projects / GitHub / mt8127 / android_kernel_alcatel_ttab.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
memcg: fix memory.memsw.usage_in_bytes for root cgroup
[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
f021a3c2
MW		 195
196/* Convenience macros for the sake of set_task_state */
197#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
198#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
199#define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
1da177e4 200
92a1f4bc
MW		 201/* Convenience macros for the sake of wake_up */
202#define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
f021a3c2 203#define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
92a1f4bc
MW		 204
205/* get_task_state() */
206#define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
f021a3c2
MW		 207 TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
208 __TASK_TRACED)
92a1f4bc 209
f021a3c2
MW		 210#define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
211#define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
92a1f4bc 212#define task_is_stopped_or_traced(task) \
f021a3c2 213 ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
92a1f4bc 214#define task_contributes_to_load(task) \
e3c8ca83 215 ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
6301cb95 216 (task->flags & PF_FREEZING) == 0)
1da177e4
LT		 217
218#define __set_task_state(tsk, state_value) \
219 do { (tsk)->state = (state_value); } while (0)
220#define set_task_state(tsk, state_value) \
221 set_mb((tsk)->state, (state_value))
222
498d0c57
AM		 223/*
224 * set_current_state() includes a barrier so that the write of current->state
225 * is correctly serialised wrt the caller's subsequent test of whether to
226 * actually sleep:
227 *
228 * set_current_state(TASK_UNINTERRUPTIBLE);
229 * if (do_i_need_to_sleep())
230 * schedule();
231 *
232 * If the caller does not need such serialisation then use __set_current_state()
233 */
1da177e4
LT		 234#define __set_current_state(state_value) \
235 do { current->state = (state_value); } while (0)
236#define set_current_state(state_value) \
237 set_mb(current->state, (state_value))
238
239/* Task command name length */
240#define TASK_COMM_LEN 16
241
1da177e4
LT		 242#include <linux/spinlock.h>
243
244/*
245 * This serializes "schedule()" and also protects
246 * the run-queue from deletions/modifications (but
247 * _adding_ to the beginning of the run-queue has
248 * a separate lock).
249 */
250extern rwlock_t tasklist_lock;
251extern spinlock_t mmlist_lock;
252
36c8b586 253struct task_struct;
1da177e4
LT		 254
255extern void sched_init(void);
256extern void sched_init_smp(void);
2d07b255 257extern asmlinkage void schedule_tail(struct task_struct *prev);
36c8b586 258extern void init_idle(struct task_struct *idle, int cpu);
1df21055 259extern void init_idle_bootup_task(struct task_struct *idle);
1da177e4 260
89f19f04 261extern int runqueue_is_locked(int cpu);
ad474cac 262extern void task_rq_unlock_wait(struct task_struct *p);
017730c1 263
6a7b3dc3 264extern cpumask_var_t nohz_cpu_mask;
46cb4b7c
SS		 265#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
266extern int select_nohz_load_balancer(int cpu);
eea08f32 267extern int get_nohz_load_balancer(void);
46cb4b7c
SS		 268#else
269static inline int select_nohz_load_balancer(int cpu)
270{
271 return 0;
272}
273#endif
1da177e4 274
e59e2ae2 275/*
39bc89fd 276 * Only dump TASK_* tasks. (0 for all tasks)
e59e2ae2
IM		 277 */
278extern void show_state_filter(unsigned long state_filter);
279
280static inline void show_state(void)
281{
39bc89fd 282 show_state_filter(0);
e59e2ae2
IM		 283}
284
1da177e4
LT		 285extern void show_regs(struct pt_regs *);
286
287/*
288 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
289 * task), SP is the stack pointer of the first frame that should be shown in the back
290 * trace (or NULL if the entire call-chain of the task should be shown).
291 */
292extern void show_stack(struct task_struct *task, unsigned long *sp);
293
294void io_schedule(void);
295long io_schedule_timeout(long timeout);
296
297extern void cpu_init (void);
298extern void trap_init(void);
299extern void update_process_times(int user);
300extern void scheduler_tick(void);
301
82a1fcb9
IM		 302extern void sched_show_task(struct task_struct *p);
303
8446f1d3 304#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 305extern void softlockup_tick(void);
8446f1d3 306extern void touch_softlockup_watchdog(void);
04c9167f 307extern void touch_all_softlockup_watchdogs(void);
baf48f65 308extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
8d65af78 309 void __user *buffer,
baf48f65 310 size_t *lenp, loff_t *ppos);
9c44bc03 311extern unsigned int softlockup_panic;
9383d967 312extern int softlockup_thresh;
8446f1d3 313#else
6687a97d 314static inline void softlockup_tick(void)
8446f1d3
IM		 315{
316}
8446f1d3
IM		 317static inline void touch_softlockup_watchdog(void)
318{
319}
04c9167f
JF		 320static inline void touch_all_softlockup_watchdogs(void)
321{
322}
8446f1d3
IM		 323#endif
324
e162b39a
MSB		 325#ifdef CONFIG_DETECT_HUNG_TASK
326extern unsigned int sysctl_hung_task_panic;
327extern unsigned long sysctl_hung_task_check_count;
328extern unsigned long sysctl_hung_task_timeout_secs;
329extern unsigned long sysctl_hung_task_warnings;
330extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
8d65af78 331 void __user *buffer,
e162b39a
MSB		 332 size_t *lenp, loff_t *ppos);
333#endif
8446f1d3 334
1da177e4
LT		 335/* Attach to any functions which should be ignored in wchan output. */
336#define __sched __attribute__((__section__(".sched.text")))
deaf2227
IM		 337
338/* Linker adds these: start and end of __sched functions */
339extern char __sched_text_start[], __sched_text_end[];
340
1da177e4
LT		 341/* Is this address in the __sched functions? */
342extern int in_sched_functions(unsigned long addr);
343
344#define MAX_SCHEDULE_TIMEOUT LONG_MAX
b3c97528 345extern signed long schedule_timeout(signed long timeout);
64ed93a2 346extern signed long schedule_timeout_interruptible(signed long timeout);
294d5cc2 347extern signed long schedule_timeout_killable(signed long timeout);
64ed93a2 348extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4 349asmlinkage void schedule(void);
0d66bf6d 350extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
1da177e4 351
ab516013 352struct nsproxy;
acce292c 353struct user_namespace;
1da177e4 354
341c87bf
KH		 355/*
356 * Default maximum number of active map areas, this limits the number of vmas
357 * per mm struct. Users can overwrite this number by sysctl but there is a
358 * problem.
359 *
360 * When a program's coredump is generated as ELF format, a section is created
361 * per a vma. In ELF, the number of sections is represented in unsigned short.
362 * This means the number of sections should be smaller than 65535 at coredump.
363 * Because the kernel adds some informative sections to a image of program at
364 * generating coredump, we need some margin. The number of extra sections is
365 * 1-3 now and depends on arch. We use "5" as safe margin, here.
366 */
367#define MAPCOUNT_ELF_CORE_MARGIN (5)
368#define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
1da177e4
LT		 369
370extern int sysctl_max_map_count;
371
372#include <linux/aio.h>
373
374extern unsigned long
375arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
376 unsigned long, unsigned long);
377extern unsigned long
378arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
379 unsigned long len, unsigned long pgoff,
380 unsigned long flags);
1363c3cd
WW		 381extern void arch_unmap_area(struct mm_struct *, unsigned long);
382extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 383
f7d0b926 384#if USE_SPLIT_PTLOCKS
f412ac08
HD		 385/*
386 * The mm counters are not protected by its page_table_lock,
387 * so must be incremented atomically.
388 */
d3cb4871
CL		 389#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
390#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
391#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
392#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
393#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
f412ac08 394
f7d0b926 395#else /* !USE_SPLIT_PTLOCKS */
f412ac08
HD		 396/*
397 * The mm counters are protected by its page_table_lock,
398 * so can be incremented directly.
399 */
1da177e4
LT		 400#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
401#define get_mm_counter(mm, member) ((mm)->_##member)
402#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
403#define inc_mm_counter(mm, member) (mm)->_##member++
404#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08 405
f7d0b926 406#endif /* !USE_SPLIT_PTLOCKS */
4294621f 407
f412ac08
HD		 408#define get_mm_rss(mm) \
409 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD		 410#define update_hiwater_rss(mm) do { \
411 unsigned long _rss = get_mm_rss(mm); \
412 if ((mm)->hiwater_rss < _rss) \
413 (mm)->hiwater_rss = _rss; \
414} while (0)
415#define update_hiwater_vm(mm) do { \
416 if ((mm)->hiwater_vm < (mm)->total_vm) \
417 (mm)->hiwater_vm = (mm)->total_vm; \
418} while (0)
419
9de1581e
ON		 420static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
421{
422 return max(mm->hiwater_rss, get_mm_rss(mm));
423}
424
1f10206c
JP		 425static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
426 struct mm_struct *mm)
427{
428 unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
429
430 if (*maxrss < hiwater_rss)
431 *maxrss = hiwater_rss;
432}
433
9de1581e
ON		 434static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
435{
436 return max(mm->hiwater_vm, mm->total_vm);
437}
901608d9 438
6c5d5238
KH		 439extern void set_dumpable(struct mm_struct *mm, int value);
440extern int get_dumpable(struct mm_struct *mm);
441
442/* mm flags */
3cb4a0bb 443/* dumpable bits */
6c5d5238
KH		 444#define MMF_DUMPABLE 0 /* core dump is permitted */
445#define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
f8af4da3 446
3cb4a0bb 447#define MMF_DUMPABLE_BITS 2
f8af4da3 448#define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
3cb4a0bb
KH		 449
450/* coredump filter bits */
451#define MMF_DUMP_ANON_PRIVATE 2
452#define MMF_DUMP_ANON_SHARED 3
453#define MMF_DUMP_MAPPED_PRIVATE 4
454#define MMF_DUMP_MAPPED_SHARED 5
82df3973 455#define MMF_DUMP_ELF_HEADERS 6
e575f111
KM		 456#define MMF_DUMP_HUGETLB_PRIVATE 7
457#define MMF_DUMP_HUGETLB_SHARED 8
f8af4da3 458
3cb4a0bb 459#define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
e575f111 460#define MMF_DUMP_FILTER_BITS 7
3cb4a0bb
KH		 461#define MMF_DUMP_FILTER_MASK \
462 (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
463#define MMF_DUMP_FILTER_DEFAULT \
e575f111 464 ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
656eb2cd
RM		 465 (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
466
467#ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
468# define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
469#else
470# define MMF_DUMP_MASK_DEFAULT_ELF 0
471#endif
f8af4da3
HD		 472 /* leave room for more dump flags */
473#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
474
475#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
6c5d5238 476
1da177e4
LT		 477struct sighand_struct {
478 atomic_t count;
479 struct k_sigaction action[_NSIG];
480 spinlock_t siglock;
b8fceee1 481 wait_queue_head_t signalfd_wqh;
1da177e4
LT		 482};
483
0e464814 484struct pacct_struct {
f6ec29a4
KK		 485 int ac_flag;
486 long ac_exitcode;
0e464814 487 unsigned long ac_mem;
77787bfb
KK		 488 cputime_t ac_utime, ac_stime;
489 unsigned long ac_minflt, ac_majflt;
0e464814
KK		 490};
491
42c4ab41
SG		 492struct cpu_itimer {
493 cputime_t expires;
494 cputime_t incr;
8356b5f9
SG		 495 u32 error;
496 u32 incr_error;
42c4ab41
SG		 497};
498
f06febc9
FM		 499/**
500 * struct task_cputime - collected CPU time counts
501 * @utime: time spent in user mode, in &cputime_t units
502 * @stime: time spent in kernel mode, in &cputime_t units
503 * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
5ce73a4a 504 *
f06febc9
FM		 505 * This structure groups together three kinds of CPU time that are
506 * tracked for threads and thread groups. Most things considering
507 * CPU time want to group these counts together and treat all three
508 * of them in parallel.
