-domains (multiprocessor balancing) declarations:
854 */
2dd73a4f 855#ifdef CONFIG_SMP
b5d978e0
PZ
856#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
857#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
858#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
859#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 860#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 861#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
b5d978e0 862#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
b5d978e0
PZ
863#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
864#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 865#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 866#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
e3589f6c 867#define SD_OVERLAP 0x2000 /* sched_domains of this level overlap */
3a7053b3 868#define SD_NUMA 0x4000 /* cross-node balancing */
5c45bf27 869
532cb4c4
MN
870extern int __weak arch_sd_sibiling_asym_packing(void);
871
1d3504fc
HS
872struct sched_domain_attr {
873 int relax_domain_level;
874};
875
876#define SD_ATTR_INIT (struct sched_domain_attr) { \
877 .relax_domain_level = -1, \
878}
879
60495e77
PZ
880extern int sched_domain_level_max;
881
5e6521ea
LZ
882struct sched_group;
883
1da177e4
LT
884struct sched_domain {
885 /* These fields must be setup */
886 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 887 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 888 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
889 unsigned long min_interval; /* Minimum balance interval ms */
890 unsigned long max_interval; /* Maximum balance interval ms */
891 unsigned int busy_factor; /* less balancing by factor if busy */
892 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 893 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
894 unsigned int busy_idx;
895 unsigned int idle_idx;
896 unsigned int newidle_idx;
897 unsigned int wake_idx;
147cbb4b 898 unsigned int forkexec_idx;
a52bfd73 899 unsigned int smt_gain;
25f55d9d
VG
900
901 int nohz_idle; /* NOHZ IDLE status */
1da177e4 902 int flags; /* See SD_* */
60495e77 903 int level;
1da177e4
LT
904
905 /* Runtime fields. */
906 unsigned long last_balance; /* init to jiffies. units in jiffies */
907 unsigned int balance_interval; /* initialise to 1. units in ms. */
908 unsigned int nr_balance_failed; /* initialise to 0 */
909
f48627e6 910 /* idle_balance() stats */
9bd721c5 911 u64 max_newidle_lb_cost;
f48627e6 912 unsigned long next_decay_max_lb_cost;
2398f2c6 913
1da177e4
LT
914#ifdef CONFIG_SCHEDSTATS
915 /* load_balance() stats */
480b9434
KC
916 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
917 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
918 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
919 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
920 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
921 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
922 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
923 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
924
925 /* Active load balancing */
480b9434
KC
926 unsigned int alb_count;
927 unsigned int alb_failed;
928 unsigned int alb_pushed;
1da177e4 929
68767a0a 930 /* SD_BALANCE_EXEC stats */
480b9434
KC
931 unsigned int sbe_count;
932 unsigned int sbe_balanced;
933 unsigned int sbe_pushed;
1da177e4 934
68767a0a 935 /* SD_BALANCE_FORK stats */
480b9434
KC
936 unsigned int sbf_count;
937 unsigned int sbf_balanced;
938 unsigned int sbf_pushed;
68767a0a 939
1da177e4 940 /* try_to_wake_up() stats */
480b9434
KC
941 unsigned int ttwu_wake_remote;
942 unsigned int ttwu_move_affine;
943 unsigned int ttwu_move_balance;
1da177e4 944#endif
a5d8c348
IM
945#ifdef CONFIG_SCHED_DEBUG
946 char *name;
947#endif
dce840a0
PZ
948 union {
949 void *private; /* used during construction */
950 struct rcu_head rcu; /* used during destruction */
951 };
6c99e9ad 952
669c55e9 953 unsigned int span_weight;
4200efd9
IM
954 /*
955 * Span of all CPUs in this domain.
956 *
957 * NOTE: this field is variable length. (Allocated dynamically
958 * by attaching extra space to the end of the structure,
959 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
960 */
961 unsigned long span[0];
1da177e4
LT
962};
963
758b2cdc
RR
964static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
965{
6c99e9ad 966 return to_cpumask(sd->span);
758b2cdc
RR
967}
968
acc3f5d7 969extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 970 struct sched_domain_attr *dattr_new);
029190c5 971
acc3f5d7
RR
972/* Allocate an array of sched domains, for partition_sched_domains(). */
973cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
974void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
975
39be3501
PZ
976bool cpus_share_cache(int this_cpu, int that_cpu);
977
1b427c15 978#else /* CONFIG_SMP */
1da177e4 979
1b427c15 980struct sched_domain_attr;
d02c7a8c 981
1b427c15 982static inline void
acc3f5d7 983partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
984 struct sched_domain_attr *dattr_new)
985{
d02c7a8c 986}
39be3501
PZ
987
988static inline bool cpus_share_cache(int this_cpu, int that_cpu)
989{
990 return true;
991}
992
1b427c15 993#endif /* !CONFIG_SMP */
1da177e4 994
47fe38fc 995
1da177e4 996struct io_context; /* See blkdev.h */
1da177e4 997
1da177e4 998
383f2835 999#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1000extern void prefetch_stack(struct task_struct *t);
383f2835
CK
1001#else
1002static inline void prefetch_stack(struct task_struct *t) { }
1003#endif
1da177e4
LT
1004
1005struct audit_context; /* See audit.c */
1006struct mempolicy;
b92ce558 1007struct pipe_inode_info;
4865ecf1 1008struct uts_namespace;
1da177e4 1009
20b8a59f 1010struct load_weight {
9dbdb155
PZ
1011 unsigned long weight;
1012 u32 inv_weight;
20b8a59f
IM
1013};
1014
9d85f21c
PT
1015struct sched_avg {
1016 /*
1017 * These sums represent an infinite geometric series and so are bound
239003ea 1018 * above by 1024/(1-y). Thus we only need a u32 to store them for all
9d85f21c
PT
1019 * choices of y < 1-2^(-32)*1024.
1020 */
1021 u32 runnable_avg_sum, runnable_avg_period;
1022 u64 last_runnable_update;
9ee474f5 1023 s64 decay_count;
2dac754e 1024 unsigned long load_avg_contrib;
9d85f21c
PT
1025};
1026
94c18227 1027#ifdef CONFIG_SCHEDSTATS
41acab88 1028struct sched_statistics {
20b8a59f 1029 u64 wait_start;
94c18227 1030 u64 wait_max;
6d082592
AV
1031 u64 wait_count;
1032 u64 wait_sum;
8f0dfc34
AV
1033 u64 iowait_count;
1034 u64 iowait_sum;
94c18227 1035
20b8a59f 1036 u64 sleep_start;
20b8a59f 1037 u64 sleep_max;
94c18227
IM
1038 s64 sum_sleep_runtime;
1039
1040 u64 block_start;
20b8a59f
IM
1041 u64 block_max;
1042 u64 exec_max;
eba1ed4b 1043 u64 slice_max;
cc367732 1044
cc367732
IM
1045 u64 nr_migrations_cold;
1046 u64 nr_failed_migrations_affine;
1047 u64 nr_failed_migrations_running;
1048 u64 nr_failed_migrations_hot;
1049 u64 nr_forced_migrations;
cc367732
IM
1050
1051 u64 nr_wakeups;
1052 u64 nr_wakeups_sync;
1053 u64 nr_wakeups_migrate;
1054 u64 nr_wakeups_local;
1055 u64 nr_wakeups_remote;
1056 u64 nr_wakeups_affine;
1057 u64 nr_wakeups_affine_attempts;
1058 u64 nr_wakeups_passive;
1059 u64 nr_wakeups_idle;
41acab88
LDM
1060};
1061#endif
1062
1063struct sched_entity {
1064 struct load_weight load; /* for load-balancing */
1065 struct rb_node run_node;
1066 struct list_head group_node;
1067 unsigned int on_rq;
1068
1069 u64 exec_start;
1070 u64 sum_exec_runtime;
1071 u64 vruntime;
1072 u64 prev_sum_exec_runtime;
1073
41acab88
LDM
1074 u64 nr_migrations;
1075
41acab88
LDM
1076#ifdef CONFIG_SCHEDSTATS
1077 struct sched_statistics statistics;
94c18227
IM
1078#endif
1079
20b8a59f
IM
1080#ifdef CONFIG_FAIR_GROUP_SCHED
1081 struct sched_entity *parent;
1082 /* rq on which this entity is (to be) queued: */
1083 struct cfs_rq *cfs_rq;
1084 /* rq "owned" by this entity/group: */
1085 struct cfs_rq *my_q;
1086#endif
8bd75c77 1087
141965c7 1088#ifdef CONFIG_SMP
f4e26b12 1089 /* Per-entity load-tracking */
9d85f21c
PT
1090 struct sched_avg avg;
1091#endif
20b8a59f 1092};
70b97a7f 1093
fa717060
PZ
1094struct sched_rt_entity {
1095 struct list_head run_list;
78f2c7db 1096 unsigned long timeout;
57d2aa00 1097 unsigned long watchdog_stamp;
bee367ed 1098 unsigned int time_slice;
6f505b16 1099
58d6c2d7 1100 struct sched_rt_entity *back;
052f1dc7 1101#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1102 struct sched_rt_entity *parent;
1103 /* rq on which this entity is (to be) queued: */
1104 struct rt_rq *rt_rq;
1105 /* rq "owned" by this entity/group: */
1106 struct rt_rq *my_q;
1107#endif
fa717060
PZ
1108};
1109
aab03e05
DF
1110struct sched_dl_entity {
1111 struct rb_node rb_node;
1112
1113 /*
1114 * Original scheduling parameters. Copied here from sched_attr
1115 * during sched_setscheduler2(), they will remain the same until
1116 * the next sched_setscheduler2().
