]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/linux/sched.h
sched: Improve balance_cpu() to consider other cpus in its group as target of (pinned...
[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 65
1da177e4
LT
66#include <asm/page.h>
67#include <asm/ptrace.h>
1da177e4
LT
68#include <asm/cputime.h>
69
70#include <linux/smp.h>
71#include <linux/sem.h>
72#include <linux/signal.h>
1da177e4
LT
73#include <linux/compiler.h>
74#include <linux/completion.h>
75#include <linux/pid.h>
76#include <linux/percpu.h>
77#include <linux/topology.h>
3e26c149 78#include <linux/proportions.h>
1da177e4 79#include <linux/seccomp.h>
e56d0903 80#include <linux/rcupdate.h>
05725f7e 81#include <linux/rculist.h>
23f78d4a 82#include <linux/rtmutex.h>
1da177e4 83
a3b6714e
DW
84#include <linux/time.h>
85#include <linux/param.h>
86#include <linux/resource.h>
87#include <linux/timer.h>
88#include <linux/hrtimer.h>
7c3ab738 89#include <linux/task_io_accounting.h>
9745512c 90#include <linux/latencytop.h>
9e2b2dc4 91#include <linux/cred.h>
fa14ff4a 92#include <linux/llist.h>
7b44ab97 93#include <linux/uidgid.h>
a3b6714e
DW
94
95#include <asm/processor.h>
36d57ac4 96
1da177e4 97struct exec_domain;
c87e2837 98struct futex_pi_state;
286100a6 99struct robust_list_head;
bddd87c7 100struct bio_list;
5ad4e53b 101struct fs_struct;
cdd6c482 102struct perf_event_context;
73c10101 103struct blk_plug;
1da177e4 104
1da177e4
LT
105/*
106 * List of flags we want to share for kernel threads,
107 * if only because they are not used by them anyway.
108 */
109#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
110
111/*
112 * These are the constant used to fake the fixed-point load-average
113 * counting. Some notes:
114 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
115 * a load-average precision of 10 bits integer + 11 bits fractional
116 * - if you want to count load-averages more often, you need more
117 * precision, or rounding will get you. With 2-second counting freq,
118 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 * 11 bit fractions.
120 */
121extern unsigned long avenrun[]; /* Load averages */
2d02494f 122extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
123
124#define FSHIFT 11 /* nr of bits of precision */
125#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 126#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
127#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
128#define EXP_5 2014 /* 1/exp(5sec/5min) */
129#define EXP_15 2037 /* 1/exp(5sec/15min) */
130
131#define CALC_LOAD(load,exp,n) \
132 load *= exp; \
133 load += n*(FIXED_1-exp); \
134 load >>= FSHIFT;
135
136extern unsigned long total_forks;
137extern int nr_threads;
1da177e4
LT
138DECLARE_PER_CPU(unsigned long, process_counts);
139extern int nr_processes(void);
140extern unsigned long nr_running(void);
141extern unsigned long nr_uninterruptible(void);
142extern unsigned long nr_iowait(void);
8c215bd3 143extern unsigned long nr_iowait_cpu(int cpu);
69d25870
AV
144extern unsigned long this_cpu_load(void);
145
146
0f004f5a 147extern void calc_global_load(unsigned long ticks);
5aaa0b7a 148extern void update_cpu_load_nohz(void);
1da177e4 149
7e49fcce
SR
150extern unsigned long get_parent_ip(unsigned long addr);
151
43ae34cb
IM
152struct seq_file;
153struct cfs_rq;
4cf86d77 154struct task_group;
43ae34cb
IM
155#ifdef CONFIG_SCHED_DEBUG
156extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
157extern void proc_sched_set_task(struct task_struct *p);
158extern void
5cef9eca 159print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
160#else
161static inline void
162proc_sched_show_task(struct task_struct *p, struct seq_file *m)
163{
164}
165static inline void proc_sched_set_task(struct task_struct *p)
166{
167}
168static inline void
5cef9eca 169print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
170{
171}
172#endif
1da177e4 173
4a8342d2
LT
174/*
175 * Task state bitmask. NOTE! These bits are also
176 * encoded in fs/proc/array.c: get_task_state().
177 *
178 * We have two separate sets of flags: task->state
179 * is about runnability, while task->exit_state are
180 * about the task exiting. Confusing, but this way
181 * modifying one set can't modify the other one by
182 * mistake.
183 */
1da177e4
LT
184#define TASK_RUNNING 0
185#define TASK_INTERRUPTIBLE 1
186#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
187#define __TASK_STOPPED 4
188#define __TASK_TRACED 8
4a8342d2
LT
189/* in tsk->exit_state */
190#define EXIT_ZOMBIE 16
191#define EXIT_DEAD 32
192/* in tsk->state again */
af927232 193#define TASK_DEAD 64
f021a3c2 194#define TASK_WAKEKILL 128
e9c84311 195#define TASK_WAKING 256
e1781538 196#define TASK_STATE_MAX 512
f021a3c2 197
44d90df6 198#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
73342151 199
e1781538
PZ
200extern char ___assert_task_state[1 - 2*!!(
201 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
202
203/* Convenience macros for the sake of set_task_state */
204#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
205#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
206#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 207
92a1f4bc
MW
208/* Convenience macros for the sake of wake_up */
209#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 210#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
211
212/* get_task_state() */
213#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
214 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
215 __TASK_TRACED)
92a1f4bc 216
f021a3c2
MW
217#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
218#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
8f92054e 219#define task_is_dead(task) ((task)->exit_state != 0)
92a1f4bc 220#define task_is_stopped_or_traced(task) \
f021a3c2 221 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 222#define task_contributes_to_load(task) \
e3c8ca83 223 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
376fede8 224 (task->flags & PF_FROZEN) == 0)
1da177e4
LT
225
226#define __set_task_state(tsk, state_value) \
227 do { (tsk)->state = (state_value); } while (0)
228#define set_task_state(tsk, state_value) \
229 set_mb((tsk)->state, (state_value))
230
498d0c57
AM
231/*
232 * set_current_state() includes a barrier so that the write of current->state
233 * is correctly serialised wrt the caller's subsequent test of whether to
234 * actually sleep:
235 *
236 * set_current_state(TASK_UNINTERRUPTIBLE);
237 * if (do_i_need_to_sleep())
238 * schedule();
239 *
240 * If the caller does not need such serialisation then use __set_current_state()
241 */
1da177e4
LT
242#define __set_current_state(state_value) \
243 do { current->state = (state_value); } while (0)
244#define set_current_state(state_value) \
245 set_mb(current->state, (state_value))
246
247/* Task command name length */
248#define TASK_COMM_LEN 16
249
1da177e4
LT
250#include <linux/spinlock.h>
251
252/*
253 * This serializes "schedule()" and also protects
254 * the run-queue from deletions/modifications (but
255 * _adding_ to the beginning of the run-queue has
256 * a separate lock).
257 */
258extern rwlock_t tasklist_lock;
259extern spinlock_t mmlist_lock;
260
36c8b586 261struct task_struct;
1da177e4 262
db1466b3
PM
263#ifdef CONFIG_PROVE_RCU
264extern int lockdep_tasklist_lock_is_held(void);
265#endif /* #ifdef CONFIG_PROVE_RCU */
266
1da177e4
LT
267extern void sched_init(void);
268extern void sched_init_smp(void);
2d07b255 269extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 270extern void init_idle(struct task_struct *idle, int cpu);
1df21055 271extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 272
89f19f04 273extern int runqueue_is_locked(int cpu);
017730c1 274
46cb4b7c 275#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
83cd4fe2 276extern void select_nohz_load_balancer(int stop_tick);
69e1e811 277extern void set_cpu_sd_state_idle(void);
83cd4fe2 278extern int get_nohz_timer_target(void);
46cb4b7c 279#else
83cd4fe2 280static inline void select_nohz_load_balancer(int stop_tick) { }
fdaabd80 281static inline void set_cpu_sd_state_idle(void) { }
46cb4b7c 282#endif
1da177e4 283
e59e2ae2 284/*
39bc89fd 285 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
286 */
287extern void show_state_filter(unsigned long state_filter);
288
289static inline void show_state(void)
290{
39bc89fd 291 show_state_filter(0);
e59e2ae2
IM
292}
293
1da177e4
LT
294extern void show_regs(struct pt_regs *);
295
296/*
297 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
298 * task), SP is the stack pointer of the first frame that should be shown in the back
299 * trace (or NULL if the entire call-chain of the task should be shown).
300 */
301extern void show_stack(struct task_struct *task, unsigned long *sp);
302
303void io_schedule(void);
304long io_schedule_timeout(long timeout);
305
306extern void cpu_init (void);
307extern void trap_init(void);
308extern void update_process_times(int user);
309extern void scheduler_tick(void);
310
82a1fcb9
IM
311extern void sched_show_task(struct task_struct *p);
312
19cc36c0 313#ifdef CONFIG_LOCKUP_DETECTOR
8446f1d3 314extern void touch_softlockup_watchdog(void);
d6ad3e28 315extern void touch_softlockup_watchdog_sync(void);
04c9167f 316extern void touch_all_softlockup_watchdogs(void);
332fbdbc
DZ
317extern int proc_dowatchdog_thresh(struct ctl_table *table, int write,
318 void __user *buffer,
319 size_t *lenp, loff_t *ppos);
9c44bc03 320extern unsigned int softlockup_panic;
004417a6 321void lockup_detector_init(void);
8446f1d3 322#else
8446f1d3
IM
323static inline void touch_softlockup_watchdog(void)
324{
325}
d6ad3e28
JW
326static inline void touch_softlockup_watchdog_sync(void)
327{
328}
04c9167f
JF
329static inline void touch_all_softlockup_watchdogs(void)
330{
331}
004417a6
PZ
332static inline void lockup_detector_init(void)
333{
334}
8446f1d3
IM
335#endif
336
e162b39a
MSB
337#ifdef CONFIG_DETECT_HUNG_TASK
338extern unsigned int sysctl_hung_task_panic;
339extern unsigned long sysctl_hung_task_check_count;
340extern unsigned long sysctl_hung_task_timeout_secs;
341extern unsigned long sysctl_hung_task_warnings;
342extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 343 void __user *buffer,
e162b39a 344 size_t *lenp, loff_t *ppos);
e4ecda1b
ML
345#else
346/* Avoid need for ifdefs elsewhere in the code */
347enum { sysctl_hung_task_timeout_secs = 0 };
e162b39a 348#endif
8446f1d3 349
1da177e4
LT
350/* Attach to any functions which should be ignored in wchan output. */
351#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
352
353/* Linker adds these: start and end of __sched functions */
354extern char __sched_text_start[], __sched_text_end[];
355
1da177e4
LT
356/* Is this address in the __sched functions? */
357extern int in_sched_functions(unsigned long addr);
358
359#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 360extern signed long schedule_timeout(signed long timeout);
64ed93a2 361extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 362extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 363extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 364asmlinkage void schedule(void);
c5491ea7 365extern void schedule_preempt_disabled(void);
c6eb3dda 366extern int mutex_spin_on_owner(struct mutex *lock, struct task_struct *owner);
1da177e4 367
ab516013 368struct nsproxy;
acce292c 369struct user_namespace;
1da177e4 370
341c87bf
KH
371/*
372 * Default maximum number of active map areas, this limits the number of vmas
373 * per mm struct. Users can overwrite this number by sysctl but there is a
374 * problem.
375 *
376 * When a program's coredump is generated as ELF format, a section is created
377 * per a vma. In ELF, the number of sections is represented in unsigned short.
378 * This means the number of sections should be smaller than 65535 at coredump.
379 * Because the kernel adds some informative sections to a image of program at
380 * generating coredump, we need some margin. The number of extra sections is
381 * 1-3 now and depends on arch. We use "5" as safe margin, here.