509 */
510struct task_cputime {
511 cputime_t utime;
512 cputime_t stime;
513 unsigned long long sum_exec_runtime;
514};
515/* Alternate field names when used to cache expirations. */
516#define prof_exp stime
517#define virt_exp utime
518#define sched_exp sum_exec_runtime
519
4cd4c1b4
PZ		 520#define INIT_CPUTIME \
521 (struct task_cputime) { \
522 .utime = cputime_zero, \
523 .stime = cputime_zero, \
524 .sum_exec_runtime = 0, \
525 }
526
c99e6efe
PZ		 527/*
528 * Disable preemption until the scheduler is running.
529 * Reset by start_kernel()->sched_init()->init_idle().
d86ee480
PZ		 530 *
531 * We include PREEMPT_ACTIVE to avoid cond_resched() from working
532 * before the scheduler is active -- see should_resched().
c99e6efe 533 */
d86ee480 534#define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
c99e6efe 535
f06febc9 536/**
4cd4c1b4
PZ		 537 * struct thread_group_cputimer - thread group interval timer counts
538 * @cputime: thread group interval timers.
539 * @running: non-zero when there are timers running and
540 * @cputime receives updates.
541 * @lock: lock for fields in this struct.
f06febc9
FM		 542 *
543 * This structure contains the version of task_cputime, above, that is
4cd4c1b4 544 * used for thread group CPU timer calculations.
f06febc9 545 */
4cd4c1b4
PZ		 546struct thread_group_cputimer {
547 struct task_cputime cputime;
548 int running;
549 spinlock_t lock;
f06febc9 550};
f06febc9 551
1da177e4
LT		 552/*
553 * NOTE! "signal_struct" does not have it's own
554 * locking, because a shared signal_struct always
555 * implies a shared sighand_struct, so locking
556 * sighand_struct is always a proper superset of
557 * the locking of signal_struct.
558 */
559struct signal_struct {
560 atomic_t count;
561 atomic_t live;
562
563 wait_queue_head_t wait_chldexit; /* for wait4() */
564
565 /* current thread group signal load-balancing target: */
36c8b586 566 struct task_struct *curr_target;
1da177e4
LT		 567
568 /* shared signal handling: */
569 struct sigpending shared_pending;
570
571 /* thread group exit support */
572 int group_exit_code;
573 /* overloaded:
574 * - notify group_exit_task when ->count is equal to notify_count
575 * - everyone except group_exit_task is stopped during signal delivery
576 * of fatal signals, group_exit_task processes the signal.
577 */
1da177e4 578 int notify_count;
07dd20e0 579 struct task_struct *group_exit_task;
1da177e4
LT		 580
581 /* thread group stop support, overloads group_exit_code too */
582 int group_stop_count;
583 unsigned int flags; /* see SIGNAL_* flags below */
584
585 /* POSIX.1b Interval Timers */
586 struct list_head posix_timers;
587
588 /* ITIMER_REAL timer for the process */
2ff678b8 589 struct hrtimer real_timer;
fea9d175 590 struct pid *leader_pid;
2ff678b8 591 ktime_t it_real_incr;
1da177e4 592
42c4ab41
SG		 593 /*
594 * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
595 * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
596 * values are defined to 0 and 1 respectively
597 */
598 struct cpu_itimer it[2];
1da177e4 599
f06febc9 600 /*
4cd4c1b4
PZ		 601 * Thread group totals for process CPU timers.
602 * See thread_group_cputimer(), et al, for details.
f06febc9 603 */
4cd4c1b4 604 struct thread_group_cputimer cputimer;
f06febc9
FM		 605
606 /* Earliest-expiration cache. */
607 struct task_cputime cputime_expires;
608
609 struct list_head cpu_timers[3];
610
ab521dc0 611 struct pid *tty_old_pgrp;
1ec320af 612
1da177e4
LT		 613 /* boolean value for session group leader */
614 int leader;
615
616 struct tty_struct *tty; /* NULL if no tty */
617
618 /*
619 * Cumulative resource counters for dead threads in the group,
620 * and for reaped dead child processes forked by this group.
621 * Live threads maintain their own counters and add to these
622 * in __exit_signal, except for the group leader.
623 */
32bd671d 624 cputime_t utime, stime, cutime, cstime;
9ac52315
LV		 625 cputime_t gtime;
626 cputime_t cgtime;
0cf55e1e
HS		 627#ifndef CONFIG_VIRT_CPU_ACCOUNTING
628 cputime_t prev_utime, prev_stime;
629#endif
1da177e4
LT		 630 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
631 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
6eaeeaba 632 unsigned long inblock, oublock, cinblock, coublock;
1f10206c 633 unsigned long maxrss, cmaxrss;
940389b8 634 struct task_io_accounting ioac;
1da177e4 635
32bd671d
PZ		 636 /*
637 * Cumulative ns of schedule CPU time fo dead threads in the
638 * group, not including a zombie group leader, (This only differs
639 * from jiffies_to_ns(utime + stime) if sched_clock uses something
640 * other than jiffies.)
641 */
642 unsigned long long sum_sched_runtime;
643
1da177e4
LT		 644 /*
645 * We don't bother to synchronize most readers of this at all,
646 * because there is no reader checking a limit that actually needs
647 * to get both rlim_cur and rlim_max atomically, and either one
648 * alone is a single word that can safely be read normally.
649 * getrlimit/setrlimit use task_lock(current->group_leader) to
650 * protect this instead of the siglock, because they really
651 * have no need to disable irqs.
652 */
653 struct rlimit rlim[RLIM_NLIMITS];
654
0e464814
KK		 655#ifdef CONFIG_BSD_PROCESS_ACCT
656 struct pacct_struct pacct; /* per-process accounting information */
657#endif
ad4ecbcb 658#ifdef CONFIG_TASKSTATS
ad4ecbcb
SN		 659 struct taskstats *stats;
660#endif
522ed776
MT		 661#ifdef CONFIG_AUDIT
662 unsigned audit_tty;
663 struct tty_audit_buf *tty_audit_buf;
664#endif
28b83c51
KM		 665
666 int oom_adj; /* OOM kill score adjustment (bit shift) */
1da177e4
LT		 667};
668
4866cde0
NP		 669/* Context switch must be unlocked if interrupts are to be enabled */
670#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
671# define __ARCH_WANT_UNLOCKED_CTXSW
672#endif
673
1da177e4
LT		 674/*
675 * Bits in flags field of signal_struct.
676 */
677#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
678#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
679#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
680#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
e4420551
ON		 681/*
682 * Pending notifications to parent.
683 */
684#define SIGNAL_CLD_STOPPED 0x00000010
685#define SIGNAL_CLD_CONTINUED 0x00000020
686#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
1da177e4 687
fae5fa44
ON		 688#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
689
ed5d2cac
ON		 690/* If true, all threads except ->group_exit_task have pending SIGKILL */
691static inline int signal_group_exit(const struct signal_struct *sig)
692{
693 return (sig->flags & SIGNAL_GROUP_EXIT) ||
694 (sig->group_exit_task != NULL);
695}
696
1da177e4
LT		 697/*
698 * Some day this will be a full-fledged user tracking system..
699 */
700struct user_struct {
701 atomic_t __count; /* reference count */
702 atomic_t processes; /* How many processes does this user have? */
703 atomic_t files; /* How many open files does this user have? */
704 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 705#ifdef CONFIG_INOTIFY_USER
0eeca283
RL		 706 atomic_t inotify_watches; /* How many inotify watches does this user have? */
707 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
708#endif
7ef9964e 709#ifdef CONFIG_EPOLL
7ef9964e
DL		 710 atomic_t epoll_watches; /* The number of file descriptors currently watched */
711#endif
970a8645 712#ifdef CONFIG_POSIX_MQUEUE
1da177e4
LT		 713 /* protected by mq_lock */
714 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
970a8645 715#endif
1da177e4
LT		 716 unsigned long locked_shm; /* How many pages of mlocked shm ? */
717
718#ifdef CONFIG_KEYS
719 struct key *uid_keyring; /* UID specific keyring */
720 struct key *session_keyring; /* UID's default session keyring */
721#endif
722
723 /* Hash table maintenance information */
735de223 724 struct hlist_node uidhash_node;
1da177e4 725 uid_t uid;
18b6e041 726 struct user_namespace *user_ns;
24e377a8 727
052f1dc7 728#ifdef CONFIG_USER_SCHED
4cf86d77 729 struct task_group *tg;
b1a8c172 730#ifdef CONFIG_SYSFS
eb41d946 731 struct kobject kobj;
3959214f 732 struct delayed_work work;
24e377a8 733#endif
b1a8c172 734#endif
789f90fc 735
cdd6c482 736#ifdef CONFIG_PERF_EVENTS
789f90fc
PZ		 737 atomic_long_t locked_vm;
738#endif
1da177e4
LT		 739};
740
eb41d946 741extern int uids_sysfs_init(void);
5cb350ba 742
1da177e4
LT		 743extern struct user_struct *find_user(uid_t);
744
745extern struct user_struct root_user;
746#define INIT_USER (&root_user)
747
b6dff3ec 748
1da177e4
LT		 749struct backing_dev_info;
750struct reclaim_state;
751
52f17b6c 752#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT		 753struct sched_info {
754 /* cumulative counters */
2d72376b 755 unsigned long pcount; /* # of times run on this cpu */
9c2c4802 756 unsigned long long run_delay; /* time spent waiting on a runqueue */
1da177e4
LT		 757
758 /* timestamps */
172ba844
BS		 759 unsigned long long last_arrival,/* when we last ran on a cpu */
760 last_queued; /* when we were last queued to run */
b8efb561
IM		 761#ifdef CONFIG_SCHEDSTATS
762 /* BKL stats */
480b9434 763 unsigned int bkl_count;
b8efb561 764#endif
1da177e4 765};
52f17b6c 766#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
1da177e4 767
ca74e92b
SN		 768#ifdef CONFIG_TASK_DELAY_ACCT
769struct task_delay_info {
770 spinlock_t lock;
771 unsigned int flags; /* Private per-task flags */
772
773 /* For each stat XXX, add following, aligned appropriately
774 *
775 * struct timespec XXX_start, XXX_end;
776 * u64 XXX_delay;
777 * u32 XXX_count;
778 *
779 * Atomicity of updates to XXX_delay, XXX_count protected by
780 * single lock above (split into XXX_lock if contention is an issue).
781 */
0ff92245
SN		 782
783 /*
784 * XXX_count is incremented on every XXX operation, the delay
785 * associated with the operation is added to XXX_delay.
786 * XXX_delay contains the accumulated delay time in nanoseconds.