1117 */
1118 u64 dl_runtime; /* maximum runtime for each instance */
1119 u64 dl_deadline; /* relative deadline of each instance */
755378a4 1120 u64 dl_period; /* separation of two instances (period) */
332ac17e 1121 u64 dl_bw; /* dl_runtime / dl_deadline */
aab03e05
DF
1122
1123 /*
1124 * Actual scheduling parameters. Initialized with the values above,
1125 * they are continously updated during task execution. Note that
1126 * the remaining runtime could be < 0 in case we are in overrun.
1127 */
1128 s64 runtime; /* remaining runtime for this instance */
1129 u64 deadline; /* absolute deadline for this instance */
1130 unsigned int flags; /* specifying the scheduler behaviour */
1131
1132 /*
1133 * Some bool flags:
1134 *
1135 * @dl_throttled tells if we exhausted the runtime. If so, the
1136 * task has to wait for a replenishment to be performed at the
1137 * next firing of dl_timer.
1138 *
1139 * @dl_new tells if a new instance arrived. If so we must
1140 * start executing it with full runtime and reset its absolute
1141 * deadline;
2d3d891d
DF
1142 *
1143 * @dl_boosted tells if we are boosted due to DI. If so we are
1144 * outside bandwidth enforcement mechanism (but only until we
1145 * exit the critical section).
aab03e05 1146 */
2d3d891d 1147 int dl_throttled, dl_new, dl_boosted;
aab03e05
DF
1148
1149 /*
1150 * Bandwidth enforcement timer. Each -deadline task has its
1151 * own bandwidth to be enforced, thus we need one timer per task.
1152 */
1153 struct hrtimer dl_timer;
1154};
8bd75c77 1155
86848966
PM
1156struct rcu_node;
1157
8dc85d54
PZ
1158enum perf_event_task_context {
1159 perf_invalid_context = -1,
1160 perf_hw_context = 0,
89a1e187 1161 perf_sw_context,
8dc85d54
PZ
1162 perf_nr_task_contexts,
1163};
1164
1da177e4
LT
1165struct task_struct {
1166 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1167 void *stack;
1da177e4 1168 atomic_t usage;
97dc32cd
WC
1169 unsigned int flags; /* per process flags, defined below */
1170 unsigned int ptrace;
1da177e4 1171
2dd73a4f 1172#ifdef CONFIG_SMP
fa14ff4a 1173 struct llist_node wake_entry;
3ca7a440 1174 int on_cpu;
62470419
MW
1175 struct task_struct *last_wakee;
1176 unsigned long wakee_flips;
1177 unsigned long wakee_flip_decay_ts;
ac66f547
PZ
1178
1179 int wake_cpu;
2dd73a4f 1180#endif
fd2f4419 1181 int on_rq;
50e645a8 1182
b29739f9 1183 int prio, static_prio, normal_prio;
c7aceaba 1184 unsigned int rt_priority;
5522d5d5 1185 const struct sched_class *sched_class;
20b8a59f 1186 struct sched_entity se;
fa717060 1187 struct sched_rt_entity rt;
8323f26c
PZ
1188#ifdef CONFIG_CGROUP_SCHED
1189 struct task_group *sched_task_group;
1190#endif
aab03e05 1191 struct sched_dl_entity dl;
1da177e4 1192
e107be36
AK
1193#ifdef CONFIG_PREEMPT_NOTIFIERS
1194 /* list of struct preempt_notifier: */
1195 struct hlist_head preempt_notifiers;
1196#endif
1197
6c5c9341 1198#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1199 unsigned int btrace_seq;
6c5c9341 1200#endif
1da177e4 1201
97dc32cd 1202 unsigned int policy;
29baa747 1203 int nr_cpus_allowed;
1da177e4 1204 cpumask_t cpus_allowed;
1da177e4 1205
a57eb940 1206#ifdef CONFIG_PREEMPT_RCU
e260be67 1207 int rcu_read_lock_nesting;
f41d911f 1208 char rcu_read_unlock_special;
f41d911f 1209 struct list_head rcu_node_entry;
a57eb940
PM
1210#endif /* #ifdef CONFIG_PREEMPT_RCU */
1211#ifdef CONFIG_TREE_PREEMPT_RCU
1212 struct rcu_node *rcu_blocked_node;
f41d911f 1213#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1214#ifdef CONFIG_RCU_BOOST
1215 struct rt_mutex *rcu_boost_mutex;
1216#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1217
52f17b6c 1218#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1219 struct sched_info sched_info;
1220#endif
1221
1222 struct list_head tasks;
806c09a7 1223#ifdef CONFIG_SMP
917b627d 1224 struct plist_node pushable_tasks;
1baca4ce 1225 struct rb_node pushable_dl_tasks;
806c09a7 1226#endif
1da177e4
LT
1227
1228 struct mm_struct *mm, *active_mm;
4471a675
JK
1229#ifdef CONFIG_COMPAT_BRK
1230 unsigned brk_randomized:1;
1231#endif
34e55232
KH
1232#if defined(SPLIT_RSS_COUNTING)
1233 struct task_rss_stat rss_stat;
1234#endif
1da177e4 1235/* task state */
97dc32cd 1236 int exit_state;
1da177e4
LT
1237 int exit_code, exit_signal;
1238 int pdeath_signal; /* The signal sent when the parent dies */
a8f072c1 1239 unsigned int jobctl; /* JOBCTL_*, siglock protected */
9b89f6ba
AE
1240
1241 /* Used for emulating ABI behavior of previous Linux versions */
97dc32cd 1242 unsigned int personality;
9b89f6ba 1243
f9ce1f1c
KT
1244 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1245 * execve */
8f0dfc34
AV
1246 unsigned in_iowait:1;
1247
259e5e6c
AL
1248 /* task may not gain privileges */
1249 unsigned no_new_privs:1;
ca94c442
LP
1250
1251 /* Revert to default priority/policy when forking */
1252 unsigned sched_reset_on_fork:1;
a8e4f2ea 1253 unsigned sched_contributes_to_load:1;
ca94c442 1254
1da177e4
LT
1255 pid_t pid;
1256 pid_t tgid;
0a425405 1257
1314562a 1258#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1259 /* Canary value for the -fstack-protector gcc feature */
1260 unsigned long stack_canary;
1314562a 1261#endif
4d1d61a6 1262 /*
1da177e4 1263 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 1264 * older sibling, respectively. (p->father can be replaced with
f470021a 1265 * p->real_parent->pid)
1da177e4 1266 */
abd63bc3
KC
1267 struct task_struct __rcu *real_parent; /* real parent process */
1268 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1269 /*
f470021a 1270 * children/sibling forms the list of my natural children
1da177e4
LT
1271 */
1272 struct list_head children; /* list of my children */
1273 struct list_head sibling; /* linkage in my parent's children list */
1274 struct task_struct *group_leader; /* threadgroup leader */
1275
f470021a
RM
1276 /*
1277 * ptraced is the list of tasks this task is using ptrace on.
1278 * This includes both natural children and PTRACE_ATTACH targets.
