HWPOISON: shmem: call set_page_dirty() with locked page
[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;
a63eaf34 103struct perf_counter_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);
dce48a84 143extern void calc_global_load(void);
23a185ca 144extern u64 cpu_nr_migrations(int cpu);
1da177e4 145
7e49fcce
SR
146extern unsigned long get_parent_ip(unsigned long addr);
147
43ae34cb
IM
148struct seq_file;
149struct cfs_rq;
4cf86d77 150struct task_group;
43ae34cb
IM
151#ifdef CONFIG_SCHED_DEBUG
152extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
153extern void proc_sched_set_task(struct task_struct *p);
154extern void
5cef9eca 155print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
156#else
157static inline void
158proc_sched_show_task(struct task_struct *p, struct seq_file *m)
159{
160}
161static inline void proc_sched_set_task(struct task_struct *p)
162{
163}
164static inline void
5cef9eca 165print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
166{
167}
168#endif
1da177e4 169
690229a0
IM
170extern unsigned long long time_sync_thresh;
171
4a8342d2
LT
172/*
173 * Task state bitmask. NOTE! These bits are also
174 * encoded in fs/proc/array.c: get_task_state().
175 *
176 * We have two separate sets of flags: task->state
177 * is about runnability, while task->exit_state are
178 * about the task exiting. Confusing, but this way
179 * modifying one set can't modify the other one by
180 * mistake.
181 */
1da177e4
LT
182#define TASK_RUNNING 0
183#define TASK_INTERRUPTIBLE 1
184#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
185#define __TASK_STOPPED 4
186#define __TASK_TRACED 8
4a8342d2
LT
187/* in tsk->exit_state */
188#define EXIT_ZOMBIE 16
189#define EXIT_DEAD 32
190/* in tsk->state again */
af927232 191#define TASK_DEAD 64
f021a3c2
MW
192#define TASK_WAKEKILL 128
193
194/* Convenience macros for the sake of set_task_state */
195#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
196#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
197#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 198
92a1f4bc
MW
199/* Convenience macros for the sake of wake_up */
200#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 201#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
202
203/* get_task_state() */
204#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
205 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
206 __TASK_TRACED)
92a1f4bc 207
f021a3c2
MW
208#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
209#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 210#define task_is_stopped_or_traced(task) \
f021a3c2 211 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 212#define task_contributes_to_load(task) \
e3c8ca83 213 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
6301cb95 214 (task->flags & PF_FREEZING) == 0)
1da177e4
LT
215
216#define __set_task_state(tsk, state_value) \
217 do { (tsk)->state = (state_value); } while (0)
218#define set_task_state(tsk, state_value) \
219 set_mb((tsk)->state, (state_value))
220
498d0c57
AM
221/*
222 * set_current_state() includes a barrier so that the write of current->state
223 * is correctly serialised wrt the caller's subsequent test of whether to
224 * actually sleep:
225 *
226 * set_current_state(TASK_UNINTERRUPTIBLE);
227 * if (do_i_need_to_sleep())
228 * schedule();
229 *
230 * If the caller does not need such serialisation then use __set_current_state()
231 */
1da177e4
LT
232#define __set_current_state(state_value) \
233 do { current->state = (state_value); } while (0)
234#define set_current_state(state_value) \
235 set_mb(current->state, (state_value))
236
237/* Task command name length */
238#define TASK_COMM_LEN 16
239
1da177e4
LT
240#include <linux/spinlock.h>
241
242/*
243 * This serializes "schedule()" and also protects
244 * the run-queue from deletions/modifications (but
245 * _adding_ to the beginning of the run-queue has
246 * a separate lock).
247 */
248extern rwlock_t tasklist_lock;
249extern spinlock_t mmlist_lock;
250
36c8b586 251struct task_struct;
1da177e4
LT
252
253extern void sched_init(void);
254extern void sched_init_smp(void);
2d07b255 255extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 256extern void init_idle(struct task_struct *idle, int cpu);
1df21055 257extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 258
017730c1 259extern int runqueue_is_locked(void);
ad474cac 260extern void task_rq_unlock_wait(struct task_struct *p);
017730c1 261
6a7b3dc3 262extern cpumask_var_t nohz_cpu_mask;
46cb4b7c
SS
263#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
264extern int select_nohz_load_balancer(int cpu);
eea08f32 265extern int get_nohz_load_balancer(void);
46cb4b7c
SS
266#else
267static inline int select_nohz_load_balancer(int cpu)
268{
269 return 0;
270}
271#endif
1da177e4 272
e59e2ae2 273/*
39bc89fd 274 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
275 */
276extern void show_state_filter(unsigned long state_filter);
277
278static inline void show_state(void)
279{
39bc89fd 280 show_state_filter(0);
e59e2ae2
IM
281}
282
1da177e4
LT
283extern void show_regs(struct pt_regs *);
284
285/*
286 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
287 * task), SP is the stack pointer of the first frame that should be shown in the back
288 * trace (or NULL if the entire call-chain of the task should be shown).
289 */
290extern void show_stack(struct task_struct *task, unsigned long *sp);
291
292void io_schedule(void);
293long io_schedule_timeout(long timeout);
294
295extern void cpu_init (void);
296extern void trap_init(void);
297extern void update_process_times(int user);
298extern void scheduler_tick(void);
299
82a1fcb9
IM
300extern void sched_show_task(struct task_struct *p);
301
8446f1d3 302#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 303extern void softlockup_tick(void);
8446f1d3 304extern void touch_softlockup_watchdog(void);
04c9167f 305extern void touch_all_softlockup_watchdogs(void);
baf48f65
MSB
306extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
307 struct file *filp, void __user *buffer,
308 size_t *lenp, loff_t *ppos);
9c44bc03 309extern unsigned int softlockup_panic;
9383d967 310extern int softlockup_thresh;
8446f1d3 311#else
6687a97d 312static inline void softlockup_tick(void)
8446f1d3
IM
313{
314}
8446f1d3
IM
315static inline void touch_softlockup_watchdog(void)
316{
317}
04c9167f
JF
318static inline void touch_all_softlockup_watchdogs(void)
319{
320}
8446f1d3
IM
321#endif
322
e162b39a
MSB
323#ifdef CONFIG_DETECT_HUNG_TASK
324extern unsigned int sysctl_hung_task_panic;
325extern unsigned long sysctl_hung_task_check_count;
326extern unsigned long sysctl_hung_task_timeout_secs;
327extern unsigned long sysctl_hung_task_warnings;
328extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
329 struct file *filp, void __user *buffer,
330 size_t *lenp, loff_t *ppos);
331#endif
8446f1d3 332
1da177e4
LT
333/* Attach to any functions which should be ignored in wchan output. */
334#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
335
336/* Linker adds these: start and end of __sched functions */
337extern char __sched_text_start[], __sched_text_end[];
338
1da177e4
LT
339/* Is this address in the __sched functions? */
340extern int in_sched_functions(unsigned long addr);
341
342#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 343extern signed long schedule_timeout(signed long timeout);
64ed93a2 344extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 345extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 346extern signed long schedule_timeout_uninterruptible(signed long timeout);
41719b03 347asmlinkage void __schedule(void);
1da177e4 348asmlinkage void schedule(void);
0d66bf6d 349extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 350
ab516013 351struct nsproxy;
acce292c 352struct user_namespace;
1da177e4 353
341c87bf
KH
354/*
355 * Default maximum number of active map areas, this limits the number of vmas
356 * per mm struct. Users can overwrite this number by sysctl but there is a
357 * problem.
358 *
359 * When a program's coredump is generated as ELF format, a section is created
360 * per a vma. In ELF, the number of sections is represented in unsigned short.
361 * This means the number of sections should be smaller than 65535 at coredump.
362 * Because the kernel adds some informative sections to a image of program at
363 * generating coredump, we need some margin. The number of extra sections is
364 * 1-3 now and depends on arch. We use "5" as safe margin, here.
365 */
366#define MAPCOUNT_ELF_CORE_MARGIN (5)
367#define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
368
369extern int sysctl_max_map_count;
370
371#include <linux/aio.h>
372
373extern unsigned long
374arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
375 unsigned long, unsigned long);
376extern unsigned long
377arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
378 unsigned long len, unsigned long pgoff,
379 unsigned long flags);
1363c3cd
WW
380extern void arch_unmap_area(struct mm_struct *, unsigned long);
381extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 382
f7d0b926 383#if USE_SPLIT_PTLOCKS
f412ac08
HD
384/*
385 * The mm counters are not protected by its page_table_lock,
386 * so must be incremented atomically.
387 */
d3cb4871
CL
388#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
389#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
390#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
391#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
392#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
f412ac08 393
f7d0b926 394#else /* !USE_SPLIT_PTLOCKS */
f412ac08
HD
395/*
396 * The mm counters are protected by its page_table_lock,
397 * so can be incremented directly.
