cciss: export linux/cciss_defs.h header
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4/*
5 * cloning flags:
6 */
7#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
8#define CLONE_VM 0x00000100 /* set if VM shared between processes */
9#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
10#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
11#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
12#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
13#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
14#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
15#define CLONE_THREAD 0x00010000 /* Same thread group? */
16#define CLONE_NEWNS 0x00020000 /* New namespace group? */
17#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
18#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
19#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
20#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
21#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
22#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
23#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
24#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
071df104 25#define CLONE_NEWUTS 0x04000000 /* New utsname group? */
25b21cb2 26#define CLONE_NEWIPC 0x08000000 /* New ipcs */
77ec739d 27#define CLONE_NEWUSER 0x10000000 /* New user namespace */
30e49c26 28#define CLONE_NEWPID 0x20000000 /* New pid namespace */
169e3674 29#define CLONE_NEWNET 0x40000000 /* New network namespace */
fadad878 30#define CLONE_IO 0x80000000 /* Clone io context */
b7b3c76a
DW
31
32/*
33 * Scheduling policies
34 */
35#define SCHED_NORMAL 0
36#define SCHED_FIFO 1
37#define SCHED_RR 2
38#define SCHED_BATCH 3
0e6aca43
IM
39/* SCHED_ISO: reserved but not implemented yet */
40#define SCHED_IDLE 5
ca94c442
LP
41/* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
42#define SCHED_RESET_ON_FORK 0x40000000
b7b3c76a 43
a3b6714e 44#ifdef __KERNEL__
b7b3c76a
DW
45
46struct sched_param {
47 int sched_priority;
48};
49
1da177e4
LT
50#include <asm/param.h> /* for HZ */
51
1da177e4
LT
52#include <linux/capability.h>
53#include <linux/threads.h>
54#include <linux/kernel.h>
55#include <linux/types.h>
56#include <linux/timex.h>
57#include <linux/jiffies.h>
58#include <linux/rbtree.h>
59#include <linux/thread_info.h>
60#include <linux/cpumask.h>
61#include <linux/errno.h>
62#include <linux/nodemask.h>
c92ff1bd 63#include <linux/mm_types.h>
1da177e4
LT
64
65#include <asm/system.h>
1da177e4
LT
66#include <asm/page.h>
67#include <asm/ptrace.h>
1da177e4
LT
68#include <asm/cputime.h>
69
70#include <linux/smp.h>
71#include <linux/sem.h>
72#include <linux/signal.h>
5ad4e53b 73#include <linux/path.h>
1da177e4
LT
74#include <linux/compiler.h>
75#include <linux/completion.h>
76#include <linux/pid.h>
77#include <linux/percpu.h>
78#include <linux/topology.h>
3e26c149 79#include <linux/proportions.h>
1da177e4 80#include <linux/seccomp.h>
e56d0903 81#include <linux/rcupdate.h>
05725f7e 82#include <linux/rculist.h>
23f78d4a 83#include <linux/rtmutex.h>
1da177e4 84
a3b6714e
DW
85#include <linux/time.h>
86#include <linux/param.h>
87#include <linux/resource.h>
88#include <linux/timer.h>
89#include <linux/hrtimer.h>
7c3ab738 90#include <linux/task_io_accounting.h>
5cb350ba 91#include <linux/kobject.h>
9745512c 92#include <linux/latencytop.h>
9e2b2dc4 93#include <linux/cred.h>
a3b6714e
DW
94
95#include <asm/processor.h>
36d57ac4 96
1da177e4 97struct exec_domain;
c87e2837 98struct futex_pi_state;
286100a6 99struct robust_list_head;
d89d8796 100struct bio;
5ad4e53b 101struct fs_struct;
e2b371f0 102struct bts_context;
cdd6c482 103struct perf_event_context;
1da177e4 104
1da177e4
LT
105/*
106 * List of flags we want to share for kernel threads,
107 * if only because they are not used by them anyway.
108 */
109#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
110
111/*
112 * These are the constant used to fake the fixed-point load-average
113 * counting. Some notes:
114 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
115 * a load-average precision of 10 bits integer + 11 bits fractional
116 * - if you want to count load-averages more often, you need more
117 * precision, or rounding will get you. With 2-second counting freq,
118 * the EXP_n values would be 1981, 2034 and 2043 if still using only
119 * 11 bit fractions.
120 */
121extern unsigned long avenrun[]; /* Load averages */
2d02494f 122extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
1da177e4
LT
123
124#define FSHIFT 11 /* nr of bits of precision */
125#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
0c2043ab 126#define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
1da177e4
LT
127#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
128#define EXP_5 2014 /* 1/exp(5sec/5min) */
129#define EXP_15 2037 /* 1/exp(5sec/15min) */
130
131#define CALC_LOAD(load,exp,n) \
132 load *= exp; \
133 load += n*(FIXED_1-exp); \
134 load >>= FSHIFT;
135
136extern unsigned long total_forks;
137extern int nr_threads;
1da177e4
LT
138DECLARE_PER_CPU(unsigned long, process_counts);
139extern int nr_processes(void);
140extern unsigned long nr_running(void);
141extern unsigned long nr_uninterruptible(void);
142extern unsigned long nr_iowait(void);
69d25870
AV
143extern unsigned long nr_iowait_cpu(void);
144extern unsigned long this_cpu_load(void);
145
146
dce48a84 147extern void calc_global_load(void);
1da177e4 148
7e49fcce
SR
149extern unsigned long get_parent_ip(unsigned long addr);
150
43ae34cb
IM
151struct seq_file;
152struct cfs_rq;
4cf86d77 153struct task_group;
43ae34cb
IM
154#ifdef CONFIG_SCHED_DEBUG
155extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
156extern void proc_sched_set_task(struct task_struct *p);
157extern void
5cef9eca 158print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
43ae34cb
IM
159#else
160static inline void
161proc_sched_show_task(struct task_struct *p, struct seq_file *m)
162{
163}
164static inline void proc_sched_set_task(struct task_struct *p)
165{
166}
167static inline void
5cef9eca 168print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
43ae34cb
IM
169{
170}
171#endif
1da177e4 172
4a8342d2
LT
173/*
174 * Task state bitmask. NOTE! These bits are also
175 * encoded in fs/proc/array.c: get_task_state().
176 *
177 * We have two separate sets of flags: task->state
178 * is about runnability, while task->exit_state are
179 * about the task exiting. Confusing, but this way
180 * modifying one set can't modify the other one by
181 * mistake.
182 */
1da177e4
LT
183#define TASK_RUNNING 0
184#define TASK_INTERRUPTIBLE 1
185#define TASK_UNINTERRUPTIBLE 2
f021a3c2
MW
186#define __TASK_STOPPED 4
187#define __TASK_TRACED 8
4a8342d2
LT
188/* in tsk->exit_state */
189#define EXIT_ZOMBIE 16
190#define EXIT_DEAD 32
191/* in tsk->state again */
af927232 192#define TASK_DEAD 64
f021a3c2 193#define TASK_WAKEKILL 128
e9c84311 194#define TASK_WAKING 256
e1781538 195#define TASK_STATE_MAX 512
f021a3c2 196
44d90df6 197#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
73342151 198
e1781538
PZ
199extern char ___assert_task_state[1 - 2*!!(
200 sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
f021a3c2
MW
201
202/* Convenience macros for the sake of set_task_state */
203#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
204#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
205#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 206
92a1f4bc
MW
207/* Convenience macros for the sake of wake_up */
208#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 209#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW
210
211/* get_task_state() */
212#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW
213 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
214 __TASK_TRACED)
92a1f4bc 215
f021a3c2
MW
216#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
217#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 218#define task_is_stopped_or_traced(task) \
f021a3c2 219 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 220#define task_contributes_to_load(task) \
e3c8ca83 221 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
6301cb95 222 (task->flags & PF_FREEZING) == 0)
1da177e4
LT
223
224#define __set_task_state(tsk, state_value) \
225 do { (tsk)->state = (state_value); } while (0)
226#define set_task_state(tsk, state_value) \
227 set_mb((tsk)->state, (state_value))
228
498d0c57
AM
229/*
230 * set_current_state() includes a barrier so that the write of current->state
231 * is correctly serialised wrt the caller's subsequent test of whether to
232 * actually sleep:
233 *
234 * set_current_state(TASK_UNINTERRUPTIBLE);
235 * if (do_i_need_to_sleep())
236 * schedule();
237 *
238 * If the caller does not need such serialisation then use __set_current_state()
239 */
1da177e4
LT
240#define __set_current_state(state_value) \
241 do { current->state = (state_value); } while (0)
242#define set_current_state(state_value) \
243 set_mb(current->state, (state_value))
244
245/* Task command name length */
246#define TASK_COMM_LEN 16
247
1da177e4
LT
248#include <linux/spinlock.h>
249
250/*
251 * This serializes "schedule()" and also protects
252 * the run-queue from deletions/modifications (but
253 * _adding_ to the beginning of the run-queue has
254 * a separate lock).
255 */
256extern rwlock_t tasklist_lock;
257extern spinlock_t mmlist_lock;
258
36c8b586 259struct task_struct;
1da177e4
LT
260
261extern void sched_init(void);
262extern void sched_init_smp(void);
2d07b255 263extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 264extern void init_idle(struct task_struct *idle, int cpu);
1df21055 265extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 266
89f19f04 267extern int runqueue_is_locked(int cpu);
ad474cac 268extern void task_rq_unlock_wait(struct task_struct *p);
017730c1 269
6a7b3dc3 270extern cpumask_var_t nohz_cpu_mask;
46cb4b7c
SS
271#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
272extern int select_nohz_load_balancer(int cpu);
eea08f32 273extern int get_nohz_load_balancer(void);
46cb4b7c
SS
274#else
275static inline int select_nohz_load_balancer(int cpu)
276{
277 return 0;
278}
279#endif
1da177e4 280
e59e2ae2 281/*
39bc89fd 282 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM
283 */
284extern void show_state_filter(unsigned long state_filter);
285
286static inline void show_state(void)
287{
39bc89fd 288 show_state_filter(0);
e59e2ae2
IM
289}
290
1da177e4
LT
291extern void show_regs(struct pt_regs *);
292
293/*
294 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
295 * task), SP is the stack pointer of the first frame that should be shown in the back
296 * trace (or NULL if the entire call-chain of the task should be shown).