1279 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1280 */
1281 struct list_head ptraced;
1282 struct list_head ptrace_entry;
1283
1da177e4 1284 /* PID/PID hash table linkage. */
92476d7f 1285 struct pid_link pids[PIDTYPE_MAX];
47e65328 1286 struct list_head thread_group;
0c740d0a 1287 struct list_head thread_node;
1da177e4
LT
1288
1289 struct completion *vfork_done; /* for vfork() */
1290 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1291 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1292
c66f08be 1293 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1294 cputime_t gtime;
9fbc42ea 1295#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
d37f761d 1296 struct cputime prev_cputime;
6a61671b
FW
1297#endif
1298#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1299 seqlock_t vtime_seqlock;
1300 unsigned long long vtime_snap;
1301 enum {
1302 VTIME_SLEEPING = 0,
1303 VTIME_USER,
1304 VTIME_SYS,
1305 } vtime_snap_whence;
d99ca3b9 1306#endif
1da177e4 1307 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1308 struct timespec start_time; /* monotonic time */
1309 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1310/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1311 unsigned long min_flt, maj_flt;
1312
f06febc9 1313 struct task_cputime cputime_expires;
1da177e4
LT
1314 struct list_head cpu_timers[3];
1315
1316/* process credentials */
1b0ba1c9 1317 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1318 * credentials (COW) */
1b0ba1c9 1319 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1320 * credentials (COW) */
36772092
PBG
1321 char comm[TASK_COMM_LEN]; /* executable name excluding path
1322 - access with [gs]et_task_comm (which lock
1323 it with task_lock())
221af7f8 1324 - initialized normally by setup_new_exec */
1da177e4
LT
1325/* file system info */
1326 int link_count, total_link_count;
3d5b6fcc 1327#ifdef CONFIG_SYSVIPC
1da177e4
LT
1328/* ipc stuff */
1329 struct sysv_sem sysvsem;
3d5b6fcc 1330#endif
e162b39a 1331#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1332/* hung task detection */
82a1fcb9
IM
1333 unsigned long last_switch_count;
1334#endif
1da177e4
LT
1335/* CPU-specific state of this task */
1336 struct thread_struct thread;
1337/* filesystem information */
1338 struct fs_struct *fs;
1339/* open file information */
1340 struct files_struct *files;
1651e14e 1341/* namespaces */
ab516013 1342 struct nsproxy *nsproxy;
1da177e4
LT
1343/* signal handlers */
1344 struct signal_struct *signal;
1345 struct sighand_struct *sighand;
1346
1347 sigset_t blocked, real_blocked;
f3de272b 1348 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1349 struct sigpending pending;
1350
1351 unsigned long sas_ss_sp;
1352 size_t sas_ss_size;
1353 int (*notifier)(void *priv);
1354 void *notifier_data;
1355 sigset_t *notifier_mask;
67d12145 1356 struct callback_head *task_works;
e73f8959 1357
1da177e4 1358 struct audit_context *audit_context;
bfef93a5 1359#ifdef CONFIG_AUDITSYSCALL
e1760bd5 1360 kuid_t loginuid;
4746ec5b 1361 unsigned int sessionid;
bfef93a5 1362#endif
932ecebb 1363 struct seccomp seccomp;
1da177e4
LT
1364
1365/* Thread group tracking */
1366 u32 parent_exec_id;
1367 u32 self_exec_id;
58568d2a
MX
1368/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1369 * mempolicy */
1da177e4 1370 spinlock_t alloc_lock;
1da177e4 1371
b29739f9 1372 /* Protection of the PI data structures: */
1d615482 1373 raw_spinlock_t pi_lock;
b29739f9 1374
23f78d4a
IM
1375#ifdef CONFIG_RT_MUTEXES
1376 /* PI waiters blocked on a rt_mutex held by this task */
fb00aca4
PZ
1377 struct rb_root pi_waiters;
1378 struct rb_node *pi_waiters_leftmost;
23f78d4a
IM
1379 /* Deadlock detection and priority inheritance handling */
1380 struct rt_mutex_waiter *pi_blocked_on;
2d3d891d
DF
1381 /* Top pi_waiters task */
1382 struct task_struct *pi_top_task;
23f78d4a
IM
1383#endif
1384
408894ee
IM
1385#ifdef CONFIG_DEBUG_MUTEXES
1386 /* mutex deadlock detection */
1387 struct mutex_waiter *blocked_on;
1388#endif
de30a2b3
IM
1389#ifdef CONFIG_TRACE_IRQFLAGS
1390 unsigned int irq_events;
de30a2b3 1391 unsigned long hardirq_enable_ip;
de30a2b3 1392 unsigned long hardirq_disable_ip;
fa1452e8 1393 unsigned int hardirq_enable_event;
de30a2b3 1394 unsigned int hardirq_disable_event;
fa1452e8
HS
1395 int hardirqs_enabled;
1396 int hardirq_context;
de30a2b3 1397 unsigned long softirq_disable_ip;
de30a2b3 1398 unsigned long softirq_enable_ip;
fa1452e8 1399 unsigned int softirq_disable_event;
de30a2b3 1400 unsigned int softirq_enable_event;
fa1452e8 1401 int softirqs_enabled;
de30a2b3
IM
1402 int softirq_context;
1403#endif
fbb9ce95 1404#ifdef CONFIG_LOCKDEP
bdb9441e 1405# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1406 u64 curr_chain_key;
1407 int lockdep_depth;
fbb9ce95 1408 unsigned int lockdep_recursion;
c7aceaba 1409 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1410 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1411#endif
408894ee 1412
1da177e4
LT
1413/* journalling filesystem info */
1414 void *journal_info;
1415
d89d8796 1416/* stacked block device info */
bddd87c7 1417 struct bio_list *bio_list;
d89d8796 1418
73c10101
JA
1419#ifdef CONFIG_BLOCK
1420/* stack plugging */
1421 struct blk_plug *plug;
1422#endif
1423
1da177e4
LT
1424/* VM state */
1425 struct reclaim_state *reclaim_state;
1426
1da177e4
LT
1427 struct backing_dev_info *backing_dev_info;
1428
1429 struct io_context *io_context;
1430
1431 unsigned long ptrace_message;
1432 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1433 struct task_io_accounting ioac;
8f0ab514 1434#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1435 u64 acct_rss_mem1; /* accumulated rss usage */
1436 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1437 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1438#endif
1439#ifdef CONFIG_CPUSETS
58568d2a 1440 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 1441 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 1442 int cpuset_mem_spread_rotor;
6adef3eb 1443 int cpuset_slab_spread_rotor;
1da177e4 1444#endif
ddbcc7e8 1445#ifdef CONFIG_CGROUPS
817929ec 1446 /* Control Group info protected by css_set_lock */
2c392b8c 1447 struct css_set __rcu *cgroups;
817929ec
PM
1448 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1449 struct list_head cg_list;
ddbcc7e8 1450#endif
42b2dd0a 1451#ifdef CONFIG_FUTEX
0771dfef 1452 struct robust_list_head __user *robust_list;
34f192c6
IM
1453#ifdef CONFIG_COMPAT
1454 struct compat_robust_list_head __user *compat_robust_list;
1455#endif
c87e2837
IM
1456 struct list_head pi_state_list;
1457 struct futex_pi_state *pi_state_cache;
c7aceaba 1458#endif
cdd6c482 1459#ifdef CONFIG_PERF_EVENTS
8dc85d54 1460 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1461 struct mutex perf_event_mutex;
1462 struct list_head perf_event_list;
a63eaf34 1463#endif
c7aceaba 1464#ifdef CONFIG_NUMA
58568d2a 1465 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1466 short il_next;
207205a2 1467 short pref_node_fork;
42b2dd0a 1468#endif
cbee9f88
PZ
1469#ifdef CONFIG_NUMA_BALANCING
1470 int numa_scan_seq;
cbee9f88 1471 unsigned int numa_scan_period;
598f0ec0 1472 unsigned int numa_scan_period_max;
de1c9ce6 1473 int numa_preferred_nid;
6b9a7460 1474 unsigned long numa_migrate_retry;
cbee9f88 1475 u64 node_stamp; /* migration stamp */
7e2703e6
RR
1476 u64 last_task_numa_placement;
1477 u64 last_sum_exec_runtime;
cbee9f88 1478 struct callback_head numa_work;
f809ca9a 1479
8c8a743c
PZ
1480 struct list_head numa_entry;
1481 struct numa_group *numa_group;
1482
745d6147
MG
1483 /*
1484 * Exponential decaying average of faults on a per-node basis.
1485 * Scheduling placement decisions are made based on the these counts.
1486 * The values remain static for the duration of a PTE scan
1487 */
ff1df896 1488 unsigned long *numa_faults_memory;
83e1d2cd 1489 unsigned long total_numa_faults;
745d6147
MG
1490
1491 /*
1492 * numa_faults_buffer records faults per node during the current
ff1df896
RR
1493 * scan window. When the scan completes, the counts in
1494 * numa_faults_memory decay and these values are copied.
745d6147 1495 */
ff1df896 1496 unsigned long *numa_faults_buffer_memory;
745d6147 1497
50ec8a40
RR
1498 /*
1499 * Track the nodes the process was running on when a NUMA hinting
1500 * fault was incurred.
1501 */
1502 unsigned long *numa_faults_cpu;
1503 unsigned long *numa_faults_buffer_cpu;
1504
04bb2f94
RR
1505 /*
1506 * numa_faults_locality tracks if faults recorded during the last
1507 * scan window were remote/local. The task scan period is adapted
1508 * based on the locality of the faults with different weights
1509 * depending on whether they were shared or private faults
1510 */
1511 unsigned long numa_faults_locality[2];
1512
b32e86b4 1513 unsigned long numa_pages_migrated;
cbee9f88
PZ
1514#endif /* CONFIG_NUMA_BALANCING */
1515
e56d0903 1516 struct rcu_head rcu;
b92ce558
JA
1517
1518 /*
1519 * cache last used pipe for splice
1520 */
1521 struct pipe_inode_info *splice_pipe;
5640f768
ED
1522
1523 struct page_frag task_frag;
1524
ca74e92b
SN
1525#ifdef CONFIG_TASK_DELAY_ACCT
1526 struct task_delay_info *delays;
f4f154fd
AM
1527#endif
1528#ifdef CONFIG_FAULT_INJECTION
1529 int make_it_fail;
ca74e92b 1530#endif
9d823e8f
WF
1531 /*
1532 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
1533 * balance_dirty_pages() for some dirty throttling pause
1534 */
1535 int nr_dirtied;
1536 int nr_dirtied_pause;
83712358 1537 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 1538
9745512c
AV
1539#ifdef CONFIG_LATENCYTOP
1540 int latency_record_count;
1541 struct latency_record latency_record[LT_SAVECOUNT];
1542#endif
6976675d
AV
1543 /*
1544 * time slack values; these are used to round up poll() and
1545 * select() etc timeout values. These are in nanoseconds.
1546 */
1547 unsigned long timer_slack_ns;
1548 unsigned long default_timer_slack_ns;
f8d570a4 1549
fb52607a 1550#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1551 /* Index of current stored address in ret_stack */
f201ae23
FW
1552 int curr_ret_stack;
1553 /* Stack of return addresses for return function tracing */
1554 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1555 /* time stamp for last schedule */
1556 unsigned long long ftrace_timestamp;
f201ae23
FW
1557 /*
1558 * Number of functions that haven't been traced
1559 * because of depth overrun.
