]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/linux/sched.h
scsi: convert to blk_delay_queue()
[mirror_ubuntu-artful-kernel.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4/*
5 * cloning flags:
6 */
7#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8#define CLONE_VM 0x00000100 /* set if VM shared between processes */
9#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15#define CLONE_THREAD 0x00010000 /* Same thread group? */
16#define CLONE_NEWNS 0x00020000 /* New namespace group? */
17#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
43bb40c9
DJ
24/* 0x02000000 was previously the unused CLONE_STOPPED (Start in stopped state)
25 and is now available for re-use. */
071df104 26#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 27#define CLONE_NEWIPC 0x08000000 /* New ipcs */
77ec739d 28#define CLONE_NEWUSER 0x10000000 /* New user namespace */
30e49c26 29#define CLONE_NEWPID 0x20000000 /* New pid namespace */
169e3674 30#define CLONE_NEWNET 0x40000000 /* New network namespace */
fadad878 31#define CLONE_IO 0x80000000 /* Clone io context */
b7b3c76a
DW
32
33/*
34 * Scheduling policies
35 */
36#define SCHED_NORMAL 0
37#define SCHED_FIFO 1
38#define SCHED_RR 2
39#define SCHED_BATCH 3
0e6aca43
IM
40/* SCHED_ISO: reserved but not implemented yet */
41#define SCHED_IDLE 5
ca94c442
LP
42/* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
43#define SCHED_RESET_ON_FORK 0x40000000
b7b3c76a 44
a3b6714e 45#ifdef __KERNEL__
b7b3c76a
DW
46
47struct sched_param {
48 int sched_priority;
49};
50
1da177e4
LT
51#include <asm/param.h> /* for HZ */
52
1da177e4
LT
53#include <linux/capability.h>
54#include <linux/threads.h>
55#include <linux/kernel.h>
56#include <linux/types.h>
57#include <linux/timex.h>
58#include <linux/jiffies.h>
59#include <linux/rbtree.h>
60#include <linux/thread_info.h>
61#include <linux/cpumask.h>
62#include <linux/errno.h>
63#include <linux/nodemask.h>
c92ff1bd 64#include <linux/mm_types.h>
1da177e4
LT
65
66#include <asm/system.h>
1da177e4
LT
67#include <asm/page.h>
68#include <asm/ptrace.h>
1da177e4
LT
69#include <asm/cputime.h>
70
71#include <linux/smp.h>
72#include <linux/sem.h>
73#include <linux/signal.h>
1da177e4
LT
74#include <linux/compiler.h>
75#include <linux/completion.h>
76#include <linux/pid.h>
77#include <linux/percpu.h>
78#include <linux/topology.h>
3e26c149 79#include <linux/proportions.h>
1da177e4 80#include <linux/seccomp.h>
e56d0903 81#include <linux/rcupdate.h>
05725f7e 82#include <linux/rculist.h>
23f78d4a 83#include <linux/rtmutex.h>
1da177e4 84
a3b6714e
DW
85#include <linux/time.h>
86#include <linux/param.h>
87#include <linux/resource.h>
88#include <linux/timer.h>
89#include <linux/hrtimer.h>
7c3ab738 90#include <linux/task_io_accounting.h>
9745512c 91#include <linux/latencytop.h>
9e2b2dc4 92#include <linux/cred.h>
a3b6714e
DW
93
94#include <asm/processor.h>
36d57ac4 95
1da177e4 96struct exec_domain;
c87e2837 97struct futex_pi_state;
286100a6 98struct robust_list_head;
bddd87c7 99struct bio_list;
5ad4e53b 100struct fs_struct;
cdd6c482 101struct perf_event_context;
1da177e4 102
1da177e4
LT
103/*
104 * List of flags we want to share for kernel threads,
105 * if only because they are not used by them anyway.
106 */
107#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
108
109/*
110 * These are the constant used to fake the fixed-point load-average
111 * counting. Some notes:
112 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
113 * a load-average precision of 10 bits integer + 11 bits fractional
114 * - if you want to count load-averages more often, you need more
115 * precision, or rounding will get you. With 2-second counting freq,
116 * the EXP_n values would be 1981, 2034 and 2043 if still using only
117 * 11 bit fractions.
118 */
119extern unsigned long avenrun[]; /* Load averages */
2d02494f 120extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
121
122#define FSHIFT 11 /* nr of bits of precision */
123#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 124#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
125#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
126#define EXP_5 2014 /* 1/exp(5sec/5min) */
127#define EXP_15 2037 /* 1/exp(5sec/15min) */
128
129#define CALC_LOAD(load,exp,n) \
130 load *= exp; \
131 load += n*(FIXED_1-exp); \
132 load >>= FSHIFT;
133
134extern unsigned long total_forks;
135extern int nr_threads;
1da177e4
LT
136DECLARE_PER_CPU(unsigned long, process_counts);
137extern int nr_processes(void);
138extern unsigned long nr_running(void);
139extern unsigned long nr_uninterruptible(void);
140extern unsigned long nr_iowait(void);
8c215bd3 141extern unsigned long nr_iowait_cpu(int cpu);
69d25870
AV
142extern unsigned long this_cpu_load(void);
143
144
0f004f5a 145extern void calc_global_load(unsigned long ticks);
1da177e4 146
7e49fcce
SR
147extern unsigned long get_parent_ip(unsigned long addr);
148
43ae34cb
IM
149struct seq_file;
150struct cfs_rq;
4cf86d77 151struct task_group;
43ae34cb
IM
152#ifdef CONFIG_SCHED_DEBUG
153extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
154extern void proc_sched_set_task(struct task_struct *p);
155extern void
5cef9eca 156print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
157#else
158static inline void
159proc_sched_show_task(struct task_struct *p, struct seq_file *m)
160{
161}
162static inline void proc_sched_set_task(struct task_struct *p)
163{
164}
165static inline void
5cef9eca 166print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
167{
168}
169#endif
1da177e4 170
4a8342d2
LT
171/*
172 * Task state bitmask. NOTE! These bits are also
173 * encoded in fs/proc/array.c: get_task_state().
174 *
175 * We have two separate sets of flags: task->state
176 * is about runnability, while task->exit_state are
177 * about the task exiting. Confusing, but this way
178 * modifying one set can't modify the other one by
179 * mistake.
180 */
1da177e4
LT
181#define TASK_RUNNING 0
182#define TASK_INTERRUPTIBLE 1
183#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
184#define __TASK_STOPPED 4
185#define __TASK_TRACED 8
4a8342d2
LT
186/* in tsk->exit_state */
187#define EXIT_ZOMBIE 16
188#define EXIT_DEAD 32
189/* in tsk->state again */
af927232 190#define TASK_DEAD 64
f021a3c2 191#define TASK_WAKEKILL 128
e9c84311 192#define TASK_WAKING 256
e1781538 193#define TASK_STATE_MAX 512
f021a3c2 194
44d90df6 195#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
73342151 196
e1781538
PZ
197extern char ___assert_task_state[1 - 2*!!(
198 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
199
200/* Convenience macros for the sake of set_task_state */
201#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
202#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
203#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 204
92a1f4bc
MW
205/* Convenience macros for the sake of wake_up */
206#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 207#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
208
209/* get_task_state() */
210#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
211 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
212 __TASK_TRACED)
92a1f4bc 213
f021a3c2
MW
214#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
215#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
8f92054e 216#define task_is_dead(task) ((task)->exit_state != 0)
92a1f4bc 217#define task_is_stopped_or_traced(task) \
f021a3c2 218 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 219#define task_contributes_to_load(task) \
e3c8ca83 220 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
6301cb95 221 (task->flags & PF_FREEZING) == 0)
1da177e4
LT
222
223#define __set_task_state(tsk, state_value) \
224 do { (tsk)->state = (state_value); } while (0)
225#define set_task_state(tsk, state_value) \
226 set_mb((tsk)->state, (state_value))
227
498d0c57
AM
228/*
229 * set_current_state() includes a barrier so that the write of current->state
230 * is correctly serialised wrt the caller's subsequent test of whether to
231 * actually sleep:
232 *
233 * set_current_state(TASK_UNINTERRUPTIBLE);
234 * if (do_i_need_to_sleep())
235 * schedule();
236 *
237 * If the caller does not need such serialisation then use __set_current_state()
238 */
1da177e4
LT
239#define __set_current_state(state_value) \
240 do { current->state = (state_value); } while (0)
241#define set_current_state(state_value) \
242 set_mb(current->state, (state_value))
243
244/* Task command name length */
245#define TASK_COMM_LEN 16
246
1da177e4
LT
247#include <linux/spinlock.h>
248
249/*
250 * This serializes "schedule()" and also protects
251 * the run-queue from deletions/modifications (but
252 * _adding_ to the beginning of the run-queue has
253 * a separate lock).
254 */
255extern rwlock_t tasklist_lock;
256extern spinlock_t mmlist_lock;
257
36c8b586 258struct task_struct;
1da177e4 259
db1466b3
PM
260#ifdef CONFIG_PROVE_RCU
261extern int lockdep_tasklist_lock_is_held(void);
262#endif /* #ifdef CONFIG_PROVE_RCU */
263
1da177e4
LT
264extern void sched_init(void);
265extern void sched_init_smp(void);
2d07b255 266extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 267extern void init_idle(struct task_struct *idle, int cpu);
1df21055 268extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 269
89f19f04 270extern int runqueue_is_locked(int cpu);
017730c1 271
6a7b3dc3 272extern cpumask_var_t nohz_cpu_mask;
46cb4b7c 273#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
83cd4fe2
VP
274extern void select_nohz_load_balancer(int stop_tick);
275extern int get_nohz_timer_target(void);
46cb4b7c 276#else
83cd4fe2 277static inline void select_nohz_load_balancer(int stop_tick) { }
46cb4b7c 278#endif
1da177e4 279
e59e2ae2 280/*
39bc89fd 281 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
282 */
283extern void show_state_filter(unsigned long state_filter);
284
285static inline void show_state(void)
286{
39bc89fd 287 show_state_filter(0);
e59e2ae2
IM
288}
289
1da177e4
LT
290extern void show_regs(struct pt_regs *);
291
292/*
293 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
294 * task), SP is the stack pointer of the first frame that should be shown in the back
295 * trace (or NULL if the entire call-chain of the task should be shown).
