blob: 3b7c99265acedca0e2ad7a21cd709bb19393b119 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
David Woodhouseb7b3c762006-04-27 00:12:56 +01004#include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
5
6/*
7 * cloning flags:
8 */
9#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
10#define CLONE_VM 0x00000100 /* set if VM shared between processes */
11#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
12#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
13#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
14#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
15#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
16#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
17#define CLONE_THREAD 0x00010000 /* Same thread group? */
18#define CLONE_NEWNS 0x00020000 /* New namespace group? */
19#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
20#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
21#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
22#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
23#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
24#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
25#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
26#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
27
28/*
29 * Scheduling policies
30 */
31#define SCHED_NORMAL 0
32#define SCHED_FIFO 1
33#define SCHED_RR 2
34#define SCHED_BATCH 3
35
David Woodhousea3b67142006-04-25 14:54:40 +010036#ifdef __KERNEL__
David Woodhouseb7b3c762006-04-27 00:12:56 +010037
38struct sched_param {
39 int sched_priority;
40};
41
Linus Torvalds1da177e2005-04-16 15:20:36 -070042#include <asm/param.h> /* for HZ */
43
Linus Torvalds1da177e2005-04-16 15:20:36 -070044#include <linux/capability.h>
45#include <linux/threads.h>
46#include <linux/kernel.h>
47#include <linux/types.h>
48#include <linux/timex.h>
49#include <linux/jiffies.h>
50#include <linux/rbtree.h>
51#include <linux/thread_info.h>
52#include <linux/cpumask.h>
53#include <linux/errno.h>
54#include <linux/nodemask.h>
55
56#include <asm/system.h>
57#include <asm/semaphore.h>
58#include <asm/page.h>
59#include <asm/ptrace.h>
60#include <asm/mmu.h>
61#include <asm/cputime.h>
62
63#include <linux/smp.h>
64#include <linux/sem.h>
65#include <linux/signal.h>
66#include <linux/securebits.h>
67#include <linux/fs_struct.h>
68#include <linux/compiler.h>
69#include <linux/completion.h>
70#include <linux/pid.h>
71#include <linux/percpu.h>
72#include <linux/topology.h>
73#include <linux/seccomp.h>
Ingo Molnare56d0902006-01-08 01:01:37 -080074#include <linux/rcupdate.h>
Ingo Molnar0771dfe2006-03-27 01:16:22 -080075#include <linux/futex.h>
Ingo Molnar23f78d4a2006-06-27 02:54:53 -070076#include <linux/rtmutex.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070077
David Woodhousea3b67142006-04-25 14:54:40 +010078#include <linux/time.h>
79#include <linux/param.h>
80#include <linux/resource.h>
81#include <linux/timer.h>
82#include <linux/hrtimer.h>
83
84#include <asm/processor.h>
H. J. Lu36d57ac2005-09-06 15:16:49 -070085
Linus Torvalds1da177e2005-04-16 15:20:36 -070086struct exec_domain;
Ingo Molnarc87e2832006-06-27 02:54:58 -070087struct futex_pi_state;
Linus Torvalds1da177e2005-04-16 15:20:36 -070088
89/*
Linus Torvalds1da177e2005-04-16 15:20:36 -070090 * List of flags we want to share for kernel threads,
91 * if only because they are not used by them anyway.
92 */
93#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
94
95/*
96 * These are the constant used to fake the fixed-point load-average
97 * counting. Some notes:
98 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
99 * a load-average precision of 10 bits integer + 11 bits fractional
100 * - if you want to count load-averages more often, you need more
101 * precision, or rounding will get you. With 2-second counting freq,
102 * the EXP_n values would be 1981, 2034 and 2043 if still using only
103 * 11 bit fractions.
104 */
105extern unsigned long avenrun[]; /* Load averages */
106
107#define FSHIFT 11 /* nr of bits of precision */
108#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
109#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
110#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
111#define EXP_5 2014 /* 1/exp(5sec/5min) */
112#define EXP_15 2037 /* 1/exp(5sec/15min) */
113
114#define CALC_LOAD(load,exp,n) \
115 load *= exp; \
116 load += n*(FIXED_1-exp); \
117 load >>= FSHIFT;
118
119extern unsigned long total_forks;
120extern int nr_threads;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700121DECLARE_PER_CPU(unsigned long, process_counts);
122extern int nr_processes(void);
123extern unsigned long nr_running(void);
124extern unsigned long nr_uninterruptible(void);
Jack Steinerdb1b1fe2006-03-31 02:31:21 -0800125extern unsigned long nr_active(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700126extern unsigned long nr_iowait(void);
Peter Williams2dd73a42006-06-27 02:54:34 -0700127extern unsigned long weighted_cpuload(const int cpu);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700128
Linus Torvalds1da177e2005-04-16 15:20:36 -0700129
Linus Torvalds4a8342d2005-09-29 15:18:21 -0700130/*
131 * Task state bitmask. NOTE! These bits are also
132 * encoded in fs/proc/array.c: get_task_state().
