sched: Move sched domain storage into the topology list
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4/*
5 * cloning flags:
6 */
7#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8#define CLONE_VM 0x00000100 /* set if VM shared between processes */
9#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15#define CLONE_THREAD 0x00010000 /* Same thread group? */
16#define CLONE_NEWNS 0x00020000 /* New namespace group? */
17#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
43bb40c9
DJ
24/* 0x02000000 was previously the unused CLONE_STOPPED (Start in stopped state)
25 and is now available for re-use. */
071df104 26#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 27#define CLONE_NEWIPC 0x08000000 /* New ipcs */
77ec739d 28#define CLONE_NEWUSER 0x10000000 /* New user namespace */
30e49c26 29#define CLONE_NEWPID 0x20000000 /* New pid namespace */
169e3674 30#define CLONE_NEWNET 0x40000000 /* New network namespace */
fadad878 31#define CLONE_IO 0x80000000 /* Clone io context */
b7b3c76a
DW
32
33/*
34 * Scheduling policies
35 */
36#define SCHED_NORMAL 0
37#define SCHED_FIFO 1
38#define SCHED_RR 2
39#define SCHED_BATCH 3
0e6aca43
IM
40/* SCHED_ISO: reserved but not implemented yet */
41#define SCHED_IDLE 5
ca94c442
LP
42/* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
43#define SCHED_RESET_ON_FORK 0x40000000
b7b3c76a 44
a3b6714e 45#ifdef __KERNEL__
b7b3c76a
DW
46
47struct sched_param {
48 int sched_priority;
49};
50
1da177e4
LT
51#include <asm/param.h> /* for HZ */
52
1da177e4
LT
53#include <linux/capability.h>
54#include <linux/threads.h>
55#include <linux/kernel.h>
56#include <linux/types.h>
57#include <linux/timex.h>
58#include <linux/jiffies.h>
59#include <linux/rbtree.h>
60#include <linux/thread_info.h>
61#include <linux/cpumask.h>
62#include <linux/errno.h>
63#include <linux/nodemask.h>
c92ff1bd 64#include <linux/mm_types.h>
1da177e4
LT
65
66#include <asm/system.h>
1da177e4
LT
67#include <asm/page.h>
68#include <asm/ptrace.h>
1da177e4
LT
69#include <asm/cputime.h>
70
71#include <linux/smp.h>
72#include <linux/sem.h>
73#include <linux/signal.h>
1da177e4
LT
74#include <linux/compiler.h>
75#include <linux/completion.h>
76#include <linux/pid.h>
77#include <linux/percpu.h>
78#include <linux/topology.h>
3e26c149 79#include <linux/proportions.h>
1da177e4 80#include <linux/seccomp.h>
e56d0903 81#include <linux/rcupdate.h>
05725f7e 82#include <linux/rculist.h>
23f78d4a 83#include <linux/rtmutex.h>
1da177e4 84
a3b6714e
DW
85#include <linux/time.h>
86#include <linux/param.h>
87#include <linux/resource.h>
88#include <linux/timer.h>
89#include <linux/hrtimer.h>
7c3ab738 90#include <linux/task_io_accounting.h>
9745512c 91#include <linux/latencytop.h>
9e2b2dc4 92#include <linux/cred.h>
a3b6714e
DW
93
94#include <asm/processor.h>
36d57ac4 95
1da177e4 96struct exec_domain;
c87e2837 97struct futex_pi_state;
286100a6 98struct robust_list_head;
bddd87c7 99struct bio_list;
5ad4e53b 100struct fs_struct;
cdd6c482 101struct perf_event_context;
73c10101 102struct blk_plug;
1da177e4 103
1da177e4
LT
104/*
105 * List of flags we want to share for kernel threads,
106 * if only because they are not used by them anyway.
107 */
108#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
109
110/*
111 * These are the constant used to fake the fixed-point load-average
112 * counting. Some notes:
113 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
114 * a load-average precision of 10 bits integer + 11 bits fractional
115 * - if you want to count load-averages more often, you need more
116 * precision, or rounding will get you. With 2-second counting freq,
117 * the EXP_n values would be 1981, 2034 and 2043 if still using only
118 * 11 bit fractions.
119 */
120extern unsigned long avenrun[]; /* Load averages */
2d02494f 121extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
122
123#define FSHIFT 11 /* nr of bits of precision */
124#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 125#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
126#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
127#define EXP_5 2014 /* 1/exp(5sec/5min) */
128#define EXP_15 2037 /* 1/exp(5sec/15min) */
129
130#define CALC_LOAD(load,exp,n) \
131 load *= exp; \
132 load += n*(FIXED_1-exp); \
133 load >>= FSHIFT;
134
135extern unsigned long total_forks;
136extern int nr_threads;
1da177e4
LT
137DECLARE_PER_CPU(unsigned long, process_counts);
138extern int nr_processes(void);
139extern unsigned long nr_running(void);
140extern unsigned long nr_uninterruptible(void);
141extern unsigned long nr_iowait(void);
8c215bd3 142extern unsigned long nr_iowait_cpu(int cpu);
69d25870
AV
143extern unsigned long this_cpu_load(void);
144
145
0f004f5a 146extern void calc_global_load(unsigned long ticks);
1da177e4 147
7e49fcce
SR
148extern unsigned long get_parent_ip(unsigned long addr);
149
43ae34cb
IM
150struct seq_file;
151struct cfs_rq;
4cf86d77 152struct task_group;
43ae34cb
IM
153#ifdef CONFIG_SCHED_DEBUG
154extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
155extern void proc_sched_set_task(struct task_struct *p);
156extern void
5cef9eca 157print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
158#else
159static inline void
160proc_sched_show_task(struct task_struct *p, struct seq_file *m)
161{
162}
163static inline void proc_sched_set_task(struct task_struct *p)
164{
165}
166static inline void
5cef9eca 167print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
168{
169}
170#endif
1da177e4 171
4a8342d2
LT
172/*
173 * Task state bitmask. NOTE! These bits are also
174 * encoded in fs/proc/array.c: get_task_state().
175 *
176 * We have two separate sets of flags: task->state
177 * is about runnability, while task->exit_state are
178 * about the task exiting. Confusing, but this way
179 * modifying one set can't modify the other one by
180 * mistake.
181 */
1da177e4
LT
182#define TASK_RUNNING 0
183#define TASK_INTERRUPTIBLE 1
184#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
185#define __TASK_STOPPED 4
186#define __TASK_TRACED 8
4a8342d2
LT
187/* in tsk->exit_state */
188#define EXIT_ZOMBIE 16
189#define EXIT_DEAD 32
190/* in tsk->state again */
af927232 191#define TASK_DEAD 64
f021a3c2 192#define TASK_WAKEKILL 128
e9c84311 193#define TASK_WAKING 256
e1781538 194#define TASK_STATE_MAX 512
f021a3c2 195
44d90df6 196#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
73342151 197
e1781538
PZ
198extern char ___assert_task_state[1 - 2*!!(
199 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
200
201/* Convenience macros for the sake of set_task_state */
202#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
203#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
204#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 205
92a1f4bc
MW
206/* Convenience macros for the sake of wake_up */
207#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 208#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
209
210/* get_task_state() */
211#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
212 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
213 __TASK_TRACED)
92a1f4bc 214
f021a3c2
MW
215#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
216#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
8f92054e 217#define task_is_dead(task) ((task)->exit_state != 0)
92a1f4bc 218#define task_is_stopped_or_traced(task) \
f021a3c2 219 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 220#define task_contributes_to_load(task) \
e3c8ca83 221 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
6301cb95 222 (task->flags & PF_FREEZING) == 0)
1da177e4
LT
223
224#define __set_task_state(tsk, state_value) \
225 do { (tsk)->state = (state_value); } while (0)
226#define set_task_state(tsk, state_value) \
227 set_mb((tsk)->state, (state_value))
228
498d0c57
AM
229/*
230 * set_current_state() includes a barrier so that the write of current->state
231 * is correctly serialised wrt the caller's subsequent test of whether to
232 * actually sleep:
233 *
234 * set_current_state(TASK_UNINTERRUPTIBLE);
235 * if (do_i_need_to_sleep())
236 * schedule();
237 *
238 * If the caller does not need such serialisation then use __set_current_state()
239 */
1da177e4
LT
240#define __set_current_state(state_value) \
241 do { current->state = (state_value); } while (0)
242#define set_current_state(state_value) \
243 set_mb(current->state, (state_value))
244
245/* Task command name length */
246#define TASK_COMM_LEN 16
247
1da177e4
LT
248#include <linux/spinlock.h>
249
250/*
251 * This serializes "schedule()" and also protects
252 * the run-queue from deletions/modifications (but
253 * _adding_ to the beginning of the run-queue has
254 * a separate lock).
255 */
256extern rwlock_t tasklist_lock;
257extern spinlock_t mmlist_lock;
258
36c8b586 259struct task_struct;
1da177e4 260
db1466b3
PM
261#ifdef CONFIG_PROVE_RCU
262extern int lockdep_tasklist_lock_is_held(void);
263#endif /* #ifdef CONFIG_PROVE_RCU */
264
1da177e4
LT
265extern void sched_init(void);
266extern void sched_init_smp(void);
2d07b255 267extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 268extern void init_idle(struct task_struct *idle, int cpu);
1df21055 269extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 270
89f19f04 271extern int runqueue_is_locked(int cpu);
017730c1 272
6a7b3dc3 273extern cpumask_var_t nohz_cpu_mask;
46cb4b7c 274#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
83cd4fe2
VP
275extern void select_nohz_load_balancer(int stop_tick);
276extern int get_nohz_timer_target(void);
46cb4b7c 277#else
83cd4fe2 278static inline void select_nohz_load_balancer(int stop_tick) { }
46cb4b7c 279#endif
1da177e4 280
e59e2ae2 281/*
39bc89fd 282 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
283 */
284extern void show_state_filter(unsigned long state_filter);
285
286static inline void show_state(void)
287{
39bc89fd 288 show_state_filter(0);
e59e2ae2
IM
289}
290
1da177e4
LT
291extern void show_regs(struct pt_regs *);
292
293/*
294 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
295 * task), SP is the stack pointer of the first frame that should be shown in the back
296 * trace (or NULL if the entire call-chain of the task should be shown).