398 */
1da177e4
LT
399#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
400#define get_mm_counter(mm, member) ((mm)->_##member)
401#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
402#define inc_mm_counter(mm, member) (mm)->_##member++
403#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08 404
f7d0b926 405#endif /* !USE_SPLIT_PTLOCKS */
4294621f 406
f412ac08
HD
407#define get_mm_rss(mm) \
408 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
409#define update_hiwater_rss(mm) do { \
410 unsigned long _rss = get_mm_rss(mm); \
411 if ((mm)->hiwater_rss < _rss) \
412 (mm)->hiwater_rss = _rss; \
413} while (0)
414#define update_hiwater_vm(mm) do { \
415 if ((mm)->hiwater_vm < (mm)->total_vm) \
416 (mm)->hiwater_vm = (mm)->total_vm; \
417} while (0)
418
9de1581e
ON
419static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
420{
421 return max(mm->hiwater_rss, get_mm_rss(mm));
422}
423
424static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
425{
426 return max(mm->hiwater_vm, mm->total_vm);
427}
901608d9 428
6c5d5238
KH
429extern void set_dumpable(struct mm_struct *mm, int value);
430extern int get_dumpable(struct mm_struct *mm);
431
432/* mm flags */
3cb4a0bb 433/* dumpable bits */
6c5d5238
KH
434#define MMF_DUMPABLE 0 /* core dump is permitted */
435#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
3cb4a0bb
KH
436#define MMF_DUMPABLE_BITS 2
437
438/* coredump filter bits */
439#define MMF_DUMP_ANON_PRIVATE 2
440#define MMF_DUMP_ANON_SHARED 3
441#define MMF_DUMP_MAPPED_PRIVATE 4
442#define MMF_DUMP_MAPPED_SHARED 5
82df3973 443#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
444#define MMF_DUMP_HUGETLB_PRIVATE 7
445#define MMF_DUMP_HUGETLB_SHARED 8
3cb4a0bb 446#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 447#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
448#define MMF_DUMP_FILTER_MASK \
449 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
450#define MMF_DUMP_FILTER_DEFAULT \
e575f111 451 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
452 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
453
454#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
455# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
456#else
457# define MMF_DUMP_MASK_DEFAULT_ELF 0
458#endif
6c5d5238 459
1da177e4
LT
460struct sighand_struct {
461 atomic_t count;
462 struct k_sigaction action[_NSIG];
463 spinlock_t siglock;
b8fceee1 464 wait_queue_head_t signalfd_wqh;
1da177e4
LT
465};
466
0e464814 467struct pacct_struct {
f6ec29a4
KK
468 int ac_flag;
469 long ac_exitcode;
0e464814 470 unsigned long ac_mem;
77787bfb
KK
471 cputime_t ac_utime, ac_stime;
472 unsigned long ac_minflt, ac_majflt;
0e464814
KK
473};
474
f06febc9
FM
475/**
476 * struct task_cputime - collected CPU time counts
477 * @utime: time spent in user mode, in &cputime_t units
478 * @stime: time spent in kernel mode, in &cputime_t units
479 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 480 *
f06febc9
FM
481 * This structure groups together three kinds of CPU time that are
482 * tracked for threads and thread groups. Most things considering
483 * CPU time want to group these counts together and treat all three
484 * of them in parallel.
485 */
486struct task_cputime {
487 cputime_t utime;
488 cputime_t stime;
489 unsigned long long sum_exec_runtime;
490};
491/* Alternate field names when used to cache expirations. */
492#define prof_exp stime
493#define virt_exp utime
494#define sched_exp sum_exec_runtime
495
4cd4c1b4
PZ
496#define INIT_CPUTIME \
497 (struct task_cputime) { \
498 .utime = cputime_zero, \
499 .stime = cputime_zero, \
500 .sum_exec_runtime = 0, \
501 }
502
c99e6efe
PZ
503/*
504 * Disable preemption until the scheduler is running.
505 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
506 *
507 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
508 * before the scheduler is active -- see should_resched().
c99e6efe 509 */
d86ee480 510#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 511
f06febc9 512/**
4cd4c1b4
PZ
513 * struct thread_group_cputimer - thread group interval timer counts
514 * @cputime: thread group interval timers.
515 * @running: non-zero when there are timers running and
516 * @cputime receives updates.
517 * @lock: lock for fields in this struct.
f06febc9
FM
518 *
519 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 520 * used for thread group CPU timer calculations.
f06febc9 521 */
4cd4c1b4
PZ
522struct thread_group_cputimer {
523 struct task_cputime cputime;
524 int running;
525 spinlock_t lock;
f06febc9 526};
f06febc9 527
1da177e4
LT
528/*
529 * NOTE! "signal_struct" does not have it's own
530 * locking, because a shared signal_struct always
531 * implies a shared sighand_struct, so locking
532 * sighand_struct is always a proper superset of
533 * the locking of signal_struct.
534 */
535struct signal_struct {
536 atomic_t count;
537 atomic_t live;
538
539 wait_queue_head_t wait_chldexit; /* for wait4() */
540
541 /* current thread group signal load-balancing target: */
36c8b586 542 struct task_struct *curr_target;
1da177e4
LT
543
544 /* shared signal handling: */
545 struct sigpending shared_pending;
546
547 /* thread group exit support */
548 int group_exit_code;
549 /* overloaded:
550 * - notify group_exit_task when ->count is equal to notify_count
551 * - everyone except group_exit_task is stopped during signal delivery
552 * of fatal signals, group_exit_task processes the signal.
553 */
1da177e4 554 int notify_count;
07dd20e0 555 struct task_struct *group_exit_task;
1da177e4
LT
556
557 /* thread group stop support, overloads group_exit_code too */
558 int group_stop_count;
559 unsigned int flags; /* see SIGNAL_* flags below */
560
561 /* POSIX.1b Interval Timers */
562 struct list_head posix_timers;
563
564 /* ITIMER_REAL timer for the process */
2ff678b8 565 struct hrtimer real_timer;
fea9d175 566 struct pid *leader_pid;
2ff678b8 567 ktime_t it_real_incr;
1da177e4
LT
568
569 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
570 cputime_t it_prof_expires, it_virt_expires;
571 cputime_t it_prof_incr, it_virt_incr;
572
f06febc9 573 /*
4cd4c1b4
PZ
574 * Thread group totals for process CPU timers.
575 * See thread_group_cputimer(), et al, for details.
f06febc9 576 */
4cd4c1b4 577 struct thread_group_cputimer cputimer;
f06febc9
FM
578
579 /* Earliest-expiration cache. */
580 struct task_cputime cputime_expires;
581
582 struct list_head cpu_timers[3];
583
ab521dc0 584 struct pid *tty_old_pgrp;
1ec320af 585
1da177e4
LT
586 /* boolean value for session group leader */
587 int leader;
588
589 struct tty_struct *tty; /* NULL if no tty */
590
591 /*
592 * Cumulative resource counters for dead threads in the group,
593 * and for reaped dead child processes forked by this group.
594 * Live threads maintain their own counters and add to these
595 * in __exit_signal, except for the group leader.
596 */
32bd671d 597 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
598 cputime_t gtime;
599 cputime_t cgtime;
1da177e4
LT
600 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
601 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 602 unsigned long inblock, oublock, cinblock, coublock;
940389b8 603 struct task_io_accounting ioac;
1da177e4 604
32bd671d
PZ
605 /*
606 * Cumulative ns of schedule CPU time fo dead threads in the
607 * group, not including a zombie group leader, (This only differs
608 * from jiffies_to_ns(utime + stime) if sched_clock uses something
609 * other than jiffies.)
610 */
611 unsigned long long sum_sched_runtime;
612
1da177e4
LT
613 /*
614 * We don't bother to synchronize most readers of this at all,
615 * because there is no reader checking a limit that actually needs
616 * to get both rlim_cur and rlim_max atomically, and either one
617 * alone is a single word that can safely be read normally.
618 * getrlimit/setrlimit use task_lock(current->group_leader) to
619 * protect this instead of the siglock, because they really
620 * have no need to disable irqs.
621 */
622 struct rlimit rlim[RLIM_NLIMITS];
623
0e464814
KK
624#ifdef CONFIG_BSD_PROCESS_ACCT
625 struct pacct_struct pacct; /* per-process accounting information */
626#endif
ad4ecbcb 627#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
628 struct taskstats *stats;
629#endif
522ed776
MT
630#ifdef CONFIG_AUDIT
631 unsigned audit_tty;
632 struct tty_audit_buf *tty_audit_buf;
633#endif
1da177e4
LT
634};
635
4866cde0
NP
636/* Context switch must be unlocked if interrupts are to be enabled */
637#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
638# define __ARCH_WANT_UNLOCKED_CTXSW
639#endif
640
1da177e4
LT
641/*
642 * Bits in flags field of signal_struct.
643 */
644#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
645#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
646#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
647#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
648/*
649 * Pending notifications to parent.
650 */
651#define SIGNAL_CLD_STOPPED 0x00000010
652#define SIGNAL_CLD_CONTINUED 0x00000020
653#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 654
fae5fa44
ON
655#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
656
ed5d2cac
ON
657/* If true, all threads except ->group_exit_task have pending SIGKILL */
658static inline int signal_group_exit(const struct signal_struct *sig)
659{
660 return (sig->flags & SIGNAL_GROUP_EXIT) ||
661 (sig->group_exit_task != NULL);
662}
663
1da177e4
LT
664/*
665 * Some day this will be a full-fledged user tracking system..
666 */
667struct user_struct {
668 atomic_t __count; /* reference count */
669 atomic_t processes; /* How many processes does this user have? */
670 atomic_t files; /* How many open files does this user have? */
671 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 672#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
673 atomic_t inotify_watches; /* How many inotify watches does this user have? */
674 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
675#endif
7ef9964e 676#ifdef CONFIG_EPOLL
7ef9964e
DL
677 atomic_t epoll_watches; /* The number of file descriptors currently watched */
678#endif
970a8645 679#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
680 /* protected by mq_lock */
681 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 682#endif
1da177e4
LT
683 unsigned long locked_shm; /* How many pages of mlocked shm ? */
684
685#ifdef CONFIG_KEYS
686 struct key *uid_keyring; /* UID specific keyring */
687 struct key *session_keyring; /* UID's default session keyring */
688#endif
689
690 /* Hash table maintenance information */
735de223 691 struct hlist_node uidhash_node;
1da177e4 692 uid_t uid;
18b6e041 693 struct user_namespace *user_ns;
24e377a8 694
052f1dc7 695#ifdef CONFIG_USER_SCHED
4cf86d77 696 struct task_group *tg;
b1a8c172 697#ifdef CONFIG_SYSFS
eb41d946 698 struct kobject kobj;
3959214f 699 struct delayed_work work;
24e377a8 700#endif
b1a8c172 701#endif
789f90fc
PZ
702
703#ifdef CONFIG_PERF_COUNTERS
704 atomic_long_t locked_vm;
705#endif
1da177e4
LT
706};
707
eb41d946 708extern int uids_sysfs_init(void);
5cb350ba 709
1da177e4
LT
710extern struct user_struct *find_user(uid_t);
711
712extern struct user_struct root_user;
713#define INIT_USER (&root_user)
714
b6dff3ec 715
1da177e4
LT
716struct backing_dev_info;
717struct reclaim_state;
718
52f17b6c 719#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
720struct sched_info {
721 /* cumulative counters */
2d72376b 722 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 723 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
724
725 /* timestamps */
172ba844
BS
726 unsigned long long last_arrival,/* when we last ran on a cpu */
727 last_queued; /* when we were last queued to run */
b8efb561
IM
728#ifdef CONFIG_SCHEDSTATS
729 /* BKL stats */
480b9434 730 unsigned int bkl_count;
b8efb561 731#endif
1da177e4 732};
52f17b6c 733#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 734
ca74e92b
SN
735#ifdef CONFIG_TASK_DELAY_ACCT
736struct task_delay_info {
737 spinlock_t lock;
738 unsigned int flags; /* Private per-task flags */
739
740 /* For each stat XXX, add following, aligned appropriately
741 *
742 * struct timespec XXX_start, XXX_end;
743 * u64 XXX_delay;
744 * u32 XXX_count;
745 *
746 * Atomicity of updates to XXX_delay, XXX_count protected by
747 * single lock above (split into XXX_lock if contention is an issue).