133 *
134 * We have two separate sets of flags: task->state
135 * is about runnability, while task->exit_state are
136 * about the task exiting. Confusing, but this way
137 * modifying one set can't modify the other one by
138 * mistake.
139 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700140#define TASK_RUNNING 0
141#define TASK_INTERRUPTIBLE 1
142#define TASK_UNINTERRUPTIBLE 2
Linus Torvalds4a8342d2005-09-29 15:18:21 -0700143#define TASK_STOPPED 4
144#define TASK_TRACED 8
145/* in tsk->exit_state */
146#define EXIT_ZOMBIE 16
147#define EXIT_DEAD 32
148/* in tsk->state again */
149#define TASK_NONINTERACTIVE 64
Oleg Nesterovc394cc92006-09-29 02:01:11 -0700150#define TASK_DEAD 128
Linus Torvalds1da177e2005-04-16 15:20:36 -0700151
152#define __set_task_state(tsk, state_value) \
153 do { (tsk)->state = (state_value); } while (0)
154#define set_task_state(tsk, state_value) \
155 set_mb((tsk)->state, (state_value))
156
Andrew Morton498d0c52005-09-13 01:25:14 -0700157/*
158 * set_current_state() includes a barrier so that the write of current->state
159 * is correctly serialised wrt the caller's subsequent test of whether to
160 * actually sleep:
161 *
162 * set_current_state(TASK_UNINTERRUPTIBLE);
163 * if (do_i_need_to_sleep())
164 * schedule();
165 *
166 * If the caller does not need such serialisation then use __set_current_state()
167 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700168#define __set_current_state(state_value) \
169 do { current->state = (state_value); } while (0)
170#define set_current_state(state_value) \
171 set_mb(current->state, (state_value))
172
173/* Task command name length */
174#define TASK_COMM_LEN 16
175
Linus Torvalds1da177e2005-04-16 15:20:36 -0700176#include <linux/spinlock.h>
177
178/*
179 * This serializes "schedule()" and also protects
180 * the run-queue from deletions/modifications (but
181 * _adding_ to the beginning of the run-queue has
182 * a separate lock).
183 */
184extern rwlock_t tasklist_lock;
185extern spinlock_t mmlist_lock;
186
Ingo Molnar36c8b582006-07-03 00:25:41 -0700187struct task_struct;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700188
189extern void sched_init(void);
190extern void sched_init_smp(void);
Ingo Molnar36c8b582006-07-03 00:25:41 -0700191extern void init_idle(struct task_struct *idle, int cpu);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700192
193extern cpumask_t nohz_cpu_mask;
194
195extern void show_state(void);
196extern void show_regs(struct pt_regs *);
197
198/*
199 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
200 * task), SP is the stack pointer of the first frame that should be shown in the back
201 * trace (or NULL if the entire call-chain of the task should be shown).