382 */
383#define MAPCOUNT_ELF_CORE_MARGIN (5)
4be929be 384#define DEFAULT_MAX_MAP_COUNT (USHRT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
385
386extern int sysctl_max_map_count;
387
388#include <linux/aio.h>
389
efc1a3b1
DH
390#ifdef CONFIG_MMU
391extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
392extern unsigned long
393arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
394 unsigned long, unsigned long);
395extern unsigned long
396arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
397 unsigned long len, unsigned long pgoff,
398 unsigned long flags);
1363c3cd
WW
399extern void arch_unmap_area(struct mm_struct *, unsigned long);
400extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
efc1a3b1
DH
401#else
402static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
403#endif
1da177e4 404
901608d9 405
6c5d5238
KH
406extern void set_dumpable(struct mm_struct *mm, int value);
407extern int get_dumpable(struct mm_struct *mm);
408
409/* mm flags */
3cb4a0bb 410/* dumpable bits */
6c5d5238
KH
411#define MMF_DUMPABLE 0 /* core dump is permitted */
412#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 413
3cb4a0bb 414#define MMF_DUMPABLE_BITS 2
f8af4da3 415#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH
416
417/* coredump filter bits */
418#define MMF_DUMP_ANON_PRIVATE 2
419#define MMF_DUMP_ANON_SHARED 3
420#define MMF_DUMP_MAPPED_PRIVATE 4
421#define MMF_DUMP_MAPPED_SHARED 5
82df3973 422#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
423#define MMF_DUMP_HUGETLB_PRIVATE 7
424#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 425
3cb4a0bb 426#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 427#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
428#define MMF_DUMP_FILTER_MASK \
429 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
430#define MMF_DUMP_FILTER_DEFAULT \
e575f111 431 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
432 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
433
434#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
435# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
436#else
437# define MMF_DUMP_MASK_DEFAULT_ELF 0
438#endif
f8af4da3
HD
439 /* leave room for more dump flags */
440#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
ba76149f 441#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
bafb282d 442#define MMF_EXE_FILE_CHANGED 18 /* see prctl_set_mm_exe_file() */
f8af4da3
HD
443
444#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 445
1da177e4
LT
446struct sighand_struct {
447 atomic_t count;
448 struct k_sigaction action[_NSIG];
449 spinlock_t siglock;
b8fceee1 450 wait_queue_head_t signalfd_wqh;
1da177e4
LT
451};
452
0e464814 453struct pacct_struct {
f6ec29a4
KK
454 int ac_flag;
455 long ac_exitcode;
0e464814 456 unsigned long ac_mem;
77787bfb
KK
457 cputime_t ac_utime, ac_stime;
458 unsigned long ac_minflt, ac_majflt;
0e464814
KK
459};
460
42c4ab41
SG
461struct cpu_itimer {
462 cputime_t expires;
463 cputime_t incr;
8356b5f9
SG
464 u32 error;
465 u32 incr_error;
42c4ab41
SG
466};
467
f06febc9
FM
468/**
469 * struct task_cputime - collected CPU time counts
470 * @utime: time spent in user mode, in &cputime_t units
471 * @stime: time spent in kernel mode, in &cputime_t units
472 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 473 *
f06febc9
FM
474 * This structure groups together three kinds of CPU time that are
475 * tracked for threads and thread groups. Most things considering
476 * CPU time want to group these counts together and treat all three
477 * of them in parallel.
478 */
479struct task_cputime {
480 cputime_t utime;
481 cputime_t stime;
482 unsigned long long sum_exec_runtime;
483};
484/* Alternate field names when used to cache expirations. */
485#define prof_exp stime
486#define virt_exp utime
487#define sched_exp sum_exec_runtime
488
4cd4c1b4
PZ
489#define INIT_CPUTIME \
490 (struct task_cputime) { \
64861634
MS
491 .utime = 0, \
492 .stime = 0, \
4cd4c1b4
PZ
493 .sum_exec_runtime = 0, \
494 }
495
c99e6efe
PZ
496/*
497 * Disable preemption until the scheduler is running.
498 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
499 *
500 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
501 * before the scheduler is active -- see should_resched().
c99e6efe 502 */
d86ee480 503#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 504
f06febc9 505/**
4cd4c1b4
PZ
506 * struct thread_group_cputimer - thread group interval timer counts
507 * @cputime: thread group interval timers.
508 * @running: non-zero when there are timers running and
509 * @cputime receives updates.
510 * @lock: lock for fields in this struct.
f06febc9
FM
511 *
512 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 513 * used for thread group CPU timer calculations.
f06febc9 514 */
4cd4c1b4
PZ
515struct thread_group_cputimer {
516 struct task_cputime cputime;
517 int running;
ee30a7b2 518 raw_spinlock_t lock;
f06febc9 519};
f06febc9 520
4714d1d3 521#include <linux/rwsem.h>
5091faa4
MG
522struct autogroup;
523
1da177e4 524/*
e815f0a8 525 * NOTE! "signal_struct" does not have its own
1da177e4
LT
526 * locking, because a shared signal_struct always
527 * implies a shared sighand_struct, so locking
528 * sighand_struct is always a proper superset of
529 * the locking of signal_struct.
530 */
531struct signal_struct {
ea6d290c 532 atomic_t sigcnt;
1da177e4 533 atomic_t live;
b3ac022c 534 int nr_threads;
1da177e4
LT
535
536 wait_queue_head_t wait_chldexit; /* for wait4() */
537
538 /* current thread group signal load-balancing target: */
36c8b586 539 struct task_struct *curr_target;
1da177e4
LT
540
541 /* shared signal handling: */
542 struct sigpending shared_pending;
543
544 /* thread group exit support */
545 int group_exit_code;
546 /* overloaded:
547 * - notify group_exit_task when ->count is equal to notify_count
548 * - everyone except group_exit_task is stopped during signal delivery
549 * of fatal signals, group_exit_task processes the signal.
550 */
1da177e4 551 int notify_count;
07dd20e0 552 struct task_struct *group_exit_task;
1da177e4
LT
553
554 /* thread group stop support, overloads group_exit_code too */
555 int group_stop_count;
556 unsigned int flags; /* see SIGNAL_* flags below */
557
ebec18a6
LP
558 /*
559 * PR_SET_CHILD_SUBREAPER marks a process, like a service
560 * manager, to re-parent orphan (double-forking) child processes
561 * to this process instead of 'init'. The service manager is
562 * able to receive SIGCHLD signals and is able to investigate
563 * the process until it calls wait(). All children of this
564 * process will inherit a flag if they should look for a
565 * child_subreaper process at exit.
566 */
567 unsigned int is_child_subreaper:1;
568 unsigned int has_child_subreaper:1;
569
1da177e4
LT
570 /* POSIX.1b Interval Timers */
571 struct list_head posix_timers;
572
573 /* ITIMER_REAL timer for the process */
2ff678b8 574 struct hrtimer real_timer;
fea9d175 575 struct pid *leader_pid;
2ff678b8 576 ktime_t it_real_incr;
1da177e4 577
42c4ab41
SG
578 /*
579 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
580 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
581 * values are defined to 0 and 1 respectively
582 */
583 struct cpu_itimer it[2];
1da177e4 584
f06febc9 585 /*
4cd4c1b4
PZ
586 * Thread group totals for process CPU timers.
587 * See thread_group_cputimer(), et al, for details.
f06febc9 588 */
4cd4c1b4 589 struct thread_group_cputimer cputimer;
f06febc9
FM
590
591 /* Earliest-expiration cache. */
592 struct task_cputime cputime_expires;
593
594 struct list_head cpu_timers[3];
595
ab521dc0 596 struct pid *tty_old_pgrp;
1ec320af 597
1da177e4
LT
598 /* boolean value for session group leader */
599 int leader;
600
601 struct tty_struct *tty; /* NULL if no tty */
602
5091faa4
MG
603#ifdef CONFIG_SCHED_AUTOGROUP
604 struct autogroup *autogroup;
605#endif
1da177e4
LT
606 /*
607 * Cumulative resource counters for dead threads in the group,
608 * and for reaped dead child processes forked by this group.
609 * Live threads maintain their own counters and add to these
610 * in __exit_signal, except for the group leader.
611 */
32bd671d 612 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
613 cputime_t gtime;
614 cputime_t cgtime;
0cf55e1e
HS
615#ifndef CONFIG_VIRT_CPU_ACCOUNTING
616 cputime_t prev_utime, prev_stime;
617#endif
1da177e4
LT
618 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
619 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 620 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 621 unsigned long maxrss, cmaxrss;
940389b8 622 struct task_io_accounting ioac;
1da177e4 623
32bd671d
PZ
624 /*
625 * Cumulative ns of schedule CPU time fo dead threads in the
626 * group, not including a zombie group leader, (This only differs
627 * from jiffies_to_ns(utime + stime) if sched_clock uses something
628 * other than jiffies.)
629 */
630 unsigned long long sum_sched_runtime;
631
1da177e4
LT
632 /*
633 * We don't bother to synchronize most readers of this at all,
634 * because there is no reader checking a limit that actually needs
635 * to get both rlim_cur and rlim_max atomically, and either one
636 * alone is a single word that can safely be read normally.
637 * getrlimit/setrlimit use task_lock(current->group_leader) to
638 * protect this instead of the siglock, because they really
639 * have no need to disable irqs.
640 */
641 struct rlimit rlim[RLIM_NLIMITS];
642
0e464814
KK
643#ifdef CONFIG_BSD_PROCESS_ACCT
644 struct pacct_struct pacct; /* per-process accounting information */
645#endif
ad4ecbcb 646#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
647 struct taskstats *stats;
648#endif
522ed776
MT
649#ifdef CONFIG_AUDIT
650 unsigned audit_tty;
651 struct tty_audit_buf *tty_audit_buf;
652#endif
4714d1d3
BB
653#ifdef CONFIG_CGROUPS
654 /*
77e4ef99
TH
655 * group_rwsem prevents new tasks from entering the threadgroup and
656 * member tasks from exiting,a more specifically, setting of
657 * PF_EXITING. fork and exit paths are protected with this rwsem
658 * using threadgroup_change_begin/end(). Users which require
659 * threadgroup to remain stable should use threadgroup_[un]lock()
660 * which also takes care of exec path. Currently, cgroup is the
661 * only user.
4714d1d3 662 */
257058ae 663 struct rw_semaphore group_rwsem;
4714d1d3 664#endif
28b83c51 665
a63d83f4
DR
666 int oom_adj; /* OOM kill score adjustment (bit shift) */
667 int oom_score_adj; /* OOM kill score adjustment */
dabb16f6
MSB
668 int oom_score_adj_min; /* OOM kill score adjustment minimum value.
669 * Only settable by CAP_SYS_RESOURCE. */
9b1bf12d
KM
670
671 struct mutex cred_guard_mutex; /* guard against foreign influences on
672 * credential calculations
673 * (notably. ptrace) */
1da177e4
LT
674};
675
4866cde0
NP
676/* Context switch must be unlocked if interrupts are to be enabled */
677#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
678# define __ARCH_WANT_UNLOCKED_CTXSW
679#endif
680
1da177e4
LT
681/*
682 * Bits in flags field of signal_struct.
683 */
684#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
ee77f075
ON
685#define SIGNAL_STOP_CONTINUED 0x00000002 /* SIGCONT since WCONTINUED reap */
686#define SIGNAL_GROUP_EXIT 0x00000004 /* group exit in progress */
e4420551
ON
687/*
688 * Pending notifications to parent.
689 */
690#define SIGNAL_CLD_STOPPED 0x00000010
691#define SIGNAL_CLD_CONTINUED 0x00000020
692#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 693
fae5fa44
ON
694#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
695
ed5d2cac
ON
696/* If true, all threads except ->group_exit_task have pending SIGKILL */
697static inline int signal_group_exit(const struct signal_struct *sig)
698{
699 return (sig->flags & SIGNAL_GROUP_EXIT) ||
700 (sig->group_exit_task != NULL);
701}
702
1da177e4
LT
703/*
704 * Some day this will be a full-fledged user tracking system..
705 */
706struct user_struct {
707 atomic_t __count; /* reference count */
708 atomic_t processes; /* How many processes does this user have? */
709 atomic_t files; /* How many open files does this user have? */
710 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 711#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
712 atomic_t inotify_watches; /* How many inotify watches does this user have? */
713 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
714#endif
4afeff85
EP
715#ifdef CONFIG_FANOTIFY
716 atomic_t fanotify_listeners;
717#endif
7ef9964e 718#ifdef CONFIG_EPOLL
52bd19f7 719 atomic_long_t epoll_watches; /* The number of file descriptors currently watched */
7ef9964e 720#endif
970a8645 721#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
722 /* protected by mq_lock */
723 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 724#endif
1da177e4
LT
725 unsigned long locked_shm; /* How many pages of mlocked shm ? */
726
727#ifdef CONFIG_KEYS
728 struct key *uid_keyring; /* UID specific keyring */
729 struct key *session_keyring; /* UID's default session keyring */
730#endif
731
732 /* Hash table maintenance information */
735de223 733 struct hlist_node uidhash_node;
7b44ab97 734 kuid_t uid;
24e377a8 735
cdd6c482 736#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
737 atomic_long_t locked_vm;
738#endif
1da177e4
LT
739};
740
eb41d946 741extern int uids_sysfs_init(void);
5cb350ba 742
7b44ab97 743extern struct user_struct *find_user(kuid_t);
1da177e4
LT
744
745extern struct user_struct root_user;
746#define INIT_USER (&root_user)
747
b6dff3ec 748
1da177e4
LT
749struct backing_dev_info;
750struct reclaim_state;
751
52f17b6c 752#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
753struct sched_info {
754 /* cumulative counters */
2d72376b 755 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 756 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
757
758 /* timestamps */
172ba844
BS
759 unsigned long long last_arrival,/* when we last ran on a cpu */
760 last_queued; /* when we were last queued to run */
1da177e4 761};
52f17b6c 762#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 763
ca74e92b
SN
764#ifdef CONFIG_TASK_DELAY_ACCT
765struct task_delay_info {
766 spinlock_t lock;
767 unsigned int flags; /* Private per-task flags */
768
769 /* For each stat XXX, add following, aligned appropriately
770 *
771 * struct timespec XXX_start, XXX_end;
772 * u64 XXX_delay;
773 * u32 XXX_count;
774 *
775 * Atomicity of updates to XXX_delay, XXX_count protected by
776 * single lock above (split into XXX_lock if contention is an issue).