748 */
0ff92245
SN
749
750 /*
751 * XXX_count is incremented on every XXX operation, the delay
752 * associated with the operation is added to XXX_delay.
753 * XXX_delay contains the accumulated delay time in nanoseconds.
754 */
755 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
756 u64 blkio_delay; /* wait for sync block io completion */
757 u64 swapin_delay; /* wait for swapin block io completion */
758 u32 blkio_count; /* total count of the number of sync block */
759 /* io operations performed */
760 u32 swapin_count; /* total count of the number of swapin block */
761 /* io operations performed */
873b4771
KK
762
763 struct timespec freepages_start, freepages_end;
764 u64 freepages_delay; /* wait for memory reclaim */
765 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 766};
52f17b6c
CS
767#endif /* CONFIG_TASK_DELAY_ACCT */
768
769static inline int sched_info_on(void)
770{
771#ifdef CONFIG_SCHEDSTATS
772 return 1;
773#elif defined(CONFIG_TASK_DELAY_ACCT)
774 extern int delayacct_on;
775 return delayacct_on;
776#else
777 return 0;
ca74e92b 778#endif
52f17b6c 779}
ca74e92b 780
d15bcfdb
IM
781enum cpu_idle_type {
782 CPU_IDLE,
783 CPU_NOT_IDLE,
784 CPU_NEWLY_IDLE,
785 CPU_MAX_IDLE_TYPES
1da177e4
LT
786};
787
788/*
789 * sched-domains (multiprocessor balancing) declarations:
790 */
9aa7b369
IM
791
792/*
793 * Increase resolution of nice-level calculations:
794 */
795#define SCHED_LOAD_SHIFT 10
796#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
797
f8700df7 798#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 799
2dd73a4f 800#ifdef CONFIG_SMP
b5d978e0
PZ
801#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
802#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
803#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
804#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
805#define SD_WAKE_IDLE 0x0010 /* Wake to idle CPU on task wakeup */
806#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
807#define SD_WAKE_BALANCE 0x0040 /* Perform balancing at task wakeup */
808#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
809#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
810#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
811#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
812#define SD_WAKE_IDLE_FAR 0x0800 /* Gain latency sacrificing cache hit */
813#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 814
afb8a9b7
GS
815enum powersavings_balance_level {
816 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
817 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
818 * first for long running threads
819 */
820 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
821 * cpu package for power savings
822 */
823 MAX_POWERSAVINGS_BALANCE_LEVELS
824};
89c4710e 825
716707b2 826extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 827
716707b2
VS
828static inline int sd_balance_for_mc_power(void)
829{
830 if (sched_smt_power_savings)
831 return SD_POWERSAVINGS_BALANCE;
5c45bf27 832
b5d978e0 833 return SD_PREFER_SIBLING;
716707b2 834}
89c4710e 835
716707b2
VS
836static inline int sd_balance_for_package_power(void)
837{
838 if (sched_mc_power_savings | sched_smt_power_savings)
839 return SD_POWERSAVINGS_BALANCE;
840
b5d978e0 841 return SD_PREFER_SIBLING;
716707b2 842}
5c45bf27 843
100fdaee
VS
844/*
845 * Optimise SD flags for power savings:
846 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
847 * Keep default SD flags if sched_{smt,mc}_power_saving=0
848 */
849
850static inline int sd_power_saving_flags(void)
851{
852 if (sched_mc_power_savings | sched_smt_power_savings)
853 return SD_BALANCE_NEWIDLE;
854
855 return 0;
856}
1da177e4
LT
857
858struct sched_group {
859 struct sched_group *next; /* Must be a circular list */
1da177e4
LT
860
861 /*
862 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 863 * single CPU.
5517d86b 864 */
18a3885f 865 unsigned int cpu_power;
6c99e9ad 866
4200efd9
IM
867 /*
868 * The CPUs this group covers.
869 *
870 * NOTE: this field is variable length. (Allocated dynamically
871 * by attaching extra space to the end of the structure,
872 * depending on how many CPUs the kernel has booted up with)
873 *
874 * It is also be embedded into static data structures at build
875 * time. (See 'struct static_sched_group' in kernel/sched.c)
876 */
877 unsigned long cpumask[0];
1da177e4
LT
878};
879
758b2cdc
RR
880static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
881{
6c99e9ad 882 return to_cpumask(sg->cpumask);
758b2cdc
RR
883}
884
1d3504fc
HS
885enum sched_domain_level {
886 SD_LV_NONE = 0,
887 SD_LV_SIBLING,
888 SD_LV_MC,
889 SD_LV_CPU,
890 SD_LV_NODE,
891 SD_LV_ALLNODES,
892 SD_LV_MAX
893};
894
895struct sched_domain_attr {
896 int relax_domain_level;
897};
898
899#define SD_ATTR_INIT (struct sched_domain_attr) { \
900 .relax_domain_level = -1, \
901}
902
1da177e4
LT
903struct sched_domain {
904 /* These fields must be setup */
905 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 906 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 907 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
908 unsigned long min_interval; /* Minimum balance interval ms */
909 unsigned long max_interval; /* Maximum balance interval ms */
910 unsigned int busy_factor; /* less balancing by factor if busy */
911 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 912 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
913 unsigned int busy_idx;
914 unsigned int idle_idx;
915 unsigned int newidle_idx;
916 unsigned int wake_idx;
147cbb4b 917 unsigned int forkexec_idx;
a52bfd73 918 unsigned int smt_gain;
1da177e4 919 int flags; /* See SD_* */
1d3504fc 920 enum sched_domain_level level;
1da177e4
LT
921
922 /* Runtime fields. */
923 unsigned long last_balance; /* init to jiffies. units in jiffies */
924 unsigned int balance_interval; /* initialise to 1. units in ms. */
925 unsigned int nr_balance_failed; /* initialise to 0 */
926
2398f2c6
PZ
927 u64 last_update;
928
1da177e4
LT
929#ifdef CONFIG_SCHEDSTATS
930 /* load_balance() stats */
480b9434
KC
931 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
932 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
933 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
934 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
935 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
936 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
937 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
938 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
939
940 /* Active load balancing */
480b9434
KC
941 unsigned int alb_count;
942 unsigned int alb_failed;
943 unsigned int alb_pushed;
1da177e4 944
68767a0a 945 /* SD_BALANCE_EXEC stats */
480b9434
KC
946 unsigned int sbe_count;
947 unsigned int sbe_balanced;
948 unsigned int sbe_pushed;
1da177e4 949
68767a0a 950 /* SD_BALANCE_FORK stats */
480b9434
KC
951 unsigned int sbf_count;
952 unsigned int sbf_balanced;
953 unsigned int sbf_pushed;
68767a0a 954
1da177e4 955 /* try_to_wake_up() stats */
480b9434
KC
956 unsigned int ttwu_wake_remote;
957 unsigned int ttwu_move_affine;
958 unsigned int ttwu_move_balance;
1da177e4 959#endif
a5d8c348
IM
960#ifdef CONFIG_SCHED_DEBUG
961 char *name;
962#endif
6c99e9ad 963
4200efd9
IM
964 /*
965 * Span of all CPUs in this domain.