296 */
297extern void show_stack(struct task_struct *task, unsigned long *sp);
298
299void io_schedule(void);
300long io_schedule_timeout(long timeout);
301
302extern void cpu_init (void);
303extern void trap_init(void);
304extern void update_process_times(int user);
305extern void scheduler_tick(void);
306
82a1fcb9
IM
307extern void sched_show_task(struct task_struct *p);
308
19cc36c0 309#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 310extern void touch_softlockup_watchdog(void);
d6ad3e28 311extern void touch_softlockup_watchdog_sync(void);
04c9167f 312extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
313extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
314 void __user *buffer,
315 size_t *lenp, loff_t *ppos);
9c44bc03 316extern unsigned int softlockup_panic;
9383d967 317extern int softlockup_thresh;
004417a6 318void lockup_detector_init(void);
8446f1d3 319#else
8446f1d3
IM
320static inline void touch_softlockup_watchdog(void)
321{
322}
d6ad3e28
JW
323static inline void touch_softlockup_watchdog_sync(void)
324{
325}
04c9167f
JF
326static inline void touch_all_softlockup_watchdogs(void)
327{
328}
004417a6
PZ
329static inline void lockup_detector_init(void)
330{
331}
8446f1d3
IM
332#endif
333
e162b39a
MSB
334#ifdef CONFIG_DETECT_HUNG_TASK
335extern unsigned int sysctl_hung_task_panic;
336extern unsigned long sysctl_hung_task_check_count;
337extern unsigned long sysctl_hung_task_timeout_secs;
338extern unsigned long sysctl_hung_task_warnings;
339extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 340 void __user *buffer,
e162b39a 341 size_t *lenp, loff_t *ppos);
e4ecda1b
ML
342#else
343/* Avoid need for ifdefs elsewhere in the code */
344enum { sysctl_hung_task_timeout_secs = 0 };
e162b39a 345#endif
8446f1d3 346
1da177e4
LT
347/* Attach to any functions which should be ignored in wchan output. */
348#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
349
350/* Linker adds these: start and end of __sched functions */
351extern char __sched_text_start[], __sched_text_end[];
352
1da177e4
LT
353/* Is this address in the __sched functions? */
354extern int in_sched_functions(unsigned long addr);
355
356#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 357extern signed long schedule_timeout(signed long timeout);
64ed93a2 358extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 359extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 360extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 361asmlinkage void schedule(void);
0d66bf6d 362extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 363
ab516013 364struct nsproxy;
acce292c 365struct user_namespace;
1da177e4 366
341c87bf
KH
367/*
368 * Default maximum number of active map areas, this limits the number of vmas
369 * per mm struct. Users can overwrite this number by sysctl but there is a
370 * problem.
371 *
372 * When a program's coredump is generated as ELF format, a section is created
373 * per a vma. In ELF, the number of sections is represented in unsigned short.
374 * This means the number of sections should be smaller than 65535 at coredump.
375 * Because the kernel adds some informative sections to a image of program at
376 * generating coredump, we need some margin. The number of extra sections is
377 * 1-3 now and depends on arch. We use "5" as safe margin, here.
378 */
379#define MAPCOUNT_ELF_CORE_MARGIN (5)
4be929be 380#define DEFAULT_MAX_MAP_COUNT (USHRT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
381
382extern int sysctl_max_map_count;
383
384#include <linux/aio.h>
385
efc1a3b1
DH
386#ifdef CONFIG_MMU
387extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
388extern unsigned long
389arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
390 unsigned long, unsigned long);
391extern unsigned long
392arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
393 unsigned long len, unsigned long pgoff,
394 unsigned long flags);
1363c3cd
WW
395extern void arch_unmap_area(struct mm_struct *, unsigned long);
396extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
efc1a3b1
DH
397#else
398static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
399#endif
1da177e4 400
901608d9 401
6c5d5238
KH
402extern void set_dumpable(struct mm_struct *mm, int value);
403extern int get_dumpable(struct mm_struct *mm);
404
405/* mm flags */
3cb4a0bb 406/* dumpable bits */
6c5d5238
KH
407#define MMF_DUMPABLE 0 /* core dump is permitted */
408#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 409
3cb4a0bb 410#define MMF_DUMPABLE_BITS 2
f8af4da3 411#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH
412
413/* coredump filter bits */
414#define MMF_DUMP_ANON_PRIVATE 2
415#define MMF_DUMP_ANON_SHARED 3
416#define MMF_DUMP_MAPPED_PRIVATE 4
417#define MMF_DUMP_MAPPED_SHARED 5
82df3973 418#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
419#define MMF_DUMP_HUGETLB_PRIVATE 7
420#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 421
3cb4a0bb 422#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 423#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
424#define MMF_DUMP_FILTER_MASK \
425 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
426#define MMF_DUMP_FILTER_DEFAULT \
e575f111 427 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
428 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
429
430#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
431# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
432#else
433# define MMF_DUMP_MASK_DEFAULT_ELF 0
434#endif
f8af4da3
HD
435 /* leave room for more dump flags */
436#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 437#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
f8af4da3
HD
438
439#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 440
1da177e4
LT
441struct sighand_struct {
442 atomic_t count;
443 struct k_sigaction action[_NSIG];
444 spinlock_t siglock;
b8fceee1 445 wait_queue_head_t signalfd_wqh;
1da177e4
LT
446};
447
0e464814 448struct pacct_struct {
f6ec29a4
KK
449 int ac_flag;
450 long ac_exitcode;
0e464814 451 unsigned long ac_mem;
77787bfb
KK
452 cputime_t ac_utime, ac_stime;
453 unsigned long ac_minflt, ac_majflt;
0e464814
KK
454};
455
42c4ab41
SG
456struct cpu_itimer {
457 cputime_t expires;
458 cputime_t incr;
8356b5f9
SG
459 u32 error;
460 u32 incr_error;
42c4ab41
SG
461};
462
f06febc9
FM
463/**
464 * struct task_cputime - collected CPU time counts
465 * @utime: time spent in user mode, in &cputime_t units
466 * @stime: time spent in kernel mode, in &cputime_t units
467 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 468 *
f06febc9
FM
469 * This structure groups together three kinds of CPU time that are
470 * tracked for threads and thread groups. Most things considering
471 * CPU time want to group these counts together and treat all three
472 * of them in parallel.
473 */
474struct task_cputime {
475 cputime_t utime;
476 cputime_t stime;
477 unsigned long long sum_exec_runtime;
478};
479/* Alternate field names when used to cache expirations. */
480#define prof_exp stime
481#define virt_exp utime
482#define sched_exp sum_exec_runtime
483
4cd4c1b4
PZ
484#define INIT_CPUTIME \
485 (struct task_cputime) { \
486 .utime = cputime_zero, \
487 .stime = cputime_zero, \
488 .sum_exec_runtime = 0, \
489 }
490
c99e6efe
PZ
491/*
492 * Disable preemption until the scheduler is running.
493 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
494 *
495 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
496 * before the scheduler is active -- see should_resched().
c99e6efe 497 */
d86ee480 498#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 499
f06febc9 500/**
4cd4c1b4
PZ
501 * struct thread_group_cputimer - thread group interval timer counts
502 * @cputime: thread group interval timers.
503 * @running: non-zero when there are timers running and
504 * @cputime receives updates.
505 * @lock: lock for fields in this struct.
f06febc9
FM
506 *
507 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 508 * used for thread group CPU timer calculations.
f06febc9 509 */
4cd4c1b4
PZ
510struct thread_group_cputimer {
511 struct task_cputime cputime;
512 int running;
513 spinlock_t lock;
f06febc9 514};
f06febc9 515
5091faa4
MG
516struct autogroup;
517
1da177e4
LT
518/*
519 * NOTE! "signal_struct" does not have it's own
520 * locking, because a shared signal_struct always
521 * implies a shared sighand_struct, so locking
522 * sighand_struct is always a proper superset of
523 * the locking of signal_struct.
524 */
525struct signal_struct {
ea6d290c 526 atomic_t sigcnt;
1da177e4 527 atomic_t live;
b3ac022c 528 int nr_threads;
1da177e4
LT
529
530 wait_queue_head_t wait_chldexit; /* for wait4() */
531
532 /* current thread group signal load-balancing target: */
36c8b586 533 struct task_struct *curr_target;
1da177e4
LT
534
535 /* shared signal handling: */
536 struct sigpending shared_pending;
537
538 /* thread group exit support */
539 int group_exit_code;
540 /* overloaded:
541 * - notify group_exit_task when ->count is equal to notify_count
542 * - everyone except group_exit_task is stopped during signal delivery
543 * of fatal signals, group_exit_task processes the signal.
544 */
1da177e4 545 int notify_count;
07dd20e0 546 struct task_struct *group_exit_task;
1da177e4
LT
547
548 /* thread group stop support, overloads group_exit_code too */
549 int group_stop_count;
550 unsigned int flags; /* see SIGNAL_* flags below */
551
552 /* POSIX.1b Interval Timers */
553 struct list_head posix_timers;
554
555 /* ITIMER_REAL timer for the process */
2ff678b8 556 struct hrtimer real_timer;
fea9d175 557 struct pid *leader_pid;
2ff678b8 558 ktime_t it_real_incr;
1da177e4 559
42c4ab41
SG
560 /*
561 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
562 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
563 * values are defined to 0 and 1 respectively
564 */
565 struct cpu_itimer it[2];
1da177e4 566
f06febc9 567 /*
4cd4c1b4
PZ
568 * Thread group totals for process CPU timers.
569 * See thread_group_cputimer(), et al, for details.
f06febc9 570 */
4cd4c1b4 571 struct thread_group_cputimer cputimer;
f06febc9
FM
572
573 /* Earliest-expiration cache. */
574 struct task_cputime cputime_expires;
575
576 struct list_head cpu_timers[3];
577
ab521dc0 578 struct pid *tty_old_pgrp;
1ec320af 579
1da177e4
LT
580 /* boolean value for session group leader */
581 int leader;
582
583 struct tty_struct *tty; /* NULL if no tty */
584
5091faa4
MG
585#ifdef CONFIG_SCHED_AUTOGROUP
586 struct autogroup *autogroup;
587#endif
1da177e4
LT
588 /*
589 * Cumulative resource counters for dead threads in the group,
590 * and for reaped dead child processes forked by this group.
591 * Live threads maintain their own counters and add to these
592 * in __exit_signal, except for the group leader.
593 */
32bd671d 594 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
595 cputime_t gtime;
596 cputime_t cgtime;
0cf55e1e
HS
597#ifndef CONFIG_VIRT_CPU_ACCOUNTING
598 cputime_t prev_utime, prev_stime;
599#endif
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;
1f10206c 603 unsigned long maxrss, cmaxrss;
940389b8 604 struct task_io_accounting ioac;
1da177e4 605
32bd671d
PZ
606 /*
607 * Cumulative ns of schedule CPU time fo dead threads in the
608 * group, not including a zombie group leader, (This only differs
609 * from jiffies_to_ns(utime + stime) if sched_clock uses something
610 * other than jiffies.)
611 */
612 unsigned long long sum_sched_runtime;
613
1da177e4
LT
614 /*
615 * We don't bother to synchronize most readers of this at all,
616 * because there is no reader checking a limit that actually needs
617 * to get both rlim_cur and rlim_max atomically, and either one
618 * alone is a single word that can safely be read normally.
619 * getrlimit/setrlimit use task_lock(current->group_leader) to
620 * protect this instead of the siglock, because they really
621 * have no need to disable irqs.