1560 */
1561 atomic_t trace_overrun;
380c4b14
FW
1562 /* Pause for the tracing */
1563 atomic_t tracing_graph_pause;
f201ae23 1564#endif
ea4e2bc4
SR
1565#ifdef CONFIG_TRACING
1566 /* state flags for use by tracers */
1567 unsigned long trace;
b1cff0ad 1568 /* bitmask and counter of trace recursion */
261842b7
SR
1569 unsigned long trace_recursion;
1570#endif /* CONFIG_TRACING */
c255a458 1571#ifdef CONFIG_MEMCG /* memcg uses this to do batch job */
569b846d
KH
1572 struct memcg_batch_info {
1573 int do_batch; /* incremented when batch uncharge started */
1574 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1575 unsigned long nr_pages; /* uncharged usage */
1576 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d 1577 } memcg_batch;
0e9d92f2 1578 unsigned int memcg_kmem_skip_account;
519e5247 1579 struct memcg_oom_info {
49426420
JW
1580 struct mem_cgroup *memcg;
1581 gfp_t gfp_mask;
1582 int order;
519e5247
JW
1583 unsigned int may_oom:1;
1584 } memcg_oom;
569b846d 1585#endif
0326f5a9
SD
1586#ifdef CONFIG_UPROBES
1587 struct uprobe_task *utask;
0326f5a9 1588#endif
cafe5635
KO
1589#if defined(CONFIG_BCACHE) || defined(CONFIG_BCACHE_MODULE)
1590 unsigned int sequential_io;
1591 unsigned int sequential_io_avg;
1592#endif
1da177e4
LT
1593};
1594
76e6eee0 1595/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1596#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1597
6688cc05
PZ
1598#define TNF_MIGRATED 0x01
1599#define TNF_NO_GROUP 0x02
dabe1d99 1600#define TNF_SHARED 0x04
04bb2f94 1601#define TNF_FAULT_LOCAL 0x08
6688cc05 1602
cbee9f88 1603#ifdef CONFIG_NUMA_BALANCING
6688cc05 1604extern void task_numa_fault(int last_node, int node, int pages, int flags);
e29cf08b 1605extern pid_t task_numa_group_id(struct task_struct *p);
1a687c2e 1606extern void set_numabalancing_state(bool enabled);
82727018 1607extern void task_numa_free(struct task_struct *p);
10f39042
RR
1608extern bool should_numa_migrate_memory(struct task_struct *p, struct page *page,
1609 int src_nid, int dst_cpu);
cbee9f88 1610#else
ac8e895b 1611static inline void task_numa_fault(int last_node, int node, int pages,
6688cc05 1612 int flags)
cbee9f88
PZ
1613{
1614}
e29cf08b
MG
1615static inline pid_t task_numa_group_id(struct task_struct *p)
1616{
1617 return 0;
1618}
1a687c2e
MG
1619static inline void set_numabalancing_state(bool enabled)
1620{
1621}
82727018
RR
1622static inline void task_numa_free(struct task_struct *p)
1623{
1624}
10f39042
RR
1625static inline bool should_numa_migrate_memory(struct task_struct *p,
1626 struct page *page, int src_nid, int dst_cpu)
1627{
1628 return true;
1629}
cbee9f88
PZ
1630#endif
1631
e868171a 1632static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1633{
1634 return task->pids[PIDTYPE_PID].pid;
1635}
1636
e868171a 1637static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1638{
1639 return task->group_leader->pids[PIDTYPE_PID].pid;
1640}
1641
6dda81f4
ON
1642/*
1643 * Without tasklist or rcu lock it is not safe to dereference
1644 * the result of task_pgrp/task_session even if task == current,
1645 * we can race with another thread doing sys_setsid/sys_setpgid.
1646 */
e868171a 1647static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1648{
1649 return task->group_leader->pids[PIDTYPE_PGID].pid;
1650}
1651
e868171a 1652static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1653{
1654 return task->group_leader->pids[PIDTYPE_SID].pid;
1655}
1656
7af57294
PE
1657struct pid_namespace;
1658
1659/*
1660 * the helpers to get the task's different pids as they are seen
1661 * from various namespaces
1662 *
1663 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1664 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1665 * current.
7af57294
PE
1666 * task_xid_nr_ns() : id seen from the ns specified;
1667 *
1668 * set_task_vxid() : assigns a virtual id to a task;
1669 *
7af57294
PE
1670 * see also pid_nr() etc in include/linux/pid.h
1671 */
52ee2dfd
ON
1672pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1673 struct pid_namespace *ns);
7af57294 1674
e868171a 1675static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1676{
1677 return tsk->pid;
1678}
1679
52ee2dfd
ON
1680static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1681 struct pid_namespace *ns)
1682{
1683 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1684}
7af57294
PE
1685
1686static inline pid_t task_pid_vnr(struct task_struct *tsk)
1687{
52ee2dfd 1688 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1689}
1690
1691
e868171a 1692static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1693{
1694 return tsk->tgid;
1695}
1696
2f2a3a46 1697pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1698
1699static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1700{
1701 return pid_vnr(task_tgid(tsk));
1702}
1703
1704
52ee2dfd
ON
1705static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1706 struct pid_namespace *ns)
7af57294 1707{
52ee2dfd 1708 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1709}
1710
7af57294
PE
1711static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1712{
52ee2dfd 1713 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1714}
1715
1716
52ee2dfd
ON
1717static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1718 struct pid_namespace *ns)
7af57294 1719{
52ee2dfd 1720 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1721}
1722
7af57294
PE
1723static inline pid_t task_session_vnr(struct task_struct *tsk)
1724{
52ee2dfd 1725 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1726}
1727
1b0f7ffd
ON
1728/* obsolete, do not use */
1729static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1730{
1731 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1732}
7af57294 1733
1da177e4
LT
1734/**
1735 * pid_alive - check that a task structure is not stale
1736 * @p: Task structure to be checked.
1737 *
1738 * Test if a process is not yet dead (at most zombie state)
1739 * If pid_alive fails, then pointers within the task structure
1740 * can be stale and must not be dereferenced.
e69f6186
YB
1741 *
1742 * Return: 1 if the process is alive. 0 otherwise.
1da177e4 1743 */
e868171a 1744static inline int pid_alive(struct task_struct *p)
1da177e4 1745{
92476d7f 1746 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1747}
1748
f400e198 1749/**
b460cbc5 1750 * is_global_init - check if a task structure is init
3260259f
H
1751 * @tsk: Task structure to be checked.
1752 *
1753 * Check if a task structure is the first user space task the kernel created.
e69f6186
YB
1754 *
1755 * Return: 1 if the task structure is init. 0 otherwise.
b460cbc5 1756 */
e868171a 1757static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1758{
1759 return tsk->pid == 1;
1760}
b460cbc5 1761
9ec52099
CLG
1762extern struct pid *cad_pid;
1763
1da177e4 1764extern void free_task(struct task_struct *tsk);
1da177e4 1765#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1766
158d9ebd 1767extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1768
1769static inline void put_task_struct(struct task_struct *t)
1770{
1771 if (atomic_dec_and_test(&t->usage))
8c7904a0 1772 __put_task_struct(t);
e56d0903 1773}
1da177e4 1774
6a61671b
FW
1775#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1776extern void task_cputime(struct task_struct *t,
1777 cputime_t *utime, cputime_t *stime);
1778extern void task_cputime_scaled(struct task_struct *t,
1779 cputime_t *utimescaled, cputime_t *stimescaled);
1780extern cputime_t task_gtime(struct task_struct *t);
1781#else
6fac4829
FW
1782static inline void task_cputime(struct task_struct *t,
1783 cputime_t *utime, cputime_t *stime)
1784{
1785 if (utime)
1786 *utime = t->utime;
1787 if (stime)
1788 *stime = t->stime;
1789}
1790
1791static inline void task_cputime_scaled(struct task_struct *t,
1792 cputime_t *utimescaled,
1793 cputime_t *stimescaled)
1794{
1795 if (utimescaled)
1796 *utimescaled = t->utimescaled;
1797 if (stimescaled)
1798 *stimescaled = t->stimescaled;
1799}
6a61671b
FW
1800
1801static inline cputime_t task_gtime(struct task_struct *t)
1802{
1803 return t->gtime;
1804}
1805#endif
e80d0a1a
FW
1806extern void task_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
1807extern void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1808
1da177e4
LT
1809/*
1810 * Per process flags
1811 */
1da177e4 1812#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1813#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1814#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1815#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1816#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1817#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1818#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1819#define PF_DUMPCORE 0x00000200 /* dumped core */
1820#define PF_SIGNALED 0x00000400 /* killed by a signal */
1821#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1822#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1823#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
774a1221 1824#define PF_USED_ASYNC 0x00004000 /* used async_schedule*(), used by module init */
1da177e4
LT
1825#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1826#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1827#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1828#define PF_KSWAPD 0x00040000 /* I am kswapd */
21caf2fc 1829#define PF_MEMALLOC_NOIO 0x00080000 /* Allocating memory without IO involved */
1da177e4 1830#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1831#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1832#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1833#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1834#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1835#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
14a40ffc 1836#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_allowed */
4db96cf0 1837#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1838#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1839#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1840#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
2b44c4db 1841#define PF_SUSPEND_TASK 0x80000000 /* this thread called freeze_processes and should not be frozen */
1da177e4
LT
1842
1843/*
1844 * Only the _current_ task can read/write to tsk->flags, but other
1845 * tasks can access tsk->flags in readonly mode for example
1846 * with tsk_used_math (like during threaded core dumping).
1847 * There is however an exception to this rule during ptrace
1848 * or during fork: the ptracer task is allowed to write to the
1849 * child->flags of its traced child (same goes for fork, the parent
1850 * can write to the child->flags), because we're guaranteed the
1851 * child is not running and in turn not changing child->flags
1852 * at the same time the parent does it.