297 */
298extern void show_stack(struct task_struct *task, unsigned long *sp);
299
300void io_schedule(void);
301long io_schedule_timeout(long timeout);
302
303extern void cpu_init (void);
304extern void trap_init(void);
305extern void update_process_times(int user);
306extern void scheduler_tick(void);
307
82a1fcb9
IM
308extern void sched_show_task(struct task_struct *p);
309
19cc36c0 310#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 311extern void touch_softlockup_watchdog(void);
d6ad3e28 312extern void touch_softlockup_watchdog_sync(void);
04c9167f 313extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
314extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
315 void __user *buffer,
316 size_t *lenp, loff_t *ppos);
9c44bc03 317extern unsigned int softlockup_panic;
9383d967 318extern int softlockup_thresh;
004417a6 319void lockup_detector_init(void);
8446f1d3 320#else
8446f1d3
IM
321static inline void touch_softlockup_watchdog(void)
322{
323}
d6ad3e28
JW
324static inline void touch_softlockup_watchdog_sync(void)
325{
326}
04c9167f
JF
327static inline void touch_all_softlockup_watchdogs(void)
328{
329}
004417a6
PZ
330static inline void lockup_detector_init(void)
331{
332}
8446f1d3
IM
333#endif
334
e162b39a
MSB
335#ifdef CONFIG_DETECT_HUNG_TASK
336extern unsigned int sysctl_hung_task_panic;
337extern unsigned long sysctl_hung_task_check_count;
338extern unsigned long sysctl_hung_task_timeout_secs;
339extern unsigned long sysctl_hung_task_warnings;
340extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 341 void __user *buffer,
e162b39a 342 size_t *lenp, loff_t *ppos);
e4ecda1b
ML
343#else
344/* Avoid need for ifdefs elsewhere in the code */
345enum { sysctl_hung_task_timeout_secs = 0 };
e162b39a 346#endif
8446f1d3 347
1da177e4
LT
348/* Attach to any functions which should be ignored in wchan output. */
349#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
350
351/* Linker adds these: start and end of __sched functions */
352extern char __sched_text_start[], __sched_text_end[];
353
1da177e4
LT
354/* Is this address in the __sched functions? */
355extern int in_sched_functions(unsigned long addr);
356
357#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 358extern signed long schedule_timeout(signed long timeout);
64ed93a2 359extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 360extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 361extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 362asmlinkage void schedule(void);
0d66bf6d 363extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 364
ab516013 365struct nsproxy;
acce292c 366struct user_namespace;
1da177e4 367
341c87bf
KH
368/*
369 * Default maximum number of active map areas, this limits the number of vmas
370 * per mm struct. Users can overwrite this number by sysctl but there is a
371 * problem.
372 *
373 * When a program's coredump is generated as ELF format, a section is created
374 * per a vma. In ELF, the number of sections is represented in unsigned short.
375 * This means the number of sections should be smaller than 65535 at coredump.
376 * Because the kernel adds some informative sections to a image of program at
377 * generating coredump, we need some margin. The number of extra sections is
378 * 1-3 now and depends on arch. We use "5" as safe margin, here.
379 */
380#define MAPCOUNT_ELF_CORE_MARGIN (5)
4be929be 381#define DEFAULT_MAX_MAP_COUNT (USHRT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
382
383extern int sysctl_max_map_count;
384
385#include <linux/aio.h>
386
efc1a3b1
DH
387#ifdef CONFIG_MMU
388extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
389extern unsigned long
390arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
391 unsigned long, unsigned long);
392extern unsigned long
393arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
394 unsigned long len, unsigned long pgoff,
395 unsigned long flags);
1363c3cd
WW
396extern void arch_unmap_area(struct mm_struct *, unsigned long);
397extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
efc1a3b1
DH
398#else
399static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
400#endif
1da177e4 401
901608d9 402
6c5d5238
KH
403extern void set_dumpable(struct mm_struct *mm, int value);
404extern int get_dumpable(struct mm_struct *mm);
405
406/* mm flags */
3cb4a0bb 407/* dumpable bits */
6c5d5238
KH
408#define MMF_DUMPABLE 0 /* core dump is permitted */
409#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 410
3cb4a0bb 411#define MMF_DUMPABLE_BITS 2
f8af4da3 412#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH
413
414/* coredump filter bits */
415#define MMF_DUMP_ANON_PRIVATE 2
416#define MMF_DUMP_ANON_SHARED 3
417#define MMF_DUMP_MAPPED_PRIVATE 4
418#define MMF_DUMP_MAPPED_SHARED 5
82df3973 419#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
420#define MMF_DUMP_HUGETLB_PRIVATE 7
421#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 422
3cb4a0bb 423#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 424#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
425#define MMF_DUMP_FILTER_MASK \
426 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
427#define MMF_DUMP_FILTER_DEFAULT \
e575f111 428 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
429 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
430
431#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
432# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
433#else
434# define MMF_DUMP_MASK_DEFAULT_ELF 0
435#endif
f8af4da3
HD
436 /* leave room for more dump flags */
437#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 438#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
f8af4da3
HD
439
440#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 441
1da177e4
LT
442struct sighand_struct {
443 atomic_t count;
444 struct k_sigaction action[_NSIG];
445 spinlock_t siglock;
b8fceee1 446 wait_queue_head_t signalfd_wqh;
1da177e4
LT
447};
448
0e464814 449struct pacct_struct {
f6ec29a4
KK
450 int ac_flag;
451 long ac_exitcode;
0e464814 452 unsigned long ac_mem;
77787bfb
KK
453 cputime_t ac_utime, ac_stime;
454 unsigned long ac_minflt, ac_majflt;
0e464814
KK
455};
456
42c4ab41
SG
457struct cpu_itimer {
458 cputime_t expires;
459 cputime_t incr;
8356b5f9
SG
460 u32 error;
461 u32 incr_error;
42c4ab41
SG
462};
463
f06febc9
FM
464/**
465 * struct task_cputime - collected CPU time counts
466 * @utime: time spent in user mode, in &cputime_t units
467 * @stime: time spent in kernel mode, in &cputime_t units
468 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 469 *
f06febc9
FM
470 * This structure groups together three kinds of CPU time that are
471 * tracked for threads and thread groups. Most things considering
472 * CPU time want to group these counts together and treat all three
473 * of them in parallel.
474 */
475struct task_cputime {
476 cputime_t utime;
477 cputime_t stime;
478 unsigned long long sum_exec_runtime;
479};
480/* Alternate field names when used to cache expirations. */
481#define prof_exp stime
482#define virt_exp utime
483#define sched_exp sum_exec_runtime
484
4cd4c1b4
PZ
485#define INIT_CPUTIME \
486 (struct task_cputime) { \
487 .utime = cputime_zero, \
488 .stime = cputime_zero, \
489 .sum_exec_runtime = 0, \
490 }
491
c99e6efe
PZ
492/*
493 * Disable preemption until the scheduler is running.
494 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
495 *
496 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
497 * before the scheduler is active -- see should_resched().
c99e6efe 498 */
d86ee480 499#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 500
f06febc9 501/**
4cd4c1b4
PZ
502 * struct thread_group_cputimer - thread group interval timer counts
503 * @cputime: thread group interval timers.
504 * @running: non-zero when there are timers running and
505 * @cputime receives updates.
506 * @lock: lock for fields in this struct.
f06febc9
FM
507 *
508 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 509 * used for thread group CPU timer calculations.
f06febc9 510 */
4cd4c1b4
PZ
511struct thread_group_cputimer {
512 struct task_cputime cputime;
513 int running;
514 spinlock_t lock;
f06febc9 515};
f06febc9 516
5091faa4
MG
517struct autogroup;
518
1da177e4 519/*
e815f0a8 520 * NOTE! "signal_struct" does not have its own
1da177e4
LT
521 * locking, because a shared signal_struct always
522 * implies a shared sighand_struct, so locking
523 * sighand_struct is always a proper superset of
524 * the locking of signal_struct.
525 */
526struct signal_struct {
ea6d290c 527 atomic_t sigcnt;
1da177e4 528 atomic_t live;
b3ac022c 529 int nr_threads;
1da177e4
LT
530
531 wait_queue_head_t wait_chldexit; /* for wait4() */
532
533 /* current thread group signal load-balancing target: */
36c8b586 534 struct task_struct *curr_target;
1da177e4
LT
535
536 /* shared signal handling: */
537 struct sigpending shared_pending;
538
539 /* thread group exit support */
540 int group_exit_code;
541 /* overloaded:
542 * - notify group_exit_task when ->count is equal to notify_count
543 * - everyone except group_exit_task is stopped during signal delivery
544 * of fatal signals, group_exit_task processes the signal.
545 */
1da177e4 546 int notify_count;
07dd20e0 547 struct task_struct *group_exit_task;
1da177e4
LT
548
549 /* thread group stop support, overloads group_exit_code too */
550 int group_stop_count;
551 unsigned int flags; /* see SIGNAL_* flags below */
552
553 /* POSIX.1b Interval Timers */
554 struct list_head posix_timers;
555
556 /* ITIMER_REAL timer for the process */
2ff678b8 557 struct hrtimer real_timer;
fea9d175 558 struct pid *leader_pid;
2ff678b8 559 ktime_t it_real_incr;
1da177e4 560
42c4ab41
SG
561 /*
562 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
563 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
564 * values are defined to 0 and 1 respectively
565 */
566 struct cpu_itimer it[2];
1da177e4 567
f06febc9 568 /*
4cd4c1b4
PZ
569 * Thread group totals for process CPU timers.
570 * See thread_group_cputimer(), et al, for details.
f06febc9 571 */
4cd4c1b4 572 struct thread_group_cputimer cputimer;
f06febc9
FM
573
574 /* Earliest-expiration cache. */
575 struct task_cputime cputime_expires;
576
577 struct list_head cpu_timers[3];
578
ab521dc0 579 struct pid *tty_old_pgrp;
1ec320af 580
1da177e4
LT
581 /* boolean value for session group leader */
582 int leader;
583
584 struct tty_struct *tty; /* NULL if no tty */
585
5091faa4
MG
586#ifdef CONFIG_SCHED_AUTOGROUP
587 struct autogroup *autogroup;
588#endif
1da177e4
LT
589 /*
590 * Cumulative resource counters for dead threads in the group,
591 * and for reaped dead child processes forked by this group.
592 * Live threads maintain their own counters and add to these
593 * in __exit_signal, except for the group leader.
594 */
32bd671d 595 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
596 cputime_t gtime;
597 cputime_t cgtime;
0cf55e1e
HS
598#ifndef CONFIG_VIRT_CPU_ACCOUNTING
599 cputime_t prev_utime, prev_stime;
600#endif
1da177e4
LT
601 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
602 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 603 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 604 unsigned long maxrss, cmaxrss;
940389b8 605 struct task_io_accounting ioac;
1da177e4 606
32bd671d
PZ
607 /*
608 * Cumulative ns of schedule CPU time fo dead threads in the
609 * group, not including a zombie group leader, (This only differs
610 * from jiffies_to_ns(utime + stime) if sched_clock uses something
611 * other than jiffies.)
612 */
613 unsigned long long sum_sched_runtime;
614
1da177e4
LT
615 /*
616 * We don't bother to synchronize most readers of this at all,
617 * because there is no reader checking a limit that actually needs
618 * to get both rlim_cur and rlim_max atomically, and either one
619 * alone is a single word that can safely be read normally.
620 * getrlimit/setrlimit use task_lock(current->group_leader) to
621 * protect this instead of the siglock, because they really
622 * have no need to disable irqs.