777 */
0ff92245
SN
778
779 /*
780 * XXX_count is incremented on every XXX operation, the delay
781 * associated with the operation is added to XXX_delay.
782 * XXX_delay contains the accumulated delay time in nanoseconds.
783 */
784 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
785 u64 blkio_delay; /* wait for sync block io completion */
786 u64 swapin_delay; /* wait for swapin block io completion */
787 u32 blkio_count; /* total count of the number of sync block */
788 /* io operations performed */
789 u32 swapin_count; /* total count of the number of swapin block */
790 /* io operations performed */
873b4771
KK
791
792 struct timespec freepages_start, freepages_end;
793 u64 freepages_delay; /* wait for memory reclaim */
794 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 795};
52f17b6c
CS
796#endif /* CONFIG_TASK_DELAY_ACCT */
797
798static inline int sched_info_on(void)
799{
800#ifdef CONFIG_SCHEDSTATS
801 return 1;
802#elif defined(CONFIG_TASK_DELAY_ACCT)
803 extern int delayacct_on;
804 return delayacct_on;
805#else
806 return 0;
ca74e92b 807#endif
52f17b6c 808}
ca74e92b 809
d15bcfdb
IM
810enum cpu_idle_type {
811 CPU_IDLE,
812 CPU_NOT_IDLE,
813 CPU_NEWLY_IDLE,
814 CPU_MAX_IDLE_TYPES
1da177e4
LT
815};
816
817/*
c8b28116
NR
818 * Increase resolution of nice-level calculations for 64-bit architectures.
819 * The extra resolution improves shares distribution and load balancing of
820 * low-weight task groups (eg. nice +19 on an autogroup), deeper taskgroup
821 * hierarchies, especially on larger systems. This is not a user-visible change
822 * and does not change the user-interface for setting shares/weights.
823 *
824 * We increase resolution only if we have enough bits to allow this increased
825 * resolution (i.e. BITS_PER_LONG > 32). The costs for increasing resolution
826 * when BITS_PER_LONG <= 32 are pretty high and the returns do not justify the
827 * increased costs.
1da177e4 828 */
e4c2fb0d 829#if 0 /* BITS_PER_LONG > 32 -- currently broken: it increases power usage under light load */
c8b28116
NR
830# define SCHED_LOAD_RESOLUTION 10
831# define scale_load(w) ((w) << SCHED_LOAD_RESOLUTION)
832# define scale_load_down(w) ((w) >> SCHED_LOAD_RESOLUTION)
833#else
834# define SCHED_LOAD_RESOLUTION 0
835# define scale_load(w) (w)
836# define scale_load_down(w) (w)
837#endif
9aa7b369 838
c8b28116 839#define SCHED_LOAD_SHIFT (10 + SCHED_LOAD_RESOLUTION)
9aa7b369
IM
840#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
841
1399fa78
NR
842/*
843 * Increase resolution of cpu_power calculations
844 */
845#define SCHED_POWER_SHIFT 10
846#define SCHED_POWER_SCALE (1L << SCHED_POWER_SHIFT)
1da177e4 847
1399fa78
NR
848/*
849 * sched-domains (multiprocessor balancing) declarations:
850 */
2dd73a4f 851#ifdef CONFIG_SMP
b5d978e0
PZ
852#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
853#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
854#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
855#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 856#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 857#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 858#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0 859#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
b5d978e0
PZ
860#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
861#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
532cb4c4 862#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
b5d978e0 863#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
e3589f6c 864#define SD_OVERLAP 0x2000 /* sched_domains of this level overlap */
5c45bf27 865
532cb4c4
MN
866extern int __weak arch_sd_sibiling_asym_packing(void);
867
9c3f75cb 868struct sched_group_power {
e3589f6c 869 atomic_t ref;
1da177e4
LT
870 /*
871 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 872 * single CPU.
5517d86b 873 */
9c3f75cb 874 unsigned int power, power_orig;
4ec4412e 875 unsigned long next_update;
69e1e811
SS
876 /*
877 * Number of busy cpus in this group.
878 */
879 atomic_t nr_busy_cpus;
c1174876
PZ
880
881 unsigned long cpumask[0]; /* iteration mask */
9c3f75cb
PZ
882};
883
884struct sched_group {
885 struct sched_group *next; /* Must be a circular list */
886 atomic_t ref;
887
aae6d3dd 888 unsigned int group_weight;
9c3f75cb 889 struct sched_group_power *sgp;
6c99e9ad 890
4200efd9
IM
891 /*
892 * The CPUs this group covers.
893 *
894 * NOTE: this field is variable length. (Allocated dynamically
895 * by attaching extra space to the end of the structure,
896 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
897 */
898 unsigned long cpumask[0];
1da177e4
LT
899};
900
758b2cdc
RR
901static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
902{
6c99e9ad 903 return to_cpumask(sg->cpumask);
758b2cdc
RR
904}
905
c1174876
PZ
906/*
907 * cpumask masking which cpus in the group are allowed to iterate up the domain
908 * tree.
909 */
910static inline struct cpumask *sched_group_mask(struct sched_group *sg)
911{
912 return to_cpumask(sg->sgp->cpumask);
913}
914
029632fb
PZ
915/**
916 * group_first_cpu - Returns the first cpu in the cpumask of a sched_group.
917 * @group: The group whose first cpu is to be returned.
918 */
919static inline unsigned int group_first_cpu(struct sched_group *group)
920{
921 return cpumask_first(sched_group_cpus(group));
922}
923
1d3504fc
HS
924struct sched_domain_attr {
925 int relax_domain_level;
926};
927
928#define SD_ATTR_INIT (struct sched_domain_attr) { \
929 .relax_domain_level = -1, \
930}
931
60495e77
PZ
932extern int sched_domain_level_max;
933
1da177e4
LT
934struct sched_domain {
935 /* These fields must be setup */
936 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 937 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 938 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
939 unsigned long min_interval; /* Minimum balance interval ms */
940 unsigned long max_interval; /* Maximum balance interval ms */
941 unsigned int busy_factor; /* less balancing by factor if busy */
942 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 943 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
944 unsigned int busy_idx;
945 unsigned int idle_idx;
946 unsigned int newidle_idx;
947 unsigned int wake_idx;
147cbb4b 948 unsigned int forkexec_idx;
a52bfd73 949 unsigned int smt_gain;
1da177e4 950 int flags; /* See SD_* */
60495e77 951 int level;
970e1789 952 int idle_buddy; /* cpu assigned to select_idle_sibling() */
1da177e4
LT
953
954 /* Runtime fields. */
955 unsigned long last_balance; /* init to jiffies. units in jiffies */
956 unsigned int balance_interval; /* initialise to 1. units in ms. */
957 unsigned int nr_balance_failed; /* initialise to 0 */
958
2398f2c6
PZ
959 u64 last_update;
960
1da177e4
LT
961#ifdef CONFIG_SCHEDSTATS
962 /* load_balance() stats */
480b9434
KC
963 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
964 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
965 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
966 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
967 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
968 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
969 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
970 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
971
972 /* Active load balancing */
480b9434
KC
973 unsigned int alb_count;
974 unsigned int alb_failed;
975 unsigned int alb_pushed;
1da177e4 976
68767a0a 977 /* SD_BALANCE_EXEC stats */
480b9434
KC
978 unsigned int sbe_count;
979 unsigned int sbe_balanced;
980 unsigned int sbe_pushed;
1da177e4 981
68767a0a 982 /* SD_BALANCE_FORK stats */
480b9434
KC
983 unsigned int sbf_count;
984 unsigned int sbf_balanced;
985 unsigned int sbf_pushed;
68767a0a 986
1da177e4 987 /* try_to_wake_up() stats */
480b9434
KC
988 unsigned int ttwu_wake_remote;
989 unsigned int ttwu_move_affine;
990 unsigned int ttwu_move_balance;
1da177e4 991#endif
a5d8c348
IM
992#ifdef CONFIG_SCHED_DEBUG
993 char *name;
994#endif
dce840a0
PZ
995 union {
996 void *private; /* used during construction */
997 struct rcu_head rcu; /* used during destruction */
998 };
6c99e9ad 999
669c55e9 1000 unsigned int span_weight;
4200efd9
IM
1001 /*
1002 * Span of all CPUs in this domain.