202 */
203extern void show_stack(struct task_struct *task, unsigned long *sp);
204
205void io_schedule(void);
206long io_schedule_timeout(long timeout);
207
208extern void cpu_init (void);
209extern void trap_init(void);
210extern void update_process_times(int user);
211extern void scheduler_tick(void);
212
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700213#ifdef CONFIG_DETECT_SOFTLOCKUP
Ingo Molnar6687a972006-03-24 03:18:41 -0800214extern void softlockup_tick(void);
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700215extern void spawn_softlockup_task(void);
216extern void touch_softlockup_watchdog(void);
217#else
Ingo Molnar6687a972006-03-24 03:18:41 -0800218static inline void softlockup_tick(void)
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700219{
220}
221static inline void spawn_softlockup_task(void)
222{
223}
224static inline void touch_softlockup_watchdog(void)
225{
226}
227#endif
228
229
Linus Torvalds1da177e2005-04-16 15:20:36 -0700230/* Attach to any functions which should be ignored in wchan output. */
231#define __sched __attribute__((__section__(".sched.text")))
232/* Is this address in the __sched functions? */
233extern int in_sched_functions(unsigned long addr);
234
235#define MAX_SCHEDULE_TIMEOUT LONG_MAX
236extern signed long FASTCALL(schedule_timeout(signed long timeout));
Nishanth Aravamudan64ed93a2005-09-10 00:27:21 -0700237extern signed long schedule_timeout_interruptible(signed long timeout);
238extern signed long schedule_timeout_uninterruptible(signed long timeout);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700239asmlinkage void schedule(void);
240
241struct namespace;
242
243/* Maximum number of active map areas.. This is a random (large) number */
244#define DEFAULT_MAX_MAP_COUNT 65536
245
246extern int sysctl_max_map_count;
247
248#include <linux/aio.h>
249
250extern unsigned long
251arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
252 unsigned long, unsigned long);
253extern unsigned long
254arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
255 unsigned long len, unsigned long pgoff,
256 unsigned long flags);
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700257extern void arch_unmap_area(struct mm_struct *, unsigned long);
258extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700259
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700260#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
261/*
262 * The mm counters are not protected by its page_table_lock,
263 * so must be incremented atomically.
264 */
Christoph Lameterd3cb4872006-01-06 00:11:20 -0800265#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
266#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
267#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
268#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
269#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
270typedef atomic_long_t mm_counter_t;
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700271
272#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
273/*
274 * The mm counters are protected by its page_table_lock,
275 * so can be incremented directly.
276 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700277#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
278#define get_mm_counter(mm, member) ((mm)->_##member)
279#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
280#define inc_mm_counter(mm, member) (mm)->_##member++
281#define dec_mm_counter(mm, member) (mm)->_##member--
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700282typedef unsigned long mm_counter_t;
Hugh Dickins42946212005-10-29 18:16:05 -0700283
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700284#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
285
286#define get_mm_rss(mm) \
287 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
Hugh Dickins365e9c872005-10-29 18:16:18 -0700288#define update_hiwater_rss(mm) do { \
289 unsigned long _rss = get_mm_rss(mm); \
290 if ((mm)->hiwater_rss < _rss) \
291 (mm)->hiwater_rss = _rss; \
292} while (0)
293#define update_hiwater_vm(mm) do { \
294 if ((mm)->hiwater_vm < (mm)->total_vm) \
295 (mm)->hiwater_vm = (mm)->total_vm; \
296} while (0)
297
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298struct mm_struct {
299 struct vm_area_struct * mmap; /* list of VMAs */
300 struct rb_root mm_rb;
301 struct vm_area_struct * mmap_cache; /* last find_vma result */
302 unsigned long (*get_unmapped_area) (struct file *filp,
303 unsigned long addr, unsigned long len,
304 unsigned long pgoff, unsigned long flags);
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700305 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
Benjamin Herrenschmidt0551fbd2006-02-28 16:59:19 -0800306 unsigned long mmap_base; /* base of mmap area */
307 unsigned long task_size; /* size of task vm space */
308 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700309 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700310 pgd_t * pgd;
311 atomic_t mm_users; /* How many users with user space? */
312 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
313 int map_count; /* number of VMAs */
314 struct rw_semaphore mmap_sem;
315 spinlock_t page_table_lock; /* Protects page tables and some counters */
316
317 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
318 * together off init_mm.mmlist, and are protected
319 * by mmlist_lock
320 */
321
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700322 /* Special counters, in some configurations protected by the
323 * page_table_lock, in other configurations by being atomic.