297 */
298extern void show_stack(struct task_struct *task, unsigned long *sp);
299
300void io_schedule(void);
301long io_schedule_timeout(long timeout);
302
303extern void cpu_init (void);
304extern void trap_init(void);
305extern void update_process_times(int user);
306extern void scheduler_tick(void);
307
82a1fcb9
IM
308extern void sched_show_task(struct task_struct *p);
309
8446f1d3 310#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 311extern void softlockup_tick(void);
8446f1d3 312extern void touch_softlockup_watchdog(void);
d6ad3e28 313extern void touch_softlockup_watchdog_sync(void);
04c9167f 314extern void touch_all_softlockup_watchdogs(void);
baf48f65 315extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
8d65af78 316 void __user *buffer,
baf48f65 317 size_t *lenp, loff_t *ppos);
9c44bc03 318extern unsigned int softlockup_panic;
9383d967 319extern int softlockup_thresh;
8446f1d3 320#else
6687a97d 321static inline void softlockup_tick(void)
8446f1d3
IM
322{
323}
8446f1d3
IM
324static inline void touch_softlockup_watchdog(void)
325{
326}
d6ad3e28
JW
327static inline void touch_softlockup_watchdog_sync(void)
328{
329}
04c9167f
JF
330static inline void touch_all_softlockup_watchdogs(void)
331{
332}
8446f1d3
IM
333#endif
334
e162b39a
MSB
335#ifdef CONFIG_DETECT_HUNG_TASK
336extern unsigned int sysctl_hung_task_panic;
337extern unsigned long sysctl_hung_task_check_count;
338extern unsigned long sysctl_hung_task_timeout_secs;
339extern unsigned long sysctl_hung_task_warnings;
340extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 341 void __user *buffer,
e162b39a
MSB
342 size_t *lenp, loff_t *ppos);
343#endif
8446f1d3 344
1da177e4
LT
345/* Attach to any functions which should be ignored in wchan output. */
346#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM
347
348/* Linker adds these: start and end of __sched functions */
349extern char __sched_text_start[], __sched_text_end[];
350
1da177e4
LT
351/* Is this address in the __sched functions? */
352extern int in_sched_functions(unsigned long addr);
353
354#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 355extern signed long schedule_timeout(signed long timeout);
64ed93a2 356extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 357extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 358extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 359asmlinkage void schedule(void);
0d66bf6d 360extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 361
ab516013 362struct nsproxy;
acce292c 363struct user_namespace;
1da177e4 364
341c87bf
KH
365/*
366 * Default maximum number of active map areas, this limits the number of vmas
367 * per mm struct. Users can overwrite this number by sysctl but there is a
368 * problem.
369 *
370 * When a program's coredump is generated as ELF format, a section is created
371 * per a vma. In ELF, the number of sections is represented in unsigned short.
372 * This means the number of sections should be smaller than 65535 at coredump.
373 * Because the kernel adds some informative sections to a image of program at
374 * generating coredump, we need some margin. The number of extra sections is
375 * 1-3 now and depends on arch. We use "5" as safe margin, here.
376 */
377#define MAPCOUNT_ELF_CORE_MARGIN (5)
378#define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT
379
380extern int sysctl_max_map_count;
381
382#include <linux/aio.h>
383
efc1a3b1
DH
384#ifdef CONFIG_MMU
385extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4
LT
386extern unsigned long
387arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
388 unsigned long, unsigned long);
389extern unsigned long
390arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
391 unsigned long len, unsigned long pgoff,
392 unsigned long flags);
1363c3cd
WW
393extern void arch_unmap_area(struct mm_struct *, unsigned long);
394extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
efc1a3b1
DH
395#else
396static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
397#endif
1da177e4 398
f7d0b926 399#if USE_SPLIT_PTLOCKS
f412ac08
HD
400/*
401 * The mm counters are not protected by its page_table_lock,
402 * so must be incremented atomically.
403 */
d3cb4871
CL
404#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
405#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
406#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
407#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
408#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
f412ac08 409
f7d0b926 410#else /* !USE_SPLIT_PTLOCKS */
f412ac08
HD
411/*
412 * The mm counters are protected by its page_table_lock,
413 * so can be incremented directly.
414 */
1da177e4
LT
415#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
416#define get_mm_counter(mm, member) ((mm)->_##member)
417#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
418#define inc_mm_counter(mm, member) (mm)->_##member++
419#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08 420
f7d0b926 421#endif /* !USE_SPLIT_PTLOCKS */
4294621f 422
f412ac08
HD
423#define get_mm_rss(mm) \
424 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
425#define update_hiwater_rss(mm) do { \
426 unsigned long _rss = get_mm_rss(mm); \
427 if ((mm)->hiwater_rss < _rss) \
428 (mm)->hiwater_rss = _rss; \
429} while (0)
430#define update_hiwater_vm(mm) do { \
431 if ((mm)->hiwater_vm < (mm)->total_vm) \
432 (mm)->hiwater_vm = (mm)->total_vm; \
433} while (0)
434
9de1581e
ON
435static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
436{
437 return max(mm->hiwater_rss, get_mm_rss(mm));
438}
439
1f10206c
JP
440static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
441 struct mm_struct *mm)
442{
443 unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
444
445 if (*maxrss < hiwater_rss)
446 *maxrss = hiwater_rss;
447}
448
9de1581e
ON
449static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
450{
451 return max(mm->hiwater_vm, mm->total_vm);
452}
901608d9 453
6c5d5238
KH
454extern void set_dumpable(struct mm_struct *mm, int value);
455extern int get_dumpable(struct mm_struct *mm);
456
457/* mm flags */
3cb4a0bb 458/* dumpable bits */
6c5d5238
KH
459#define MMF_DUMPABLE 0 /* core dump is permitted */
460#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 461
3cb4a0bb 462#define MMF_DUMPABLE_BITS 2
f8af4da3 463#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH
464
465/* coredump filter bits */
466#define MMF_DUMP_ANON_PRIVATE 2
467#define MMF_DUMP_ANON_SHARED 3
468#define MMF_DUMP_MAPPED_PRIVATE 4
469#define MMF_DUMP_MAPPED_SHARED 5
82df3973 470#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM
471#define MMF_DUMP_HUGETLB_PRIVATE 7
472#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 473
3cb4a0bb 474#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 475#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH
476#define MMF_DUMP_FILTER_MASK \
477 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
478#define MMF_DUMP_FILTER_DEFAULT \
e575f111 479 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM
480 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
481
482#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
483# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
484#else
485# define MMF_DUMP_MASK_DEFAULT_ELF 0
486#endif
f8af4da3
HD
487 /* leave room for more dump flags */
488#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
489
490#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 491
1da177e4
LT
492struct sighand_struct {
493 atomic_t count;
494 struct k_sigaction action[_NSIG];
495 spinlock_t siglock;
b8fceee1 496 wait_queue_head_t signalfd_wqh;
1da177e4
LT
497};
498
0e464814 499struct pacct_struct {
f6ec29a4
KK
500 int ac_flag;
501 long ac_exitcode;
0e464814 502 unsigned long ac_mem;
77787bfb
KK
503 cputime_t ac_utime, ac_stime;
504 unsigned long ac_minflt, ac_majflt;
0e464814
KK
505};
506
42c4ab41
SG
507struct cpu_itimer {
508 cputime_t expires;
509 cputime_t incr;
8356b5f9
SG
510 u32 error;
511 u32 incr_error;
42c4ab41
SG
512};
513
f06febc9
FM
514/**
515 * struct task_cputime - collected CPU time counts
516 * @utime: time spent in user mode, in &cputime_t units
517 * @stime: time spent in kernel mode, in &cputime_t units
518 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 519 *
f06febc9
FM
520 * This structure groups together three kinds of CPU time that are
521 * tracked for threads and thread groups. Most things considering
522 * CPU time want to group these counts together and treat all three
523 * of them in parallel.
524 */
525struct task_cputime {
526 cputime_t utime;
527 cputime_t stime;
528 unsigned long long sum_exec_runtime;
529};
530/* Alternate field names when used to cache expirations. */
531#define prof_exp stime
532#define virt_exp utime
533#define sched_exp sum_exec_runtime
534
4cd4c1b4
PZ
535#define INIT_CPUTIME \
536 (struct task_cputime) { \
537 .utime = cputime_zero, \
538 .stime = cputime_zero, \
539 .sum_exec_runtime = 0, \
540 }
541
c99e6efe
PZ
542/*
543 * Disable preemption until the scheduler is running.
544 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ
545 *
546 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
547 * before the scheduler is active -- see should_resched().
c99e6efe 548 */
d86ee480 549#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 550
f06febc9 551/**
4cd4c1b4
PZ
552 * struct thread_group_cputimer - thread group interval timer counts
553 * @cputime: thread group interval timers.
554 * @running: non-zero when there are timers running and
555 * @cputime receives updates.
556 * @lock: lock for fields in this struct.
f06febc9
FM
557 *
558 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 559 * used for thread group CPU timer calculations.
f06febc9 560 */
4cd4c1b4
PZ
561struct thread_group_cputimer {
562 struct task_cputime cputime;
563 int running;
564 spinlock_t lock;
f06febc9 565};
f06febc9 566
1da177e4
LT
567/*
568 * NOTE! "signal_struct" does not have it's own
569 * locking, because a shared signal_struct always
570 * implies a shared sighand_struct, so locking
571 * sighand_struct is always a proper superset of
572 * the locking of signal_struct.
573 */
574struct signal_struct {
575 atomic_t count;
576 atomic_t live;
577
578 wait_queue_head_t wait_chldexit; /* for wait4() */
579
580 /* current thread group signal load-balancing target: */
36c8b586 581 struct task_struct *curr_target;
1da177e4
LT
582
583 /* shared signal handling: */
584 struct sigpending shared_pending;
585
586 /* thread group exit support */
587 int group_exit_code;
588 /* overloaded:
589 * - notify group_exit_task when ->count is equal to notify_count
590 * - everyone except group_exit_task is stopped during signal delivery
591 * of fatal signals, group_exit_task processes the signal.
592 */
1da177e4 593 int notify_count;
07dd20e0 594 struct task_struct *group_exit_task;
1da177e4
LT
595
596 /* thread group stop support, overloads group_exit_code too */
597 int group_stop_count;
598 unsigned int flags; /* see SIGNAL_* flags below */
599
600 /* POSIX.1b Interval Timers */
601 struct list_head posix_timers;
602
603 /* ITIMER_REAL timer for the process */
2ff678b8 604 struct hrtimer real_timer;
fea9d175 605 struct pid *leader_pid;
2ff678b8 606 ktime_t it_real_incr;
1da177e4 607
42c4ab41
SG
608 /*
609 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
610 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
611 * values are defined to 0 and 1 respectively
612 */
613 struct cpu_itimer it[2];
1da177e4 614
f06febc9 615 /*
4cd4c1b4
PZ
616 * Thread group totals for process CPU timers.
617 * See thread_group_cputimer(), et al, for details.
f06febc9 618 */
4cd4c1b4 619 struct thread_group_cputimer cputimer;
f06febc9
FM
620
621 /* Earliest-expiration cache. */
622 struct task_cputime cputime_expires;
623
624 struct list_head cpu_timers[3];
625
ab521dc0 626 struct pid *tty_old_pgrp;
1ec320af 627
1da177e4
LT
628 /* boolean value for session group leader */
629 int leader;
630
631 struct tty_struct *tty; /* NULL if no tty */
632
633 /*
634 * Cumulative resource counters for dead threads in the group,
635 * and for reaped dead child processes forked by this group.
636 * Live threads maintain their own counters and add to these
637 * in __exit_signal, except for the group leader.