623 */
624 struct rlimit rlim[RLIM_NLIMITS];
625
0e464814
KK
626#ifdef CONFIG_BSD_PROCESS_ACCT
627 struct pacct_struct pacct; /* per-process accounting information */
628#endif
ad4ecbcb 629#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
630 struct taskstats *stats;
631#endif
522ed776
MT
632#ifdef CONFIG_AUDIT
633 unsigned audit_tty;
634 struct tty_audit_buf *tty_audit_buf;
635#endif
28b83c51 636
a63d83f4
DR
637 int oom_adj; /* OOM kill score adjustment (bit shift) */
638 int oom_score_adj; /* OOM kill score adjustment */
dabb16f6
MSB
639 int oom_score_adj_min; /* OOM kill score adjustment minimum value.
640 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
641
642 struct mutex cred_guard_mutex; /* guard against foreign influences on
643 * credential calculations
644 * (notably. ptrace) */
1da177e4
LT
645};
646
4866cde0
NP
647/* Context switch must be unlocked if interrupts are to be enabled */
648#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
649# define __ARCH_WANT_UNLOCKED_CTXSW
650#endif
651
1da177e4
LT
652/*
653 * Bits in flags field of signal_struct.
654 */
655#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
656#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
657#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
658#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
659/*
660 * Pending notifications to parent.
661 */
662#define SIGNAL_CLD_STOPPED 0x00000010
663#define SIGNAL_CLD_CONTINUED 0x00000020
664#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 665
fae5fa44
ON
666#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
667
ed5d2cac
ON
668/* If true, all threads except ->group_exit_task have pending SIGKILL */
669static inline int signal_group_exit(const struct signal_struct *sig)
670{
671 return (sig->flags & SIGNAL_GROUP_EXIT) ||
672 (sig->group_exit_task != NULL);
673}
674
1da177e4
LT
675/*
676 * Some day this will be a full-fledged user tracking system..
677 */
678struct user_struct {
679 atomic_t __count; /* reference count */
680 atomic_t processes; /* How many processes does this user have? */
681 atomic_t files; /* How many open files does this user have? */
682 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 683#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
684 atomic_t inotify_watches; /* How many inotify watches does this user have? */
685 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
686#endif
4afeff85
EP
687#ifdef CONFIG_FANOTIFY
688 atomic_t fanotify_listeners;
689#endif
7ef9964e 690#ifdef CONFIG_EPOLL
52bd19f7 691 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 692#endif
970a8645 693#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
694 /* protected by mq_lock */
695 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 696#endif
1da177e4
LT
697 unsigned long locked_shm; /* How many pages of mlocked shm ? */
698
699#ifdef CONFIG_KEYS
700 struct key *uid_keyring; /* UID specific keyring */
701 struct key *session_keyring; /* UID's default session keyring */
702#endif
703
704 /* Hash table maintenance information */
735de223 705 struct hlist_node uidhash_node;
1da177e4 706 uid_t uid;
18b6e041 707 struct user_namespace *user_ns;
24e377a8 708
cdd6c482 709#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
710 atomic_long_t locked_vm;
711#endif
1da177e4
LT
712};
713
eb41d946 714extern int uids_sysfs_init(void);
5cb350ba 715
1da177e4
LT
716extern struct user_struct *find_user(uid_t);
717
718extern struct user_struct root_user;
719#define INIT_USER (&root_user)
720
b6dff3ec 721
1da177e4
LT
722struct backing_dev_info;
723struct reclaim_state;
724
52f17b6c 725#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
726struct sched_info {
727 /* cumulative counters */
2d72376b 728 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 729 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
730
731 /* timestamps */
172ba844
BS
732 unsigned long long last_arrival,/* when we last ran on a cpu */
733 last_queued; /* when we were last queued to run */
b8efb561
IM
734#ifdef CONFIG_SCHEDSTATS
735 /* BKL stats */
480b9434 736 unsigned int bkl_count;
b8efb561 737#endif
1da177e4 738};
52f17b6c 739#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 740
ca74e92b
SN
741#ifdef CONFIG_TASK_DELAY_ACCT
742struct task_delay_info {
743 spinlock_t lock;
744 unsigned int flags; /* Private per-task flags */
745
746 /* For each stat XXX, add following, aligned appropriately
747 *
748 * struct timespec XXX_start, XXX_end;
749 * u64 XXX_delay;
750 * u32 XXX_count;
751 *
752 * Atomicity of updates to XXX_delay, XXX_count protected by
753 * single lock above (split into XXX_lock if contention is an issue).
754 */
0ff92245
SN
755
756 /*
757 * XXX_count is incremented on every XXX operation, the delay
758 * associated with the operation is added to XXX_delay.
759 * XXX_delay contains the accumulated delay time in nanoseconds.
760 */
761 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
762 u64 blkio_delay; /* wait for sync block io completion */
763 u64 swapin_delay; /* wait for swapin block io completion */
764 u32 blkio_count; /* total count of the number of sync block */
765 /* io operations performed */
766 u32 swapin_count; /* total count of the number of swapin block */
767 /* io operations performed */
873b4771
KK
768
769 struct timespec freepages_start, freepages_end;
770 u64 freepages_delay; /* wait for memory reclaim */
771 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 772};
52f17b6c
CS
773#endif /* CONFIG_TASK_DELAY_ACCT */
774
775static inline int sched_info_on(void)
776{
777#ifdef CONFIG_SCHEDSTATS
778 return 1;
779#elif defined(CONFIG_TASK_DELAY_ACCT)
780 extern int delayacct_on;
781 return delayacct_on;
782#else
783 return 0;
ca74e92b 784#endif
52f17b6c 785}
ca74e92b 786
d15bcfdb
IM
787enum cpu_idle_type {
788 CPU_IDLE,
789 CPU_NOT_IDLE,
790 CPU_NEWLY_IDLE,
791 CPU_MAX_IDLE_TYPES
1da177e4
LT
792};
793
794/*
795 * sched-domains (multiprocessor balancing) declarations:
796 */
9aa7b369
IM
797
798/*
799 * Increase resolution of nice-level calculations:
800 */
801#define SCHED_LOAD_SHIFT 10
802#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
803
f8700df7 804#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 805
2dd73a4f 806#ifdef CONFIG_SMP
b5d978e0
PZ
807#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
808#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
809#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
810#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 811#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 812#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 813#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0
PZ
814#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
815#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
816#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
817#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 818#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 819#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 820
afb8a9b7
GS
821enum powersavings_balance_level {
822 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
823 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
824 * first for long running threads
825 */
826 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
827 * cpu package for power savings
828 */
829 MAX_POWERSAVINGS_BALANCE_LEVELS
830};
89c4710e 831
716707b2 832extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 833
716707b2
VS
834static inline int sd_balance_for_mc_power(void)
835{
836 if (sched_smt_power_savings)
837 return SD_POWERSAVINGS_BALANCE;
5c45bf27 838
28f53181
VS
839 if (!sched_mc_power_savings)
840 return SD_PREFER_SIBLING;
841
842 return 0;
716707b2 843}
89c4710e 844
716707b2
VS
845static inline int sd_balance_for_package_power(void)
846{
847 if (sched_mc_power_savings | sched_smt_power_savings)
848 return SD_POWERSAVINGS_BALANCE;
849
b5d978e0 850 return SD_PREFER_SIBLING;
716707b2 851}
5c45bf27 852
532cb4c4
MN
853extern int __weak arch_sd_sibiling_asym_packing(void);
854
100fdaee
VS
855/*
856 * Optimise SD flags for power savings:
25985edc 857 * SD_BALANCE_NEWIDLE helps aggressive task consolidation and power savings.
100fdaee
VS
858 * Keep default SD flags if sched_{smt,mc}_power_saving=0
859 */
860
861static inline int sd_power_saving_flags(void)
862{
863 if (sched_mc_power_savings | sched_smt_power_savings)
864 return SD_BALANCE_NEWIDLE;
865
866 return 0;
867}
1da177e4
LT
868
869struct sched_group {
870 struct sched_group *next; /* Must be a circular list */
dce840a0 871 atomic_t ref;
1da177e4
LT
872
873 /*
874 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 875 * single CPU.
5517d86b 876 */
9d5efe05 877 unsigned int cpu_power, cpu_power_orig;
aae6d3dd 878 unsigned int group_weight;
6c99e9ad 879
4200efd9
IM
880 /*
881 * The CPUs this group covers.
882 *
883 * NOTE: this field is variable length. (Allocated dynamically
884 * by attaching extra space to the end of the structure,
885 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
886 */
887 unsigned long cpumask[0];
1da177e4
LT
888};
889
758b2cdc
RR
890static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
891{
6c99e9ad 892 return to_cpumask(sg->cpumask);
758b2cdc
RR
893}
894
1d3504fc
HS
895enum sched_domain_level {
896 SD_LV_NONE = 0,
3859173d 897#ifdef CONFIG_SCHED_SMT
1d3504fc 898 SD_LV_SIBLING,
3859173d
PZ
899#endif
900#ifdef CONFIG_SCHED_MC
1d3504fc 901 SD_LV_MC,
3859173d
PZ
902#endif
903#ifdef CONFIG_SCHED_BOOK
01a08546 904 SD_LV_BOOK,
3859173d 905#endif
1d3504fc 906 SD_LV_CPU,
3859173d 907#ifdef CONFIG_NUMA
1d3504fc
HS
908 SD_LV_NODE,
909 SD_LV_ALLNODES,
3859173d 910#endif
1d3504fc
HS
911 SD_LV_MAX
912};
913
914struct sched_domain_attr {
915 int relax_domain_level;
916};
917
918#define SD_ATTR_INIT (struct sched_domain_attr) { \
919 .relax_domain_level = -1, \
920}
921
1da177e4
LT
922struct sched_domain {
923 /* These fields must be setup */
924 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 925 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 926 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
927 unsigned long min_interval; /* Minimum balance interval ms */
928 unsigned long max_interval; /* Maximum balance interval ms */
929 unsigned int busy_factor; /* less balancing by factor if busy */
930 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 931 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
932 unsigned int busy_idx;
933 unsigned int idle_idx;
934 unsigned int newidle_idx;
935 unsigned int wake_idx;
147cbb4b 936 unsigned int forkexec_idx;
a52bfd73 937 unsigned int smt_gain;
1da177e4 938 int flags; /* See SD_* */
1d3504fc 939 enum sched_domain_level level;
1da177e4
LT
940
941 /* Runtime fields. */
942 unsigned long last_balance; /* init to jiffies. units in jiffies */
943 unsigned int balance_interval; /* initialise to 1. units in ms. */
944 unsigned int nr_balance_failed; /* initialise to 0 */
945
2398f2c6
PZ
946 u64 last_update;
947
1da177e4
LT
948#ifdef CONFIG_SCHEDSTATS
949 /* load_balance() stats */
480b9434
KC
950 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
951 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
952 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
953 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
954 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
955 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
956 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
957 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
958
959 /* Active load balancing */
480b9434
KC
960 unsigned int alb_count;
961 unsigned int alb_failed;
962 unsigned int alb_pushed;
1da177e4 963
68767a0a 964 /* SD_BALANCE_EXEC stats */
480b9434
KC
965 unsigned int sbe_count;
966 unsigned int sbe_balanced;
967 unsigned int sbe_pushed;
1da177e4 968
68767a0a 969 /* SD_BALANCE_FORK stats */
480b9434
KC
970 unsigned int sbf_count;
971 unsigned int sbf_balanced;
972 unsigned int sbf_pushed;
68767a0a 973
1da177e4 974 /* try_to_wake_up() stats */
480b9434
KC
975 unsigned int ttwu_wake_remote;
976 unsigned int ttwu_move_affine;
977 unsigned int ttwu_move_balance;
1da177e4 978#endif
a5d8c348
IM
979#ifdef CONFIG_SCHED_DEBUG
980 char *name;
981#endif
dce840a0
PZ
982 union {
983 void *private; /* used during construction */
984 struct rcu_head rcu; /* used during destruction */
985 };
6c99e9ad 986
669c55e9 987 unsigned int span_weight;
4200efd9
IM
988 /*
989 * Span of all CPUs in this domain.