324 */
Hugh Dickins42946212005-10-29 18:16:05 -0700325 mm_counter_t _file_rss;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700326 mm_counter_t _anon_rss;
327
Hugh Dickinsf4499522005-10-29 18:16:19 -0700328 unsigned long hiwater_rss; /* High-watermark of RSS usage */
329 unsigned long hiwater_vm; /* High-water virtual memory usage */
330
331 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
332 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
333 unsigned long start_code, end_code, start_data, end_data;
334 unsigned long start_brk, brk, start_stack;
335 unsigned long arg_start, arg_end, env_start, env_end;
336
H. J. Lu36d57ac2005-09-06 15:16:49 -0700337 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338
Alan Coxd6e71142005-06-23 00:09:43 -0700339 unsigned dumpable:2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700340 cpumask_t cpu_vm_mask;
341
342 /* Architecture-specific MM context */
343 mm_context_t context;
344
345 /* Token based thrashing protection. */
346 unsigned long swap_token_time;
347 char recent_pagein;
348
349 /* coredumping support */
350 int core_waiters;
351 struct completion *core_startup_done, core_done;
352
353 /* aio bits */
354 rwlock_t ioctx_list_lock;
355 struct kioctx *ioctx_list;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700356};
357
358struct sighand_struct {
359 atomic_t count;
360 struct k_sigaction action[_NSIG];
361 spinlock_t siglock;
362};
363
KaiGai Kohei0e464812006-06-25 05:49:24 -0700364struct pacct_struct {
KaiGai Koheif6ec29a2006-06-25 05:49:25 -0700365 int ac_flag;
366 long ac_exitcode;
KaiGai Kohei0e464812006-06-25 05:49:24 -0700367 unsigned long ac_mem;
KaiGai Kohei77787bf2006-06-25 05:49:26 -0700368 cputime_t ac_utime, ac_stime;
369 unsigned long ac_minflt, ac_majflt;
KaiGai Kohei0e464812006-06-25 05:49:24 -0700370};
371
Linus Torvalds1da177e2005-04-16 15:20:36 -0700372/*
373 * NOTE! "signal_struct" does not have it's own
374 * locking, because a shared signal_struct always
375 * implies a shared sighand_struct, so locking
376 * sighand_struct is always a proper superset of
377 * the locking of signal_struct.
378 */
379struct signal_struct {
380 atomic_t count;
381 atomic_t live;
382
383 wait_queue_head_t wait_chldexit; /* for wait4() */
384
385 /* current thread group signal load-balancing target: */
Ingo Molnar36c8b582006-07-03 00:25:41 -0700386 struct task_struct *curr_target;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700387
388 /* shared signal handling: */
389 struct sigpending shared_pending;
390
391 /* thread group exit support */
392 int group_exit_code;
393 /* overloaded:
394 * - notify group_exit_task when ->count is equal to notify_count
395 * - everyone except group_exit_task is stopped during signal delivery
396 * of fatal signals, group_exit_task processes the signal.
397 */
398 struct task_struct *group_exit_task;
399 int notify_count;
400
401 /* thread group stop support, overloads group_exit_code too */
402 int group_stop_count;
403 unsigned int flags; /* see SIGNAL_* flags below */
404
405 /* POSIX.1b Interval Timers */
406 struct list_head posix_timers;
407
408 /* ITIMER_REAL timer for the process */
Thomas Gleixner2ff678b2006-01-09 20:52:34 -0800409 struct hrtimer real_timer;
Roman Zippel05cfb612006-03-26 01:38:12 -0800410 struct task_struct *tsk;
Thomas Gleixner2ff678b2006-01-09 20:52:34 -0800411 ktime_t it_real_incr;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700412
413 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
414 cputime_t it_prof_expires, it_virt_expires;
415 cputime_t it_prof_incr, it_virt_incr;
416
417 /* job control IDs */
418 pid_t pgrp;
419 pid_t tty_old_pgrp;
420 pid_t session;
421 /* boolean value for session group leader */
422 int leader;
423
424 struct tty_struct *tty; /* NULL if no tty */
425
426 /*
427 * Cumulative resource counters for dead threads in the group,
428 * and for reaped dead child processes forked by this group.
429 * Live threads maintain their own counters and add to these
430 * in __exit_signal, except for the group leader.
431 */
432 cputime_t utime, stime, cutime, cstime;
433 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
434 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
435
436 /*
437 * Cumulative ns of scheduled CPU time for dead threads in the
438 * group, not including a zombie group leader. (This only differs
439 * from jiffies_to_ns(utime + stime) if sched_clock uses something
440 * other than jiffies.)
441 */
442 unsigned long long sched_time;
443
444 /*
445 * We don't bother to synchronize most readers of this at all,
446 * because there is no reader checking a limit that actually needs
447 * to get both rlim_cur and rlim_max atomically, and either one
448 * alone is a single word that can safely be read normally.
449 * getrlimit/setrlimit use task_lock(current->group_leader) to
450 * protect this instead of the siglock, because they really
451 * have no need to disable irqs.