966 *
967 * NOTE: this field is variable length. (Allocated dynamically
968 * by attaching extra space to the end of the structure,
969 * depending on how many CPUs the kernel has booted up with)
970 *
971 * It is also be embedded into static data structures at build
972 * time. (See 'struct static_sched_domain' in kernel/sched.c)
973 */
974 unsigned long span[0];
1da177e4
LT
975};
976
758b2cdc
RR
977static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
978{
6c99e9ad 979 return to_cpumask(sd->span);
758b2cdc
RR
980}
981
96f874e2 982extern void partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1d3504fc 983 struct sched_domain_attr *dattr_new);
029190c5 984
06aaf76a
IM
985/* Test a flag in parent sched domain */
986static inline int test_sd_parent(struct sched_domain *sd, int flag)
987{
988 if (sd->parent && (sd->parent->flags & flag))
989 return 1;
990
991 return 0;
992}
029190c5 993
1b427c15 994#else /* CONFIG_SMP */
1da177e4 995
1b427c15 996struct sched_domain_attr;
d02c7a8c 997
1b427c15 998static inline void
96f874e2 999partition_sched_domains(int ndoms_new, struct cpumask *doms_new,
1b427c15
IM
1000 struct sched_domain_attr *dattr_new)
1001{
d02c7a8c 1002}
1b427c15 1003#endif /* !CONFIG_SMP */
1da177e4
LT
1004
1005struct io_context; /* See blkdev.h */
1da177e4 1006
1da177e4 1007
383f2835 1008#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1009extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1010#else
1011static inline void prefetch_stack(struct task_struct *t) { }
1012#endif
1da177e4
LT
1013
1014struct audit_context; /* See audit.c */
1015struct mempolicy;
b92ce558 1016struct pipe_inode_info;
4865ecf1 1017struct uts_namespace;
1da177e4 1018
20b8a59f
IM
1019struct rq;
1020struct sched_domain;
1021
1022struct sched_class {
5522d5d5 1023 const struct sched_class *next;
20b8a59f 1024
fd390f6a 1025 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
f02231e5 1026 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
4530d7ab 1027 void (*yield_task) (struct rq *rq);
20b8a59f 1028
15afe09b 1029 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int sync);
20b8a59f 1030
fb8d4724 1031 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1032 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1033
681f3e68 1034#ifdef CONFIG_SMP
4ce72a2c
LZ
1035 int (*select_task_rq)(struct task_struct *p, int sync);
1036
43010659 1037 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
e1d1484f 1038 struct rq *busiest, unsigned long max_load_move,
20b8a59f 1039 struct sched_domain *sd, enum cpu_idle_type idle,
a4ac01c3 1040 int *all_pinned, int *this_best_prio);
20b8a59f 1041
e1d1484f
PW
1042 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1043 struct rq *busiest, struct sched_domain *sd,
1044 enum cpu_idle_type idle);
9a897c5a
SR
1045 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1046 void (*post_schedule) (struct rq *this_rq);
1047 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1048
cd8ba7cd 1049 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1050 const struct cpumask *newmask);
57d885fe 1051
1f11eb6a
GH
1052 void (*rq_online)(struct rq *rq);
1053 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1054#endif
1055
1056 void (*set_curr_task) (struct rq *rq);
1057 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
1058 void (*task_new) (struct rq *rq, struct task_struct *p);
cb469845
SR
1059
1060 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1061 int running);
1062 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1063 int running);
1064 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1065 int oldprio, int running);
810b3817
PZ
1066
1067#ifdef CONFIG_FAIR_GROUP_SCHED
1068 void (*moved_group) (struct task_struct *p);
1069#endif
20b8a59f
IM
1070};
1071
1072struct load_weight {
1073 unsigned long weight, inv_weight;
1074};
1075
1076/*
1077 * CFS stats for a schedulable entity (task, task-group etc)
1078 *
1079 * Current field usage histogram:
1080 *
1081 * 4 se->block_start
1082 * 4 se->run_node
1083 * 4 se->sleep_start
20b8a59f 1084 * 6 se->load.weight
20b8a59f
IM
1085 */
1086struct sched_entity {
20b8a59f
IM
1087 struct load_weight load; /* for load-balancing */
1088 struct rb_node run_node;
4a55bd5e 1089 struct list_head group_node;
20b8a59f
IM
1090 unsigned int on_rq;
1091
94c18227
IM
1092 u64 exec_start;
1093 u64 sum_exec_runtime;
e9acbff6 1094 u64 vruntime;
f6cf891c 1095 u64 prev_sum_exec_runtime;
94c18227 1096
4ae7d5ce
IM
1097 u64 last_wakeup;
1098 u64 avg_overlap;
1099
6c594c21
IM
1100 u64 nr_migrations;
1101
34cb6135
IM
1102 u64 start_runtime;
1103 u64 avg_wakeup;
34cb6135 1104
94c18227 1105#ifdef CONFIG_SCHEDSTATS
20b8a59f 1106 u64 wait_start;
94c18227 1107 u64 wait_max;
6d082592
AV
1108 u64 wait_count;
1109 u64 wait_sum;
8f0dfc34
AV
1110 u64 iowait_count;
1111 u64 iowait_sum;
94c18227 1112
20b8a59f 1113 u64 sleep_start;
20b8a59f 1114 u64 sleep_max;
94c18227
IM
1115 s64 sum_sleep_runtime;
1116
1117 u64 block_start;
20b8a59f
IM
1118 u64 block_max;
1119 u64 exec_max;
eba1ed4b 1120 u64 slice_max;
cc367732 1121
cc367732
IM
1122 u64 nr_migrations_cold;
1123 u64 nr_failed_migrations_affine;
1124 u64 nr_failed_migrations_running;
1125 u64 nr_failed_migrations_hot;
1126 u64 nr_forced_migrations;
1127 u64 nr_forced2_migrations;
1128
1129 u64 nr_wakeups;
1130 u64 nr_wakeups_sync;
1131 u64 nr_wakeups_migrate;
1132 u64 nr_wakeups_local;
1133 u64 nr_wakeups_remote;
1134 u64 nr_wakeups_affine;
1135 u64 nr_wakeups_affine_attempts;
1136 u64 nr_wakeups_passive;
1137 u64 nr_wakeups_idle;
94c18227
IM
1138#endif
1139
20b8a59f
IM
1140#ifdef CONFIG_FAIR_GROUP_SCHED
1141 struct sched_entity *parent;
1142 /* rq on which this entity is (to be) queued: */
1143 struct cfs_rq *cfs_rq;
1144 /* rq "owned" by this entity/group: */
1145 struct cfs_rq *my_q;
1146#endif
1147};
70b97a7f 1148
fa717060
PZ
1149struct sched_rt_entity {
1150 struct list_head run_list;
78f2c7db 1151 unsigned long timeout;
bee367ed 1152 unsigned int time_slice;
6f505b16
PZ
1153 int nr_cpus_allowed;
1154
58d6c2d7 1155 struct sched_rt_entity *back;
052f1dc7 1156#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1157 struct sched_rt_entity *parent;
1158 /* rq on which this entity is (to be) queued: */
1159 struct rt_rq *rt_rq;
1160 /* rq "owned" by this entity/group: */
1161 struct rt_rq *my_q;
1162#endif
fa717060
PZ
1163};
1164
86848966
PM
1165struct rcu_node;
1166
1da177e4
LT
1167struct task_struct {
1168 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1169 void *stack;
1da177e4 1170 atomic_t usage;
97dc32cd
WC
1171 unsigned int flags; /* per process flags, defined below */
1172 unsigned int ptrace;
1da177e4 1173
36772092 1174 int lock_depth; /* BKL lock depth */
1da177e4 1175
2dd73a4f
PW
1176#ifdef CONFIG_SMP
1177#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1178 int oncpu;
1179#endif
2dd73a4f 1180#endif
50e645a8 1181
b29739f9 1182 int prio, static_prio, normal_prio;
c7aceaba 1183 unsigned int rt_priority;
5522d5d5 1184 const struct sched_class *sched_class;
20b8a59f 1185 struct sched_entity se;
fa717060 1186 struct sched_rt_entity rt;
1da177e4 1187
e107be36
AK
1188#ifdef CONFIG_PREEMPT_NOTIFIERS
1189 /* list of struct preempt_notifier: */
1190 struct hlist_head preempt_notifiers;
1191#endif
1192
18796aa0
AD
1193 /*
1194 * fpu_counter contains the number of consecutive context switches
1195 * that the FPU is used. If this is over a threshold, the lazy fpu
1196 * saving becomes unlazy to save the trap. This is an unsigned char
1197 * so that after 256 times the counter wraps and the behavior turns
1198 * lazy again; this to deal with bursty apps that only use FPU for
1199 * a short time
1200 */
1201 unsigned char fpu_counter;
0753ba01 1202 s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
6c5c9341 1203#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1204 unsigned int btrace_seq;
6c5c9341 1205#endif
1da177e4 1206
97dc32cd 1207 unsigned int policy;
1da177e4 1208 cpumask_t cpus_allowed;
1da177e4 1209
f41d911f 1210#ifdef CONFIG_TREE_PREEMPT_RCU
e260be67 1211 int rcu_read_lock_nesting;
f41d911f 1212 char rcu_read_unlock_special;
86848966 1213 struct rcu_node *rcu_blocked_node;
f41d911f
PM
1214 struct list_head rcu_node_entry;
1215#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
e260be67 1216
52f17b6c 1217#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1218 struct sched_info sched_info;
1219#endif
1220
1221 struct list_head tasks;
917b627d 1222 struct plist_node pushable_tasks;
1da177e4
LT
1223
1224 struct mm_struct *mm, *active_mm;
1225
1226/* task state */
1227 struct linux_binfmt *binfmt;
97dc32cd 1228 int exit_state;
1da177e4
LT
1229 int exit_code, exit_signal;
1230 int pdeath_signal; /* The signal sent when the parent dies */
1231 /* ??? */
97dc32cd 1232 unsigned int personality;
1da177e4 1233 unsigned did_exec:1;
f9ce1f1c
KT
1234 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1235 * execve */
8f0dfc34
AV
1236 unsigned in_iowait:1;
1237
ca94c442
LP
1238
1239 /* Revert to default priority/policy when forking */
1240 unsigned sched_reset_on_fork:1;
1241
1da177e4
LT
1242 pid_t pid;
1243 pid_t tgid;
0a425405 1244
1314562a 1245#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1246 /* Canary value for the -fstack-protector gcc feature */
1247 unsigned long stack_canary;
1314562a 1248#endif
e0032087 1249
1da177e4
LT
1250 /*
1251 * pointers to (original) parent process, youngest child, younger sibling,
1252 * older sibling, respectively. (p->father can be replaced with
f470021a 1253 * p->real_parent->pid)
1da177e4 1254 */
f470021a
RM
1255 struct task_struct *real_parent; /* real parent process */
1256 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1257 /*
f470021a 1258 * children/sibling forms the list of my natural children
1da177e4
LT
1259 */
1260 struct list_head children; /* list of my children */
1261 struct list_head sibling; /* linkage in my parent's children list */
1262 struct task_struct *group_leader; /* threadgroup leader */
1263
f470021a
RM
1264 /*
1265 * ptraced is the list of tasks this task is using ptrace on.
1266 * This includes both natural children and PTRACE_ATTACH targets.
1267 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1268 */
1269 struct list_head ptraced;
1270 struct list_head ptrace_entry;
1271
ca0002a1
MM
1272 /*
1273 * This is the tracer handle for the ptrace BTS extension.
1274 * This field actually belongs to the ptracer task.