1853 */
1854#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1855#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1856#define clear_used_math() clear_stopped_child_used_math(current)
1857#define set_used_math() set_stopped_child_used_math(current)
1858#define conditional_stopped_child_used_math(condition, child) \
1859 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1860#define conditional_used_math(condition) \
1861 conditional_stopped_child_used_math(condition, current)
1862#define copy_to_stopped_child_used_math(child) \
1863 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1864/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1865#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1866#define used_math() tsk_used_math(current)
1867
21caf2fc
ML
1868/* __GFP_IO isn't allowed if PF_MEMALLOC_NOIO is set in current->flags */
1869static inline gfp_t memalloc_noio_flags(gfp_t flags)
1870{
1871 if (unlikely(current->flags & PF_MEMALLOC_NOIO))
1872 flags &= ~__GFP_IO;
1873 return flags;
1874}
1875
1876static inline unsigned int memalloc_noio_save(void)
1877{
1878 unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
1879 current->flags |= PF_MEMALLOC_NOIO;
1880 return flags;
1881}
1882
1883static inline void memalloc_noio_restore(unsigned int flags)
1884{
1885 current->flags = (current->flags & ~PF_MEMALLOC_NOIO) | flags;
1886}
1887
e5c1902e 1888/*
a8f072c1 1889 * task->jobctl flags
e5c1902e 1890 */
a8f072c1 1891#define JOBCTL_STOP_SIGMASK 0xffff /* signr of the last group stop */
e5c1902e 1892
a8f072c1
TH
1893#define JOBCTL_STOP_DEQUEUED_BIT 16 /* stop signal dequeued */
1894#define JOBCTL_STOP_PENDING_BIT 17 /* task should stop for group stop */
1895#define JOBCTL_STOP_CONSUME_BIT 18 /* consume group stop count */
73ddff2b 1896#define JOBCTL_TRAP_STOP_BIT 19 /* trap for STOP */
fb1d910c 1897#define JOBCTL_TRAP_NOTIFY_BIT 20 /* trap for NOTIFY */
a8f072c1 1898#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */
544b2c91 1899#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */
a8f072c1
TH
1900
1901#define JOBCTL_STOP_DEQUEUED (1 << JOBCTL_STOP_DEQUEUED_BIT)
1902#define JOBCTL_STOP_PENDING (1 << JOBCTL_STOP_PENDING_BIT)
1903#define JOBCTL_STOP_CONSUME (1 << JOBCTL_STOP_CONSUME_BIT)
73ddff2b 1904#define JOBCTL_TRAP_STOP (1 << JOBCTL_TRAP_STOP_BIT)
fb1d910c 1905#define JOBCTL_TRAP_NOTIFY (1 << JOBCTL_TRAP_NOTIFY_BIT)
a8f072c1 1906#define JOBCTL_TRAPPING (1 << JOBCTL_TRAPPING_BIT)
544b2c91 1907#define JOBCTL_LISTENING (1 << JOBCTL_LISTENING_BIT)
a8f072c1 1908
fb1d910c 1909#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
73ddff2b 1910#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK)
3759a0d9 1911
7dd3db54
TH
1912extern bool task_set_jobctl_pending(struct task_struct *task,
1913 unsigned int mask);
73ddff2b 1914extern void task_clear_jobctl_trapping(struct task_struct *task);
3759a0d9
TH
1915extern void task_clear_jobctl_pending(struct task_struct *task,
1916 unsigned int mask);
39efa3ef 1917
a57eb940 1918#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1919
1920#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1aa03f11 1921#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1922
1923static inline void rcu_copy_process(struct task_struct *p)
1924{
1925 p->rcu_read_lock_nesting = 0;
1926 p->rcu_read_unlock_special = 0;
a57eb940 1927#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1928 p->rcu_blocked_node = NULL;
24278d14
PM
1929#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1930#ifdef CONFIG_RCU_BOOST
1931 p->rcu_boost_mutex = NULL;
1932#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1933 INIT_LIST_HEAD(&p->rcu_node_entry);
1934}
1935
f41d911f
PM
1936#else
1937
1938static inline void rcu_copy_process(struct task_struct *p)
1939{
1940}
1941
1942#endif
1943
907aed48
MG
1944static inline void tsk_restore_flags(struct task_struct *task,
1945 unsigned long orig_flags, unsigned long flags)
1946{
1947 task->flags &= ~flags;
1948 task->flags |= orig_flags & flags;
1949}
1950
1da177e4 1951#ifdef CONFIG_SMP
1e1b6c51
KM
1952extern void do_set_cpus_allowed(struct task_struct *p,
1953 const struct cpumask *new_mask);
1954
cd8ba7cd 1955extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1956 const struct cpumask *new_mask);
1da177e4 1957#else
1e1b6c51
KM
1958static inline void do_set_cpus_allowed(struct task_struct *p,
1959 const struct cpumask *new_mask)
1960{
1961}
cd8ba7cd 1962static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1963 const struct cpumask *new_mask)
1da177e4 1964{
96f874e2 1965 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1966 return -EINVAL;
1967 return 0;
1968}
1969#endif
e0ad9556 1970
3451d024 1971#ifdef CONFIG_NO_HZ_COMMON
5167e8d5
PZ
1972void calc_load_enter_idle(void);
1973void calc_load_exit_idle(void);
1974#else
1975static inline void calc_load_enter_idle(void) { }
1976static inline void calc_load_exit_idle(void) { }
3451d024 1977#endif /* CONFIG_NO_HZ_COMMON */
5167e8d5 1978
e0ad9556 1979#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1980static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1981{
1982 return set_cpus_allowed_ptr(p, &new_mask);
1983}
e0ad9556 1984#endif
1da177e4 1985
b342501c 1986/*
c676329a
PZ
1987 * Do not use outside of architecture code which knows its limitations.
1988 *
1989 * sched_clock() has no promise of monotonicity or bounded drift between
1990 * CPUs, use (which you should not) requires disabling IRQs.
1991 *
1992 * Please use one of the three interfaces below.
b342501c 1993 */
1bbfa6f2 1994extern unsigned long long notrace sched_clock(void);
c676329a 1995/*
489a71b0 1996 * See the comment in kernel/sched/clock.c
c676329a
PZ
1997 */
1998extern u64 cpu_clock(int cpu);
1999extern u64 local_clock(void);
2000extern u64 sched_clock_cpu(int cpu);
2001
e436d800 2002
c1955a3d 2003extern void sched_clock_init(void);
3e51f33f 2004
c1955a3d 2005#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
2006static inline void sched_clock_tick(void)
2007{
2008}
2009
2010static inline void sched_clock_idle_sleep_event(void)
2011{
2012}
2013
2014static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
2015{
2016}
2017#else
c676329a
PZ
2018/*
2019 * Architectures can set this to 1 if they have specified
2020 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
2021 * but then during bootup it turns out that sched_clock()
2022 * is reliable after all:
2023 */
35af99e6
PZ
2024extern int sched_clock_stable(void);
2025extern void set_sched_clock_stable(void);
2026extern void clear_sched_clock_stable(void);
c676329a 2027
3e51f33f
PZ
2028extern void sched_clock_tick(void);
2029extern void sched_clock_idle_sleep_event(void);
2030extern void sched_clock_idle_wakeup_event(u64 delta_ns);
2031#endif
2032
b52bfee4
VP
2033#ifdef CONFIG_IRQ_TIME_ACCOUNTING
2034/*
2035 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
2036 * The reason for this explicit opt-in is not to have perf penalty with
2037 * slow sched_clocks.
2038 */
2039extern void enable_sched_clock_irqtime(void);
2040extern void disable_sched_clock_irqtime(void);
2041#else
2042static inline void enable_sched_clock_irqtime(void) {}
2043static inline void disable_sched_clock_irqtime(void) {}
2044#endif
2045
36c8b586 2046extern unsigned long long
41b86e9c 2047task_sched_runtime(struct task_struct *task);
1da177e4
LT
2048
2049/* sched_exec is called by processes performing an exec */
2050#ifdef CONFIG_SMP
2051extern void sched_exec(void);
2052#else
2053#define sched_exec() {}
2054#endif
2055
2aa44d05
IM
2056extern void sched_clock_idle_sleep_event(void);
2057extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 2058
1da177e4
LT
2059#ifdef CONFIG_HOTPLUG_CPU
2060extern void idle_task_exit(void);
2061#else
2062static inline void idle_task_exit(void) {}
2063#endif
2064
3451d024 2065#if defined(CONFIG_NO_HZ_COMMON) && defined(CONFIG_SMP)
1c20091e 2066extern void wake_up_nohz_cpu(int cpu);
06d8308c 2067#else
1c20091e 2068static inline void wake_up_nohz_cpu(int cpu) { }
06d8308c
TG
2069#endif
2070
ce831b38
FW
2071#ifdef CONFIG_NO_HZ_FULL
2072extern bool sched_can_stop_tick(void);
265f22a9 2073extern u64 scheduler_tick_max_deferment(void);
ce831b38
FW
2074#else
2075static inline bool sched_can_stop_tick(void) { return false; }
06d8308c
TG
2076#endif
2077
5091faa4 2078#ifdef CONFIG_SCHED_AUTOGROUP
5091faa4
MG
2079extern void sched_autogroup_create_attach(struct task_struct *p);
2080extern void sched_autogroup_detach(struct task_struct *p);
2081extern void sched_autogroup_fork(struct signal_struct *sig);
2082extern void sched_autogroup_exit(struct signal_struct *sig);
2083#ifdef CONFIG_PROC_FS
2084extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
2e5b5b3a 2085extern int proc_sched_autogroup_set_nice(struct task_struct *p, int nice);
5091faa4
MG
2086#endif
2087#else
2088static inline void sched_autogroup_create_attach(struct task_struct *p) { }
2089static inline void sched_autogroup_detach(struct task_struct *p) { }
2090static inline void sched_autogroup_fork(struct signal_struct *sig) { }
2091static inline void sched_autogroup_exit(struct signal_struct *sig) { }
2092#endif
2093
d95f4122 2094extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
2095extern void set_user_nice(struct task_struct *p, long nice);
2096extern int task_prio(const struct task_struct *p);
d0ea0268
DY
2097/**
2098 * task_nice - return the nice value of a given task.
2099 * @p: the task in question.
2100 *
2101 * Return: The nice value [ -20 ... 0 ... 19 ].