787 */
788 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
789 u64 blkio_delay; /* wait for sync block io completion */
790 u64 swapin_delay; /* wait for swapin block io completion */
791 u32 blkio_count; /* total count of the number of sync block */
792 /* io operations performed */
793 u32 swapin_count; /* total count of the number of swapin block */
794 /* io operations performed */
873b4771
KK		 795
796 struct timespec freepages_start, freepages_end;
797 u64 freepages_delay; /* wait for memory reclaim */
798 u32 freepages_count; /* total count of memory reclaim */
ca74e92b 799};
52f17b6c
CS		 800#endif /* CONFIG_TASK_DELAY_ACCT */
801
802static inline int sched_info_on(void)
803{
804#ifdef CONFIG_SCHEDSTATS
805 return 1;
806#elif defined(CONFIG_TASK_DELAY_ACCT)
807 extern int delayacct_on;
808 return delayacct_on;
809#else
810 return 0;
ca74e92b 811#endif
52f17b6c 812}
ca74e92b 813
d15bcfdb
IM		 814enum cpu_idle_type {
815 CPU_IDLE,
816 CPU_NOT_IDLE,
817 CPU_NEWLY_IDLE,
818 CPU_MAX_IDLE_TYPES
1da177e4
LT		 819};
820
821/*
822 * sched-domains (multiprocessor balancing) declarations:
823 */
9aa7b369
IM		 824
825/*
826 * Increase resolution of nice-level calculations:
827 */
828#define SCHED_LOAD_SHIFT 10
829#define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
830
f8700df7 831#define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
1da177e4 832
2dd73a4f 833#ifdef CONFIG_SMP
b5d978e0
PZ		 834#define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
835#define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
836#define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
837#define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
c88d5910 838#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
b5d978e0 839#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
59abf026 840#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
b5d978e0
PZ		 841#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
842#define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
843#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
844#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
c88d5910 845
b5d978e0 846#define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
5c45bf27 847
afb8a9b7
GS		 848enum powersavings_balance_level {
849 POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
850 POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
851 * first for long running threads
852 */
853 POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
854 * cpu package for power savings
855 */
856 MAX_POWERSAVINGS_BALANCE_LEVELS
857};
89c4710e 858
716707b2 859extern int sched_mc_power_savings, sched_smt_power_savings;
89c4710e 860
716707b2
VS		 861static inline int sd_balance_for_mc_power(void)
862{
863 if (sched_smt_power_savings)
864 return SD_POWERSAVINGS_BALANCE;
5c45bf27 865
b5d978e0 866 return SD_PREFER_SIBLING;
716707b2 867}
89c4710e 868
716707b2
VS		 869static inline int sd_balance_for_package_power(void)
870{
871 if (sched_mc_power_savings | sched_smt_power_savings)
872 return SD_POWERSAVINGS_BALANCE;
873
b5d978e0 874 return SD_PREFER_SIBLING;
716707b2 875}
5c45bf27 876
100fdaee
VS		 877/*
878 * Optimise SD flags for power savings:
879 * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
880 * Keep default SD flags if sched_{smt,mc}_power_saving=0
881 */
882
883static inline int sd_power_saving_flags(void)
884{
885 if (sched_mc_power_savings | sched_smt_power_savings)
886 return SD_BALANCE_NEWIDLE;
887
888 return 0;
889}
1da177e4
LT		 890
891struct sched_group {
892 struct sched_group *next; /* Must be a circular list */
1da177e4
LT		 893
894 /*
895 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
18a3885f 896 * single CPU.
5517d86b 897 */
18a3885f 898 unsigned int cpu_power;
6c99e9ad 899
4200efd9
IM		 900 /*
901 * The CPUs this group covers.
902 *
903 * NOTE: this field is variable length. (Allocated dynamically
904 * by attaching extra space to the end of the structure,
905 * depending on how many CPUs the kernel has booted up with)
906 *
907 * It is also be embedded into static data structures at build
908 * time. (See 'struct static_sched_group' in kernel/sched.c)
909 */
910 unsigned long cpumask[0];
1da177e4
LT		 911};
912
758b2cdc
RR		 913static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
914{
6c99e9ad 915 return to_cpumask(sg->cpumask);
758b2cdc
RR		 916}
917
1d3504fc
HS		 918enum sched_domain_level {
919 SD_LV_NONE = 0,
920 SD_LV_SIBLING,
921 SD_LV_MC,
922 SD_LV_CPU,
923 SD_LV_NODE,
924 SD_LV_ALLNODES,
925 SD_LV_MAX
926};
927
928struct sched_domain_attr {
929 int relax_domain_level;
930};
931
932#define SD_ATTR_INIT (struct sched_domain_attr) { \
933 .relax_domain_level = -1, \
934}
935
1da177e4
LT		 936struct sched_domain {
937 /* These fields must be setup */
938 struct sched_domain *parent; /* top domain must be null terminated */
1a848870 939 struct sched_domain *child; /* bottom domain must be null terminated */
1da177e4 940 struct sched_group *groups; /* the balancing groups of the domain */
1da177e4
LT		 941 unsigned long min_interval; /* Minimum balance interval ms */
942 unsigned long max_interval; /* Maximum balance interval ms */
943 unsigned int busy_factor; /* less balancing by factor if busy */
944 unsigned int imbalance_pct; /* No balance until over watermark */
1da177e4 945 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
7897986b
NP		 946 unsigned int busy_idx;
947 unsigned int idle_idx;
948 unsigned int newidle_idx;
949 unsigned int wake_idx;
147cbb4b 950 unsigned int forkexec_idx;
a52bfd73 951 unsigned int smt_gain;
1da177e4 952 int flags; /* See SD_* */
1d3504fc 953 enum sched_domain_level level;
1da177e4
LT		 954
955 /* Runtime fields. */
956 unsigned long last_balance; /* init to jiffies. units in jiffies */
957 unsigned int balance_interval; /* initialise to 1. units in ms. */
958 unsigned int nr_balance_failed; /* initialise to 0 */
959
2398f2c6
PZ		 960 u64 last_update;
961
1da177e4
LT		 962#ifdef CONFIG_SCHEDSTATS
963 /* load_balance() stats */
480b9434
KC		 964 unsigned int lb_count[CPU_MAX_IDLE_TYPES];
965 unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
966 unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
967 unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
968 unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
969 unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
970 unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
971 unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
1da177e4
LT		 972
973 /* Active load balancing */
480b9434
KC		 974 unsigned int alb_count;
975 unsigned int alb_failed;
976 unsigned int alb_pushed;
1da177e4 977
68767a0a 978 /* SD_BALANCE_EXEC stats */
480b9434
KC		 979 unsigned int sbe_count;
980 unsigned int sbe_balanced;
981 unsigned int sbe_pushed;
1da177e4 982
68767a0a 983 /* SD_BALANCE_FORK stats */
480b9434
KC		 984 unsigned int sbf_count;
985 unsigned int sbf_balanced;
986 unsigned int sbf_pushed;
68767a0a 987
1da177e4 988 /* try_to_wake_up() stats */
480b9434
KC		 989 unsigned int ttwu_wake_remote;
990 unsigned int ttwu_move_affine;
991 unsigned int ttwu_move_balance;
1da177e4 992#endif
a5d8c348
IM		 993#ifdef CONFIG_SCHED_DEBUG
994 char *name;
995#endif
6c99e9ad 996
4200efd9
IM		 997 /*
998 * Span of all CPUs in this domain.
999 *
1000 * NOTE: this field is variable length. (Allocated dynamically
1001 * by attaching extra space to the end of the structure,
1002 * depending on how many CPUs the kernel has booted up with)
1003 *
1004 * It is also be embedded into static data structures at build
1005 * time. (See 'struct static_sched_domain' in kernel/sched.c)
1006 */
1007 unsigned long span[0];
1da177e4
LT		 1008};
1009
758b2cdc
RR		 1010static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
1011{
6c99e9ad 1012 return to_cpumask(sd->span);
758b2cdc
RR		 1013}
1014
acc3f5d7 1015extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1d3504fc 1016 struct sched_domain_attr *dattr_new);
029190c5 1017
acc3f5d7
RR		 1018/* Allocate an array of sched domains, for partition_sched_domains(). */
1019cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
1020void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
1021
06aaf76a
IM		 1022/* Test a flag in parent sched domain */
1023static inline int test_sd_parent(struct sched_domain *sd, int flag)
1024{
1025 if (sd->parent && (sd->parent->flags & flag))
1026 return 1;
1027
1028 return 0;
1029}
029190c5 1030
47fe38fc
PZ		 1031unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
1032unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
1033
1b427c15 1034#else /* CONFIG_SMP */
1da177e4 1035
1b427c15 1036struct sched_domain_attr;
d02c7a8c 1037
1b427c15 1038static inline void
acc3f5d7 1039partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
1b427c15
IM		 1040 struct sched_domain_attr *dattr_new)
1041{
d02c7a8c 1042}
1b427c15 1043#endif /* !CONFIG_SMP */
1da177e4 1044
47fe38fc 1045
1da177e4 1046struct io_context; /* See blkdev.h */
1da177e4 1047
1da177e4 1048
383f2835 1049#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
36c8b586 1050extern void prefetch_stack(struct task_struct *t);
383f2835
KC		 1051#else
1052static inline void prefetch_stack(struct task_struct *t) { }
1053#endif