990 *
991 * NOTE: this field is variable length. (Allocated dynamically
992 * by attaching extra space to the end of the structure,
993 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
994 */
995 unsigned long span[0];
1da177e4
LT
996};
997
758b2cdc
RR
998static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
999{
6c99e9ad 1000 return to_cpumask(sd->span);
758b2cdc
RR
1001}
1002
acc3f5d7 1003extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 1004 struct sched_domain_attr *dattr_new);
029190c5 1005
acc3f5d7
RR
1006/* Allocate an array of sched domains, for partition_sched_domains(). */
1007cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1008void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1009
06aaf76a
IM
1010/* Test a flag in parent sched domain */
1011static inline int test_sd_parent(struct sched_domain *sd, int flag)
1012{
1013 if (sd->parent && (sd->parent->flags & flag))
1014 return 1;
1015
1016 return 0;
1017}
029190c5 1018
47fe38fc
PZ
1019unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1020unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1021
1b427c15 1022#else /* CONFIG_SMP */
1da177e4 1023
1b427c15 1024struct sched_domain_attr;
d02c7a8c 1025
1b427c15 1026static inline void
acc3f5d7 1027partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1028 struct sched_domain_attr *dattr_new)
1029{
d02c7a8c 1030}
1b427c15 1031#endif /* !CONFIG_SMP */
1da177e4 1032
47fe38fc 1033
1da177e4 1034struct io_context; /* See blkdev.h */
1da177e4 1035
1da177e4 1036
383f2835 1037#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1038extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1039#else
1040static inline void prefetch_stack(struct task_struct *t) { }
1041#endif
1da177e4
LT
1042
1043struct audit_context; /* See audit.c */
1044struct mempolicy;
b92ce558 1045struct pipe_inode_info;
4865ecf1 1046struct uts_namespace;
1da177e4 1047
20b8a59f
IM
1048struct rq;
1049struct sched_domain;
1050
7d478721
PZ
1051/*
1052 * wake flags
1053 */
1054#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1055#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1056
371fd7e7
PZ
1057#define ENQUEUE_WAKEUP 1
1058#define ENQUEUE_WAKING 2
1059#define ENQUEUE_HEAD 4
1060
1061#define DEQUEUE_SLEEP 1
1062
20b8a59f 1063struct sched_class {
5522d5d5 1064 const struct sched_class *next;
20b8a59f 1065
371fd7e7
PZ
1066 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int flags);
1067 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int flags);
4530d7ab 1068 void (*yield_task) (struct rq *rq);
d95f4122 1069 bool (*yield_to_task) (struct rq *rq, struct task_struct *p, bool preempt);
20b8a59f 1070
7d478721 1071 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1072
fb8d4724 1073 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1074 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1075
681f3e68 1076#ifdef CONFIG_SMP
0017d735
PZ
1077 int (*select_task_rq)(struct rq *rq, struct task_struct *p,
1078 int sd_flag, int flags);
4ce72a2c 1079
9a897c5a
SR
1080 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1081 void (*post_schedule) (struct rq *this_rq);
efbbd05a
PZ
1082 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1083 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1084
cd8ba7cd 1085 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1086 const struct cpumask *newmask);
57d885fe 1087
1f11eb6a
GH
1088 void (*rq_online)(struct rq *rq);
1089 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1090#endif
1091
1092 void (*set_curr_task) (struct rq *rq);
1093 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1094 void (*task_fork) (struct task_struct *p);
cb469845 1095
da7a735e
PZ
1096 void (*switched_from) (struct rq *this_rq, struct task_struct *task);
1097 void (*switched_to) (struct rq *this_rq, struct task_struct *task);
cb469845 1098 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
da7a735e 1099 int oldprio);
810b3817 1100
dba091b9
TG
1101 unsigned int (*get_rr_interval) (struct rq *rq,
1102 struct task_struct *task);
0d721cea 1103
810b3817 1104#ifdef CONFIG_FAIR_GROUP_SCHED
b2b5ce02 1105 void (*task_move_group) (struct task_struct *p, int on_rq);
810b3817 1106#endif
20b8a59f
IM
1107};
1108
1109struct load_weight {
1110 unsigned long weight, inv_weight;
1111};
1112
94c18227 1113#ifdef CONFIG_SCHEDSTATS
41acab88 1114struct sched_statistics {
20b8a59f 1115 u64 wait_start;
94c18227 1116 u64 wait_max;
6d082592
AV
1117 u64 wait_count;
1118 u64 wait_sum;
8f0dfc34
AV
1119 u64 iowait_count;
1120 u64 iowait_sum;
94c18227 1121
20b8a59f 1122 u64 sleep_start;
20b8a59f 1123 u64 sleep_max;
94c18227
IM
1124 s64 sum_sleep_runtime;
1125
1126 u64 block_start;
20b8a59f
IM
1127 u64 block_max;
1128 u64 exec_max;
eba1ed4b 1129 u64 slice_max;
cc367732 1130
cc367732
IM
1131 u64 nr_migrations_cold;
1132 u64 nr_failed_migrations_affine;
1133 u64 nr_failed_migrations_running;
1134 u64 nr_failed_migrations_hot;
1135 u64 nr_forced_migrations;
cc367732
IM
1136
1137 u64 nr_wakeups;
1138 u64 nr_wakeups_sync;
1139 u64 nr_wakeups_migrate;
1140 u64 nr_wakeups_local;
1141 u64 nr_wakeups_remote;
1142 u64 nr_wakeups_affine;
1143 u64 nr_wakeups_affine_attempts;
1144 u64 nr_wakeups_passive;
1145 u64 nr_wakeups_idle;
41acab88
LDM
1146};
1147#endif
1148
1149struct sched_entity {
1150 struct load_weight load; /* for load-balancing */
1151 struct rb_node run_node;
1152 struct list_head group_node;
1153 unsigned int on_rq;
1154
1155 u64 exec_start;
1156 u64 sum_exec_runtime;
1157 u64 vruntime;
1158 u64 prev_sum_exec_runtime;
1159
41acab88
LDM
1160 u64 nr_migrations;
1161
41acab88
LDM
1162#ifdef CONFIG_SCHEDSTATS
1163 struct sched_statistics statistics;
94c18227
IM
1164#endif
1165
20b8a59f
IM
1166#ifdef CONFIG_FAIR_GROUP_SCHED
1167 struct sched_entity *parent;
1168 /* rq on which this entity is (to be) queued: */
1169 struct cfs_rq *cfs_rq;
1170 /* rq "owned" by this entity/group: */
1171 struct cfs_rq *my_q;
1172#endif
1173};
70b97a7f 1174
fa717060
PZ
1175struct sched_rt_entity {
1176 struct list_head run_list;
78f2c7db 1177 unsigned long timeout;
bee367ed 1178 unsigned int time_slice;
6f505b16
PZ
1179 int nr_cpus_allowed;
1180
58d6c2d7 1181 struct sched_rt_entity *back;
052f1dc7 1182#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1183 struct sched_rt_entity *parent;
1184 /* rq on which this entity is (to be) queued: */
1185 struct rt_rq *rt_rq;
1186 /* rq "owned" by this entity/group: */
1187 struct rt_rq *my_q;
1188#endif
fa717060
PZ
1189};
1190
86848966
PM
1191struct rcu_node;
1192
8dc85d54
PZ
1193enum perf_event_task_context {
1194 perf_invalid_context = -1,
1195 perf_hw_context = 0,
89a1e187 1196 perf_sw_context,
8dc85d54
PZ
1197 perf_nr_task_contexts,
1198};
1199
1da177e4
LT
1200struct task_struct {
1201 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1202 void *stack;
1da177e4 1203 atomic_t usage;
97dc32cd
WC
1204 unsigned int flags; /* per process flags, defined below */
1205 unsigned int ptrace;
1da177e4 1206
36772092 1207 int lock_depth; /* BKL lock depth */
1da177e4 1208
2dd73a4f
PW
1209#ifdef CONFIG_SMP
1210#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1211 int oncpu;
1212#endif
2dd73a4f 1213#endif
50e645a8 1214
b29739f9 1215 int prio, static_prio, normal_prio;
c7aceaba 1216 unsigned int rt_priority;
5522d5d5 1217 const struct sched_class *sched_class;
20b8a59f 1218 struct sched_entity se;
fa717060 1219 struct sched_rt_entity rt;
1da177e4 1220
e107be36
AK
1221#ifdef CONFIG_PREEMPT_NOTIFIERS
1222 /* list of struct preempt_notifier: */
1223 struct hlist_head preempt_notifiers;
1224#endif
1225
18796aa0
AD
1226 /*
1227 * fpu_counter contains the number of consecutive context switches
1228 * that the FPU is used. If this is over a threshold, the lazy fpu
1229 * saving becomes unlazy to save the trap. This is an unsigned char
1230 * so that after 256 times the counter wraps and the behavior turns
1231 * lazy again; this to deal with bursty apps that only use FPU for
1232 * a short time
1233 */
1234 unsigned char fpu_counter;
6c5c9341 1235#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1236 unsigned int btrace_seq;
6c5c9341 1237#endif
1da177e4 1238
97dc32cd 1239 unsigned int policy;
1da177e4 1240 cpumask_t cpus_allowed;
1da177e4 1241
a57eb940 1242#ifdef CONFIG_PREEMPT_RCU
e260be67 1243 int rcu_read_lock_nesting;
f41d911f 1244 char rcu_read_unlock_special;
f41d911f 1245 struct list_head rcu_node_entry;
a57eb940
PM
1246#endif /* #ifdef CONFIG_PREEMPT_RCU */
1247#ifdef CONFIG_TREE_PREEMPT_RCU
1248 struct rcu_node *rcu_blocked_node;
f41d911f 1249#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1250#ifdef CONFIG_RCU_BOOST
1251 struct rt_mutex *rcu_boost_mutex;
1252#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1253
52f17b6c 1254#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1255 struct sched_info sched_info;
1256#endif
1257
1258 struct list_head tasks;
806c09a7 1259#ifdef CONFIG_SMP
917b627d 1260 struct plist_node pushable_tasks;
806c09a7 1261#endif
1da177e4
LT
1262
1263 struct mm_struct *mm, *active_mm;
34e55232
KH
1264#if defined(SPLIT_RSS_COUNTING)
1265 struct task_rss_stat rss_stat;
1266#endif
1da177e4 1267/* task state */
97dc32cd 1268 int exit_state;
1da177e4
LT
1269 int exit_code, exit_signal;
1270 int pdeath_signal; /* The signal sent when the parent dies */
1271 /* ??? */
97dc32cd 1272 unsigned int personality;
1da177e4 1273 unsigned did_exec:1;
f9ce1f1c
KT
1274 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1275 * execve */
8f0dfc34
AV
1276 unsigned in_iowait:1;
1277
ca94c442
LP
1278
1279 /* Revert to default priority/policy when forking */
1280 unsigned sched_reset_on_fork:1;
1281
1da177e4
LT
1282 pid_t pid;
1283 pid_t tgid;
0a425405 1284
1314562a 1285#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1286 /* Canary value for the -fstack-protector gcc feature */
1287 unsigned long stack_canary;
1314562a 1288#endif
e0032087 1289
1da177e4
LT
1290 /*
1291 * pointers to (original) parent process, youngest child, younger sibling,
1292 * older sibling, respectively. (p->father can be replaced with
f470021a 1293 * p->real_parent->pid)
1da177e4 1294 */
f470021a
RM
1295 struct task_struct *real_parent; /* real parent process */
1296 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1297 /*
f470021a 1298 * children/sibling forms the list of my natural children
1da177e4
LT
1299 */
1300 struct list_head children; /* list of my children */
1301 struct list_head sibling; /* linkage in my parent's children list */
1302 struct task_struct *group_leader; /* threadgroup leader */
1303
f470021a
RM
1304 /*
1305 * ptraced is the list of tasks this task is using ptrace on.
