resource: move kernel function inside __KERNEL__
[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 */
24#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
071df104 25#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 26#define CLONE_NEWIPC 0x08000000 /* New ipcs */
77ec739d 27#define CLONE_NEWUSER 0x10000000 /* New user namespace */
30e49c26 28#define CLONE_NEWPID 0x20000000 /* New pid namespace */
169e3674 29#define CLONE_NEWNET 0x40000000 /* New network namespace */
fadad878 30#define CLONE_IO 0x80000000 /* Clone io context */
b7b3c76a
DW
31
32/*
33 * Scheduling policies
34 */
35#define SCHED_NORMAL 0
36#define SCHED_FIFO 1
37#define SCHED_RR 2
38#define SCHED_BATCH 3
0e6aca43
IM
39/* SCHED_ISO: reserved but not implemented yet */
40#define SCHED_IDLE 5
ca94c442
LP
41/* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
42#define SCHED_RESET_ON_FORK 0x40000000
b7b3c76a 43
a3b6714e 44#ifdef __KERNEL__
b7b3c76a
DW
45
46struct sched_param {
47 int sched_priority;
48};
49
1da177e4
LT
50#include <asm/param.h> /* for HZ */
51
1da177e4
LT
52#include <linux/capability.h>
53#include <linux/threads.h>
54#include <linux/kernel.h>
55#include <linux/types.h>
56#include <linux/timex.h>
57#include <linux/jiffies.h>
58#include <linux/rbtree.h>
59#include <linux/thread_info.h>
60#include <linux/cpumask.h>
61#include <linux/errno.h>
62#include <linux/nodemask.h>
c92ff1bd 63#include <linux/mm_types.h>
1da177e4
LT
64
65#include <asm/system.h>
1da177e4
LT
66#include <asm/page.h>
67#include <asm/ptrace.h>
1da177e4
LT
68#include <asm/cputime.h>
69
70#include <linux/smp.h>
71#include <linux/sem.h>
72#include <linux/signal.h>
5ad4e53b 73#include <linux/path.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>
5cb350ba 91#include <linux/kobject.h>
9745512c 92#include <linux/latencytop.h>
9e2b2dc4 93#include <linux/cred.h>
a3b6714e
DW
94
95#include <asm/processor.h>
36d57ac4 96
1da177e4 97struct exec_domain;
c87e2837 98struct futex_pi_state;
286100a6 99struct robust_list_head;
d89d8796 100struct bio;
5ad4e53b 101struct fs_struct;
e2b371f0 102struct bts_context;
cdd6c482 103struct perf_event_context;
1da177e4 104
1da177e4
LT
105/*
106 * List of flags we want to share for kernel threads,
107 * if only because they are not used by them anyway.
108 */
109#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
110
111/*
112 * These are the constant used to fake the fixed-point load-average
113 * counting. Some notes:
114 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
115 * a load-average precision of 10 bits integer + 11 bits fractional
116 * - if you want to count load-averages more often, you need more
117 * precision, or rounding will get you. With 2-second counting freq,
118 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 * 11 bit fractions.
120 */
121extern unsigned long avenrun[]; /* Load averages */
2d02494f 122extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
123
124#define FSHIFT 11 /* nr of bits of precision */
125#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 126#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
127#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
128#define EXP_5 2014 /* 1/exp(5sec/5min) */
129#define EXP_15 2037 /* 1/exp(5sec/15min) */
130
131#define CALC_LOAD(load,exp,n) \
132 load *= exp; \
133 load += n*(FIXED_1-exp); \
134 load >>= FSHIFT;
135
136extern unsigned long total_forks;
137extern int nr_threads;
1da177e4
LT
138DECLARE_PER_CPU(unsigned long, process_counts);
139extern int nr_processes(void);
140extern unsigned long nr_running(void);
141extern unsigned long nr_uninterruptible(void);
142extern unsigned long nr_iowait(void);
69d25870
AV
143extern unsigned long nr_iowait_cpu(void);
144extern unsigned long this_cpu_load(void);
145
146
dce48a84 147extern void calc_global_load(void);
1da177e4 148
7e49fcce
SR
149extern unsigned long get_parent_ip(unsigned long addr);
150
43ae34cb
IM
151struct seq_file;
152struct cfs_rq;
4cf86d77 153struct task_group;
43ae34cb
IM
154#ifdef CONFIG_SCHED_DEBUG
155extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
156extern void proc_sched_set_task(struct task_struct *p);
157extern void
5cef9eca 158print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
159#else
160static inline void
161proc_sched_show_task(struct task_struct *p, struct seq_file *m)
162{
163}
164static inline void proc_sched_set_task(struct task_struct *p)
165{
166}
167static inline void
5cef9eca 168print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
169{
170}
171#endif
1da177e4 172
4a8342d2
LT
173/*
174 * Task state bitmask. NOTE! These bits are also
175 * encoded in fs/proc/array.c: get_task_state().
176 *
177 * We have two separate sets of flags: task->state
178 * is about runnability, while task->exit_state are
179 * about the task exiting. Confusing, but this way
180 * modifying one set can't modify the other one by
181 * mistake.
182 */
1da177e4
LT
183#define TASK_RUNNING 0
184#define TASK_INTERRUPTIBLE 1
185#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
186#define __TASK_STOPPED 4
187#define __TASK_TRACED 8
4a8342d2
LT
188/* in tsk->exit_state */
189#define EXIT_ZOMBIE 16
190#define EXIT_DEAD 32
191/* in tsk->state again */
af927232 192#define TASK_DEAD 64
f021a3c2 193#define TASK_WAKEKILL 128
e9c84311 194#define TASK_WAKING 256
e1781538 195#define TASK_STATE_MAX 512
f021a3c2 196
44d90df6 197#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
73342151 198
e1781538
PZ
199extern char ___assert_task_state[1 - 2*!!(
200 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
201
202/* Convenience macros for the sake of set_task_state */
203#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
204#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
205#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 206
92a1f4bc
MW
207/* Convenience macros for the sake of wake_up */
208#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 209#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
210
211/* get_task_state() */
212#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
213 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
214 __TASK_TRACED)
92a1f4bc 215
f021a3c2
MW
216#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
217#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 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
LT
260
261extern void sched_init(void);
262extern void sched_init_smp(void);
2d07b255 263extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 264extern void init_idle(struct task_struct *idle, int cpu);
1df21055 265extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 266
89f19f04 267extern int runqueue_is_locked(int cpu);
ad474cac 268extern void task_rq_unlock_wait(struct task_struct *p);
017730c1 269
6a7b3dc3 270extern cpumask_var_t nohz_cpu_mask;
46cb4b7c
SS
271#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
272extern int select_nohz_load_balancer(int cpu);
eea08f32 273extern int get_nohz_load_balancer(void);
46cb4b7c
SS
274#else
275static inline int select_nohz_load_balancer(int cpu)
276{
277 return 0;
278}
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
8446f1d3 310#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 311extern void softlockup_tick(void);
8446f1d3 312extern void touch_softlockup_watchdog(void);
04c9167f 313extern void touch_all_softlockup_watchdogs(void);
baf48f65 314extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
8d65af78 315 void __user *buffer,
baf48f65 316 size_t *lenp, loff_t *ppos);
9c44bc03 317extern unsigned int softlockup_panic;
9383d967 318extern int softlockup_thresh;
8446f1d3 319#else
6687a97d 320static inline void softlockup_tick(void)
8446f1d3
IM
321{
322}
8446f1d3
IM
323static inline void touch_softlockup_watchdog(void)
324{
325}
04c9167f
JF
326static inline void touch_all_softlockup_watchdogs(void)
327{
328}
8446f1d3
IM
329#endif
330
e162b39a
MSB
331#ifdef CONFIG_DETECT_HUNG_TASK
332extern unsigned int sysctl_hung_task_panic;
333extern unsigned long sysctl_hung_task_check_count;
334extern unsigned long sysctl_hung_task_timeout_secs;
335extern unsigned long sysctl_hung_task_warnings;
336extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 337 void __user *buffer,
e162b39a
MSB
338 size_t *lenp, loff_t *ppos);
339#endif
8446f1d3 340
1da177e4
LT
341/* Attach to any functions which should be ignored in wchan output. */
342#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
343
344/* Linker adds these: start and end of __sched functions */
345extern char __sched_text_start[], __sched_text_end[];
346
1da177e4
LT
347/* Is this address in the __sched functions? */
348extern int in_sched_functions(unsigned long addr);
349
350#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 351extern signed long schedule_timeout(signed long timeout);
64ed93a2 352extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 353extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 354extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 355asmlinkage void schedule(void);
0d66bf6d 356extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 357
ab516013 358struct nsproxy;
acce292c 359struct user_namespace;
1da177e4 360
341c87bf
KH
361/*
362 * Default maximum number of active map areas, this limits the number of vmas
363 * per mm struct. Users can overwrite this number by sysctl but there is a
364 * problem.
365 *
366 * When a program's coredump is generated as ELF format, a section is created
367 * per a vma. In ELF, the number of sections is represented in unsigned short.
368 * This means the number of sections should be smaller than 65535 at coredump.
369 * Because the kernel adds some informative sections to a image of program at
370 * generating coredump, we need some margin. The number of extra sections is
371 * 1-3 now and depends on arch. We use "5" as safe margin, here.
372 */
373#define MAPCOUNT_ELF_CORE_MARGIN (5)
374#define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
375
376extern int sysctl_max_map_count;
377
378#include <linux/aio.h>
379
380extern unsigned long
381arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
382 unsigned long, unsigned long);
383extern unsigned long
384arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
385 unsigned long len, unsigned long pgoff,
386 unsigned long flags);
1363c3cd
WW
387extern void arch_unmap_area(struct mm_struct *, unsigned long);
388extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 389
f7d0b926 390#if USE_SPLIT_PTLOCKS
f412ac08
HD
391/*
392 * The mm counters are not protected by its page_table_lock,
393 * so must be incremented atomically.
394 */
d3cb4871
CL
395#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
396#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
397#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
398#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
399#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
f412ac08 400
f7d0b926 401#else /* !USE_SPLIT_PTLOCKS */
f412ac08
HD
402/*
403 * The mm counters are protected by its page_table_lock,
404 * so can be incremented directly.