1da177e4
LT		 1054
1055struct audit_context; /* See audit.c */
1056struct mempolicy;
b92ce558 1057struct pipe_inode_info;
4865ecf1 1058struct uts_namespace;
1da177e4 1059
20b8a59f
IM		 1060struct rq;
1061struct sched_domain;
1062
7d478721
PZ		 1063/*
1064 * wake flags
1065 */
1066#define WF_SYNC 0x01 /* waker goes to sleep after wakup */
a7558e01 1067#define WF_FORK 0x02 /* child wakeup after fork */
7d478721 1068
20b8a59f 1069struct sched_class {
5522d5d5 1070 const struct sched_class *next;
20b8a59f 1071
fd390f6a 1072 void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
f02231e5 1073 void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
4530d7ab 1074 void (*yield_task) (struct rq *rq);
20b8a59f 1075
7d478721 1076 void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
20b8a59f 1077
fb8d4724 1078 struct task_struct * (*pick_next_task) (struct rq *rq);
31ee529c 1079 void (*put_prev_task) (struct rq *rq, struct task_struct *p);
20b8a59f 1080
681f3e68 1081#ifdef CONFIG_SMP
7d478721 1082 int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
4ce72a2c 1083
43010659 1084 unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
e1d1484f 1085 struct rq *busiest, unsigned long max_load_move,
20b8a59f 1086 struct sched_domain *sd, enum cpu_idle_type idle,
a4ac01c3 1087 int *all_pinned, int *this_best_prio);
20b8a59f 1088
e1d1484f
PW		 1089 int (*move_one_task) (struct rq *this_rq, int this_cpu,
1090 struct rq *busiest, struct sched_domain *sd,
1091 enum cpu_idle_type idle);
9a897c5a
SR		 1092 void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
1093 void (*post_schedule) (struct rq *this_rq);
1094 void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
e1d1484f 1095
cd8ba7cd 1096 void (*set_cpus_allowed)(struct task_struct *p,
96f874e2 1097 const struct cpumask *newmask);
57d885fe 1098
1f11eb6a
GH		 1099 void (*rq_online)(struct rq *rq);
1100 void (*rq_offline)(struct rq *rq);
4ce72a2c
LZ		 1101#endif
1102
1103 void (*set_curr_task) (struct rq *rq);
1104 void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
cd29fe6f 1105 void (*task_fork) (struct task_struct *p);
cb469845
SR		 1106
1107 void (*switched_from) (struct rq *this_rq, struct task_struct *task,
1108 int running);
1109 void (*switched_to) (struct rq *this_rq, struct task_struct *task,
1110 int running);
1111 void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1112 int oldprio, int running);
810b3817 1113
dba091b9
TG		 1114 unsigned int (*get_rr_interval) (struct rq *rq,
1115 struct task_struct *task);
0d721cea 1116
810b3817
PZ		 1117#ifdef CONFIG_FAIR_GROUP_SCHED
1118 void (*moved_group) (struct task_struct *p);
1119#endif
20b8a59f
IM		 1120};
1121
1122struct load_weight {
1123 unsigned long weight, inv_weight;
1124};
1125
1126/*
1127 * CFS stats for a schedulable entity (task, task-group etc)
1128 *
1129 * Current field usage histogram:
1130 *
1131 * 4 se->block_start
1132 * 4 se->run_node
1133 * 4 se->sleep_start
20b8a59f 1134 * 6 se->load.weight
20b8a59f
IM		 1135 */
1136struct sched_entity {
20b8a59f
IM		 1137 struct load_weight load; /* for load-balancing */
1138 struct rb_node run_node;
4a55bd5e 1139 struct list_head group_node;
20b8a59f
IM		 1140 unsigned int on_rq;
1141
94c18227
IM		 1142 u64 exec_start;
1143 u64 sum_exec_runtime;
e9acbff6 1144 u64 vruntime;
f6cf891c 1145 u64 prev_sum_exec_runtime;
94c18227 1146
4ae7d5ce
IM		 1147 u64 last_wakeup;
1148 u64 avg_overlap;
1149
6c594c21
IM		 1150 u64 nr_migrations;
1151
34cb6135
IM		 1152 u64 start_runtime;
1153 u64 avg_wakeup;
34cb6135 1154
94c18227 1155#ifdef CONFIG_SCHEDSTATS
20b8a59f 1156 u64 wait_start;
94c18227 1157 u64 wait_max;
6d082592
AV		 1158 u64 wait_count;
1159 u64 wait_sum;
8f0dfc34
AV		 1160 u64 iowait_count;
1161 u64 iowait_sum;
94c18227 1162
20b8a59f 1163 u64 sleep_start;
20b8a59f 1164 u64 sleep_max;
94c18227
IM		 1165 s64 sum_sleep_runtime;
1166
1167 u64 block_start;
20b8a59f
IM		 1168 u64 block_max;
1169 u64 exec_max;
eba1ed4b 1170 u64 slice_max;
cc367732 1171
cc367732
IM		 1172 u64 nr_migrations_cold;
1173 u64 nr_failed_migrations_affine;
1174 u64 nr_failed_migrations_running;
1175 u64 nr_failed_migrations_hot;
1176 u64 nr_forced_migrations;
cc367732
IM		 1177
1178 u64 nr_wakeups;
1179 u64 nr_wakeups_sync;
1180 u64 nr_wakeups_migrate;
1181 u64 nr_wakeups_local;
1182 u64 nr_wakeups_remote;
1183 u64 nr_wakeups_affine;
1184 u64 nr_wakeups_affine_attempts;
1185 u64 nr_wakeups_passive;
1186 u64 nr_wakeups_idle;
94c18227
IM		 1187#endif
1188
20b8a59f
IM		 1189#ifdef CONFIG_FAIR_GROUP_SCHED
1190 struct sched_entity *parent;
1191 /* rq on which this entity is (to be) queued: */
1192 struct cfs_rq *cfs_rq;
1193 /* rq "owned" by this entity/group: */
1194 struct cfs_rq *my_q;
1195#endif
1196};
70b97a7f 1197
fa717060
PZ		 1198struct sched_rt_entity {
1199 struct list_head run_list;
78f2c7db 1200 unsigned long timeout;
bee367ed 1201 unsigned int time_slice;
6f505b16
PZ		 1202 int nr_cpus_allowed;
1203
58d6c2d7 1204 struct sched_rt_entity *back;
052f1dc7 1205#ifdef CONFIG_RT_GROUP_SCHED
6f505b16
PZ		 1206 struct sched_rt_entity *parent;
1207 /* rq on which this entity is (to be) queued: */
1208 struct rt_rq *rt_rq;
1209 /* rq "owned" by this entity/group: */
1210 struct rt_rq *my_q;
1211#endif
fa717060
PZ		 1212};
1213
86848966
PM		 1214struct rcu_node;
1215
1da177e4
LT		 1216struct task_struct {
1217 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
f7e4217b 1218 void *stack;
1da177e4 1219 atomic_t usage;
97dc32cd
WC		 1220 unsigned int flags; /* per process flags, defined below */
1221 unsigned int ptrace;
1da177e4 1222
36772092 1223 int lock_depth; /* BKL lock depth */
1da177e4 1224
2dd73a4f
PW		 1225#ifdef CONFIG_SMP
1226#ifdef __ARCH_WANT_UNLOCKED_CTXSW
4866cde0
NP		 1227 int oncpu;
1228#endif
2dd73a4f 1229#endif
50e645a8 1230
b29739f9 1231 int prio, static_prio, normal_prio;
c7aceaba 1232 unsigned int rt_priority;
5522d5d5 1233 const struct sched_class *sched_class;
20b8a59f 1234 struct sched_entity se;
fa717060 1235 struct sched_rt_entity rt;
1da177e4 1236
e107be36
AK		 1237#ifdef CONFIG_PREEMPT_NOTIFIERS
1238 /* list of struct preempt_notifier: */
1239 struct hlist_head preempt_notifiers;
1240#endif
1241
18796aa0
AD		 1242 /*
1243 * fpu_counter contains the number of consecutive context switches
1244 * that the FPU is used. If this is over a threshold, the lazy fpu
1245 * saving becomes unlazy to save the trap. This is an unsigned char
1246 * so that after 256 times the counter wraps and the behavior turns
1247 * lazy again; this to deal with bursty apps that only use FPU for
1248 * a short time
1249 */
1250 unsigned char fpu_counter;
6c5c9341 1251#ifdef CONFIG_BLK_DEV_IO_TRACE
2056a782 1252 unsigned int btrace_seq;
6c5c9341 1253#endif
1da177e4 1254
97dc32cd 1255 unsigned int policy;
1da177e4 1256 cpumask_t cpus_allowed;
1da177e4 1257
f41d911f 1258#ifdef CONFIG_TREE_PREEMPT_RCU
e260be67 1259 int rcu_read_lock_nesting;
f41d911f 1260 char rcu_read_unlock_special;
86848966 1261 struct rcu_node *rcu_blocked_node;
f41d911f
PM		 1262 struct list_head rcu_node_entry;
1263#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
e260be67 1264
52f17b6c 1265#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1da177e4
LT		 1266 struct sched_info sched_info;
1267#endif
1268
1269 struct list_head tasks;
917b627d 1270 struct plist_node pushable_tasks;
1da177e4
LT		 1271
1272 struct mm_struct *mm, *active_mm;
1273
1274/* task state */
97dc32cd 1275 int exit_state;
1da177e4
LT		 1276 int exit_code, exit_signal;
1277 int pdeath_signal; /* The signal sent when the parent dies */
1278 /* ??? */
97dc32cd 1279 unsigned int personality;
1da177e4 1280 unsigned did_exec:1;
f9ce1f1c
KT		 1281 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
1282 * execve */
8f0dfc34
AV		 1283 unsigned in_iowait:1;
1284
ca94c442
LP		 1285
1286 /* Revert to default priority/policy when forking */
1287 unsigned sched_reset_on_fork:1;
1288
1da177e4
LT		 1289 pid_t pid;
1290 pid_t tgid;
0a425405 1291
1314562a 1292#ifdef CONFIG_CC_STACKPROTECTOR
0a425405
AV		 1293 /* Canary value for the -fstack-protector gcc feature */
1294 unsigned long stack_canary;
1314562a 1295#endif
e0032087 1296
1da177e4
LT		 1297 /*
1298 * pointers to (original) parent process, youngest child, younger sibling,
1299 * older sibling, respectively. (p->father can be replaced with
f470021a 1300 * p->real_parent->pid)
1da177e4 1301 */
f470021a
RM		 1302 struct task_struct *real_parent; /* real parent process */
1303 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
1da177e4 1304 /*
f470021a 1305 * children/sibling forms the list of my natural children
1da177e4
LT		 1306 */
1307 struct list_head children; /* list of my children */
1308 struct list_head sibling; /* linkage in my parent's children list */
1309 struct task_struct *group_leader; /* threadgroup leader */
1310
f470021a
RM		 1311 /*
1312 * ptraced is the list of tasks this task is using ptrace on.
1313 * This includes both natural children and PTRACE_ATTACH targets.
1314 * p->ptrace_entry is p's link on the p->parent->ptraced list.
1315 */
1316 struct list_head ptraced;
1317 struct list_head ptrace_entry;
1318
ca0002a1
MM		 1319 /*
1320 * This is the tracer handle for the ptrace BTS extension.
1321 * This field actually belongs to the ptracer task.
