sched: Deal with non-atomic min_vruntime reads on 32bits
[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);
c6eb3dda 363extern int mutex_spin_on_owner(struct mutex *lock, struct task_struct *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 */
1da177e4
LT
871
872 /*
873 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 874 * single CPU.
5517d86b 875 */
9d5efe05 876 unsigned int cpu_power, cpu_power_orig;
aae6d3dd 877 unsigned int group_weight;
6c99e9ad 878
4200efd9
IM
879 /*
880 * The CPUs this group covers.
881 *
882 * NOTE: this field is variable length. (Allocated dynamically
883 * by attaching extra space to the end of the structure,
884 * depending on how many CPUs the kernel has booted up with)
885 *
886 * It is also be embedded into static data structures at build
887 * time. (See 'struct static_sched_group' in kernel/sched.c)
888 */
889 unsigned long cpumask[0];
1da177e4
LT
890};
891
758b2cdc
RR
892static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
893{
6c99e9ad 894 return to_cpumask(sg->cpumask);
758b2cdc
RR
895}
896
1d3504fc
HS
897enum sched_domain_level {
898 SD_LV_NONE = 0,
899 SD_LV_SIBLING,
900 SD_LV_MC,
01a08546 901 SD_LV_BOOK,
1d3504fc
HS
902 SD_LV_CPU,
903 SD_LV_NODE,
904 SD_LV_ALLNODES,
905 SD_LV_MAX
906};
907
908struct sched_domain_attr {
909 int relax_domain_level;
910};
911
912#define SD_ATTR_INIT (struct sched_domain_attr) { \
913 .relax_domain_level = -1, \
914}
915
1da177e4
LT
916struct sched_domain {
917 /* These fields must be setup */
918 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 919 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 920 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
921 unsigned long min_interval; /* Minimum balance interval ms */
922 unsigned long max_interval; /* Maximum balance interval ms */
923 unsigned int busy_factor; /* less balancing by factor if busy */
924 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 925 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
926 unsigned int busy_idx;
927 unsigned int idle_idx;
928 unsigned int newidle_idx;
929 unsigned int wake_idx;
147cbb4b 930 unsigned int forkexec_idx;
a52bfd73 931 unsigned int smt_gain;
1da177e4 932 int flags; /* See SD_* */
1d3504fc 933 enum sched_domain_level level;
1da177e4
LT
934
935 /* Runtime fields. */
936 unsigned long last_balance; /* init to jiffies. units in jiffies */
937 unsigned int balance_interval; /* initialise to 1. units in ms. */
938 unsigned int nr_balance_failed; /* initialise to 0 */
939
2398f2c6
PZ
940 u64 last_update;
941
1da177e4
LT
942#ifdef CONFIG_SCHEDSTATS
943 /* load_balance() stats */
480b9434
KC
944 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
945 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
946 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
947 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
948 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
949 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
950 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
951 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
952
953 /* Active load balancing */
480b9434
KC
954 unsigned int alb_count;
955 unsigned int alb_failed;
956 unsigned int alb_pushed;
1da177e4 957
68767a0a 958 /* SD_BALANCE_EXEC stats */
480b9434
KC
959 unsigned int sbe_count;
960 unsigned int sbe_balanced;
961 unsigned int sbe_pushed;
1da177e4 962
68767a0a 963 /* SD_BALANCE_FORK stats */
480b9434
KC
964 unsigned int sbf_count;
965 unsigned int sbf_balanced;
966 unsigned int sbf_pushed;
68767a0a 967
1da177e4 968 /* try_to_wake_up() stats */
480b9434
KC
969 unsigned int ttwu_wake_remote;
970 unsigned int ttwu_move_affine;
971 unsigned int ttwu_move_balance;
1da177e4 972#endif
a5d8c348
IM
973#ifdef CONFIG_SCHED_DEBUG
974 char *name;
975#endif
6c99e9ad 976
669c55e9 977 unsigned int span_weight;
4200efd9
IM
978 /*
979 * Span of all CPUs in this domain.
980 *
981 * NOTE: this field is variable length. (Allocated dynamically
982 * by attaching extra space to the end of the structure,
983 * depending on how many CPUs the kernel has booted up with)
984 *
985 * It is also be embedded into static data structures at build
986 * time. (See 'struct static_sched_domain' in kernel/sched.c)
987 */
988 unsigned long span[0];
1da177e4
LT
989};
990
758b2cdc
RR
991static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
992{
6c99e9ad 993 return to_cpumask(sd->span);
758b2cdc
RR
994}
995
acc3f5d7 996extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 997 struct sched_domain_attr *dattr_new);
029190c5 998
acc3f5d7
RR
999/* Allocate an array of sched domains, for partition_sched_domains(). */
1000cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1001void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1002
06aaf76a
IM
1003/* Test a flag in parent sched domain */
1004static inline int test_sd_parent(struct sched_domain *sd, int flag)
1005{
1006 if (sd->parent && (sd->parent->flags & flag))
1007 return 1;
1008
1009 return 0;
1010}
029190c5 1011
47fe38fc
PZ
1012unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1013unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1014
1b427c15 1015#else /* CONFIG_SMP */
1da177e4 1016
1b427c15 1017struct sched_domain_attr;
d02c7a8c 1018
1b427c15 1019static inline void
acc3f5d7 1020partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1021 struct sched_domain_attr *dattr_new)
1022{
d02c7a8c 1023}
1b427c15 1024#endif /* !CONFIG_SMP */
1da177e4 1025
47fe38fc 1026
1da177e4 1027struct io_context; /* See blkdev.h */
1da177e4 1028
1da177e4 1029
383f2835 1030#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1031extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1032#else
1033static inline void prefetch_stack(struct task_struct *t) { }
1034#endif
1da177e4
LT
1035
1036struct audit_context; /* See audit.c */
1037struct mempolicy;
b92ce558 1038struct pipe_inode_info;
4865ecf1 1039struct uts_namespace;
1da177e4 1040
20b8a59f
IM
1041struct rq;
1042struct sched_domain;
1043
7d478721
PZ
1044/*
1045 * wake flags
1046 */
1047#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1048#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1049
371fd7e7 1050#define ENQUEUE_WAKEUP 1
74f8e4b2
PZ
1051#define ENQUEUE_HEAD 2
1052#ifdef CONFIG_SMP
1053#define ENQUEUE_WAKING 4 /* sched_class::task_waking was called */
1054#else
1055#define ENQUEUE_WAKING 0
1056#endif
371fd7e7
PZ
1057
1058#define DEQUEUE_SLEEP 1
1059
20b8a59f 1060struct sched_class {
5522d5d5 1061 const struct sched_class *next;
20b8a59f 1062
371fd7e7
PZ
1063 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int flags);
1064 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int flags);
4530d7ab 1065 void (*yield_task) (struct rq *rq);
d95f4122 1066 bool (*yield_to_task) (struct rq *rq, struct task_struct *p, bool preempt);
20b8a59f 1067
7d478721 1068 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1069
fb8d4724 1070 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1071 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1072
681f3e68 1073#ifdef CONFIG_SMP
7608dec2 1074 int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
4ce72a2c 1075
9a897c5a
SR
1076 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1077 void (*post_schedule) (struct rq *this_rq);
74f8e4b2 1078 void (*task_waking) (struct task_struct *task);