405 */
1da177e4
LT
406#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
407#define get_mm_counter(mm, member) ((mm)->_##member)
408#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
409#define inc_mm_counter(mm, member) (mm)->_##member++
410#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08 411
f7d0b926 412#endif /* !USE_SPLIT_PTLOCKS */
4294621f 413
f412ac08
HD
414#define get_mm_rss(mm) \
415 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
416#define update_hiwater_rss(mm) do { \
417 unsigned long _rss = get_mm_rss(mm); \
418 if ((mm)->hiwater_rss < _rss) \
419 (mm)->hiwater_rss = _rss; \
420} while (0)
421#define update_hiwater_vm(mm) do { \
422 if ((mm)->hiwater_vm < (mm)->total_vm) \
423 (mm)->hiwater_vm = (mm)->total_vm; \
424} while (0)
425
9de1581e
ON
426static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
427{
428 return max(mm->hiwater_rss, get_mm_rss(mm));
429}
430
1f10206c
JP
431static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
432 struct mm_struct *mm)
433{
434 unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
435
436 if (*maxrss < hiwater_rss)
437 *maxrss = hiwater_rss;
438}
439
9de1581e
ON
440static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
441{
442 return max(mm->hiwater_vm, mm->total_vm);
443}
901608d9 444
6c5d5238
KH
445extern void set_dumpable(struct mm_struct *mm, int value);
446extern int get_dumpable(struct mm_struct *mm);
447
448/* mm flags */
3cb4a0bb 449/* dumpable bits */
6c5d5238
KH
450#define MMF_DUMPABLE 0 /* core dump is permitted */
451#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 452
3cb4a0bb 453#define MMF_DUMPABLE_BITS 2
f8af4da3 454#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH
455
456/* coredump filter bits */
457#define MMF_DUMP_ANON_PRIVATE 2
458#define MMF_DUMP_ANON_SHARED 3
459#define MMF_DUMP_MAPPED_PRIVATE 4
460#define MMF_DUMP_MAPPED_SHARED 5
82df3973 461#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
462#define MMF_DUMP_HUGETLB_PRIVATE 7
463#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 464
3cb4a0bb 465#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 466#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
467#define MMF_DUMP_FILTER_MASK \
468 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
469#define MMF_DUMP_FILTER_DEFAULT \
e575f111 470 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
471 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
472
473#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
474# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
475#else
476# define MMF_DUMP_MASK_DEFAULT_ELF 0
477#endif
f8af4da3
HD
478 /* leave room for more dump flags */
479#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
480
481#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 482
1da177e4
LT
483struct sighand_struct {
484 atomic_t count;
485 struct k_sigaction action[_NSIG];
486 spinlock_t siglock;
b8fceee1 487 wait_queue_head_t signalfd_wqh;
1da177e4
LT
488};
489
0e464814 490struct pacct_struct {
f6ec29a4
KK
491 int ac_flag;
492 long ac_exitcode;
0e464814 493 unsigned long ac_mem;
77787bfb
KK
494 cputime_t ac_utime, ac_stime;
495 unsigned long ac_minflt, ac_majflt;
0e464814
KK
496};
497
42c4ab41
SG
498struct cpu_itimer {
499 cputime_t expires;
500 cputime_t incr;
8356b5f9
SG
501 u32 error;
502 u32 incr_error;
42c4ab41
SG
503};
504
f06febc9
FM
505/**
506 * struct task_cputime - collected CPU time counts
507 * @utime: time spent in user mode, in &cputime_t units
508 * @stime: time spent in kernel mode, in &cputime_t units
509 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 510 *
f06febc9
FM
511 * This structure groups together three kinds of CPU time that are
512 * tracked for threads and thread groups. Most things considering
513 * CPU time want to group these counts together and treat all three
514 * of them in parallel.
515 */
516struct task_cputime {
517 cputime_t utime;
518 cputime_t stime;
519 unsigned long long sum_exec_runtime;
520};
521/* Alternate field names when used to cache expirations. */
522#define prof_exp stime
523#define virt_exp utime
524#define sched_exp sum_exec_runtime
525
4cd4c1b4
PZ
526#define INIT_CPUTIME \
527 (struct task_cputime) { \
528 .utime = cputime_zero, \
529 .stime = cputime_zero, \
530 .sum_exec_runtime = 0, \
531 }
532
c99e6efe
PZ
533/*
534 * Disable preemption until the scheduler is running.
535 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
536 *
537 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
538 * before the scheduler is active -- see should_resched().
c99e6efe 539 */
d86ee480 540#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 541
f06febc9 542/**
4cd4c1b4
PZ
543 * struct thread_group_cputimer - thread group interval timer counts
544 * @cputime: thread group interval timers.
545 * @running: non-zero when there are timers running and
546 * @cputime receives updates.
547 * @lock: lock for fields in this struct.
f06febc9
FM
548 *
549 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 550 * used for thread group CPU timer calculations.
f06febc9 551 */
4cd4c1b4
PZ
552struct thread_group_cputimer {
553 struct task_cputime cputime;
554 int running;
555 spinlock_t lock;
f06febc9 556};
f06febc9 557
1da177e4
LT
558/*
559 * NOTE! "signal_struct" does not have it's own
560 * locking, because a shared signal_struct always
561 * implies a shared sighand_struct, so locking
562 * sighand_struct is always a proper superset of
563 * the locking of signal_struct.
564 */
565struct signal_struct {
566 atomic_t count;
567 atomic_t live;
568
569 wait_queue_head_t wait_chldexit; /* for wait4() */
570
571 /* current thread group signal load-balancing target: */
36c8b586 572 struct task_struct *curr_target;
1da177e4
LT
573
574 /* shared signal handling: */
575 struct sigpending shared_pending;
576
577 /* thread group exit support */
578 int group_exit_code;
579 /* overloaded:
580 * - notify group_exit_task when ->count is equal to notify_count
581 * - everyone except group_exit_task is stopped during signal delivery
582 * of fatal signals, group_exit_task processes the signal.
583 */
1da177e4 584 int notify_count;
07dd20e0 585 struct task_struct *group_exit_task;
1da177e4
LT
586
587 /* thread group stop support, overloads group_exit_code too */
588 int group_stop_count;
589 unsigned int flags; /* see SIGNAL_* flags below */
590
591 /* POSIX.1b Interval Timers */
592 struct list_head posix_timers;
593
594 /* ITIMER_REAL timer for the process */
2ff678b8 595 struct hrtimer real_timer;
fea9d175 596 struct pid *leader_pid;
2ff678b8 597 ktime_t it_real_incr;
1da177e4 598
42c4ab41
SG
599 /*
600 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
601 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
602 * values are defined to 0 and 1 respectively
603 */
604 struct cpu_itimer it[2];
1da177e4 605
f06febc9 606 /*
4cd4c1b4
PZ
607 * Thread group totals for process CPU timers.
608 * See thread_group_cputimer(), et al, for details.
f06febc9 609 */
4cd4c1b4 610 struct thread_group_cputimer cputimer;
f06febc9
FM
611
612 /* Earliest-expiration cache. */
613 struct task_cputime cputime_expires;
614
615 struct list_head cpu_timers[3];
616
ab521dc0 617 struct pid *tty_old_pgrp;
1ec320af 618
1da177e4
LT
619 /* boolean value for session group leader */
620 int leader;
621
622 struct tty_struct *tty; /* NULL if no tty */
623
624 /*
625 * Cumulative resource counters for dead threads in the group,
626 * and for reaped dead child processes forked by this group.
627 * Live threads maintain their own counters and add to these
628 * in __exit_signal, except for the group leader.
629 */
32bd671d 630 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
631 cputime_t gtime;
632 cputime_t cgtime;
0cf55e1e
HS
633#ifndef CONFIG_VIRT_CPU_ACCOUNTING
634 cputime_t prev_utime, prev_stime;
635#endif
1da177e4
LT
636 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
637 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 638 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 639 unsigned long maxrss, cmaxrss;
940389b8 640 struct task_io_accounting ioac;
1da177e4 641
32bd671d
PZ
642 /*
643 * Cumulative ns of schedule CPU time fo dead threads in the
644 * group, not including a zombie group leader, (This only differs
645 * from jiffies_to_ns(utime + stime) if sched_clock uses something
646 * other than jiffies.)
647 */
648 unsigned long long sum_sched_runtime;
649
1da177e4
LT
650 /*
651 * We don't bother to synchronize most readers of this at all,
652 * because there is no reader checking a limit that actually needs
653 * to get both rlim_cur and rlim_max atomically, and either one
654 * alone is a single word that can safely be read normally.
655 * getrlimit/setrlimit use task_lock(current->group_leader) to
656 * protect this instead of the siglock, because they really
657 * have no need to disable irqs.
658 */
659 struct rlimit rlim[RLIM_NLIMITS];
660
0e464814
KK
661#ifdef CONFIG_BSD_PROCESS_ACCT
662 struct pacct_struct pacct; /* per-process accounting information */
663#endif
ad4ecbcb 664#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
665 struct taskstats *stats;
666#endif
522ed776
MT
667#ifdef CONFIG_AUDIT
668 unsigned audit_tty;
669 struct tty_audit_buf *tty_audit_buf;
670#endif
28b83c51
KM
671
672 int oom_adj; /* OOM kill score adjustment (bit shift) */
1da177e4
LT
673};
674
4866cde0
NP
675/* Context switch must be unlocked if interrupts are to be enabled */
676#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
677# define __ARCH_WANT_UNLOCKED_CTXSW
678#endif
679
1da177e4
LT
680/*
681 * Bits in flags field of signal_struct.
682 */
683#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
684#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
685#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
686#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
687/*
688 * Pending notifications to parent.
689 */
690#define SIGNAL_CLD_STOPPED 0x00000010
691#define SIGNAL_CLD_CONTINUED 0x00000020
692#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 693
fae5fa44
ON
694#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
695
ed5d2cac
ON
696/* If true, all threads except ->group_exit_task have pending SIGKILL */
697static inline int signal_group_exit(const struct signal_struct *sig)
698{
699 return (sig->flags & SIGNAL_GROUP_EXIT) ||
700 (sig->group_exit_task != NULL);
701}
702
1da177e4
LT
703/*
704 * Some day this will be a full-fledged user tracking system..