1275 */
e2b371f0 1276 struct bts_context *bts;
ca0002a1 1277
1da177e4 1278 /* PID/PID hash table linkage. */
92476d7f 1279 struct pid_link pids[PIDTYPE_MAX];
47e65328 1280 struct list_head thread_group;
1da177e4
LT
1281
1282 struct completion *vfork_done; /* for vfork() */
1283 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1284 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1285
c66f08be 1286 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1287 cputime_t gtime;
9301899b 1288 cputime_t prev_utime, prev_stime;
1da177e4 1289 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1290 struct timespec start_time; /* monotonic time */
1291 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1292/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1293 unsigned long min_flt, maj_flt;
1294
f06febc9 1295 struct task_cputime cputime_expires;
1da177e4
LT
1296 struct list_head cpu_timers[3];
1297
1298/* process credentials */
3b11a1de
DH
1299 const struct cred *real_cred; /* objective and real subjective task
1300 * credentials (COW) */
1301 const struct cred *cred; /* effective (overridable) subjective task
1302 * credentials (COW) */
5e751e99
DH
1303 struct mutex cred_guard_mutex; /* guard against foreign influences on
1304 * credential calculations
1305 * (notably. ptrace) */
ee18d64c 1306 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1307
36772092
PBG
1308 char comm[TASK_COMM_LEN]; /* executable name excluding path
1309 - access with [gs]et_task_comm (which lock
1310 it with task_lock())
1311 - initialized normally by flush_old_exec */
1da177e4
LT
1312/* file system info */
1313 int link_count, total_link_count;
3d5b6fcc 1314#ifdef CONFIG_SYSVIPC
1da177e4
LT
1315/* ipc stuff */
1316 struct sysv_sem sysvsem;
3d5b6fcc 1317#endif
e162b39a 1318#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1319/* hung task detection */
82a1fcb9
IM
1320 unsigned long last_switch_count;
1321#endif
1da177e4
LT
1322/* CPU-specific state of this task */
1323 struct thread_struct thread;
1324/* filesystem information */
1325 struct fs_struct *fs;
1326/* open file information */
1327 struct files_struct *files;
1651e14e 1328/* namespaces */
ab516013 1329 struct nsproxy *nsproxy;
1da177e4
LT
1330/* signal handlers */
1331 struct signal_struct *signal;
1332 struct sighand_struct *sighand;
1333
1334 sigset_t blocked, real_blocked;
f3de272b 1335 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1336 struct sigpending pending;
1337
1338 unsigned long sas_ss_sp;
1339 size_t sas_ss_size;
1340 int (*notifier)(void *priv);
1341 void *notifier_data;
1342 sigset_t *notifier_mask;
1da177e4 1343 struct audit_context *audit_context;
bfef93a5
AV
1344#ifdef CONFIG_AUDITSYSCALL
1345 uid_t loginuid;
4746ec5b 1346 unsigned int sessionid;
bfef93a5 1347#endif
1da177e4
LT
1348 seccomp_t seccomp;
1349
1350/* Thread group tracking */
1351 u32 parent_exec_id;
1352 u32 self_exec_id;
58568d2a
MX
1353/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1354 * mempolicy */
1da177e4 1355 spinlock_t alloc_lock;
1da177e4 1356
3aa551c9
TG
1357#ifdef CONFIG_GENERIC_HARDIRQS
1358 /* IRQ handler threads */
1359 struct irqaction *irqaction;
1360#endif
1361
b29739f9
IM
1362 /* Protection of the PI data structures: */
1363 spinlock_t pi_lock;
1364
23f78d4a
IM
1365#ifdef CONFIG_RT_MUTEXES
1366 /* PI waiters blocked on a rt_mutex held by this task */
1367 struct plist_head pi_waiters;
1368 /* Deadlock detection and priority inheritance handling */
1369 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1370#endif
1371
408894ee
IM
1372#ifdef CONFIG_DEBUG_MUTEXES
1373 /* mutex deadlock detection */
1374 struct mutex_waiter *blocked_on;
1375#endif
de30a2b3
IM
1376#ifdef CONFIG_TRACE_IRQFLAGS
1377 unsigned int irq_events;
1378 int hardirqs_enabled;
1379 unsigned long hardirq_enable_ip;
1380 unsigned int hardirq_enable_event;
1381 unsigned long hardirq_disable_ip;
1382 unsigned int hardirq_disable_event;
1383 int softirqs_enabled;
1384 unsigned long softirq_disable_ip;
1385 unsigned int softirq_disable_event;
1386 unsigned long softirq_enable_ip;
1387 unsigned int softirq_enable_event;
1388 int hardirq_context;
1389 int softirq_context;
1390#endif
fbb9ce95 1391#ifdef CONFIG_LOCKDEP
bdb9441e 1392# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1393 u64 curr_chain_key;
1394 int lockdep_depth;
fbb9ce95 1395 unsigned int lockdep_recursion;
c7aceaba 1396 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1397 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1398#endif
408894ee 1399
1da177e4
LT
1400/* journalling filesystem info */
1401 void *journal_info;
1402
d89d8796
NB
1403/* stacked block device info */
1404 struct bio *bio_list, **bio_tail;
1405
1da177e4
LT
1406/* VM state */
1407 struct reclaim_state *reclaim_state;
1408
1da177e4
LT
1409 struct backing_dev_info *backing_dev_info;
1410
1411 struct io_context *io_context;
1412
1413 unsigned long ptrace_message;
1414 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1415 struct task_io_accounting ioac;
8f0ab514 1416#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1417 u64 acct_rss_mem1; /* accumulated rss usage */
1418 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1419 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1420#endif
1421#ifdef CONFIG_CPUSETS
58568d2a 1422 nodemask_t mems_allowed; /* Protected by alloc_lock */
825a46af 1423 int cpuset_mem_spread_rotor;
1da177e4 1424#endif
ddbcc7e8 1425#ifdef CONFIG_CGROUPS
817929ec
PM
1426 /* Control Group info protected by css_set_lock */
1427 struct css_set *cgroups;
1428 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1429 struct list_head cg_list;
ddbcc7e8 1430#endif
42b2dd0a 1431#ifdef CONFIG_FUTEX
0771dfef 1432 struct robust_list_head __user *robust_list;
34f192c6
IM
1433#ifdef CONFIG_COMPAT
1434 struct compat_robust_list_head __user *compat_robust_list;
1435#endif
c87e2837
IM
1436 struct list_head pi_state_list;
1437 struct futex_pi_state *pi_state_cache;
c7aceaba 1438#endif
a63eaf34
PM
1439#ifdef CONFIG_PERF_COUNTERS
1440 struct perf_counter_context *perf_counter_ctxp;
082ff5a2
PZ
1441 struct mutex perf_counter_mutex;
1442 struct list_head perf_counter_list;
a63eaf34 1443#endif
c7aceaba 1444#ifdef CONFIG_NUMA
58568d2a 1445 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1446 short il_next;
42b2dd0a 1447#endif
22e2c507 1448 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1449 struct rcu_head rcu;
b92ce558
JA
1450
1451 /*
1452 * cache last used pipe for splice
1453 */
1454 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1455#ifdef CONFIG_TASK_DELAY_ACCT
1456 struct task_delay_info *delays;
f4f154fd
AM
1457#endif
1458#ifdef CONFIG_FAULT_INJECTION
1459 int make_it_fail;
ca74e92b 1460#endif
3e26c149 1461 struct prop_local_single dirties;
9745512c
AV
1462#ifdef CONFIG_LATENCYTOP
1463 int latency_record_count;
1464 struct latency_record latency_record[LT_SAVECOUNT];
1465#endif
6976675d
AV
1466 /*
1467 * time slack values; these are used to round up poll() and
1468 * select() etc timeout values. These are in nanoseconds.
1469 */
1470 unsigned long timer_slack_ns;
1471 unsigned long default_timer_slack_ns;
f8d570a4
DM
1472
1473 struct list_head *scm_work_list;
fb52607a 1474#ifdef CONFIG_FUNCTION_GRAPH_TRACER
f201ae23
FW
1475 /* Index of current stored adress in ret_stack */
1476 int curr_ret_stack;
1477 /* Stack of return addresses for return function tracing */
1478 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1479 /* time stamp for last schedule */
1480 unsigned long long ftrace_timestamp;
f201ae23
FW
1481 /*
1482 * Number of functions that haven't been traced
1483 * because of depth overrun.
1484 */
1485 atomic_t trace_overrun;
380c4b14
FW
1486 /* Pause for the tracing */
1487 atomic_t tracing_graph_pause;
f201ae23 1488#endif
ea4e2bc4
SR
1489#ifdef CONFIG_TRACING
1490 /* state flags for use by tracers */
1491 unsigned long trace;
261842b7
SR
1492 /* bitmask of trace recursion */
1493 unsigned long trace_recursion;
1494#endif /* CONFIG_TRACING */
1da177e4
LT
1495};
1496
76e6eee0
RR
1497/* Future-safe accessor for struct task_struct's cpus_allowed. */
1498#define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1499
e05606d3
IM
1500/*
1501 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1502 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1503 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1504 * values are inverted: lower p->prio value means higher priority.
1505 *
1506 * The MAX_USER_RT_PRIO value allows the actual maximum
1507 * RT priority to be separate from the value exported to
1508 * user-space. This allows kernel threads to set their
1509 * priority to a value higher than any user task. Note:
1510 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1511 */
1512
1513#define MAX_USER_RT_PRIO 100
1514#define MAX_RT_PRIO MAX_USER_RT_PRIO
1515
1516#define MAX_PRIO (MAX_RT_PRIO + 40)
1517#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1518
1519static inline int rt_prio(int prio)
1520{
1521 if (unlikely(prio < MAX_RT_PRIO))
1522 return 1;
1523 return 0;
1524}
1525
e868171a 1526static inline int rt_task(struct task_struct *p)
e05606d3
IM
1527{
1528 return rt_prio(p->prio);
1529}
1530
e868171a 1531static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1532{
1533 return task->pids[PIDTYPE_PID].pid;
1534}
1535
e868171a 1536static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1537{
1538 return task->group_leader->pids[PIDTYPE_PID].pid;
1539}
1540
6dda81f4
ON
1541/*
1542 * Without tasklist or rcu lock it is not safe to dereference
1543 * the result of task_pgrp/task_session even if task == current,
1544 * we can race with another thread doing sys_setsid/sys_setpgid.