1306 * This includes both natural children and PTRACE_ATTACH targets.
1307 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1308 */
1309 struct list_head ptraced;
1310 struct list_head ptrace_entry;
1311
1da177e4 1312 /* PID/PID hash table linkage. */
92476d7f 1313 struct pid_link pids[PIDTYPE_MAX];
47e65328 1314 struct list_head thread_group;
1da177e4
LT
1315
1316 struct completion *vfork_done; /* for vfork() */
1317 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1318 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1319
c66f08be 1320 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1321 cputime_t gtime;
d99ca3b9 1322#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1323 cputime_t prev_utime, prev_stime;
d99ca3b9 1324#endif
1da177e4 1325 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1326 struct timespec start_time; /* monotonic time */
1327 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1328/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1329 unsigned long min_flt, maj_flt;
1330
f06febc9 1331 struct task_cputime cputime_expires;
1da177e4
LT
1332 struct list_head cpu_timers[3];
1333
1334/* process credentials */
1b0ba1c9 1335 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1336 * credentials (COW) */
1b0ba1c9 1337 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1338 * credentials (COW) */
ee18d64c 1339 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1340
36772092
PBG
1341 char comm[TASK_COMM_LEN]; /* executable name excluding path
1342 - access with [gs]et_task_comm (which lock
1343 it with task_lock())
221af7f8 1344 - initialized normally by setup_new_exec */
1da177e4
LT
1345/* file system info */
1346 int link_count, total_link_count;
3d5b6fcc 1347#ifdef CONFIG_SYSVIPC
1da177e4
LT
1348/* ipc stuff */
1349 struct sysv_sem sysvsem;
3d5b6fcc 1350#endif
e162b39a 1351#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1352/* hung task detection */
82a1fcb9
IM
1353 unsigned long last_switch_count;
1354#endif
1da177e4
LT
1355/* CPU-specific state of this task */
1356 struct thread_struct thread;
1357/* filesystem information */
1358 struct fs_struct *fs;
1359/* open file information */
1360 struct files_struct *files;
1651e14e 1361/* namespaces */
ab516013 1362 struct nsproxy *nsproxy;
1da177e4
LT
1363/* signal handlers */
1364 struct signal_struct *signal;
1365 struct sighand_struct *sighand;
1366
1367 sigset_t blocked, real_blocked;
f3de272b 1368 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1369 struct sigpending pending;
1370
1371 unsigned long sas_ss_sp;
1372 size_t sas_ss_size;
1373 int (*notifier)(void *priv);
1374 void *notifier_data;
1375 sigset_t *notifier_mask;
1da177e4 1376 struct audit_context *audit_context;
bfef93a5
AV
1377#ifdef CONFIG_AUDITSYSCALL
1378 uid_t loginuid;
4746ec5b 1379 unsigned int sessionid;
bfef93a5 1380#endif
1da177e4
LT
1381 seccomp_t seccomp;
1382
1383/* Thread group tracking */
1384 u32 parent_exec_id;
1385 u32 self_exec_id;
58568d2a
MX
1386/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1387 * mempolicy */
1da177e4 1388 spinlock_t alloc_lock;
1da177e4 1389
3aa551c9
TG
1390#ifdef CONFIG_GENERIC_HARDIRQS
1391 /* IRQ handler threads */
1392 struct irqaction *irqaction;
1393#endif
1394
b29739f9 1395 /* Protection of the PI data structures: */
1d615482 1396 raw_spinlock_t pi_lock;
b29739f9 1397
23f78d4a
IM
1398#ifdef CONFIG_RT_MUTEXES
1399 /* PI waiters blocked on a rt_mutex held by this task */
1400 struct plist_head pi_waiters;
1401 /* Deadlock detection and priority inheritance handling */
1402 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1403#endif
1404
408894ee
IM
1405#ifdef CONFIG_DEBUG_MUTEXES
1406 /* mutex deadlock detection */
1407 struct mutex_waiter *blocked_on;
1408#endif
de30a2b3
IM
1409#ifdef CONFIG_TRACE_IRQFLAGS
1410 unsigned int irq_events;
de30a2b3 1411 unsigned long hardirq_enable_ip;
de30a2b3 1412 unsigned long hardirq_disable_ip;
fa1452e8 1413 unsigned int hardirq_enable_event;
de30a2b3 1414 unsigned int hardirq_disable_event;
fa1452e8
HS
1415 int hardirqs_enabled;
1416 int hardirq_context;
de30a2b3 1417 unsigned long softirq_disable_ip;
de30a2b3 1418 unsigned long softirq_enable_ip;
fa1452e8 1419 unsigned int softirq_disable_event;
de30a2b3 1420 unsigned int softirq_enable_event;
fa1452e8 1421 int softirqs_enabled;
de30a2b3
IM
1422 int softirq_context;
1423#endif
fbb9ce95 1424#ifdef CONFIG_LOCKDEP
bdb9441e 1425# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1426 u64 curr_chain_key;
1427 int lockdep_depth;
fbb9ce95 1428 unsigned int lockdep_recursion;
c7aceaba 1429 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1430 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1431#endif
408894ee 1432
1da177e4
LT
1433/* journalling filesystem info */
1434 void *journal_info;
1435
d89d8796 1436/* stacked block device info */
bddd87c7 1437 struct bio_list *bio_list;
d89d8796 1438
73c10101
JA
1439#ifdef CONFIG_BLOCK
1440/* stack plugging */
1441 struct blk_plug *plug;
1442#endif
1443
1da177e4
LT
1444/* VM state */
1445 struct reclaim_state *reclaim_state;
1446
1da177e4
LT
1447 struct backing_dev_info *backing_dev_info;
1448
1449 struct io_context *io_context;
1450
1451 unsigned long ptrace_message;
1452 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1453 struct task_io_accounting ioac;
8f0ab514 1454#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1455 u64 acct_rss_mem1; /* accumulated rss usage */
1456 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1457 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1458#endif
1459#ifdef CONFIG_CPUSETS
58568d2a 1460 nodemask_t mems_allowed; /* Protected by alloc_lock */
c0ff7453 1461 int mems_allowed_change_disable;
825a46af 1462 int cpuset_mem_spread_rotor;
6adef3eb 1463 int cpuset_slab_spread_rotor;
1da177e4 1464#endif
ddbcc7e8 1465#ifdef CONFIG_CGROUPS
817929ec 1466 /* Control Group info protected by css_set_lock */
2c392b8c 1467 struct css_set __rcu *cgroups;
817929ec
PM
1468 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1469 struct list_head cg_list;
ddbcc7e8 1470#endif
42b2dd0a 1471#ifdef CONFIG_FUTEX
0771dfef 1472 struct robust_list_head __user *robust_list;
34f192c6
IM
1473#ifdef CONFIG_COMPAT
1474 struct compat_robust_list_head __user *compat_robust_list;
1475#endif
c87e2837
IM
1476 struct list_head pi_state_list;
1477 struct futex_pi_state *pi_state_cache;
c7aceaba 1478#endif
cdd6c482 1479#ifdef CONFIG_PERF_EVENTS
8dc85d54 1480 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1481 struct mutex perf_event_mutex;
1482 struct list_head perf_event_list;
a63eaf34 1483#endif
c7aceaba 1484#ifdef CONFIG_NUMA
58568d2a 1485 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1486 short il_next;
207205a2 1487 short pref_node_fork;
42b2dd0a 1488#endif
22e2c507 1489 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1490 struct rcu_head rcu;
b92ce558
JA
1491
1492 /*
1493 * cache last used pipe for splice
1494 */
1495 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1496#ifdef CONFIG_TASK_DELAY_ACCT
1497 struct task_delay_info *delays;
f4f154fd
AM
1498#endif
1499#ifdef CONFIG_FAULT_INJECTION
1500 int make_it_fail;
ca74e92b 1501#endif
3e26c149 1502 struct prop_local_single dirties;
9745512c
AV
1503#ifdef CONFIG_LATENCYTOP
1504 int latency_record_count;
1505 struct latency_record latency_record[LT_SAVECOUNT];
1506#endif
6976675d
AV
1507 /*
1508 * time slack values; these are used to round up poll() and
1509 * select() etc timeout values. These are in nanoseconds.
1510 */
1511 unsigned long timer_slack_ns;
1512 unsigned long default_timer_slack_ns;
f8d570a4
DM
1513
1514 struct list_head *scm_work_list;
fb52607a 1515#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1516 /* Index of current stored address in ret_stack */
f201ae23
FW
1517 int curr_ret_stack;
1518 /* Stack of return addresses for return function tracing */
1519 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1520 /* time stamp for last schedule */
1521 unsigned long long ftrace_timestamp;
f201ae23
FW
1522 /*
1523 * Number of functions that haven't been traced
1524 * because of depth overrun.
1525 */
1526 atomic_t trace_overrun;
380c4b14
FW
1527 /* Pause for the tracing */
1528 atomic_t tracing_graph_pause;
f201ae23 1529#endif
ea4e2bc4
SR
1530#ifdef CONFIG_TRACING
1531 /* state flags for use by tracers */
1532 unsigned long trace;
261842b7
SR
1533 /* bitmask of trace recursion */
1534 unsigned long trace_recursion;
1535#endif /* CONFIG_TRACING */
569b846d
KH
1536#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1537 struct memcg_batch_info {
1538 int do_batch; /* incremented when batch uncharge started */
1539 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1540 unsigned long nr_pages; /* uncharged usage */
1541 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d
KH
1542 } memcg_batch;
1543#endif
1da177e4
LT
1544};
1545
76e6eee0 1546/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1547#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1548
e05606d3
IM
1549/*
1550 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1551 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1552 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1553 * values are inverted: lower p->prio value means higher priority.