2102 */
2103static inline int task_nice(const struct task_struct *p)
2104{
2105 return PRIO_TO_NICE((p)->static_prio);
2106}
36c8b586
IM
2107extern int can_nice(const struct task_struct *p, const int nice);
2108extern int task_curr(const struct task_struct *p);
1da177e4 2109extern int idle_cpu(int cpu);
fe7de49f
KM
2110extern int sched_setscheduler(struct task_struct *, int,
2111 const struct sched_param *);
961ccddd 2112extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 2113 const struct sched_param *);
d50dde5a
DF
2114extern int sched_setattr(struct task_struct *,
2115 const struct sched_attr *);
36c8b586 2116extern struct task_struct *idle_task(int cpu);
c4f30608
PM
2117/**
2118 * is_idle_task - is the specified task an idle task?
fa757281 2119 * @p: the task in question.
e69f6186
YB
2120 *
2121 * Return: 1 if @p is an idle task. 0 otherwise.
c4f30608 2122 */
7061ca3b 2123static inline bool is_idle_task(const struct task_struct *p)
c4f30608
PM
2124{
2125 return p->pid == 0;
2126}
36c8b586
IM
2127extern struct task_struct *curr_task(int cpu);
2128extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
2129
2130void yield(void);
2131
2132/*
2133 * The default (Linux) execution domain.
2134 */
2135extern struct exec_domain default_exec_domain;
2136
2137union thread_union {
2138 struct thread_info thread_info;
2139 unsigned long stack[THREAD_SIZE/sizeof(long)];
2140};
2141
2142#ifndef __HAVE_ARCH_KSTACK_END
2143static inline int kstack_end(void *addr)
2144{
2145 /* Reliable end of stack detection:
2146 * Some APM bios versions misalign the stack
2147 */
2148 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2149}
2150#endif
2151
2152extern union thread_union init_thread_union;
2153extern struct task_struct init_task;
2154
2155extern struct mm_struct init_mm;
2156
198fe21b
PE
2157extern struct pid_namespace init_pid_ns;
2158
2159/*
2160 * find a task by one of its numerical ids
2161 *
198fe21b
PE
2162 * find_task_by_pid_ns():
2163 * finds a task by its pid in the specified namespace
228ebcbe
PE
2164 * find_task_by_vpid():
2165 * finds a task by its virtual pid
198fe21b 2166 *
e49859e7 2167 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2168 */
2169
228ebcbe
PE
2170extern struct task_struct *find_task_by_vpid(pid_t nr);
2171extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2172 struct pid_namespace *ns);
198fe21b 2173
1da177e4 2174/* per-UID process charging. */
7b44ab97 2175extern struct user_struct * alloc_uid(kuid_t);
1da177e4
LT
2176static inline struct user_struct *get_uid(struct user_struct *u)
2177{
2178 atomic_inc(&u->__count);
2179 return u;
2180}
2181extern void free_uid(struct user_struct *);
1da177e4
LT
2182
2183#include <asm/current.h>
2184
f0af911a 2185extern void xtime_update(unsigned long ticks);
1da177e4 2186
b3c97528
HH
2187extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2188extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 2189extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
2190#ifdef CONFIG_SMP
2191 extern void kick_process(struct task_struct *tsk);
2192#else
2193 static inline void kick_process(struct task_struct *tsk) { }
2194#endif
aab03e05 2195extern int sched_fork(unsigned long clone_flags, struct task_struct *p);
ad46c2c4 2196extern void sched_dead(struct task_struct *p);
1da177e4 2197
1da177e4
LT
2198extern void proc_caches_init(void);
2199extern void flush_signals(struct task_struct *);
3bcac026 2200extern void __flush_signals(struct task_struct *);
10ab825b 2201extern void ignore_signals(struct task_struct *);
1da177e4
LT
2202extern void flush_signal_handlers(struct task_struct *, int force_default);
2203extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2204
2205static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2206{
2207 unsigned long flags;
2208 int ret;
2209
2210 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2211 ret = dequeue_signal(tsk, mask, info);
2212 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2213
2214 return ret;
53c8f9f1 2215}
1da177e4
LT
2216
2217extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2218 sigset_t *mask);
2219extern void unblock_all_signals(void);
2220extern void release_task(struct task_struct * p);
2221extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2222extern int force_sigsegv(int, struct task_struct *);
2223extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2224extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2225extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
d178bc3a
SH
2226extern int kill_pid_info_as_cred(int, struct siginfo *, struct pid *,
2227 const struct cred *, u32);
c4b92fc1
EB
2228extern int kill_pgrp(struct pid *pid, int sig, int priv);
2229extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2230extern int kill_proc_info(int, struct siginfo *, pid_t);
86773473 2231extern __must_check bool do_notify_parent(struct task_struct *, int);
a7f0765e 2232extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2233extern void force_sig(int, struct task_struct *);
1da177e4 2234extern int send_sig(int, struct task_struct *, int);
09faef11 2235extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2236extern struct sigqueue *sigqueue_alloc(void);
2237extern void sigqueue_free(struct sigqueue *);
ac5c2153 2238extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2239extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4 2240
51a7b448
AV
2241static inline void restore_saved_sigmask(void)
2242{
2243 if (test_and_clear_restore_sigmask())
77097ae5 2244 __set_current_blocked(&current->saved_sigmask);
51a7b448
AV
2245}
2246
b7f9a11a
AV
2247static inline sigset_t *sigmask_to_save(void)
2248{
2249 sigset_t *res = &current->blocked;
2250 if (unlikely(test_restore_sigmask()))
2251 res = &current->saved_sigmask;
2252 return res;
2253}
2254
9ec52099
CLG
2255static inline int kill_cad_pid(int sig, int priv)
2256{
2257 return kill_pid(cad_pid, sig, priv);
2258}
2259
1da177e4
LT
2260/* These can be the second arg to send_sig_info/send_group_sig_info. */
2261#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2262#define SEND_SIG_PRIV ((struct siginfo *) 1)
2263#define SEND_SIG_FORCED ((struct siginfo *) 2)
2264
2a855dd0
SAS
2265/*
2266 * True if we are on the alternate signal stack.
2267 */
1da177e4
LT
2268static inline int on_sig_stack(unsigned long sp)
2269{
2a855dd0
SAS
2270#ifdef CONFIG_STACK_GROWSUP
2271 return sp >= current->sas_ss_sp &&
2272 sp - current->sas_ss_sp < current->sas_ss_size;
2273#else
2274 return sp > current->sas_ss_sp &&
2275 sp - current->sas_ss_sp <= current->sas_ss_size;
2276#endif
1da177e4
LT
2277}
2278
2279static inline int sas_ss_flags(unsigned long sp)
2280{
2281 return (current->sas_ss_size == 0 ? SS_DISABLE
2282 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2283}
2284
5a1b98d3
AV
2285static inline unsigned long sigsp(unsigned long sp, struct ksignal *ksig)
2286{
2287 if (unlikely((ksig->ka.sa.sa_flags & SA_ONSTACK)) && ! sas_ss_flags(sp))
2288#ifdef CONFIG_STACK_GROWSUP
2289 return current->sas_ss_sp;
2290#else
2291 return current->sas_ss_sp + current->sas_ss_size;
2292#endif
2293 return sp;
2294}
2295
1da177e4
LT
2296/*
2297 * Routines for handling mm_structs
2298 */
2299extern struct mm_struct * mm_alloc(void);
2300
2301/* mmdrop drops the mm and the page tables */
b3c97528 2302extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2303static inline void mmdrop(struct mm_struct * mm)
2304{
6fb43d7b 2305 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2306 __mmdrop(mm);
2307}
2308
2309/* mmput gets rid of the mappings and all user-space */
2310extern void mmput(struct mm_struct *);
2311/* Grab a reference to a task's mm, if it is not already going away */
2312extern struct mm_struct *get_task_mm(struct task_struct *task);
8cdb878d
CY
2313/*
2314 * Grab a reference to a task's mm, if it is not already going away
2315 * and ptrace_may_access with the mode parameter passed to it
2316 * succeeds.
2317 */
2318extern struct mm_struct *mm_access(struct task_struct *task, unsigned int mode);
1da177e4
LT
2319/* Remove the current tasks stale references to the old mm_struct */
2320extern void mm_release(struct task_struct *, struct mm_struct *);
2321
6f2c55b8 2322extern int copy_thread(unsigned long, unsigned long, unsigned long,
afa86fc4 2323 struct task_struct *);
1da177e4
LT
2324extern void flush_thread(void);
2325extern void exit_thread(void);
2326
1da177e4 2327extern void exit_files(struct task_struct *);
a7e5328a 2328extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2329
1da177e4 2330extern void exit_itimers(struct signal_struct *);
cbaffba1 2331extern void flush_itimer_signals(void);
1da177e4 2332
9402c95f 2333extern void do_group_exit(int);
1da177e4 2334
1da177e4
LT
2335extern int allow_signal(int);
2336extern int disallow_signal(int);
1da177e4 2337
d7627467
DH
2338extern int do_execve(const char *,
2339 const char __user * const __user *,
da3d4c5f 2340 const char __user * const __user *);
e80d6661 2341extern long do_fork(unsigned long, unsigned long, unsigned long, int __user *, int __user *);
36c8b586 2342struct task_struct *fork_idle(int);
2aa3a7f8 2343extern pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
1da177e4
LT
2344
2345extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2346extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2347
2348#ifdef CONFIG_SMP
317f3941 2349void scheduler_ipi(void);
85ba2d86 2350extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2351#else
184748cc 2352static inline void scheduler_ipi(void) { }
85ba2d86
RM
2353static inline unsigned long wait_task_inactive(struct task_struct *p,
2354 long match_state)
2355{
2356 return 1;
2357}
1da177e4
LT
2358#endif
2359
05725f7e
JP
2360#define next_task(p) \
2361 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2362
2363#define for_each_process(p) \
2364 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2365
5bb459bb 2366extern bool current_is_single_threaded(void);
d84f4f99 2367
1da177e4
LT
2368/*
2369 * Careful: do_each_thread/while_each_thread is a double loop so
2370 * 'break' will not work as expected - use goto instead.