1003 *
1004 * NOTE: this field is variable length. (Allocated dynamically
1005 * by attaching extra space to the end of the structure,
1006 * depending on how many CPUs the kernel has booted up with)
4200efd9
IM
1007 */
1008 unsigned long span[0];
1da177e4
LT
1009};
1010
758b2cdc
RR
1011static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1012{
6c99e9ad 1013 return to_cpumask(sd->span);
758b2cdc
RR
1014}
1015
acc3f5d7 1016extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 1017 struct sched_domain_attr *dattr_new);
029190c5 1018
acc3f5d7
RR
1019/* Allocate an array of sched domains, for partition_sched_domains(). */
1020cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1021void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1022
06aaf76a
IM
1023/* Test a flag in parent sched domain */
1024static inline int test_sd_parent(struct sched_domain *sd, int flag)
1025{
1026 if (sd->parent && (sd->parent->flags & flag))
1027 return 1;
1028
1029 return 0;
1030}
029190c5 1031
47fe38fc
PZ
1032unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1033unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1034
39be3501
PZ
1035bool cpus_share_cache(int this_cpu, int that_cpu);
1036
1b427c15 1037#else /* CONFIG_SMP */
1da177e4 1038
1b427c15 1039struct sched_domain_attr;
d02c7a8c 1040
1b427c15 1041static inline void
acc3f5d7 1042partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1043 struct sched_domain_attr *dattr_new)
1044{
d02c7a8c 1045}
39be3501
PZ
1046
1047static inline bool cpus_share_cache(int this_cpu, int that_cpu)
1048{
1049 return true;
1050}
1051
1b427c15 1052#endif /* !CONFIG_SMP */
1da177e4 1053
47fe38fc 1054
1da177e4 1055struct io_context; /* See blkdev.h */
1da177e4 1056
1da177e4 1057
383f2835 1058#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1059extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1060#else
1061static inline void prefetch_stack(struct task_struct *t) { }
1062#endif
1da177e4
LT
1063
1064struct audit_context; /* See audit.c */
1065struct mempolicy;
b92ce558 1066struct pipe_inode_info;
4865ecf1 1067struct uts_namespace;
1da177e4 1068
20b8a59f
IM
1069struct rq;
1070struct sched_domain;
1071
7d478721
PZ
1072/*
1073 * wake flags
1074 */
1075#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1076#define WF_FORK 0x02 /* child wakeup after fork */
f339b9dc 1077#define WF_MIGRATED 0x04 /* internal use, task got migrated */
7d478721 1078
371fd7e7 1079#define ENQUEUE_WAKEUP 1
74f8e4b2
PZ
1080#define ENQUEUE_HEAD 2
1081#ifdef CONFIG_SMP
1082#define ENQUEUE_WAKING 4 /* sched_class::task_waking was called */
1083#else
1084#define ENQUEUE_WAKING 0
1085#endif
371fd7e7
PZ
1086
1087#define DEQUEUE_SLEEP 1
1088
20b8a59f 1089struct sched_class {
5522d5d5 1090 const struct sched_class *next;
20b8a59f 1091
371fd7e7
PZ
1092 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int flags);
1093 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int flags);
4530d7ab 1094 void (*yield_task) (struct rq *rq);
d95f4122 1095 bool (*yield_to_task) (struct rq *rq, struct task_struct *p, bool preempt);
20b8a59f 1096
7d478721 1097 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1098
fb8d4724 1099 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1100 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1101
681f3e68 1102#ifdef CONFIG_SMP
7608dec2 1103 int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
4ce72a2c 1104
9a897c5a
SR
1105 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1106 void (*post_schedule) (struct rq *this_rq);
74f8e4b2 1107 void (*task_waking) (struct task_struct *task);
efbbd05a 1108 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1109
cd8ba7cd 1110 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1111 const struct cpumask *newmask);
57d885fe 1112
1f11eb6a
GH
1113 void (*rq_online)(struct rq *rq);
1114 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1115#endif
1116
1117 void (*set_curr_task) (struct rq *rq);
1118 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1119 void (*task_fork) (struct task_struct *p);
cb469845 1120
da7a735e
PZ
1121 void (*switched_from) (struct rq *this_rq, struct task_struct *task);
1122 void (*switched_to) (struct rq *this_rq, struct task_struct *task);
cb469845 1123 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
da7a735e 1124 int oldprio);
810b3817 1125
dba091b9
TG
1126 unsigned int (*get_rr_interval) (struct rq *rq,
1127 struct task_struct *task);
0d721cea 1128
810b3817 1129#ifdef CONFIG_FAIR_GROUP_SCHED
b2b5ce02 1130 void (*task_move_group) (struct task_struct *p, int on_rq);
810b3817 1131#endif
20b8a59f
IM
1132};
1133
1134struct load_weight {
1135 unsigned long weight, inv_weight;
1136};
1137
94c18227 1138#ifdef CONFIG_SCHEDSTATS
41acab88 1139struct sched_statistics {
20b8a59f 1140 u64 wait_start;
94c18227 1141 u64 wait_max;
6d082592
AV
1142 u64 wait_count;
1143 u64 wait_sum;
8f0dfc34
AV
1144 u64 iowait_count;
1145 u64 iowait_sum;
94c18227 1146
20b8a59f 1147 u64 sleep_start;
20b8a59f 1148 u64 sleep_max;
94c18227
IM
1149 s64 sum_sleep_runtime;
1150
1151 u64 block_start;
20b8a59f
IM
1152 u64 block_max;
1153 u64 exec_max;
eba1ed4b 1154 u64 slice_max;
cc367732 1155
cc367732
IM
1156 u64 nr_migrations_cold;
1157 u64 nr_failed_migrations_affine;
1158 u64 nr_failed_migrations_running;
1159 u64 nr_failed_migrations_hot;
1160 u64 nr_forced_migrations;
cc367732
IM
1161
1162 u64 nr_wakeups;
1163 u64 nr_wakeups_sync;
1164 u64 nr_wakeups_migrate;
1165 u64 nr_wakeups_local;
1166 u64 nr_wakeups_remote;
1167 u64 nr_wakeups_affine;
1168 u64 nr_wakeups_affine_attempts;
1169 u64 nr_wakeups_passive;
1170 u64 nr_wakeups_idle;
41acab88
LDM
1171};
1172#endif
1173
1174struct sched_entity {
1175 struct load_weight load; /* for load-balancing */
1176 struct rb_node run_node;
1177 struct list_head group_node;
1178 unsigned int on_rq;
1179
1180 u64 exec_start;
1181 u64 sum_exec_runtime;
1182 u64 vruntime;
1183 u64 prev_sum_exec_runtime;
1184
41acab88
LDM
1185 u64 nr_migrations;
1186
41acab88
LDM
1187#ifdef CONFIG_SCHEDSTATS
1188 struct sched_statistics statistics;
94c18227
IM
1189#endif
1190
20b8a59f
IM
1191#ifdef CONFIG_FAIR_GROUP_SCHED
1192 struct sched_entity *parent;
1193 /* rq on which this entity is (to be) queued: */
1194 struct cfs_rq *cfs_rq;
1195 /* rq "owned" by this entity/group: */
1196 struct cfs_rq *my_q;
1197#endif
1198};
70b97a7f 1199
fa717060
PZ
1200struct sched_rt_entity {
1201 struct list_head run_list;
78f2c7db 1202 unsigned long timeout;
bee367ed 1203 unsigned int time_slice;
6f505b16 1204
58d6c2d7 1205 struct sched_rt_entity *back;
052f1dc7 1206#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1207 struct sched_rt_entity *parent;
1208 /* rq on which this entity is (to be) queued: */
1209 struct rt_rq *rt_rq;
1210 /* rq "owned" by this entity/group: */
1211 struct rt_rq *my_q;
1212#endif
fa717060
PZ
1213};
1214
de5bdff7
HS
1215/*
1216 * default timeslice is 100 msecs (used only for SCHED_RR tasks).
1217 * Timeslices get refilled after they expire.
1218 */
1219#define RR_TIMESLICE (100 * HZ / 1000)
1220
86848966
PM
1221struct rcu_node;
1222
8dc85d54
PZ
1223enum perf_event_task_context {
1224 perf_invalid_context = -1,
1225 perf_hw_context = 0,
89a1e187 1226 perf_sw_context,
8dc85d54
PZ
1227 perf_nr_task_contexts,
1228};
1229
1da177e4
LT
1230struct task_struct {
1231 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1232 void *stack;
1da177e4 1233 atomic_t usage;
97dc32cd
WC
1234 unsigned int flags; /* per process flags, defined below */
1235 unsigned int ptrace;
1da177e4 1236
2dd73a4f 1237#ifdef CONFIG_SMP
fa14ff4a 1238 struct llist_node wake_entry;
3ca7a440 1239 int on_cpu;
2dd73a4f 1240#endif
fd2f4419 1241 int on_rq;
50e645a8 1242
b29739f9 1243 int prio, static_prio, normal_prio;
c7aceaba 1244 unsigned int rt_priority;
5522d5d5 1245 const struct sched_class *sched_class;
20b8a59f 1246 struct sched_entity se;
fa717060 1247 struct sched_rt_entity rt;
1da177e4 1248
e107be36
AK
1249#ifdef CONFIG_PREEMPT_NOTIFIERS
1250 /* list of struct preempt_notifier: */
1251 struct hlist_head preempt_notifiers;
1252#endif
1253
18796aa0
AD
1254 /*
1255 * fpu_counter contains the number of consecutive context switches
1256 * that the FPU is used. If this is over a threshold, the lazy fpu
1257 * saving becomes unlazy to save the trap. This is an unsigned char
1258 * so that after 256 times the counter wraps and the behavior turns
1259 * lazy again; this to deal with bursty apps that only use FPU for
1260 * a short time
1261 */
1262 unsigned char fpu_counter;
6c5c9341 1263#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1264 unsigned int btrace_seq;
6c5c9341 1265#endif
1da177e4 1266
97dc32cd 1267 unsigned int policy;
29baa747 1268 int nr_cpus_allowed;
1da177e4 1269 cpumask_t cpus_allowed;
1da177e4 1270
a57eb940 1271#ifdef CONFIG_PREEMPT_RCU
e260be67 1272 int rcu_read_lock_nesting;
f41d911f 1273 char rcu_read_unlock_special;
f41d911f 1274 struct list_head rcu_node_entry;
a57eb940
PM
1275#endif /* #ifdef CONFIG_PREEMPT_RCU */
1276#ifdef CONFIG_TREE_PREEMPT_RCU
1277 struct rcu_node *rcu_blocked_node;
f41d911f 1278#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
24278d14
PM
1279#ifdef CONFIG_RCU_BOOST
1280 struct rt_mutex *rcu_boost_mutex;
1281#endif /* #ifdef CONFIG_RCU_BOOST */
e260be67 1282
52f17b6c 1283#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1284 struct sched_info sched_info;
1285#endif
1286
1287 struct list_head tasks;
806c09a7 1288#ifdef CONFIG_SMP
917b627d 1289 struct plist_node pushable_tasks;
806c09a7 1290#endif
1da177e4
LT
1291
1292 struct mm_struct *mm, *active_mm;
4471a675
JK
1293#ifdef CONFIG_COMPAT_BRK
1294 unsigned brk_randomized:1;
1295#endif
34e55232
KH
1296#if defined(SPLIT_RSS_COUNTING)
1297 struct task_rss_stat rss_stat;
1298#endif
1da177e4 1299/* task state */
97dc32cd 1300 int exit_state;
1da177e4
LT
1301 int exit_code, exit_signal;
1302 int pdeath_signal; /* The signal sent when the parent dies */
a8f072c1 1303 unsigned int jobctl; /* JOBCTL_*, siglock protected */
1da177e4 1304 /* ??? */
97dc32cd 1305 unsigned int personality;
1da177e4 1306 unsigned did_exec:1;
f9ce1f1c
KT
1307 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1308 * execve */
8f0dfc34
AV
1309 unsigned in_iowait:1;
1310
259e5e6c
AL
1311 /* task may not gain privileges */
1312 unsigned no_new_privs:1;
ca94c442
LP
1313
1314 /* Revert to default priority/policy when forking */
1315 unsigned sched_reset_on_fork:1;
a8e4f2ea 1316 unsigned sched_contributes_to_load:1;
ca94c442 1317
1da177e4
LT
1318 pid_t pid;
1319 pid_t tgid;
0a425405 1320
1314562a 1321#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1322 /* Canary value for the -fstack-protector gcc feature */
1323 unsigned long stack_canary;
1314562a 1324#endif
4d1d61a6 1325 /*
1da177e4 1326 * pointers to (original) parent process, youngest child, younger sibling,
4d1d61a6 1327 * older sibling, respectively. (p->father can be replaced with
f470021a 1328 * p->real_parent->pid)
1da177e4 1329 */
abd63bc3
KC
1330 struct task_struct __rcu *real_parent; /* real parent process */
1331 struct task_struct __rcu *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1332 /*
f470021a 1333 * children/sibling forms the list of my natural children
1da177e4
LT
1334 */
1335 struct list_head children; /* list of my children */
1336 struct list_head sibling; /* linkage in my parent's children list */
1337 struct task_struct *group_leader; /* threadgroup leader */
1338
f470021a
RM
1339 /*
1340 * ptraced is the list of tasks this task is using ptrace on.
1341 * This includes both natural children and PTRACE_ATTACH targets.