638 */
32bd671d 639 cputime_t utime, stime, cutime, cstime;
9ac52315
LV
640 cputime_t gtime;
641 cputime_t cgtime;
0cf55e1e
HS
642#ifndef CONFIG_VIRT_CPU_ACCOUNTING
643 cputime_t prev_utime, prev_stime;
644#endif
1da177e4
LT
645 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
646 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 647 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 648 unsigned long maxrss, cmaxrss;
940389b8 649 struct task_io_accounting ioac;
1da177e4 650
32bd671d
PZ
651 /*
652 * Cumulative ns of schedule CPU time fo dead threads in the
653 * group, not including a zombie group leader, (This only differs
654 * from jiffies_to_ns(utime + stime) if sched_clock uses something
655 * other than jiffies.)
656 */
657 unsigned long long sum_sched_runtime;
658
1da177e4
LT
659 /*
660 * We don't bother to synchronize most readers of this at all,
661 * because there is no reader checking a limit that actually needs
662 * to get both rlim_cur and rlim_max atomically, and either one
663 * alone is a single word that can safely be read normally.
664 * getrlimit/setrlimit use task_lock(current->group_leader) to
665 * protect this instead of the siglock, because they really
666 * have no need to disable irqs.
667 */
668 struct rlimit rlim[RLIM_NLIMITS];
669
0e464814
KK
670#ifdef CONFIG_BSD_PROCESS_ACCT
671 struct pacct_struct pacct; /* per-process accounting information */
672#endif
ad4ecbcb 673#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN
674 struct taskstats *stats;
675#endif
522ed776
MT
676#ifdef CONFIG_AUDIT
677 unsigned audit_tty;
678 struct tty_audit_buf *tty_audit_buf;
679#endif
28b83c51
KM
680
681 int oom_adj; /* OOM kill score adjustment (bit shift) */
1da177e4
LT
682};
683
4866cde0
NP
684/* Context switch must be unlocked if interrupts are to be enabled */
685#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
686# define __ARCH_WANT_UNLOCKED_CTXSW
687#endif
688
1da177e4
LT
689/*
690 * Bits in flags field of signal_struct.
691 */
692#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
693#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
694#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
695#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON
696/*
697 * Pending notifications to parent.
698 */
699#define SIGNAL_CLD_STOPPED 0x00000010
700#define SIGNAL_CLD_CONTINUED 0x00000020
701#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 702
fae5fa44
ON
703#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
704
ed5d2cac
ON
705/* If true, all threads except ->group_exit_task have pending SIGKILL */
706static inline int signal_group_exit(const struct signal_struct *sig)
707{
708 return (sig->flags & SIGNAL_GROUP_EXIT) ||
709 (sig->group_exit_task != NULL);
710}
711
1da177e4
LT
712/*
713 * Some day this will be a full-fledged user tracking system..
714 */
715struct user_struct {
716 atomic_t __count; /* reference count */
717 atomic_t processes; /* How many processes does this user have? */
718 atomic_t files; /* How many open files does this user have? */
719 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 720#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
721 atomic_t inotify_watches; /* How many inotify watches does this user have? */
722 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
723#endif
7ef9964e 724#ifdef CONFIG_EPOLL
7ef9964e
DL
725 atomic_t epoll_watches; /* The number of file descriptors currently watched */
726#endif
970a8645 727#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT
728 /* protected by mq_lock */
729 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 730#endif
1da177e4
LT
731 unsigned long locked_shm; /* How many pages of mlocked shm ? */
732
733#ifdef CONFIG_KEYS
734 struct key *uid_keyring; /* UID specific keyring */
735 struct key *session_keyring; /* UID's default session keyring */
736#endif
737
738 /* Hash table maintenance information */
735de223 739 struct hlist_node uidhash_node;
1da177e4 740 uid_t uid;
18b6e041 741 struct user_namespace *user_ns;
24e377a8 742
052f1dc7 743#ifdef CONFIG_USER_SCHED
4cf86d77 744 struct task_group *tg;
b1a8c172 745#ifdef CONFIG_SYSFS
eb41d946 746 struct kobject kobj;
3959214f 747 struct delayed_work work;
24e377a8 748#endif
b1a8c172 749#endif
789f90fc 750
cdd6c482 751#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ
752 atomic_long_t locked_vm;
753#endif
1da177e4
LT
754};
755
eb41d946 756extern int uids_sysfs_init(void);
5cb350ba 757
1da177e4
LT
758extern struct user_struct *find_user(uid_t);
759
760extern struct user_struct root_user;
761#define INIT_USER (&root_user)
762
b6dff3ec 763
1da177e4
LT
764struct backing_dev_info;
765struct reclaim_state;
766
52f17b6c 767#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
768struct sched_info {
769 /* cumulative counters */
2d72376b 770 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 771 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT
772
773 /* timestamps */
172ba844
BS
774 unsigned long long last_arrival,/* when we last ran on a cpu */
775 last_queued; /* when we were last queued to run */
b8efb561
IM
776#ifdef CONFIG_SCHEDSTATS
777 /* BKL stats */
480b9434 778 unsigned int bkl_count;
b8efb561 779#endif
1da177e4 780};
52f17b6c 781#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 782
ca74e92b
SN
783#ifdef CONFIG_TASK_DELAY_ACCT
784struct task_delay_info {
785 spinlock_t lock;
786 unsigned int flags; /* Private per-task flags */
787
788 /* For each stat XXX, add following, aligned appropriately
789 *
790 * struct timespec XXX_start, XXX_end;
791 * u64 XXX_delay;
792 * u32 XXX_count;
793 *
794 * Atomicity of updates to XXX_delay, XXX_count protected by
795 * single lock above (split into XXX_lock if contention is an issue).
796 */
0ff92245
SN
797
798 /*
799 * XXX_count is incremented on every XXX operation, the delay
800 * associated with the operation is added to XXX_delay.
801 * XXX_delay contains the accumulated delay time in nanoseconds.
802 */
803 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
804 u64 blkio_delay; /* wait for sync block io completion */
805 u64 swapin_delay; /* wait for swapin block io completion */
806 u32 blkio_count; /* total count of the number of sync block */
807 /* io operations performed */
808 u32 swapin_count; /* total count of the number of swapin block */
809 /* io operations performed */
873b4771
KK
810
811 struct timespec freepages_start, freepages_end;
812 u64 freepages_delay; /* wait for memory reclaim */
813 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 814};
52f17b6c
CS
815#endif /* CONFIG_TASK_DELAY_ACCT */
816
817static inline int sched_info_on(void)
818{
819#ifdef CONFIG_SCHEDSTATS
820 return 1;
821#elif defined(CONFIG_TASK_DELAY_ACCT)
822 extern int delayacct_on;
823 return delayacct_on;
824#else
825 return 0;
ca74e92b 826#endif
52f17b6c 827}
ca74e92b 828
d15bcfdb
IM
829enum cpu_idle_type {
830 CPU_IDLE,
831 CPU_NOT_IDLE,
832 CPU_NEWLY_IDLE,
833 CPU_MAX_IDLE_TYPES
1da177e4
LT
834};
835
836/*
837 * sched-domains (multiprocessor balancing) declarations:
838 */
9aa7b369
IM
839
840/*
841 * Increase resolution of nice-level calculations:
842 */
843#define SCHED_LOAD_SHIFT 10
844#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
845
f8700df7 846#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 847
2dd73a4f 848#ifdef CONFIG_SMP
b5d978e0
PZ
849#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
850#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
851#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
852#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 853#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 854#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 855#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0
PZ
856#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
857#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
858#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
859#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
c88d5910 860
b5d978e0 861#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 862
afb8a9b7
GS
863enum powersavings_balance_level {
864 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
865 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
866 * first for long running threads
867 */
868 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
869 * cpu package for power savings
870 */
871 MAX_POWERSAVINGS_BALANCE_LEVELS
872};
89c4710e 873
716707b2 874extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 875
716707b2
VS
876static inline int sd_balance_for_mc_power(void)
877{
878 if (sched_smt_power_savings)
879 return SD_POWERSAVINGS_BALANCE;
5c45bf27 880
b5d978e0 881 return SD_PREFER_SIBLING;
716707b2 882}
89c4710e 883
716707b2
VS
884static inline int sd_balance_for_package_power(void)
885{
886 if (sched_mc_power_savings | sched_smt_power_savings)
887 return SD_POWERSAVINGS_BALANCE;
888
b5d978e0 889 return SD_PREFER_SIBLING;
716707b2 890}
5c45bf27 891
100fdaee
VS
892/*
893 * Optimise SD flags for power savings:
894 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
895 * Keep default SD flags if sched_{smt,mc}_power_saving=0
896 */
897
898static inline int sd_power_saving_flags(void)
899{
900 if (sched_mc_power_savings | sched_smt_power_savings)
901 return SD_BALANCE_NEWIDLE;
902
903 return 0;
904}
1da177e4
LT
905
906struct sched_group {
907 struct sched_group *next; /* Must be a circular list */
1da177e4
LT
908
909 /*
910 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 911 * single CPU.
5517d86b 912 */
18a3885f 913 unsigned int cpu_power;
6c99e9ad 914
4200efd9
IM
915 /*
916 * The CPUs this group covers.
917 *
918 * NOTE: this field is variable length. (Allocated dynamically
919 * by attaching extra space to the end of the structure,
920 * depending on how many CPUs the kernel has booted up with)
921 *
922 * It is also be embedded into static data structures at build
923 * time. (See 'struct static_sched_group' in kernel/sched.c)
924 */
925 unsigned long cpumask[0];
1da177e4
LT
926};
927
758b2cdc
RR
928static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
929{
6c99e9ad 930 return to_cpumask(sg->cpumask);
758b2cdc
RR
931}
932
1d3504fc
HS
933enum sched_domain_level {
934 SD_LV_NONE = 0,
935 SD_LV_SIBLING,
936 SD_LV_MC,
937 SD_LV_CPU,
938 SD_LV_NODE,
939 SD_LV_ALLNODES,
940 SD_LV_MAX
941};
942
943struct sched_domain_attr {
944 int relax_domain_level;
945};
946
947#define SD_ATTR_INIT (struct sched_domain_attr) { \
948 .relax_domain_level = -1, \
949}
950
1da177e4
LT
951struct sched_domain {
952 /* These fields must be setup */
953 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 954 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 955 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT
956 unsigned long min_interval; /* Minimum balance interval ms */
957 unsigned long max_interval; /* Maximum balance interval ms */
958 unsigned int busy_factor; /* less balancing by factor if busy */
959 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 960 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP
961 unsigned int busy_idx;
962 unsigned int idle_idx;
963 unsigned int newidle_idx;
964 unsigned int wake_idx;
147cbb4b 965 unsigned int forkexec_idx;
a52bfd73 966 unsigned int smt_gain;
1da177e4 967 int flags; /* See SD_* */
1d3504fc 968 enum sched_domain_level level;
1da177e4
LT
969
970 /* Runtime fields. */
971 unsigned long last_balance; /* init to jiffies. units in jiffies */
972 unsigned int balance_interval; /* initialise to 1. units in ms. */
973 unsigned int nr_balance_failed; /* initialise to 0 */
974
2398f2c6
PZ
975 u64 last_update;
976
1da177e4
LT
977#ifdef CONFIG_SCHEDSTATS
978 /* load_balance() stats */
480b9434
KC
979 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
980 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
981 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
982 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
983 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
984 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
985 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
986 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT
987
988 /* Active load balancing */
480b9434
KC
989 unsigned int alb_count;
990 unsigned int alb_failed;
991 unsigned int alb_pushed;
1da177e4 992
68767a0a 993 /* SD_BALANCE_EXEC stats */
480b9434
KC
994 unsigned int sbe_count;
995 unsigned int sbe_balanced;
996 unsigned int sbe_pushed;
1da177e4 997
68767a0a 998 /* SD_BALANCE_FORK stats */
480b9434
KC
999 unsigned int sbf_count;
1000 unsigned int sbf_balanced;
1001 unsigned int sbf_pushed;
68767a0a 1002
1da177e4 1003 /* try_to_wake_up() stats */
480b9434
KC
1004 unsigned int ttwu_wake_remote;
1005 unsigned int ttwu_move_affine;
1006 unsigned int ttwu_move_balance;
1da177e4 1007#endif
a5d8c348
IM
1008#ifdef CONFIG_SCHED_DEBUG
1009 char *name;
1010#endif
6c99e9ad 1011
4200efd9
IM
1012 /*
1013 * Span of all CPUs in this domain.