705 */
706struct user_struct {
707 atomic_t __count; /* reference count */
708 atomic_t processes; /* How many processes does this user have? */
709 atomic_t files; /* How many open files does this user have? */
710 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 711#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
712 atomic_t inotify_watches; /* How many inotify watches does this user have? */
713 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
714#endif
7ef9964e 715#ifdef CONFIG_EPOLL
7ef9964e
DL
716 atomic_t epoll_watches; /* The number of file descriptors currently watched */
717#endif
970a8645 718#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
719 /* protected by mq_lock */
720 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 721#endif
1da177e4
LT
722 unsigned long locked_shm; /* How many pages of mlocked shm ? */
723
724#ifdef CONFIG_KEYS
725 struct key *uid_keyring; /* UID specific keyring */
726 struct key *session_keyring; /* UID's default session keyring */
727#endif
728
729 /* Hash table maintenance information */
735de223 730 struct hlist_node uidhash_node;
1da177e4 731 uid_t uid;
18b6e041 732 struct user_namespace *user_ns;
24e377a8 733
052f1dc7 734#ifdef CONFIG_USER_SCHED
4cf86d77 735 struct task_group *tg;
b1a8c172 736#ifdef CONFIG_SYSFS
eb41d946 737 struct kobject kobj;
3959214f 738 struct delayed_work work;
24e377a8 739#endif
b1a8c172 740#endif
789f90fc 741
cdd6c482 742#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
743 atomic_long_t locked_vm;
744#endif
1da177e4
LT
745};
746
eb41d946 747extern int uids_sysfs_init(void);
5cb350ba 748
1da177e4
LT
749extern struct user_struct *find_user(uid_t);
750
751extern struct user_struct root_user;
752#define INIT_USER (&root_user)
753
b6dff3ec 754
1da177e4
LT
755struct backing_dev_info;
756struct reclaim_state;
757
52f17b6c 758#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
759struct sched_info {
760 /* cumulative counters */
2d72376b 761 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 762 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
763
764 /* timestamps */
172ba844
BS
765 unsigned long long last_arrival,/* when we last ran on a cpu */
766 last_queued; /* when we were last queued to run */
b8efb561
IM
767#ifdef CONFIG_SCHEDSTATS
768 /* BKL stats */
480b9434 769 unsigned int bkl_count;
b8efb561 770#endif
1da177e4 771};
52f17b6c 772#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 773
ca74e92b
SN
774#ifdef CONFIG_TASK_DELAY_ACCT
775struct task_delay_info {
776 spinlock_t lock;
777 unsigned int flags; /* Private per-task flags */
778
779 /* For each stat XXX, add following, aligned appropriately
780 *
781 * struct timespec XXX_start, XXX_end;
782 * u64 XXX_delay;
783 * u32 XXX_count;
784 *
785 * Atomicity of updates to XXX_delay, XXX_count protected by
786 * single lock above (split into XXX_lock if contention is an issue).
787 */
0ff92245
SN
788
789 /*
790 * XXX_count is incremented on every XXX operation, the delay
791 * associated with the operation is added to XXX_delay.
792 * XXX_delay contains the accumulated delay time in nanoseconds.
793 */
794 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
795 u64 blkio_delay; /* wait for sync block io completion */
796 u64 swapin_delay; /* wait for swapin block io completion */
797 u32 blkio_count; /* total count of the number of sync block */
798 /* io operations performed */
799 u32 swapin_count; /* total count of the number of swapin block */
800 /* io operations performed */
873b4771
KK
801
802 struct timespec freepages_start, freepages_end;
803 u64 freepages_delay; /* wait for memory reclaim */
804 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 805};
52f17b6c
CS
806#endif /* CONFIG_TASK_DELAY_ACCT */
807
808static inline int sched_info_on(void)
809{
810#ifdef CONFIG_SCHEDSTATS
811 return 1;
812#elif defined(CONFIG_TASK_DELAY_ACCT)
813 extern int delayacct_on;
814 return delayacct_on;
815#else
816 return 0;
ca74e92b 817#endif
52f17b6c 818}
ca74e92b 819
d15bcfdb
IM
820enum cpu_idle_type {
821 CPU_IDLE,
822 CPU_NOT_IDLE,
823 CPU_NEWLY_IDLE,
824 CPU_MAX_IDLE_TYPES
1da177e4
LT
825};
826
827/*
828 * sched-domains (multiprocessor balancing) declarations:
829 */
9aa7b369
IM
830
831/*
832 * Increase resolution of nice-level calculations:
833 */
834#define SCHED_LOAD_SHIFT 10
835#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
836
f8700df7 837#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 838
2dd73a4f 839#ifdef CONFIG_SMP
b5d978e0
PZ
840#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
841#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
842#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
843#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 844#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 845#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 846#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0
PZ
847#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
848#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
849#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
850#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
c88d5910 851
b5d978e0 852#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 853
afb8a9b7
GS
854enum powersavings_balance_level {
855 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
856 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
857 * first for long running threads
858 */
859 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
860 * cpu package for power savings
861 */
862 MAX_POWERSAVINGS_BALANCE_LEVELS
863};
89c4710e 864
716707b2 865extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 866
716707b2
VS
867static inline int sd_balance_for_mc_power(void)
868{
869 if (sched_smt_power_savings)
870 return SD_POWERSAVINGS_BALANCE;
5c45bf27 871
b5d978e0 872 return SD_PREFER_SIBLING;
716707b2 873}
89c4710e 874
716707b2
VS
875static inline int sd_balance_for_package_power(void)
876{
877 if (sched_mc_power_savings | sched_smt_power_savings)
878 return SD_POWERSAVINGS_BALANCE;
879
b5d978e0 880 return SD_PREFER_SIBLING;
716707b2 881}
5c45bf27 882
100fdaee
VS
883/*
884 * Optimise SD flags for power savings:
885 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
886 * Keep default SD flags if sched_{smt,mc}_power_saving=0
887 */
888
889static inline int sd_power_saving_flags(void)
890{
891 if (sched_mc_power_savings | sched_smt_power_savings)
892 return SD_BALANCE_NEWIDLE;
893
894 return 0;
895}
1da177e4
LT
896
897struct sched_group {
898 struct sched_group *next; /* Must be a circular list */
1da177e4
LT
899
900 /*
901 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 902 * single CPU.
5517d86b 903 */
18a3885f 904 unsigned int cpu_power;
6c99e9ad 905
4200efd9
IM
906 /*
907 * The CPUs this group covers.
908 *
909 * NOTE: this field is variable length. (Allocated dynamically
910 * by attaching extra space to the end of the structure,
911 * depending on how many CPUs the kernel has booted up with)
912 *
913 * It is also be embedded into static data structures at build
914 * time. (See 'struct static_sched_group' in kernel/sched.c)
915 */
916 unsigned long cpumask[0];
1da177e4
LT
917};
918
758b2cdc
RR
919static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
920{
6c99e9ad 921 return to_cpumask(sg->cpumask);
758b2cdc
RR
922}
923
1d3504fc
HS
924enum sched_domain_level {
925 SD_LV_NONE = 0,
926 SD_LV_SIBLING,
927 SD_LV_MC,
928 SD_LV_CPU,
929 SD_LV_NODE,
930 SD_LV_ALLNODES,
931 SD_LV_MAX
932};
933
934struct sched_domain_attr {
935 int relax_domain_level;
936};
937
938#define SD_ATTR_INIT (struct sched_domain_attr) { \
939 .relax_domain_level = -1, \
940}
941
1da177e4
LT
942struct sched_domain {
943 /* These fields must be setup */
944 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 945 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 946 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
947 unsigned long min_interval; /* Minimum balance interval ms */
948 unsigned long max_interval; /* Maximum balance interval ms */
949 unsigned int busy_factor; /* less balancing by factor if busy */
950 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 951 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
952 unsigned int busy_idx;
953 unsigned int idle_idx;
954 unsigned int newidle_idx;
955 unsigned int wake_idx;
147cbb4b 956 unsigned int forkexec_idx;
a52bfd73 957 unsigned int smt_gain;
1da177e4 958 int flags; /* See SD_* */
1d3504fc 959 enum sched_domain_level level;
1da177e4
LT
960
961 /* Runtime fields. */
962 unsigned long last_balance; /* init to jiffies. units in jiffies */
963 unsigned int balance_interval; /* initialise to 1. units in ms. */
964 unsigned int nr_balance_failed; /* initialise to 0 */
965
2398f2c6
PZ
966 u64 last_update;
967
1da177e4
LT
968#ifdef CONFIG_SCHEDSTATS
969 /* load_balance() stats */
480b9434
KC
970 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
971 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
972 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
973 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
974 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
975 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
976 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
977 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
978
979 /* Active load balancing */
480b9434
KC
980 unsigned int alb_count;
981 unsigned int alb_failed;
982 unsigned int alb_pushed;
1da177e4 983
68767a0a 984 /* SD_BALANCE_EXEC stats */
480b9434
KC
985 unsigned int sbe_count;
986 unsigned int sbe_balanced;
987 unsigned int sbe_pushed;
1da177e4 988
68767a0a 989 /* SD_BALANCE_FORK stats */
480b9434
KC
990 unsigned int sbf_count;
991 unsigned int sbf_balanced;
992 unsigned int sbf_pushed;
68767a0a 993
1da177e4 994 /* try_to_wake_up() stats */
480b9434
KC
995 unsigned int ttwu_wake_remote;
996 unsigned int ttwu_move_affine;
997 unsigned int ttwu_move_balance;
1da177e4 998#endif
a5d8c348
IM
999#ifdef CONFIG_SCHED_DEBUG
1000 char *name;
1001#endif
6c99e9ad 1002
4200efd9
IM
1003 /*
1004 * Span of all CPUs in this domain.