1342 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1343 */
1344 struct list_head ptraced;
1345 struct list_head ptrace_entry;
1346
1da177e4 1347 /* PID/PID hash table linkage. */
92476d7f 1348 struct pid_link pids[PIDTYPE_MAX];
47e65328 1349 struct list_head thread_group;
1da177e4
LT
1350
1351 struct completion *vfork_done; /* for vfork() */
1352 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1353 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1354
c66f08be 1355 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1356 cputime_t gtime;
d99ca3b9 1357#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1358 cputime_t prev_utime, prev_stime;
d99ca3b9 1359#endif
1da177e4 1360 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1361 struct timespec start_time; /* monotonic time */
1362 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1363/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1364 unsigned long min_flt, maj_flt;
1365
f06febc9 1366 struct task_cputime cputime_expires;
1da177e4
LT
1367 struct list_head cpu_timers[3];
1368
1369/* process credentials */
1b0ba1c9 1370 const struct cred __rcu *real_cred; /* objective and real subjective task
3b11a1de 1371 * credentials (COW) */
1b0ba1c9 1372 const struct cred __rcu *cred; /* effective (overridable) subjective task
3b11a1de 1373 * credentials (COW) */
36772092
PBG
1374 char comm[TASK_COMM_LEN]; /* executable name excluding path
1375 - access with [gs]et_task_comm (which lock
1376 it with task_lock())
221af7f8 1377 - initialized normally by setup_new_exec */
1da177e4
LT
1378/* file system info */
1379 int link_count, total_link_count;
3d5b6fcc 1380#ifdef CONFIG_SYSVIPC
1da177e4
LT
1381/* ipc stuff */
1382 struct sysv_sem sysvsem;
3d5b6fcc 1383#endif
e162b39a 1384#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1385/* hung task detection */
82a1fcb9
IM
1386 unsigned long last_switch_count;
1387#endif
1da177e4
LT
1388/* CPU-specific state of this task */
1389 struct thread_struct thread;
1390/* filesystem information */
1391 struct fs_struct *fs;
1392/* open file information */
1393 struct files_struct *files;
1651e14e 1394/* namespaces */
ab516013 1395 struct nsproxy *nsproxy;
1da177e4
LT
1396/* signal handlers */
1397 struct signal_struct *signal;
1398 struct sighand_struct *sighand;
1399
1400 sigset_t blocked, real_blocked;
f3de272b 1401 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1402 struct sigpending pending;
1403
1404 unsigned long sas_ss_sp;
1405 size_t sas_ss_size;
1406 int (*notifier)(void *priv);
1407 void *notifier_data;
1408 sigset_t *notifier_mask;
e73f8959
ON
1409 struct hlist_head task_works;
1410
1da177e4 1411 struct audit_context *audit_context;
bfef93a5
AV
1412#ifdef CONFIG_AUDITSYSCALL
1413 uid_t loginuid;
4746ec5b 1414 unsigned int sessionid;
bfef93a5 1415#endif
932ecebb 1416 struct seccomp seccomp;
1da177e4
LT
1417
1418/* Thread group tracking */
1419 u32 parent_exec_id;
1420 u32 self_exec_id;
58568d2a
MX
1421/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1422 * mempolicy */
1da177e4 1423 spinlock_t alloc_lock;
1da177e4 1424
b29739f9 1425 /* Protection of the PI data structures: */
1d615482 1426 raw_spinlock_t pi_lock;
b29739f9 1427
23f78d4a
IM
1428#ifdef CONFIG_RT_MUTEXES
1429 /* PI waiters blocked on a rt_mutex held by this task */
1430 struct plist_head pi_waiters;
1431 /* Deadlock detection and priority inheritance handling */
1432 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1433#endif
1434
408894ee
IM
1435#ifdef CONFIG_DEBUG_MUTEXES
1436 /* mutex deadlock detection */
1437 struct mutex_waiter *blocked_on;
1438#endif
de30a2b3
IM
1439#ifdef CONFIG_TRACE_IRQFLAGS
1440 unsigned int irq_events;
de30a2b3 1441 unsigned long hardirq_enable_ip;
de30a2b3 1442 unsigned long hardirq_disable_ip;
fa1452e8 1443 unsigned int hardirq_enable_event;
de30a2b3 1444 unsigned int hardirq_disable_event;
fa1452e8
HS
1445 int hardirqs_enabled;
1446 int hardirq_context;
de30a2b3 1447 unsigned long softirq_disable_ip;
de30a2b3 1448 unsigned long softirq_enable_ip;
fa1452e8 1449 unsigned int softirq_disable_event;
de30a2b3 1450 unsigned int softirq_enable_event;
fa1452e8 1451 int softirqs_enabled;
de30a2b3
IM
1452 int softirq_context;
1453#endif
fbb9ce95 1454#ifdef CONFIG_LOCKDEP
bdb9441e 1455# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1456 u64 curr_chain_key;
1457 int lockdep_depth;
fbb9ce95 1458 unsigned int lockdep_recursion;
c7aceaba 1459 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1460 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1461#endif
408894ee 1462
1da177e4
LT
1463/* journalling filesystem info */
1464 void *journal_info;
1465
d89d8796 1466/* stacked block device info */
bddd87c7 1467 struct bio_list *bio_list;
d89d8796 1468
73c10101
JA
1469#ifdef CONFIG_BLOCK
1470/* stack plugging */
1471 struct blk_plug *plug;
1472#endif
1473
1da177e4
LT
1474/* VM state */
1475 struct reclaim_state *reclaim_state;
1476
1da177e4
LT
1477 struct backing_dev_info *backing_dev_info;
1478
1479 struct io_context *io_context;
1480
1481 unsigned long ptrace_message;
1482 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1483 struct task_io_accounting ioac;
8f0ab514 1484#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1485 u64 acct_rss_mem1; /* accumulated rss usage */
1486 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1487 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1488#endif
1489#ifdef CONFIG_CPUSETS
58568d2a 1490 nodemask_t mems_allowed; /* Protected by alloc_lock */
cc9a6c87 1491 seqcount_t mems_allowed_seq; /* Seqence no to catch updates */
825a46af 1492 int cpuset_mem_spread_rotor;
6adef3eb 1493 int cpuset_slab_spread_rotor;
1da177e4 1494#endif
ddbcc7e8 1495#ifdef CONFIG_CGROUPS
817929ec 1496 /* Control Group info protected by css_set_lock */
2c392b8c 1497 struct css_set __rcu *cgroups;
817929ec
PM
1498 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1499 struct list_head cg_list;
ddbcc7e8 1500#endif
42b2dd0a 1501#ifdef CONFIG_FUTEX
0771dfef 1502 struct robust_list_head __user *robust_list;
34f192c6
IM
1503#ifdef CONFIG_COMPAT
1504 struct compat_robust_list_head __user *compat_robust_list;
1505#endif
c87e2837
IM
1506 struct list_head pi_state_list;
1507 struct futex_pi_state *pi_state_cache;
c7aceaba 1508#endif
cdd6c482 1509#ifdef CONFIG_PERF_EVENTS
8dc85d54 1510 struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts];
cdd6c482
IM
1511 struct mutex perf_event_mutex;
1512 struct list_head perf_event_list;
a63eaf34 1513#endif
c7aceaba 1514#ifdef CONFIG_NUMA
58568d2a 1515 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1516 short il_next;
207205a2 1517 short pref_node_fork;
42b2dd0a 1518#endif
e56d0903 1519 struct rcu_head rcu;
b92ce558
JA
1520
1521 /*
1522 * cache last used pipe for splice
1523 */
1524 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1525#ifdef CONFIG_TASK_DELAY_ACCT
1526 struct task_delay_info *delays;
f4f154fd
AM
1527#endif
1528#ifdef CONFIG_FAULT_INJECTION
1529 int make_it_fail;
ca74e92b 1530#endif
9d823e8f
WF
1531 /*
1532 * when (nr_dirtied >= nr_dirtied_pause), it's time to call
1533 * balance_dirty_pages() for some dirty throttling pause
1534 */
1535 int nr_dirtied;
1536 int nr_dirtied_pause;
83712358 1537 unsigned long dirty_paused_when; /* start of a write-and-pause period */
9d823e8f 1538
9745512c
AV
1539#ifdef CONFIG_LATENCYTOP
1540 int latency_record_count;
1541 struct latency_record latency_record[LT_SAVECOUNT];
1542#endif
6976675d
AV
1543 /*
1544 * time slack values; these are used to round up poll() and
1545 * select() etc timeout values. These are in nanoseconds.
1546 */
1547 unsigned long timer_slack_ns;
1548 unsigned long default_timer_slack_ns;
f8d570a4
DM
1549
1550 struct list_head *scm_work_list;
fb52607a 1551#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3ad2f3fb 1552 /* Index of current stored address in ret_stack */
f201ae23
FW
1553 int curr_ret_stack;
1554 /* Stack of return addresses for return function tracing */
1555 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1556 /* time stamp for last schedule */
1557 unsigned long long ftrace_timestamp;
f201ae23
FW
1558 /*
1559 * Number of functions that haven't been traced
1560 * because of depth overrun.
1561 */
1562 atomic_t trace_overrun;
380c4b14
FW
1563 /* Pause for the tracing */
1564 atomic_t tracing_graph_pause;
f201ae23 1565#endif
ea4e2bc4
SR
1566#ifdef CONFIG_TRACING
1567 /* state flags for use by tracers */
1568 unsigned long trace;
b1cff0ad 1569 /* bitmask and counter of trace recursion */
261842b7
SR
1570 unsigned long trace_recursion;
1571#endif /* CONFIG_TRACING */
569b846d
KH
1572#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1573 struct memcg_batch_info {
1574 int do_batch; /* incremented when batch uncharge started */
1575 struct mem_cgroup *memcg; /* target memcg of uncharge */
7ffd4ca7
JW
1576 unsigned long nr_pages; /* uncharged usage */
1577 unsigned long memsw_nr_pages; /* uncharged mem+swap usage */
569b846d
KH
1578 } memcg_batch;
1579#endif
bf26c018
FW
1580#ifdef CONFIG_HAVE_HW_BREAKPOINT
1581 atomic_t ptrace_bp_refcnt;
1582#endif
0326f5a9
SD
1583#ifdef CONFIG_UPROBES
1584 struct uprobe_task *utask;
1585 int uprobe_srcu_id;
1586#endif
1da177e4
LT
1587};
1588
76e6eee0 1589/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1590#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1591
e05606d3
IM
1592/*
1593 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1594 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1595 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1596 * values are inverted: lower p->prio value means higher priority.
1597 *
1598 * The MAX_USER_RT_PRIO value allows the actual maximum
1599 * RT priority to be separate from the value exported to
1600 * user-space. This allows kernel threads to set their
1601 * priority to a value higher than any user task. Note:
1602 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1603 */
1604
1605#define MAX_USER_RT_PRIO 100
1606#define MAX_RT_PRIO MAX_USER_RT_PRIO
1607
1608#define MAX_PRIO (MAX_RT_PRIO + 40)
1609#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1610
1611static inline int rt_prio(int prio)
1612{
1613 if (unlikely(prio < MAX_RT_PRIO))
1614 return 1;
1615 return 0;
1616}
1617
e868171a 1618static inline int rt_task(struct task_struct *p)
e05606d3
IM
1619{
1620 return rt_prio(p->prio);
1621}
1622
e868171a 1623static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1624{
1625 return task->pids[PIDTYPE_PID].pid;
1626}
1627
e868171a 1628static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1629{
1630 return task->group_leader->pids[PIDTYPE_PID].pid;
1631}
1632
6dda81f4
ON
1633/*
1634 * Without tasklist or rcu lock it is not safe to dereference
1635 * the result of task_pgrp/task_session even if task == current,
1636 * we can race with another thread doing sys_setsid/sys_setpgid.
1637 */
e868171a 1638static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1639{
1640 return task->group_leader->pids[PIDTYPE_PGID].pid;
1641}
1642
e868171a 1643static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1644{
1645 return task->group_leader->pids[PIDTYPE_SID].pid;
1646}
1647
7af57294
PE
1648struct pid_namespace;
1649
1650/*
1651 * the helpers to get the task's different pids as they are seen
1652 * from various namespaces
1653 *
1654 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1655 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1656 * current.
7af57294
PE
1657 * task_xid_nr_ns() : id seen from the ns specified;
1658 *
1659 * set_task_vxid() : assigns a virtual id to a task;
1660 *
7af57294
PE
1661 * see also pid_nr() etc in include/linux/pid.h
1662 */
52ee2dfd
ON
1663pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1664 struct pid_namespace *ns);
7af57294 1665
e868171a 1666static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1667{
1668 return tsk->pid;
1669}
1670
52ee2dfd
ON
1671static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1672 struct pid_namespace *ns)
1673{
1674 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1675}
7af57294
PE
1676
1677static inline pid_t task_pid_vnr(struct task_struct *tsk)
1678{
52ee2dfd 1679 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1680}
1681
1682
e868171a 1683static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1684{
1685 return tsk->tgid;
1686}
1687
2f2a3a46 1688pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1689
1690static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1691{
1692 return pid_vnr(task_tgid(tsk));
1693}
1694
1695
52ee2dfd
ON
1696static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1697 struct pid_namespace *ns)
7af57294 1698{
52ee2dfd 1699 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1700}
1701
7af57294
PE
1702static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1703{
52ee2dfd 1704 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1705}
1706
1707
52ee2dfd
ON
1708static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1709 struct pid_namespace *ns)
7af57294 1710{
52ee2dfd 1711 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1712}
1713
7af57294
PE
1714static inline pid_t task_session_vnr(struct task_struct *tsk)
1715{
52ee2dfd 1716 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1717}
1718
1b0f7ffd
ON
1719/* obsolete, do not use */
1720static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1721{
1722 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1723}
7af57294 1724
1da177e4
LT
1725/**
1726 * pid_alive - check that a task structure is not stale
1727 * @p: Task structure to be checked.
1728 *
1729 * Test if a process is not yet dead (at most zombie state)
1730 * If pid_alive fails, then pointers within the task structure
1731 * can be stale and must not be dereferenced.
1732 */
e868171a 1733static inline int pid_alive(struct task_struct *p)
1da177e4 1734{
92476d7f 1735 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1736}
1737
f400e198 1738/**
b460cbc5 1739 * is_global_init - check if a task structure is init
3260259f
HK
1740 * @tsk: Task structure to be checked.