efbbd05a 1079 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1080
cd8ba7cd 1081 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1082 const struct cpumask *newmask);
57d885fe 1083
1f11eb6a
GH
1084 void (*rq_online)(struct rq *rq);
1085 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1086#endif
1087
1088 void (*set_curr_task) (struct rq *rq);
1089 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1090 void (*task_fork) (struct task_struct *p);
cb469845 1091
da7a735e
PZ
1092 void (*switched_from) (struct rq *this_rq, struct task_struct *task);
1093 void (*switched_to) (struct rq *this_rq, struct task_struct *task);
cb469845 1094 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
da7a735e 1095 int oldprio);
810b3817 1096
dba091b9
TG
1097 unsigned int (*get_rr_interval) (struct rq *rq,
1098 struct task_struct *task);
0d721cea 1099
810b3817 1100#ifdef CONFIG_FAIR_GROUP_SCHED
b2b5ce02 1101 void (*task_move_group) (struct task_struct *p, int on_rq);
810b3817 1102#endif
20b8a59f
IM
1103};
1104
1105struct load_weight {
1106 unsigned long weight, inv_weight;
1107};
1108
94c18227 1109#ifdef CONFIG_SCHEDSTATS
41acab88 1110struct sched_statistics {
20b8a59f 1111 u64 wait_start;
94c18227 1112 u64 wait_max;
6d082592
AV
1113 u64 wait_count;
1114 u64 wait_sum;
8f0dfc34
AV
1115 u64 iowait_count;
1116 u64 iowait_sum;
94c18227 1117
20b8a59f 1118 u64 sleep_start;
20b8a59f 1119 u64 sleep_max;
94c18227
IM
1120 s64 sum_sleep_runtime;
1121
1122 u64 block_start;
20b8a59f
IM
1123 u64 block_max;
1124 u64 exec_max;
eba1ed4b 1125 u64 slice_max;
cc367732 1126
cc367732
IM
1127 u64 nr_migrations_cold;
1128 u64 nr_failed_migrations_affine;
1129 u64 nr_failed_migrations_running;
1130 u64 nr_failed_migrations_hot;
1131 u64 nr_forced_migrations;
cc367732
IM
1132
1133 u64 nr_wakeups;
1134 u64 nr_wakeups_sync;
1135 u64 nr_wakeups_migrate;
1136 u64 nr_wakeups_local;
1137 u64 nr_wakeups_remote;
1138 u64 nr_wakeups_affine;
1139 u64 nr_wakeups_affine_attempts;
1140 u64 nr_wakeups_passive;
1141 u64 nr_wakeups_idle;
41acab88
LDM
1142};
1143#endif
1144
1145struct sched_entity {
1146 struct load_weight load; /* for load-balancing */
1147 struct rb_node run_node;
1148 struct list_head group_node;
1149 unsigned int on_rq;
1150
1151 u64 exec_start;
1152 u64 sum_exec_runtime;
1153 u64 vruntime;
1154 u64 prev_sum_exec_runtime;
1155
41acab88
LDM
1156 u64 nr_migrations;
1157
41acab88
LDM
1158#ifdef CONFIG_SCHEDSTATS
1159 struct sched_statistics statistics;
94c18227
IM
1160#endif
1161
20b8a59f
IM
1162#ifdef CONFIG_FAIR_GROUP_SCHED
1163 struct sched_entity *parent;
1164 /* rq on which this entity is (to be) queued: */
1165 struct cfs_rq *cfs_rq;
1166 /* rq "owned" by this entity/group: */
1167 struct cfs_rq *my_q;
1168#endif
1169};
70b97a7f 1170
fa717060
PZ
1171struct sched_rt_entity {
1172 struct list_head run_list;
78f2c7db 1173 unsigned long timeout;
bee367ed 1174 unsigned int time_slice;
6f505b16
PZ
1175 int nr_cpus_allowed;
1176
58d6c2d7 1177 struct sched_rt_entity *back;
052f1dc7 1178#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1179 struct sched_rt_entity *parent;
1180 /* rq on which this entity is (to be) queued: */
1181 struct rt_rq *rt_rq;
1182 /* rq "owned" by this entity/group: */
1183 struct rt_rq *my_q;
1184#endif
fa717060
PZ
1185};
1186
86848966
PM
1187struct rcu_node;
1188
8dc85d54
PZ
1189enum perf_event_task_context {
1190 perf_invalid_context = -1,
1191 perf_hw_context = 0,
89a1e187 1192 perf_sw_context,
8dc85d54
PZ
1193 perf_nr_task_contexts,
1194};
1195
1da177e4
LT
1196struct task_struct {
1197 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1198 void *stack;
1da177e4 1199 atomic_t usage;
97dc32cd
WC
1200 unsigned int flags; /* per process flags, defined below */
1201 unsigned int ptrace;
1da177e4 1202
36772092 1203 int lock_depth; /* BKL lock depth */
1da177e4 1204
2dd73a4f 1205#ifdef CONFIG_SMP
3ca7a440 1206 int on_cpu;
2dd73a4f 1207#endif
fd2f4419 1208 int on_rq;
50e645a8 1209
b29739f9 1210 int prio, static_prio, normal_prio;
c7aceaba 1211 unsigned int rt_priority;
5522d5d5 1212 const struct sched_class *sched_class;
20b8a59f 1213 struct sched_entity se;
fa717060 1214 struct sched_rt_entity rt;
1da177e4 1215
e107be36
AK
1216#ifdef CONFIG_PREEMPT_NOTIFIERS
1217 /* list of struct preempt_notifier: */
1218 struct hlist_head preempt_notifiers;
1219#endif
1220
18796aa0
AD
1221 /*
1222 * fpu_counter contains the number of consecutive context switches
1223 * that the FPU is used. If this is over a threshold, the lazy fpu
1224 * saving becomes unlazy to save the trap. This is an unsigned char
1225 * so that after 256 times the counter wraps and the behavior turns
1226 * lazy again; this to deal with bursty apps that only use FPU for
1227 * a short time
1228 */
1229 unsigned char fpu_counter;
6c5c9341 1230#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1231 unsigned int btrace_seq;
6c5c9341 1232#endif
1da177e4 1233
97dc32cd 1234 unsigned int policy;
1da177e4 1235 cpumask_t cpus_allowed;
1da177e4 1236
a57eb940 1237#ifdef CONFIG_PREEMPT_RCU
e260be67 1238 int rcu_read_lock_nesting;
f41d911f 1239 char rcu_read_unlock_special;
f41d911f 1240 struct list_head rcu_node_entry;
a57eb940
PM
1241#endif /* #ifdef CONFIG_PREEMPT_RCU */
1242#ifdef CONFIG_TREE_PREEMPT_RCU
1243 struct rcu_node *rcu_blocked_node;
f41d911f 1244#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1245#ifdef CONFIG_RCU_BOOST
1246 struct rt_mutex *rcu_boost_mutex;
1247#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1248
52f17b6c 1249#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1250 struct sched_info sched_info;
1251#endif
1252
1253 struct list_head tasks;
806c09a7 1254#ifdef CONFIG_SMP
917b627d 1255 struct plist_node pushable_tasks;
806c09a7 1256#endif
1da177e4
LT
1257
1258 struct mm_struct *mm, *active_mm;
34e55232
KH
1259#if defined(SPLIT_RSS_COUNTING)
1260 struct task_rss_stat rss_stat;
1261#endif
1da177e4 1262/* task state */
97dc32cd 1263 int exit_state;
1da177e4
LT
1264 int exit_code, exit_signal;
1265 int pdeath_signal; /* The signal sent when the parent dies */
1266 /* ??? */
97dc32cd 1267 unsigned int personality;
1da177e4 1268 unsigned did_exec:1;
f9ce1f1c
KT
1269 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1270 * execve */
8f0dfc34
AV
1271 unsigned in_iowait:1;
1272
ca94c442
LP
1273
1274 /* Revert to default priority/policy when forking */
1275 unsigned sched_reset_on_fork:1;
1276
1da177e4
LT
1277 pid_t pid;
1278 pid_t tgid;
0a425405 1279
1314562a 1280#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1281 /* Canary value for the -fstack-protector gcc feature */
1282 unsigned long stack_canary;
1314562a 1283#endif
e0032087 1284
1da177e4
LT
1285 /*
1286 * pointers to (original) parent process, youngest child, younger sibling,
1287 * older sibling, respectively. (p->father can be replaced with
f470021a 1288 * p->real_parent->pid)
1da177e4 1289 */
f470021a
RM
1290 struct task_struct *real_parent; /* real parent process */
1291 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1292 /*
f470021a 1293 * children/sibling forms the list of my natural children
1da177e4
LT
1294 */
1295 struct list_head children; /* list of my children */
1296 struct list_head sibling; /* linkage in my parent's children list */
1297 struct task_struct *group_leader; /* threadgroup leader */
1298
f470021a
RM
1299 /*
1300 * ptraced is the list of tasks this task is using ptrace on.