1554 *
1555 * The MAX_USER_RT_PRIO value allows the actual maximum
1556 * RT priority to be separate from the value exported to
1557 * user-space. This allows kernel threads to set their
1558 * priority to a value higher than any user task. Note:
1559 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1560 */
1561
1562#define MAX_USER_RT_PRIO 100
1563#define MAX_RT_PRIO MAX_USER_RT_PRIO
1564
1565#define MAX_PRIO (MAX_RT_PRIO + 40)
1566#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1567
1568static inline int rt_prio(int prio)
1569{
1570 if (unlikely(prio < MAX_RT_PRIO))
1571 return 1;
1572 return 0;
1573}
1574
e868171a 1575static inline int rt_task(struct task_struct *p)
e05606d3
IM
1576{
1577 return rt_prio(p->prio);
1578}
1579
e868171a 1580static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1581{
1582 return task->pids[PIDTYPE_PID].pid;
1583}
1584
e868171a 1585static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1586{
1587 return task->group_leader->pids[PIDTYPE_PID].pid;
1588}
1589
6dda81f4
ON
1590/*
1591 * Without tasklist or rcu lock it is not safe to dereference
1592 * the result of task_pgrp/task_session even if task == current,
1593 * we can race with another thread doing sys_setsid/sys_setpgid.
1594 */
e868171a 1595static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1596{
1597 return task->group_leader->pids[PIDTYPE_PGID].pid;
1598}
1599
e868171a 1600static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1601{
1602 return task->group_leader->pids[PIDTYPE_SID].pid;
1603}
1604
7af57294
PE
1605struct pid_namespace;
1606
1607/*
1608 * the helpers to get the task's different pids as they are seen
1609 * from various namespaces
1610 *
1611 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1612 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1613 * current.
7af57294
PE
1614 * task_xid_nr_ns() : id seen from the ns specified;
1615 *
1616 * set_task_vxid() : assigns a virtual id to a task;
1617 *
7af57294
PE
1618 * see also pid_nr() etc in include/linux/pid.h
1619 */
52ee2dfd
ON
1620pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1621 struct pid_namespace *ns);
7af57294 1622
e868171a 1623static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1624{
1625 return tsk->pid;
1626}
1627
52ee2dfd
ON
1628static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1629 struct pid_namespace *ns)
1630{
1631 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1632}
7af57294
PE
1633
1634static inline pid_t task_pid_vnr(struct task_struct *tsk)
1635{
52ee2dfd 1636 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1637}
1638
1639
e868171a 1640static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1641{
1642 return tsk->tgid;
1643}
1644
2f2a3a46 1645pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1646
1647static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1648{
1649 return pid_vnr(task_tgid(tsk));
1650}
1651
1652
52ee2dfd
ON
1653static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1654 struct pid_namespace *ns)
7af57294 1655{
52ee2dfd 1656 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1657}
1658
7af57294
PE
1659static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1660{
52ee2dfd 1661 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1662}
1663
1664
52ee2dfd
ON
1665static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1666 struct pid_namespace *ns)
7af57294 1667{
52ee2dfd 1668 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1669}
1670
7af57294
PE
1671static inline pid_t task_session_vnr(struct task_struct *tsk)
1672{
52ee2dfd 1673 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1674}
1675
1b0f7ffd
ON
1676/* obsolete, do not use */
1677static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1678{
1679 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1680}
7af57294 1681
1da177e4
LT
1682/**
1683 * pid_alive - check that a task structure is not stale
1684 * @p: Task structure to be checked.
1685 *
1686 * Test if a process is not yet dead (at most zombie state)
1687 * If pid_alive fails, then pointers within the task structure
1688 * can be stale and must not be dereferenced.
1689 */
e868171a 1690static inline int pid_alive(struct task_struct *p)
1da177e4 1691{
92476d7f 1692 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1693}
1694
f400e198 1695/**
b460cbc5 1696 * is_global_init - check if a task structure is init
3260259f
HK
1697 * @tsk: Task structure to be checked.
1698 *
1699 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1700 */
e868171a 1701static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1702{
1703 return tsk->pid == 1;
1704}
b460cbc5
SH
1705
1706/*
1707 * is_container_init:
1708 * check whether in the task is init in its own pid namespace.
f400e198 1709 */
b461cc03 1710extern int is_container_init(struct task_struct *tsk);
f400e198 1711
9ec52099
CLG
1712extern struct pid *cad_pid;
1713
1da177e4 1714extern void free_task(struct task_struct *tsk);
1da177e4 1715#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1716
158d9ebd 1717extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1718
1719static inline void put_task_struct(struct task_struct *t)
1720{
1721 if (atomic_dec_and_test(&t->usage))
8c7904a0 1722 __put_task_struct(t);
e56d0903 1723}
1da177e4 1724
d180c5bc 1725extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1726extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1727
1da177e4
LT
1728/*
1729 * Per process flags
1730 */
1da177e4
LT
1731#define PF_STARTING 0x00000002 /* being created */
1732#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1733#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1734#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1735#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1736#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1737#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1738#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1739#define PF_DUMPCORE 0x00000200 /* dumped core */
1740#define PF_SIGNALED 0x00000400 /* killed by a signal */
1741#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1da177e4 1742#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1743#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1744#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1745#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1746#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1747#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1748#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1749#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1750#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1751#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1752#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1753#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1754#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1755#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1756#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1757#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1758#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1759#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
ebb12db5 1760#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1761
1762/*
1763 * Only the _current_ task can read/write to tsk->flags, but other
1764 * tasks can access tsk->flags in readonly mode for example
1765 * with tsk_used_math (like during threaded core dumping).
1766 * There is however an exception to this rule during ptrace
1767 * or during fork: the ptracer task is allowed to write to the
1768 * child->flags of its traced child (same goes for fork, the parent
1769 * can write to the child->flags), because we're guaranteed the
1770 * child is not running and in turn not changing child->flags
1771 * at the same time the parent does it.
1772 */
1773#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1774#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1775#define clear_used_math() clear_stopped_child_used_math(current)
1776#define set_used_math() set_stopped_child_used_math(current)
1777#define conditional_stopped_child_used_math(condition, child) \
1778 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1779#define conditional_used_math(condition) \
1780 conditional_stopped_child_used_math(condition, current)
1781#define copy_to_stopped_child_used_math(child) \
1782 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1783/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1784#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1785#define used_math() tsk_used_math(current)
1786
a57eb940 1787#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1788
1789#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
24278d14
PM
1790#define RCU_READ_UNLOCK_BOOSTED (1 << 1) /* boosted while in RCU read-side. */
1791#define RCU_READ_UNLOCK_NEED_QS (1 << 2) /* RCU core needs CPU response. */
f41d911f
PM
1792
1793static inline void rcu_copy_process(struct task_struct *p)
1794{
1795 p->rcu_read_lock_nesting = 0;
1796 p->rcu_read_unlock_special = 0;
a57eb940 1797#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1798 p->rcu_blocked_node = NULL;
24278d14
PM
1799#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1800#ifdef CONFIG_RCU_BOOST
1801 p->rcu_boost_mutex = NULL;
1802#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1803 INIT_LIST_HEAD(&p->rcu_node_entry);
1804}
1805
f41d911f
PM
1806#else
1807
1808static inline void rcu_copy_process(struct task_struct *p)
1809{
1810}
1811
1812#endif
1813
1da177e4 1814#ifdef CONFIG_SMP
cd8ba7cd 1815extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1816 const struct cpumask *new_mask);
1da177e4 1817#else
cd8ba7cd 1818static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1819 const struct cpumask *new_mask)
1da177e4 1820{
96f874e2 1821 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1822 return -EINVAL;
1823 return 0;
1824}
1825#endif
e0ad9556
RR
1826
1827#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1828static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1829{
1830 return set_cpus_allowed_ptr(p, &new_mask);
1831}
e0ad9556 1832#endif
1da177e4 1833
b342501c 1834/*
c676329a
PZ
1835 * Do not use outside of architecture code which knows its limitations.
1836 *
1837 * sched_clock() has no promise of monotonicity or bounded drift between
1838 * CPUs, use (which you should not) requires disabling IRQs.
1839 *
1840 * Please use one of the three interfaces below.
b342501c 1841 */
1bbfa6f2 1842extern unsigned long long notrace sched_clock(void);
c676329a
PZ
1843/*
1844 * See the comment in kernel/sched_clock.c
1845 */
1846extern u64 cpu_clock(int cpu);
1847extern u64 local_clock(void);
1848extern u64 sched_clock_cpu(int cpu);
1849
e436d800 1850
c1955a3d 1851extern void sched_clock_init(void);
3e51f33f 1852
c1955a3d 1853#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1854static inline void sched_clock_tick(void)
1855{
1856}
1857
1858static inline void sched_clock_idle_sleep_event(void)
1859{
1860}
1861
1862static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1863{
1864}
1865#else
c676329a
PZ
1866/*
1867 * Architectures can set this to 1 if they have specified
1868 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1869 * but then during bootup it turns out that sched_clock()
1870 * is reliable after all:
1871 */
1872extern int sched_clock_stable;
1873
3e51f33f
PZ
1874extern void sched_clock_tick(void);
1875extern void sched_clock_idle_sleep_event(void);
1876extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1877#endif
1878
b52bfee4
VP
1879#ifdef CONFIG_IRQ_TIME_ACCOUNTING
1880/*
1881 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
1882 * The reason for this explicit opt-in is not to have perf penalty with
1883 * slow sched_clocks.