1741 *
1742 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1743 */
e868171a 1744static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1745{
1746 return tsk->pid == 1;
1747}
b460cbc5
SH
1748
1749/*
1750 * is_container_init:
1751 * check whether in the task is init in its own pid namespace.
f400e198 1752 */
b461cc03 1753extern int is_container_init(struct task_struct *tsk);
f400e198 1754
9ec52099
CLG
1755extern struct pid *cad_pid;
1756
1da177e4 1757extern void free_task(struct task_struct *tsk);
1da177e4 1758#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1759
158d9ebd 1760extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1761
1762static inline void put_task_struct(struct task_struct *t)
1763{
1764 if (atomic_dec_and_test(&t->usage))
8c7904a0 1765 __put_task_struct(t);
e56d0903 1766}
1da177e4 1767
d180c5bc 1768extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1769extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1770
1da177e4
LT
1771/*
1772 * Per process flags
1773 */
1da177e4 1774#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1775#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1776#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
21aa9af0 1777#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
1da177e4 1778#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1779#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1780#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1781#define PF_DUMPCORE 0x00000200 /* dumped core */
1782#define PF_SIGNALED 0x00000400 /* killed by a signal */
1783#define PF_MEMALLOC 0x00000800 /* Allocating memory */
72fa5997 1784#define PF_NPROC_EXCEEDED 0x00001000 /* set_user noticed that RLIMIT_NPROC was exceeded */
1da177e4 1785#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1da177e4
LT
1786#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1787#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1788#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1789#define PF_KSWAPD 0x00040000 /* I am kswapd */
1da177e4 1790#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1791#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1792#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1793#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1794#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1795#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1796#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1797#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1798#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1799#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
58a69cb4 1800#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
1da177e4
LT
1801
1802/*
1803 * Only the _current_ task can read/write to tsk->flags, but other
1804 * tasks can access tsk->flags in readonly mode for example
1805 * with tsk_used_math (like during threaded core dumping).
1806 * There is however an exception to this rule during ptrace
1807 * or during fork: the ptracer task is allowed to write to the
1808 * child->flags of its traced child (same goes for fork, the parent
1809 * can write to the child->flags), because we're guaranteed the
1810 * child is not running and in turn not changing child->flags
1811 * at the same time the parent does it.
1812 */
1813#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1814#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1815#define clear_used_math() clear_stopped_child_used_math(current)
1816#define set_used_math() set_stopped_child_used_math(current)
1817#define conditional_stopped_child_used_math(condition, child) \
1818 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1819#define conditional_used_math(condition) \
1820 conditional_stopped_child_used_math(condition, current)
1821#define copy_to_stopped_child_used_math(child) \
1822 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1823/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1824#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1825#define used_math() tsk_used_math(current)
1826
e5c1902e 1827/*
a8f072c1 1828 * task->jobctl flags
e5c1902e 1829 */
a8f072c1 1830#define JOBCTL_STOP_SIGMASK 0xffff /* signr of the last group stop */
e5c1902e 1831
a8f072c1
TH
1832#define JOBCTL_STOP_DEQUEUED_BIT 16 /* stop signal dequeued */
1833#define JOBCTL_STOP_PENDING_BIT 17 /* task should stop for group stop */
1834#define JOBCTL_STOP_CONSUME_BIT 18 /* consume group stop count */
73ddff2b 1835#define JOBCTL_TRAP_STOP_BIT 19 /* trap for STOP */
fb1d910c 1836#define JOBCTL_TRAP_NOTIFY_BIT 20 /* trap for NOTIFY */
a8f072c1 1837#define JOBCTL_TRAPPING_BIT 21 /* switching to TRACED */
544b2c91 1838#define JOBCTL_LISTENING_BIT 22 /* ptracer is listening for events */
a8f072c1
TH
1839
1840#define JOBCTL_STOP_DEQUEUED (1 << JOBCTL_STOP_DEQUEUED_BIT)
1841#define JOBCTL_STOP_PENDING (1 << JOBCTL_STOP_PENDING_BIT)
1842#define JOBCTL_STOP_CONSUME (1 << JOBCTL_STOP_CONSUME_BIT)
73ddff2b 1843#define JOBCTL_TRAP_STOP (1 << JOBCTL_TRAP_STOP_BIT)
fb1d910c 1844#define JOBCTL_TRAP_NOTIFY (1 << JOBCTL_TRAP_NOTIFY_BIT)
a8f072c1 1845#define JOBCTL_TRAPPING (1 << JOBCTL_TRAPPING_BIT)
544b2c91 1846#define JOBCTL_LISTENING (1 << JOBCTL_LISTENING_BIT)
a8f072c1 1847
fb1d910c 1848#define JOBCTL_TRAP_MASK (JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
73ddff2b 1849#define JOBCTL_PENDING_MASK (JOBCTL_STOP_PENDING | JOBCTL_TRAP_MASK)
3759a0d9 1850
7dd3db54
TH
1851extern bool task_set_jobctl_pending(struct task_struct *task,
1852 unsigned int mask);
73ddff2b 1853extern void task_clear_jobctl_trapping(struct task_struct *task);
3759a0d9
TH
1854extern void task_clear_jobctl_pending(struct task_struct *task,
1855 unsigned int mask);
39efa3ef 1856
a57eb940 1857#ifdef CONFIG_PREEMPT_RCU
f41d911f
PM
1858
1859#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1aa03f11 1860#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1861
1862static inline void rcu_copy_process(struct task_struct *p)
1863{
1864 p->rcu_read_lock_nesting = 0;
1865 p->rcu_read_unlock_special = 0;
a57eb940 1866#ifdef CONFIG_TREE_PREEMPT_RCU
dd5d19ba 1867 p->rcu_blocked_node = NULL;
24278d14
PM
1868#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1869#ifdef CONFIG_RCU_BOOST
1870 p->rcu_boost_mutex = NULL;
1871#endif /* #ifdef CONFIG_RCU_BOOST */
f41d911f
PM
1872 INIT_LIST_HEAD(&p->rcu_node_entry);
1873}
1874
f41d911f
PM
1875#else
1876
1877static inline void rcu_copy_process(struct task_struct *p)
1878{
1879}
1880
1881#endif
1882
1da177e4 1883#ifdef CONFIG_SMP
1e1b6c51
KM
1884extern void do_set_cpus_allowed(struct task_struct *p,
1885 const struct cpumask *new_mask);
1886
cd8ba7cd 1887extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1888 const struct cpumask *new_mask);
1da177e4 1889#else
1e1b6c51
KM
1890static inline void do_set_cpus_allowed(struct task_struct *p,
1891 const struct cpumask *new_mask)
1892{
1893}
cd8ba7cd 1894static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1895 const struct cpumask *new_mask)
1da177e4 1896{
96f874e2 1897 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1898 return -EINVAL;
1899 return 0;
1900}
1901#endif
e0ad9556 1902
5167e8d5
PZ
1903#ifdef CONFIG_NO_HZ
1904void calc_load_enter_idle(void);
1905void calc_load_exit_idle(void);
1906#else
1907static inline void calc_load_enter_idle(void) { }
1908static inline void calc_load_exit_idle(void) { }
1909#endif /* CONFIG_NO_HZ */
1910
e0ad9556 1911#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1912static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1913{
1914 return set_cpus_allowed_ptr(p, &new_mask);
1915}
e0ad9556 1916#endif
1da177e4 1917
b342501c 1918/*
c676329a
PZ
1919 * Do not use outside of architecture code which knows its limitations.
1920 *
1921 * sched_clock() has no promise of monotonicity or bounded drift between
1922 * CPUs, use (which you should not) requires disabling IRQs.
1923 *
1924 * Please use one of the three interfaces below.
b342501c 1925 */
1bbfa6f2 1926extern unsigned long long notrace sched_clock(void);
c676329a 1927/*
489a71b0 1928 * See the comment in kernel/sched/clock.c
c676329a
PZ
1929 */
1930extern u64 cpu_clock(int cpu);
1931extern u64 local_clock(void);
1932extern u64 sched_clock_cpu(int cpu);
1933
e436d800 1934
c1955a3d 1935extern void sched_clock_init(void);
3e51f33f 1936
c1955a3d 1937#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1938static inline void sched_clock_tick(void)
1939{
1940}
1941
1942static inline void sched_clock_idle_sleep_event(void)
1943{
1944}
1945
1946static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1947{
1948}
1949#else
c676329a
PZ
1950/*
1951 * Architectures can set this to 1 if they have specified
1952 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1953 * but then during bootup it turns out that sched_clock()
1954 * is reliable after all:
1955 */
1956extern int sched_clock_stable;
1957
3e51f33f
PZ
1958extern void sched_clock_tick(void);
1959extern void sched_clock_idle_sleep_event(void);
1960extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1961#endif
1962
b52bfee4
VP
1963#ifdef CONFIG_IRQ_TIME_ACCOUNTING
1964/*
1965 * An i/f to runtime opt-in for irq time accounting based off of sched_clock.
1966 * The reason for this explicit opt-in is not to have perf penalty with
1967 * slow sched_clocks.
1968 */
1969extern void enable_sched_clock_irqtime(void);
1970extern void disable_sched_clock_irqtime(void);
1971#else
1972static inline void enable_sched_clock_irqtime(void) {}
1973static inline void disable_sched_clock_irqtime(void) {}
1974#endif
1975
36c8b586 1976extern unsigned long long
41b86e9c 1977task_sched_runtime(struct task_struct *task);
1da177e4
LT
1978
1979/* sched_exec is called by processes performing an exec */
1980#ifdef CONFIG_SMP
1981extern void sched_exec(void);
1982#else
1983#define sched_exec() {}
1984#endif
1985
2aa44d05
IM
1986extern void sched_clock_idle_sleep_event(void);
1987extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1988
1da177e4
LT
1989#ifdef CONFIG_HOTPLUG_CPU
1990extern void idle_task_exit(void);
1991#else
1992static inline void idle_task_exit(void) {}
1993#endif
1994
06d8308c
TG
1995#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1996extern void wake_up_idle_cpu(int cpu);
1997#else
1998static inline void wake_up_idle_cpu(int cpu) { }
1999#endif
2000
21805085 2001extern unsigned int sysctl_sched_latency;
b2be5e96 2002extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 2003extern unsigned int sysctl_sched_wakeup_granularity;
bf0f6f24 2004extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
2005
2006enum sched_tunable_scaling {
2007 SCHED_TUNABLESCALING_NONE,
2008 SCHED_TUNABLESCALING_LOG,
2009 SCHED_TUNABLESCALING_LINEAR,
2010 SCHED_TUNABLESCALING_END,
2011};
2012extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
2013
2bba22c5 2014#ifdef CONFIG_SCHED_DEBUG
da84d961 2015extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 2016extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 2017extern unsigned int sysctl_sched_time_avg;
cd1bb94b 2018extern unsigned int sysctl_timer_migration;
a7a4f8a7 2019extern unsigned int sysctl_sched_shares_window;
b2be5e96 2020
1983a922 2021int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 2022 void __user *buffer, size_t *length,
b2be5e96 2023 loff_t *ppos);
2bd8e6d4 2024#endif
eea08f32
AB
2025#ifdef CONFIG_SCHED_DEBUG
2026static inline unsigned int get_sysctl_timer_migration(void)
2027{
2028 return sysctl_timer_migration;
2029}
2030#else
2031static inline unsigned int get_sysctl_timer_migration(void)
2032{
2033 return 1;
2034}
2035#endif
9f0c1e56
PZ
2036extern unsigned int sysctl_sched_rt_period;
2037extern int sysctl_sched_rt_runtime;
2bd8e6d4 2038
d0b27fa7 2039int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 2040 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
2041 loff_t *ppos);
2042
5091faa4
MG
2043#ifdef CONFIG_SCHED_AUTOGROUP
2044extern unsigned int sysctl_sched_autogroup_enabled;
2045
2046extern void sched_autogroup_create_attach(struct task_struct *p);
2047extern void sched_autogroup_detach(struct task_struct *p);
2048extern void sched_autogroup_fork(struct signal_struct *sig);
2049extern void sched_autogroup_exit(struct signal_struct *sig);
2050#ifdef CONFIG_PROC_FS
2051extern void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m);
2e5b5b3a 2052extern int proc_sched_autogroup_set_nice(struct task_struct *p, int nice);
5091faa4
MG
2053#endif
2054#else
2055static inline void sched_autogroup_create_attach(struct task_struct *p) { }
2056static inline void sched_autogroup_detach(struct task_struct *p) { }
2057static inline void sched_autogroup_fork(struct signal_struct *sig) { }
2058static inline void sched_autogroup_exit(struct signal_struct *sig) { }
2059#endif
2060
ec12cb7f
PT
2061#ifdef CONFIG_CFS_BANDWIDTH
2062extern unsigned int sysctl_sched_cfs_bandwidth_slice;
2063#endif
2064
b29739f9 2065#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
2066extern int rt_mutex_getprio(struct task_struct *p);
2067extern void rt_mutex_setprio(struct task_struct *p, int prio);
2068extern void rt_mutex_adjust_pi(struct task_struct *p);
3c7d5184
TG
2069static inline bool tsk_is_pi_blocked(struct task_struct *tsk)
2070{
2071 return tsk->pi_blocked_on != NULL;
2072}
b29739f9 2073#else
e868171a 2074static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
2075{
2076 return p->normal_prio;
2077}
95e02ca9 2078# define rt_mutex_adjust_pi(p) do { } while (0)
3c7d5184
TG
2079static inline bool tsk_is_pi_blocked(struct task_struct *tsk)
2080{
2081 return false;
2082}
b29739f9
IM
2083#endif
2084
d95f4122 2085extern bool yield_to(struct task_struct *p, bool preempt);
36c8b586
IM
2086extern void set_user_nice(struct task_struct *p, long nice);
2087extern int task_prio(const struct task_struct *p);
2088extern int task_nice(const struct task_struct *p);
2089extern int can_nice(const struct task_struct *p, const int nice);
2090extern int task_curr(const struct task_struct *p);
1da177e4 2091extern int idle_cpu(int cpu);
fe7de49f
KM
2092extern int sched_setscheduler(struct task_struct *, int,
2093 const struct sched_param *);
961ccddd 2094extern int sched_setscheduler_nocheck(struct task_struct *, int,
fe7de49f 2095 const struct sched_param *);
36c8b586 2096extern struct task_struct *idle_task(int cpu);
c4f30608
PM
2097/**
2098 * is_idle_task - is the specified task an idle task?