1301 * This includes both natural children and PTRACE_ATTACH targets.
1302 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1303 */
1304 struct list_head ptraced;
1305 struct list_head ptrace_entry;
1306
1da177e4 1307 /* PID/PID hash table linkage. */
92476d7f 1308 struct pid_link pids[PIDTYPE_MAX];
47e65328 1309 struct list_head thread_group;
1da177e4
LT
1310
1311 struct completion *vfork_done; /* for vfork() */
1312 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1313 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1314
c66f08be 1315 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1316 cputime_t gtime;
d99ca3b9 1317#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1318 cputime_t prev_utime, prev_stime;
d99ca3b9 1319#endif
1da177e4 1320 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1321 struct timespec start_time; /* monotonic time */
1322 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1323/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1324 unsigned long min_flt, maj_flt;
1325
f06febc9 1326 struct task_cputime cputime_expires;
1da177e4
LT
1327 struct list_head cpu_timers[3];
1328
1329/* process credentials */
1b0ba1c9 1330 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1331 * credentials (COW) */
1b0ba1c9 1332 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1333 * credentials (COW) */
ee18d64c 1334 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1335
36772092
PBG
1336 char comm[TASK_COMM_LEN]; /* executable name excluding path
1337 - access with [gs]et_task_comm (which lock
1338 it with task_lock())
221af7f8 1339 - initialized normally by setup_new_exec */
1da177e4
LT
1340/* file system info */
1341 int link_count, total_link_count;
3d5b6fcc 1342#ifdef CONFIG_SYSVIPC
1da177e4
LT
1343/* ipc stuff */
1344 struct sysv_sem sysvsem;
3d5b6fcc 1345#endif
e162b39a 1346#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1347/* hung task detection */
82a1fcb9
IM
1348 unsigned long last_switch_count;
1349#endif
1da177e4
LT
1350/* CPU-specific state of this task */
1351 struct thread_struct thread;
1352/* filesystem information */
1353 struct fs_struct *fs;
1354/* open file information */
1355 struct files_struct *files;
1651e14e 1356/* namespaces */
ab516013 1357 struct nsproxy *nsproxy;
1da177e4
LT
1358/* signal handlers */
1359 struct signal_struct *signal;
1360 struct sighand_struct *sighand;
1361
1362 sigset_t blocked, real_blocked;
f3de272b 1363 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1364 struct sigpending pending;
1365
1366 unsigned long sas_ss_sp;
1367 size_t sas_ss_size;
1368 int (*notifier)(void *priv);
1369 void *notifier_data;
1370 sigset_t *notifier_mask;
1da177e4 1371 struct audit_context *audit_context;
bfef93a5
AV
1372#ifdef CONFIG_AUDITSYSCALL
1373 uid_t loginuid;
4746ec5b 1374 unsigned int sessionid;
bfef93a5 1375#endif
1da177e4
LT
1376 seccomp_t seccomp;
1377
1378/* Thread group tracking */
1379 u32 parent_exec_id;
1380 u32 self_exec_id;
58568d2a
MX
1381/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1382 * mempolicy */
1da177e4 1383 spinlock_t alloc_lock;
1da177e4 1384
3aa551c9
TG
1385#ifdef CONFIG_GENERIC_HARDIRQS
1386 /* IRQ handler threads */
1387 struct irqaction *irqaction;
1388#endif
1389
b29739f9 1390 /* Protection of the PI data structures: */
1d615482 1391 raw_spinlock_t pi_lock;
b29739f9 1392
23f78d4a
IM
1393#ifdef CONFIG_RT_MUTEXES
1394 /* PI waiters blocked on a rt_mutex held by this task */
1395 struct plist_head pi_waiters;
1396 /* Deadlock detection and priority inheritance handling */
1397 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1398#endif
1399
408894ee
IM
1400#ifdef CONFIG_DEBUG_MUTEXES
1401 /* mutex deadlock detection */
1402 struct mutex_waiter *blocked_on;
1403#endif
de30a2b3
IM
1404#ifdef CONFIG_TRACE_IRQFLAGS
1405 unsigned int irq_events;
de30a2b3 1406 unsigned long hardirq_enable_ip;
de30a2b3 1407 unsigned long hardirq_disable_ip;
fa1452e8 1408 unsigned int hardirq_enable_event;
de30a2b3 1409 unsigned int hardirq_disable_event;
fa1452e8
HS
1410 int hardirqs_enabled;
1411 int hardirq_context;
de30a2b3 1412 unsigned long softirq_disable_ip;
de30a2b3 1413 unsigned long softirq_enable_ip;
fa1452e8 1414 unsigned int softirq_disable_event;
de30a2b3 1415 unsigned int softirq_enable_event;
fa1452e8 1416 int softirqs_enabled;
de30a2b3
IM
1417 int softirq_context;
1418#endif
fbb9ce95 1419#ifdef CONFIG_LOCKDEP
bdb9441e 1420# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1421 u64 curr_chain_key;
1422 int lockdep_depth;
fbb9ce95 1423 unsigned int lockdep_recursion;
c7aceaba 1424 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1425 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1426#endif
408894ee 1427
1da177e4
LT
1428/* journalling filesystem info */
1429 void *journal_info;
1430
d89d8796 1431/* stacked block device info */
bddd87c7 1432 struct bio_list *bio_list;
d89d8796 1433
73c10101
JA
1434#ifdef CONFIG_BLOCK
1435/* stack plugging */
1436 struct blk_plug *plug;
1437#endif
1438
1da177e4
LT
1439/* VM state */
1440 struct reclaim_state *reclaim_state;
1441
1da177e4
LT
1442 struct backing_dev_info *backing_dev_info;
1443
1444 struct io_context *io_context;
1445
1446 unsigned long ptrace_message;
1447 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1448 struct task_io_accounting ioac;
8f0ab514 1449#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1450 u64 acct_rss_mem1; /* accumulated rss usage */
1451 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1452 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1453#endif
1454#ifdef CONFIG_CPUSETS
58568d2a 1455 nodemask_t mems_allowed; /* Protected by alloc_lock */
c0ff7453 1456 int mems_allowed_change_disable;
825a46af 1457 int cpuset_mem_spread_rotor;
6adef3eb 1458 int cpuset_slab_spread_rotor;
1da177e4 1459#endif
ddbcc7e8 1460#ifdef CONFIG_CGROUPS
817929ec 1461 /* Control Group info protected by css_set_lock */
2c392b8c 1462 struct css_set __rcu *cgroups;
817929ec
PM
1463 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1464 struct list_head cg_list;
ddbcc7e8 1465#endif
42b2dd0a 1466#ifdef CONFIG_FUTEX
0771dfef 1467 struct robust_list_head __user *robust_list;
34f192c6
IM
1468#ifdef CONFIG_COMPAT
1469 struct compat_robust_list_head __user *compat_robust_list;
1470#endif
c87e2837
IM
1471 struct list_head pi_state_list;
1472 struct futex_pi_state *pi_state_cache;
c7aceaba 1473#endif
cdd6c482 1474#ifdef CONFIG_PERF_EVENTS
8dc85d54 1475 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1476 struct mutex perf_event_mutex;
1477 struct list_head perf_event_list;
a63eaf34 1478#endif
c7aceaba 1479#ifdef CONFIG_NUMA
58568d2a 1480 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1481 short il_next;