1014 *
1015 * NOTE: this field is variable length. (Allocated dynamically
1016 * by attaching extra space to the end of the structure,
1017 * depending on how many CPUs the kernel has booted up with)
1018 *
1019 * It is also be embedded into static data structures at build
1020 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1021 */
1022 unsigned long span[0];
1da177e4
LT
1023};
1024
758b2cdc
RR
1025static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1026{
6c99e9ad 1027 return to_cpumask(sd->span);
758b2cdc
RR
1028}
1029
acc3f5d7 1030extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 1031 struct sched_domain_attr *dattr_new);
029190c5 1032
acc3f5d7
RR
1033/* Allocate an array of sched domains, for partition_sched_domains(). */
1034cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1035void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1036
06aaf76a
IM
1037/* Test a flag in parent sched domain */
1038static inline int test_sd_parent(struct sched_domain *sd, int flag)
1039{
1040 if (sd->parent && (sd->parent->flags & flag))
1041 return 1;
1042
1043 return 0;
1044}
029190c5 1045
47fe38fc
PZ
1046unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1047unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1048
1b427c15 1049#else /* CONFIG_SMP */
1da177e4 1050
1b427c15 1051struct sched_domain_attr;
d02c7a8c 1052
1b427c15 1053static inline void
acc3f5d7 1054partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM
1055 struct sched_domain_attr *dattr_new)
1056{
d02c7a8c 1057}
1b427c15 1058#endif /* !CONFIG_SMP */
1da177e4 1059
47fe38fc 1060
1da177e4 1061struct io_context; /* See blkdev.h */
1da177e4 1062
1da177e4 1063
383f2835 1064#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1065extern void prefetch_stack(struct task_struct *t);
383f2835
KC
1066#else
1067static inline void prefetch_stack(struct task_struct *t) { }
1068#endif
1da177e4
LT
1069
1070struct audit_context; /* See audit.c */
1071struct mempolicy;
b92ce558 1072struct pipe_inode_info;
4865ecf1 1073struct uts_namespace;
1da177e4 1074
20b8a59f
IM
1075struct rq;
1076struct sched_domain;
1077
7d478721
PZ
1078/*
1079 * wake flags
1080 */
1081#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1082#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1083
20b8a59f 1084struct sched_class {
5522d5d5 1085 const struct sched_class *next;
20b8a59f 1086
fd390f6a 1087 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
f02231e5 1088 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
4530d7ab 1089 void (*yield_task) (struct rq *rq);
20b8a59f 1090
7d478721 1091 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1092
fb8d4724 1093 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1094 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1095
681f3e68 1096#ifdef CONFIG_SMP
7d478721 1097 int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
4ce72a2c 1098
43010659 1099 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
e1d1484f 1100 struct rq *busiest, unsigned long max_load_move,
20b8a59f 1101 struct sched_domain *sd, enum cpu_idle_type idle,
a4ac01c3 1102 int *all_pinned, int *this_best_prio);
20b8a59f 1103
e1d1484f
PW
1104 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1105 struct rq *busiest, struct sched_domain *sd,
1106 enum cpu_idle_type idle);
9a897c5a
SR
1107 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1108 void (*post_schedule) (struct rq *this_rq);
efbbd05a
PZ
1109 void (*task_waking) (struct rq *this_rq, struct task_struct *task);
1110 void (*task_woken) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1111
cd8ba7cd 1112 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1113 const struct cpumask *newmask);
57d885fe 1114
1f11eb6a
GH
1115 void (*rq_online)(struct rq *rq);
1116 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ
1117#endif
1118
1119 void (*set_curr_task) (struct rq *rq);
1120 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1121 void (*task_fork) (struct task_struct *p);
cb469845
SR
1122
1123 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1124 int running);
1125 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1126 int running);
1127 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1128 int oldprio, int running);
810b3817 1129
dba091b9
TG
1130 unsigned int (*get_rr_interval) (struct rq *rq,
1131 struct task_struct *task);
0d721cea 1132
810b3817 1133#ifdef CONFIG_FAIR_GROUP_SCHED
88ec22d3 1134 void (*moved_group) (struct task_struct *p, int on_rq);
810b3817 1135#endif
20b8a59f
IM
1136};
1137
1138struct load_weight {
1139 unsigned long weight, inv_weight;
1140};
1141
1142/*
1143 * CFS stats for a schedulable entity (task, task-group etc)
1144 *
1145 * Current field usage histogram:
1146 *
1147 * 4 se->block_start
1148 * 4 se->run_node
1149 * 4 se->sleep_start
20b8a59f 1150 * 6 se->load.weight
20b8a59f
IM
1151 */
1152struct sched_entity {
20b8a59f
IM
1153 struct load_weight load; /* for load-balancing */
1154 struct rb_node run_node;
4a55bd5e 1155 struct list_head group_node;
20b8a59f
IM
1156 unsigned int on_rq;
1157
94c18227
IM
1158 u64 exec_start;
1159 u64 sum_exec_runtime;
e9acbff6 1160 u64 vruntime;
f6cf891c 1161 u64 prev_sum_exec_runtime;
94c18227 1162
4ae7d5ce
IM
1163 u64 last_wakeup;
1164 u64 avg_overlap;
1165
6c594c21
IM
1166 u64 nr_migrations;
1167
34cb6135
IM
1168 u64 start_runtime;
1169 u64 avg_wakeup;
34cb6135 1170
94c18227 1171#ifdef CONFIG_SCHEDSTATS
20b8a59f 1172 u64 wait_start;
94c18227 1173 u64 wait_max;
6d082592
AV
1174 u64 wait_count;
1175 u64 wait_sum;
8f0dfc34
AV
1176 u64 iowait_count;
1177 u64 iowait_sum;
94c18227 1178
20b8a59f 1179 u64 sleep_start;
20b8a59f 1180 u64 sleep_max;
94c18227
IM
1181 s64 sum_sleep_runtime;
1182
1183 u64 block_start;
20b8a59f
IM
1184 u64 block_max;
1185 u64 exec_max;
eba1ed4b 1186 u64 slice_max;
cc367732 1187
cc367732
IM
1188 u64 nr_migrations_cold;
1189 u64 nr_failed_migrations_affine;
1190 u64 nr_failed_migrations_running;
1191 u64 nr_failed_migrations_hot;
1192 u64 nr_forced_migrations;
cc367732
IM
1193
1194 u64 nr_wakeups;
1195 u64 nr_wakeups_sync;
1196 u64 nr_wakeups_migrate;
1197 u64 nr_wakeups_local;
1198 u64 nr_wakeups_remote;
1199 u64 nr_wakeups_affine;
1200 u64 nr_wakeups_affine_attempts;
1201 u64 nr_wakeups_passive;
1202 u64 nr_wakeups_idle;
94c18227
IM
1203#endif
1204
20b8a59f
IM
1205#ifdef CONFIG_FAIR_GROUP_SCHED
1206 struct sched_entity *parent;
1207 /* rq on which this entity is (to be) queued: */
1208 struct cfs_rq *cfs_rq;
1209 /* rq "owned" by this entity/group: */
1210 struct cfs_rq *my_q;
1211#endif
1212};
70b97a7f 1213
fa717060
PZ
1214struct sched_rt_entity {
1215 struct list_head run_list;
78f2c7db 1216 unsigned long timeout;
bee367ed 1217 unsigned int time_slice;
6f505b16
PZ
1218 int nr_cpus_allowed;
1219
58d6c2d7 1220 struct sched_rt_entity *back;
052f1dc7 1221#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ
1222 struct sched_rt_entity *parent;
1223 /* rq on which this entity is (to be) queued: */
1224 struct rt_rq *rt_rq;
1225 /* rq "owned" by this entity/group: */
1226 struct rt_rq *my_q;
1227#endif
fa717060
PZ
1228};
1229
86848966
PM
1230struct rcu_node;
1231
1da177e4
LT
1232struct task_struct {
1233 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1234 void *stack;
1da177e4 1235 atomic_t usage;
97dc32cd
WC
1236 unsigned int flags; /* per process flags, defined below */
1237 unsigned int ptrace;
1da177e4 1238
36772092 1239 int lock_depth; /* BKL lock depth */
1da177e4 1240
2dd73a4f
PW
1241#ifdef CONFIG_SMP
1242#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP
1243 int oncpu;
1244#endif
2dd73a4f 1245#endif
50e645a8 1246
b29739f9 1247 int prio, static_prio, normal_prio;
c7aceaba 1248 unsigned int rt_priority;
5522d5d5 1249 const struct sched_class *sched_class;
20b8a59f 1250 struct sched_entity se;
fa717060 1251 struct sched_rt_entity rt;
1da177e4 1252
e107be36
AK
1253#ifdef CONFIG_PREEMPT_NOTIFIERS
1254 /* list of struct preempt_notifier: */
1255 struct hlist_head preempt_notifiers;
1256#endif
1257
18796aa0
AD
1258 /*
1259 * fpu_counter contains the number of consecutive context switches
1260 * that the FPU is used. If this is over a threshold, the lazy fpu
1261 * saving becomes unlazy to save the trap. This is an unsigned char
1262 * so that after 256 times the counter wraps and the behavior turns
1263 * lazy again; this to deal with bursty apps that only use FPU for
1264 * a short time
1265 */
1266 unsigned char fpu_counter;
6c5c9341 1267#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1268 unsigned int btrace_seq;
6c5c9341 1269#endif
1da177e4 1270
97dc32cd 1271 unsigned int policy;
1da177e4 1272 cpumask_t cpus_allowed;
1da177e4 1273
f41d911f 1274#ifdef CONFIG_TREE_PREEMPT_RCU
e260be67 1275 int rcu_read_lock_nesting;
f41d911f 1276 char rcu_read_unlock_special;
86848966 1277 struct rcu_node *rcu_blocked_node;
f41d911f
PM
1278 struct list_head rcu_node_entry;
1279#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
e260be67 1280
52f17b6c 1281#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT
1282 struct sched_info sched_info;
1283#endif
1284
1285 struct list_head tasks;
917b627d 1286 struct plist_node pushable_tasks;
1da177e4
LT
1287
1288 struct mm_struct *mm, *active_mm;
1289
1290/* task state */
97dc32cd 1291 int exit_state;
1da177e4
LT
1292 int exit_code, exit_signal;
1293 int pdeath_signal; /* The signal sent when the parent dies */
1294 /* ??? */
97dc32cd 1295 unsigned int personality;
1da177e4 1296 unsigned did_exec:1;
f9ce1f1c
KT
1297 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1298 * execve */
8f0dfc34
AV
1299 unsigned in_iowait:1;
1300
ca94c442
LP
1301
1302 /* Revert to default priority/policy when forking */
1303 unsigned sched_reset_on_fork:1;
1304
1da177e4
LT
1305 pid_t pid;
1306 pid_t tgid;
0a425405 1307
1314562a 1308#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV
1309 /* Canary value for the -fstack-protector gcc feature */
1310 unsigned long stack_canary;
1314562a 1311#endif
e0032087 1312
1da177e4
LT
1313 /*
1314 * pointers to (original) parent process, youngest child, younger sibling,
1315 * older sibling, respectively. (p->father can be replaced with
f470021a 1316 * p->real_parent->pid)
1da177e4 1317 */
f470021a
RM
1318 struct task_struct *real_parent; /* real parent process */
1319 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1320 /*
f470021a 1321 * children/sibling forms the list of my natural children
1da177e4
LT
1322 */
1323 struct list_head children; /* list of my children */
1324 struct list_head sibling; /* linkage in my parent's children list */
1325 struct task_struct *group_leader; /* threadgroup leader */
1326
f470021a
RM
1327 /*
1328 * ptraced is the list of tasks this task is using ptrace on.