1884 */
1885extern void enable_sched_clock_irqtime(void);
1886extern void disable_sched_clock_irqtime(void);
1887#else
1888static inline void enable_sched_clock_irqtime(void) {}
1889static inline void disable_sched_clock_irqtime(void) {}
1890#endif
1891
36c8b586 1892extern unsigned long long
41b86e9c 1893task_sched_runtime(struct task_struct *task);
f06febc9 1894extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1895
1896/* sched_exec is called by processes performing an exec */
1897#ifdef CONFIG_SMP
1898extern void sched_exec(void);
1899#else
1900#define sched_exec() {}
1901#endif
1902
2aa44d05
IM
1903extern void sched_clock_idle_sleep_event(void);
1904extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1905
1da177e4
LT
1906#ifdef CONFIG_HOTPLUG_CPU
1907extern void idle_task_exit(void);
1908#else
1909static inline void idle_task_exit(void) {}
1910#endif
1911
06d8308c
TG
1912#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1913extern void wake_up_idle_cpu(int cpu);
1914#else
1915static inline void wake_up_idle_cpu(int cpu) { }
1916#endif
1917
21805085 1918extern unsigned int sysctl_sched_latency;
b2be5e96 1919extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1920extern unsigned int sysctl_sched_wakeup_granularity;
bf0f6f24 1921extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
1922
1923enum sched_tunable_scaling {
1924 SCHED_TUNABLESCALING_NONE,
1925 SCHED_TUNABLESCALING_LOG,
1926 SCHED_TUNABLESCALING_LINEAR,
1927 SCHED_TUNABLESCALING_END,
1928};
1929extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1930
2bba22c5 1931#ifdef CONFIG_SCHED_DEBUG
da84d961 1932extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1933extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1934extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1935extern unsigned int sysctl_timer_migration;
a7a4f8a7 1936extern unsigned int sysctl_sched_shares_window;
b2be5e96 1937
1983a922 1938int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1939 void __user *buffer, size_t *length,
b2be5e96 1940 loff_t *ppos);
2bd8e6d4 1941#endif
eea08f32
AB
1942#ifdef CONFIG_SCHED_DEBUG
1943static inline unsigned int get_sysctl_timer_migration(void)
1944{
1945 return sysctl_timer_migration;
1946}
1947#else
1948static inline unsigned int get_sysctl_timer_migration(void)
1949{
1950 return 1;
1951}
1952#endif
9f0c1e56
PZ
1953extern unsigned int sysctl_sched_rt_period;
1954extern int sysctl_sched_rt_runtime;
2bd8e6d4 1955
d0b27fa7 1956int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1957 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
1958 loff_t *ppos);
1959
5091faa4
MG
1960#ifdef CONFIG_SCHED_AUTOGROUP
1961extern unsigned int sysctl_sched_autogroup_enabled;
1962
1963extern void sched_autogroup_create_attach(struct task_struct *p);
1964extern void sched_autogroup_detach(struct task_struct *p);
1965extern void sched_autogroup_fork(struct signal_struct *sig);
1966extern void sched_autogroup_exit(struct signal_struct *sig);
1967#ifdef CONFIG_PROC_FS
1968extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
1969extern int proc_sched_autogroup_set_nice(struct task_struct *p, int *nice);
1970#endif
1971#else
1972static inline void sched_autogroup_create_attach(struct task_struct *p) { }
1973static inline void sched_autogroup_detach(struct task_struct *p) { }
1974static inline void sched_autogroup_fork(struct signal_struct *sig) { }
1975static inline void sched_autogroup_exit(struct signal_struct *sig) { }
1976#endif
1977
b29739f9 1978#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1979extern int rt_mutex_getprio(struct task_struct *p);
1980extern void rt_mutex_setprio(struct task_struct *p, int prio);
1981extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1982#else
e868171a 1983static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1984{
1985 return p->normal_prio;
1986}
95e02ca9 1987# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1988#endif
1989
d95f4122 1990extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
1991extern void set_user_nice(struct task_struct *p, long nice);
1992extern int task_prio(const struct task_struct *p);
1993extern int task_nice(const struct task_struct *p);
1994extern int can_nice(const struct task_struct *p, const int nice);
1995extern int task_curr(const struct task_struct *p);
1da177e4 1996extern int idle_cpu(int cpu);
fe7de49f
KM
1997extern int sched_setscheduler(struct task_struct *, int,
1998 const struct sched_param *);
961ccddd 1999extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 2000 const struct sched_param *);
36c8b586
IM
2001extern struct task_struct *idle_task(int cpu);
2002extern struct task_struct *curr_task(int cpu);
2003extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
2004
2005void yield(void);
2006
2007/*
2008 * The default (Linux) execution domain.
2009 */
2010extern struct exec_domain default_exec_domain;
2011
2012union thread_union {
2013 struct thread_info thread_info;
2014 unsigned long stack[THREAD_SIZE/sizeof(long)];
2015};
2016
2017#ifndef __HAVE_ARCH_KSTACK_END
2018static inline int kstack_end(void *addr)
2019{
2020 /* Reliable end of stack detection:
2021 * Some APM bios versions misalign the stack
2022 */
2023 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2024}
2025#endif
2026
2027extern union thread_union init_thread_union;
2028extern struct task_struct init_task;
2029
2030extern struct mm_struct init_mm;
2031
198fe21b
PE
2032extern struct pid_namespace init_pid_ns;
2033
2034/*
2035 * find a task by one of its numerical ids
2036 *
198fe21b
PE
2037 * find_task_by_pid_ns():
2038 * finds a task by its pid in the specified namespace
228ebcbe
PE
2039 * find_task_by_vpid():
2040 * finds a task by its virtual pid
198fe21b 2041 *
e49859e7 2042 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2043 */
2044
228ebcbe
PE
2045extern struct task_struct *find_task_by_vpid(pid_t nr);
2046extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2047 struct pid_namespace *ns);
198fe21b 2048
8520d7c7 2049extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2050
2051/* per-UID process charging. */
acce292c 2052extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
2053static inline struct user_struct *get_uid(struct user_struct *u)
2054{
2055 atomic_inc(&u->__count);
2056 return u;
2057}
2058extern void free_uid(struct user_struct *);
28f300d2 2059extern void release_uids(struct user_namespace *ns);
1da177e4
LT
2060
2061#include <asm/current.h>
2062
f0af911a 2063extern void xtime_update(unsigned long ticks);
1da177e4 2064
b3c97528
HH
2065extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2066extern int wake_up_process(struct task_struct *tsk);
2067extern void wake_up_new_task(struct task_struct *tsk,
2068 unsigned long clone_flags);
1da177e4
LT
2069#ifdef CONFIG_SMP
2070 extern void kick_process(struct task_struct *tsk);
2071#else
2072 static inline void kick_process(struct task_struct *tsk) { }
2073#endif
ad46c2c4
IM
2074extern void sched_fork(struct task_struct *p, int clone_flags);
2075extern void sched_dead(struct task_struct *p);
1da177e4 2076
1da177e4
LT
2077extern void proc_caches_init(void);
2078extern void flush_signals(struct task_struct *);
3bcac026 2079extern void __flush_signals(struct task_struct *);
10ab825b 2080extern void ignore_signals(struct task_struct *);
1da177e4
LT
2081extern void flush_signal_handlers(struct task_struct *, int force_default);
2082extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2083
2084static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2085{
2086 unsigned long flags;
2087 int ret;
2088
2089 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2090 ret = dequeue_signal(tsk, mask, info);
2091 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2092
2093 return ret;
2094}
2095
2096extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2097 sigset_t *mask);
2098extern void unblock_all_signals(void);
2099extern void release_task(struct task_struct * p);
2100extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2101extern int force_sigsegv(int, struct task_struct *);
2102extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2103extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2104extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2105extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2106extern int kill_pgrp(struct pid *pid, int sig, int priv);
2107extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2108extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2109extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2110extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2111extern void force_sig(int, struct task_struct *);
1da177e4 2112extern int send_sig(int, struct task_struct *, int);
09faef11 2113extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2114extern struct sigqueue *sigqueue_alloc(void);
2115extern void sigqueue_free(struct sigqueue *);
ac5c2153 2116extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2117extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2118extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2119
9ec52099
CLG
2120static inline int kill_cad_pid(int sig, int priv)
2121{
2122 return kill_pid(cad_pid, sig, priv);
2123}
2124
1da177e4
LT
2125/* These can be the second arg to send_sig_info/send_group_sig_info. */
2126#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2127#define SEND_SIG_PRIV ((struct siginfo *) 1)
2128#define SEND_SIG_FORCED ((struct siginfo *) 2)
2129
2a855dd0
SAS
2130/*
2131 * True if we are on the alternate signal stack.
2132 */
1da177e4
LT
2133static inline int on_sig_stack(unsigned long sp)
2134{
2a855dd0
SAS
2135#ifdef CONFIG_STACK_GROWSUP
2136 return sp >= current->sas_ss_sp &&
2137 sp - current->sas_ss_sp < current->sas_ss_size;
2138#else
2139 return sp > current->sas_ss_sp &&
2140 sp - current->sas_ss_sp <= current->sas_ss_size;
2141#endif
1da177e4
LT
2142}
2143
2144static inline int sas_ss_flags(unsigned long sp)
2145{
2146 return (current->sas_ss_size == 0 ? SS_DISABLE
2147 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2148}
2149
1da177e4
LT
2150/*
2151 * Routines for handling mm_structs
2152 */
2153extern struct mm_struct * mm_alloc(void);
2154
2155/* mmdrop drops the mm and the page tables */
b3c97528 2156extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2157static inline void mmdrop(struct mm_struct * mm)
2158{
6fb43d7b 2159 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2160 __mmdrop(mm);
2161}
2162
2163/* mmput gets rid of the mappings and all user-space */
2164extern void mmput(struct mm_struct *);
2165/* Grab a reference to a task's mm, if it is not already going away */
2166extern struct mm_struct *get_task_mm(struct task_struct *task);
2167/* Remove the current tasks stale references to the old mm_struct */
2168extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2169/* Allocate a new mm structure and copy contents from tsk->mm */
2170extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2171
6f2c55b8
AD
2172extern int copy_thread(unsigned long, unsigned long, unsigned long,
2173 struct task_struct *, struct pt_regs *);
1da177e4
LT
2174extern void flush_thread(void);
2175extern void exit_thread(void);
2176
1da177e4 2177extern void exit_files(struct task_struct *);
a7e5328a 2178extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2179
1da177e4 2180extern void exit_itimers(struct signal_struct *);
cbaffba1 2181extern void flush_itimer_signals(void);
1da177e4
LT
2182
2183extern NORET_TYPE void do_group_exit(int);
2184
1da177e4
LT
2185extern void daemonize(const char *, ...);
2186extern int allow_signal(int);
2187extern int disallow_signal(int);
1da177e4 2188
d7627467
DH
2189extern int do_execve(const char *,
2190 const char __user * const __user *,
2191 const char __user * const __user *, struct pt_regs *);
1da177e4 2192extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2193struct task_struct *fork_idle(int);
1da177e4
LT
2194
2195extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2196extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2197
2198#ifdef CONFIG_SMP
85ba2d86 2199extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2200#else
85ba2d86
RM
2201static inline unsigned long wait_task_inactive(struct task_struct *p,
2202 long match_state)
2203{
2204 return 1;
2205}
1da177e4
LT
2206#endif
2207
05725f7e
JP
2208#define next_task(p) \
2209 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2210
2211#define for_each_process(p) \
2212 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2213
5bb459bb 2214extern bool current_is_single_threaded(void);
d84f4f99 2215
1da177e4
LT
2216/*
2217 * Careful: do_each_thread/while_each_thread is a double loop so
2218 * 'break' will not work as expected - use goto instead.
2219 */
2220#define do_each_thread(g, t) \
2221 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2222
2223#define while_each_thread(g, t) \
2224 while ((t = next_thread(t)) != g)
2225
7e49827c
ON
2226static inline int get_nr_threads(struct task_struct *tsk)
2227{
b3ac022c 2228 return tsk->signal->nr_threads;
7e49827c
ON
2229}
2230
de12a787
EB
2231/* de_thread depends on thread_group_leader not being a pid based check */
2232#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2233
0804ef4b
EB
2234/* Do to the insanities of de_thread it is possible for a process
2235 * to have the pid of the thread group leader without actually being
2236 * the thread group leader. For iteration through the pids in proc
2237 * all we care about is that we have a task with the appropriate
2238 * pid, we don't actually care if we have the right task.