1545 */
e868171a 1546static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1547{
1548 return task->group_leader->pids[PIDTYPE_PGID].pid;
1549}
1550
e868171a 1551static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1552{
1553 return task->group_leader->pids[PIDTYPE_SID].pid;
1554}
1555
7af57294
PE
1556struct pid_namespace;
1557
1558/*
1559 * the helpers to get the task's different pids as they are seen
1560 * from various namespaces
1561 *
1562 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1563 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1564 * current.
7af57294
PE
1565 * task_xid_nr_ns() : id seen from the ns specified;
1566 *
1567 * set_task_vxid() : assigns a virtual id to a task;
1568 *
7af57294
PE
1569 * see also pid_nr() etc in include/linux/pid.h
1570 */
52ee2dfd
ON
1571pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1572 struct pid_namespace *ns);
7af57294 1573
e868171a 1574static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1575{
1576 return tsk->pid;
1577}
1578
52ee2dfd
ON
1579static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1580 struct pid_namespace *ns)
1581{
1582 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1583}
7af57294
PE
1584
1585static inline pid_t task_pid_vnr(struct task_struct *tsk)
1586{
52ee2dfd 1587 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1588}
1589
1590
e868171a 1591static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1592{
1593 return tsk->tgid;
1594}
1595
2f2a3a46 1596pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1597
1598static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1599{
1600 return pid_vnr(task_tgid(tsk));
1601}
1602
1603
52ee2dfd
ON
1604static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1605 struct pid_namespace *ns)
7af57294 1606{
52ee2dfd 1607 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1608}
1609
7af57294
PE
1610static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1611{
52ee2dfd 1612 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1613}
1614
1615
52ee2dfd
ON
1616static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1617 struct pid_namespace *ns)
7af57294 1618{
52ee2dfd 1619 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1620}
1621
7af57294
PE
1622static inline pid_t task_session_vnr(struct task_struct *tsk)
1623{
52ee2dfd 1624 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1625}
1626
1b0f7ffd
ON
1627/* obsolete, do not use */
1628static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1629{
1630 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1631}
7af57294 1632
1da177e4
LT
1633/**
1634 * pid_alive - check that a task structure is not stale
1635 * @p: Task structure to be checked.
1636 *
1637 * Test if a process is not yet dead (at most zombie state)
1638 * If pid_alive fails, then pointers within the task structure
1639 * can be stale and must not be dereferenced.
1640 */
e868171a 1641static inline int pid_alive(struct task_struct *p)
1da177e4 1642{
92476d7f 1643 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1644}
1645
f400e198 1646/**
b460cbc5 1647 * is_global_init - check if a task structure is init
3260259f
HK
1648 * @tsk: Task structure to be checked.
1649 *
1650 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1651 */
e868171a 1652static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1653{
1654 return tsk->pid == 1;
1655}
b460cbc5
SH
1656
1657/*
1658 * is_container_init:
1659 * check whether in the task is init in its own pid namespace.
f400e198 1660 */
b461cc03 1661extern int is_container_init(struct task_struct *tsk);
f400e198 1662
9ec52099
CLG
1663extern struct pid *cad_pid;
1664
1da177e4 1665extern void free_task(struct task_struct *tsk);
1da177e4 1666#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1667
158d9ebd 1668extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1669
1670static inline void put_task_struct(struct task_struct *t)
1671{
1672 if (atomic_dec_and_test(&t->usage))
8c7904a0 1673 __put_task_struct(t);
e56d0903 1674}
1da177e4 1675
49048622
BS
1676extern cputime_t task_utime(struct task_struct *p);
1677extern cputime_t task_stime(struct task_struct *p);
1678extern cputime_t task_gtime(struct task_struct *p);
1679
1da177e4
LT
1680/*
1681 * Per process flags
1682 */
1683#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1684 /* Not implemented yet, only for 486*/
1685#define PF_STARTING 0x00000002 /* being created */
1686#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1687#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1688#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1da177e4
LT
1689#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1690#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1691#define PF_DUMPCORE 0x00000200 /* dumped core */
1692#define PF_SIGNALED 0x00000400 /* killed by a signal */
1693#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1694#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1695#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1696#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1697#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1698#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1699#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1700#define PF_KSWAPD 0x00040000 /* I am kswapd */
1701#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1702#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1703#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1704#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1705#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1706#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1707#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1708#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
c61afb18 1709#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1710#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
ba96a0c8 1711#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
ebb12db5 1712#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1713
1714/*
1715 * Only the _current_ task can read/write to tsk->flags, but other
1716 * tasks can access tsk->flags in readonly mode for example
1717 * with tsk_used_math (like during threaded core dumping).
1718 * There is however an exception to this rule during ptrace
1719 * or during fork: the ptracer task is allowed to write to the
1720 * child->flags of its traced child (same goes for fork, the parent
1721 * can write to the child->flags), because we're guaranteed the
1722 * child is not running and in turn not changing child->flags
1723 * at the same time the parent does it.
1724 */
1725#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1726#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1727#define clear_used_math() clear_stopped_child_used_math(current)
1728#define set_used_math() set_stopped_child_used_math(current)
1729#define conditional_stopped_child_used_math(condition, child) \
1730 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1731#define conditional_used_math(condition) \
1732 conditional_stopped_child_used_math(condition, current)
1733#define copy_to_stopped_child_used_math(child) \
1734 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1735/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1736#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1737#define used_math() tsk_used_math(current)
1738
f41d911f
PM
1739#ifdef CONFIG_TREE_PREEMPT_RCU
1740
1741#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1742#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
1743#define RCU_READ_UNLOCK_GOT_QS (1 << 2) /* CPU has responded to RCU core. */
1744
1745static inline void rcu_copy_process(struct task_struct *p)
1746{
1747 p->rcu_read_lock_nesting = 0;
1748 p->rcu_read_unlock_special = 0;
dd5d19ba 1749 p->rcu_blocked_node = NULL;
f41d911f
PM
1750 INIT_LIST_HEAD(&p->rcu_node_entry);
1751}
1752
f41d911f
PM
1753#else
1754
1755static inline void rcu_copy_process(struct task_struct *p)
1756{
1757}
1758
1759#endif
1760
1da177e4 1761#ifdef CONFIG_SMP
cd8ba7cd 1762extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1763 const struct cpumask *new_mask);
1da177e4 1764#else
cd8ba7cd 1765static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1766 const struct cpumask *new_mask)
1da177e4 1767{
96f874e2 1768 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1769 return -EINVAL;
1770 return 0;
1771}
1772#endif
cd8ba7cd
MT
1773static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1774{
1775 return set_cpus_allowed_ptr(p, &new_mask);
1776}
1da177e4 1777
b342501c
IM
1778/*
1779 * Architectures can set this to 1 if they have specified
1780 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1781 * but then during bootup it turns out that sched_clock()
1782 * is reliable after all:
1783 */
1784#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1785extern int sched_clock_stable;
1786#endif
1787
1da177e4 1788extern unsigned long long sched_clock(void);
e436d800 1789
c1955a3d
PZ
1790extern void sched_clock_init(void);
1791extern u64 sched_clock_cpu(int cpu);
3e51f33f 1792
c1955a3d 1793#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1794static inline void sched_clock_tick(void)
1795{
1796}
1797
1798static inline void sched_clock_idle_sleep_event(void)
1799{
1800}
1801
1802static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1803{
1804}
1805#else
3e51f33f
PZ
1806extern void sched_clock_tick(void);
1807extern void sched_clock_idle_sleep_event(void);
1808extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1809#endif
1810
e436d800
IM
1811/*
1812 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1813 * clock constructed from sched_clock():
1814 */
1815extern unsigned long long cpu_clock(int cpu);
1816
36c8b586 1817extern unsigned long long
41b86e9c 1818task_sched_runtime(struct task_struct *task);
f06febc9 1819extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1820
1821/* sched_exec is called by processes performing an exec */
1822#ifdef CONFIG_SMP
1823extern void sched_exec(void);
1824#else
1825#define sched_exec() {}
1826#endif
1827
2aa44d05
IM
1828extern void sched_clock_idle_sleep_event(void);
1829extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1830
1da177e4
LT
1831#ifdef CONFIG_HOTPLUG_CPU
1832extern void idle_task_exit(void);
1833#else
1834static inline void idle_task_exit(void) {}
1835#endif
1836
1837extern void sched_idle_next(void);
b29739f9 1838
06d8308c
TG
1839#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1840extern void wake_up_idle_cpu(int cpu);
1841#else
1842static inline void wake_up_idle_cpu(int cpu) { }
1843#endif
1844
21805085 1845extern unsigned int sysctl_sched_latency;
b2be5e96 1846extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1847extern unsigned int sysctl_sched_wakeup_granularity;
47fea2ad
JSR
1848extern unsigned int sysctl_sched_shares_ratelimit;
1849extern unsigned int sysctl_sched_shares_thresh;
bf0f6f24 1850extern unsigned int sysctl_sched_child_runs_first;
2bba22c5 1851#ifdef CONFIG_SCHED_DEBUG
bf0f6f24 1852extern unsigned int sysctl_sched_features;
da84d961 1853extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1854extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1855extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1856extern unsigned int sysctl_timer_migration;
b2be5e96
PZ
1857
1858int sched_nr_latency_handler(struct ctl_table *table, int write,
1859 struct file *file, void __user *buffer, size_t *length,
1860 loff_t *ppos);
2bd8e6d4 1861#endif
eea08f32
AB
1862#ifdef CONFIG_SCHED_DEBUG
1863static inline unsigned int get_sysctl_timer_migration(void)
1864{
1865 return sysctl_timer_migration;
1866}
1867#else
1868static inline unsigned int get_sysctl_timer_migration(void)
1869{
1870 return 1;
1871}
1872#endif
9f0c1e56
PZ
1873extern unsigned int sysctl_sched_rt_period;
1874extern int sysctl_sched_rt_runtime;
2bd8e6d4 1875
d0b27fa7
PZ
1876int sched_rt_handler(struct ctl_table *table, int write,
1877 struct file *filp, void __user *buffer, size_t *lenp,
1878 loff_t *ppos);
1879
2bd8e6d4 1880extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1881
b29739f9 1882#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1883extern int rt_mutex_getprio(struct task_struct *p);
1884extern void rt_mutex_setprio(struct task_struct *p, int prio);
1885extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1886#else
e868171a 1887static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1888{
1889 return p->normal_prio;
1890}
95e02ca9 1891# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1892#endif
1893
36c8b586
IM
1894extern void set_user_nice(struct task_struct *p, long nice);
1895extern int task_prio(const struct task_struct *p);
1896extern int task_nice(const struct task_struct *p);
1897extern int can_nice(const struct task_struct *p, const int nice);
1898extern int task_curr(const struct task_struct *p);
1da177e4
LT
1899extern int idle_cpu(int cpu);
1900extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
961ccddd
RR
1901extern int sched_setscheduler_nocheck(struct task_struct *, int,
1902 struct sched_param *);
36c8b586
IM
1903extern struct task_struct *idle_task(int cpu);
1904extern struct task_struct *curr_task(int cpu);
1905extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1906
1907void yield(void);
1908
1909/*
1910 * The default (Linux) execution domain.