622 */
623 struct rlimit rlim[RLIM_NLIMITS];
624
0e464814
KK
625#ifdef CONFIG_BSD_PROCESS_ACCT
626 struct pacct_struct pacct; /* per-process accounting information */
627#endif
ad4ecbcb 628#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
629 struct taskstats *stats;
630#endif
522ed776
MT
631#ifdef CONFIG_AUDIT
632 unsigned audit_tty;
633 struct tty_audit_buf *tty_audit_buf;
634#endif
28b83c51 635
a63d83f4
DR
636 int oom_adj; /* OOM kill score adjustment (bit shift) */
637 int oom_score_adj; /* OOM kill score adjustment */
dabb16f6
MSB
638 int oom_score_adj_min; /* OOM kill score adjustment minimum value.
639 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
640
641 struct mutex cred_guard_mutex; /* guard against foreign influences on
642 * credential calculations
643 * (notably. ptrace) */
1da177e4
LT
644};
645
4866cde0
NP
646/* Context switch must be unlocked if interrupts are to be enabled */
647#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
648# define __ARCH_WANT_UNLOCKED_CTXSW
649#endif
650
1da177e4
LT
651/*
652 * Bits in flags field of signal_struct.
653 */
654#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
655#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
656#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
657#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
658/*
659 * Pending notifications to parent.
660 */
661#define SIGNAL_CLD_STOPPED 0x00000010
662#define SIGNAL_CLD_CONTINUED 0x00000020
663#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 664
fae5fa44
ON
665#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
666
ed5d2cac
ON
667/* If true, all threads except ->group_exit_task have pending SIGKILL */
668static inline int signal_group_exit(const struct signal_struct *sig)
669{
670 return (sig->flags & SIGNAL_GROUP_EXIT) ||
671 (sig->group_exit_task != NULL);
672}
673
1da177e4
LT
674/*
675 * Some day this will be a full-fledged user tracking system..
676 */
677struct user_struct {
678 atomic_t __count; /* reference count */
679 atomic_t processes; /* How many processes does this user have? */
680 atomic_t files; /* How many open files does this user have? */
681 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 682#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
683 atomic_t inotify_watches; /* How many inotify watches does this user have? */
684 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
685#endif
4afeff85
EP
686#ifdef CONFIG_FANOTIFY
687 atomic_t fanotify_listeners;
688#endif
7ef9964e 689#ifdef CONFIG_EPOLL
52bd19f7 690 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 691#endif
970a8645 692#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
693 /* protected by mq_lock */
694 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 695#endif
1da177e4
LT
696 unsigned long locked_shm; /* How many pages of mlocked shm ? */
697
698#ifdef CONFIG_KEYS
699 struct key *uid_keyring; /* UID specific keyring */
700 struct key *session_keyring; /* UID's default session keyring */
701#endif
702
703 /* Hash table maintenance information */
735de223 704 struct hlist_node uidhash_node;
1da177e4 705 uid_t uid;
18b6e041 706 struct user_namespace *user_ns;
24e377a8 707
cdd6c482 708#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
709 atomic_long_t locked_vm;
710#endif
1da177e4
LT
711};
712
eb41d946 713extern int uids_sysfs_init(void);
5cb350ba 714
1da177e4
LT
715extern struct user_struct *find_user(uid_t);
716
717extern struct user_struct root_user;
718#define INIT_USER (&root_user)
719
b6dff3ec 720
1da177e4
LT
721struct backing_dev_info;
722struct reclaim_state;
723
52f17b6c 724#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
725struct sched_info {
726 /* cumulative counters */
2d72376b 727 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 728 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
729
730 /* timestamps */
172ba844
BS
731 unsigned long long last_arrival,/* when we last ran on a cpu */
732 last_queued; /* when we were last queued to run */
b8efb561
IM
733#ifdef CONFIG_SCHEDSTATS
734 /* BKL stats */
480b9434 735 unsigned int bkl_count;
b8efb561 736#endif
1da177e4 737};
52f17b6c 738#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 739
ca74e92b
SN
740#ifdef CONFIG_TASK_DELAY_ACCT
741struct task_delay_info {
742 spinlock_t lock;
743 unsigned int flags; /* Private per-task flags */
744
745 /* For each stat XXX, add following, aligned appropriately
746 *
747 * struct timespec XXX_start, XXX_end;
748 * u64 XXX_delay;
749 * u32 XXX_count;
750 *
751 * Atomicity of updates to XXX_delay, XXX_count protected by
752 * single lock above (split into XXX_lock if contention is an issue).
753 */
0ff92245
SN
754
755 /*
756 * XXX_count is incremented on every XXX operation, the delay
757 * associated with the operation is added to XXX_delay.
758 * XXX_delay contains the accumulated delay time in nanoseconds.
759 */
760 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
761 u64 blkio_delay; /* wait for sync block io completion */
762 u64 swapin_delay; /* wait for swapin block io completion */
763 u32 blkio_count; /* total count of the number of sync block */
764 /* io operations performed */
765 u32 swapin_count; /* total count of the number of swapin block */
766 /* io operations performed */
873b4771
KK
767
768 struct timespec freepages_start, freepages_end;
769 u64 freepages_delay; /* wait for memory reclaim */
770 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 771};
52f17b6c
CS
772#endif /* CONFIG_TASK_DELAY_ACCT */
773
774static inline int sched_info_on(void)
775{
776#ifdef CONFIG_SCHEDSTATS
777 return 1;
778#elif defined(CONFIG_TASK_DELAY_ACCT)
779 extern int delayacct_on;
780 return delayacct_on;
781#else
782 return 0;
ca74e92b 783#endif
52f17b6c 784}
ca74e92b 785
d15bcfdb
IM
786enum cpu_idle_type {
787 CPU_IDLE,
788 CPU_NOT_IDLE,
789 CPU_NEWLY_IDLE,
790 CPU_MAX_IDLE_TYPES
1da177e4
LT
791};
792
793/*
794 * sched-domains (multiprocessor balancing) declarations:
795 */
9aa7b369
IM
796
797/*
798 * Increase resolution of nice-level calculations:
799 */
800#define SCHED_LOAD_SHIFT 10
801#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
802
f8700df7 803#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 804
2dd73a4f 805#ifdef CONFIG_SMP
b5d978e0
PZ
806#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
807#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
808#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
809#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 810#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 811#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 812#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0
PZ
813#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
814#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
815#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
816#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 817#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 818#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 819
afb8a9b7
GS
820enum powersavings_balance_level {
821 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
822 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
823 * first for long running threads
824 */
825 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
826 * cpu package for power savings
827 */
828 MAX_POWERSAVINGS_BALANCE_LEVELS
829};
89c4710e 830
716707b2 831extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 832
716707b2
VS
833static inline int sd_balance_for_mc_power(void)
834{
835 if (sched_smt_power_savings)
836 return SD_POWERSAVINGS_BALANCE;
5c45bf27 837
28f53181
VS
838 if (!sched_mc_power_savings)
839 return SD_PREFER_SIBLING;
840
841 return 0;
716707b2 842}
89c4710e 843
716707b2
VS
844static inline int sd_balance_for_package_power(void)
845{
846 if (sched_mc_power_savings | sched_smt_power_savings)
847 return SD_POWERSAVINGS_BALANCE;
848
b5d978e0 849 return SD_PREFER_SIBLING;
716707b2 850}
5c45bf27 851
532cb4c4
MN
852extern int __weak arch_sd_sibiling_asym_packing(void);
853
100fdaee
VS
854/*
855 * Optimise SD flags for power savings:
856 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
857 * Keep default SD flags if sched_{smt,mc}_power_saving=0
858 */
859
860static inline int sd_power_saving_flags(void)
861{
862 if (sched_mc_power_savings | sched_smt_power_savings)
863 return SD_BALANCE_NEWIDLE;
864
865 return 0;
866}
1da177e4
LT
867
868struct sched_group {
869 struct sched_group *next; /* Must be a circular list */
1da177e4
LT
870
871 /*
872 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 873 * single CPU.
5517d86b 874 */
9d5efe05 875 unsigned int cpu_power, cpu_power_orig;
aae6d3dd 876 unsigned int group_weight;
6c99e9ad 877
4200efd9
IM
878 /*
879 * The CPUs this group covers.
880 *
881 * NOTE: this field is variable length. (Allocated dynamically
882 * by attaching extra space to the end of the structure,
883 * depending on how many CPUs the kernel has booted up with)
884 *
885 * It is also be embedded into static data structures at build
886 * time. (See 'struct static_sched_group' in kernel/sched.c)
887 */
888 unsigned long cpumask[0];
1da177e4
LT
889};
890
758b2cdc
RR
891static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
892{
6c99e9ad 893 return to_cpumask(sg->cpumask);
758b2cdc
RR
894}
895
1d3504fc
HS
896enum sched_domain_level {
897 SD_LV_NONE = 0,
898 SD_LV_SIBLING,
899 SD_LV_MC,
01a08546 900 SD_LV_BOOK,
1d3504fc
HS
901 SD_LV_CPU,
902 SD_LV_NODE,
903 SD_LV_ALLNODES,
904 SD_LV_MAX
905};
906
907struct sched_domain_attr {
908 int relax_domain_level;
909};
910
911#define SD_ATTR_INIT (struct sched_domain_attr) { \
912 .relax_domain_level = -1, \
913}
914
1da177e4
LT
915struct sched_domain {
916 /* These fields must be setup */
917 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 918 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 919 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
920 unsigned long min_interval; /* Minimum balance interval ms */
921 unsigned long max_interval; /* Maximum balance interval ms */
922 unsigned int busy_factor; /* less balancing by factor if busy */
923 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 924 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
925 unsigned int busy_idx;
926 unsigned int idle_idx;
927 unsigned int newidle_idx;
928 unsigned int wake_idx;
147cbb4b 929 unsigned int forkexec_idx;
a52bfd73 930 unsigned int smt_gain;
1da177e4 931 int flags; /* See SD_* */
1d3504fc 932 enum sched_domain_level level;
1da177e4
LT
933
934 /* Runtime fields. */
935 unsigned long last_balance; /* init to jiffies. units in jiffies */
936 unsigned int balance_interval; /* initialise to 1. units in ms. */
937 unsigned int nr_balance_failed; /* initialise to 0 */
938
2398f2c6
PZ
939 u64 last_update;
940
1da177e4
LT
941#ifdef CONFIG_SCHEDSTATS
942 /* load_balance() stats */
480b9434
KC
943 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
944 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
945 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
946 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
947 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
948 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
949 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
950 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
951
952 /* Active load balancing */
480b9434
KC
953 unsigned int alb_count;
954 unsigned int alb_failed;
955 unsigned int alb_pushed;
1da177e4 956
68767a0a 957 /* SD_BALANCE_EXEC stats */
480b9434
KC
958 unsigned int sbe_count;
959 unsigned int sbe_balanced;
960 unsigned int sbe_pushed;
1da177e4 961
68767a0a 962 /* SD_BALANCE_FORK stats */
480b9434
KC
963 unsigned int sbf_count;
964 unsigned int sbf_balanced;
965 unsigned int sbf_pushed;
68767a0a 966
1da177e4 967 /* try_to_wake_up() stats */
480b9434
KC
968 unsigned int ttwu_wake_remote;
969 unsigned int ttwu_move_affine;
970 unsigned int ttwu_move_balance;
1da177e4 971#endif
a5d8c348
IM
972#ifdef CONFIG_SCHED_DEBUG
973 char *name;
974#endif
6c99e9ad 975
669c55e9 976 unsigned int span_weight;
4200efd9
IM
977 /*
978 * Span of all CPUs in this domain.