2239 */
e868171a 2240static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2241{
2242 return p->pid == p->tgid;
2243}
2244
bac0abd6
PE
2245static inline
2246int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2247{
2248 return p1->tgid == p2->tgid;
2249}
2250
36c8b586 2251static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2252{
05725f7e
JP
2253 return list_entry_rcu(p->thread_group.next,
2254 struct task_struct, thread_group);
47e65328
ON
2255}
2256
e868171a 2257static inline int thread_group_empty(struct task_struct *p)
1da177e4 2258{
47e65328 2259 return list_empty(&p->thread_group);
1da177e4
LT
2260}
2261
2262#define delay_group_leader(p) \
2263 (thread_group_leader(p) && !thread_group_empty(p))
2264
39c626ae
ON
2265static inline int task_detached(struct task_struct *p)
2266{
2267 return p->exit_signal == -1;
2268}
2269
1da177e4 2270/*
260ea101 2271 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2272 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2273 * pins the final release of task.io_context. Also protects ->cpuset and
2274 * ->cgroup.subsys[].
1da177e4
LT
2275 *
2276 * Nests both inside and outside of read_lock(&tasklist_lock).
2277 * It must not be nested with write_lock_irq(&tasklist_lock),
2278 * neither inside nor outside.
2279 */
2280static inline void task_lock(struct task_struct *p)
2281{
2282 spin_lock(&p->alloc_lock);
2283}
2284
2285static inline void task_unlock(struct task_struct *p)
2286{
2287 spin_unlock(&p->alloc_lock);
2288}
2289
b8ed374e 2290extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2291 unsigned long *flags);
2292
b8ed374e
NK
2293#define lock_task_sighand(tsk, flags) \
2294({ struct sighand_struct *__ss; \
2295 __cond_lock(&(tsk)->sighand->siglock, \
2296 (__ss = __lock_task_sighand(tsk, flags))); \
2297 __ss; \
2298}) \
2299
f63ee72e
ON
2300static inline void unlock_task_sighand(struct task_struct *tsk,
2301 unsigned long *flags)
2302{
2303 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2304}
2305
f037360f
AV
2306#ifndef __HAVE_THREAD_FUNCTIONS
2307
f7e4217b
RZ
2308#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2309#define task_stack_page(task) ((task)->stack)
a1261f54 2310
10ebffde
AV
2311static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2312{
2313 *task_thread_info(p) = *task_thread_info(org);
2314 task_thread_info(p)->task = p;
2315}
2316
2317static inline unsigned long *end_of_stack(struct task_struct *p)
2318{
f7e4217b 2319 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2320}
2321
f037360f
AV
2322#endif
2323
8b05c7e6
FT
2324static inline int object_is_on_stack(void *obj)
2325{
2326 void *stack = task_stack_page(current);
2327
2328 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2329}
2330
8c9843e5
BH
2331extern void thread_info_cache_init(void);
2332
7c9f8861
ES
2333#ifdef CONFIG_DEBUG_STACK_USAGE
2334static inline unsigned long stack_not_used(struct task_struct *p)
2335{
2336 unsigned long *n = end_of_stack(p);
2337
2338 do { /* Skip over canary */
2339 n++;
2340 } while (!*n);
2341
2342 return (unsigned long)n - (unsigned long)end_of_stack(p);
2343}
2344#endif
2345
1da177e4
LT
2346/* set thread flags in other task's structures
2347 * - see asm/thread_info.h for TIF_xxxx flags available
2348 */
2349static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2350{
a1261f54 2351 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2352}
2353
2354static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2355{
a1261f54 2356 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2357}
2358
2359static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2360{
a1261f54 2361 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2362}
2363
2364static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2365{
a1261f54 2366 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2367}
2368
2369static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2370{
a1261f54 2371 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2372}
2373
2374static inline void set_tsk_need_resched(struct task_struct *tsk)
2375{
2376 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2377}
2378
2379static inline void clear_tsk_need_resched(struct task_struct *tsk)
2380{
2381 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2382}
2383
8ae121ac
GH
2384static inline int test_tsk_need_resched(struct task_struct *tsk)
2385{
2386 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2387}
2388
690cc3ff
EB
2389static inline int restart_syscall(void)
2390{
2391 set_tsk_thread_flag(current, TIF_SIGPENDING);
2392 return -ERESTARTNOINTR;
2393}
2394
1da177e4
LT
2395static inline int signal_pending(struct task_struct *p)
2396{
2397 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2398}
f776d12d 2399
d9588725
RM
2400static inline int __fatal_signal_pending(struct task_struct *p)
2401{
2402 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2403}
f776d12d
MW
2404
2405static inline int fatal_signal_pending(struct task_struct *p)
2406{
2407 return signal_pending(p) && __fatal_signal_pending(p);
2408}
2409
16882c1e
ON
2410static inline int signal_pending_state(long state, struct task_struct *p)
2411{
2412 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2413 return 0;
2414 if (!signal_pending(p))
2415 return 0;
2416
16882c1e
ON
2417 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2418}
2419
1da177e4
LT
2420static inline int need_resched(void)
2421{
9404ef02 2422 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2423}
2424
2425/*
2426 * cond_resched() and cond_resched_lock(): latency reduction via
2427 * explicit rescheduling in places that are safe. The return
2428 * value indicates whether a reschedule was done in fact.
2429 * cond_resched_lock() will drop the spinlock before scheduling,
2430 * cond_resched_softirq() will enable bhs before scheduling.
2431 */
c3921ab7 2432extern int _cond_resched(void);
6f80bd98 2433
613afbf8
FW
2434#define cond_resched() ({ \
2435 __might_sleep(__FILE__, __LINE__, 0); \
2436 _cond_resched(); \
2437})
6f80bd98 2438
613afbf8
FW
2439extern int __cond_resched_lock(spinlock_t *lock);
2440
716a4234
FW
2441#ifdef CONFIG_PREEMPT
2442#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2443#else
716a4234 2444#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2445#endif
716a4234 2446
613afbf8 2447#define cond_resched_lock(lock) ({ \
716a4234 2448 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2449 __cond_resched_lock(lock); \
2450})
2451
2452extern int __cond_resched_softirq(void);
2453
75e1056f
VP
2454#define cond_resched_softirq() ({ \
2455 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2456 __cond_resched_softirq(); \
613afbf8 2457})
1da177e4
LT
2458
2459/*
2460 * Does a critical section need to be broken due to another
95c354fe
NP
2461 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2462 * but a general need for low latency)
1da177e4 2463 */
95c354fe 2464static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2465{
95c354fe
NP
2466#ifdef CONFIG_PREEMPT
2467 return spin_is_contended(lock);
2468#else
1da177e4 2469 return 0;
95c354fe 2470#endif
1da177e4
LT
2471}
2472
f06febc9
FM
2473/*
2474 * Thread group CPU time accounting.
2475 */
4cd4c1b4 2476void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2477void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2478
490dea45 2479static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2480{
4cd4c1b4 2481 spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2482}
2483
7bb44ade
RM
2484/*
2485 * Reevaluate whether the task has signals pending delivery.
2486 * Wake the task if so.
2487 * This is required every time the blocked sigset_t changes.
2488 * callers must hold sighand->siglock.
2489 */
2490extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2491extern void recalc_sigpending(void);
2492
2493extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2494
2495/*
2496 * Wrappers for p->thread_info->cpu access. No-op on UP.
2497 */
2498#ifdef CONFIG_SMP
2499
2500static inline unsigned int task_cpu(const struct task_struct *p)
2501{
a1261f54 2502 return task_thread_info(p)->cpu;
1da177e4
LT
2503}
2504
c65cc870 2505extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2506
2507#else
2508
2509static inline unsigned int task_cpu(const struct task_struct *p)
2510{
2511 return 0;
2512}
2513
2514static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2515{
2516}
2517
2518#endif /* CONFIG_SMP */
2519
96f874e2
RR
2520extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2521extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2522
1da177e4
LT
2523extern void normalize_rt_tasks(void);
2524
7c941438 2525#ifdef CONFIG_CGROUP_SCHED
9b5b7751 2526
07e06b01 2527extern struct task_group root_task_group;
9b5b7751 2528
ec7dc8ac 2529extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2530extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2531extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2532#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2533extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2534extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2535#endif
2536#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2537extern int sched_group_set_rt_runtime(struct task_group *tg,
2538 long rt_runtime_us);
2539extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2540extern int sched_group_set_rt_period(struct task_group *tg,
2541 long rt_period_us);
2542extern long sched_group_rt_period(struct task_group *tg);
54e99124 2543extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2544#endif
9b5b7751
SV
2545#endif
2546
54e99124
DG
2547extern int task_can_switch_user(struct user_struct *up,
2548 struct task_struct *tsk);
2549
4b98d11b
AD
2550#ifdef CONFIG_TASK_XACCT
2551static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2552{
940389b8 2553 tsk->ioac.rchar += amt;
4b98d11b
AD
2554}
2555
2556static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2557{
940389b8 2558 tsk->ioac.wchar += amt;
4b98d11b
AD
2559}
2560
2561static inline void inc_syscr(struct task_struct *tsk)
2562{
940389b8 2563 tsk->ioac.syscr++;
4b98d11b
AD
2564}
2565
2566static inline void inc_syscw(struct task_struct *tsk)
2567{
940389b8 2568 tsk->ioac.syscw++;
4b98d11b
AD
2569}
2570#else
2571static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2572{
2573}
2574
2575static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2576{
2577}
2578
2579static inline void inc_syscr(struct task_struct *tsk)
2580{
2581}
2582
2583static inline void inc_syscw(struct task_struct *tsk)
2584{
2585}
2586#endif
2587
82455257
DH
2588#ifndef TASK_SIZE_OF
2589#define TASK_SIZE_OF(tsk) TASK_SIZE
2590#endif
2591
cf475ad2
BS
2592#ifdef CONFIG_MM_OWNER
2593extern void mm_update_next_owner(struct mm_struct *mm);
2594extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2595#else
2596static inline void mm_update_next_owner(struct mm_struct *mm)
2597{
2598}
2599
2600static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2601{
2602}
2603#endif /* CONFIG_MM_OWNER */
2604
3e10e716
JS
2605static inline unsigned long task_rlimit(const struct task_struct *tsk,
2606 unsigned int limit)
2607{
2608 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2609}
2610
2611static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2612 unsigned int limit)
2613{
2614 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2615}
2616
2617static inline unsigned long rlimit(unsigned int limit)
2618{
2619 return task_rlimit(current, limit);
2620}
2621
2622static inline unsigned long rlimit_max(unsigned int limit)
2623{
2624 return task_rlimit_max(current, limit);
2625}
2626
1da177e4
LT
2627#endif /* __KERNEL__ */
2628
2629#endif