207205a2 1482 short pref_node_fork;
42b2dd0a 1483#endif
22e2c507 1484 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1485 struct rcu_head rcu;
b92ce558
JA
1486
1487 /*
1488 * cache last used pipe for splice
1489 */
1490 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1491#ifdef CONFIG_TASK_DELAY_ACCT
1492 struct task_delay_info *delays;
f4f154fd
AM
1493#endif
1494#ifdef CONFIG_FAULT_INJECTION
1495 int make_it_fail;
ca74e92b 1496#endif
3e26c149 1497 struct prop_local_single dirties;
9745512c
AV
1498#ifdef CONFIG_LATENCYTOP
1499 int latency_record_count;
1500 struct latency_record latency_record[LT_SAVECOUNT];
1501#endif
6976675d
AV
1502 /*
1503 * time slack values; these are used to round up poll() and
1504 * select() etc timeout values. These are in nanoseconds.
1505 */
1506 unsigned long timer_slack_ns;
1507 unsigned long default_timer_slack_ns;
f8d570a4
DM
1508
1509 struct list_head *scm_work_list;
fb52607a 1510#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1511 /* Index of current stored address in ret_stack */
f201ae23
FW
1512 int curr_ret_stack;
1513 /* Stack of return addresses for return function tracing */
1514 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1515 /* time stamp for last schedule */
1516 unsigned long long ftrace_timestamp;
f201ae23
FW
1517 /*
1518 * Number of functions that haven't been traced
1519 * because of depth overrun.
1520 */
1521 atomic_t trace_overrun;
380c4b14
FW
1522 /* Pause for the tracing */
1523 atomic_t tracing_graph_pause;
f201ae23 1524#endif
ea4e2bc4
SR
1525#ifdef CONFIG_TRACING
1526 /* state flags for use by tracers */
1527 unsigned long trace;
261842b7
SR
1528 /* bitmask of trace recursion */
1529 unsigned long trace_recursion;
1530#endif /* CONFIG_TRACING */
569b846d
KH
1531#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1532 struct memcg_batch_info {
1533 int do_batch; /* incremented when batch uncharge started */
1534 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1535 unsigned long nr_pages; /* uncharged usage */
1536 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d
KH
1537 } memcg_batch;
1538#endif
1da177e4
LT
1539};
1540
76e6eee0 1541/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1542#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1543
e05606d3
IM
1544/*
1545 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1546 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1547 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1548 * values are inverted: lower p->prio value means higher priority.
1549 *
1550 * The MAX_USER_RT_PRIO value allows the actual maximum
1551 * RT priority to be separate from the value exported to
1552 * user-space. This allows kernel threads to set their
1553 * priority to a value higher than any user task. Note:
1554 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1555 */
1556
1557#define MAX_USER_RT_PRIO 100
1558#define MAX_RT_PRIO MAX_USER_RT_PRIO
1559
1560#define MAX_PRIO (MAX_RT_PRIO + 40)
1561#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1562
1563static inline int rt_prio(int prio)
1564{
1565 if (unlikely(prio < MAX_RT_PRIO))
1566 return 1;
1567 return 0;
1568}
1569
e868171a 1570static inline int rt_task(struct task_struct *p)
e05606d3
IM
1571{
1572 return rt_prio(p->prio);
1573}
1574
e868171a 1575static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1576{
1577 return task->pids[PIDTYPE_PID].pid;
1578}
1579
e868171a 1580static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1581{
1582 return task->group_leader->pids[PIDTYPE_PID].pid;
1583}
1584
6dda81f4
ON
1585/*
1586 * Without tasklist or rcu lock it is not safe to dereference
1587 * the result of task_pgrp/task_session even if task == current,
1588 * we can race with another thread doing sys_setsid/sys_setpgid.
1589 */
e868171a 1590static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1591{
1592 return task->group_leader->pids[PIDTYPE_PGID].pid;
1593}
1594
e868171a 1595static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1596{
1597 return task->group_leader->pids[PIDTYPE_SID].pid;
1598}
1599
7af57294
PE
1600struct pid_namespace;
1601
1602/*
1603 * the helpers to get the task's different pids as they are seen
1604 * from various namespaces
1605 *
1606 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1607 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1608 * current.
7af57294
PE
1609 * task_xid_nr_ns() : id seen from the ns specified;
1610 *
1611 * set_task_vxid() : assigns a virtual id to a task;
1612 *
7af57294
PE
1613 * see also pid_nr() etc in include/linux/pid.h
1614 */
52ee2dfd
ON
1615pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1616 struct pid_namespace *ns);
7af57294 1617
e868171a 1618static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1619{
1620 return tsk->pid;
1621}
1622
52ee2dfd
ON
1623static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1624 struct pid_namespace *ns)
1625{
1626 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1627}
7af57294
PE
1628
1629static inline pid_t task_pid_vnr(struct task_struct *tsk)
1630{
52ee2dfd 1631 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1632}
1633
1634
e868171a 1635static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1636{
1637 return tsk->tgid;
1638}
1639
2f2a3a46 1640pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1641
1642static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1643{
1644 return pid_vnr(task_tgid(tsk));
1645}
1646
1647
52ee2dfd
ON
1648static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1649 struct pid_namespace *ns)
7af57294 1650{
52ee2dfd 1651 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1652}
1653
7af57294
PE
1654static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1655{
52ee2dfd 1656 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1657}
1658
1659
52ee2dfd
ON
1660static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1661 struct pid_namespace *ns)
7af57294 1662{
52ee2dfd 1663 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1664}
1665
7af57294
PE
1666static inline pid_t task_session_vnr(struct task_struct *tsk)
1667{
52ee2dfd 1668 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1669}
1670
1b0f7ffd
ON
1671/* obsolete, do not use */
1672static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1673{
1674 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1675}
7af57294 1676
1da177e4
LT
1677/**
1678 * pid_alive - check that a task structure is not stale
1679 * @p: Task structure to be checked.
1680 *
1681 * Test if a process is not yet dead (at most zombie state)
1682 * If pid_alive fails, then pointers within the task structure
1683 * can be stale and must not be dereferenced.