452 */
453 struct rlimit rlim[RLIM_NLIMITS];
454
455 struct list_head cpu_timers[3];
456
457 /* keep the process-shared keyrings here so that they do the right
458 * thing in threads created with CLONE_THREAD */
459#ifdef CONFIG_KEYS
460 struct key *session_keyring; /* keyring inherited over fork */
461 struct key *process_keyring; /* keyring private to this process */
462#endif
KaiGai Kohei0e464812006-06-25 05:49:24 -0700463#ifdef CONFIG_BSD_PROCESS_ACCT
464 struct pacct_struct pacct; /* per-process accounting information */
465#endif
Shailabh Nagarad4ecbc2006-07-14 00:24:44 -0700466#ifdef CONFIG_TASKSTATS
467 spinlock_t stats_lock;
468 struct taskstats *stats;
469#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700470};
471
Nick Piggin4866cde2005-06-25 14:57:23 -0700472/* Context switch must be unlocked if interrupts are to be enabled */
473#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
474# define __ARCH_WANT_UNLOCKED_CTXSW
475#endif
476
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477/*
478 * Bits in flags field of signal_struct.
479 */
480#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
481#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
482#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
483#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
484
485
486/*
487 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
Ingo Molnarb0a94992006-01-14 13:20:41 -0800488 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
489 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
490 * values are inverted: lower p->prio value means higher priority.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700491 *
492 * The MAX_USER_RT_PRIO value allows the actual maximum
493 * RT priority to be separate from the value exported to
494 * user-space. This allows kernel threads to set their
495 * priority to a value higher than any user task. Note:
496 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
497 */
498
499#define MAX_USER_RT_PRIO 100
500#define MAX_RT_PRIO MAX_USER_RT_PRIO
501
502#define MAX_PRIO (MAX_RT_PRIO + 40)
503
Ingo Molnarb29739f2006-06-27 02:54:51 -0700504#define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
505#define rt_task(p) rt_prio((p)->prio)
Con Kolivasd425b272006-03-31 02:31:29 -0800506#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
Oleg Nesterov57a6f51c2006-09-29 02:00:49 -0700507#define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
508#define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700509
510/*
511 * Some day this will be a full-fledged user tracking system..
512 */
513struct user_struct {
514 atomic_t __count; /* reference count */
515 atomic_t processes; /* How many processes does this user have? */
516 atomic_t files; /* How many open files does this user have? */
517 atomic_t sigpending; /* How many pending signals does this user have? */
Amy Griffis2d9048e2006-06-01 13:10:59 -0700518#ifdef CONFIG_INOTIFY_USER
Robert Love0eeca282005-07-12 17:06:03 -0400519 atomic_t inotify_watches; /* How many inotify watches does this user have? */
520 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
521#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700522 /* protected by mq_lock */
523 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
524 unsigned long locked_shm; /* How many pages of mlocked shm ? */
525
526#ifdef CONFIG_KEYS
527 struct key *uid_keyring; /* UID specific keyring */
528 struct key *session_keyring; /* UID's default session keyring */
529#endif
530
531 /* Hash table maintenance information */
532 struct list_head uidhash_list;
533 uid_t uid;
534};
535
536extern struct user_struct *find_user(uid_t);
537
538extern struct user_struct root_user;
539#define INIT_USER (&root_user)
540
Linus Torvalds1da177e2005-04-16 15:20:36 -0700541struct backing_dev_info;
542struct reclaim_state;
543
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700544#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700545struct sched_info {
546 /* cumulative counters */
547 unsigned long cpu_time, /* time spent on the cpu */
548 run_delay, /* time spent waiting on a runqueue */
549 pcnt; /* # of timeslices run on this cpu */
550
551 /* timestamps */
552 unsigned long last_arrival, /* when we last ran on a cpu */
553 last_queued; /* when we were last queued to run */
554};
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700555#endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700556
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700557#ifdef CONFIG_SCHEDSTATS
Linus Torvalds1da177e2005-04-16 15:20:36 -0700558extern struct file_operations proc_schedstat_operations;
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700559#endif /* CONFIG_SCHEDSTATS */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700560
Shailabh Nagarca74e922006-07-14 00:24:36 -0700561#ifdef CONFIG_TASK_DELAY_ACCT
562struct task_delay_info {
563 spinlock_t lock;
564 unsigned int flags; /* Private per-task flags */
565
566 /* For each stat XXX, add following, aligned appropriately
567 *
568 * struct timespec XXX_start, XXX_end;
569 * u64 XXX_delay;
570 * u32 XXX_count;
571 *
572 * Atomicity of updates to XXX_delay, XXX_count protected by
573 * single lock above (split into XXX_lock if contention is an issue).