1911 */
1912extern struct exec_domain default_exec_domain;
1913
1914union thread_union {
1915 struct thread_info thread_info;
1916 unsigned long stack[THREAD_SIZE/sizeof(long)];
1917};
1918
1919#ifndef __HAVE_ARCH_KSTACK_END
1920static inline int kstack_end(void *addr)
1921{
1922 /* Reliable end of stack detection:
1923 * Some APM bios versions misalign the stack
1924 */
1925 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1926}
1927#endif
1928
1929extern union thread_union init_thread_union;
1930extern struct task_struct init_task;
1931
1932extern struct mm_struct init_mm;
1933
198fe21b
PE
1934extern struct pid_namespace init_pid_ns;
1935
1936/*
1937 * find a task by one of its numerical ids
1938 *
198fe21b
PE
1939 * find_task_by_pid_ns():
1940 * finds a task by its pid in the specified namespace
228ebcbe
PE
1941 * find_task_by_vpid():
1942 * finds a task by its virtual pid
198fe21b 1943 *
e49859e7 1944 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
1945 */
1946
228ebcbe
PE
1947extern struct task_struct *find_task_by_vpid(pid_t nr);
1948extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1949 struct pid_namespace *ns);
198fe21b 1950
8520d7c7 1951extern void __set_special_pids(struct pid *pid);
1da177e4
LT
1952
1953/* per-UID process charging. */
acce292c 1954extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
1955static inline struct user_struct *get_uid(struct user_struct *u)
1956{
1957 atomic_inc(&u->__count);
1958 return u;
1959}
1960extern void free_uid(struct user_struct *);
28f300d2 1961extern void release_uids(struct user_namespace *ns);
1da177e4
LT
1962
1963#include <asm/current.h>
1964
3171a030 1965extern void do_timer(unsigned long ticks);
1da177e4 1966
b3c97528
HH
1967extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1968extern int wake_up_process(struct task_struct *tsk);
1969extern void wake_up_new_task(struct task_struct *tsk,
1970 unsigned long clone_flags);
1da177e4
LT
1971#ifdef CONFIG_SMP
1972 extern void kick_process(struct task_struct *tsk);
1973#else
1974 static inline void kick_process(struct task_struct *tsk) { }
1975#endif
ad46c2c4
IM
1976extern void sched_fork(struct task_struct *p, int clone_flags);
1977extern void sched_dead(struct task_struct *p);
1da177e4 1978
1da177e4
LT
1979extern void proc_caches_init(void);
1980extern void flush_signals(struct task_struct *);
3bcac026 1981extern void __flush_signals(struct task_struct *);
10ab825b 1982extern void ignore_signals(struct task_struct *);
1da177e4
LT
1983extern void flush_signal_handlers(struct task_struct *, int force_default);
1984extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1985
1986static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1987{
1988 unsigned long flags;
1989 int ret;
1990
1991 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1992 ret = dequeue_signal(tsk, mask, info);
1993 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1994
1995 return ret;
1996}
1997
1998extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1999 sigset_t *mask);
2000extern void unblock_all_signals(void);
2001extern void release_task(struct task_struct * p);
2002extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2003extern int force_sigsegv(int, struct task_struct *);
2004extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2005extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2006extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2007extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2008extern int kill_pgrp(struct pid *pid, int sig, int priv);
2009extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2010extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2011extern int do_notify_parent(struct task_struct *, int);
1da177e4
LT
2012extern void force_sig(int, struct task_struct *);
2013extern void force_sig_specific(int, struct task_struct *);
2014extern int send_sig(int, struct task_struct *, int);
2015extern void zap_other_threads(struct task_struct *p);
1da177e4
LT
2016extern struct sigqueue *sigqueue_alloc(void);
2017extern void sigqueue_free(struct sigqueue *);
ac5c2153 2018extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2019extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2020extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2021
9ec52099
CLG
2022static inline int kill_cad_pid(int sig, int priv)
2023{
2024 return kill_pid(cad_pid, sig, priv);
2025}
2026
1da177e4
LT
2027/* These can be the second arg to send_sig_info/send_group_sig_info. */
2028#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2029#define SEND_SIG_PRIV ((struct siginfo *) 1)
2030#define SEND_SIG_FORCED ((struct siginfo *) 2)
2031
621d3121
ON
2032static inline int is_si_special(const struct siginfo *info)
2033{
2034 return info <= SEND_SIG_FORCED;
2035}
2036
1da177e4
LT
2037/* True if we are on the alternate signal stack. */
2038
2039static inline int on_sig_stack(unsigned long sp)
2040{
2041 return (sp - current->sas_ss_sp < current->sas_ss_size);
2042}
2043
2044static inline int sas_ss_flags(unsigned long sp)
2045{
2046 return (current->sas_ss_size == 0 ? SS_DISABLE
2047 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2048}
2049
1da177e4
LT
2050/*
2051 * Routines for handling mm_structs
2052 */
2053extern struct mm_struct * mm_alloc(void);
2054
2055/* mmdrop drops the mm and the page tables */
b3c97528 2056extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2057static inline void mmdrop(struct mm_struct * mm)
2058{
6fb43d7b 2059 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2060 __mmdrop(mm);
2061}
2062
2063/* mmput gets rid of the mappings and all user-space */
2064extern void mmput(struct mm_struct *);
2065/* Grab a reference to a task's mm, if it is not already going away */
2066extern struct mm_struct *get_task_mm(struct task_struct *task);
2067/* Remove the current tasks stale references to the old mm_struct */
2068extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2069/* Allocate a new mm structure and copy contents from tsk->mm */
2070extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2071
6f2c55b8
AD
2072extern int copy_thread(unsigned long, unsigned long, unsigned long,
2073 struct task_struct *, struct pt_regs *);
1da177e4
LT
2074extern void flush_thread(void);
2075extern void exit_thread(void);
2076
1da177e4 2077extern void exit_files(struct task_struct *);
6b3934ef 2078extern void __cleanup_signal(struct signal_struct *);
a7e5328a 2079extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2080
1da177e4 2081extern void exit_itimers(struct signal_struct *);
cbaffba1 2082extern void flush_itimer_signals(void);
1da177e4
LT
2083
2084extern NORET_TYPE void do_group_exit(int);
2085
1da177e4
LT
2086extern void daemonize(const char *, ...);
2087extern int allow_signal(int);
2088extern int disallow_signal(int);
1da177e4
LT
2089
2090extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2091extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2092struct task_struct *fork_idle(int);
1da177e4
LT
2093
2094extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2095extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2096
2097#ifdef CONFIG_SMP
a26b89f0 2098extern void wait_task_context_switch(struct task_struct *p);
85ba2d86 2099extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2100#else
a26b89f0 2101static inline void wait_task_context_switch(struct task_struct *p) {}
85ba2d86
RM
2102static inline unsigned long wait_task_inactive(struct task_struct *p,
2103 long match_state)
2104{
2105 return 1;
2106}
1da177e4
LT
2107#endif
2108
05725f7e
JP
2109#define next_task(p) \
2110 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2111
2112#define for_each_process(p) \
2113 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2114
5bb459bb 2115extern bool current_is_single_threaded(void);
d84f4f99 2116
1da177e4
LT
2117/*
2118 * Careful: do_each_thread/while_each_thread is a double loop so
2119 * 'break' will not work as expected - use goto instead.
2120 */
2121#define do_each_thread(g, t) \
2122 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2123
2124#define while_each_thread(g, t) \
2125 while ((t = next_thread(t)) != g)
2126
de12a787
EB
2127/* de_thread depends on thread_group_leader not being a pid based check */
2128#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2129
0804ef4b
EB
2130/* Do to the insanities of de_thread it is possible for a process
2131 * to have the pid of the thread group leader without actually being
2132 * the thread group leader. For iteration through the pids in proc
2133 * all we care about is that we have a task with the appropriate
2134 * pid, we don't actually care if we have the right task.
2135 */
e868171a 2136static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2137{
2138 return p->pid == p->tgid;
2139}
2140
bac0abd6
PE
2141static inline
2142int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2143{
2144 return p1->tgid == p2->tgid;
2145}
2146
36c8b586 2147static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2148{
05725f7e
JP
2149 return list_entry_rcu(p->thread_group.next,
2150 struct task_struct, thread_group);
47e65328
ON
2151}
2152
e868171a 2153static inline int thread_group_empty(struct task_struct *p)
1da177e4 2154{
47e65328 2155 return list_empty(&p->thread_group);
1da177e4
LT
2156}
2157
2158#define delay_group_leader(p) \
2159 (thread_group_leader(p) && !thread_group_empty(p))
2160
39c626ae
ON
2161static inline int task_detached(struct task_struct *p)
2162{
2163 return p->exit_signal == -1;
2164}
2165
1da177e4 2166/*
260ea101 2167 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2168 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2169 * pins the final release of task.io_context. Also protects ->cpuset and
2170 * ->cgroup.subsys[].
1da177e4
LT
2171 *
2172 * Nests both inside and outside of read_lock(&tasklist_lock).