1684 */
e868171a 1685static inline int pid_alive(struct task_struct *p)
1da177e4 1686{
92476d7f 1687 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1688}
1689
f400e198 1690/**
b460cbc5 1691 * is_global_init - check if a task structure is init
3260259f
HK
1692 * @tsk: Task structure to be checked.
1693 *
1694 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1695 */
e868171a 1696static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1697{
1698 return tsk->pid == 1;
1699}
b460cbc5
SH
1700
1701/*
1702 * is_container_init:
1703 * check whether in the task is init in its own pid namespace.
f400e198 1704 */
b461cc03 1705extern int is_container_init(struct task_struct *tsk);
f400e198 1706
9ec52099
CLG
1707extern struct pid *cad_pid;
1708
1da177e4 1709extern void free_task(struct task_struct *tsk);
1da177e4 1710#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1711
158d9ebd 1712extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1713
1714static inline void put_task_struct(struct task_struct *t)
1715{
1716 if (atomic_dec_and_test(&t->usage))
8c7904a0 1717 __put_task_struct(t);
e56d0903 1718}
1da177e4 1719
d180c5bc 1720extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1721extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1722
1da177e4
LT
1723/*
1724 * Per process flags
1725 */
1da177e4
LT
1726#define PF_STARTING 0x00000002 /* being created */
1727#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1728#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1729#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1730#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1731#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1732#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1733#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1734#define PF_DUMPCORE 0x00000200 /* dumped core */
1735#define PF_SIGNALED 0x00000400 /* killed by a signal */
1736#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1da177e4 1737#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1738#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1739#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1740#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1741#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1742#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1743#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1744#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1745#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1746#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1747#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1748#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1749#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1750#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1751#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1752#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1753#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1754#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
ebb12db5 1755#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1756
1757/*
1758 * Only the _current_ task can read/write to tsk->flags, but other
1759 * tasks can access tsk->flags in readonly mode for example
1760 * with tsk_used_math (like during threaded core dumping).
1761 * There is however an exception to this rule during ptrace
1762 * or during fork: the ptracer task is allowed to write to the
1763 * child->flags of its traced child (same goes for fork, the parent
1764 * can write to the child->flags), because we're guaranteed the
1765 * child is not running and in turn not changing child->flags
1766 * at the same time the parent does it.
1767 */
1768#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1769#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1770#define clear_used_math() clear_stopped_child_used_math(current)
1771#define set_used_math() set_stopped_child_used_math(current)
1772#define conditional_stopped_child_used_math(condition, child) \
1773 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1774#define conditional_used_math(condition) \
1775 conditional_stopped_child_used_math(condition, current)
1776#define copy_to_stopped_child_used_math(child) \
1777 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1778/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1779#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1780#define used_math() tsk_used_math(current)
1781
a57eb940 1782#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1783
1784#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
24278d14
PM
1785#define RCU_READ_UNLOCK_BOOSTED (1 << 1) /* boosted while in RCU read-side. */
1786#define RCU_READ_UNLOCK_NEED_QS (1 << 2) /* RCU core needs CPU response. */
f41d911f
PM
1787
1788static inline void rcu_copy_process(struct task_struct *p)
1789{
1790 p->rcu_read_lock_nesting = 0;
1791 p->rcu_read_unlock_special = 0;
a57eb940 1792#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1793 p->rcu_blocked_node = NULL;
24278d14
PM
1794#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1795#ifdef CONFIG_RCU_BOOST
1796 p->rcu_boost_mutex = NULL;
1797#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1798 INIT_LIST_HEAD(&p->rcu_node_entry);
1799}
1800
f41d911f
PM
1801#else
1802
1803static inline void rcu_copy_process(struct task_struct *p)
1804{
1805}
1806
1807#endif
1808
1da177e4 1809#ifdef CONFIG_SMP
cd8ba7cd 1810extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1811 const struct cpumask *new_mask);
1da177e4 1812#else
cd8ba7cd 1813static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1814 const struct cpumask *new_mask)
1da177e4 1815{
96f874e2 1816 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1817 return -EINVAL;
1818 return 0;
1819}
1820#endif
e0ad9556
RR
1821
1822#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1823static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1824{
1825 return set_cpus_allowed_ptr(p, &new_mask);
1826}
e0ad9556 1827#endif
1da177e4 1828
b342501c 1829/*
c676329a
PZ
1830 * Do not use outside of architecture code which knows its limitations.
1831 *
1832 * sched_clock() has no promise of monotonicity or bounded drift between
1833 * CPUs, use (which you should not) requires disabling IRQs.
1834 *
1835 * Please use one of the three interfaces below.
b342501c 1836 */
1bbfa6f2 1837extern unsigned long long notrace sched_clock(void);
c676329a
PZ
1838/*
1839 * See the comment in kernel/sched_clock.c
1840 */
1841extern u64 cpu_clock(int cpu);
1842extern u64 local_clock(void);
1843extern u64 sched_clock_cpu(int cpu);
1844
e436d800 1845
c1955a3d 1846extern void sched_clock_init(void);
3e51f33f 1847
c1955a3d 1848#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1849static inline void sched_clock_tick(void)
1850{
1851}
1852
1853static inline void sched_clock_idle_sleep_event(void)
1854{
1855}
1856
1857static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1858{
1859}
1860#else
c676329a
PZ
1861/*
1862 * Architectures can set this to 1 if they have specified
1863 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1864 * but then during bootup it turns out that sched_clock()
1865 * is reliable after all:
1866 */
1867extern int sched_clock_stable;
1868
3e51f33f
PZ
1869extern void sched_clock_tick(void);
1870extern void sched_clock_idle_sleep_event(void);
1871extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1872#endif
1873
b52bfee4
VP
1874#ifdef CONFIG_IRQ_TIME_ACCOUNTING
1875/*
1876 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
1877 * The reason for this explicit opt-in is not to have perf penalty with
1878 * slow sched_clocks.