979 *
980 * NOTE: this field is variable length. (Allocated dynamically
981 * by attaching extra space to the end of the structure,
982 * depending on how many CPUs the kernel has booted up with)
983 *
984 * It is also be embedded into static data structures at build
985 * time. (See 'struct static_sched_domain' in kernel/sched.c)
986 */
987 unsigned long span[0];
1da177e4
LT
988};
989
758b2cdc
RR
990static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
991{
6c99e9ad 992 return to_cpumask(sd->span);
758b2cdc
RR
993}
994
acc3f5d7 995extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 996 struct sched_domain_attr *dattr_new);
029190c5 997
acc3f5d7
RR
998/* Allocate an array of sched domains, for partition_sched_domains(). */
999cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1000void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1001
06aaf76a
IM
1002/* Test a flag in parent sched domain */
1003static inline int test_sd_parent(struct sched_domain *sd, int flag)
1004{
1005 if (sd->parent && (sd->parent->flags & flag))
1006 return 1;
1007
1008 return 0;
1009}
029190c5 1010
47fe38fc
PZ
1011unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1012unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1013
1b427c15 1014#else /* CONFIG_SMP */
1da177e4 1015
1b427c15 1016struct sched_domain_attr;
d02c7a8c 1017
1b427c15 1018static inline void
acc3f5d7 1019partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1020 struct sched_domain_attr *dattr_new)
1021{
d02c7a8c 1022}
1b427c15 1023#endif /* !CONFIG_SMP */
1da177e4 1024
47fe38fc 1025
1da177e4 1026struct io_context; /* See blkdev.h */
1da177e4 1027
1da177e4 1028
383f2835 1029#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1030extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1031#else
1032static inline void prefetch_stack(struct task_struct *t) { }
1033#endif
1da177e4
LT
1034
1035struct audit_context; /* See audit.c */
1036struct mempolicy;
b92ce558 1037struct pipe_inode_info;
4865ecf1 1038struct uts_namespace;
1da177e4 1039
20b8a59f
IM
1040struct rq;
1041struct sched_domain;
1042
7d478721
PZ
1043/*
1044 * wake flags
1045 */
1046#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1047#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1048
371fd7e7
PZ
1049#define ENQUEUE_WAKEUP 1
1050#define ENQUEUE_WAKING 2
1051#define ENQUEUE_HEAD 4
1052
1053#define DEQUEUE_SLEEP 1
1054
20b8a59f 1055struct sched_class {
5522d5d5 1056 const struct sched_class *next;
20b8a59f 1057
371fd7e7
PZ
1058 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int flags);
1059 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int flags);
4530d7ab 1060 void (*yield_task) (struct rq *rq);
20b8a59f 1061
7d478721 1062 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1063
fb8d4724 1064 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1065 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1066
681f3e68 1067#ifdef CONFIG_SMP
0017d735
PZ
1068 int (*select_task_rq)(struct rq *rq, struct task_struct *p,
1069 int sd_flag, int flags);
4ce72a2c 1070
9a897c5a
SR
1071 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1072 void (*post_schedule) (struct rq *this_rq);
efbbd05a
PZ
1073 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1074 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1075
cd8ba7cd 1076 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1077 const struct cpumask *newmask);
57d885fe 1078
1f11eb6a
GH
1079 void (*rq_online)(struct rq *rq);
1080 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1081#endif
1082
1083 void (*set_curr_task) (struct rq *rq);
1084 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1085 void (*task_fork) (struct task_struct *p);
cb469845
SR
1086
1087 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1088 int running);
1089 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1090 int running);
1091 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1092 int oldprio, int running);
810b3817 1093
dba091b9
TG
1094 unsigned int (*get_rr_interval) (struct rq *rq,
1095 struct task_struct *task);
0d721cea 1096
810b3817 1097#ifdef CONFIG_FAIR_GROUP_SCHED
b2b5ce02 1098 void (*task_move_group) (struct task_struct *p, int on_rq);
810b3817 1099#endif
20b8a59f
IM
1100};
1101
1102struct load_weight {
1103 unsigned long weight, inv_weight;
1104};
1105
94c18227 1106#ifdef CONFIG_SCHEDSTATS
41acab88 1107struct sched_statistics {
20b8a59f 1108 u64 wait_start;
94c18227 1109 u64 wait_max;
6d082592
AV
1110 u64 wait_count;
1111 u64 wait_sum;
8f0dfc34
AV
1112 u64 iowait_count;
1113 u64 iowait_sum;
94c18227 1114
20b8a59f 1115 u64 sleep_start;
20b8a59f 1116 u64 sleep_max;
94c18227
IM
1117 s64 sum_sleep_runtime;
1118
1119 u64 block_start;
20b8a59f
IM
1120 u64 block_max;
1121 u64 exec_max;
eba1ed4b 1122 u64 slice_max;
cc367732 1123
cc367732
IM
1124 u64 nr_migrations_cold;
1125 u64 nr_failed_migrations_affine;
1126 u64 nr_failed_migrations_running;
1127 u64 nr_failed_migrations_hot;
1128 u64 nr_forced_migrations;
cc367732
IM
1129
1130 u64 nr_wakeups;
1131 u64 nr_wakeups_sync;
1132 u64 nr_wakeups_migrate;
1133 u64 nr_wakeups_local;
1134 u64 nr_wakeups_remote;
1135 u64 nr_wakeups_affine;
1136 u64 nr_wakeups_affine_attempts;
1137 u64 nr_wakeups_passive;
1138 u64 nr_wakeups_idle;
41acab88
LDM
1139};
1140#endif
1141
1142struct sched_entity {
1143 struct load_weight load; /* for load-balancing */
1144 struct rb_node run_node;
1145 struct list_head group_node;
1146 unsigned int on_rq;
1147
1148 u64 exec_start;
1149 u64 sum_exec_runtime;
1150 u64 vruntime;
1151 u64 prev_sum_exec_runtime;
1152
41acab88
LDM
1153 u64 nr_migrations;
1154
41acab88
LDM
1155#ifdef CONFIG_SCHEDSTATS
1156 struct sched_statistics statistics;
94c18227
IM
1157#endif
1158
20b8a59f
IM
1159#ifdef CONFIG_FAIR_GROUP_SCHED
1160 struct sched_entity *parent;
1161 /* rq on which this entity is (to be) queued: */
1162 struct cfs_rq *cfs_rq;
1163 /* rq "owned" by this entity/group: */
1164 struct cfs_rq *my_q;
1165#endif
1166};
70b97a7f 1167
fa717060
PZ
1168struct sched_rt_entity {
1169 struct list_head run_list;
78f2c7db 1170 unsigned long timeout;
bee367ed 1171 unsigned int time_slice;
6f505b16
PZ
1172 int nr_cpus_allowed;
1173
58d6c2d7 1174 struct sched_rt_entity *back;
052f1dc7 1175#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1176 struct sched_rt_entity *parent;
1177 /* rq on which this entity is (to be) queued: */
1178 struct rt_rq *rt_rq;
1179 /* rq "owned" by this entity/group: */
1180 struct rt_rq *my_q;
1181#endif
fa717060
PZ
1182};
1183
86848966
PM
1184struct rcu_node;
1185
8dc85d54
PZ
1186enum perf_event_task_context {
1187 perf_invalid_context = -1,
1188 perf_hw_context = 0,
89a1e187 1189 perf_sw_context,
8dc85d54
PZ
1190 perf_nr_task_contexts,
1191};
1192
1da177e4
LT
1193struct task_struct {
1194 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1195 void *stack;
1da177e4 1196 atomic_t usage;
97dc32cd
WC
1197 unsigned int flags; /* per process flags, defined below */
1198 unsigned int ptrace;
1da177e4 1199
36772092 1200 int lock_depth; /* BKL lock depth */
1da177e4 1201
2dd73a4f
PW
1202#ifdef CONFIG_SMP
1203#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1204 int oncpu;
1205#endif
2dd73a4f 1206#endif
50e645a8 1207
b29739f9 1208 int prio, static_prio, normal_prio;
c7aceaba 1209 unsigned int rt_priority;
5522d5d5 1210 const struct sched_class *sched_class;
20b8a59f 1211 struct sched_entity se;
fa717060 1212 struct sched_rt_entity rt;
1da177e4 1213
e107be36
AK
1214#ifdef CONFIG_PREEMPT_NOTIFIERS
1215 /* list of struct preempt_notifier: */
1216 struct hlist_head preempt_notifiers;
1217#endif
1218
18796aa0
AD
1219 /*
1220 * fpu_counter contains the number of consecutive context switches
1221 * that the FPU is used. If this is over a threshold, the lazy fpu
1222 * saving becomes unlazy to save the trap. This is an unsigned char
1223 * so that after 256 times the counter wraps and the behavior turns
1224 * lazy again; this to deal with bursty apps that only use FPU for
1225 * a short time
1226 */
1227 unsigned char fpu_counter;
6c5c9341 1228#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1229 unsigned int btrace_seq;
6c5c9341 1230#endif
1da177e4 1231
97dc32cd 1232 unsigned int policy;
1da177e4 1233 cpumask_t cpus_allowed;
1da177e4 1234
a57eb940 1235#ifdef CONFIG_PREEMPT_RCU
e260be67 1236 int rcu_read_lock_nesting;
f41d911f 1237 char rcu_read_unlock_special;
f41d911f 1238 struct list_head rcu_node_entry;
a57eb940
PM
1239#endif /* #ifdef CONFIG_PREEMPT_RCU */
1240#ifdef CONFIG_TREE_PREEMPT_RCU
1241 struct rcu_node *rcu_blocked_node;
f41d911f 1242#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1243#ifdef CONFIG_RCU_BOOST
1244 struct rt_mutex *rcu_boost_mutex;
1245#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1246
52f17b6c 1247#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1248 struct sched_info sched_info;
1249#endif
1250
1251 struct list_head tasks;
806c09a7 1252#ifdef CONFIG_SMP
917b627d 1253 struct plist_node pushable_tasks;
806c09a7 1254#endif
1da177e4
LT
1255
1256 struct mm_struct *mm, *active_mm;
34e55232
KH
1257#if defined(SPLIT_RSS_COUNTING)
1258 struct task_rss_stat rss_stat;
1259#endif
1da177e4 1260/* task state */
97dc32cd 1261 int exit_state;
1da177e4
LT
1262 int exit_code, exit_signal;
1263 int pdeath_signal; /* The signal sent when the parent dies */
1264 /* ??? */
97dc32cd 1265 unsigned int personality;
1da177e4 1266 unsigned did_exec:1;
f9ce1f1c
KT
1267 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1268 * execve */
8f0dfc34
AV
1269 unsigned in_iowait:1;
1270
ca94c442
LP
1271
1272 /* Revert to default priority/policy when forking */
1273 unsigned sched_reset_on_fork:1;
1274
1da177e4
LT
1275 pid_t pid;
1276 pid_t tgid;
0a425405 1277
1314562a 1278#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1279 /* Canary value for the -fstack-protector gcc feature */
1280 unsigned long stack_canary;
1314562a 1281#endif
e0032087 1282
1da177e4
LT
1283 /*
1284 * pointers to (original) parent process, youngest child, younger sibling,
1285 * older sibling, respectively. (p->father can be replaced with
f470021a 1286 * p->real_parent->pid)
1da177e4 1287 */
f470021a
RM
1288 struct task_struct *real_parent; /* real parent process */
1289 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1290 /*
f470021a 1291 * children/sibling forms the list of my natural children
1da177e4
LT
1292 */
1293 struct list_head children; /* list of my children */
1294 struct list_head sibling; /* linkage in my parent's children list */
1295 struct task_struct *group_leader; /* threadgroup leader */
1296
f470021a
RM
1297 /*
1298 * ptraced is the list of tasks this task is using ptrace on.