fa757281 2099 * @p: the task in question.
c4f30608 2100 */
7061ca3b 2101static inline bool is_idle_task(const struct task_struct *p)
c4f30608
PM
2102{
2103 return p->pid == 0;
2104}
36c8b586
IM
2105extern struct task_struct *curr_task(int cpu);
2106extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
2107
2108void yield(void);
2109
2110/*
2111 * The default (Linux) execution domain.
2112 */
2113extern struct exec_domain default_exec_domain;
2114
2115union thread_union {
2116 struct thread_info thread_info;
2117 unsigned long stack[THREAD_SIZE/sizeof(long)];
2118};
2119
2120#ifndef __HAVE_ARCH_KSTACK_END
2121static inline int kstack_end(void *addr)
2122{
2123 /* Reliable end of stack detection:
2124 * Some APM bios versions misalign the stack
2125 */
2126 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2127}
2128#endif
2129
2130extern union thread_union init_thread_union;
2131extern struct task_struct init_task;
2132
2133extern struct mm_struct init_mm;
2134
198fe21b
PE
2135extern struct pid_namespace init_pid_ns;
2136
2137/*
2138 * find a task by one of its numerical ids
2139 *
198fe21b
PE
2140 * find_task_by_pid_ns():
2141 * finds a task by its pid in the specified namespace
228ebcbe
PE
2142 * find_task_by_vpid():
2143 * finds a task by its virtual pid
198fe21b 2144 *
e49859e7 2145 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2146 */
2147
228ebcbe
PE
2148extern struct task_struct *find_task_by_vpid(pid_t nr);
2149extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2150 struct pid_namespace *ns);
198fe21b 2151
8520d7c7 2152extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2153
2154/* per-UID process charging. */
7b44ab97 2155extern struct user_struct * alloc_uid(kuid_t);
1da177e4
LT
2156static inline struct user_struct *get_uid(struct user_struct *u)
2157{
2158 atomic_inc(&u->__count);
2159 return u;
2160}
2161extern void free_uid(struct user_struct *);
1da177e4
LT
2162
2163#include <asm/current.h>
2164
f0af911a 2165extern void xtime_update(unsigned long ticks);
1da177e4 2166
b3c97528
HH
2167extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2168extern int wake_up_process(struct task_struct *tsk);
3e51e3ed 2169extern void wake_up_new_task(struct task_struct *tsk);
1da177e4
LT
2170#ifdef CONFIG_SMP
2171 extern void kick_process(struct task_struct *tsk);
2172#else
2173 static inline void kick_process(struct task_struct *tsk) { }
2174#endif
3e51e3ed 2175extern void sched_fork(struct task_struct *p);
ad46c2c4 2176extern void sched_dead(struct task_struct *p);
1da177e4 2177
1da177e4
LT
2178extern void proc_caches_init(void);
2179extern void flush_signals(struct task_struct *);
3bcac026 2180extern void __flush_signals(struct task_struct *);
10ab825b 2181extern void ignore_signals(struct task_struct *);
1da177e4
LT
2182extern void flush_signal_handlers(struct task_struct *, int force_default);
2183extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2184
2185static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2186{
2187 unsigned long flags;
2188 int ret;
2189
2190 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2191 ret = dequeue_signal(tsk, mask, info);
2192 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2193
2194 return ret;
53c8f9f1 2195}
1da177e4
LT
2196
2197extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2198 sigset_t *mask);
2199extern void unblock_all_signals(void);
2200extern void release_task(struct task_struct * p);
2201extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2202extern int force_sigsegv(int, struct task_struct *);
2203extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2204extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2205extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
d178bc3a
SH
2206extern int kill_pid_info_as_cred(int, struct siginfo *, struct pid *,
2207 const struct cred *, u32);
c4b92fc1
EB
2208extern int kill_pgrp(struct pid *pid, int sig, int priv);
2209extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2210extern int kill_proc_info(int, struct siginfo *, pid_t);
86773473 2211extern __must_check bool do_notify_parent(struct task_struct *, int);
a7f0765e 2212extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2213extern void force_sig(int, struct task_struct *);
1da177e4 2214extern int send_sig(int, struct task_struct *, int);
09faef11 2215extern int zap_other_threads(struct task_struct *p);
1da177e4
LT
2216extern struct sigqueue *sigqueue_alloc(void);
2217extern void sigqueue_free(struct sigqueue *);
ac5c2153 2218extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2219extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2220extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2221
51a7b448
AV
2222static inline void restore_saved_sigmask(void)
2223{
2224 if (test_and_clear_restore_sigmask())
77097ae5 2225 __set_current_blocked(&current->saved_sigmask);
51a7b448
AV
2226}
2227
b7f9a11a
AV
2228static inline sigset_t *sigmask_to_save(void)
2229{
2230 sigset_t *res = &current->blocked;
2231 if (unlikely(test_restore_sigmask()))
2232 res = &current->saved_sigmask;
2233 return res;
2234}
2235
9ec52099
CLG
2236static inline int kill_cad_pid(int sig, int priv)
2237{
2238 return kill_pid(cad_pid, sig, priv);
2239}
2240
1da177e4
LT
2241/* These can be the second arg to send_sig_info/send_group_sig_info. */
2242#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2243#define SEND_SIG_PRIV ((struct siginfo *) 1)
2244#define SEND_SIG_FORCED ((struct siginfo *) 2)
2245
2a855dd0
SAS
2246/*
2247 * True if we are on the alternate signal stack.
2248 */
1da177e4
LT
2249static inline int on_sig_stack(unsigned long sp)
2250{
2a855dd0
SAS
2251#ifdef CONFIG_STACK_GROWSUP
2252 return sp >= current->sas_ss_sp &&
2253 sp - current->sas_ss_sp < current->sas_ss_size;
2254#else
2255 return sp > current->sas_ss_sp &&
2256 sp - current->sas_ss_sp <= current->sas_ss_size;
2257#endif
1da177e4
LT
2258}
2259
2260static inline int sas_ss_flags(unsigned long sp)
2261{
2262 return (current->sas_ss_size == 0 ? SS_DISABLE
2263 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2264}
2265
1da177e4
LT
2266/*
2267 * Routines for handling mm_structs
2268 */
2269extern struct mm_struct * mm_alloc(void);
2270
2271/* mmdrop drops the mm and the page tables */
b3c97528 2272extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2273static inline void mmdrop(struct mm_struct * mm)
2274{
6fb43d7b 2275 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2276 __mmdrop(mm);
2277}
2278
2279/* mmput gets rid of the mappings and all user-space */
2280extern void mmput(struct mm_struct *);
2281/* Grab a reference to a task's mm, if it is not already going away */
2282extern struct mm_struct *get_task_mm(struct task_struct *task);
8cdb878d
CY
2283/*
2284 * Grab a reference to a task's mm, if it is not already going away
2285 * and ptrace_may_access with the mode parameter passed to it
2286 * succeeds.
2287 */
2288extern struct mm_struct *mm_access(struct task_struct *task, unsigned int mode);
1da177e4
LT
2289/* Remove the current tasks stale references to the old mm_struct */
2290extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2291/* Allocate a new mm structure and copy contents from tsk->mm */
2292extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2293
6f2c55b8
AD
2294extern int copy_thread(unsigned long, unsigned long, unsigned long,
2295 struct task_struct *, struct pt_regs *);
1da177e4
LT
2296extern void flush_thread(void);
2297extern void exit_thread(void);
2298
1da177e4 2299extern void exit_files(struct task_struct *);
a7e5328a 2300extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2301
1da177e4 2302extern void exit_itimers(struct signal_struct *);
cbaffba1 2303extern void flush_itimer_signals(void);
1da177e4 2304
9402c95f 2305extern void do_group_exit(int);
1da177e4 2306
1da177e4
LT
2307extern void daemonize(const char *, ...);
2308extern int allow_signal(int);
2309extern int disallow_signal(int);
1da177e4 2310
d7627467
DH
2311extern int do_execve(const char *,
2312 const char __user * const __user *,
2313 const char __user * const __user *, struct pt_regs *);
1da177e4 2314extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2315struct task_struct *fork_idle(int);
1da177e4
LT
2316
2317extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2318extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2319
2320#ifdef CONFIG_SMP
317f3941 2321void scheduler_ipi(void);
85ba2d86 2322extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2323#else
184748cc 2324static inline void scheduler_ipi(void) { }
85ba2d86
RM
2325static inline unsigned long wait_task_inactive(struct task_struct *p,
2326 long match_state)
2327{
2328 return 1;
2329}
1da177e4
LT
2330#endif
2331
05725f7e
JP
2332#define next_task(p) \
2333 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2334
2335#define for_each_process(p) \
2336 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2337
5bb459bb 2338extern bool current_is_single_threaded(void);
d84f4f99 2339
1da177e4
LT
2340/*
2341 * Careful: do_each_thread/while_each_thread is a double loop so
2342 * 'break' will not work as expected - use goto instead.
2343 */
2344#define do_each_thread(g, t) \
2345 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2346
2347#define while_each_thread(g, t) \
2348 while ((t = next_thread(t)) != g)
2349
7e49827c
ON
2350static inline int get_nr_threads(struct task_struct *tsk)
2351{
b3ac022c 2352 return tsk->signal->nr_threads;
7e49827c
ON
2353}
2354
087806b1
ON
2355static inline bool thread_group_leader(struct task_struct *p)
2356{
2357 return p->exit_signal >= 0;
2358}
1da177e4 2359
0804ef4b
EB
2360/* Do to the insanities of de_thread it is possible for a process
2361 * to have the pid of the thread group leader without actually being
2362 * the thread group leader. For iteration through the pids in proc
2363 * all we care about is that we have a task with the appropriate
2364 * pid, we don't actually care if we have the right task.
2365 */
e868171a 2366static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2367{
2368 return p->pid == p->tgid;
2369}
2370
bac0abd6
PE
2371static inline
2372int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2373{
2374 return p1->tgid == p2->tgid;
2375}
2376
36c8b586 2377static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2378{
05725f7e
JP
2379 return list_entry_rcu(p->thread_group.next,
2380 struct task_struct, thread_group);
47e65328
ON
2381}
2382
e868171a 2383static inline int thread_group_empty(struct task_struct *p)
1da177e4 2384{
47e65328 2385 return list_empty(&p->thread_group);
1da177e4
LT
2386}
2387
2388#define delay_group_leader(p) \
2389 (thread_group_leader(p) && !thread_group_empty(p))
2390
1da177e4 2391/*
260ea101 2392 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2393 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8 2394 * pins the final release of task.io_context. Also protects ->cpuset and
d68b46fe 2395 * ->cgroup.subsys[]. And ->vfork_done.
1da177e4
LT
2396 *
2397 * Nests both inside and outside of read_lock(&tasklist_lock).
2398 * It must not be nested with write_lock_irq(&tasklist_lock),
2399 * neither inside nor outside.