1329 * This includes both natural children and PTRACE_ATTACH targets.
1330 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1331 */
1332 struct list_head ptraced;
1333 struct list_head ptrace_entry;
1334
ca0002a1
MM
1335 /*
1336 * This is the tracer handle for the ptrace BTS extension.
1337 * This field actually belongs to the ptracer task.
1338 */
e2b371f0 1339 struct bts_context *bts;
ca0002a1 1340
1da177e4 1341 /* PID/PID hash table linkage. */
92476d7f 1342 struct pid_link pids[PIDTYPE_MAX];
47e65328 1343 struct list_head thread_group;
1da177e4
LT
1344
1345 struct completion *vfork_done; /* for vfork() */
1346 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1347 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1348
c66f08be 1349 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1350 cputime_t gtime;
d99ca3b9 1351#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1352 cputime_t prev_utime, prev_stime;
d99ca3b9 1353#endif
1da177e4 1354 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ
1355 struct timespec start_time; /* monotonic time */
1356 struct timespec real_start_time; /* boot based time */
1da177e4
LT
1357/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1358 unsigned long min_flt, maj_flt;
1359
f06febc9 1360 struct task_cputime cputime_expires;
1da177e4
LT
1361 struct list_head cpu_timers[3];
1362
1363/* process credentials */
3b11a1de
DH
1364 const struct cred *real_cred; /* objective and real subjective task
1365 * credentials (COW) */
1366 const struct cred *cred; /* effective (overridable) subjective task
1367 * credentials (COW) */
5e751e99
DH
1368 struct mutex cred_guard_mutex; /* guard against foreign influences on
1369 * credential calculations
1370 * (notably. ptrace) */
ee18d64c 1371 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1372
36772092
PBG
1373 char comm[TASK_COMM_LEN]; /* executable name excluding path
1374 - access with [gs]et_task_comm (which lock
1375 it with task_lock())
221af7f8 1376 - initialized normally by setup_new_exec */
1da177e4
LT
1377/* file system info */
1378 int link_count, total_link_count;
3d5b6fcc 1379#ifdef CONFIG_SYSVIPC
1da177e4
LT
1380/* ipc stuff */
1381 struct sysv_sem sysvsem;
3d5b6fcc 1382#endif
e162b39a 1383#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1384/* hung task detection */
82a1fcb9
IM
1385 unsigned long last_switch_count;
1386#endif
1da177e4
LT
1387/* CPU-specific state of this task */
1388 struct thread_struct thread;
1389/* filesystem information */
1390 struct fs_struct *fs;
1391/* open file information */
1392 struct files_struct *files;
1651e14e 1393/* namespaces */
ab516013 1394 struct nsproxy *nsproxy;
1da177e4
LT
1395/* signal handlers */
1396 struct signal_struct *signal;
1397 struct sighand_struct *sighand;
1398
1399 sigset_t blocked, real_blocked;
f3de272b 1400 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT
1401 struct sigpending pending;
1402
1403 unsigned long sas_ss_sp;
1404 size_t sas_ss_size;
1405 int (*notifier)(void *priv);
1406 void *notifier_data;
1407 sigset_t *notifier_mask;
1da177e4 1408 struct audit_context *audit_context;
bfef93a5
AV
1409#ifdef CONFIG_AUDITSYSCALL
1410 uid_t loginuid;
4746ec5b 1411 unsigned int sessionid;
bfef93a5 1412#endif
1da177e4
LT
1413 seccomp_t seccomp;
1414
1415/* Thread group tracking */
1416 u32 parent_exec_id;
1417 u32 self_exec_id;
58568d2a
MX
1418/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1419 * mempolicy */
1da177e4 1420 spinlock_t alloc_lock;
1da177e4 1421
3aa551c9
TG
1422#ifdef CONFIG_GENERIC_HARDIRQS
1423 /* IRQ handler threads */
1424 struct irqaction *irqaction;
1425#endif
1426
b29739f9 1427 /* Protection of the PI data structures: */
1d615482 1428 raw_spinlock_t pi_lock;
b29739f9 1429
23f78d4a
IM
1430#ifdef CONFIG_RT_MUTEXES
1431 /* PI waiters blocked on a rt_mutex held by this task */
1432 struct plist_head pi_waiters;
1433 /* Deadlock detection and priority inheritance handling */
1434 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM
1435#endif
1436
408894ee
IM
1437#ifdef CONFIG_DEBUG_MUTEXES
1438 /* mutex deadlock detection */
1439 struct mutex_waiter *blocked_on;
1440#endif
de30a2b3
IM
1441#ifdef CONFIG_TRACE_IRQFLAGS
1442 unsigned int irq_events;
de30a2b3 1443 unsigned long hardirq_enable_ip;
de30a2b3 1444 unsigned long hardirq_disable_ip;
fa1452e8 1445 unsigned int hardirq_enable_event;
de30a2b3 1446 unsigned int hardirq_disable_event;
fa1452e8
HS
1447 int hardirqs_enabled;
1448 int hardirq_context;
de30a2b3 1449 unsigned long softirq_disable_ip;
de30a2b3 1450 unsigned long softirq_enable_ip;
fa1452e8 1451 unsigned int softirq_disable_event;
de30a2b3 1452 unsigned int softirq_enable_event;
fa1452e8 1453 int softirqs_enabled;
de30a2b3
IM
1454 int softirq_context;
1455#endif
fbb9ce95 1456#ifdef CONFIG_LOCKDEP
bdb9441e 1457# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM
1458 u64 curr_chain_key;
1459 int lockdep_depth;
fbb9ce95 1460 unsigned int lockdep_recursion;
c7aceaba 1461 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1462 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1463#endif
408894ee 1464
1da177e4
LT
1465/* journalling filesystem info */
1466 void *journal_info;
1467
d89d8796
NB
1468/* stacked block device info */
1469 struct bio *bio_list, **bio_tail;
1470
1da177e4
LT
1471/* VM state */
1472 struct reclaim_state *reclaim_state;
1473
1da177e4
LT
1474 struct backing_dev_info *backing_dev_info;
1475
1476 struct io_context *io_context;
1477
1478 unsigned long ptrace_message;
1479 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1480 struct task_io_accounting ioac;
8f0ab514 1481#if defined(CONFIG_TASK_XACCT)
1da177e4
LT
1482 u64 acct_rss_mem1; /* accumulated rss usage */
1483 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1484 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT
1485#endif
1486#ifdef CONFIG_CPUSETS
58568d2a 1487 nodemask_t mems_allowed; /* Protected by alloc_lock */
825a46af 1488 int cpuset_mem_spread_rotor;
1da177e4 1489#endif
ddbcc7e8 1490#ifdef CONFIG_CGROUPS
817929ec
PM
1491 /* Control Group info protected by css_set_lock */
1492 struct css_set *cgroups;
1493 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1494 struct list_head cg_list;
ddbcc7e8 1495#endif
42b2dd0a 1496#ifdef CONFIG_FUTEX
0771dfef 1497 struct robust_list_head __user *robust_list;
34f192c6
IM
1498#ifdef CONFIG_COMPAT
1499 struct compat_robust_list_head __user *compat_robust_list;
1500#endif
c87e2837
IM
1501 struct list_head pi_state_list;
1502 struct futex_pi_state *pi_state_cache;
c7aceaba 1503#endif
cdd6c482
IM
1504#ifdef CONFIG_PERF_EVENTS
1505 struct perf_event_context *perf_event_ctxp;
1506 struct mutex perf_event_mutex;
1507 struct list_head perf_event_list;
a63eaf34 1508#endif
c7aceaba 1509#ifdef CONFIG_NUMA
58568d2a 1510 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1511 short il_next;
42b2dd0a 1512#endif
22e2c507 1513 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1514 struct rcu_head rcu;
b92ce558
JA
1515
1516 /*
1517 * cache last used pipe for splice
1518 */
1519 struct pipe_inode_info *splice_pipe;
ca74e92b
SN
1520#ifdef CONFIG_TASK_DELAY_ACCT
1521 struct task_delay_info *delays;
f4f154fd
AM
1522#endif
1523#ifdef CONFIG_FAULT_INJECTION
1524 int make_it_fail;
ca74e92b 1525#endif
3e26c149 1526 struct prop_local_single dirties;
9745512c
AV
1527#ifdef CONFIG_LATENCYTOP
1528 int latency_record_count;
1529 struct latency_record latency_record[LT_SAVECOUNT];
1530#endif
6976675d
AV
1531 /*
1532 * time slack values; these are used to round up poll() and
1533 * select() etc timeout values. These are in nanoseconds.
1534 */
1535 unsigned long timer_slack_ns;
1536 unsigned long default_timer_slack_ns;
f8d570a4
DM
1537
1538 struct list_head *scm_work_list;
fb52607a 1539#ifdef CONFIG_FUNCTION_GRAPH_TRACER
f201ae23
FW
1540 /* Index of current stored adress in ret_stack */
1541 int curr_ret_stack;
1542 /* Stack of return addresses for return function tracing */
1543 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR
1544 /* time stamp for last schedule */
1545 unsigned long long ftrace_timestamp;
f201ae23
FW
1546 /*
1547 * Number of functions that haven't been traced
1548 * because of depth overrun.