1322 */
e2b371f0 1323 struct bts_context *bts;
ca0002a1 1324
1da177e4 1325 /* PID/PID hash table linkage. */
92476d7f 1326 struct pid_link pids[PIDTYPE_MAX];
47e65328 1327 struct list_head thread_group;
1da177e4
LT		 1328
1329 struct completion *vfork_done; /* for vfork() */
1330 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
1331 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
1332
c66f08be 1333 cputime_t utime, stime, utimescaled, stimescaled;
9ac52315 1334 cputime_t gtime;
d99ca3b9 1335#ifndef CONFIG_VIRT_CPU_ACCOUNTING
9301899b 1336 cputime_t prev_utime, prev_stime;
d99ca3b9 1337#endif
1da177e4 1338 unsigned long nvcsw, nivcsw; /* context switch counts */
924b42d5
TJ		 1339 struct timespec start_time; /* monotonic time */
1340 struct timespec real_start_time; /* boot based time */
1da177e4
LT		 1341/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1342 unsigned long min_flt, maj_flt;
1343
f06febc9 1344 struct task_cputime cputime_expires;
1da177e4
LT		 1345 struct list_head cpu_timers[3];
1346
1347/* process credentials */
3b11a1de
DH		 1348 const struct cred *real_cred; /* objective and real subjective task
1349 * credentials (COW) */
1350 const struct cred *cred; /* effective (overridable) subjective task
1351 * credentials (COW) */
5e751e99
DH		 1352 struct mutex cred_guard_mutex; /* guard against foreign influences on
1353 * credential calculations
1354 * (notably. ptrace) */
ee18d64c 1355 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
b6dff3ec 1356
36772092
PBG		 1357 char comm[TASK_COMM_LEN]; /* executable name excluding path
1358 - access with [gs]et_task_comm (which lock
1359 it with task_lock())
1360 - initialized normally by flush_old_exec */
1da177e4
LT		 1361/* file system info */
1362 int link_count, total_link_count;
3d5b6fcc 1363#ifdef CONFIG_SYSVIPC
1da177e4
LT		 1364/* ipc stuff */
1365 struct sysv_sem sysvsem;
3d5b6fcc 1366#endif
e162b39a 1367#ifdef CONFIG_DETECT_HUNG_TASK
82a1fcb9 1368/* hung task detection */
82a1fcb9
IM		 1369 unsigned long last_switch_count;
1370#endif
1da177e4
LT		 1371/* CPU-specific state of this task */
1372 struct thread_struct thread;
1373/* filesystem information */
1374 struct fs_struct *fs;
1375/* open file information */
1376 struct files_struct *files;
1651e14e 1377/* namespaces */
ab516013 1378 struct nsproxy *nsproxy;
1da177e4
LT		 1379/* signal handlers */
1380 struct signal_struct *signal;
1381 struct sighand_struct *sighand;
1382
1383 sigset_t blocked, real_blocked;
f3de272b 1384 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
1da177e4
LT		 1385 struct sigpending pending;
1386
1387 unsigned long sas_ss_sp;
1388 size_t sas_ss_size;
1389 int (*notifier)(void *priv);
1390 void *notifier_data;
1391 sigset_t *notifier_mask;
1da177e4 1392 struct audit_context *audit_context;
bfef93a5
AV		 1393#ifdef CONFIG_AUDITSYSCALL
1394 uid_t loginuid;
4746ec5b 1395 unsigned int sessionid;
bfef93a5 1396#endif
1da177e4
LT		 1397 seccomp_t seccomp;
1398
1399/* Thread group tracking */
1400 u32 parent_exec_id;
1401 u32 self_exec_id;
58568d2a
MX		 1402/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
1403 * mempolicy */
1da177e4 1404 spinlock_t alloc_lock;
1da177e4 1405
3aa551c9
TG		 1406#ifdef CONFIG_GENERIC_HARDIRQS
1407 /* IRQ handler threads */
1408 struct irqaction *irqaction;
1409#endif
1410
b29739f9 1411 /* Protection of the PI data structures: */
1d615482 1412 raw_spinlock_t pi_lock;
b29739f9 1413
23f78d4a
IM		 1414#ifdef CONFIG_RT_MUTEXES
1415 /* PI waiters blocked on a rt_mutex held by this task */
1416 struct plist_head pi_waiters;
1417 /* Deadlock detection and priority inheritance handling */
1418 struct rt_mutex_waiter *pi_blocked_on;
23f78d4a
IM		 1419#endif
1420
408894ee
IM		 1421#ifdef CONFIG_DEBUG_MUTEXES
1422 /* mutex deadlock detection */
1423 struct mutex_waiter *blocked_on;
1424#endif
de30a2b3
IM		 1425#ifdef CONFIG_TRACE_IRQFLAGS
1426 unsigned int irq_events;
de30a2b3 1427 unsigned long hardirq_enable_ip;
de30a2b3 1428 unsigned long hardirq_disable_ip;
fa1452e8 1429 unsigned int hardirq_enable_event;
de30a2b3 1430 unsigned int hardirq_disable_event;
fa1452e8
HS		 1431 int hardirqs_enabled;
1432 int hardirq_context;
de30a2b3 1433 unsigned long softirq_disable_ip;
de30a2b3 1434 unsigned long softirq_enable_ip;
fa1452e8 1435 unsigned int softirq_disable_event;
de30a2b3 1436 unsigned int softirq_enable_event;
fa1452e8 1437 int softirqs_enabled;
de30a2b3
IM		 1438 int softirq_context;
1439#endif
fbb9ce95 1440#ifdef CONFIG_LOCKDEP
bdb9441e 1441# define MAX_LOCK_DEPTH 48UL
fbb9ce95
IM		 1442 u64 curr_chain_key;
1443 int lockdep_depth;
fbb9ce95 1444 unsigned int lockdep_recursion;
c7aceaba 1445 struct held_lock held_locks[MAX_LOCK_DEPTH];
cf40bd16 1446 gfp_t lockdep_reclaim_gfp;
fbb9ce95 1447#endif
408894ee 1448
e4c570c4 1449#ifdef CONFIG_FS_JOURNAL_INFO
1da177e4
LT		 1450/* journalling filesystem info */
1451 void *journal_info;
e4c570c4 1452#endif
1da177e4 1453
d89d8796
NB		 1454/* stacked block device info */
1455 struct bio *bio_list, **bio_tail;
1456
1da177e4
LT		 1457/* VM state */
1458 struct reclaim_state *reclaim_state;
1459
1da177e4
LT		 1460 struct backing_dev_info *backing_dev_info;
1461
1462 struct io_context *io_context;
1463
1464 unsigned long ptrace_message;
1465 siginfo_t *last_siginfo; /* For ptrace use. */
7c3ab738 1466 struct task_io_accounting ioac;
8f0ab514 1467#if defined(CONFIG_TASK_XACCT)
1da177e4
LT		 1468 u64 acct_rss_mem1; /* accumulated rss usage */
1469 u64 acct_vm_mem1; /* accumulated virtual memory usage */
49b5cf34 1470 cputime_t acct_timexpd; /* stime + utime since last update */
1da177e4
LT		 1471#endif
1472#ifdef CONFIG_CPUSETS
58568d2a 1473 nodemask_t mems_allowed; /* Protected by alloc_lock */
825a46af 1474 int cpuset_mem_spread_rotor;
1da177e4 1475#endif
ddbcc7e8 1476#ifdef CONFIG_CGROUPS
817929ec
PM		 1477 /* Control Group info protected by css_set_lock */
1478 struct css_set *cgroups;
1479 /* cg_list protected by css_set_lock and tsk->alloc_lock */
1480 struct list_head cg_list;
ddbcc7e8 1481#endif
42b2dd0a 1482#ifdef CONFIG_FUTEX
0771dfef 1483 struct robust_list_head __user *robust_list;
34f192c6
IM		 1484#ifdef CONFIG_COMPAT
1485 struct compat_robust_list_head __user *compat_robust_list;
1486#endif
c87e2837
IM		 1487 struct list_head pi_state_list;
1488 struct futex_pi_state *pi_state_cache;
c7aceaba 1489#endif
cdd6c482
IM		 1490#ifdef CONFIG_PERF_EVENTS
1491 struct perf_event_context *perf_event_ctxp;
1492 struct mutex perf_event_mutex;
1493 struct list_head perf_event_list;
a63eaf34 1494#endif
c7aceaba 1495#ifdef CONFIG_NUMA
58568d2a 1496 struct mempolicy *mempolicy; /* Protected by alloc_lock */
c7aceaba 1497 short il_next;
42b2dd0a 1498#endif
22e2c507 1499 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 1500 struct rcu_head rcu;
b92ce558
JA		 1501
1502 /*
1503 * cache last used pipe for splice
1504 */
1505 struct pipe_inode_info *splice_pipe;
ca74e92b
SN		 1506#ifdef CONFIG_TASK_DELAY_ACCT
1507 struct task_delay_info *delays;
f4f154fd
AM		 1508#endif
1509#ifdef CONFIG_FAULT_INJECTION
1510 int make_it_fail;
ca74e92b 1511#endif
3e26c149 1512 struct prop_local_single dirties;
9745512c
AV		 1513#ifdef CONFIG_LATENCYTOP
1514 int latency_record_count;
1515 struct latency_record latency_record[LT_SAVECOUNT];
1516#endif
6976675d
AV		 1517 /*
1518 * time slack values; these are used to round up poll() and
1519 * select() etc timeout values. These are in nanoseconds.
1520 */
1521 unsigned long timer_slack_ns;
1522 unsigned long default_timer_slack_ns;
f8d570a4
DM		 1523
1524 struct list_head *scm_work_list;
fb52607a 1525#ifdef CONFIG_FUNCTION_GRAPH_TRACER
f201ae23
FW		 1526 /* Index of current stored adress in ret_stack */
1527 int curr_ret_stack;
1528 /* Stack of return addresses for return function tracing */
1529 struct ftrace_ret_stack *ret_stack;
8aef2d28
SR		 1530 /* time stamp for last schedule */
1531 unsigned long long ftrace_timestamp;
f201ae23
FW		 1532 /*
1533 * Number of functions that haven't been traced
1534 * because of depth overrun.
1535 */
1536 atomic_t trace_overrun;
380c4b14
FW		 1537 /* Pause for the tracing */
1538 atomic_t tracing_graph_pause;
f201ae23 1539#endif
ea4e2bc4
SR		 1540#ifdef CONFIG_TRACING
1541 /* state flags for use by tracers */
1542 unsigned long trace;
261842b7
SR		 1543 /* bitmask of trace recursion */
1544 unsigned long trace_recursion;
1545#endif /* CONFIG_TRACING */
d899bf7b 1546 unsigned long stack_start;
1da177e4
LT		 1547};
1548
76e6eee0
RR		 1549/* Future-safe accessor for struct task_struct's cpus_allowed. */
1550#define tsk_cpumask(tsk) (&(tsk)->cpus_allowed)
1551
e05606d3
IM		 1552/*
1553 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1554 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1555 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1556 * values are inverted: lower p->prio value means higher priority.
1557 *
1558 * The MAX_USER_RT_PRIO value allows the actual maximum
1559 * RT priority to be separate from the value exported to
1560 * user-space. This allows kernel threads to set their
1561 * priority to a value higher than any user task. Note:
1562 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1563 */
1564
1565#define MAX_USER_RT_PRIO 100
1566#define MAX_RT_PRIO MAX_USER_RT_PRIO
1567
1568#define MAX_PRIO (MAX_RT_PRIO + 40)
1569#define DEFAULT_PRIO (MAX_RT_PRIO + 20)
1570
1571static inline int rt_prio(int prio)
1572{
1573 if (unlikely(prio < MAX_RT_PRIO))
1574 return 1;
1575 return 0;
1576}
1577
e868171a 1578static inline int rt_task(struct task_struct *p)
e05606d3
IM		 1579{
1580 return rt_prio(p->prio);
1581}
1582
e868171a 1583static inline struct pid *task_pid(struct task_struct *task)
22c935f4
EB		 1584{
1585 return task->pids[PIDTYPE_PID].pid;
1586}
1587
e868171a 1588static inline struct pid *task_tgid(struct task_struct *task)
22c935f4
EB		 1589{
1590 return task->group_leader->pids[PIDTYPE_PID].pid;
1591}
1592
6dda81f4
ON		 1593/*
1594 * Without tasklist or rcu lock it is not safe to dereference
1595 * the result of task_pgrp/task_session even if task == current,
1596 * we can race with another thread doing sys_setsid/sys_setpgid.
1597 */
e868171a 1598static inline struct pid *task_pgrp(struct task_struct *task)
22c935f4
EB		 1599{
1600 return task->group_leader->pids[PIDTYPE_PGID].pid;
1601}
1602
e868171a 1603static inline struct pid *task_session(struct task_struct *task)
22c935f4
EB		 1604{
1605 return task->group_leader->pids[PIDTYPE_SID].pid;
1606}
1607
7af57294
PE		 1608struct pid_namespace;
1609
1610/*
1611 * the helpers to get the task's different pids as they are seen
1612 * from various namespaces
1613 *
1614 * task_xid_nr() : global id, i.e. the id seen from the init namespace;
44c4e1b2
EB		 1615 * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
1616 * current.
7af57294
PE		 1617 * task_xid_nr_ns() : id seen from the ns specified;
1618 *
1619 * set_task_vxid() : assigns a virtual id to a task;
1620 *
7af57294
PE		 1621 * see also pid_nr() etc in include/linux/pid.h
1622 */
52ee2dfd
ON		 1623pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
1624 struct pid_namespace *ns);
7af57294 1625
e868171a 1626static inline pid_t task_pid_nr(struct task_struct *tsk)
7af57294
PE		 1627{
1628 return tsk->pid;
1629}
1630
52ee2dfd
ON		 1631static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
1632 struct pid_namespace *ns)
1633{
1634 return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
1635}
7af57294
PE		 1636
1637static inline pid_t task_pid_vnr(struct task_struct *tsk)
1638{
52ee2dfd 1639 return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
7af57294
PE		 1640}
1641
1642
e868171a 1643static inline pid_t task_tgid_nr(struct task_struct *tsk)
7af57294
PE		 1644{
1645 return tsk->tgid;
1646}
1647
2f2a3a46 1648pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
7af57294
PE		 1649
1650static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1651{
1652 return pid_vnr(task_tgid(tsk));
1653}
1654
1655
52ee2dfd
ON		 1656static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
1657 struct pid_namespace *ns)
7af57294 1658{
52ee2dfd 1659 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
7af57294
PE		 1660}
1661
7af57294
PE		 1662static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1663{
52ee2dfd 1664 return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
7af57294
PE		 1665}
1666
1667
52ee2dfd
ON		 1668static inline pid_t task_session_nr_ns(struct task_struct *tsk,
1669 struct pid_namespace *ns)
7af57294 1670{
52ee2dfd 1671 return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
7af57294
PE		 1672}
1673
7af57294
PE		 1674static inline pid_t task_session_vnr(struct task_struct *tsk)
1675{
52ee2dfd 1676 return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
7af57294
PE		 1677}
1678
1b0f7ffd
ON		 1679/* obsolete, do not use */
1680static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1681{
1682 return task_pgrp_nr_ns(tsk, &init_pid_ns);
1683}
7af57294 1684
1da177e4
LT		 1685/**
1686 * pid_alive - check that a task structure is not stale
1687 * @p: Task structure to be checked.