1299 * This includes both natural children and PTRACE_ATTACH targets.
1300 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1301 */
1302 struct list_head ptraced;
1303 struct list_head ptrace_entry;
1304
1da177e4 1305 /* PID/PID hash table linkage. */
92476d7f 1306 struct pid_link pids[PIDTYPE_MAX];
47e65328 1307 struct list_head thread_group;
1da177e4
LT
1308
1309 struct completion *vfork_done; /* for vfork() */
1310 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1311 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1312
c66f08be 1313 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1314 cputime_t gtime;
d99ca3b9 1315#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1316 cputime_t prev_utime, prev_stime;
d99ca3b9 1317#endif
1da177e4 1318 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1319 struct timespec start_time; /* monotonic time */
1320 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1321/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1322 unsigned long min_flt, maj_flt;
1323
f06febc9 1324 struct task_cputime cputime_expires;
1da177e4
LT
1325 struct list_head cpu_timers[3];
1326
1327/* process credentials */
1b0ba1c9 1328 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1329 * credentials (COW) */
1b0ba1c9 1330 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1331 * credentials (COW) */
ee18d64c 1332 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1333
36772092
PBG
1334 char comm[TASK_COMM_LEN]; /* executable name excluding path
1335 - access with [gs]et_task_comm (which lock
1336 it with task_lock())
221af7f8 1337 - initialized normally by setup_new_exec */
1da177e4
LT
1338/* file system info */
1339 int link_count, total_link_count;
3d5b6fcc 1340#ifdef CONFIG_SYSVIPC
1da177e4
LT
1341/* ipc stuff */
1342 struct sysv_sem sysvsem;
3d5b6fcc 1343#endif
e162b39a 1344#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1345/* hung task detection */
82a1fcb9
IM
1346 unsigned long last_switch_count;
1347#endif
1da177e4
LT
1348/* CPU-specific state of this task */
1349 struct thread_struct thread;
1350/* filesystem information */
1351 struct fs_struct *fs;
1352/* open file information */
1353 struct files_struct *files;
1651e14e 1354/* namespaces */
ab516013 1355 struct nsproxy *nsproxy;
1da177e4
LT
1356/* signal handlers */
1357 struct signal_struct *signal;
1358 struct sighand_struct *sighand;
1359
1360 sigset_t blocked, real_blocked;
f3de272b 1361 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1362 struct sigpending pending;
1363
1364 unsigned long sas_ss_sp;
1365 size_t sas_ss_size;
1366 int (*notifier)(void *priv);
1367 void *notifier_data;
1368 sigset_t *notifier_mask;
1da177e4 1369 struct audit_context *audit_context;
bfef93a5
AV
1370#ifdef CONFIG_AUDITSYSCALL
1371 uid_t loginuid;
4746ec5b 1372 unsigned int sessionid;
bfef93a5 1373#endif
1da177e4
LT
1374 seccomp_t seccomp;
1375
1376/* Thread group tracking */
1377 u32 parent_exec_id;
1378 u32 self_exec_id;
58568d2a
MX
1379/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1380 * mempolicy */
1da177e4 1381 spinlock_t alloc_lock;
1da177e4 1382
3aa551c9
TG
1383#ifdef CONFIG_GENERIC_HARDIRQS
1384 /* IRQ handler threads */
1385 struct irqaction *irqaction;
1386#endif
1387
b29739f9 1388 /* Protection of the PI data structures: */
1d615482 1389 raw_spinlock_t pi_lock;
b29739f9 1390
23f78d4a
IM
1391#ifdef CONFIG_RT_MUTEXES
1392 /* PI waiters blocked on a rt_mutex held by this task */
1393 struct plist_head pi_waiters;
1394 /* Deadlock detection and priority inheritance handling */
1395 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1396#endif
1397
408894ee
IM
1398#ifdef CONFIG_DEBUG_MUTEXES
1399 /* mutex deadlock detection */
1400 struct mutex_waiter *blocked_on;
1401#endif
de30a2b3
IM
1402#ifdef CONFIG_TRACE_IRQFLAGS
1403 unsigned int irq_events;
de30a2b3 1404 unsigned long hardirq_enable_ip;
de30a2b3 1405 unsigned long hardirq_disable_ip;
fa1452e8 1406 unsigned int hardirq_enable_event;
de30a2b3 1407 unsigned int hardirq_disable_event;
fa1452e8
HS
1408 int hardirqs_enabled;
1409 int hardirq_context;
de30a2b3 1410 unsigned long softirq_disable_ip;
de30a2b3 1411 unsigned long softirq_enable_ip;
fa1452e8 1412 unsigned int softirq_disable_event;
de30a2b3 1413 unsigned int softirq_enable_event;
fa1452e8 1414 int softirqs_enabled;
de30a2b3
IM
1415 int softirq_context;
1416#endif
fbb9ce95 1417#ifdef CONFIG_LOCKDEP
bdb9441e 1418# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1419 u64 curr_chain_key;
1420 int lockdep_depth;
fbb9ce95 1421 unsigned int lockdep_recursion;
c7aceaba 1422 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1423 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1424#endif
408894ee 1425
1da177e4
LT
1426/* journalling filesystem info */
1427 void *journal_info;
1428
d89d8796 1429/* stacked block device info */
bddd87c7 1430 struct bio_list *bio_list;
d89d8796 1431
1da177e4
LT
1432/* VM state */
1433 struct reclaim_state *reclaim_state;
1434
1da177e4
LT
1435 struct backing_dev_info *backing_dev_info;
1436
1437 struct io_context *io_context;
1438
1439 unsigned long ptrace_message;
1440 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1441 struct task_io_accounting ioac;
8f0ab514 1442#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1443 u64 acct_rss_mem1; /* accumulated rss usage */
1444 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1445 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1446#endif
1447#ifdef CONFIG_CPUSETS
58568d2a 1448 nodemask_t mems_allowed; /* Protected by alloc_lock */
c0ff7453 1449 int mems_allowed_change_disable;
825a46af 1450 int cpuset_mem_spread_rotor;
6adef3eb 1451 int cpuset_slab_spread_rotor;
1da177e4 1452#endif
ddbcc7e8 1453#ifdef CONFIG_CGROUPS
817929ec 1454 /* Control Group info protected by css_set_lock */
2c392b8c 1455 struct css_set __rcu *cgroups;
817929ec
PM
1456 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1457 struct list_head cg_list;
ddbcc7e8 1458#endif
42b2dd0a 1459#ifdef CONFIG_FUTEX
0771dfef 1460 struct robust_list_head __user *robust_list;
34f192c6
IM
1461#ifdef CONFIG_COMPAT
1462 struct compat_robust_list_head __user *compat_robust_list;
1463#endif
c87e2837
IM
1464 struct list_head pi_state_list;
1465 struct futex_pi_state *pi_state_cache;
c7aceaba 1466#endif
cdd6c482 1467#ifdef CONFIG_PERF_EVENTS
8dc85d54 1468 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1469 struct mutex perf_event_mutex;
1470 struct list_head perf_event_list;
a63eaf34 1471#endif
c7aceaba 1472#ifdef CONFIG_NUMA
58568d2a 1473 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1474 short il_next;
42b2dd0a 1475#endif
22e2c507 1476 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1477 struct rcu_head rcu;
b92ce558
JA
1478
1479 /*
1480 * cache last used pipe for splice
1481 */
1482 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1483#ifdef CONFIG_TASK_DELAY_ACCT
1484 struct task_delay_info *delays;
f4f154fd
AM
1485#endif
1486#ifdef CONFIG_FAULT_INJECTION
1487 int make_it_fail;
ca74e92b 1488#endif
3e26c149 1489 struct prop_local_single dirties;
9745512c
AV
1490#ifdef CONFIG_LATENCYTOP
1491 int latency_record_count;
1492 struct latency_record latency_record[LT_SAVECOUNT];
1493#endif
6976675d
AV
1494 /*
1495 * time slack values; these are used to round up poll() and
1496 * select() etc timeout values. These are in nanoseconds.
1497 */
1498 unsigned long timer_slack_ns;
1499 unsigned long default_timer_slack_ns;
f8d570a4
DM
1500
1501 struct list_head *scm_work_list;
fb52607a 1502#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1503 /* Index of current stored address in ret_stack */
f201ae23
FW
1504 int curr_ret_stack;
1505 /* Stack of return addresses for return function tracing */
1506 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1507 /* time stamp for last schedule */
1508 unsigned long long ftrace_timestamp;
f201ae23
FW
1509 /*
1510 * Number of functions that haven't been traced
1511 * because of depth overrun.
1512 */
1513 atomic_t trace_overrun;
380c4b14
FW
1514 /* Pause for the tracing */
1515 atomic_t tracing_graph_pause;
f201ae23 1516#endif
ea4e2bc4
SR
1517#ifdef CONFIG_TRACING
1518 /* state flags for use by tracers */
1519 unsigned long trace;
261842b7
SR
1520 /* bitmask of trace recursion */
1521 unsigned long trace_recursion;
1522#endif /* CONFIG_TRACING */
569b846d
KH
1523#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1524 struct memcg_batch_info {
1525 int do_batch; /* incremented when batch uncharge started */
1526 struct mem_cgroup *memcg; /* target memcg of uncharge */
1527 unsigned long bytes; /* uncharged usage */
1528 unsigned long memsw_bytes; /* uncharged mem+swap usage */
1529 } memcg_batch;
1530#endif
1da177e4
LT
1531};
1532
76e6eee0 1533/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1534#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1535
e05606d3
IM
1536/*
1537 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1538 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1539 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1540 * values are inverted: lower p->prio value means higher priority.