1549 */
1550 atomic_t trace_overrun;
380c4b14
FW
1551 /* Pause for the tracing */
1552 atomic_t tracing_graph_pause;
f201ae23 1553#endif
ea4e2bc4
SR
1554#ifdef CONFIG_TRACING
1555 /* state flags for use by tracers */
1556 unsigned long trace;
261842b7
SR
1557 /* bitmask of trace recursion */
1558 unsigned long trace_recursion;
1559#endif /* CONFIG_TRACING */
d899bf7b 1560 unsigned long stack_start;
569b846d
KH
1561#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
1562 struct memcg_batch_info {
1563 int do_batch; /* incremented when batch uncharge started */
1564 struct mem_cgroup *memcg; /* target memcg of uncharge */
1565 unsigned long bytes; /* uncharged usage */
1566 unsigned long memsw_bytes; /* uncharged mem+swap usage */
1567 } memcg_batch;
1568#endif
1da177e4
LT
1569};
1570
76e6eee0 1571/* Future-safe accessor for struct task_struct's cpus_allowed. */
a4636818 1572#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
76e6eee0 1573
e05606d3
IM
1574/*
1575 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1576 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1577 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1578 * values are inverted: lower p->prio value means higher priority.
1579 *
1580 * The MAX_USER_RT_PRIO value allows the actual maximum
1581 * RT priority to be separate from the value exported to
1582 * user-space. This allows kernel threads to set their
1583 * priority to a value higher than any user task. Note:
1584 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1585 */
1586
1587#define MAX_USER_RT_PRIO 100
1588#define MAX_RT_PRIO MAX_USER_RT_PRIO
1589
1590#define MAX_PRIO (MAX_RT_PRIO + 40)
1591#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1592
1593static inline int rt_prio(int prio)
1594{
1595 if (unlikely(prio < MAX_RT_PRIO))
1596 return 1;
1597 return 0;
1598}
1599
e868171a 1600static inline int rt_task(struct task_struct *p)
e05606d3
IM
1601{
1602 return rt_prio(p->prio);
1603}
1604
e868171a 1605static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB
1606{
1607 return task->pids[PIDTYPE_PID].pid;
1608}
1609
e868171a 1610static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB
1611{
1612 return task->group_leader->pids[PIDTYPE_PID].pid;
1613}
1614
6dda81f4
ON
1615/*
1616 * Without tasklist or rcu lock it is not safe to dereference
1617 * the result of task_pgrp/task_session even if task == current,
1618 * we can race with another thread doing sys_setsid/sys_setpgid.
1619 */
e868171a 1620static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB
1621{
1622 return task->group_leader->pids[PIDTYPE_PGID].pid;
1623}
1624
e868171a 1625static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB
1626{
1627 return task->group_leader->pids[PIDTYPE_SID].pid;
1628}
1629
7af57294
PE
1630struct pid_namespace;
1631
1632/*
1633 * the helpers to get the task's different pids as they are seen
1634 * from various namespaces
1635 *
1636 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB
1637 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1638 * current.
7af57294
PE
1639 * task_xid_nr_ns() : id seen from the ns specified;
1640 *
1641 * set_task_vxid() : assigns a virtual id to a task;
1642 *
7af57294
PE
1643 * see also pid_nr() etc in include/linux/pid.h
1644 */
52ee2dfd
ON
1645pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1646 struct pid_namespace *ns);
7af57294 1647
e868171a 1648static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE
1649{
1650 return tsk->pid;
1651}
1652
52ee2dfd
ON
1653static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1654 struct pid_namespace *ns)
1655{
1656 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1657}
7af57294
PE
1658
1659static inline pid_t task_pid_vnr(struct task_struct *tsk)
1660{
52ee2dfd 1661 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE
1662}
1663
1664
e868171a 1665static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE
1666{
1667 return tsk->tgid;
1668}
1669
2f2a3a46 1670pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE
1671
1672static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1673{
1674 return pid_vnr(task_tgid(tsk));
1675}
1676
1677
52ee2dfd
ON
1678static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1679 struct pid_namespace *ns)
7af57294 1680{
52ee2dfd 1681 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE
1682}
1683
7af57294
PE
1684static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1685{
52ee2dfd 1686 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE
1687}
1688
1689
52ee2dfd
ON
1690static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1691 struct pid_namespace *ns)
7af57294 1692{
52ee2dfd 1693 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE
1694}
1695
7af57294
PE
1696static inline pid_t task_session_vnr(struct task_struct *tsk)
1697{
52ee2dfd 1698 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE
1699}
1700
1b0f7ffd
ON
1701/* obsolete, do not use */
1702static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1703{
1704 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1705}
7af57294 1706
1da177e4
LT
1707/**
1708 * pid_alive - check that a task structure is not stale
1709 * @p: Task structure to be checked.
1710 *
1711 * Test if a process is not yet dead (at most zombie state)
1712 * If pid_alive fails, then pointers within the task structure
1713 * can be stale and must not be dereferenced.
1714 */
e868171a 1715static inline int pid_alive(struct task_struct *p)
1da177e4 1716{
92476d7f 1717 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
1718}
1719
f400e198 1720/**
b460cbc5 1721 * is_global_init - check if a task structure is init
3260259f
HK
1722 * @tsk: Task structure to be checked.
1723 *
1724 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1725 */
e868171a 1726static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE
1727{
1728 return tsk->pid == 1;
1729}
b460cbc5
SH
1730
1731/*
1732 * is_container_init:
1733 * check whether in the task is init in its own pid namespace.
f400e198 1734 */
b461cc03 1735extern int is_container_init(struct task_struct *tsk);
f400e198 1736
9ec52099
CLG
1737extern struct pid *cad_pid;
1738
1da177e4 1739extern void free_task(struct task_struct *tsk);
1da177e4 1740#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1741
158d9ebd 1742extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
1743
1744static inline void put_task_struct(struct task_struct *t)
1745{
1746 if (atomic_dec_and_test(&t->usage))
8c7904a0 1747 __put_task_struct(t);
e56d0903 1748}
1da177e4 1749
d180c5bc 1750extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1751extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1752
1da177e4
LT
1753/*
1754 * Per process flags
1755 */
1756#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1757 /* Not implemented yet, only for 486*/
1758#define PF_STARTING 0x00000002 /* being created */
1759#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1760#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1761#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1da177e4 1762#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1763#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT
1764#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1765#define PF_DUMPCORE 0x00000200 /* dumped core */
1766#define PF_SIGNALED 0x00000400 /* killed by a signal */
1767#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1768#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1769#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1770#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT
1771#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1772#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1773#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1774#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1775#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1776#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1777#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA
1778#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1779#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1780#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1781#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1782#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1783#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1784#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1785#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
ba96a0c8 1786#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
ebb12db5 1787#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT
1788
1789/*
1790 * Only the _current_ task can read/write to tsk->flags, but other
1791 * tasks can access tsk->flags in readonly mode for example
1792 * with tsk_used_math (like during threaded core dumping).
1793 * There is however an exception to this rule during ptrace
1794 * or during fork: the ptracer task is allowed to write to the
1795 * child->flags of its traced child (same goes for fork, the parent
1796 * can write to the child->flags), because we're guaranteed the
1797 * child is not running and in turn not changing child->flags
1798 * at the same time the parent does it.
1799 */
1800#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1801#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1802#define clear_used_math() clear_stopped_child_used_math(current)
1803#define set_used_math() set_stopped_child_used_math(current)
1804#define conditional_stopped_child_used_math(condition, child) \
1805 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1806#define conditional_used_math(condition) \
1807 conditional_stopped_child_used_math(condition, current)
1808#define copy_to_stopped_child_used_math(child) \
1809 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1810/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1811#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1812#define used_math() tsk_used_math(current)
1813
f41d911f
PM
1814#ifdef CONFIG_TREE_PREEMPT_RCU
1815
1816#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1817#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM
1818
1819static inline void rcu_copy_process(struct task_struct *p)
1820{
1821 p->rcu_read_lock_nesting = 0;
1822 p->rcu_read_unlock_special = 0;
dd5d19ba 1823 p->rcu_blocked_node = NULL;
f41d911f
PM
1824 INIT_LIST_HEAD(&p->rcu_node_entry);
1825}
1826
f41d911f
PM
1827#else
1828
1829static inline void rcu_copy_process(struct task_struct *p)
1830{
1831}
1832
1833#endif
1834
1da177e4 1835#ifdef CONFIG_SMP
cd8ba7cd 1836extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1837 const struct cpumask *new_mask);
1da177e4 1838#else
cd8ba7cd 1839static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1840 const struct cpumask *new_mask)
1da177e4 1841{
96f874e2 1842 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT
1843 return -EINVAL;
1844 return 0;
1845}
1846#endif
e0ad9556
RR
1847
1848#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT
1849static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1850{
1851 return set_cpus_allowed_ptr(p, &new_mask);
1852}
e0ad9556 1853#endif
1da177e4 1854
b342501c
IM
1855/*
1856 * Architectures can set this to 1 if they have specified
1857 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1858 * but then during bootup it turns out that sched_clock()
1859 * is reliable after all:
1860 */
1861#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1862extern int sched_clock_stable;
1863#endif
1864
1bbfa6f2
MF
1865/* ftrace calls sched_clock() directly */
1866extern unsigned long long notrace sched_clock(void);
e436d800 1867
c1955a3d
PZ
1868extern void sched_clock_init(void);
1869extern u64 sched_clock_cpu(int cpu);
3e51f33f 1870
c1955a3d 1871#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ
1872static inline void sched_clock_tick(void)
1873{
1874}
1875
1876static inline void sched_clock_idle_sleep_event(void)
1877{
1878}
1879
1880static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1881{
1882}
1883#else
3e51f33f
PZ
1884extern void sched_clock_tick(void);
1885extern void sched_clock_idle_sleep_event(void);
1886extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1887#endif
1888
e436d800
IM
1889/*
1890 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1891 * clock constructed from sched_clock():
1892 */
1893extern unsigned long long cpu_clock(int cpu);
1894
36c8b586 1895extern unsigned long long
41b86e9c 1896task_sched_runtime(struct task_struct *task);
f06febc9 1897extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT
1898
1899/* sched_exec is called by processes performing an exec */
1900#ifdef CONFIG_SMP
1901extern void sched_exec(void);
1902#else
1903#define sched_exec() {}
1904#endif
1905
2aa44d05
IM
1906extern void sched_clock_idle_sleep_event(void);
1907extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1908
1da177e4
LT
1909#ifdef CONFIG_HOTPLUG_CPU
1910extern void idle_task_exit(void);
1911#else
1912static inline void idle_task_exit(void) {}
1913#endif
1914
1915extern void sched_idle_next(void);
b29739f9 1916
06d8308c
TG
1917#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1918extern void wake_up_idle_cpu(int cpu);
1919#else
1920static inline void wake_up_idle_cpu(int cpu) { }
1921#endif
1922
21805085 1923extern unsigned int sysctl_sched_latency;
b2be5e96 1924extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1925extern unsigned int sysctl_sched_wakeup_granularity;
47fea2ad
JSR
1926extern unsigned int sysctl_sched_shares_ratelimit;
1927extern unsigned int sysctl_sched_shares_thresh;
bf0f6f24 1928extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE
1929
1930enum sched_tunable_scaling {
1931 SCHED_TUNABLESCALING_NONE,
1932 SCHED_TUNABLESCALING_LOG,
1933 SCHED_TUNABLESCALING_LINEAR,
1934 SCHED_TUNABLESCALING_END,
1935};
1936extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1937
2bba22c5 1938#ifdef CONFIG_SCHED_DEBUG
da84d961 1939extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1940extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1941extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1942extern unsigned int sysctl_timer_migration;
b2be5e96 1943
1983a922 1944int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1945 void __user *buffer, size_t *length,
b2be5e96 1946 loff_t *ppos);
2bd8e6d4 1947#endif
eea08f32
AB
1948#ifdef CONFIG_SCHED_DEBUG
1949static inline unsigned int get_sysctl_timer_migration(void)
1950{
1951 return sysctl_timer_migration;
1952}
1953#else
1954static inline unsigned int get_sysctl_timer_migration(void)
1955{
1956 return 1;
1957}
1958#endif
9f0c1e56
PZ
1959extern unsigned int sysctl_sched_rt_period;
1960extern int sysctl_sched_rt_runtime;
2bd8e6d4 1961
d0b27fa7 1962int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1963 void __user *buffer, size_t *lenp,
d0b27fa7
PZ
1964 loff_t *ppos);
1965
2bd8e6d4 1966extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1967
b29739f9 1968#ifdef CONFIG_RT_MUTEXES
36c8b586
IM
1969extern int rt_mutex_getprio(struct task_struct *p);
1970extern void rt_mutex_setprio(struct task_struct *p, int prio);
1971extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1972#else
e868171a 1973static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM
1974{
1975 return p->normal_prio;
1976}
95e02ca9 1977# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM
1978#endif
1979
36c8b586
IM
1980extern void set_user_nice(struct task_struct *p, long nice);
1981extern int task_prio(const struct task_struct *p);
1982extern int task_nice(const struct task_struct *p);
1983extern int can_nice(const struct task_struct *p, const int nice);
1984extern int task_curr(const struct task_struct *p);
1da177e4
LT
1985extern int idle_cpu(int cpu);
1986extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
961ccddd
RR
1987extern int sched_setscheduler_nocheck(struct task_struct *, int,
1988 struct sched_param *);
36c8b586
IM
1989extern struct task_struct *idle_task(int cpu);
1990extern struct task_struct *curr_task(int cpu);
1991extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT
1992
1993void yield(void);
1994
1995/*
1996 * The default (Linux) execution domain.