574 */
Shailabh Nagar0ff92242006-07-14 00:24:37 -0700575
576 /*
577 * XXX_count is incremented on every XXX operation, the delay
578 * associated with the operation is added to XXX_delay.
579 * XXX_delay contains the accumulated delay time in nanoseconds.
580 */
581 struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
582 u64 blkio_delay; /* wait for sync block io completion */
583 u64 swapin_delay; /* wait for swapin block io completion */
584 u32 blkio_count; /* total count of the number of sync block */
585 /* io operations performed */
586 u32 swapin_count; /* total count of the number of swapin block */
587 /* io operations performed */
Shailabh Nagarca74e922006-07-14 00:24:36 -0700588};
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700589#endif /* CONFIG_TASK_DELAY_ACCT */
590
591static inline int sched_info_on(void)
592{
593#ifdef CONFIG_SCHEDSTATS
594 return 1;
595#elif defined(CONFIG_TASK_DELAY_ACCT)
596 extern int delayacct_on;
597 return delayacct_on;
598#else
599 return 0;
Shailabh Nagarca74e922006-07-14 00:24:36 -0700600#endif
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700601}
Shailabh Nagarca74e922006-07-14 00:24:36 -0700602
Linus Torvalds1da177e2005-04-16 15:20:36 -0700603enum idle_type
604{
605 SCHED_IDLE,
606 NOT_IDLE,
607 NEWLY_IDLE,
608 MAX_IDLE_TYPES
609};
610
611/*
612 * sched-domains (multiprocessor balancing) declarations:
613 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700614#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
615
Peter Williams2dd73a42006-06-27 02:54:34 -0700616#ifdef CONFIG_SMP
Linus Torvalds1da177e2005-04-16 15:20:36 -0700617#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
618#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
619#define SD_BALANCE_EXEC 4 /* Balance on exec */
Nick Piggin147cbb42005-06-25 14:57:19 -0700620#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
621#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
622#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
623#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
624#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
Siddha, Suresh B5c45bf22006-06-27 02:54:42 -0700625#define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
626
627#define BALANCE_FOR_POWER ((sched_mc_power_savings || sched_smt_power_savings) \
628 ? SD_POWERSAVINGS_BALANCE : 0)
629
Linus Torvalds1da177e2005-04-16 15:20:36 -0700630
631struct sched_group {
632 struct sched_group *next; /* Must be a circular list */
633 cpumask_t cpumask;
634
635 /*
636 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
637 * single CPU. This is read only (except for setup, hotplug CPU).
638 */
639 unsigned long cpu_power;
640};
641
642struct sched_domain {
643 /* These fields must be setup */
644 struct sched_domain *parent; /* top domain must be null terminated */
645 struct sched_group *groups; /* the balancing groups of the domain */
646 cpumask_t span; /* span of all CPUs in this domain */
647 unsigned long min_interval; /* Minimum balance interval ms */
648 unsigned long max_interval; /* Maximum balance interval ms */
649 unsigned int busy_factor; /* less balancing by factor if busy */
650 unsigned int imbalance_pct; /* No balance until over watermark */
651 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
652 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
653 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
Nick Piggin78979862005-06-25 14:57:13 -0700654 unsigned int busy_idx;
655 unsigned int idle_idx;
656 unsigned int newidle_idx;
657 unsigned int wake_idx;
Nick Piggin147cbb42005-06-25 14:57:19 -0700658 unsigned int forkexec_idx;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700659 int flags; /* See SD_* */
660
661 /* Runtime fields. */
662 unsigned long last_balance; /* init to jiffies. units in jiffies */
663 unsigned int balance_interval; /* initialise to 1. units in ms. */
664 unsigned int nr_balance_failed; /* initialise to 0 */
665
666#ifdef CONFIG_SCHEDSTATS
667 /* load_balance() stats */
668 unsigned long lb_cnt[MAX_IDLE_TYPES];
669 unsigned long lb_failed[MAX_IDLE_TYPES];
670 unsigned long lb_balanced[MAX_IDLE_TYPES];
671 unsigned long lb_imbalance[MAX_IDLE_TYPES];
672 unsigned long lb_gained[MAX_IDLE_TYPES];
673 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