1005 *
1006 * NOTE: this field is variable length. (Allocated dynamically
1007 * by attaching extra space to the end of the structure,
1008 * depending on how many CPUs the kernel has booted up with)
1009 *
1010 * It is also be embedded into static data structures at build
1011 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1012 */
1013 unsigned long span[0];
1da177e4
LT
1014};
1015
758b2cdc
RR
1016static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1017{
6c99e9ad 1018 return to_cpumask(sd->span);
758b2cdc
RR
1019}
1020
acc3f5d7 1021extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 1022 struct sched_domain_attr *dattr_new);
029190c5 1023
acc3f5d7
RR
1024/* Allocate an array of sched domains, for partition_sched_domains(). */
1025cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1026void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1027
06aaf76a
IM
1028/* Test a flag in parent sched domain */
1029static inline int test_sd_parent(struct sched_domain *sd, int flag)
1030{
1031 if (sd->parent && (sd->parent->flags & flag))
1032 return 1;
1033
1034 return 0;
1035}
029190c5 1036
47fe38fc
PZ
1037unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1038unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1039
1b427c15 1040#else /* CONFIG_SMP */
1da177e4 1041
1b427c15 1042struct sched_domain_attr;
d02c7a8c 1043
1b427c15 1044static inline void
acc3f5d7 1045partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1046 struct sched_domain_attr *dattr_new)
1047{
d02c7a8c 1048}
1b427c15 1049#endif /* !CONFIG_SMP */
1da177e4 1050
47fe38fc 1051
1da177e4 1052struct io_context; /* See blkdev.h */
1da177e4 1053
1da177e4 1054
383f2835 1055#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1056extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1057#else
1058static inline void prefetch_stack(struct task_struct *t) { }
1059#endif
1da177e4
LT
1060
1061struct audit_context; /* See audit.c */
1062struct mempolicy;
b92ce558 1063struct pipe_inode_info;
4865ecf1 1064struct uts_namespace;
1da177e4 1065
20b8a59f
IM
1066struct rq;
1067struct sched_domain;
1068
7d478721
PZ
1069/*
1070 * wake flags
1071 */
1072#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1073#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1074
20b8a59f 1075struct sched_class {
5522d5d5 1076 const struct sched_class *next;
20b8a59f 1077
fd390f6a 1078 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
f02231e5 1079 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
4530d7ab 1080 void (*yield_task) (struct rq *rq);
20b8a59f 1081
7d478721 1082 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1083
fb8d4724 1084 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1085 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1086
681f3e68 1087#ifdef CONFIG_SMP
7d478721 1088 int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
4ce72a2c 1089
43010659 1090 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
e1d1484f 1091 struct rq *busiest, unsigned long max_load_move,
20b8a59f 1092 struct sched_domain *sd, enum cpu_idle_type idle,
a4ac01c3 1093 int *all_pinned, int *this_best_prio);
20b8a59f 1094
e1d1484f
PW
1095 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1096 struct rq *busiest, struct sched_domain *sd,
1097 enum cpu_idle_type idle);
9a897c5a
SR
1098 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1099 void (*post_schedule) (struct rq *this_rq);
efbbd05a
PZ
1100 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1101 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1102
cd8ba7cd 1103 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1104 const struct cpumask *newmask);
57d885fe 1105
1f11eb6a
GH
1106 void (*rq_online)(struct rq *rq);
1107 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1108#endif
1109
1110 void (*set_curr_task) (struct rq *rq);
1111 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1112 void (*task_fork) (struct task_struct *p);
cb469845
SR
1113
1114 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1115 int running);
1116 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1117 int running);
1118 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1119 int oldprio, int running);
810b3817 1120
dba091b9
TG
1121 unsigned int (*get_rr_interval) (struct rq *rq,
1122 struct task_struct *task);
0d721cea 1123
810b3817 1124#ifdef CONFIG_FAIR_GROUP_SCHED
88ec22d3 1125 void (*moved_group) (struct task_struct *p, int on_rq);
810b3817 1126#endif
20b8a59f
IM
1127};
1128
1129struct load_weight {
1130 unsigned long weight, inv_weight;
1131};
1132
1133/*
1134 * CFS stats for a schedulable entity (task, task-group etc)
1135 *
1136 * Current field usage histogram:
1137 *
1138 * 4 se->block_start
1139 * 4 se->run_node
1140 * 4 se->sleep_start
20b8a59f 1141 * 6 se->load.weight
20b8a59f
IM
1142 */
1143struct sched_entity {
20b8a59f
IM
1144 struct load_weight load; /* for load-balancing */
1145 struct rb_node run_node;
4a55bd5e 1146 struct list_head group_node;
20b8a59f
IM
1147 unsigned int on_rq;
1148
94c18227
IM
1149 u64 exec_start;
1150 u64 sum_exec_runtime;
e9acbff6 1151 u64 vruntime;
f6cf891c 1152 u64 prev_sum_exec_runtime;
94c18227 1153
4ae7d5ce
IM
1154 u64 last_wakeup;
1155 u64 avg_overlap;
1156
6c594c21
IM
1157 u64 nr_migrations;
1158
34cb6135
IM
1159 u64 start_runtime;
1160 u64 avg_wakeup;
34cb6135 1161
94c18227 1162#ifdef CONFIG_SCHEDSTATS
20b8a59f 1163 u64 wait_start;
94c18227 1164 u64 wait_max;
6d082592
AV
1165 u64 wait_count;
1166 u64 wait_sum;
8f0dfc34
AV
1167 u64 iowait_count;
1168 u64 iowait_sum;
94c18227 1169
20b8a59f 1170 u64 sleep_start;
20b8a59f 1171 u64 sleep_max;
94c18227
IM
1172 s64 sum_sleep_runtime;
1173
1174 u64 block_start;
20b8a59f
IM
1175 u64 block_max;
1176 u64 exec_max;
eba1ed4b 1177 u64 slice_max;
cc367732 1178
cc367732
IM
1179 u64 nr_migrations_cold;
1180 u64 nr_failed_migrations_affine;
1181 u64 nr_failed_migrations_running;
1182 u64 nr_failed_migrations_hot;
1183 u64 nr_forced_migrations;
cc367732
IM
1184
1185 u64 nr_wakeups;
1186 u64 nr_wakeups_sync;
1187 u64 nr_wakeups_migrate;
1188 u64 nr_wakeups_local;
1189 u64 nr_wakeups_remote;
1190 u64 nr_wakeups_affine;
1191 u64 nr_wakeups_affine_attempts;
1192 u64 nr_wakeups_passive;
1193 u64 nr_wakeups_idle;
94c18227
IM
1194#endif
1195
20b8a59f
IM
1196#ifdef CONFIG_FAIR_GROUP_SCHED
1197 struct sched_entity *parent;
1198 /* rq on which this entity is (to be) queued: */
1199 struct cfs_rq *cfs_rq;
1200 /* rq "owned" by this entity/group: */
1201 struct cfs_rq *my_q;
1202#endif
1203};
70b97a7f 1204
fa717060
PZ
1205struct sched_rt_entity {
1206 struct list_head run_list;
78f2c7db 1207 unsigned long timeout;
bee367ed 1208 unsigned int time_slice;
6f505b16
PZ
1209 int nr_cpus_allowed;
1210
58d6c2d7 1211 struct sched_rt_entity *back;
052f1dc7 1212#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1213 struct sched_rt_entity *parent;
1214 /* rq on which this entity is (to be) queued: */
1215 struct rt_rq *rt_rq;
1216 /* rq "owned" by this entity/group: */
1217 struct rt_rq *my_q;
1218#endif
fa717060
PZ
1219};
1220
86848966
PM
1221struct rcu_node;
1222
1da177e4
LT
1223struct task_struct {
1224 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1225 void *stack;
1da177e4 1226 atomic_t usage;
97dc32cd
WC
1227 unsigned int flags; /* per process flags, defined below */
1228 unsigned int ptrace;
1da177e4 1229
36772092 1230 int lock_depth; /* BKL lock depth */
1da177e4 1231
2dd73a4f
PW
1232#ifdef CONFIG_SMP
1233#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1234 int oncpu;
1235#endif
2dd73a4f 1236#endif
50e645a8 1237
b29739f9 1238 int prio, static_prio, normal_prio;
c7aceaba 1239 unsigned int rt_priority;
5522d5d5 1240 const struct sched_class *sched_class;
20b8a59f 1241 struct sched_entity se;
fa717060 1242 struct sched_rt_entity rt;
1da177e4 1243
e107be36
AK
1244#ifdef CONFIG_PREEMPT_NOTIFIERS
1245 /* list of struct preempt_notifier: */
1246 struct hlist_head preempt_notifiers;
1247#endif
1248
18796aa0
AD
1249 /*
1250 * fpu_counter contains the number of consecutive context switches
1251 * that the FPU is used. If this is over a threshold, the lazy fpu
1252 * saving becomes unlazy to save the trap. This is an unsigned char
1253 * so that after 256 times the counter wraps and the behavior turns
1254 * lazy again; this to deal with bursty apps that only use FPU for
1255 * a short time
1256 */
1257 unsigned char fpu_counter;
6c5c9341 1258#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1259 unsigned int btrace_seq;
6c5c9341 1260#endif
1da177e4 1261
97dc32cd 1262 unsigned int policy;
1da177e4 1263 cpumask_t cpus_allowed;
1da177e4 1264
f41d911f 1265#ifdef CONFIG_TREE_PREEMPT_RCU
e260be67 1266 int rcu_read_lock_nesting;
f41d911f 1267 char rcu_read_unlock_special;
86848966 1268 struct rcu_node *rcu_blocked_node;
f41d911f
PM
1269 struct list_head rcu_node_entry;
1270#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
e260be67 1271
52f17b6c 1272#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1273 struct sched_info sched_info;
1274#endif
1275
1276 struct list_head tasks;
917b627d 1277 struct plist_node pushable_tasks;
1da177e4
LT
1278
1279 struct mm_struct *mm, *active_mm;
1280
1281/* task state */
97dc32cd 1282 int exit_state;
1da177e4
LT
1283 int exit_code, exit_signal;
1284 int pdeath_signal; /* The signal sent when the parent dies */
1285 /* ??? */
97dc32cd 1286 unsigned int personality;
1da177e4 1287 unsigned did_exec:1;
f9ce1f1c
KT
1288 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1289 * execve */
8f0dfc34
AV
1290 unsigned in_iowait:1;
1291
ca94c442
LP
1292
1293 /* Revert to default priority/policy when forking */
1294 unsigned sched_reset_on_fork:1;
1295
1da177e4
LT
1296 pid_t pid;
1297 pid_t tgid;
0a425405 1298
1314562a 1299#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1300 /* Canary value for the -fstack-protector gcc feature */
1301 unsigned long stack_canary;
1314562a 1302#endif
e0032087 1303
1da177e4
LT
1304 /*
1305 * pointers to (original) parent process, youngest child, younger sibling,
1306 * older sibling, respectively. (p->father can be replaced with
f470021a 1307 * p->real_parent->pid)
1da177e4 1308 */
f470021a
RM
1309 struct task_struct *real_parent; /* real parent process */
1310 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1311 /*
f470021a 1312 * children/sibling forms the list of my natural children
1da177e4
LT
1313 */
1314 struct list_head children; /* list of my children */
1315 struct list_head sibling; /* linkage in my parent's children list */
1316 struct task_struct *group_leader; /* threadgroup leader */
1317
f470021a
RM
1318 /*
1319 * ptraced is the list of tasks this task is using ptrace on.