1997 */
1998extern struct exec_domain default_exec_domain;
1999
2000union thread_union {
2001 struct thread_info thread_info;
2002 unsigned long stack[THREAD_SIZE/sizeof(long)];
2003};
2004
2005#ifndef __HAVE_ARCH_KSTACK_END
2006static inline int kstack_end(void *addr)
2007{
2008 /* Reliable end of stack detection:
2009 * Some APM bios versions misalign the stack
2010 */
2011 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
2012}
2013#endif
2014
2015extern union thread_union init_thread_union;
2016extern struct task_struct init_task;
2017
2018extern struct mm_struct init_mm;
2019
198fe21b
PE
2020extern struct pid_namespace init_pid_ns;
2021
2022/*
2023 * find a task by one of its numerical ids
2024 *
198fe21b
PE
2025 * find_task_by_pid_ns():
2026 * finds a task by its pid in the specified namespace
228ebcbe
PE
2027 * find_task_by_vpid():
2028 * finds a task by its virtual pid
198fe21b 2029 *
e49859e7 2030 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE
2031 */
2032
228ebcbe
PE
2033extern struct task_struct *find_task_by_vpid(pid_t nr);
2034extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2035 struct pid_namespace *ns);
198fe21b 2036
8520d7c7 2037extern void __set_special_pids(struct pid *pid);
1da177e4
LT
2038
2039/* per-UID process charging. */
acce292c 2040extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT
2041static inline struct user_struct *get_uid(struct user_struct *u)
2042{
2043 atomic_inc(&u->__count);
2044 return u;
2045}
2046extern void free_uid(struct user_struct *);
28f300d2 2047extern void release_uids(struct user_namespace *ns);
1da177e4
LT
2048
2049#include <asm/current.h>
2050
3171a030 2051extern void do_timer(unsigned long ticks);
1da177e4 2052
b3c97528
HH
2053extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2054extern int wake_up_process(struct task_struct *tsk);
2055extern void wake_up_new_task(struct task_struct *tsk,
2056 unsigned long clone_flags);
1da177e4
LT
2057#ifdef CONFIG_SMP
2058 extern void kick_process(struct task_struct *tsk);
2059#else
2060 static inline void kick_process(struct task_struct *tsk) { }
2061#endif
ad46c2c4
IM
2062extern void sched_fork(struct task_struct *p, int clone_flags);
2063extern void sched_dead(struct task_struct *p);
1da177e4 2064
1da177e4
LT
2065extern void proc_caches_init(void);
2066extern void flush_signals(struct task_struct *);
3bcac026 2067extern void __flush_signals(struct task_struct *);
10ab825b 2068extern void ignore_signals(struct task_struct *);
1da177e4
LT
2069extern void flush_signal_handlers(struct task_struct *, int force_default);
2070extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2071
2072static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2073{
2074 unsigned long flags;
2075 int ret;
2076
2077 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2078 ret = dequeue_signal(tsk, mask, info);
2079 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2080
2081 return ret;
2082}
2083
2084extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2085 sigset_t *mask);
2086extern void unblock_all_signals(void);
2087extern void release_task(struct task_struct * p);
2088extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT
2089extern int force_sigsegv(int, struct task_struct *);
2090extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2091extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2092extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2093extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB
2094extern int kill_pgrp(struct pid *pid, int sig, int priv);
2095extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2096extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2097extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2098extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4 2099extern void force_sig(int, struct task_struct *);
1da177e4
LT
2100extern int send_sig(int, struct task_struct *, int);
2101extern void zap_other_threads(struct task_struct *p);
1da177e4
LT
2102extern struct sigqueue *sigqueue_alloc(void);
2103extern void sigqueue_free(struct sigqueue *);
ac5c2153 2104extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2105extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
2106extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2107
9ec52099
CLG
2108static inline int kill_cad_pid(int sig, int priv)
2109{
2110 return kill_pid(cad_pid, sig, priv);
2111}
2112
1da177e4
LT
2113/* These can be the second arg to send_sig_info/send_group_sig_info. */
2114#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2115#define SEND_SIG_PRIV ((struct siginfo *) 1)
2116#define SEND_SIG_FORCED ((struct siginfo *) 2)
2117
2a855dd0
SAS
2118/*
2119 * True if we are on the alternate signal stack.
2120 */
1da177e4
LT
2121static inline int on_sig_stack(unsigned long sp)
2122{
2a855dd0
SAS
2123#ifdef CONFIG_STACK_GROWSUP
2124 return sp >= current->sas_ss_sp &&
2125 sp - current->sas_ss_sp < current->sas_ss_size;
2126#else
2127 return sp > current->sas_ss_sp &&
2128 sp - current->sas_ss_sp <= current->sas_ss_size;
2129#endif
1da177e4
LT
2130}
2131
2132static inline int sas_ss_flags(unsigned long sp)
2133{
2134 return (current->sas_ss_size == 0 ? SS_DISABLE
2135 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2136}
2137
1da177e4
LT
2138/*
2139 * Routines for handling mm_structs
2140 */
2141extern struct mm_struct * mm_alloc(void);
2142
2143/* mmdrop drops the mm and the page tables */
b3c97528 2144extern void __mmdrop(struct mm_struct *);
1da177e4
LT
2145static inline void mmdrop(struct mm_struct * mm)
2146{
6fb43d7b 2147 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT
2148 __mmdrop(mm);
2149}
2150
2151/* mmput gets rid of the mappings and all user-space */
2152extern void mmput(struct mm_struct *);
2153/* Grab a reference to a task's mm, if it is not already going away */
2154extern struct mm_struct *get_task_mm(struct task_struct *task);
2155/* Remove the current tasks stale references to the old mm_struct */
2156extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO
2157/* Allocate a new mm structure and copy contents from tsk->mm */
2158extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2159
6f2c55b8
AD
2160extern int copy_thread(unsigned long, unsigned long, unsigned long,
2161 struct task_struct *, struct pt_regs *);
1da177e4
LT
2162extern void flush_thread(void);
2163extern void exit_thread(void);
2164
1da177e4 2165extern void exit_files(struct task_struct *);
6b3934ef 2166extern void __cleanup_signal(struct signal_struct *);
a7e5328a 2167extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2168
1da177e4 2169extern void exit_itimers(struct signal_struct *);
cbaffba1 2170extern void flush_itimer_signals(void);
1da177e4
LT
2171
2172extern NORET_TYPE void do_group_exit(int);
2173
1da177e4
LT
2174extern void daemonize(const char *, ...);
2175extern int allow_signal(int);
2176extern int disallow_signal(int);
1da177e4
LT
2177
2178extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2179extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2180struct task_struct *fork_idle(int);
1da177e4
LT
2181
2182extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2183extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT
2184
2185#ifdef CONFIG_SMP
a26b89f0 2186extern void wait_task_context_switch(struct task_struct *p);
85ba2d86 2187extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2188#else
a26b89f0 2189static inline void wait_task_context_switch(struct task_struct *p) {}
85ba2d86
RM
2190static inline unsigned long wait_task_inactive(struct task_struct *p,
2191 long match_state)
2192{
2193 return 1;
2194}
1da177e4
LT
2195#endif
2196
05725f7e
JP
2197#define next_task(p) \
2198 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT
2199
2200#define for_each_process(p) \
2201 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2202
5bb459bb 2203extern bool current_is_single_threaded(void);
d84f4f99 2204
1da177e4
LT
2205/*
2206 * Careful: do_each_thread/while_each_thread is a double loop so
2207 * 'break' will not work as expected - use goto instead.
2208 */
2209#define do_each_thread(g, t) \
2210 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2211
2212#define while_each_thread(g, t) \
2213 while ((t = next_thread(t)) != g)
2214
de12a787
EB
2215/* de_thread depends on thread_group_leader not being a pid based check */
2216#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2217
0804ef4b
EB
2218/* Do to the insanities of de_thread it is possible for a process
2219 * to have the pid of the thread group leader without actually being
2220 * the thread group leader. For iteration through the pids in proc
2221 * all we care about is that we have a task with the appropriate
2222 * pid, we don't actually care if we have the right task.
2223 */
e868171a 2224static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB
2225{
2226 return p->pid == p->tgid;
2227}
2228
bac0abd6
PE
2229static inline
2230int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2231{
2232 return p1->tgid == p2->tgid;
2233}
2234
36c8b586 2235static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2236{
05725f7e
JP
2237 return list_entry_rcu(p->thread_group.next,
2238 struct task_struct, thread_group);
47e65328
ON
2239}
2240
e868171a 2241static inline int thread_group_empty(struct task_struct *p)
1da177e4 2242{
47e65328 2243 return list_empty(&p->thread_group);
1da177e4
LT
2244}
2245
2246#define delay_group_leader(p) \
2247 (thread_group_leader(p) && !thread_group_empty(p))
2248
39c626ae
ON
2249static inline int task_detached(struct task_struct *p)
2250{
2251 return p->exit_signal == -1;
2252}
2253
1da177e4 2254/*
260ea101 2255 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2256 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM
2257 * pins the final release of task.io_context. Also protects ->cpuset and
2258 * ->cgroup.subsys[].