674 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
675 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
676
677 /* Active load balancing */
678 unsigned long alb_cnt;
679 unsigned long alb_failed;
680 unsigned long alb_pushed;
681
Nick Piggin68767a02005-06-25 14:57:20 -0700682 /* SD_BALANCE_EXEC stats */
683 unsigned long sbe_cnt;
684 unsigned long sbe_balanced;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700685 unsigned long sbe_pushed;
686
Nick Piggin68767a02005-06-25 14:57:20 -0700687 /* SD_BALANCE_FORK stats */
688 unsigned long sbf_cnt;
689 unsigned long sbf_balanced;
690 unsigned long sbf_pushed;
691
Linus Torvalds1da177e2005-04-16 15:20:36 -0700692 /* try_to_wake_up() stats */
693 unsigned long ttwu_wake_remote;
694 unsigned long ttwu_move_affine;
695 unsigned long ttwu_move_balance;
696#endif
697};
698
Srivatsa Vaddagiri51888ca2006-06-27 02:54:38 -0700699extern int partition_sched_domains(cpumask_t *partition1,
Dinakar Guniguntala1a20ff22005-06-25 14:57:33 -0700700 cpumask_t *partition2);
akpm@osdl.org198e2f12006-01-12 01:05:30 -0800701
702/*
703 * Maximum cache size the migration-costs auto-tuning code will
704 * search from:
705 */
706extern unsigned int max_cache_size;
707
708#endif /* CONFIG_SMP */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700709
710
711struct io_context; /* See blkdev.h */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700712struct cpuset;
713
714#define NGROUPS_SMALL 32
715#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
716struct group_info {
717 int ngroups;
718 atomic_t usage;
719 gid_t small_block[NGROUPS_SMALL];
720 int nblocks;
721 gid_t *blocks[0];
722};
723
724/*
725 * get_group_info() must be called with the owning task locked (via task_lock())
726 * when task != current. The reason being that the vast majority of callers are
727 * looking at current->group_info, which can not be changed except by the
728 * current task. Changing current->group_info requires the task lock, too.
729 */
730#define get_group_info(group_info) do { \
731 atomic_inc(&(group_info)->usage); \
732} while (0)
733
734#define put_group_info(group_info) do { \
735 if (atomic_dec_and_test(&(group_info)->usage)) \
736 groups_free(group_info); \
737} while (0)
738
David Howells3e301482005-06-23 22:00:56 -0700739extern struct group_info *groups_alloc(int gidsetsize);
740extern void groups_free(struct group_info *group_info);
741extern int set_current_groups(struct group_info *group_info);
742extern int groups_search(struct group_info *group_info, gid_t grp);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700743/* access the groups "array" with this macro */
744#define GROUP_AT(gi, i) \
745 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
746
Chen, Kenneth W383f2832005-09-09 13:02:02 -0700747#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
Ingo Molnar36c8b582006-07-03 00:25:41 -0700748extern void prefetch_stack(struct task_struct *t);
Chen, Kenneth W383f2832005-09-09 13:02:02 -0700749#else
750static inline void prefetch_stack(struct task_struct *t) { }
751#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700752
753struct audit_context; /* See audit.c */
754struct mempolicy;
Jens Axboeb92ce552006-04-11 13:52:07 +0200755struct pipe_inode_info;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700756
Con Kolivas3dee3862006-03-31 02:31:23 -0800757enum sleep_type {
758 SLEEP_NORMAL,
759 SLEEP_NONINTERACTIVE,
760 SLEEP_INTERACTIVE,
761 SLEEP_INTERRUPTED,
762};
763
Ingo Molnar70b97a72006-07-03 00:25:42 -0700764struct prio_array;
765
Linus Torvalds1da177e2005-04-16 15:20:36 -0700766struct task_struct {
767 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
768 struct thread_info *thread_info;
769 atomic_t usage;
770 unsigned long flags; /* per process flags, defined below */
771 unsigned long ptrace;
772
Paolo 'Blaisorblade' Giarrusso36772092005-05-05 16:16:12 -0700773 int lock_depth; /* BKL lock depth */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700774
Peter Williams2dd73a42006-06-27 02:54:34 -0700775#ifdef CONFIG_SMP
776#ifdef __ARCH_WANT_UNLOCKED_CTXSW
Nick Piggin4866cde2005-06-25 14:57:23 -0700777 int oncpu;
778#endif
Peter Williams2dd73a42006-06-27 02:54:34 -0700779#endif
780 int load_weight; /* for niceness load balancing purposes */