2173 * It must not be nested with write_lock_irq(&tasklist_lock),
2174 * neither inside nor outside.
2175 */
2176static inline void task_lock(struct task_struct *p)
2177{
2178 spin_lock(&p->alloc_lock);
2179}
2180
2181static inline void task_unlock(struct task_struct *p)
2182{
2183 spin_unlock(&p->alloc_lock);
2184}
2185
f63ee72e
ON
2186extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2187 unsigned long *flags);
2188
2189static inline void unlock_task_sighand(struct task_struct *tsk,
2190 unsigned long *flags)
2191{
2192 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2193}
2194
f037360f
AV
2195#ifndef __HAVE_THREAD_FUNCTIONS
2196
f7e4217b
RZ
2197#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2198#define task_stack_page(task) ((task)->stack)
a1261f54 2199
10ebffde
AV
2200static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2201{
2202 *task_thread_info(p) = *task_thread_info(org);
2203 task_thread_info(p)->task = p;
2204}
2205
2206static inline unsigned long *end_of_stack(struct task_struct *p)
2207{
f7e4217b 2208 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2209}
2210
f037360f
AV
2211#endif
2212
8b05c7e6
FT
2213static inline int object_is_on_stack(void *obj)
2214{
2215 void *stack = task_stack_page(current);
2216
2217 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2218}
2219
8c9843e5
BH
2220extern void thread_info_cache_init(void);
2221
7c9f8861
ES
2222#ifdef CONFIG_DEBUG_STACK_USAGE
2223static inline unsigned long stack_not_used(struct task_struct *p)
2224{
2225 unsigned long *n = end_of_stack(p);
2226
2227 do { /* Skip over canary */
2228 n++;
2229 } while (!*n);
2230
2231 return (unsigned long)n - (unsigned long)end_of_stack(p);
2232}
2233#endif
2234
1da177e4
LT
2235/* set thread flags in other task's structures
2236 * - see asm/thread_info.h for TIF_xxxx flags available
2237 */
2238static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2239{
a1261f54 2240 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2241}
2242
2243static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2244{
a1261f54 2245 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2246}
2247
2248static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2249{
a1261f54 2250 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2251}
2252
2253static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2254{
a1261f54 2255 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2256}
2257
2258static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2259{
a1261f54 2260 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2261}
2262
2263static inline void set_tsk_need_resched(struct task_struct *tsk)
2264{
2265 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2266}
2267
2268static inline void clear_tsk_need_resched(struct task_struct *tsk)
2269{
2270 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2271}
2272
8ae121ac
GH
2273static inline int test_tsk_need_resched(struct task_struct *tsk)
2274{
2275 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2276}
2277
690cc3ff
EB
2278static inline int restart_syscall(void)
2279{
2280 set_tsk_thread_flag(current, TIF_SIGPENDING);
2281 return -ERESTARTNOINTR;
2282}
2283
1da177e4
LT
2284static inline int signal_pending(struct task_struct *p)
2285{
2286 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2287}
f776d12d 2288
b3c97528 2289extern int __fatal_signal_pending(struct task_struct *p);
f776d12d
MW
2290
2291static inline int fatal_signal_pending(struct task_struct *p)
2292{
2293 return signal_pending(p) && __fatal_signal_pending(p);
2294}
2295
16882c1e
ON
2296static inline int signal_pending_state(long state, struct task_struct *p)
2297{
2298 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2299 return 0;
2300 if (!signal_pending(p))
2301 return 0;
2302
16882c1e
ON
2303 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2304}
2305
1da177e4
LT
2306static inline int need_resched(void)
2307{
9404ef02 2308 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2309}
2310
2311/*
2312 * cond_resched() and cond_resched_lock(): latency reduction via
2313 * explicit rescheduling in places that are safe. The return
2314 * value indicates whether a reschedule was done in fact.
2315 * cond_resched_lock() will drop the spinlock before scheduling,
2316 * cond_resched_softirq() will enable bhs before scheduling.
2317 */
c3921ab7 2318extern int _cond_resched(void);
6f80bd98 2319
613afbf8
FW
2320#define cond_resched() ({ \
2321 __might_sleep(__FILE__, __LINE__, 0); \
2322 _cond_resched(); \
2323})
6f80bd98 2324
613afbf8
FW
2325extern int __cond_resched_lock(spinlock_t *lock);
2326
716a4234
FW
2327#ifdef CONFIG_PREEMPT
2328#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2329#else
716a4234 2330#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2331#endif
716a4234 2332
613afbf8 2333#define cond_resched_lock(lock) ({ \
716a4234 2334 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2335 __cond_resched_lock(lock); \
2336})
2337
2338extern int __cond_resched_softirq(void);
2339
2340#define cond_resched_softirq() ({ \
2341 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2342 __cond_resched_softirq(); \
2343})
1da177e4
LT
2344
2345/*
2346 * Does a critical section need to be broken due to another
95c354fe
NP
2347 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2348 * but a general need for low latency)
1da177e4 2349 */
95c354fe 2350static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2351{
95c354fe
NP
2352#ifdef CONFIG_PREEMPT
2353 return spin_is_contended(lock);
2354#else
1da177e4 2355 return 0;
95c354fe 2356#endif
1da177e4
LT
2357}
2358
f06febc9
FM
2359/*
2360 * Thread group CPU time accounting.
2361 */
4cd4c1b4 2362void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2363void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2364
490dea45 2365static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2366{
4cd4c1b4
PZ
2367 sig->cputimer.cputime = INIT_CPUTIME;
2368 spin_lock_init(&sig->cputimer.lock);
2369 sig->cputimer.running = 0;
f06febc9
FM
2370}
2371
f06febc9
FM
2372static inline void thread_group_cputime_free(struct signal_struct *sig)
2373{
f06febc9
FM
2374}
2375
7bb44ade
RM
2376/*
2377 * Reevaluate whether the task has signals pending delivery.
2378 * Wake the task if so.
2379 * This is required every time the blocked sigset_t changes.
2380 * callers must hold sighand->siglock.
2381 */
2382extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2383extern void recalc_sigpending(void);
2384
2385extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2386
2387/*
2388 * Wrappers for p->thread_info->cpu access. No-op on UP.
2389 */
2390#ifdef CONFIG_SMP
2391
2392static inline unsigned int task_cpu(const struct task_struct *p)
2393{
a1261f54 2394 return task_thread_info(p)->cpu;
1da177e4
LT
2395}
2396
c65cc870 2397extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2398
2399#else
2400
2401static inline unsigned int task_cpu(const struct task_struct *p)
2402{
2403 return 0;
2404}
2405
2406static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2407{
2408}
2409
2410#endif /* CONFIG_SMP */
2411
1da177e4 2412extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4 2413
1a3c3034
IM
2414#ifdef CONFIG_TRACING
2415extern void
2416__trace_special(void *__tr, void *__data,
2417 unsigned long arg1, unsigned long arg2, unsigned long arg3);
1da177e4 2418#else
1a3c3034
IM
2419static inline void
2420__trace_special(void *__tr, void *__data,
2421 unsigned long arg1, unsigned long arg2, unsigned long arg3)
1da177e4 2422{
1da177e4
LT
2423}
2424#endif
2425
96f874e2
RR
2426extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2427extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2428
1da177e4
LT
2429extern void normalize_rt_tasks(void);
2430
052f1dc7 2431#ifdef CONFIG_GROUP_SCHED
9b5b7751 2432
4cf86d77 2433extern struct task_group init_task_group;
eff766a6
PZ
2434#ifdef CONFIG_USER_SCHED
2435extern struct task_group root_task_group;
6c415b92 2436extern void set_tg_uid(struct user_struct *user);
eff766a6 2437#endif
9b5b7751 2438
ec7dc8ac 2439extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2440extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2441extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2442#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2443extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2444extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2445#endif
2446#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2447extern int sched_group_set_rt_runtime(struct task_group *tg,
2448 long rt_runtime_us);
2449extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2450extern int sched_group_set_rt_period(struct task_group *tg,
2451 long rt_period_us);
2452extern long sched_group_rt_period(struct task_group *tg);
54e99124 2453extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2454#endif
9b5b7751
SV
2455#endif
2456
54e99124
DG
2457extern int task_can_switch_user(struct user_struct *up,
2458 struct task_struct *tsk);
2459
4b98d11b
AD
2460#ifdef CONFIG_TASK_XACCT
2461static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2462{
940389b8 2463 tsk->ioac.rchar += amt;
4b98d11b
AD
2464}
2465
2466static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2467{
940389b8 2468 tsk->ioac.wchar += amt;
4b98d11b
AD
2469}
2470
2471static inline void inc_syscr(struct task_struct *tsk)
2472{
940389b8 2473 tsk->ioac.syscr++;
4b98d11b
AD
2474}
2475
2476static inline void inc_syscw(struct task_struct *tsk)
2477{
940389b8 2478 tsk->ioac.syscw++;
4b98d11b
AD
2479}
2480#else
2481static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2482{
2483}
2484
2485static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2486{
2487}
2488
2489static inline void inc_syscr(struct task_struct *tsk)
2490{
2491}
2492
2493static inline void inc_syscw(struct task_struct *tsk)
2494{
2495}
2496#endif
2497
82455257
DH
2498#ifndef TASK_SIZE_OF
2499#define TASK_SIZE_OF(tsk) TASK_SIZE
2500#endif
2501
0793a61d
TG
2502/*
2503 * Call the function if the target task is executing on a CPU right now:
2504 */
2505extern void task_oncpu_function_call(struct task_struct *p,
2506 void (*func) (void *info), void *info);
2507
2508
cf475ad2
BS
2509#ifdef CONFIG_MM_OWNER
2510extern void mm_update_next_owner(struct mm_struct *mm);
2511extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2512#else
2513static inline void mm_update_next_owner(struct mm_struct *mm)
2514{
2515}
2516
2517static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2518{
2519}
2520#endif /* CONFIG_MM_OWNER */
2521
7c731e0a
SR
2522#define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2523
1da177e4
LT
2524#endif /* __KERNEL__ */
2525
2526#endif