2371 */
2372#define do_each_thread(g, t) \
2373 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2374
2375#define while_each_thread(g, t) \
2376 while ((t = next_thread(t)) != g)
2377
0c740d0a
ON
2378#define __for_each_thread(signal, t) \
2379 list_for_each_entry_rcu(t, &(signal)->thread_head, thread_node)
2380
2381#define for_each_thread(p, t) \
2382 __for_each_thread((p)->signal, t)
2383
2384/* Careful: this is a double loop, 'break' won't work as expected. */
2385#define for_each_process_thread(p, t) \
2386 for_each_process(p) for_each_thread(p, t)
2387
7e49827c
ON
2388static inline int get_nr_threads(struct task_struct *tsk)
2389{
b3ac022c 2390 return tsk->signal->nr_threads;
7e49827c
ON
2391}
2392
087806b1
ON
2393static inline bool thread_group_leader(struct task_struct *p)
2394{
2395 return p->exit_signal >= 0;
2396}
1da177e4 2397
0804ef4b
EB
2398/* Do to the insanities of de_thread it is possible for a process
2399 * to have the pid of the thread group leader without actually being
2400 * the thread group leader. For iteration through the pids in proc
2401 * all we care about is that we have a task with the appropriate
2402 * pid, we don't actually care if we have the right task.
2403 */
e1403b8e 2404static inline bool has_group_leader_pid(struct task_struct *p)
0804ef4b 2405{
e1403b8e 2406 return task_pid(p) == p->signal->leader_pid;
0804ef4b
EB
2407}
2408
bac0abd6 2409static inline
e1403b8e 2410bool same_thread_group(struct task_struct *p1, struct task_struct *p2)
bac0abd6 2411{
e1403b8e 2412 return p1->signal == p2->signal;
bac0abd6
PE
2413}
2414
36c8b586 2415static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2416{
05725f7e
JP
2417 return list_entry_rcu(p->thread_group.next,
2418 struct task_struct, thread_group);
47e65328
ON
2419}
2420
e868171a 2421static inline int thread_group_empty(struct task_struct *p)
1da177e4 2422{
47e65328 2423 return list_empty(&p->thread_group);
1da177e4
LT
2424}
2425
2426#define delay_group_leader(p) \
2427 (thread_group_leader(p) && !thread_group_empty(p))
2428
1da177e4 2429/*
260ea101 2430 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2431 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8 2432 * pins the final release of task.io_context. Also protects ->cpuset and
d68b46fe 2433 * ->cgroup.subsys[]. And ->vfork_done.
1da177e4
LT
2434 *
2435 * Nests both inside and outside of read_lock(&tasklist_lock).
2436 * It must not be nested with write_lock_irq(&tasklist_lock),
2437 * neither inside nor outside.
2438 */
2439static inline void task_lock(struct task_struct *p)
2440{
2441 spin_lock(&p->alloc_lock);
2442}
2443
2444static inline void task_unlock(struct task_struct *p)
2445{
2446 spin_unlock(&p->alloc_lock);
2447}
2448
b8ed374e 2449extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2450 unsigned long *flags);
2451
9388dc30
AV
2452static inline struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2453 unsigned long *flags)
2454{
2455 struct sighand_struct *ret;
2456
2457 ret = __lock_task_sighand(tsk, flags);
2458 (void)__cond_lock(&tsk->sighand->siglock, ret);
2459 return ret;
2460}
b8ed374e 2461
f63ee72e
ON
2462static inline void unlock_task_sighand(struct task_struct *tsk,
2463 unsigned long *flags)
2464{
2465 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2466}
2467
4714d1d3 2468#ifdef CONFIG_CGROUPS
257058ae 2469static inline void threadgroup_change_begin(struct task_struct *tsk)
4714d1d3 2470{
257058ae 2471 down_read(&tsk->signal->group_rwsem);
4714d1d3 2472}
257058ae 2473static inline void threadgroup_change_end(struct task_struct *tsk)
4714d1d3 2474{
257058ae 2475 up_read(&tsk->signal->group_rwsem);
4714d1d3 2476}
77e4ef99
TH
2477
2478/**
2479 * threadgroup_lock - lock threadgroup
2480 * @tsk: member task of the threadgroup to lock
2481 *
2482 * Lock the threadgroup @tsk belongs to. No new task is allowed to enter
2483 * and member tasks aren't allowed to exit (as indicated by PF_EXITING) or
e56fb287
ON
2484 * change ->group_leader/pid. This is useful for cases where the threadgroup
2485 * needs to stay stable across blockable operations.
77e4ef99
TH
2486 *
2487 * fork and exit paths explicitly call threadgroup_change_{begin|end}() for
2488 * synchronization. While held, no new task will be added to threadgroup
2489 * and no existing live task will have its PF_EXITING set.
2490 *
e56fb287
ON
2491 * de_thread() does threadgroup_change_{begin|end}() when a non-leader
2492 * sub-thread becomes a new leader.
77e4ef99 2493 */
257058ae 2494static inline void threadgroup_lock(struct task_struct *tsk)
4714d1d3 2495{
257058ae 2496 down_write(&tsk->signal->group_rwsem);
4714d1d3 2497}
77e4ef99
TH
2498
2499/**
2500 * threadgroup_unlock - unlock threadgroup
2501 * @tsk: member task of the threadgroup to unlock
2502 *
2503 * Reverse threadgroup_lock().
2504 */
257058ae 2505static inline void threadgroup_unlock(struct task_struct *tsk)
4714d1d3 2506{
257058ae 2507 up_write(&tsk->signal->group_rwsem);
4714d1d3
BB
2508}
2509#else
257058ae
TH
2510static inline void threadgroup_change_begin(struct task_struct *tsk) {}
2511static inline void threadgroup_change_end(struct task_struct *tsk) {}
2512static inline void threadgroup_lock(struct task_struct *tsk) {}
2513static inline void threadgroup_unlock(struct task_struct *tsk) {}
4714d1d3
BB
2514#endif
2515
f037360f
AV
2516#ifndef __HAVE_THREAD_FUNCTIONS
2517
f7e4217b
RZ
2518#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2519#define task_stack_page(task) ((task)->stack)
a1261f54 2520
10ebffde
AV
2521static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2522{
2523 *task_thread_info(p) = *task_thread_info(org);
2524 task_thread_info(p)->task = p;
2525}
2526
2527static inline unsigned long *end_of_stack(struct task_struct *p)
2528{
f7e4217b 2529 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2530}
2531
f037360f
AV
2532#endif
2533
8b05c7e6
FT
2534static inline int object_is_on_stack(void *obj)
2535{
2536 void *stack = task_stack_page(current);
2537
2538 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2539}
2540
8c9843e5
BH
2541extern void thread_info_cache_init(void);
2542
7c9f8861
ES
2543#ifdef CONFIG_DEBUG_STACK_USAGE
2544static inline unsigned long stack_not_used(struct task_struct *p)
2545{
2546 unsigned long *n = end_of_stack(p);
2547
2548 do { /* Skip over canary */
2549 n++;
2550 } while (!*n);
2551
2552 return (unsigned long)n - (unsigned long)end_of_stack(p);
2553}
2554#endif
2555
1da177e4
LT
2556/* set thread flags in other task's structures
2557 * - see asm/thread_info.h for TIF_xxxx flags available
2558 */
2559static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2560{
a1261f54 2561 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2562}
2563
2564static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2565{
a1261f54 2566 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2567}
2568
2569static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2570{
a1261f54 2571 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2572}
2573
2574static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2575{
a1261f54 2576 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2577}
2578
2579static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2580{
a1261f54 2581 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2582}
2583
2584static inline void set_tsk_need_resched(struct task_struct *tsk)
2585{
2586 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2587}
2588
2589static inline void clear_tsk_need_resched(struct task_struct *tsk)
2590{
2591 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2592}
2593
8ae121ac
GH
2594static inline int test_tsk_need_resched(struct task_struct *tsk)
2595{
2596 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2597}
2598
690cc3ff
EB
2599static inline int restart_syscall(void)
2600{
2601 set_tsk_thread_flag(current, TIF_SIGPENDING);
2602 return -ERESTARTNOINTR;
2603}
2604
1da177e4
LT
2605static inline int signal_pending(struct task_struct *p)
2606{
2607 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2608}
f776d12d 2609
d9588725
RM
2610static inline int __fatal_signal_pending(struct task_struct *p)
2611{
2612 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2613}
f776d12d
MW
2614
2615static inline int fatal_signal_pending(struct task_struct *p)
2616{
2617 return signal_pending(p) && __fatal_signal_pending(p);
2618}
2619
16882c1e
ON
2620static inline int signal_pending_state(long state, struct task_struct *p)
2621{
2622 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2623 return 0;
2624 if (!signal_pending(p))
2625 return 0;
2626
16882c1e
ON
2627 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2628}
2629
1da177e4
LT
2630/*
2631 * cond_resched() and cond_resched_lock(): latency reduction via
2632 * explicit rescheduling in places that are safe. The return
2633 * value indicates whether a reschedule was done in fact.
2634 * cond_resched_lock() will drop the spinlock before scheduling,
2635 * cond_resched_softirq() will enable bhs before scheduling.