1879 */
1880extern void enable_sched_clock_irqtime(void);
1881extern void disable_sched_clock_irqtime(void);
1882#else
1883static inline void enable_sched_clock_irqtime(void) {}
1884static inline void disable_sched_clock_irqtime(void) {}
1885#endif
1886
36c8b586 1887extern unsigned long long
41b86e9c 1888task_sched_runtime(struct task_struct *task);
f06febc9 1889extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1890
1891/* sched_exec is called by processes performing an exec */
1892#ifdef CONFIG_SMP
1893extern void sched_exec(void);
1894#else
1895#define sched_exec() {}
1896#endif
1897
2aa44d05
IM
1898extern void sched_clock_idle_sleep_event(void);
1899extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1900
1da177e4
LT
1901#ifdef CONFIG_HOTPLUG_CPU
1902extern void idle_task_exit(void);
1903#else
1904static inline void idle_task_exit(void) {}
1905#endif
1906
06d8308c
TG
1907#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1908extern void wake_up_idle_cpu(int cpu);
1909#else
1910static inline void wake_up_idle_cpu(int cpu) { }
1911#endif
1912
21805085 1913extern unsigned int sysctl_sched_latency;
b2be5e96 1914extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1915extern unsigned int sysctl_sched_wakeup_granularity;
bf0f6f24 1916extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
1917
1918enum sched_tunable_scaling {
1919 SCHED_TUNABLESCALING_NONE,
1920 SCHED_TUNABLESCALING_LOG,
1921 SCHED_TUNABLESCALING_LINEAR,
1922 SCHED_TUNABLESCALING_END,
1923};
1924extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1925
2bba22c5 1926#ifdef CONFIG_SCHED_DEBUG
da84d961 1927extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1928extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1929extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1930extern unsigned int sysctl_timer_migration;
a7a4f8a7 1931extern unsigned int sysctl_sched_shares_window;
b2be5e96 1932
1983a922 1933int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1934 void __user *buffer, size_t *length,
b2be5e96 1935 loff_t *ppos);
2bd8e6d4 1936#endif
eea08f32
AB
1937#ifdef CONFIG_SCHED_DEBUG
1938static inline unsigned int get_sysctl_timer_migration(void)
1939{
1940 return sysctl_timer_migration;
1941}
1942#else
1943static inline unsigned int get_sysctl_timer_migration(void)
1944{
1945 return 1;
1946}
1947#endif
9f0c1e56
PZ
1948extern unsigned int sysctl_sched_rt_period;
1949extern int sysctl_sched_rt_runtime;
2bd8e6d4 1950
d0b27fa7 1951int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1952 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
1953 loff_t *ppos);
1954
5091faa4
MG
1955#ifdef CONFIG_SCHED_AUTOGROUP
1956extern unsigned int sysctl_sched_autogroup_enabled;
1957
1958extern void sched_autogroup_create_attach(struct task_struct *p);
1959extern void sched_autogroup_detach(struct task_struct *p);
1960extern void sched_autogroup_fork(struct signal_struct *sig);
1961extern void sched_autogroup_exit(struct signal_struct *sig);
1962#ifdef CONFIG_PROC_FS
1963extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
1964extern int proc_sched_autogroup_set_nice(struct task_struct *p, int *nice);
1965#endif
1966#else
1967static inline void sched_autogroup_create_attach(struct task_struct *p) { }
1968static inline void sched_autogroup_detach(struct task_struct *p) { }
1969static inline void sched_autogroup_fork(struct signal_struct *sig) { }
1970static inline void sched_autogroup_exit(struct signal_struct *sig) { }
1971#endif
1972
b29739f9 1973#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1974extern int rt_mutex_getprio(struct task_struct *p);
1975extern void rt_mutex_setprio(struct task_struct *p, int prio);
1976extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1977#else
e868171a 1978static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1979{
1980 return p->normal_prio;
1981}
95e02ca9 1982# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1983#endif
1984
d95f4122 1985extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
1986extern void set_user_nice(struct task_struct *p, long nice);
1987extern int task_prio(const struct task_struct *p);
1988extern int task_nice(const struct task_struct *p);
1989extern int can_nice(const struct task_struct *p, const int nice);
1990extern int task_curr(const struct task_struct *p);
1da177e4 1991extern int idle_cpu(int cpu);
fe7de49f
KM
1992extern int sched_setscheduler(struct task_struct *, int,
1993 const struct sched_param *);
961ccddd 1994extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 1995 const struct sched_param *);
36c8b586
IM
1996extern struct task_struct *idle_task(int cpu);
1997extern struct task_struct *curr_task(int cpu);
1998extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1999
2000void yield(void);
2001
2002/*
2003 * The default (Linux) execution domain.
2004 */
2005extern struct exec_domain default_exec_domain;
2006
2007union thread_union {
2008 struct thread_info thread_info;
2009 unsigned long stack[THREAD_SIZE/sizeof(long)];
2010};
2011
2012#ifndef __HAVE_ARCH_KSTACK_END
2013static inline int kstack_end(void *addr)
2014{
2015 /* Reliable end of stack detection:
2016 * Some APM bios versions misalign the stack
2017 */
2018 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2019}
2020#endif
2021
2022extern union thread_union init_thread_union;
2023extern struct task_struct init_task;
2024
2025extern struct mm_struct init_mm;
2026
198fe21b
PE
2027extern struct pid_namespace init_pid_ns;
2028
2029/*
2030 * find a task by one of its numerical ids
2031 *
198fe21b
PE
2032 * find_task_by_pid_ns():
2033 * finds a task by its pid in the specified namespace
228ebcbe
PE
2034 * find_task_by_vpid():
2035 * finds a task by its virtual pid
198fe21b 2036 *
e49859e7 2037 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2038 */
2039
228ebcbe
PE
2040extern struct task_struct *find_task_by_vpid(pid_t nr);
2041extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2042 struct pid_namespace *ns);
198fe21b 2043
8520d7c7 2044extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2045
2046/* per-UID process charging. */
acce292c 2047extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
2048static inline struct user_struct *get_uid(struct user_struct *u)
2049{
2050 atomic_inc(&u->__count);
2051 return u;
2052}
2053extern void free_uid(struct user_struct *);
28f300d2 2054extern void release_uids(struct user_namespace *ns);
1da177e4
LT
2055
2056#include <asm/current.h>
2057
f0af911a 2058extern void xtime_update(unsigned long ticks);
1da177e4 2059
b3c97528
HH
2060extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2061extern int wake_up_process(struct task_struct *tsk);
2062extern void wake_up_new_task(struct task_struct *tsk,
2063 unsigned long clone_flags);
1da177e4
LT
2064#ifdef CONFIG_SMP
2065 extern void kick_process(struct task_struct *tsk);
2066#else
2067 static inline void kick_process(struct task_struct *tsk) { }
2068#endif
ad46c2c4
IM
2069extern void sched_fork(struct task_struct *p, int clone_flags);
2070extern void sched_dead(struct task_struct *p);
1da177e4 2071
1da177e4
LT
2072extern void proc_caches_init(void);
2073extern void flush_signals(struct task_struct *);
3bcac026 2074extern void __flush_signals(struct task_struct *);
10ab825b 2075extern void ignore_signals(struct task_struct *);
1da177e4
LT
2076extern void flush_signal_handlers(struct task_struct *, int force_default);
2077extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2078
2079static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2080{
2081 unsigned long flags;
2082 int ret;
2083
2084 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2085 ret = dequeue_signal(tsk, mask, info);
2086 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2087
2088 return ret;
2089}
2090
2091extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2092 sigset_t *mask);
2093extern void unblock_all_signals(void);
2094extern void release_task(struct task_struct * p);
2095extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2096extern int force_sigsegv(int, struct task_struct *);
2097extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2098extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2099extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2100extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2101extern int kill_pgrp(struct pid *pid, int sig, int priv);
2102extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2103extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2104extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2105extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2106extern void force_sig(int, struct task_struct *);
1da177e4 2107extern int send_sig(int, struct task_struct *, int);
09faef11 2108extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2109extern struct sigqueue *sigqueue_alloc(void);
2110extern void sigqueue_free(struct sigqueue *);
ac5c2153 2111extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2112extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2113extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2114
9ec52099
CLG
2115static inline int kill_cad_pid(int sig, int priv)
2116{
2117 return kill_pid(cad_pid, sig, priv);
2118}
2119
1da177e4
LT
2120/* These can be the second arg to send_sig_info/send_group_sig_info. */
2121#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2122#define SEND_SIG_PRIV ((struct siginfo *) 1)
2123#define SEND_SIG_FORCED ((struct siginfo *) 2)
2124
2a855dd0
SAS
2125/*
2126 * True if we are on the alternate signal stack.