1541 *
1542 * The MAX_USER_RT_PRIO value allows the actual maximum
1543 * RT priority to be separate from the value exported to
1544 * user-space. This allows kernel threads to set their
1545 * priority to a value higher than any user task. Note:
1546 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1547 */
1548
1549#define MAX_USER_RT_PRIO 100
1550#define MAX_RT_PRIO MAX_USER_RT_PRIO
1551
1552#define MAX_PRIO (MAX_RT_PRIO + 40)
1553#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1554
1555static inline int rt_prio(int prio)
1556{
1557 if (unlikely(prio < MAX_RT_PRIO))
1558 return 1;
1559 return 0;
1560}
1561
e868171a 1562static inline int rt_task(struct task_struct *p)
e05606d3
IM
1563{
1564 return rt_prio(p->prio);
1565}
1566
e868171a 1567static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1568{
1569 return task->pids[PIDTYPE_PID].pid;
1570}
1571
e868171a 1572static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1573{
1574 return task->group_leader->pids[PIDTYPE_PID].pid;
1575}
1576
6dda81f4
ON
1577/*
1578 * Without tasklist or rcu lock it is not safe to dereference
1579 * the result of task_pgrp/task_session even if task == current,
1580 * we can race with another thread doing sys_setsid/sys_setpgid.
1581 */
e868171a 1582static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1583{
1584 return task->group_leader->pids[PIDTYPE_PGID].pid;
1585}
1586
e868171a 1587static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1588{
1589 return task->group_leader->pids[PIDTYPE_SID].pid;
1590}
1591
7af57294
PE
1592struct pid_namespace;
1593
1594/*
1595 * the helpers to get the task's different pids as they are seen
1596 * from various namespaces
1597 *
1598 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1599 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1600 * current.
7af57294
PE
1601 * task_xid_nr_ns() : id seen from the ns specified;
1602 *
1603 * set_task_vxid() : assigns a virtual id to a task;
1604 *
7af57294
PE
1605 * see also pid_nr() etc in include/linux/pid.h
1606 */
52ee2dfd
ON
1607pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1608 struct pid_namespace *ns);
7af57294 1609
e868171a 1610static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1611{
1612 return tsk->pid;
1613}
1614
52ee2dfd
ON
1615static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1616 struct pid_namespace *ns)
1617{
1618 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1619}
7af57294
PE
1620
1621static inline pid_t task_pid_vnr(struct task_struct *tsk)
1622{
52ee2dfd 1623 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1624}
1625
1626
e868171a 1627static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1628{
1629 return tsk->tgid;
1630}
1631
2f2a3a46 1632pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1633
1634static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1635{
1636 return pid_vnr(task_tgid(tsk));
1637}
1638
1639
52ee2dfd
ON
1640static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1641 struct pid_namespace *ns)
7af57294 1642{
52ee2dfd 1643 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1644}
1645
7af57294
PE
1646static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1647{
52ee2dfd 1648 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1649}
1650
1651
52ee2dfd
ON
1652static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1653 struct pid_namespace *ns)
7af57294 1654{
52ee2dfd 1655 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1656}
1657
7af57294
PE
1658static inline pid_t task_session_vnr(struct task_struct *tsk)
1659{
52ee2dfd 1660 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1661}
1662
1b0f7ffd
ON
1663/* obsolete, do not use */
1664static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1665{
1666 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1667}
7af57294 1668
1da177e4
LT
1669/**
1670 * pid_alive - check that a task structure is not stale
1671 * @p: Task structure to be checked.
1672 *
1673 * Test if a process is not yet dead (at most zombie state)
1674 * If pid_alive fails, then pointers within the task structure
1675 * can be stale and must not be dereferenced.
1676 */
e868171a 1677static inline int pid_alive(struct task_struct *p)
1da177e4 1678{
92476d7f 1679 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1680}
1681
f400e198 1682/**
b460cbc5 1683 * is_global_init - check if a task structure is init
3260259f
HK
1684 * @tsk: Task structure to be checked.
1685 *
1686 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1687 */
e868171a 1688static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1689{
1690 return tsk->pid == 1;
1691}
b460cbc5
SH
1692
1693/*
1694 * is_container_init:
1695 * check whether in the task is init in its own pid namespace.
f400e198 1696 */
b461cc03 1697extern int is_container_init(struct task_struct *tsk);
f400e198 1698
9ec52099
CLG
1699extern struct pid *cad_pid;
1700
1da177e4 1701extern void free_task(struct task_struct *tsk);
1da177e4 1702#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1703
158d9ebd 1704extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1705
1706static inline void put_task_struct(struct task_struct *t)
1707{
1708 if (atomic_dec_and_test(&t->usage))
8c7904a0 1709 __put_task_struct(t);
e56d0903 1710}
1da177e4 1711
d180c5bc 1712extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1713extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1714
1da177e4
LT
1715/*
1716 * Per process flags
1717 */
6cdd5199 1718#define PF_KSOFTIRQD 0x00000001 /* I am ksoftirqd */
1da177e4
LT
1719#define PF_STARTING 0x00000002 /* being created */
1720#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1721#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1722#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1723#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1724#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1725#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1726#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1727#define PF_DUMPCORE 0x00000200 /* dumped core */
1728#define PF_SIGNALED 0x00000400 /* killed by a signal */
1729#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1da177e4 1730#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1731#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1732#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1733#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1734#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1735#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1736#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1737#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1738#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1739#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1740#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1741#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1742#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1743#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1744#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1745#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1746#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1747#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
ebb12db5 1748#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1749
1750/*
1751 * Only the _current_ task can read/write to tsk->flags, but other
1752 * tasks can access tsk->flags in readonly mode for example
1753 * with tsk_used_math (like during threaded core dumping).
1754 * There is however an exception to this rule during ptrace
1755 * or during fork: the ptracer task is allowed to write to the
1756 * child->flags of its traced child (same goes for fork, the parent
1757 * can write to the child->flags), because we're guaranteed the
1758 * child is not running and in turn not changing child->flags
1759 * at the same time the parent does it.
1760 */
1761#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1762#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1763#define clear_used_math() clear_stopped_child_used_math(current)
1764#define set_used_math() set_stopped_child_used_math(current)
1765#define conditional_stopped_child_used_math(condition, child) \
1766 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1767#define conditional_used_math(condition) \
1768 conditional_stopped_child_used_math(condition, current)
1769#define copy_to_stopped_child_used_math(child) \
1770 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1771/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1772#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1773#define used_math() tsk_used_math(current)
1774
a57eb940 1775#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1776
1777#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
24278d14
PM
1778#define RCU_READ_UNLOCK_BOOSTED (1 << 1) /* boosted while in RCU read-side. */
1779#define RCU_READ_UNLOCK_NEED_QS (1 << 2) /* RCU core needs CPU response. */
f41d911f
PM
1780
1781static inline void rcu_copy_process(struct task_struct *p)
1782{
1783 p->rcu_read_lock_nesting = 0;
1784 p->rcu_read_unlock_special = 0;
a57eb940 1785#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1786 p->rcu_blocked_node = NULL;
24278d14
PM
1787#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1788#ifdef CONFIG_RCU_BOOST
1789 p->rcu_boost_mutex = NULL;
1790#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1791 INIT_LIST_HEAD(&p->rcu_node_entry);
1792}
1793
f41d911f
PM
1794#else
1795
1796static inline void rcu_copy_process(struct task_struct *p)
1797{
1798}
1799
1800#endif
1801
1da177e4 1802#ifdef CONFIG_SMP
cd8ba7cd 1803extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1804 const struct cpumask *new_mask);
1da177e4 1805#else
cd8ba7cd 1806static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1807 const struct cpumask *new_mask)
1da177e4 1808{
96f874e2 1809 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1810 return -EINVAL;
1811 return 0;
1812}
1813#endif
e0ad9556
RR
1814
1815#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1816static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1817{
1818 return set_cpus_allowed_ptr(p, &new_mask);
1819}
e0ad9556 1820#endif
1da177e4 1821
b342501c 1822/*
c676329a
PZ
1823 * Do not use outside of architecture code which knows its limitations.
1824 *
1825 * sched_clock() has no promise of monotonicity or bounded drift between
1826 * CPUs, use (which you should not) requires disabling IRQs.
1827 *
1828 * Please use one of the three interfaces below.
b342501c 1829 */
1bbfa6f2 1830extern unsigned long long notrace sched_clock(void);
c676329a
PZ
1831/*
1832 * See the comment in kernel/sched_clock.c
1833 */
1834extern u64 cpu_clock(int cpu);
1835extern u64 local_clock(void);
1836extern u64 sched_clock_cpu(int cpu);
1837
e436d800 1838
c1955a3d 1839extern void sched_clock_init(void);
3e51f33f 1840
c1955a3d 1841#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1842static inline void sched_clock_tick(void)
1843{
1844}
1845
1846static inline void sched_clock_idle_sleep_event(void)
1847{
1848}
1849
1850static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1851{
1852}
1853#else
c676329a
PZ
1854/*
1855 * Architectures can set this to 1 if they have specified
1856 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1857 * but then during bootup it turns out that sched_clock()
1858 * is reliable after all:
1859 */
1860extern int sched_clock_stable;
1861
3e51f33f
PZ
1862extern void sched_clock_tick(void);
1863extern void sched_clock_idle_sleep_event(void);
1864extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1865#endif
1866
b52bfee4
VP
1867#ifdef CONFIG_IRQ_TIME_ACCOUNTING
1868/*
1869 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
1870 * The reason for this explicit opt-in is not to have perf penalty with
1871 * slow sched_clocks.