2400 */
2401static inline void task_lock(struct task_struct *p)
2402{
2403 spin_lock(&p->alloc_lock);
2404}
2405
2406static inline void task_unlock(struct task_struct *p)
2407{
2408 spin_unlock(&p->alloc_lock);
2409}
2410
b8ed374e 2411extern struct sighand_struct *__lock_task_sighand(struct task_struct *tsk,
f63ee72e
ON
2412 unsigned long *flags);
2413
9388dc30
AV
2414static inline struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2415 unsigned long *flags)
2416{
2417 struct sighand_struct *ret;
2418
2419 ret = __lock_task_sighand(tsk, flags);
2420 (void)__cond_lock(&tsk->sighand->siglock, ret);
2421 return ret;
2422}
b8ed374e 2423
f63ee72e
ON
2424static inline void unlock_task_sighand(struct task_struct *tsk,
2425 unsigned long *flags)
2426{
2427 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2428}
2429
4714d1d3 2430#ifdef CONFIG_CGROUPS
257058ae 2431static inline void threadgroup_change_begin(struct task_struct *tsk)
4714d1d3 2432{
257058ae 2433 down_read(&tsk->signal->group_rwsem);
4714d1d3 2434}
257058ae 2435static inline void threadgroup_change_end(struct task_struct *tsk)
4714d1d3 2436{
257058ae 2437 up_read(&tsk->signal->group_rwsem);
4714d1d3 2438}
77e4ef99
TH
2439
2440/**
2441 * threadgroup_lock - lock threadgroup
2442 * @tsk: member task of the threadgroup to lock
2443 *
2444 * Lock the threadgroup @tsk belongs to. No new task is allowed to enter
2445 * and member tasks aren't allowed to exit (as indicated by PF_EXITING) or
2446 * perform exec. This is useful for cases where the threadgroup needs to
2447 * stay stable across blockable operations.
2448 *
2449 * fork and exit paths explicitly call threadgroup_change_{begin|end}() for
2450 * synchronization. While held, no new task will be added to threadgroup
2451 * and no existing live task will have its PF_EXITING set.
2452 *
2453 * During exec, a task goes and puts its thread group through unusual
2454 * changes. After de-threading, exclusive access is assumed to resources
2455 * which are usually shared by tasks in the same group - e.g. sighand may
2456 * be replaced with a new one. Also, the exec'ing task takes over group
2457 * leader role including its pid. Exclude these changes while locked by
2458 * grabbing cred_guard_mutex which is used to synchronize exec path.
2459 */
257058ae 2460static inline void threadgroup_lock(struct task_struct *tsk)
4714d1d3 2461{
77e4ef99
TH
2462 /*
2463 * exec uses exit for de-threading nesting group_rwsem inside
2464 * cred_guard_mutex. Grab cred_guard_mutex first.
2465 */
2466 mutex_lock(&tsk->signal->cred_guard_mutex);
257058ae 2467 down_write(&tsk->signal->group_rwsem);
4714d1d3 2468}
77e4ef99
TH
2469
2470/**
2471 * threadgroup_unlock - unlock threadgroup
2472 * @tsk: member task of the threadgroup to unlock
2473 *
2474 * Reverse threadgroup_lock().
2475 */
257058ae 2476static inline void threadgroup_unlock(struct task_struct *tsk)
4714d1d3 2477{
257058ae 2478 up_write(&tsk->signal->group_rwsem);
77e4ef99 2479 mutex_unlock(&tsk->signal->cred_guard_mutex);
4714d1d3
BB
2480}
2481#else
257058ae
TH
2482static inline void threadgroup_change_begin(struct task_struct *tsk) {}
2483static inline void threadgroup_change_end(struct task_struct *tsk) {}
2484static inline void threadgroup_lock(struct task_struct *tsk) {}
2485static inline void threadgroup_unlock(struct task_struct *tsk) {}
4714d1d3
BB
2486#endif
2487
f037360f
AV
2488#ifndef __HAVE_THREAD_FUNCTIONS
2489
f7e4217b
RZ
2490#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2491#define task_stack_page(task) ((task)->stack)
a1261f54 2492
10ebffde
AV
2493static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2494{
2495 *task_thread_info(p) = *task_thread_info(org);
2496 task_thread_info(p)->task = p;
2497}
2498
2499static inline unsigned long *end_of_stack(struct task_struct *p)
2500{
f7e4217b 2501 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2502}
2503
f037360f
AV
2504#endif
2505
8b05c7e6
FT
2506static inline int object_is_on_stack(void *obj)
2507{
2508 void *stack = task_stack_page(current);
2509
2510 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2511}
2512
8c9843e5
BH
2513extern void thread_info_cache_init(void);
2514
7c9f8861
ES
2515#ifdef CONFIG_DEBUG_STACK_USAGE
2516static inline unsigned long stack_not_used(struct task_struct *p)
2517{
2518 unsigned long *n = end_of_stack(p);
2519
2520 do { /* Skip over canary */
2521 n++;
2522 } while (!*n);
2523
2524 return (unsigned long)n - (unsigned long)end_of_stack(p);
2525}
2526#endif
2527
1da177e4
LT
2528/* set thread flags in other task's structures
2529 * - see asm/thread_info.h for TIF_xxxx flags available
2530 */
2531static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2532{
a1261f54 2533 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2534}
2535
2536static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2537{
a1261f54 2538 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2539}
2540
2541static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2542{
a1261f54 2543 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2544}
2545
2546static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2547{
a1261f54 2548 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2549}
2550
2551static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2552{
a1261f54 2553 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2554}
2555
2556static inline void set_tsk_need_resched(struct task_struct *tsk)
2557{
2558 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2559}
2560
2561static inline void clear_tsk_need_resched(struct task_struct *tsk)
2562{
2563 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2564}
2565
8ae121ac
GH
2566static inline int test_tsk_need_resched(struct task_struct *tsk)
2567{
2568 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2569}
2570
690cc3ff
EB
2571static inline int restart_syscall(void)
2572{
2573 set_tsk_thread_flag(current, TIF_SIGPENDING);
2574 return -ERESTARTNOINTR;
2575}
2576
1da177e4
LT
2577static inline int signal_pending(struct task_struct *p)
2578{
2579 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2580}
f776d12d 2581
d9588725
RM
2582static inline int __fatal_signal_pending(struct task_struct *p)
2583{
2584 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2585}
f776d12d
MW
2586
2587static inline int fatal_signal_pending(struct task_struct *p)
2588{
2589 return signal_pending(p) && __fatal_signal_pending(p);
2590}
2591
16882c1e
ON
2592static inline int signal_pending_state(long state, struct task_struct *p)
2593{
2594 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2595 return 0;
2596 if (!signal_pending(p))
2597 return 0;
2598
16882c1e
ON
2599 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2600}
2601
1da177e4
LT
2602static inline int need_resched(void)
2603{
9404ef02 2604 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2605}
2606
2607/*
2608 * cond_resched() and cond_resched_lock(): latency reduction via
2609 * explicit rescheduling in places that are safe. The return
2610 * value indicates whether a reschedule was done in fact.
2611 * cond_resched_lock() will drop the spinlock before scheduling,
2612 * cond_resched_softirq() will enable bhs before scheduling.
2613 */
c3921ab7 2614extern int _cond_resched(void);
6f80bd98 2615
613afbf8
FW
2616#define cond_resched() ({ \
2617 __might_sleep(__FILE__, __LINE__, 0); \
2618 _cond_resched(); \
2619})
6f80bd98 2620
613afbf8
FW
2621extern int __cond_resched_lock(spinlock_t *lock);
2622
bdd4e85d 2623#ifdef CONFIG_PREEMPT_COUNT
716a4234 2624#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2625#else
716a4234 2626#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2627#endif
716a4234 2628
613afbf8 2629#define cond_resched_lock(lock) ({ \
716a4234 2630 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2631 __cond_resched_lock(lock); \
2632})
2633
2634extern int __cond_resched_softirq(void);
2635
75e1056f
VP
2636#define cond_resched_softirq() ({ \
2637 __might_sleep(__FILE__, __LINE__, SOFTIRQ_DISABLE_OFFSET); \
2638 __cond_resched_softirq(); \
613afbf8 2639})
1da177e4
LT
2640
2641/*
2642 * Does a critical section need to be broken due to another
95c354fe
NP
2643 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2644 * but a general need for low latency)
1da177e4 2645 */
95c354fe 2646static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2647{
95c354fe
NP
2648#ifdef CONFIG_PREEMPT
2649 return spin_is_contended(lock);
2650#else
1da177e4 2651 return 0;
95c354fe 2652#endif
1da177e4
LT
2653}
2654
f06febc9
FM
2655/*
2656 * Thread group CPU time accounting.
2657 */
4cd4c1b4 2658void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2659void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2660
490dea45 2661static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2662{
ee30a7b2 2663 raw_spin_lock_init(&sig->cputimer.lock);
f06febc9
FM
2664}
2665
7bb44ade
RM
2666/*
2667 * Reevaluate whether the task has signals pending delivery.
2668 * Wake the task if so.
2669 * This is required every time the blocked sigset_t changes.
2670 * callers must hold sighand->siglock.
2671 */
2672extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2673extern void recalc_sigpending(void);
2674
2675extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2676
2677/*
2678 * Wrappers for p->thread_info->cpu access. No-op on UP.
2679 */
2680#ifdef CONFIG_SMP
2681
2682static inline unsigned int task_cpu(const struct task_struct *p)
2683{
a1261f54 2684 return task_thread_info(p)->cpu;
1da177e4
LT
2685}
2686
c65cc870 2687extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2688
2689#else
2690
2691static inline unsigned int task_cpu(const struct task_struct *p)
2692{
2693 return 0;
2694}
2695
2696static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2697{
2698}
2699
2700#endif /* CONFIG_SMP */
2701
96f874e2
RR
2702extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2703extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2704
1da177e4
LT
2705extern void normalize_rt_tasks(void);
2706
7c941438 2707#ifdef CONFIG_CGROUP_SCHED
9b5b7751 2708
07e06b01 2709extern struct task_group root_task_group;
9b5b7751 2710
ec7dc8ac 2711extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2712extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2713extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2714#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2715extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2716extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2717#endif
2718#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2719extern int sched_group_set_rt_runtime(struct task_group *tg,
2720 long rt_runtime_us);
2721extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2722extern int sched_group_set_rt_period(struct task_group *tg,
2723 long rt_period_us);
2724extern long sched_group_rt_period(struct task_group *tg);
54e99124 2725extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2726#endif
9b5b7751
SV
2727#endif
2728
54e99124
DG
2729extern int task_can_switch_user(struct user_struct *up,
2730 struct task_struct *tsk);
2731
4b98d11b
AD
2732#ifdef CONFIG_TASK_XACCT
2733static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2734{
940389b8 2735 tsk->ioac.rchar += amt;
4b98d11b
AD
2736}
2737
2738static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2739{
940389b8 2740 tsk->ioac.wchar += amt;
4b98d11b
AD
2741}
2742
2743static inline void inc_syscr(struct task_struct *tsk)
2744{
940389b8 2745 tsk->ioac.syscr++;
4b98d11b
AD
2746}
2747
2748static inline void inc_syscw(struct task_struct *tsk)
2749{
940389b8 2750 tsk->ioac.syscw++;
4b98d11b
AD
2751}
2752#else
2753static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2754{
2755}
2756
2757static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2758{
2759}
2760
2761static inline void inc_syscr(struct task_struct *tsk)
2762{
2763}
2764
2765static inline void inc_syscw(struct task_struct *tsk)
2766{
2767}
2768#endif
2769
82455257
DH
2770#ifndef TASK_SIZE_OF
2771#define TASK_SIZE_OF(tsk) TASK_SIZE
2772#endif
2773
cf475ad2
BS
2774#ifdef CONFIG_MM_OWNER
2775extern void mm_update_next_owner(struct mm_struct *mm);
2776extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2777#else
2778static inline void mm_update_next_owner(struct mm_struct *mm)
2779{
2780}
2781
2782static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2783{
2784}
2785#endif /* CONFIG_MM_OWNER */
2786
3e10e716
JS
2787static inline unsigned long task_rlimit(const struct task_struct *tsk,
2788 unsigned int limit)
2789{
2790 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2791}
2792
2793static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2794 unsigned int limit)
2795{
2796 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2797}
2798
2799static inline unsigned long rlimit(unsigned int limit)
2800{
2801 return task_rlimit(current, limit);
2802}
2803
2804static inline unsigned long rlimit_max(unsigned int limit)
2805{
2806 return task_rlimit_max(current, limit);
2807}
2808
1da177e4
LT
2809#endif /* __KERNEL__ */
2810
2811#endif