1688 *
1689 * Test if a process is not yet dead (at most zombie state)
1690 * If pid_alive fails, then pointers within the task structure
1691 * can be stale and must not be dereferenced.
1692 */
e868171a 1693static inline int pid_alive(struct task_struct *p)
1da177e4 1694{
92476d7f 1695 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT		 1696}
1697
f400e198 1698/**
b460cbc5 1699 * is_global_init - check if a task structure is init
3260259f
HK		 1700 * @tsk: Task structure to be checked.
1701 *
1702 * Check if a task structure is the first user space task the kernel created.
b460cbc5 1703 */
e868171a 1704static inline int is_global_init(struct task_struct *tsk)
b461cc03
PE		 1705{
1706 return tsk->pid == 1;
1707}
b460cbc5
SH		 1708
1709/*
1710 * is_container_init:
1711 * check whether in the task is init in its own pid namespace.
f400e198 1712 */
b461cc03 1713extern int is_container_init(struct task_struct *tsk);
f400e198 1714
9ec52099
CLG		 1715extern struct pid *cad_pid;
1716
1da177e4 1717extern void free_task(struct task_struct *tsk);
1da177e4 1718#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 1719
158d9ebd 1720extern void __put_task_struct(struct task_struct *t);
e56d0903
IM		 1721
1722static inline void put_task_struct(struct task_struct *t)
1723{
1724 if (atomic_dec_and_test(&t->usage))
8c7904a0 1725 __put_task_struct(t);
e56d0903 1726}
1da177e4 1727
d180c5bc 1728extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
0cf55e1e 1729extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
49048622 1730
1da177e4
LT		 1731/*
1732 * Per process flags
1733 */
1734#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1735 /* Not implemented yet, only for 486*/
1736#define PF_STARTING 0x00000002 /* being created */
1737#define PF_EXITING 0x00000004 /* getting shut down */
778e9a9c 1738#define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
94886b84 1739#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
1da177e4 1740#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
4db96cf0 1741#define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
1da177e4
LT		 1742#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1743#define PF_DUMPCORE 0x00000200 /* dumped core */
1744#define PF_SIGNALED 0x00000400 /* killed by a signal */
1745#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1746#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1747#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
6301cb95 1748#define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
1da177e4
LT		 1749#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1750#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1751#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1752#define PF_KSWAPD 0x00040000 /* I am kswapd */
35451bee 1753#define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
1da177e4 1754#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
246bb0b1 1755#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
b31dc66a
JA		 1756#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1757#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1758#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1759#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
9985b0ba 1760#define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
4db96cf0 1761#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
c61afb18 1762#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
61a87122 1763#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
ba96a0c8 1764#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
ebb12db5 1765#define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
1da177e4
LT		 1766
1767/*
1768 * Only the _current_ task can read/write to tsk->flags, but other
1769 * tasks can access tsk->flags in readonly mode for example
1770 * with tsk_used_math (like during threaded core dumping).
1771 * There is however an exception to this rule during ptrace
1772 * or during fork: the ptracer task is allowed to write to the
1773 * child->flags of its traced child (same goes for fork, the parent
1774 * can write to the child->flags), because we're guaranteed the
1775 * child is not running and in turn not changing child->flags
1776 * at the same time the parent does it.
1777 */
1778#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1779#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1780#define clear_used_math() clear_stopped_child_used_math(current)
1781#define set_used_math() set_stopped_child_used_math(current)
1782#define conditional_stopped_child_used_math(condition, child) \
1783 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1784#define conditional_used_math(condition) \
1785 conditional_stopped_child_used_math(condition, current)
1786#define copy_to_stopped_child_used_math(child) \
1787 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1788/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1789#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1790#define used_math() tsk_used_math(current)
1791
f41d911f
PM		 1792#ifdef CONFIG_TREE_PREEMPT_RCU
1793
1794#define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
1795#define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
f41d911f
PM		 1796
1797static inline void rcu_copy_process(struct task_struct *p)
1798{
1799 p->rcu_read_lock_nesting = 0;
1800 p->rcu_read_unlock_special = 0;
dd5d19ba 1801 p->rcu_blocked_node = NULL;
f41d911f
PM		 1802 INIT_LIST_HEAD(&p->rcu_node_entry);
1803}
1804
f41d911f
PM		 1805#else
1806
1807static inline void rcu_copy_process(struct task_struct *p)
1808{
1809}
1810
1811#endif
1812
1da177e4 1813#ifdef CONFIG_SMP
cd8ba7cd 1814extern int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1815 const struct cpumask *new_mask);
1da177e4 1816#else
cd8ba7cd 1817static inline int set_cpus_allowed_ptr(struct task_struct *p,
96f874e2 1818 const struct cpumask *new_mask)
1da177e4 1819{
96f874e2 1820 if (!cpumask_test_cpu(0, new_mask))
1da177e4
LT		 1821 return -EINVAL;
1822 return 0;
1823}
1824#endif
e0ad9556
RR		 1825
1826#ifndef CONFIG_CPUMASK_OFFSTACK
cd8ba7cd
MT		 1827static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1828{
1829 return set_cpus_allowed_ptr(p, &new_mask);
1830}
e0ad9556 1831#endif
1da177e4 1832
b342501c
IM		 1833/*
1834 * Architectures can set this to 1 if they have specified
1835 * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
1836 * but then during bootup it turns out that sched_clock()
1837 * is reliable after all:
1838 */
1839#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
1840extern int sched_clock_stable;
1841#endif
1842
1bbfa6f2
MF		 1843/* ftrace calls sched_clock() directly */
1844extern unsigned long long notrace sched_clock(void);
e436d800 1845
c1955a3d
PZ		 1846extern void sched_clock_init(void);
1847extern u64 sched_clock_cpu(int cpu);
3e51f33f 1848
c1955a3d 1849#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
3e51f33f
PZ		 1850static inline void sched_clock_tick(void)
1851{
1852}
1853
1854static inline void sched_clock_idle_sleep_event(void)
1855{
1856}
1857
1858static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
1859{
1860}
1861#else
3e51f33f
PZ		 1862extern void sched_clock_tick(void);
1863extern void sched_clock_idle_sleep_event(void);
1864extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1865#endif
1866
e436d800
IM		 1867/*
1868 * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1869 * clock constructed from sched_clock():
1870 */
1871extern unsigned long long cpu_clock(int cpu);
1872
36c8b586 1873extern unsigned long long
41b86e9c 1874task_sched_runtime(struct task_struct *task);
f06febc9 1875extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
1da177e4
LT		 1876
1877/* sched_exec is called by processes performing an exec */
1878#ifdef CONFIG_SMP
1879extern void sched_exec(void);
1880#else
1881#define sched_exec() {}
1882#endif
1883
2aa44d05
IM		 1884extern void sched_clock_idle_sleep_event(void);
1885extern void sched_clock_idle_wakeup_event(u64 delta_ns);
bb29ab26 1886
1da177e4
LT		 1887#ifdef CONFIG_HOTPLUG_CPU
1888extern void idle_task_exit(void);
1889#else
1890static inline void idle_task_exit(void) {}
1891#endif
1892
1893extern void sched_idle_next(void);
b29739f9 1894
06d8308c
TG		 1895#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1896extern void wake_up_idle_cpu(int cpu);
1897#else
1898static inline void wake_up_idle_cpu(int cpu) { }
1899#endif
1900
21805085 1901extern unsigned int sysctl_sched_latency;
b2be5e96 1902extern unsigned int sysctl_sched_min_granularity;
bf0f6f24 1903extern unsigned int sysctl_sched_wakeup_granularity;
47fea2ad
JSR		 1904extern unsigned int sysctl_sched_shares_ratelimit;
1905extern unsigned int sysctl_sched_shares_thresh;
bf0f6f24 1906extern unsigned int sysctl_sched_child_runs_first;
1983a922
CE		 1907
1908enum sched_tunable_scaling {
1909 SCHED_TUNABLESCALING_NONE,
1910 SCHED_TUNABLESCALING_LOG,
1911 SCHED_TUNABLESCALING_LINEAR,
1912 SCHED_TUNABLESCALING_END,
1913};
1914extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
1915
2bba22c5 1916#ifdef CONFIG_SCHED_DEBUG
da84d961 1917extern unsigned int sysctl_sched_migration_cost;
b82d9fdd 1918extern unsigned int sysctl_sched_nr_migrate;
e9e9250b 1919extern unsigned int sysctl_sched_time_avg;
cd1bb94b 1920extern unsigned int sysctl_timer_migration;
b2be5e96 1921
1983a922 1922int sched_proc_update_handler(struct ctl_table *table, int write,
8d65af78 1923 void __user *buffer, size_t *length,
b2be5e96 1924 loff_t *ppos);
2bd8e6d4 1925#endif
eea08f32
AB		 1926#ifdef CONFIG_SCHED_DEBUG
1927static inline unsigned int get_sysctl_timer_migration(void)
1928{
1929 return sysctl_timer_migration;
1930}
1931#else
1932static inline unsigned int get_sysctl_timer_migration(void)
1933{
1934 return 1;
1935}
1936#endif
9f0c1e56
PZ		 1937extern unsigned int sysctl_sched_rt_period;
1938extern int sysctl_sched_rt_runtime;
2bd8e6d4 1939
d0b27fa7 1940int sched_rt_handler(struct ctl_table *table, int write,
8d65af78 1941 void __user *buffer, size_t *lenp,
d0b27fa7
PZ		 1942 loff_t *ppos);
1943
2bd8e6d4 1944extern unsigned int sysctl_sched_compat_yield;
bf0f6f24 1945
b29739f9 1946#ifdef CONFIG_RT_MUTEXES
36c8b586
IM		 1947extern int rt_mutex_getprio(struct task_struct *p);
1948extern void rt_mutex_setprio(struct task_struct *p, int prio);
1949extern void rt_mutex_adjust_pi(struct task_struct *p);
b29739f9 1950#else
e868171a 1951static inline int rt_mutex_getprio(struct task_struct *p)
b29739f9
IM		 1952{
1953 return p->normal_prio;
1954}
95e02ca9 1955# define rt_mutex_adjust_pi(p) do { } while (0)
b29739f9
IM		 1956#endif
1957
36c8b586
IM		 1958extern void set_user_nice(struct task_struct *p, long nice);
1959extern int task_prio(const struct task_struct *p);
1960extern int task_nice(const struct task_struct *p);
1961extern int can_nice(const struct task_struct *p, const int nice);
1962extern int task_curr(const struct task_struct *p);
1da177e4
LT		 1963extern int idle_cpu(int cpu);
1964extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
961ccddd
RR		 1965extern int sched_setscheduler_nocheck(struct task_struct *, int,
1966 struct sched_param *);
36c8b586
IM		 1967extern struct task_struct *idle_task(int cpu);
1968extern struct task_struct *curr_task(int cpu);
1969extern void set_curr_task(int cpu, struct task_struct *p);
1da177e4
LT		 1970
1971void yield(void);
1972
1973/*
1974 * The default (Linux) execution domain.