1da177e4
LT
2259 *
2260 * Nests both inside and outside of read_lock(&tasklist_lock).
2261 * It must not be nested with write_lock_irq(&tasklist_lock),
2262 * neither inside nor outside.
2263 */
2264static inline void task_lock(struct task_struct *p)
2265{
2266 spin_lock(&p->alloc_lock);
2267}
2268
2269static inline void task_unlock(struct task_struct *p)
2270{
2271 spin_unlock(&p->alloc_lock);
2272}
2273
f63ee72e
ON
2274extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2275 unsigned long *flags);
2276
2277static inline void unlock_task_sighand(struct task_struct *tsk,
2278 unsigned long *flags)
2279{
2280 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2281}
2282
f037360f
AV
2283#ifndef __HAVE_THREAD_FUNCTIONS
2284
f7e4217b
RZ
2285#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2286#define task_stack_page(task) ((task)->stack)
a1261f54 2287
10ebffde
AV
2288static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2289{
2290 *task_thread_info(p) = *task_thread_info(org);
2291 task_thread_info(p)->task = p;
2292}
2293
2294static inline unsigned long *end_of_stack(struct task_struct *p)
2295{
f7e4217b 2296 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV
2297}
2298
f037360f
AV
2299#endif
2300
8b05c7e6
FT
2301static inline int object_is_on_stack(void *obj)
2302{
2303 void *stack = task_stack_page(current);
2304
2305 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2306}
2307
8c9843e5
BH
2308extern void thread_info_cache_init(void);
2309
7c9f8861
ES
2310#ifdef CONFIG_DEBUG_STACK_USAGE
2311static inline unsigned long stack_not_used(struct task_struct *p)
2312{
2313 unsigned long *n = end_of_stack(p);
2314
2315 do { /* Skip over canary */
2316 n++;
2317 } while (!*n);
2318
2319 return (unsigned long)n - (unsigned long)end_of_stack(p);
2320}
2321#endif
2322
1da177e4
LT
2323/* set thread flags in other task's structures
2324 * - see asm/thread_info.h for TIF_xxxx flags available
2325 */
2326static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2327{
a1261f54 2328 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2329}
2330
2331static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2332{
a1261f54 2333 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2334}
2335
2336static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2337{
a1261f54 2338 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2339}
2340
2341static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2342{
a1261f54 2343 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2344}
2345
2346static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2347{
a1261f54 2348 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
2349}
2350
2351static inline void set_tsk_need_resched(struct task_struct *tsk)
2352{
2353 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2354}
2355
2356static inline void clear_tsk_need_resched(struct task_struct *tsk)
2357{
2358 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2359}
2360
8ae121ac
GH
2361static inline int test_tsk_need_resched(struct task_struct *tsk)
2362{
2363 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2364}
2365
690cc3ff
EB
2366static inline int restart_syscall(void)
2367{
2368 set_tsk_thread_flag(current, TIF_SIGPENDING);
2369 return -ERESTARTNOINTR;
2370}
2371
1da177e4
LT
2372static inline int signal_pending(struct task_struct *p)
2373{
2374 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2375}
f776d12d 2376
d9588725
RM
2377static inline int __fatal_signal_pending(struct task_struct *p)
2378{
2379 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2380}
f776d12d
MW
2381
2382static inline int fatal_signal_pending(struct task_struct *p)
2383{
2384 return signal_pending(p) && __fatal_signal_pending(p);
2385}
2386
16882c1e
ON
2387static inline int signal_pending_state(long state, struct task_struct *p)
2388{
2389 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2390 return 0;
2391 if (!signal_pending(p))
2392 return 0;
2393
16882c1e
ON
2394 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2395}
2396
1da177e4
LT
2397static inline int need_resched(void)
2398{
9404ef02 2399 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT
2400}
2401
2402/*
2403 * cond_resched() and cond_resched_lock(): latency reduction via
2404 * explicit rescheduling in places that are safe. The return
2405 * value indicates whether a reschedule was done in fact.
2406 * cond_resched_lock() will drop the spinlock before scheduling,
2407 * cond_resched_softirq() will enable bhs before scheduling.
2408 */
c3921ab7 2409extern int _cond_resched(void);
6f80bd98 2410
613afbf8
FW
2411#define cond_resched() ({ \
2412 __might_sleep(__FILE__, __LINE__, 0); \
2413 _cond_resched(); \
2414})
6f80bd98 2415
613afbf8
FW
2416extern int __cond_resched_lock(spinlock_t *lock);
2417
716a4234
FW
2418#ifdef CONFIG_PREEMPT
2419#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2420#else
716a4234 2421#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2422#endif
716a4234 2423
613afbf8 2424#define cond_resched_lock(lock) ({ \
716a4234 2425 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW
2426 __cond_resched_lock(lock); \
2427})
2428
2429extern int __cond_resched_softirq(void);
2430
2431#define cond_resched_softirq() ({ \
2432 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2433 __cond_resched_softirq(); \
2434})
1da177e4
LT
2435
2436/*
2437 * Does a critical section need to be broken due to another
95c354fe
NP
2438 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2439 * but a general need for low latency)
1da177e4 2440 */
95c354fe 2441static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2442{
95c354fe
NP
2443#ifdef CONFIG_PREEMPT
2444 return spin_is_contended(lock);
2445#else
1da177e4 2446 return 0;
95c354fe 2447#endif
1da177e4
LT
2448}
2449
f06febc9
FM
2450/*
2451 * Thread group CPU time accounting.
2452 */
4cd4c1b4 2453void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2454void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2455
490dea45 2456static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2457{
4cd4c1b4
PZ
2458 sig->cputimer.cputime = INIT_CPUTIME;
2459 spin_lock_init(&sig->cputimer.lock);
2460 sig->cputimer.running = 0;
f06febc9
FM
2461}
2462
f06febc9
FM
2463static inline void thread_group_cputime_free(struct signal_struct *sig)
2464{
f06febc9
FM
2465}
2466
7bb44ade
RM
2467/*
2468 * Reevaluate whether the task has signals pending delivery.
2469 * Wake the task if so.
2470 * This is required every time the blocked sigset_t changes.
2471 * callers must hold sighand->siglock.
2472 */
2473extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT
2474extern void recalc_sigpending(void);
2475
2476extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2477
2478/*
2479 * Wrappers for p->thread_info->cpu access. No-op on UP.
2480 */
2481#ifdef CONFIG_SMP
2482
2483static inline unsigned int task_cpu(const struct task_struct *p)
2484{
a1261f54 2485 return task_thread_info(p)->cpu;
1da177e4
LT
2486}
2487
c65cc870 2488extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT
2489
2490#else
2491
2492static inline unsigned int task_cpu(const struct task_struct *p)
2493{
2494 return 0;
2495}
2496
2497static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2498{
2499}
2500
2501#endif /* CONFIG_SMP */
2502
1a3c3034
IM
2503#ifdef CONFIG_TRACING
2504extern void
2505__trace_special(void *__tr, void *__data,
2506 unsigned long arg1, unsigned long arg2, unsigned long arg3);
1da177e4 2507#else
1a3c3034
IM
2508static inline void
2509__trace_special(void *__tr, void *__data,
2510 unsigned long arg1, unsigned long arg2, unsigned long arg3)
1da177e4 2511{
1da177e4
LT
2512}
2513#endif
2514
96f874e2
RR
2515extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2516extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2517
1da177e4
LT
2518extern void normalize_rt_tasks(void);
2519
052f1dc7 2520#ifdef CONFIG_GROUP_SCHED
9b5b7751 2521
4cf86d77 2522extern struct task_group init_task_group;
eff766a6
PZ
2523#ifdef CONFIG_USER_SCHED
2524extern struct task_group root_task_group;
6c415b92 2525extern void set_tg_uid(struct user_struct *user);
eff766a6 2526#endif
9b5b7751 2527
ec7dc8ac 2528extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2529extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2530extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2531#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2532extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2533extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ
2534#endif
2535#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ
2536extern int sched_group_set_rt_runtime(struct task_group *tg,
2537 long rt_runtime_us);
2538extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ
2539extern int sched_group_set_rt_period(struct task_group *tg,
2540 long rt_period_us);
2541extern long sched_group_rt_period(struct task_group *tg);
54e99124 2542extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2543#endif
9b5b7751
SV
2544#endif
2545
54e99124
DG
2546extern int task_can_switch_user(struct user_struct *up,
2547 struct task_struct *tsk);
2548
4b98d11b
AD
2549#ifdef CONFIG_TASK_XACCT
2550static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2551{
940389b8 2552 tsk->ioac.rchar += amt;
4b98d11b
AD
2553}
2554
2555static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2556{
940389b8 2557 tsk->ioac.wchar += amt;
4b98d11b
AD
2558}
2559
2560static inline void inc_syscr(struct task_struct *tsk)
2561{
940389b8 2562 tsk->ioac.syscr++;
4b98d11b
AD
2563}
2564
2565static inline void inc_syscw(struct task_struct *tsk)
2566{
940389b8 2567 tsk->ioac.syscw++;
4b98d11b
AD
2568}
2569#else
2570static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2571{
2572}
2573
2574static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2575{
2576}
2577
2578static inline void inc_syscr(struct task_struct *tsk)
2579{
2580}
2581
2582static inline void inc_syscw(struct task_struct *tsk)
2583{
2584}
2585#endif
2586
82455257
DH
2587#ifndef TASK_SIZE_OF
2588#define TASK_SIZE_OF(tsk) TASK_SIZE
2589#endif
2590
0793a61d
TG
2591/*
2592 * Call the function if the target task is executing on a CPU right now:
2593 */
2594extern void task_oncpu_function_call(struct task_struct *p,
2595 void (*func) (void *info), void *info);
2596
2597
cf475ad2
BS
2598#ifdef CONFIG_MM_OWNER
2599extern void mm_update_next_owner(struct mm_struct *mm);
2600extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2601#else
2602static inline void mm_update_next_owner(struct mm_struct *mm)
2603{
2604}
2605
2606static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2607{
2608}
2609#endif /* CONFIG_MM_OWNER */
2610
3e10e716
JS
2611static inline unsigned long task_rlimit(const struct task_struct *tsk,
2612 unsigned int limit)
2613{
2614 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_cur);
2615}
2616
2617static inline unsigned long task_rlimit_max(const struct task_struct *tsk,
2618 unsigned int limit)
2619{
2620 return ACCESS_ONCE(tsk->signal->rlim[limit].rlim_max);
2621}
2622
2623static inline unsigned long rlimit(unsigned int limit)
2624{
2625 return task_rlimit(current, limit);
2626}
2627
2628static inline unsigned long rlimit_max(unsigned int limit)
2629{
2630 return task_rlimit_max(current, limit);
2631}
2632
1da177e4
LT
2633#endif /* __KERNEL__ */
2634
2635#endif