1320 * This includes both natural children and PTRACE_ATTACH targets.
1321 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1322 */
1323 struct list_head ptraced;
1324 struct list_head ptrace_entry;
1325
ca0002a1
MM
1326 /*
1327 * This is the tracer handle for the ptrace BTS extension.
1328 * This field actually belongs to the ptracer task.
1329 */
e2b371f0 1330 struct bts_context *bts;
ca0002a1 1331
1da177e4 1332 /* PID/PID hash table linkage. */
92476d7f 1333 struct pid_link pids[PIDTYPE_MAX];
47e65328 1334 struct list_head thread_group;
1da177e4
LT
1335
1336 struct completion *vfork_done; /* for vfork() */
1337 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1338 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1339
c66f08be 1340 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1341 cputime_t gtime;
d99ca3b9 1342#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1343 cputime_t prev_utime, prev_stime;
d99ca3b9 1344#endif
1da177e4 1345 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1346 struct timespec start_time; /* monotonic time */
1347 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1348/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1349 unsigned long min_flt, maj_flt;
1350
f06febc9 1351 struct task_cputime cputime_expires;
1da177e4
LT
1352 struct list_head cpu_timers[3];
1353
1354/* process credentials */
3b11a1de
DH
1355 const struct cred *real_cred; /* objective and real subjective task
1356 * credentials (COW) */
1357 const struct cred *cred; /* effective (overridable) subjective task
1358 * credentials (COW) */
5e751e99
DH
1359 struct mutex cred_guard_mutex; /* guard against foreign influences on
1360 * credential calculations
1361 * (notably. ptrace) */
ee18d64c 1362 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1363
36772092
PBG
1364 char comm[TASK_COMM_LEN]; /* executable name excluding path
1365 - access with [gs]et_task_comm (which lock
1366 it with task_lock())
1367 - initialized normally by flush_old_exec */
1da177e4
LT
1368/* file system info */
1369 int link_count, total_link_count;
3d5b6fcc 1370#ifdef CONFIG_SYSVIPC
1da177e4
LT
1371/* ipc stuff */
1372 struct sysv_sem sysvsem;
3d5b6fcc 1373#endif
e162b39a 1374#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1375/* hung task detection */
82a1fcb9
IM
1376 unsigned long last_switch_count;
1377#endif
1da177e4
LT
1378/* CPU-specific state of this task */
1379 struct thread_struct thread;
1380/* filesystem information */
1381 struct fs_struct *fs;
1382/* open file information */
1383 struct files_struct *files;
1651e14e 1384/* namespaces */
ab516013 1385 struct nsproxy *nsproxy;
1da177e4
LT
1386/* signal handlers */
1387 struct signal_struct *signal;
1388 struct sighand_struct *sighand;
1389
1390 sigset_t blocked, real_blocked;
f3de272b 1391 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1392 struct sigpending pending;
1393
1394 unsigned long sas_ss_sp;
1395 size_t sas_ss_size;
1396 int (*notifier)(void *priv);
1397 void *notifier_data;
1398 sigset_t *notifier_mask;
1da177e4 1399 struct audit_context *audit_context;
bfef93a5
AV
1400#ifdef CONFIG_AUDITSYSCALL
1401 uid_t loginuid;
4746ec5b 1402 unsigned int sessionid;
bfef93a5 1403#endif
1da177e4
LT
1404 seccomp_t seccomp;
1405
1406/* Thread group tracking */
1407 u32 parent_exec_id;
1408 u32 self_exec_id;
58568d2a
MX
1409/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1410 * mempolicy */
1da177e4 1411 spinlock_t alloc_lock;
1da177e4 1412
3aa551c9
TG
1413#ifdef CONFIG_GENERIC_HARDIRQS
1414 /* IRQ handler threads */
1415 struct irqaction *irqaction;
1416#endif
1417
b29739f9 1418 /* Protection of the PI data structures: */
1d615482 1419 raw_spinlock_t pi_lock;
b29739f9 1420
23f78d4a
IM
1421#ifdef CONFIG_RT_MUTEXES
1422 /* PI waiters blocked on a rt_mutex held by this task */
1423 struct plist_head pi_waiters;
1424 /* Deadlock detection and priority inheritance handling */
1425 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1426#endif
1427
408894ee
IM
1428#ifdef CONFIG_DEBUG_MUTEXES
1429 /* mutex deadlock detection */
1430 struct mutex_waiter *blocked_on;
1431#endif
de30a2b3
IM
1432#ifdef CONFIG_TRACE_IRQFLAGS
1433 unsigned int irq_events;
de30a2b3 1434 unsigned long hardirq_enable_ip;
de30a2b3 1435 unsigned long hardirq_disable_ip;
fa1452e8 1436 unsigned int hardirq_enable_event;
de30a2b3 1437 unsigned int hardirq_disable_event;
fa1452e8
HS
1438 int hardirqs_enabled;
1439 int hardirq_context;
de30a2b3 1440 unsigned long softirq_disable_ip;
de30a2b3 1441 unsigned long softirq_enable_ip;
fa1452e8 1442 unsigned int softirq_disable_event;
de30a2b3 1443 unsigned int softirq_enable_event;
fa1452e8 1444 int softirqs_enabled;
de30a2b3
IM
1445 int softirq_context;
1446#endif
fbb9ce95 1447#ifdef CONFIG_LOCKDEP
bdb9441e 1448# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1449 u64 curr_chain_key;
1450 int lockdep_depth;
fbb9ce95 1451 unsigned int lockdep_recursion;
c7aceaba 1452 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1453 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1454#endif
408894ee 1455
1da177e4
LT
1456/* journalling filesystem info */
1457 void *journal_info;
1458
d89d8796
NB
1459/* stacked block device info */
1460 struct bio *bio_list, **bio_tail;
1461
1da177e4
LT
1462/* VM state */
1463 struct reclaim_state *reclaim_state;
1464
1da177e4
LT
1465 struct backing_dev_info *backing_dev_info;
1466
1467 struct io_context *io_context;
1468
1469 unsigned long ptrace_message;
1470 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1471 struct task_io_accounting ioac;
8f0ab514 1472#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1473 u64 acct_rss_mem1; /* accumulated rss usage */
1474 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1475 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1476#endif
1477#ifdef CONFIG_CPUSETS
58568d2a 1478 nodemask_t mems_allowed; /* Protected by alloc_lock */
825a46af 1479 int cpuset_mem_spread_rotor;
1da177e4 1480#endif
ddbcc7e8 1481#ifdef CONFIG_CGROUPS
817929ec
PM
1482 /* Control Group info protected by css_set_lock */
1483 struct css_set *cgroups;
1484 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1485 struct list_head cg_list;
ddbcc7e8 1486#endif
42b2dd0a 1487#ifdef CONFIG_FUTEX
0771dfef 1488 struct robust_list_head __user *robust_list;
34f192c6
IM
1489#ifdef CONFIG_COMPAT
1490 struct compat_robust_list_head __user *compat_robust_list;
1491#endif
c87e2837
IM
1492 struct list_head pi_state_list;
1493 struct futex_pi_state *pi_state_cache;
c7aceaba 1494#endif
cdd6c482
IM
1495#ifdef CONFIG_PERF_EVENTS
1496 struct perf_event_context *perf_event_ctxp;
1497 struct mutex perf_event_mutex;
1498 struct list_head perf_event_list;
a63eaf34 1499#endif
c7aceaba 1500#ifdef CONFIG_NUMA
58568d2a 1501 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1502 short il_next;
42b2dd0a 1503#endif
22e2c507 1504 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1505 struct rcu_head rcu;
b92ce558
JA
1506
1507 /*
1508 * cache last used pipe for splice
1509 */
1510 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1511#ifdef CONFIG_TASK_DELAY_ACCT
1512 struct task_delay_info *delays;
f4f154fd
AM
1513#endif
1514#ifdef CONFIG_FAULT_INJECTION
1515 int make_it_fail;
ca74e92b 1516#endif
3e26c149 1517 struct prop_local_single dirties;
9745512c
AV
1518#ifdef CONFIG_LATENCYTOP
1519 int latency_record_count;
1520 struct latency_record latency_record[LT_SAVECOUNT];
1521#endif
6976675d
AV
1522 /*
1523 * time slack values; these are used to round up poll() and
1524 * select() etc timeout values. These are in nanoseconds.
1525 */
1526 unsigned long timer_slack_ns;
1527 unsigned long default_timer_slack_ns;
f8d570a4
DM
1528
1529 struct list_head *scm_work_list;
fb52607a 1530#ifdef CONFIG_FUNCTION_GRAPH_TRACER
f201ae23
FW
1531 /* Index of current stored adress in ret_stack */
1532 int curr_ret_stack;
1533 /* Stack of return addresses for return function tracing */
1534 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1535 /* time stamp for last schedule */
1536 unsigned long long ftrace_timestamp;
f201ae23
FW
1537 /*
1538 * Number of functions that haven't been traced
1539 * because of depth overrun.
1540 */
1541 atomic_t trace_overrun;
380c4b14
FW
1542 /* Pause for the tracing */
1543 atomic_t tracing_graph_pause;
f201ae23 1544#endif
ea4e2bc4
SR
1545#ifdef CONFIG_TRACING
1546 /* state flags for use by tracers */
1547 unsigned long trace;
261842b7
SR
1548 /* bitmask of trace recursion */
1549 unsigned long trace_recursion;
1550#endif /* CONFIG_TRACING */
d899bf7b 1551 unsigned long stack_start;
569b846d
KH
1552#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1553 struct memcg_batch_info {
1554 int do_batch; /* incremented when batch uncharge started */
1555 struct mem_cgroup *memcg; /* target memcg of uncharge */
1556 unsigned long bytes; /* uncharged usage */
1557 unsigned long memsw_bytes; /* uncharged mem+swap usage */
1558 } memcg_batch;
1559#endif
1da177e4
LT
1560};
1561
76e6eee0 1562/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1563#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1564
e05606d3
IM
1565/*
1566 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1567 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1568 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1569 * values are inverted: lower p->prio value means higher priority.