1975 */
1976extern struct exec_domain default_exec_domain;
1977
1978union thread_union {
1979 struct thread_info thread_info;
1980 unsigned long stack[THREAD_SIZE/sizeof(long)];
1981};
1982
1983#ifndef __HAVE_ARCH_KSTACK_END
1984static inline int kstack_end(void *addr)
1985{
1986 /* Reliable end of stack detection:
1987 * Some APM bios versions misalign the stack
1988 */
1989 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1990}
1991#endif
1992
1993extern union thread_union init_thread_union;
1994extern struct task_struct init_task;
1995
1996extern struct mm_struct init_mm;
1997
198fe21b
PE		 1998extern struct pid_namespace init_pid_ns;
1999
2000/*
2001 * find a task by one of its numerical ids
2002 *
198fe21b
PE		 2003 * find_task_by_pid_ns():
2004 * finds a task by its pid in the specified namespace
228ebcbe
PE		 2005 * find_task_by_vpid():
2006 * finds a task by its virtual pid
198fe21b 2007 *
e49859e7 2008 * see also find_vpid() etc in include/linux/pid.h
198fe21b
PE		 2009 */
2010
228ebcbe
PE		 2011extern struct task_struct *find_task_by_vpid(pid_t nr);
2012extern struct task_struct *find_task_by_pid_ns(pid_t nr,
2013 struct pid_namespace *ns);
198fe21b 2014
8520d7c7 2015extern void __set_special_pids(struct pid *pid);
1da177e4
LT		 2016
2017/* per-UID process charging. */
acce292c 2018extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1da177e4
LT		 2019static inline struct user_struct *get_uid(struct user_struct *u)
2020{
2021 atomic_inc(&u->__count);
2022 return u;
2023}
2024extern void free_uid(struct user_struct *);
28f300d2 2025extern void release_uids(struct user_namespace *ns);
1da177e4
LT		 2026
2027#include <asm/current.h>
2028
3171a030 2029extern void do_timer(unsigned long ticks);
1da177e4 2030
b3c97528
HH		 2031extern int wake_up_state(struct task_struct *tsk, unsigned int state);
2032extern int wake_up_process(struct task_struct *tsk);
2033extern void wake_up_new_task(struct task_struct *tsk,
2034 unsigned long clone_flags);
1da177e4
LT		 2035#ifdef CONFIG_SMP
2036 extern void kick_process(struct task_struct *tsk);
2037#else
2038 static inline void kick_process(struct task_struct *tsk) { }
2039#endif
ad46c2c4
IM		 2040extern void sched_fork(struct task_struct *p, int clone_flags);
2041extern void sched_dead(struct task_struct *p);
1da177e4 2042
1da177e4
LT		 2043extern void proc_caches_init(void);
2044extern void flush_signals(struct task_struct *);
3bcac026 2045extern void __flush_signals(struct task_struct *);
10ab825b 2046extern void ignore_signals(struct task_struct *);
1da177e4
LT		 2047extern void flush_signal_handlers(struct task_struct *, int force_default);
2048extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
2049
2050static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
2051{
2052 unsigned long flags;
2053 int ret;
2054
2055 spin_lock_irqsave(&tsk->sighand->siglock, flags);
2056 ret = dequeue_signal(tsk, mask, info);
2057 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
2058
2059 return ret;
2060}
2061
2062extern void block_all_signals(int (*notifier)(void *priv), void *priv,
2063 sigset_t *mask);
2064extern void unblock_all_signals(void);
2065extern void release_task(struct task_struct * p);
2066extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1da177e4
LT		 2067extern int force_sigsegv(int, struct task_struct *);
2068extern int force_sig_info(int, struct siginfo *, struct task_struct *);
c4b92fc1 2069extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
c4b92fc1 2070extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
2425c08b 2071extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
c4b92fc1
EB		 2072extern int kill_pgrp(struct pid *pid, int sig, int priv);
2073extern int kill_pid(struct pid *pid, int sig, int priv);
c3de4b38 2074extern int kill_proc_info(int, struct siginfo *, pid_t);
2b2a1ff6 2075extern int do_notify_parent(struct task_struct *, int);
a7f0765e 2076extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
1da177e4
LT		 2077extern void force_sig(int, struct task_struct *);
2078extern void force_sig_specific(int, struct task_struct *);
2079extern int send_sig(int, struct task_struct *, int);
2080extern void zap_other_threads(struct task_struct *p);
1da177e4
LT		 2081extern struct sigqueue *sigqueue_alloc(void);
2082extern void sigqueue_free(struct sigqueue *);
ac5c2153 2083extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
9ac95f2f 2084extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT		 2085extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
2086
9ec52099
CLG		 2087static inline int kill_cad_pid(int sig, int priv)
2088{
2089 return kill_pid(cad_pid, sig, priv);
2090}
2091
1da177e4
LT		 2092/* These can be the second arg to send_sig_info/send_group_sig_info. */
2093#define SEND_SIG_NOINFO ((struct siginfo *) 0)
2094#define SEND_SIG_PRIV ((struct siginfo *) 1)
2095#define SEND_SIG_FORCED ((struct siginfo *) 2)
2096
621d3121
ON		 2097static inline int is_si_special(const struct siginfo *info)
2098{
2099 return info <= SEND_SIG_FORCED;
2100}
2101
2a855dd0
SAS		 2102/*
2103 * True if we are on the alternate signal stack.
2104 */
1da177e4
LT		 2105static inline int on_sig_stack(unsigned long sp)
2106{
2a855dd0
SAS		 2107#ifdef CONFIG_STACK_GROWSUP
2108 return sp >= current->sas_ss_sp &&
2109 sp - current->sas_ss_sp < current->sas_ss_size;
2110#else
2111 return sp > current->sas_ss_sp &&
2112 sp - current->sas_ss_sp <= current->sas_ss_size;
2113#endif
1da177e4
LT		 2114}
2115
2116static inline int sas_ss_flags(unsigned long sp)
2117{
2118 return (current->sas_ss_size == 0 ? SS_DISABLE
2119 : on_sig_stack(sp) ? SS_ONSTACK : 0);
2120}
2121
1da177e4
LT		 2122/*
2123 * Routines for handling mm_structs
2124 */
2125extern struct mm_struct * mm_alloc(void);
2126
2127/* mmdrop drops the mm and the page tables */
b3c97528 2128extern void __mmdrop(struct mm_struct *);
1da177e4
LT		 2129static inline void mmdrop(struct mm_struct * mm)
2130{
6fb43d7b 2131 if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1da177e4
LT		 2132 __mmdrop(mm);
2133}
2134
2135/* mmput gets rid of the mappings and all user-space */
2136extern void mmput(struct mm_struct *);
2137/* Grab a reference to a task's mm, if it is not already going away */
2138extern struct mm_struct *get_task_mm(struct task_struct *task);
2139/* Remove the current tasks stale references to the old mm_struct */
2140extern void mm_release(struct task_struct *, struct mm_struct *);
402b0862
CO		 2141/* Allocate a new mm structure and copy contents from tsk->mm */
2142extern struct mm_struct *dup_mm(struct task_struct *tsk);
1da177e4 2143
6f2c55b8
AD		 2144extern int copy_thread(unsigned long, unsigned long, unsigned long,
2145 struct task_struct *, struct pt_regs *);
1da177e4
LT		 2146extern void flush_thread(void);
2147extern void exit_thread(void);
2148
1da177e4 2149extern void exit_files(struct task_struct *);
6b3934ef 2150extern void __cleanup_signal(struct signal_struct *);
a7e5328a 2151extern void __cleanup_sighand(struct sighand_struct *);
cbaffba1 2152
1da177e4 2153extern void exit_itimers(struct signal_struct *);
cbaffba1 2154extern void flush_itimer_signals(void);
1da177e4
LT		 2155
2156extern NORET_TYPE void do_group_exit(int);
2157
1da177e4
LT		 2158extern void daemonize(const char *, ...);
2159extern int allow_signal(int);
2160extern int disallow_signal(int);
1da177e4
LT		 2161
2162extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
2163extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
36c8b586 2164struct task_struct *fork_idle(int);
1da177e4
LT		 2165
2166extern void set_task_comm(struct task_struct *tsk, char *from);
59714d65 2167extern char *get_task_comm(char *to, struct task_struct *tsk);
1da177e4
LT		 2168
2169#ifdef CONFIG_SMP
a26b89f0 2170extern void wait_task_context_switch(struct task_struct *p);
85ba2d86 2171extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
1da177e4 2172#else
a26b89f0 2173static inline void wait_task_context_switch(struct task_struct *p) {}
85ba2d86
RM		 2174static inline unsigned long wait_task_inactive(struct task_struct *p,
2175 long match_state)
2176{
2177 return 1;
2178}
1da177e4
LT		 2179#endif
2180
05725f7e
JP		 2181#define next_task(p) \
2182 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
1da177e4
LT		 2183
2184#define for_each_process(p) \
2185 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
2186
5bb459bb 2187extern bool current_is_single_threaded(void);
d84f4f99 2188
1da177e4
LT		 2189/*
2190 * Careful: do_each_thread/while_each_thread is a double loop so
2191 * 'break' will not work as expected - use goto instead.
2192 */
2193#define do_each_thread(g, t) \
2194 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
2195
2196#define while_each_thread(g, t) \
2197 while ((t = next_thread(t)) != g)
2198
de12a787
EB		 2199/* de_thread depends on thread_group_leader not being a pid based check */
2200#define thread_group_leader(p) (p == p->group_leader)
1da177e4 2201
0804ef4b
EB		 2202/* Do to the insanities of de_thread it is possible for a process
2203 * to have the pid of the thread group leader without actually being
2204 * the thread group leader. For iteration through the pids in proc
2205 * all we care about is that we have a task with the appropriate
2206 * pid, we don't actually care if we have the right task.
2207 */
e868171a 2208static inline int has_group_leader_pid(struct task_struct *p)
0804ef4b
EB		 2209{
2210 return p->pid == p->tgid;
2211}
2212
bac0abd6
PE		 2213static inline
2214int same_thread_group(struct task_struct *p1, struct task_struct *p2)
2215{
2216 return p1->tgid == p2->tgid;
2217}
2218
36c8b586 2219static inline struct task_struct *next_thread(const struct task_struct *p)
47e65328 2220{
05725f7e
JP		 2221 return list_entry_rcu(p->thread_group.next,
2222 struct task_struct, thread_group);
47e65328
ON		 2223}
2224
e868171a 2225static inline int thread_group_empty(struct task_struct *p)
1da177e4 2226{
47e65328 2227 return list_empty(&p->thread_group);
1da177e4
LT		 2228}
2229
2230#define delay_group_leader(p) \
2231 (thread_group_leader(p) && !thread_group_empty(p))
2232
39c626ae
ON		 2233static inline int task_detached(struct task_struct *p)
2234{
2235 return p->exit_signal == -1;
2236}
2237
1da177e4 2238/*
260ea101 2239 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 2240 * subscriptions and synchronises with wait4(). Also used in procfs. Also
ddbcc7e8
PM		 2241 * pins the final release of task.io_context. Also protects ->cpuset and
2242 * ->cgroup.subsys[].
1da177e4
LT		 2243 *
2244 * Nests both inside and outside of read_lock(&tasklist_lock).
2245 * It must not be nested with write_lock_irq(&tasklist_lock),
2246 * neither inside nor outside.