1872 */
1873extern void enable_sched_clock_irqtime(void);
1874extern void disable_sched_clock_irqtime(void);
1875#else
1876static inline void enable_sched_clock_irqtime(void) {}
1877static inline void disable_sched_clock_irqtime(void) {}
1878#endif
1879
36c8b586 1880extern unsigned long long
41b86e9c 1881task_sched_runtime(struct task_struct *task);
f06febc9 1882extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1883
1884/* sched_exec is called by processes performing an exec */
1885#ifdef CONFIG_SMP
1886extern void sched_exec(void);
1887#else
1888#define sched_exec() {}
1889#endif
1890
2aa44d05
IM
1891extern void sched_clock_idle_sleep_event(void);
1892extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1893
1da177e4
LT
1894#ifdef CONFIG_HOTPLUG_CPU
1895extern void idle_task_exit(void);
1896#else
1897static inline void idle_task_exit(void) {}
1898#endif
1899
06d8308c
TG
1900#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1901extern void wake_up_idle_cpu(int cpu);
1902#else
1903static inline void wake_up_idle_cpu(int cpu) { }
1904#endif
1905
21805085 1906extern unsigned int sysctl_sched_latency;
b2be5e96 1907extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1908extern unsigned int sysctl_sched_wakeup_granularity;
bf0f6f24 1909extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
1910
1911enum sched_tunable_scaling {
1912 SCHED_TUNABLESCALING_NONE,
1913 SCHED_TUNABLESCALING_LOG,
1914 SCHED_TUNABLESCALING_LINEAR,
1915 SCHED_TUNABLESCALING_END,
1916};
1917extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1918
2bba22c5 1919#ifdef CONFIG_SCHED_DEBUG
da84d961 1920extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1921extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1922extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1923extern unsigned int sysctl_timer_migration;
a7a4f8a7 1924extern unsigned int sysctl_sched_shares_window;
b2be5e96 1925
1983a922 1926int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1927 void __user *buffer, size_t *length,
b2be5e96 1928 loff_t *ppos);
2bd8e6d4 1929#endif
eea08f32
AB
1930#ifdef CONFIG_SCHED_DEBUG
1931static inline unsigned int get_sysctl_timer_migration(void)
1932{
1933 return sysctl_timer_migration;
1934}
1935#else
1936static inline unsigned int get_sysctl_timer_migration(void)
1937{
1938 return 1;
1939}
1940#endif
9f0c1e56
PZ
1941extern unsigned int sysctl_sched_rt_period;
1942extern int sysctl_sched_rt_runtime;
2bd8e6d4 1943
d0b27fa7 1944int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1945 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
1946 loff_t *ppos);
1947
2bd8e6d4 1948extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1949
5091faa4
MG
1950#ifdef CONFIG_SCHED_AUTOGROUP
1951extern unsigned int sysctl_sched_autogroup_enabled;
1952
1953extern void sched_autogroup_create_attach(struct task_struct *p);
1954extern void sched_autogroup_detach(struct task_struct *p);
1955extern void sched_autogroup_fork(struct signal_struct *sig);
1956extern void sched_autogroup_exit(struct signal_struct *sig);
1957#ifdef CONFIG_PROC_FS
1958extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
1959extern int proc_sched_autogroup_set_nice(struct task_struct *p, int *nice);
1960#endif
1961#else
1962static inline void sched_autogroup_create_attach(struct task_struct *p) { }
1963static inline void sched_autogroup_detach(struct task_struct *p) { }
1964static inline void sched_autogroup_fork(struct signal_struct *sig) { }
1965static inline void sched_autogroup_exit(struct signal_struct *sig) { }
1966#endif
1967
b29739f9 1968#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1969extern int rt_mutex_getprio(struct task_struct *p);
1970extern void rt_mutex_setprio(struct task_struct *p, int prio);
1971extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1972#else
e868171a 1973static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1974{
1975 return p->normal_prio;
1976}
95e02ca9 1977# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1978#endif
1979
36c8b586
IM
1980extern void set_user_nice(struct task_struct *p, long nice);
1981extern int task_prio(const struct task_struct *p);
1982extern int task_nice(const struct task_struct *p);
1983extern int can_nice(const struct task_struct *p, const int nice);
1984extern int task_curr(const struct task_struct *p);
1da177e4 1985extern int idle_cpu(int cpu);
fe7de49f
KM
1986extern int sched_setscheduler(struct task_struct *, int,
1987 const struct sched_param *);
961ccddd 1988extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 1989 const struct sched_param *);
36c8b586
IM
1990extern struct task_struct *idle_task(int cpu);
1991extern struct task_struct *curr_task(int cpu);
1992extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1993
1994void yield(void);
1995
1996/*
1997 * The default (Linux) execution domain.
1998 */
1999extern struct exec_domain default_exec_domain;
2000
2001union thread_union {
2002 struct thread_info thread_info;
2003 unsigned long stack[THREAD_SIZE/sizeof(long)];
2004};
2005
2006#ifndef __HAVE_ARCH_KSTACK_END
2007static inline int kstack_end(void *addr)
2008{
2009 /* Reliable end of stack detection:
2010 * Some APM bios versions misalign the stack
2011 */
2012 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2013}
2014#endif
2015
2016extern union thread_union init_thread_union;
2017extern struct task_struct init_task;
2018
2019extern struct mm_struct init_mm;
2020
198fe21b
PE
2021extern struct pid_namespace init_pid_ns;
2022
2023/*
2024 * find a task by one of its numerical ids
2025 *
198fe21b
PE
2026 * find_task_by_pid_ns():
2027 * finds a task by its pid in the specified namespace
228ebcbe
PE
2028 * find_task_by_vpid():
2029 * finds a task by its virtual pid
198fe21b 2030 *
e49859e7 2031 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2032 */
2033
228ebcbe
PE
2034extern struct task_struct *find_task_by_vpid(pid_t nr);
2035extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2036 struct pid_namespace *ns);
198fe21b 2037
8520d7c7 2038extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2039
2040/* per-UID process charging. */
acce292c 2041extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
2042static inline struct user_struct *get_uid(struct user_struct *u)
2043{
2044 atomic_inc(&u->__count);
2045 return u;
2046}
2047extern void free_uid(struct user_struct *);
28f300d2 2048extern void release_uids(struct user_namespace *ns);
1da177e4
LT
2049
2050#include <asm/current.h>
2051
3171a030 2052extern void do_timer(unsigned long ticks);
1da177e4 2053
b3c97528
HH
2054extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2055extern int wake_up_process(struct task_struct *tsk);
2056extern void wake_up_new_task(struct task_struct *tsk,
2057 unsigned long clone_flags);
1da177e4
LT
2058#ifdef CONFIG_SMP
2059 extern void kick_process(struct task_struct *tsk);
2060#else
2061 static inline void kick_process(struct task_struct *tsk) { }
2062#endif
ad46c2c4
IM
2063extern void sched_fork(struct task_struct *p, int clone_flags);
2064extern void sched_dead(struct task_struct *p);
1da177e4 2065
1da177e4
LT
2066extern void proc_caches_init(void);
2067extern void flush_signals(struct task_struct *);
3bcac026 2068extern void __flush_signals(struct task_struct *);
10ab825b 2069extern void ignore_signals(struct task_struct *);
1da177e4
LT
2070extern void flush_signal_handlers(struct task_struct *, int force_default);
2071extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2072
2073static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2074{
2075 unsigned long flags;
2076 int ret;
2077
2078 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2079 ret = dequeue_signal(tsk, mask, info);
2080 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2081
2082 return ret;
2083}
2084
2085extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2086 sigset_t *mask);
2087extern void unblock_all_signals(void);
2088extern void release_task(struct task_struct * p);
2089extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2090extern int force_sigsegv(int, struct task_struct *);
2091extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2092extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2093extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2094extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2095extern int kill_pgrp(struct pid *pid, int sig, int priv);
2096extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2097extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2098extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2099extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2100extern void force_sig(int, struct task_struct *);
1da177e4 2101extern int send_sig(int, struct task_struct *, int);
09faef11 2102extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2103extern struct sigqueue *sigqueue_alloc(void);
2104extern void sigqueue_free(struct sigqueue *);
ac5c2153 2105extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2106extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2107extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2108
9ec52099
CLG
2109static inline int kill_cad_pid(int sig, int priv)
2110{
2111 return kill_pid(cad_pid, sig, priv);
2112}
2113
1da177e4
LT
2114/* These can be the second arg to send_sig_info/send_group_sig_info. */
2115#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2116#define SEND_SIG_PRIV ((struct siginfo *) 1)
2117#define SEND_SIG_FORCED ((struct siginfo *) 2)
2118
2a855dd0
SAS
2119/*
2120 * True if we are on the alternate signal stack.
2121 */
1da177e4
LT
2122static inline int on_sig_stack(unsigned long sp)
2123{
2a855dd0
SAS
2124#ifdef CONFIG_STACK_GROWSUP
2125 return sp >= current->sas_ss_sp &&
2126 sp - current->sas_ss_sp < current->sas_ss_size;
2127#else
2128 return sp > current->sas_ss_sp &&
2129 sp - current->sas_ss_sp <= current->sas_ss_size;
2130#endif
1da177e4
LT
2131}
2132
2133static inline int sas_ss_flags(unsigned long sp)
2134{
2135 return (current->sas_ss_size == 0 ? SS_DISABLE
2136 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2137}
2138
1da177e4
LT
2139/*
2140 * Routines for handling mm_structs
2141 */
2142extern struct mm_struct * mm_alloc(void);
2143
2144/* mmdrop drops the mm and the page tables */
b3c97528 2145extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2146static inline void mmdrop(struct mm_struct * mm)
2147{
6fb43d7b 2148 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2149 __mmdrop(mm);
2150}
2151
2152/* mmput gets rid of the mappings and all user-space */
2153extern void mmput(struct mm_struct *);
2154/* Grab a reference to a task's mm, if it is not already going away */
2155extern struct mm_struct *get_task_mm(struct task_struct *task);
2156/* Remove the current tasks stale references to the old mm_struct */
2157extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2158/* Allocate a new mm structure and copy contents from tsk->mm */
2159extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2160
6f2c55b8
AD
2161extern int copy_thread(unsigned long, unsigned long, unsigned long,
2162 struct task_struct *, struct pt_regs *);
1da177e4
LT
2163extern void flush_thread(void);
2164extern void exit_thread(void);
2165
1da177e4 2166extern void exit_files(struct task_struct *);
a7e5328a 2167extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2168
1da177e4 2169extern void exit_itimers(struct signal_struct *);
cbaffba1 2170extern void flush_itimer_signals(void);
1da177e4
LT
2171
2172extern NORET_TYPE void do_group_exit(int);
2173
1da177e4
LT
2174extern void daemonize(const char *, ...);
2175extern int allow_signal(int);
2176extern int disallow_signal(int);
1da177e4 2177
d7627467
DH
2178extern int do_execve(const char *,
2179 const char __user * const __user *,
2180 const char __user * const __user *, struct pt_regs *);
1da177e4 2181extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2182struct task_struct *fork_idle(int);
1da177e4
LT
2183
2184extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2185extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2186
2187#ifdef CONFIG_SMP
85ba2d86 2188extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2189#else
85ba2d86
RM
2190static inline unsigned long wait_task_inactive(struct task_struct *p,
2191 long match_state)
2192{
2193 return 1;
2194}
1da177e4
LT
2195#endif
2196
05725f7e
JP
2197#define next_task(p) \
2198 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2199
2200#define for_each_process(p) \
2201 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2202
5bb459bb 2203extern bool current_is_single_threaded(void);
d84f4f99 2204
1da177e4
LT
2205/*
2206 * Careful: do_each_thread/while_each_thread is a double loop so
2207 * 'break' will not work as expected - use goto instead.
2208 */
2209#define do_each_thread(g, t) \
2210 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2211
2212#define while_each_thread(g, t) \
2213 while ((t = next_thread(t)) != g)
2214
7e49827c
ON
2215static inline int get_nr_threads(struct task_struct *tsk)
2216{
b3ac022c 2217 return tsk->signal->nr_threads;
7e49827c
ON
2218}
2219
de12a787
EB
2220/* de_thread depends on thread_group_leader not being a pid based check */
2221#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2222
0804ef4b
EB
2223/* Do to the insanities of de_thread it is possible for a process
2224 * to have the pid of the thread group leader without actually being
2225 * the thread group leader. For iteration through the pids in proc
2226 * all we care about is that we have a task with the appropriate
2227 * pid, we don't actually care if we have the right task.