1570 *
1571 * The MAX_USER_RT_PRIO value allows the actual maximum
1572 * RT priority to be separate from the value exported to
1573 * user-space. This allows kernel threads to set their
1574 * priority to a value higher than any user task. Note:
1575 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1576 */
1577
1578#define MAX_USER_RT_PRIO 100
1579#define MAX_RT_PRIO MAX_USER_RT_PRIO
1580
1581#define MAX_PRIO (MAX_RT_PRIO + 40)
1582#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1583
1584static inline int rt_prio(int prio)
1585{
1586 if (unlikely(prio < MAX_RT_PRIO))
1587 return 1;
1588 return 0;
1589}
1590
e868171a 1591static inline int rt_task(struct task_struct *p)
e05606d3
IM
1592{
1593 return rt_prio(p->prio);
1594}
1595
e868171a 1596static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1597{
1598 return task->pids[PIDTYPE_PID].pid;
1599}
1600
e868171a 1601static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1602{
1603 return task->group_leader->pids[PIDTYPE_PID].pid;
1604}
1605
6dda81f4
ON
1606/*
1607 * Without tasklist or rcu lock it is not safe to dereference
1608 * the result of task_pgrp/task_session even if task == current,
1609 * we can race with another thread doing sys_setsid/sys_setpgid.
1610 */
e868171a 1611static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1612{
1613 return task->group_leader->pids[PIDTYPE_PGID].pid;
1614}
1615
e868171a 1616static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1617{
1618 return task->group_leader->pids[PIDTYPE_SID].pid;
1619}
1620
7af57294
PE
1621struct pid_namespace;
1622
1623/*
1624 * the helpers to get the task's different pids as they are seen
1625 * from various namespaces
1626 *
1627 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1628 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1629 * current.
7af57294
PE
1630 * task_xid_nr_ns() : id seen from the ns specified;
1631 *
1632 * set_task_vxid() : assigns a virtual id to a task;
1633 *
7af57294
PE
1634 * see also pid_nr() etc in include/linux/pid.h
1635 */
52ee2dfd
ON
1636pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1637 struct pid_namespace *ns);
7af57294 1638
e868171a 1639static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1640{
1641 return tsk->pid;
1642}
1643
52ee2dfd
ON
1644static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1645 struct pid_namespace *ns)
1646{
1647 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1648}
7af57294
PE
1649
1650static inline pid_t task_pid_vnr(struct task_struct *tsk)
1651{
52ee2dfd 1652 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1653}
1654
1655
e868171a 1656static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1657{
1658 return tsk->tgid;
1659}
1660
2f2a3a46 1661pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1662
1663static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1664{
1665 return pid_vnr(task_tgid(tsk));
1666}
1667
1668
52ee2dfd
ON
1669static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1670 struct pid_namespace *ns)
7af57294 1671{
52ee2dfd 1672 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1673}
1674
7af57294
PE
1675static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1676{
52ee2dfd 1677 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1678}
1679
1680
52ee2dfd
ON
1681static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1682 struct pid_namespace *ns)
7af57294 1683{
52ee2dfd 1684 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1685}
1686
7af57294
PE
1687static inline pid_t task_session_vnr(struct task_struct *tsk)
1688{
52ee2dfd 1689 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1690}
1691
1b0f7ffd
ON
1692/* obsolete, do not use */
1693static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1694{
1695 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1696}
7af57294 1697
1da177e4
LT
1698/**
1699 * pid_alive - check that a task structure is not stale
1700 * @p: Task structure to be checked.
1701 *
1702 * Test if a process is not yet dead (at most zombie state)
1703 * If pid_alive fails, then pointers within the task structure
1704 * can be stale and must not be dereferenced.
1705 */
e868171a 1706static inline int pid_alive(struct task_struct *p)
1da177e4 1707{
92476d7f 1708 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1709}
1710
f400e198 1711/**
b460cbc5 1712 * is_global_init - check if a task structure is init
3260259f
HK
1713 * @tsk: Task structure to be checked.
1714 *
1715 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1716 */
e868171a 1717static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1718{
1719 return tsk->pid == 1;
1720}
b460cbc5
SH
1721
1722/*
1723 * is_container_init:
1724 * check whether in the task is init in its own pid namespace.
f400e198 1725 */
b461cc03 1726extern int is_container_init(struct task_struct *tsk);
f400e198 1727
9ec52099
CLG
1728extern struct pid *cad_pid;
1729
1da177e4 1730extern void free_task(struct task_struct *tsk);
1da177e4 1731#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1732
158d9ebd 1733extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1734
1735static inline void put_task_struct(struct task_struct *t)
1736{
1737 if (atomic_dec_and_test(&t->usage))
8c7904a0 1738 __put_task_struct(t);
e56d0903 1739}
1da177e4 1740
d180c5bc 1741extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1742extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1743
1da177e4
LT
1744/*
1745 * Per process flags
1746 */
1747#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1748 /* Not implemented yet, only for 486*/
1749#define PF_STARTING 0x00000002 /* being created */
1750#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1751#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1752#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1da177e4 1753#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1754#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1755#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1756#define PF_DUMPCORE 0x00000200 /* dumped core */
1757#define PF_SIGNALED 0x00000400 /* killed by a signal */
1758#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1759#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1760#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1761#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1762#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1763#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1764#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1765#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1766#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1767#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1768#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1769#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1770#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1771#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1772#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1773#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1774#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1775#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1776#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
ba96a0c8 1777#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
ebb12db5 1778#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1779
1780/*
1781 * Only the _current_ task can read/write to tsk->flags, but other
1782 * tasks can access tsk->flags in readonly mode for example
1783 * with tsk_used_math (like during threaded core dumping).
1784 * There is however an exception to this rule during ptrace
1785 * or during fork: the ptracer task is allowed to write to the
1786 * child->flags of its traced child (same goes for fork, the parent
1787 * can write to the child->flags), because we're guaranteed the
1788 * child is not running and in turn not changing child->flags
1789 * at the same time the parent does it.
1790 */
1791#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1792#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1793#define clear_used_math() clear_stopped_child_used_math(current)
1794#define set_used_math() set_stopped_child_used_math(current)
1795#define conditional_stopped_child_used_math(condition, child) \
1796 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1797#define conditional_used_math(condition) \
1798 conditional_stopped_child_used_math(condition, current)
1799#define copy_to_stopped_child_used_math(child) \
1800 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1801/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1802#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1803#define used_math() tsk_used_math(current)
1804
f41d911f
PM
1805#ifdef CONFIG_TREE_PREEMPT_RCU
1806
1807#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1808#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1809
1810static inline void rcu_copy_process(struct task_struct *p)
1811{
1812 p->rcu_read_lock_nesting = 0;
1813 p->rcu_read_unlock_special = 0;
dd5d19ba 1814 p->rcu_blocked_node = NULL;
f41d911f
PM
1815 INIT_LIST_HEAD(&p->rcu_node_entry);
1816}
1817
f41d911f
PM
1818#else
1819
1820static inline void rcu_copy_process(struct task_struct *p)
1821{
1822}
1823
1824#endif
1825
1da177e4 1826#ifdef CONFIG_SMP
cd8ba7cd 1827extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1828 const struct cpumask *new_mask);
1da177e4 1829#else
cd8ba7cd 1830static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1831 const struct cpumask *new_mask)
1da177e4 1832{
96f874e2 1833 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1834 return -EINVAL;
1835 return 0;
1836}
1837#endif
e0ad9556
RR
1838
1839#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1840static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1841{
1842 return set_cpus_allowed_ptr(p, &new_mask);
1843}
e0ad9556 1844#endif
1da177e4 1845
b342501c
IM
1846/*
1847 * Architectures can set this to 1 if they have specified
1848 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1849 * but then during bootup it turns out that sched_clock()
1850 * is reliable after all:
1851 */
1852#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1853extern int sched_clock_stable;
1854#endif
1855
1bbfa6f2
MF
1856/* ftrace calls sched_clock() directly */
1857extern unsigned long long notrace sched_clock(void);
e436d800 1858
c1955a3d
PZ
1859extern void sched_clock_init(void);
1860extern u64 sched_clock_cpu(int cpu);
3e51f33f 1861
c1955a3d 1862#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1863static inline void sched_clock_tick(void)
1864{
1865}
1866
1867static inline void sched_clock_idle_sleep_event(void)
1868{
1869}
1870
1871static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1872{
1873}
1874#else
3e51f33f
PZ
1875extern void sched_clock_tick(void);
1876extern void sched_clock_idle_sleep_event(void);
1877extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1878#endif
1879
e436d800
IM
1880/*
1881 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1882 * clock constructed from sched_clock():
1883 */
1884extern unsigned long long cpu_clock(int cpu);
1885
36c8b586 1886extern unsigned long long
41b86e9c 1887task_sched_runtime(struct task_struct *task);
f06febc9 1888extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1889
1890/* sched_exec is called by processes performing an exec */
1891#ifdef CONFIG_SMP
1892extern void sched_exec(void);
1893#else
1894#define sched_exec() {}
1895#endif
1896
2aa44d05
IM
1897extern void sched_clock_idle_sleep_event(void);
1898extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1899
1da177e4
LT
1900#ifdef CONFIG_HOTPLUG_CPU
1901extern void idle_task_exit(void);
1902#else
1903static inline void idle_task_exit(void) {}
1904#endif
1905
1906extern void sched_idle_next(void);
b29739f9 1907
06d8308c
TG
1908#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1909extern void wake_up_idle_cpu(int cpu);
1910#else
1911static inline void wake_up_idle_cpu(int cpu) { }
1912#endif
1913
21805085 1914extern unsigned int sysctl_sched_latency;
b2be5e96 1915extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1916extern unsigned int sysctl_sched_wakeup_granularity;
47fea2ad
JSR
1917extern unsigned int sysctl_sched_shares_ratelimit;
1918extern unsigned int sysctl_sched_shares_thresh;
bf0f6f24 1919extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
1920
1921enum sched_tunable_scaling {
1922 SCHED_TUNABLESCALING_NONE,
1923 SCHED_TUNABLESCALING_LOG,
1924 SCHED_TUNABLESCALING_LINEAR,
1925 SCHED_TUNABLESCALING_END,
1926};
1927extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1928
2bba22c5 1929#ifdef CONFIG_SCHED_DEBUG
da84d961 1930extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1931extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1932extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1933extern unsigned int sysctl_timer_migration;
b2be5e96 1934
1983a922 1935int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1936 void __user *buffer, size_t *length,
b2be5e96 1937 loff_t *ppos);
2bd8e6d4 1938#endif
eea08f32
AB
1939#ifdef CONFIG_SCHED_DEBUG
1940static inline unsigned int get_sysctl_timer_migration(void)
1941{
1942 return sysctl_timer_migration;
1943}
1944#else
1945static inline unsigned int get_sysctl_timer_migration(void)
1946{
1947 return 1;
1948}
1949#endif
9f0c1e56
PZ
1950extern unsigned int sysctl_sched_rt_period;
1951extern int sysctl_sched_rt_runtime;
2bd8e6d4 1952
d0b27fa7 1953int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1954 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
1955 loff_t *ppos);
1956
2bd8e6d4 1957extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1958
b29739f9 1959#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1960extern int rt_mutex_getprio(struct task_struct *p);
1961extern void rt_mutex_setprio(struct task_struct *p, int prio);
1962extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1963#else
e868171a 1964static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1965{
1966 return p->normal_prio;
1967}
95e02ca9 1968# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1969#endif
1970
36c8b586
IM
1971extern void set_user_nice(struct task_struct *p, long nice);
1972extern int task_prio(const struct task_struct *p);
1973extern int task_nice(const struct task_struct *p);
1974extern int can_nice(const struct task_struct *p, const int nice);
1975extern int task_curr(const struct task_struct *p);
1da177e4
LT
1976extern int idle_cpu(int cpu);
1977extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
961ccddd
RR
1978extern int sched_setscheduler_nocheck(struct task_struct *, int,
1979 struct sched_param *);
36c8b586
IM
1980extern struct task_struct *idle_task(int cpu);
1981extern struct task_struct *curr_task(int cpu);
1982extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1983
1984void yield(void);
1985
1986/*
1987 * The default (Linux) execution domain.