2636 */
c3921ab7 2637extern int _cond_resched(void);
6f80bd98 2638
613afbf8
FW
2639#define cond_resched() ({ \
2640 __might_sleep(__FILE__, __LINE__, 0); \
2641 _cond_resched(); \
2642})
6f80bd98 2643
613afbf8
FW
2644extern int __cond_resched_lock(spinlock_t *lock);
2645
bdd4e85d 2646#ifdef CONFIG_PREEMPT_COUNT
716a4234 2647#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2648#else
716a4234 2649#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2650#endif
716a4234 2651
613afbf8 2652#define cond_resched_lock(lock) ({ \
716a4234 2653 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2654 __cond_resched_lock(lock); \
2655})
2656
2657extern int __cond_resched_softirq(void);
2658
75e1056f
VP
2659#define cond_resched_softirq() ({ \
2660 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2661 __cond_resched_softirq(); \
613afbf8 2662})
1da177e4 2663
f6f3c437
SH
2664static inline void cond_resched_rcu(void)
2665{
2666#if defined(CONFIG_DEBUG_ATOMIC_SLEEP) || !defined(CONFIG_PREEMPT_RCU)
2667 rcu_read_unlock();
2668 cond_resched();
2669 rcu_read_lock();
2670#endif
2671}
2672
1da177e4
LT
2673/*
2674 * Does a critical section need to be broken due to another
95c354fe
NP
2675 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2676 * but a general need for low latency)
1da177e4 2677 */
95c354fe 2678static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2679{
95c354fe
NP
2680#ifdef CONFIG_PREEMPT
2681 return spin_is_contended(lock);
2682#else
1da177e4 2683 return 0;
95c354fe 2684#endif
1da177e4
LT
2685}
2686
ee761f62
TG
2687/*
2688 * Idle thread specific functions to determine the need_resched
2689 * polling state. We have two versions, one based on TS_POLLING in
2690 * thread_info.status and one based on TIF_POLLING_NRFLAG in
2691 * thread_info.flags
2692 */
2693#ifdef TS_POLLING
2694static inline int tsk_is_polling(struct task_struct *p)
2695{
2696 return task_thread_info(p)->status & TS_POLLING;
2697}
ea811747 2698static inline void __current_set_polling(void)
3a98f871
TG
2699{
2700 current_thread_info()->status |= TS_POLLING;
2701}
2702
ea811747
PZ
2703static inline bool __must_check current_set_polling_and_test(void)
2704{
2705 __current_set_polling();
2706
2707 /*
2708 * Polling state must be visible before we test NEED_RESCHED,
2709 * paired by resched_task()
2710 */
2711 smp_mb();
2712
2713 return unlikely(tif_need_resched());
2714}
2715
2716static inline void __current_clr_polling(void)
3a98f871
TG
2717{
2718 current_thread_info()->status &= ~TS_POLLING;
ea811747
PZ
2719}
2720
2721static inline bool __must_check current_clr_polling_and_test(void)
2722{
2723 __current_clr_polling();
2724
2725 /*
2726 * Polling state must be visible before we test NEED_RESCHED,
2727 * paired by resched_task()
2728 */
2729 smp_mb();
2730
2731 return unlikely(tif_need_resched());
3a98f871 2732}
ee761f62
TG
2733#elif defined(TIF_POLLING_NRFLAG)
2734static inline int tsk_is_polling(struct task_struct *p)
2735{
2736 return test_tsk_thread_flag(p, TIF_POLLING_NRFLAG);
2737}
ea811747
PZ
2738
2739static inline void __current_set_polling(void)
3a98f871
TG
2740{
2741 set_thread_flag(TIF_POLLING_NRFLAG);
2742}
2743
ea811747
PZ
2744static inline bool __must_check current_set_polling_and_test(void)
2745{
2746 __current_set_polling();
2747
2748 /*
2749 * Polling state must be visible before we test NEED_RESCHED,
2750 * paired by resched_task()
2751 *
2752 * XXX: assumes set/clear bit are identical barrier wise.
2753 */
2754 smp_mb__after_clear_bit();
2755
2756 return unlikely(tif_need_resched());
2757}
2758
2759static inline void __current_clr_polling(void)
3a98f871
TG
2760{
2761 clear_thread_flag(TIF_POLLING_NRFLAG);
2762}
ea811747
PZ
2763
2764static inline bool __must_check current_clr_polling_and_test(void)
2765{
2766 __current_clr_polling();
2767
2768 /*
2769 * Polling state must be visible before we test NEED_RESCHED,
2770 * paired by resched_task()
2771 */
2772 smp_mb__after_clear_bit();
2773
2774 return unlikely(tif_need_resched());
2775}
2776
ee761f62
TG
2777#else
2778static inline int tsk_is_polling(struct task_struct *p) { return 0; }
ea811747
PZ
2779static inline void __current_set_polling(void) { }
2780static inline void __current_clr_polling(void) { }
2781
2782static inline bool __must_check current_set_polling_and_test(void)
2783{
2784 return unlikely(tif_need_resched());
2785}
2786static inline bool __must_check current_clr_polling_and_test(void)
2787{
2788 return unlikely(tif_need_resched());
2789}
ee761f62
TG
2790#endif
2791
8cb75e0c
PZ
2792static inline void current_clr_polling(void)
2793{
2794 __current_clr_polling();
2795
2796 /*
2797 * Ensure we check TIF_NEED_RESCHED after we clear the polling bit.
2798 * Once the bit is cleared, we'll get IPIs with every new
2799 * TIF_NEED_RESCHED and the IPI handler, scheduler_ipi(), will also
2800 * fold.
2801 */
2802 smp_mb(); /* paired with resched_task() */
2803
2804 preempt_fold_need_resched();
2805}
2806
75f93fed
PZ
2807static __always_inline bool need_resched(void)
2808{
2809 return unlikely(tif_need_resched());
2810}
2811
f06febc9
FM
2812/*
2813 * Thread group CPU time accounting.
2814 */
4cd4c1b4 2815void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2816void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2817
490dea45 2818static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2819{
ee30a7b2 2820 raw_spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2821}
2822
7bb44ade
RM
2823/*
2824 * Reevaluate whether the task has signals pending delivery.
2825 * Wake the task if so.
2826 * This is required every time the blocked sigset_t changes.
2827 * callers must hold sighand->siglock.
2828 */
2829extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2830extern void recalc_sigpending(void);
2831
910ffdb1
ON
2832extern void signal_wake_up_state(struct task_struct *t, unsigned int state);
2833
2834static inline void signal_wake_up(struct task_struct *t, bool resume)
2835{
2836 signal_wake_up_state(t, resume ? TASK_WAKEKILL : 0);
2837}
2838static inline void ptrace_signal_wake_up(struct task_struct *t, bool resume)
2839{
2840 signal_wake_up_state(t, resume ? __TASK_TRACED : 0);
2841}
1da177e4
LT
2842
2843/*
2844 * Wrappers for p->thread_info->cpu access. No-op on UP.
2845 */
2846#ifdef CONFIG_SMP
2847
2848static inline unsigned int task_cpu(const struct task_struct *p)
2849{
a1261f54 2850 return task_thread_info(p)->cpu;
1da177e4
LT
2851}
2852
b32e86b4
IM
2853static inline int task_node(const struct task_struct *p)
2854{
2855 return cpu_to_node(task_cpu(p));
2856}
2857
c65cc870 2858extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2859
2860#else
2861
2862static inline unsigned int task_cpu(const struct task_struct *p)
2863{
2864 return 0;
2865}
2866
2867static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2868{
2869}
2870
2871#endif /* CONFIG_SMP */
2872
96f874e2
RR
2873extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2874extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2875
7c941438 2876#ifdef CONFIG_CGROUP_SCHED
07e06b01 2877extern struct task_group root_task_group;
8323f26c 2878#endif /* CONFIG_CGROUP_SCHED */
9b5b7751 2879
54e99124
DG
2880extern int task_can_switch_user(struct user_struct *up,
2881 struct task_struct *tsk);
2882
4b98d11b
AD
2883#ifdef CONFIG_TASK_XACCT
2884static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2885{
940389b8 2886 tsk->ioac.rchar += amt;
4b98d11b
AD
2887}
2888
2889static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2890{
940389b8 2891 tsk->ioac.wchar += amt;
4b98d11b
AD
2892}
2893
2894static inline void inc_syscr(struct task_struct *tsk)
2895{
940389b8 2896 tsk->ioac.syscr++;
4b98d11b
AD
2897}
2898
2899static inline void inc_syscw(struct task_struct *tsk)
2900{
940389b8 2901 tsk->ioac.syscw++;
4b98d11b
AD
2902}
2903#else
2904static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2905{
2906}
2907
2908static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2909{
2910}
2911
2912static inline void inc_syscr(struct task_struct *tsk)
2913{
2914}
2915
2916static inline void inc_syscw(struct task_struct *tsk)
2917{
2918}
2919#endif
2920
82455257
DH
2921#ifndef TASK_SIZE_OF
2922#define TASK_SIZE_OF(tsk) TASK_SIZE
2923#endif
2924
cf475ad2
BS
2925#ifdef CONFIG_MM_OWNER
2926extern void mm_update_next_owner(struct mm_struct *mm);
2927extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2928#else
2929static inline void mm_update_next_owner(struct mm_struct *mm)
2930{
2931}
2932
2933static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2934{
2935}
2936#endif /* CONFIG_MM_OWNER */
2937
3e10e716
JS
2938static inline unsigned long task_rlimit(const struct task_struct *tsk,
2939 unsigned int limit)
2940{
2941 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2942}
2943
2944static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2945 unsigned int limit)
2946{
2947 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2948}
2949
2950static inline unsigned long rlimit(unsigned int limit)
2951{
2952 return task_rlimit(current, limit);
2953}
2954
2955static inline unsigned long rlimit_max(unsigned int limit)
2956{
2957 return task_rlimit_max(current, limit);
2958}
2959
1da177e4 2960#endif