2228 */
e868171a 2229static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2230{
2231 return p->pid == p->tgid;
2232}
2233
bac0abd6
PE
2234static inline
2235int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2236{
2237 return p1->tgid == p2->tgid;
2238}
2239
36c8b586 2240static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2241{
05725f7e
JP
2242 return list_entry_rcu(p->thread_group.next,
2243 struct task_struct, thread_group);
47e65328
ON
2244}
2245
e868171a 2246static inline int thread_group_empty(struct task_struct *p)
1da177e4 2247{
47e65328 2248 return list_empty(&p->thread_group);
1da177e4
LT
2249}
2250
2251#define delay_group_leader(p) \
2252 (thread_group_leader(p) && !thread_group_empty(p))
2253
39c626ae
ON
2254static inline int task_detached(struct task_struct *p)
2255{
2256 return p->exit_signal == -1;
2257}
2258
1da177e4 2259/*
260ea101 2260 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2261 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2262 * pins the final release of task.io_context. Also protects ->cpuset and
2263 * ->cgroup.subsys[].
1da177e4
LT
2264 *
2265 * Nests both inside and outside of read_lock(&tasklist_lock).
2266 * It must not be nested with write_lock_irq(&tasklist_lock),
2267 * neither inside nor outside.
2268 */
2269static inline void task_lock(struct task_struct *p)
2270{
2271 spin_lock(&p->alloc_lock);
2272}
2273
2274static inline void task_unlock(struct task_struct *p)
2275{
2276 spin_unlock(&p->alloc_lock);
2277}
2278
b8ed374e 2279extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2280 unsigned long *flags);
2281
b8ed374e
NK
2282#define lock_task_sighand(tsk, flags) \
2283({ struct sighand_struct *__ss; \
2284 __cond_lock(&(tsk)->sighand->siglock, \
2285 (__ss = __lock_task_sighand(tsk, flags))); \
2286 __ss; \
2287}) \
2288
f63ee72e
ON
2289static inline void unlock_task_sighand(struct task_struct *tsk,
2290 unsigned long *flags)
2291{
2292 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2293}
2294
f037360f
AV
2295#ifndef __HAVE_THREAD_FUNCTIONS
2296
f7e4217b
RZ
2297#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2298#define task_stack_page(task) ((task)->stack)
a1261f54 2299
10ebffde
AV
2300static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2301{
2302 *task_thread_info(p) = *task_thread_info(org);
2303 task_thread_info(p)->task = p;
2304}
2305
2306static inline unsigned long *end_of_stack(struct task_struct *p)
2307{
f7e4217b 2308 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2309}
2310
f037360f
AV
2311#endif
2312
8b05c7e6
FT
2313static inline int object_is_on_stack(void *obj)
2314{
2315 void *stack = task_stack_page(current);
2316
2317 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2318}
2319
8c9843e5
BH
2320extern void thread_info_cache_init(void);
2321
7c9f8861
ES
2322#ifdef CONFIG_DEBUG_STACK_USAGE
2323static inline unsigned long stack_not_used(struct task_struct *p)
2324{
2325 unsigned long *n = end_of_stack(p);
2326
2327 do { /* Skip over canary */
2328 n++;
2329 } while (!*n);
2330
2331 return (unsigned long)n - (unsigned long)end_of_stack(p);
2332}
2333#endif
2334
1da177e4
LT
2335/* set thread flags in other task's structures
2336 * - see asm/thread_info.h for TIF_xxxx flags available
2337 */
2338static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2339{
a1261f54 2340 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2341}
2342
2343static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2344{
a1261f54 2345 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2346}
2347
2348static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2349{
a1261f54 2350 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2351}
2352
2353static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2354{
a1261f54 2355 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2356}
2357
2358static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2359{
a1261f54 2360 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2361}
2362
2363static inline void set_tsk_need_resched(struct task_struct *tsk)
2364{
2365 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2366}
2367
2368static inline void clear_tsk_need_resched(struct task_struct *tsk)
2369{
2370 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2371}
2372
8ae121ac
GH
2373static inline int test_tsk_need_resched(struct task_struct *tsk)
2374{
2375 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2376}
2377
690cc3ff
EB
2378static inline int restart_syscall(void)
2379{
2380 set_tsk_thread_flag(current, TIF_SIGPENDING);
2381 return -ERESTARTNOINTR;
2382}
2383
1da177e4
LT
2384static inline int signal_pending(struct task_struct *p)
2385{
2386 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2387}
f776d12d 2388
d9588725
RM
2389static inline int __fatal_signal_pending(struct task_struct *p)
2390{
2391 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2392}
f776d12d
MW
2393
2394static inline int fatal_signal_pending(struct task_struct *p)
2395{
2396 return signal_pending(p) && __fatal_signal_pending(p);
2397}
2398
16882c1e
ON
2399static inline int signal_pending_state(long state, struct task_struct *p)
2400{
2401 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2402 return 0;
2403 if (!signal_pending(p))
2404 return 0;
2405
16882c1e
ON
2406 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2407}
2408
1da177e4
LT
2409static inline int need_resched(void)
2410{
9404ef02 2411 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2412}
2413
2414/*
2415 * cond_resched() and cond_resched_lock(): latency reduction via
2416 * explicit rescheduling in places that are safe. The return
2417 * value indicates whether a reschedule was done in fact.
2418 * cond_resched_lock() will drop the spinlock before scheduling,
2419 * cond_resched_softirq() will enable bhs before scheduling.
2420 */
c3921ab7 2421extern int _cond_resched(void);
6f80bd98 2422
613afbf8
FW
2423#define cond_resched() ({ \
2424 __might_sleep(__FILE__, __LINE__, 0); \
2425 _cond_resched(); \
2426})
6f80bd98 2427
613afbf8
FW
2428extern int __cond_resched_lock(spinlock_t *lock);
2429
716a4234
FW
2430#ifdef CONFIG_PREEMPT
2431#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2432#else
716a4234 2433#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2434#endif
716a4234 2435
613afbf8 2436#define cond_resched_lock(lock) ({ \
716a4234 2437 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2438 __cond_resched_lock(lock); \
2439})
2440
2441extern int __cond_resched_softirq(void);
2442
75e1056f
VP
2443#define cond_resched_softirq() ({ \
2444 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2445 __cond_resched_softirq(); \
613afbf8 2446})
1da177e4
LT
2447
2448/*
2449 * Does a critical section need to be broken due to another
95c354fe
NP
2450 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2451 * but a general need for low latency)
1da177e4 2452 */
95c354fe 2453static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2454{
95c354fe
NP
2455#ifdef CONFIG_PREEMPT
2456 return spin_is_contended(lock);
2457#else
1da177e4 2458 return 0;
95c354fe 2459#endif
1da177e4
LT
2460}
2461
f06febc9
FM
2462/*
2463 * Thread group CPU time accounting.
2464 */
4cd4c1b4 2465void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2466void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2467
490dea45 2468static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2469{
4cd4c1b4 2470 spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2471}
2472
7bb44ade
RM
2473/*
2474 * Reevaluate whether the task has signals pending delivery.
2475 * Wake the task if so.
2476 * This is required every time the blocked sigset_t changes.
2477 * callers must hold sighand->siglock.
2478 */
2479extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2480extern void recalc_sigpending(void);
2481
2482extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2483
2484/*
2485 * Wrappers for p->thread_info->cpu access. No-op on UP.
2486 */
2487#ifdef CONFIG_SMP
2488
2489static inline unsigned int task_cpu(const struct task_struct *p)
2490{
a1261f54 2491 return task_thread_info(p)->cpu;
1da177e4
LT
2492}
2493
c65cc870 2494extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2495
2496#else
2497
2498static inline unsigned int task_cpu(const struct task_struct *p)
2499{
2500 return 0;
2501}
2502
2503static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2504{
2505}
2506
2507#endif /* CONFIG_SMP */
2508
96f874e2
RR
2509extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2510extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2511
1da177e4
LT
2512extern void normalize_rt_tasks(void);
2513
7c941438 2514#ifdef CONFIG_CGROUP_SCHED
9b5b7751 2515
07e06b01 2516extern struct task_group root_task_group;
9b5b7751 2517
ec7dc8ac 2518extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2519extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2520extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2521#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2522extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2523extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2524#endif
2525#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2526extern int sched_group_set_rt_runtime(struct task_group *tg,
2527 long rt_runtime_us);
2528extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2529extern int sched_group_set_rt_period(struct task_group *tg,
2530 long rt_period_us);
2531extern long sched_group_rt_period(struct task_group *tg);
54e99124 2532extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2533#endif
9b5b7751
SV
2534#endif
2535
54e99124
DG
2536extern int task_can_switch_user(struct user_struct *up,
2537 struct task_struct *tsk);
2538
4b98d11b
AD
2539#ifdef CONFIG_TASK_XACCT
2540static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2541{
940389b8 2542 tsk->ioac.rchar += amt;
4b98d11b
AD
2543}
2544
2545static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2546{
940389b8 2547 tsk->ioac.wchar += amt;
4b98d11b
AD
2548}
2549
2550static inline void inc_syscr(struct task_struct *tsk)
2551{
940389b8 2552 tsk->ioac.syscr++;
4b98d11b
AD
2553}
2554
2555static inline void inc_syscw(struct task_struct *tsk)
2556{
940389b8 2557 tsk->ioac.syscw++;
4b98d11b
AD
2558}
2559#else
2560static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2561{
2562}
2563
2564static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2565{
2566}
2567
2568static inline void inc_syscr(struct task_struct *tsk)
2569{
2570}
2571
2572static inline void inc_syscw(struct task_struct *tsk)
2573{
2574}
2575#endif
2576
82455257
DH
2577#ifndef TASK_SIZE_OF
2578#define TASK_SIZE_OF(tsk) TASK_SIZE
2579#endif
2580
0793a61d
TG
2581/*
2582 * Call the function if the target task is executing on a CPU right now:
2583 */
2584extern void task_oncpu_function_call(struct task_struct *p,
2585 void (*func) (void *info), void *info);
2586
2587
cf475ad2
BS
2588#ifdef CONFIG_MM_OWNER
2589extern void mm_update_next_owner(struct mm_struct *mm);
2590extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2591#else
2592static inline void mm_update_next_owner(struct mm_struct *mm)
2593{
2594}
2595
2596static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2597{
2598}
2599#endif /* CONFIG_MM_OWNER */
2600
3e10e716
JS
2601static inline unsigned long task_rlimit(const struct task_struct *tsk,
2602 unsigned int limit)
2603{
2604 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2605}
2606
2607static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2608 unsigned int limit)
2609{
2610 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2611}
2612
2613static inline unsigned long rlimit(unsigned int limit)
2614{
2615 return task_rlimit(current, limit);
2616}
2617
2618static inline unsigned long rlimit_max(unsigned int limit)
2619{
2620 return task_rlimit_max(current, limit);
2621}
2622
1da177e4
LT
2623#endif /* __KERNEL__ */
2624
2625#endif