1988 */
1989extern struct exec_domain default_exec_domain;
1990
1991union thread_union {
1992 struct thread_info thread_info;
1993 unsigned long stack[THREAD_SIZE/sizeof(long)];
1994};
1995
1996#ifndef __HAVE_ARCH_KSTACK_END
1997static inline int kstack_end(void *addr)
1998{
1999 /* Reliable end of stack detection:
2000 * Some APM bios versions misalign the stack
2001 */
2002 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2003}
2004#endif
2005
2006extern union thread_union init_thread_union;
2007extern struct task_struct init_task;
2008
2009extern struct mm_struct init_mm;
2010
198fe21b
PE
2011extern struct pid_namespace init_pid_ns;
2012
2013/*
2014 * find a task by one of its numerical ids
2015 *
198fe21b
PE
2016 * find_task_by_pid_ns():
2017 * finds a task by its pid in the specified namespace
228ebcbe
PE
2018 * find_task_by_vpid():
2019 * finds a task by its virtual pid
198fe21b 2020 *
e49859e7 2021 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2022 */
2023
228ebcbe
PE
2024extern struct task_struct *find_task_by_vpid(pid_t nr);
2025extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2026 struct pid_namespace *ns);
198fe21b 2027
8520d7c7 2028extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2029
2030/* per-UID process charging. */
acce292c 2031extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
2032static inline struct user_struct *get_uid(struct user_struct *u)
2033{
2034 atomic_inc(&u->__count);
2035 return u;
2036}
2037extern void free_uid(struct user_struct *);
28f300d2 2038extern void release_uids(struct user_namespace *ns);
1da177e4
LT
2039
2040#include <asm/current.h>
2041
3171a030 2042extern void do_timer(unsigned long ticks);
1da177e4 2043
b3c97528
HH
2044extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2045extern int wake_up_process(struct task_struct *tsk);
2046extern void wake_up_new_task(struct task_struct *tsk,
2047 unsigned long clone_flags);
1da177e4
LT
2048#ifdef CONFIG_SMP
2049 extern void kick_process(struct task_struct *tsk);
2050#else
2051 static inline void kick_process(struct task_struct *tsk) { }
2052#endif
ad46c2c4
IM
2053extern void sched_fork(struct task_struct *p, int clone_flags);
2054extern void sched_dead(struct task_struct *p);
1da177e4 2055
1da177e4
LT
2056extern void proc_caches_init(void);
2057extern void flush_signals(struct task_struct *);
3bcac026 2058extern void __flush_signals(struct task_struct *);
10ab825b 2059extern void ignore_signals(struct task_struct *);
1da177e4
LT
2060extern void flush_signal_handlers(struct task_struct *, int force_default);
2061extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2062
2063static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2064{
2065 unsigned long flags;
2066 int ret;
2067
2068 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2069 ret = dequeue_signal(tsk, mask, info);
2070 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2071
2072 return ret;
2073}
2074
2075extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2076 sigset_t *mask);
2077extern void unblock_all_signals(void);
2078extern void release_task(struct task_struct * p);
2079extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2080extern int force_sigsegv(int, struct task_struct *);
2081extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2082extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2083extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2084extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2085extern int kill_pgrp(struct pid *pid, int sig, int priv);
2086extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2087extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2088extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2089extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2090extern void force_sig(int, struct task_struct *);
1da177e4
LT
2091extern int send_sig(int, struct task_struct *, int);
2092extern void zap_other_threads(struct task_struct *p);
1da177e4
LT
2093extern struct sigqueue *sigqueue_alloc(void);
2094extern void sigqueue_free(struct sigqueue *);
ac5c2153 2095extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2096extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2097extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2098
9ec52099
CLG
2099static inline int kill_cad_pid(int sig, int priv)
2100{
2101 return kill_pid(cad_pid, sig, priv);
2102}
2103
1da177e4
LT
2104/* These can be the second arg to send_sig_info/send_group_sig_info. */
2105#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2106#define SEND_SIG_PRIV ((struct siginfo *) 1)
2107#define SEND_SIG_FORCED ((struct siginfo *) 2)
2108
2a855dd0
SAS
2109/*
2110 * True if we are on the alternate signal stack.
2111 */
1da177e4
LT
2112static inline int on_sig_stack(unsigned long sp)
2113{
2a855dd0
SAS
2114#ifdef CONFIG_STACK_GROWSUP
2115 return sp >= current->sas_ss_sp &&
2116 sp - current->sas_ss_sp < current->sas_ss_size;
2117#else
2118 return sp > current->sas_ss_sp &&
2119 sp - current->sas_ss_sp <= current->sas_ss_size;
2120#endif
1da177e4
LT
2121}
2122
2123static inline int sas_ss_flags(unsigned long sp)
2124{
2125 return (current->sas_ss_size == 0 ? SS_DISABLE
2126 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2127}
2128
1da177e4
LT
2129/*
2130 * Routines for handling mm_structs
2131 */
2132extern struct mm_struct * mm_alloc(void);
2133
2134/* mmdrop drops the mm and the page tables */
b3c97528 2135extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2136static inline void mmdrop(struct mm_struct * mm)
2137{
6fb43d7b 2138 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2139 __mmdrop(mm);
2140}
2141
2142/* mmput gets rid of the mappings and all user-space */
2143extern void mmput(struct mm_struct *);
2144/* Grab a reference to a task's mm, if it is not already going away */
2145extern struct mm_struct *get_task_mm(struct task_struct *task);
2146/* Remove the current tasks stale references to the old mm_struct */
2147extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2148/* Allocate a new mm structure and copy contents from tsk->mm */
2149extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2150
6f2c55b8
AD
2151extern int copy_thread(unsigned long, unsigned long, unsigned long,
2152 struct task_struct *, struct pt_regs *);
1da177e4
LT
2153extern void flush_thread(void);
2154extern void exit_thread(void);
2155
1da177e4 2156extern void exit_files(struct task_struct *);
6b3934ef 2157extern void __cleanup_signal(struct signal_struct *);
a7e5328a 2158extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2159
1da177e4 2160extern void exit_itimers(struct signal_struct *);
cbaffba1 2161extern void flush_itimer_signals(void);
1da177e4
LT
2162
2163extern NORET_TYPE void do_group_exit(int);
2164
1da177e4
LT
2165extern void daemonize(const char *, ...);
2166extern int allow_signal(int);
2167extern int disallow_signal(int);
1da177e4
LT
2168
2169extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2170extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2171struct task_struct *fork_idle(int);
1da177e4
LT
2172
2173extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2174extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2175
2176#ifdef CONFIG_SMP
a26b89f0 2177extern void wait_task_context_switch(struct task_struct *p);
85ba2d86 2178extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2179#else
a26b89f0 2180static inline void wait_task_context_switch(struct task_struct *p) {}
85ba2d86
RM
2181static inline unsigned long wait_task_inactive(struct task_struct *p,
2182 long match_state)
2183{
2184 return 1;
2185}
1da177e4
LT
2186#endif
2187
05725f7e
JP
2188#define next_task(p) \
2189 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2190
2191#define for_each_process(p) \
2192 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2193
5bb459bb 2194extern bool current_is_single_threaded(void);
d84f4f99 2195
1da177e4
LT
2196/*
2197 * Careful: do_each_thread/while_each_thread is a double loop so
2198 * 'break' will not work as expected - use goto instead.
2199 */
2200#define do_each_thread(g, t) \
2201 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2202
2203#define while_each_thread(g, t) \
2204 while ((t = next_thread(t)) != g)
2205
de12a787
EB
2206/* de_thread depends on thread_group_leader not being a pid based check */
2207#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2208
0804ef4b
EB
2209/* Do to the insanities of de_thread it is possible for a process
2210 * to have the pid of the thread group leader without actually being
2211 * the thread group leader. For iteration through the pids in proc
2212 * all we care about is that we have a task with the appropriate
2213 * pid, we don't actually care if we have the right task.
2214 */
e868171a 2215static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2216{
2217 return p->pid == p->tgid;
2218}
2219
bac0abd6
PE
2220static inline
2221int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2222{
2223 return p1->tgid == p2->tgid;
2224}
2225
36c8b586 2226static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2227{
05725f7e
JP
2228 return list_entry_rcu(p->thread_group.next,
2229 struct task_struct, thread_group);
47e65328
ON
2230}
2231
e868171a 2232static inline int thread_group_empty(struct task_struct *p)
1da177e4 2233{
47e65328 2234 return list_empty(&p->thread_group);
1da177e4
LT
2235}
2236
2237#define delay_group_leader(p) \
2238 (thread_group_leader(p) && !thread_group_empty(p))
2239
39c626ae
ON
2240static inline int task_detached(struct task_struct *p)
2241{
2242 return p->exit_signal == -1;
2243}
2244
1da177e4 2245/*
260ea101 2246 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2247 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2248 * pins the final release of task.io_context. Also protects ->cpuset and
2249 * ->cgroup.subsys[].