2247 */
2248static inline void task_lock(struct task_struct *p)
2249{
2250 spin_lock(&p->alloc_lock);
2251}
2252
2253static inline void task_unlock(struct task_struct *p)
2254{
2255 spin_unlock(&p->alloc_lock);
2256}
2257
f63ee72e
ON		 2258extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
2259 unsigned long *flags);
2260
2261static inline void unlock_task_sighand(struct task_struct *tsk,
2262 unsigned long *flags)
2263{
2264 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
2265}
2266
f037360f
AV		 2267#ifndef __HAVE_THREAD_FUNCTIONS
2268
f7e4217b
RZ		 2269#define task_thread_info(task) ((struct thread_info *)(task)->stack)
2270#define task_stack_page(task) ((task)->stack)
a1261f54 2271
10ebffde
AV		 2272static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
2273{
2274 *task_thread_info(p) = *task_thread_info(org);
2275 task_thread_info(p)->task = p;
2276}
2277
2278static inline unsigned long *end_of_stack(struct task_struct *p)
2279{
f7e4217b 2280 return (unsigned long *)(task_thread_info(p) + 1);
10ebffde
AV		 2281}
2282
f037360f
AV		 2283#endif
2284
8b05c7e6
FT		 2285static inline int object_is_on_stack(void *obj)
2286{
2287 void *stack = task_stack_page(current);
2288
2289 return (obj >= stack) && (obj < (stack + THREAD_SIZE));
2290}
2291
8c9843e5
BH		 2292extern void thread_info_cache_init(void);
2293
7c9f8861
ES		 2294#ifdef CONFIG_DEBUG_STACK_USAGE
2295static inline unsigned long stack_not_used(struct task_struct *p)
2296{
2297 unsigned long *n = end_of_stack(p);
2298
2299 do { /* Skip over canary */
2300 n++;
2301 } while (!*n);
2302
2303 return (unsigned long)n - (unsigned long)end_of_stack(p);
2304}
2305#endif
2306
1da177e4
LT		 2307/* set thread flags in other task's structures
2308 * - see asm/thread_info.h for TIF_xxxx flags available
2309 */
2310static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
2311{
a1261f54 2312 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT		 2313}
2314
2315static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2316{
a1261f54 2317 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT		 2318}
2319
2320static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
2321{
a1261f54 2322 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT		 2323}
2324
2325static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
2326{
a1261f54 2327 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT		 2328}
2329
2330static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
2331{
a1261f54 2332 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT		 2333}
2334
2335static inline void set_tsk_need_resched(struct task_struct *tsk)
2336{
2337 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2338}
2339
2340static inline void clear_tsk_need_resched(struct task_struct *tsk)
2341{
2342 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
2343}
2344
8ae121ac
GH		 2345static inline int test_tsk_need_resched(struct task_struct *tsk)
2346{
2347 return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
2348}
2349
690cc3ff
EB		 2350static inline int restart_syscall(void)
2351{
2352 set_tsk_thread_flag(current, TIF_SIGPENDING);
2353 return -ERESTARTNOINTR;
2354}
2355
1da177e4
LT		 2356static inline int signal_pending(struct task_struct *p)
2357{
2358 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
2359}
f776d12d 2360
d9588725
RM		 2361static inline int __fatal_signal_pending(struct task_struct *p)
2362{
2363 return unlikely(sigismember(&p->pending.signal, SIGKILL));
2364}
f776d12d
MW		 2365
2366static inline int fatal_signal_pending(struct task_struct *p)
2367{
2368 return signal_pending(p) && __fatal_signal_pending(p);
2369}
2370
16882c1e
ON		 2371static inline int signal_pending_state(long state, struct task_struct *p)
2372{
2373 if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
2374 return 0;
2375 if (!signal_pending(p))
2376 return 0;
2377
16882c1e
ON		 2378 return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
2379}
2380
1da177e4
LT		 2381static inline int need_resched(void)
2382{
9404ef02 2383 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1da177e4
LT		 2384}
2385
2386/*
2387 * cond_resched() and cond_resched_lock(): latency reduction via
2388 * explicit rescheduling in places that are safe. The return
2389 * value indicates whether a reschedule was done in fact.
2390 * cond_resched_lock() will drop the spinlock before scheduling,
2391 * cond_resched_softirq() will enable bhs before scheduling.
2392 */
c3921ab7 2393extern int _cond_resched(void);
6f80bd98 2394
613afbf8
FW		 2395#define cond_resched() ({ \
2396 __might_sleep(__FILE__, __LINE__, 0); \
2397 _cond_resched(); \
2398})
6f80bd98 2399
613afbf8
FW		 2400extern int __cond_resched_lock(spinlock_t *lock);
2401
716a4234
FW		 2402#ifdef CONFIG_PREEMPT
2403#define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
02b67cc3 2404#else
716a4234 2405#define PREEMPT_LOCK_OFFSET 0
02b67cc3 2406#endif
716a4234 2407
613afbf8 2408#define cond_resched_lock(lock) ({ \
716a4234 2409 __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
613afbf8
FW		 2410 __cond_resched_lock(lock); \
2411})
2412
2413extern int __cond_resched_softirq(void);
2414
2415#define cond_resched_softirq() ({ \
2416 __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
2417 __cond_resched_softirq(); \
2418})
1da177e4
LT		 2419
2420/*
2421 * Does a critical section need to be broken due to another
95c354fe
NP		 2422 * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2423 * but a general need for low latency)
1da177e4 2424 */
95c354fe 2425static inline int spin_needbreak(spinlock_t *lock)
1da177e4 2426{
95c354fe
NP		 2427#ifdef CONFIG_PREEMPT
2428 return spin_is_contended(lock);
2429#else
1da177e4 2430 return 0;
95c354fe 2431#endif
1da177e4
LT		 2432}
2433
f06febc9
FM		 2434/*
2435 * Thread group CPU time accounting.
2436 */
4cd4c1b4 2437void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
4da94d49 2438void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
f06febc9 2439
490dea45 2440static inline void thread_group_cputime_init(struct signal_struct *sig)
f06febc9 2441{
4cd4c1b4
PZ		 2442 sig->cputimer.cputime = INIT_CPUTIME;
2443 spin_lock_init(&sig->cputimer.lock);
2444 sig->cputimer.running = 0;
f06febc9
FM		 2445}
2446
f06febc9
FM		 2447static inline void thread_group_cputime_free(struct signal_struct *sig)
2448{
f06febc9
FM		 2449}
2450
7bb44ade
RM		 2451/*
2452 * Reevaluate whether the task has signals pending delivery.
2453 * Wake the task if so.
2454 * This is required every time the blocked sigset_t changes.
2455 * callers must hold sighand->siglock.
2456 */
2457extern void recalc_sigpending_and_wake(struct task_struct *t);
1da177e4
LT		 2458extern void recalc_sigpending(void);
2459
2460extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2461
2462/*
2463 * Wrappers for p->thread_info->cpu access. No-op on UP.
2464 */
2465#ifdef CONFIG_SMP
2466
2467static inline unsigned int task_cpu(const struct task_struct *p)
2468{
a1261f54 2469 return task_thread_info(p)->cpu;
1da177e4
LT		 2470}
2471
c65cc870 2472extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
1da177e4
LT		 2473
2474#else
2475
2476static inline unsigned int task_cpu(const struct task_struct *p)
2477{
2478 return 0;
2479}
2480
2481static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2482{
2483}
2484
2485#endif /* CONFIG_SMP */
2486
1da177e4 2487extern void arch_pick_mmap_layout(struct mm_struct *mm);
1da177e4 2488
1a3c3034
IM		 2489#ifdef CONFIG_TRACING
2490extern void
2491__trace_special(void *__tr, void *__data,
2492 unsigned long arg1, unsigned long arg2, unsigned long arg3);
1da177e4 2493#else
1a3c3034
IM		 2494static inline void
2495__trace_special(void *__tr, void *__data,
2496 unsigned long arg1, unsigned long arg2, unsigned long arg3)
1da177e4 2497{
1da177e4
LT		 2498}
2499#endif
2500
96f874e2
RR		 2501extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
2502extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
5c45bf27 2503
1da177e4
LT		 2504extern void normalize_rt_tasks(void);
2505
052f1dc7 2506#ifdef CONFIG_GROUP_SCHED
9b5b7751 2507
4cf86d77 2508extern struct task_group init_task_group;
eff766a6
PZ		 2509#ifdef CONFIG_USER_SCHED
2510extern struct task_group root_task_group;
6c415b92 2511extern void set_tg_uid(struct user_struct *user);
eff766a6 2512#endif
9b5b7751 2513
ec7dc8ac 2514extern struct task_group *sched_create_group(struct task_group *parent);
4cf86d77 2515extern void sched_destroy_group(struct task_group *tg);
9b5b7751 2516extern void sched_move_task(struct task_struct *tsk);
052f1dc7 2517#ifdef CONFIG_FAIR_GROUP_SCHED
4cf86d77 2518extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
5cb350ba 2519extern unsigned long sched_group_shares(struct task_group *tg);
052f1dc7
PZ		 2520#endif
2521#ifdef CONFIG_RT_GROUP_SCHED
9f0c1e56
PZ		 2522extern int sched_group_set_rt_runtime(struct task_group *tg,
2523 long rt_runtime_us);
2524extern long sched_group_rt_runtime(struct task_group *tg);
d0b27fa7
PZ		 2525extern int sched_group_set_rt_period(struct task_group *tg,
2526 long rt_period_us);
2527extern long sched_group_rt_period(struct task_group *tg);
54e99124 2528extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
052f1dc7 2529#endif
9b5b7751
SV		 2530#endif
2531
54e99124
DG		 2532extern int task_can_switch_user(struct user_struct *up,
2533 struct task_struct *tsk);
2534
4b98d11b
AD		 2535#ifdef CONFIG_TASK_XACCT
2536static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2537{
940389b8 2538 tsk->ioac.rchar += amt;
4b98d11b
AD		 2539}
2540
2541static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2542{
940389b8 2543 tsk->ioac.wchar += amt;
4b98d11b
AD		 2544}
2545
2546static inline void inc_syscr(struct task_struct *tsk)
2547{
940389b8 2548 tsk->ioac.syscr++;
4b98d11b
AD		 2549}
2550
2551static inline void inc_syscw(struct task_struct *tsk)
2552{
940389b8 2553 tsk->ioac.syscw++;
4b98d11b
AD		 2554}
2555#else
2556static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2557{
2558}
2559
2560static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2561{
2562}
2563
2564static inline void inc_syscr(struct task_struct *tsk)
2565{
2566}
2567
2568static inline void inc_syscw(struct task_struct *tsk)
2569{
2570}
2571#endif
2572
82455257
DH		 2573#ifndef TASK_SIZE_OF
2574#define TASK_SIZE_OF(tsk) TASK_SIZE
2575#endif
2576
0793a61d
TG		 2577/*
2578 * Call the function if the target task is executing on a CPU right now:
2579 */
2580extern void task_oncpu_function_call(struct task_struct *p,
2581 void (*func) (void *info), void *info);
2582
2583
cf475ad2
BS		 2584#ifdef CONFIG_MM_OWNER
2585extern void mm_update_next_owner(struct mm_struct *mm);
2586extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2587#else
2588static inline void mm_update_next_owner(struct mm_struct *mm)
2589{
2590}
2591
2592static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2593{
2594}
2595#endif /* CONFIG_MM_OWNER */
2596
7c731e0a
SR		 2597#define TASK_STATE_TO_CHAR_STR "RSDTtZX"
2598
1da177e4
LT		 2599#endif /* __KERNEL__ */
2600
2601#endif
kernel source for telekom puls (alcatel/mediatek)