2127 */
1da177e4
LT
2128static inline int on_sig_stack(unsigned long sp)
2129{
2a855dd0
SAS
2130#ifdef CONFIG_STACK_GROWSUP
2131 return sp >= current->sas_ss_sp &&
2132 sp - current->sas_ss_sp < current->sas_ss_size;
2133#else
2134 return sp > current->sas_ss_sp &&
2135 sp - current->sas_ss_sp <= current->sas_ss_size;
2136#endif
1da177e4
LT
2137}
2138
2139static inline int sas_ss_flags(unsigned long sp)
2140{
2141 return (current->sas_ss_size == 0 ? SS_DISABLE
2142 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2143}
2144
1da177e4
LT
2145/*
2146 * Routines for handling mm_structs
2147 */
2148extern struct mm_struct * mm_alloc(void);
2149
2150/* mmdrop drops the mm and the page tables */
b3c97528 2151extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2152static inline void mmdrop(struct mm_struct * mm)
2153{
6fb43d7b 2154 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2155 __mmdrop(mm);
2156}
2157
2158/* mmput gets rid of the mappings and all user-space */
2159extern void mmput(struct mm_struct *);
2160/* Grab a reference to a task's mm, if it is not already going away */
2161extern struct mm_struct *get_task_mm(struct task_struct *task);
2162/* Remove the current tasks stale references to the old mm_struct */
2163extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2164/* Allocate a new mm structure and copy contents from tsk->mm */
2165extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2166
6f2c55b8
AD
2167extern int copy_thread(unsigned long, unsigned long, unsigned long,
2168 struct task_struct *, struct pt_regs *);
1da177e4
LT
2169extern void flush_thread(void);
2170extern void exit_thread(void);
2171
1da177e4 2172extern void exit_files(struct task_struct *);
a7e5328a 2173extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2174
1da177e4 2175extern void exit_itimers(struct signal_struct *);
cbaffba1 2176extern void flush_itimer_signals(void);
1da177e4
LT
2177
2178extern NORET_TYPE void do_group_exit(int);
2179
1da177e4
LT
2180extern void daemonize(const char *, ...);
2181extern int allow_signal(int);
2182extern int disallow_signal(int);
1da177e4 2183
d7627467
DH
2184extern int do_execve(const char *,
2185 const char __user * const __user *,
2186 const char __user * const __user *, struct pt_regs *);
1da177e4 2187extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2188struct task_struct *fork_idle(int);
1da177e4
LT
2189
2190extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2191extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2192
2193#ifdef CONFIG_SMP
184748cc 2194static inline void scheduler_ipi(void) { }
85ba2d86 2195extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2196#else
184748cc 2197static inline void scheduler_ipi(void) { }
85ba2d86
RM
2198static inline unsigned long wait_task_inactive(struct task_struct *p,
2199 long match_state)
2200{
2201 return 1;
2202}
1da177e4
LT
2203#endif
2204
05725f7e
JP
2205#define next_task(p) \
2206 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2207
2208#define for_each_process(p) \
2209 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2210
5bb459bb 2211extern bool current_is_single_threaded(void);
d84f4f99 2212
1da177e4
LT
2213/*
2214 * Careful: do_each_thread/while_each_thread is a double loop so
2215 * 'break' will not work as expected - use goto instead.
2216 */
2217#define do_each_thread(g, t) \
2218 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2219
2220#define while_each_thread(g, t) \
2221 while ((t = next_thread(t)) != g)
2222
7e49827c
ON
2223static inline int get_nr_threads(struct task_struct *tsk)
2224{
b3ac022c 2225 return tsk->signal->nr_threads;
7e49827c
ON
2226}
2227
de12a787
EB
2228/* de_thread depends on thread_group_leader not being a pid based check */
2229#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2230
0804ef4b
EB
2231/* Do to the insanities of de_thread it is possible for a process
2232 * to have the pid of the thread group leader without actually being
2233 * the thread group leader. For iteration through the pids in proc
2234 * all we care about is that we have a task with the appropriate
2235 * pid, we don't actually care if we have the right task.
2236 */
e868171a 2237static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2238{
2239 return p->pid == p->tgid;
2240}
2241
bac0abd6
PE
2242static inline
2243int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2244{
2245 return p1->tgid == p2->tgid;
2246}
2247
36c8b586 2248static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2249{
05725f7e
JP
2250 return list_entry_rcu(p->thread_group.next,
2251 struct task_struct, thread_group);
47e65328
ON
2252}
2253
e868171a 2254static inline int thread_group_empty(struct task_struct *p)
1da177e4 2255{
47e65328 2256 return list_empty(&p->thread_group);
1da177e4
LT
2257}
2258
2259#define delay_group_leader(p) \
2260 (thread_group_leader(p) && !thread_group_empty(p))
2261
39c626ae
ON
2262static inline int task_detached(struct task_struct *p)
2263{
2264 return p->exit_signal == -1;
2265}
2266
1da177e4 2267/*
260ea101 2268 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2269 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2270 * pins the final release of task.io_context. Also protects ->cpuset and
2271 * ->cgroup.subsys[].
1da177e4
LT
2272 *
2273 * Nests both inside and outside of read_lock(&tasklist_lock).
2274 * It must not be nested with write_lock_irq(&tasklist_lock),
2275 * neither inside nor outside.
2276 */
2277static inline void task_lock(struct task_struct *p)
2278{
2279 spin_lock(&p->alloc_lock);
2280}
2281
2282static inline void task_unlock(struct task_struct *p)
2283{
2284 spin_unlock(&p->alloc_lock);
2285}
2286
b8ed374e 2287extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2288 unsigned long *flags);
2289
b8ed374e
NK
2290#define lock_task_sighand(tsk, flags) \
2291({ struct sighand_struct *__ss; \
2292 __cond_lock(&(tsk)->sighand->siglock, \
2293 (__ss = __lock_task_sighand(tsk, flags))); \
2294 __ss; \
2295}) \
2296
f63ee72e
ON
2297static inline void unlock_task_sighand(struct task_struct *tsk,
2298 unsigned long *flags)
2299{
2300 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2301}
2302
f037360f
AV
2303#ifndef __HAVE_THREAD_FUNCTIONS
2304
f7e4217b
RZ
2305#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2306#define task_stack_page(task) ((task)->stack)
a1261f54 2307
10ebffde
AV
2308static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2309{
2310 *task_thread_info(p) = *task_thread_info(org);
2311 task_thread_info(p)->task = p;
2312}
2313
2314static inline unsigned long *end_of_stack(struct task_struct *p)
2315{
f7e4217b 2316 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2317}
2318
f037360f
AV
2319#endif
2320
8b05c7e6
FT
2321static inline int object_is_on_stack(void *obj)
2322{
2323 void *stack = task_stack_page(current);
2324
2325 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2326}
2327
8c9843e5
BH
2328extern void thread_info_cache_init(void);
2329
7c9f8861
ES
2330#ifdef CONFIG_DEBUG_STACK_USAGE
2331static inline unsigned long stack_not_used(struct task_struct *p)
2332{
2333 unsigned long *n = end_of_stack(p);
2334
2335 do { /* Skip over canary */
2336 n++;
2337 } while (!*n);
2338
2339 return (unsigned long)n - (unsigned long)end_of_stack(p);
2340}
2341#endif
2342
1da177e4
LT
2343/* set thread flags in other task's structures
2344 * - see asm/thread_info.h for TIF_xxxx flags available
2345 */
2346static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2347{
a1261f54 2348 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2349}
2350
2351static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2352{
a1261f54 2353 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2354}
2355
2356static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2357{
a1261f54 2358 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2359}
2360
2361static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2362{
a1261f54 2363 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2364}
2365
2366static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2367{
a1261f54 2368 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2369}
2370
2371static inline void set_tsk_need_resched(struct task_struct *tsk)
2372{
2373 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2374}
2375
2376static inline void clear_tsk_need_resched(struct task_struct *tsk)
2377{
2378 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2379}
2380
8ae121ac
GH
2381static inline int test_tsk_need_resched(struct task_struct *tsk)
2382{
2383 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2384}
2385
690cc3ff
EB
2386static inline int restart_syscall(void)
2387{
2388 set_tsk_thread_flag(current, TIF_SIGPENDING);
2389 return -ERESTARTNOINTR;
2390}
2391
1da177e4
LT
2392static inline int signal_pending(struct task_struct *p)
2393{
2394 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2395}
f776d12d 2396
d9588725
RM
2397static inline int __fatal_signal_pending(struct task_struct *p)
2398{
2399 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2400}
f776d12d
MW
2401
2402static inline int fatal_signal_pending(struct task_struct *p)
2403{
2404 return signal_pending(p) && __fatal_signal_pending(p);
2405}
2406
16882c1e
ON
2407static inline int signal_pending_state(long state, struct task_struct *p)
2408{
2409 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2410 return 0;
2411 if (!signal_pending(p))
2412 return 0;
2413
16882c1e
ON
2414 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2415}
2416
1da177e4
LT
2417static inline int need_resched(void)
2418{
9404ef02 2419 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2420}
2421
2422/*
2423 * cond_resched() and cond_resched_lock(): latency reduction via
2424 * explicit rescheduling in places that are safe. The return
2425 * value indicates whether a reschedule was done in fact.