1da177e4
LT
2250 *
2251 * Nests both inside and outside of read_lock(&tasklist_lock).
2252 * It must not be nested with write_lock_irq(&tasklist_lock),
2253 * neither inside nor outside.
2254 */
2255static inline void task_lock(struct task_struct *p)
2256{
2257 spin_lock(&p->alloc_lock);
2258}
2259
2260static inline void task_unlock(struct task_struct *p)
2261{
2262 spin_unlock(&p->alloc_lock);
2263}
2264
f63ee72e
ON
2265extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2266 unsigned long *flags);
2267
2268static inline void unlock_task_sighand(struct task_struct *tsk,
2269 unsigned long *flags)
2270{
2271 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2272}
2273
f037360f
AV
2274#ifndef __HAVE_THREAD_FUNCTIONS
2275
f7e4217b
RZ
2276#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2277#define task_stack_page(task) ((task)->stack)
a1261f54 2278
10ebffde
AV
2279static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2280{
2281 *task_thread_info(p) = *task_thread_info(org);
2282 task_thread_info(p)->task = p;
2283}
2284
2285static inline unsigned long *end_of_stack(struct task_struct *p)
2286{
f7e4217b 2287 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2288}
2289
f037360f
AV
2290#endif
2291
8b05c7e6
FT
2292static inline int object_is_on_stack(void *obj)
2293{
2294 void *stack = task_stack_page(current);
2295
2296 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2297}
2298
8c9843e5
BH
2299extern void thread_info_cache_init(void);
2300
7c9f8861
ES
2301#ifdef CONFIG_DEBUG_STACK_USAGE
2302static inline unsigned long stack_not_used(struct task_struct *p)
2303{
2304 unsigned long *n = end_of_stack(p);
2305
2306 do { /* Skip over canary */
2307 n++;
2308 } while (!*n);
2309
2310 return (unsigned long)n - (unsigned long)end_of_stack(p);
2311}
2312#endif
2313
1da177e4
LT
2314/* set thread flags in other task's structures
2315 * - see asm/thread_info.h for TIF_xxxx flags available
2316 */
2317static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2318{
a1261f54 2319 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2320}
2321
2322static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2323{
a1261f54 2324 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2325}
2326
2327static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2328{
a1261f54 2329 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2330}
2331
2332static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2333{
a1261f54 2334 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2335}
2336
2337static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2338{
a1261f54 2339 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2340}
2341
2342static inline void set_tsk_need_resched(struct task_struct *tsk)
2343{
2344 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2345}
2346
2347static inline void clear_tsk_need_resched(struct task_struct *tsk)
2348{
2349 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2350}
2351
8ae121ac
GH
2352static inline int test_tsk_need_resched(struct task_struct *tsk)
2353{
2354 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2355}
2356
690cc3ff
EB
2357static inline int restart_syscall(void)
2358{
2359 set_tsk_thread_flag(current, TIF_SIGPENDING);
2360 return -ERESTARTNOINTR;
2361}
2362
1da177e4
LT
2363static inline int signal_pending(struct task_struct *p)
2364{
2365 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2366}
f776d12d 2367
d9588725
RM
2368static inline int __fatal_signal_pending(struct task_struct *p)
2369{
2370 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2371}
f776d12d
MW
2372
2373static inline int fatal_signal_pending(struct task_struct *p)
2374{
2375 return signal_pending(p) && __fatal_signal_pending(p);
2376}
2377
16882c1e
ON
2378static inline int signal_pending_state(long state, struct task_struct *p)
2379{
2380 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2381 return 0;
2382 if (!signal_pending(p))
2383 return 0;
2384
16882c1e
ON
2385 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2386}
2387
1da177e4
LT
2388static inline int need_resched(void)
2389{
9404ef02 2390 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2391}
2392
2393/*
2394 * cond_resched() and cond_resched_lock(): latency reduction via
2395 * explicit rescheduling in places that are safe. The return
2396 * value indicates whether a reschedule was done in fact.
2397 * cond_resched_lock() will drop the spinlock before scheduling,
2398 * cond_resched_softirq() will enable bhs before scheduling.
2399 */
c3921ab7 2400extern int _cond_resched(void);
6f80bd98 2401
613afbf8
FW
2402#define cond_resched() ({ \
2403 __might_sleep(__FILE__, __LINE__, 0); \
2404 _cond_resched(); \
2405})
6f80bd98 2406
613afbf8
FW
2407extern int __cond_resched_lock(spinlock_t *lock);
2408
716a4234
FW
2409#ifdef CONFIG_PREEMPT
2410#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2411#else
716a4234 2412#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2413#endif
716a4234 2414
613afbf8 2415#define cond_resched_lock(lock) ({ \
716a4234 2416 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2417 __cond_resched_lock(lock); \
2418})
2419
2420extern int __cond_resched_softirq(void);
2421
2422#define cond_resched_softirq() ({ \
2423 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2424 __cond_resched_softirq(); \
2425})
1da177e4
LT
2426
2427/*
2428 * Does a critical section need to be broken due to another
95c354fe
NP
2429 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2430 * but a general need for low latency)
1da177e4 2431 */
95c354fe 2432static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2433{
95c354fe
NP
2434#ifdef CONFIG_PREEMPT
2435 return spin_is_contended(lock);
2436#else
1da177e4 2437 return 0;
95c354fe 2438#endif
1da177e4
LT
2439}
2440
f06febc9
FM
2441/*
2442 * Thread group CPU time accounting.
2443 */
4cd4c1b4 2444void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2445void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2446
490dea45 2447static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2448{
4cd4c1b4
PZ
2449 sig->cputimer.cputime = INIT_CPUTIME;
2450 spin_lock_init(&sig->cputimer.lock);
2451 sig->cputimer.running = 0;
f06febc9
FM
2452}
2453
f06febc9
FM
2454static inline void thread_group_cputime_free(struct signal_struct *sig)
2455{
f06febc9
FM
2456}
2457
7bb44ade
RM
2458/*
2459 * Reevaluate whether the task has signals pending delivery.
2460 * Wake the task if so.
2461 * This is required every time the blocked sigset_t changes.
2462 * callers must hold sighand->siglock.
2463 */
2464extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2465extern void recalc_sigpending(void);
2466
2467extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2468
2469/*
2470 * Wrappers for p->thread_info->cpu access. No-op on UP.
2471 */
2472#ifdef CONFIG_SMP
2473
2474static inline unsigned int task_cpu(const struct task_struct *p)
2475{
a1261f54 2476 return task_thread_info(p)->cpu;
1da177e4
LT
2477}
2478
c65cc870 2479extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2480
2481#else
2482
2483static inline unsigned int task_cpu(const struct task_struct *p)
2484{
2485 return 0;
2486}
2487
2488static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2489{
2490}
2491
2492#endif /* CONFIG_SMP */
2493
1da177e4 2494extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4 2495
1a3c3034
IM
2496#ifdef CONFIG_TRACING
2497extern void
2498__trace_special(void *__tr, void *__data,
2499 unsigned long arg1, unsigned long arg2, unsigned long arg3);
1da177e4 2500#else
1a3c3034
IM
2501static inline void
2502__trace_special(void *__tr, void *__data,
2503 unsigned long arg1, unsigned long arg2, unsigned long arg3)
1da177e4 2504{
1da177e4
LT
2505}
2506#endif
2507
96f874e2
RR
2508extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2509extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2510
1da177e4
LT
2511extern void normalize_rt_tasks(void);
2512
052f1dc7 2513#ifdef CONFIG_GROUP_SCHED
9b5b7751 2514
4cf86d77 2515extern struct task_group init_task_group;
eff766a6
PZ
2516#ifdef CONFIG_USER_SCHED
2517extern struct task_group root_task_group;
6c415b92 2518extern void set_tg_uid(struct user_struct *user);
eff766a6 2519#endif
9b5b7751 2520
ec7dc8ac 2521extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2522extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2523extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2524#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2525extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2526extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2527#endif
2528#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2529extern int sched_group_set_rt_runtime(struct task_group *tg,
2530 long rt_runtime_us);
2531extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2532extern int sched_group_set_rt_period(struct task_group *tg,
2533 long rt_period_us);
2534extern long sched_group_rt_period(struct task_group *tg);
54e99124 2535extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2536#endif
9b5b7751
SV
2537#endif
2538
54e99124
DG
2539extern int task_can_switch_user(struct user_struct *up,
2540 struct task_struct *tsk);
2541
4b98d11b
AD
2542#ifdef CONFIG_TASK_XACCT
2543static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2544{
940389b8 2545 tsk->ioac.rchar += amt;
4b98d11b
AD
2546}
2547
2548static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2549{
940389b8 2550 tsk->ioac.wchar += amt;
4b98d11b
AD
2551}
2552
2553static inline void inc_syscr(struct task_struct *tsk)
2554{
940389b8 2555 tsk->ioac.syscr++;
4b98d11b
AD
2556}
2557
2558static inline void inc_syscw(struct task_struct *tsk)
2559{
940389b8 2560 tsk->ioac.syscw++;
4b98d11b
AD
2561}
2562#else
2563static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2564{
2565}
2566
2567static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2568{
2569}
2570
2571static inline void inc_syscr(struct task_struct *tsk)
2572{
2573}
2574
2575static inline void inc_syscw(struct task_struct *tsk)
2576{
2577}
2578#endif
2579
82455257
DH
2580#ifndef TASK_SIZE_OF
2581#define TASK_SIZE_OF(tsk) TASK_SIZE
2582#endif
2583
0793a61d
TG
2584/*
2585 * Call the function if the target task is executing on a CPU right now:
2586 */
2587extern void task_oncpu_function_call(struct task_struct *p,
2588 void (*func) (void *info), void *info);
2589
2590
cf475ad2
BS
2591#ifdef CONFIG_MM_OWNER
2592extern void mm_update_next_owner(struct mm_struct *mm);
2593extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2594#else
2595static inline void mm_update_next_owner(struct mm_struct *mm)
2596{
2597}
2598
2599static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2600{
2601}
2602#endif /* CONFIG_MM_OWNER */
2603
1da177e4
LT
2604#endif /* __KERNEL__ */
2605
2606#endif