2426 * cond_resched_lock() will drop the spinlock before scheduling,
2427 * cond_resched_softirq() will enable bhs before scheduling.
2428 */
c3921ab7 2429extern int _cond_resched(void);
6f80bd98 2430
613afbf8
FW
2431#define cond_resched() ({ \
2432 __might_sleep(__FILE__, __LINE__, 0); \
2433 _cond_resched(); \
2434})
6f80bd98 2435
613afbf8
FW
2436extern int __cond_resched_lock(spinlock_t *lock);
2437
716a4234
FW
2438#ifdef CONFIG_PREEMPT
2439#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2440#else
716a4234 2441#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2442#endif
716a4234 2443
613afbf8 2444#define cond_resched_lock(lock) ({ \
716a4234 2445 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2446 __cond_resched_lock(lock); \
2447})
2448
2449extern int __cond_resched_softirq(void);
2450
75e1056f
VP
2451#define cond_resched_softirq() ({ \
2452 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2453 __cond_resched_softirq(); \
613afbf8 2454})
1da177e4
LT
2455
2456/*
2457 * Does a critical section need to be broken due to another
95c354fe
NP
2458 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2459 * but a general need for low latency)
1da177e4 2460 */
95c354fe 2461static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2462{
95c354fe
NP
2463#ifdef CONFIG_PREEMPT
2464 return spin_is_contended(lock);
2465#else
1da177e4 2466 return 0;
95c354fe 2467#endif
1da177e4
LT
2468}
2469
f06febc9
FM
2470/*
2471 * Thread group CPU time accounting.
2472 */
4cd4c1b4 2473void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2474void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2475
490dea45 2476static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2477{
4cd4c1b4 2478 spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2479}
2480
7bb44ade
RM
2481/*
2482 * Reevaluate whether the task has signals pending delivery.
2483 * Wake the task if so.
2484 * This is required every time the blocked sigset_t changes.
2485 * callers must hold sighand->siglock.
2486 */
2487extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2488extern void recalc_sigpending(void);
2489
2490extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2491
2492/*
2493 * Wrappers for p->thread_info->cpu access. No-op on UP.
2494 */
2495#ifdef CONFIG_SMP
2496
2497static inline unsigned int task_cpu(const struct task_struct *p)
2498{
a1261f54 2499 return task_thread_info(p)->cpu;
1da177e4
LT
2500}
2501
c65cc870 2502extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2503
2504#else
2505
2506static inline unsigned int task_cpu(const struct task_struct *p)
2507{
2508 return 0;
2509}
2510
2511static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2512{
2513}
2514
2515#endif /* CONFIG_SMP */
2516
96f874e2
RR
2517extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2518extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2519
1da177e4
LT
2520extern void normalize_rt_tasks(void);
2521
7c941438 2522#ifdef CONFIG_CGROUP_SCHED
9b5b7751 2523
07e06b01 2524extern struct task_group root_task_group;
9b5b7751 2525
ec7dc8ac 2526extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2527extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2528extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2529#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2530extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2531extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2532#endif
2533#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2534extern int sched_group_set_rt_runtime(struct task_group *tg,
2535 long rt_runtime_us);
2536extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2537extern int sched_group_set_rt_period(struct task_group *tg,
2538 long rt_period_us);
2539extern long sched_group_rt_period(struct task_group *tg);
54e99124 2540extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2541#endif
9b5b7751
SV
2542#endif
2543
54e99124
DG
2544extern int task_can_switch_user(struct user_struct *up,
2545 struct task_struct *tsk);
2546
4b98d11b
AD
2547#ifdef CONFIG_TASK_XACCT
2548static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2549{
940389b8 2550 tsk->ioac.rchar += amt;
4b98d11b
AD
2551}
2552
2553static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2554{
940389b8 2555 tsk->ioac.wchar += amt;
4b98d11b
AD
2556}
2557
2558static inline void inc_syscr(struct task_struct *tsk)
2559{
940389b8 2560 tsk->ioac.syscr++;
4b98d11b
AD
2561}
2562
2563static inline void inc_syscw(struct task_struct *tsk)
2564{
940389b8 2565 tsk->ioac.syscw++;
4b98d11b
AD
2566}
2567#else
2568static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2569{
2570}
2571
2572static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2573{
2574}
2575
2576static inline void inc_syscr(struct task_struct *tsk)
2577{
2578}
2579
2580static inline void inc_syscw(struct task_struct *tsk)
2581{
2582}
2583#endif
2584
82455257
DH
2585#ifndef TASK_SIZE_OF
2586#define TASK_SIZE_OF(tsk) TASK_SIZE
2587#endif
2588
cf475ad2
BS
2589#ifdef CONFIG_MM_OWNER
2590extern void mm_update_next_owner(struct mm_struct *mm);
2591extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2592#else
2593static inline void mm_update_next_owner(struct mm_struct *mm)
2594{
2595}
2596
2597static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2598{
2599}
2600#endif /* CONFIG_MM_OWNER */
2601
3e10e716
JS
2602static inline unsigned long task_rlimit(const struct task_struct *tsk,
2603 unsigned int limit)
2604{
2605 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2606}
2607
2608static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2609 unsigned int limit)
2610{
2611 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2612}
2613
2614static inline unsigned long rlimit(unsigned int limit)
2615{
2616 return task_rlimit(current, limit);
2617}
2618
2619static inline unsigned long rlimit_max(unsigned int limit)
2620{
2621 return task_rlimit_max(current, limit);
2622}
2623
1da177e4
LT
2624#endif /* __KERNEL__ */
2625
2626#endif