Ingo Molnarb29739f2006-06-27 02:54:51 -0700781 int prio, static_prio, normal_prio;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700782 struct list_head run_list;
Ingo Molnar70b97a72006-07-03 00:25:42 -0700783 struct prio_array *array;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700784
Jens Axboe22e2c502005-06-27 10:55:12 +0200785 unsigned short ioprio;
Alexey Dobriyan6c5c9342006-09-29 01:59:40 -0700786#ifdef CONFIG_BLK_DEV_IO_TRACE
Jens Axboe2056a782006-03-23 20:00:26 +0100787 unsigned int btrace_seq;
Alexey Dobriyan6c5c9342006-09-29 01:59:40 -0700788#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700789 unsigned long sleep_avg;
790 unsigned long long timestamp, last_ran;
791 unsigned long long sched_time; /* sched_clock time spent running */
Con Kolivas3dee3862006-03-31 02:31:23 -0800792 enum sleep_type sleep_type;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700793
794 unsigned long policy;
795 cpumask_t cpus_allowed;
796 unsigned int time_slice, first_time_slice;
797
Chandra Seetharaman52f17b62006-07-14 00:24:38 -0700798#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700799 struct sched_info sched_info;
800#endif
801
802 struct list_head tasks;
803 /*
804 * ptrace_list/ptrace_children forms the list of my children
805 * that were stolen by a ptracer.
806 */
807 struct list_head ptrace_children;
808 struct list_head ptrace_list;
809
810 struct mm_struct *mm, *active_mm;
811
812/* task state */
813 struct linux_binfmt *binfmt;
814 long exit_state;
815 int exit_code, exit_signal;
816 int pdeath_signal; /* The signal sent when the parent dies */
817 /* ??? */
818 unsigned long personality;
819 unsigned did_exec:1;
820 pid_t pid;
821 pid_t tgid;
Arjan van de Ven0a4254052006-09-26 10:52:38 +0200822
823#ifdef CONFIG_CC_STACKPROTECTOR
824 /* Canary value for the -fstack-protector gcc feature */
825 unsigned long stack_canary;
826#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700827 /*
828 * pointers to (original) parent process, youngest child, younger sibling,
829 * older sibling, respectively. (p->father can be replaced with
830 * p->parent->pid)
831 */
832 struct task_struct *real_parent; /* real parent process (when being debugged) */
833 struct task_struct *parent; /* parent process */
834 /*
835 * children/sibling forms the list of my children plus the
836 * tasks I'm ptracing.
837 */
838 struct list_head children; /* list of my children */
839 struct list_head sibling; /* linkage in my parent's children list */
840 struct task_struct *group_leader; /* threadgroup leader */
841
842 /* PID/PID hash table linkage. */
Eric W. Biederman92476d72006-03-31 02:31:42 -0800843 struct pid_link pids[PIDTYPE_MAX];
Oleg Nesterov47e65322006-03-28 16:11:25 -0800844 struct list_head thread_group;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700845
846 struct completion *vfork_done; /* for vfork() */
847 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
848 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
849
850 unsigned long rt_priority;
851 cputime_t utime, stime;
852 unsigned long nvcsw, nivcsw; /* context switch counts */
853 struct timespec start_time;
854/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
855 unsigned long min_flt, maj_flt;
856
857 cputime_t it_prof_expires, it_virt_expires;
858 unsigned long long it_sched_expires;
859 struct list_head cpu_timers[3];
860
861/* process credentials */
862 uid_t uid,euid,suid,fsuid;
863 gid_t gid,egid,sgid,fsgid;
864 struct group_info *group_info;
865 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
866 unsigned keep_capabilities:1;
867 struct user_struct *user;
868#ifdef CONFIG_KEYS
David Howellsb5f545c2006-01-08 01:02:47 -0800869 struct key *request_key_auth; /* assumed request_key authority */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700870 struct key *thread_keyring; /* keyring private to this thread */
David Howells3e301482005-06-23 22:00:56 -0700871 unsigned char jit_keyring; /* default keyring to attach requested keys to */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700872#endif
Arjan van de Vene07e23e2006-09-26 10:52:36 +0200873 /*
874 * fpu_counter contains the number of consecutive context switches
875 * that the FPU is used. If this is over a threshold, the lazy fpu
876 * saving becomes unlazy to save the trap. This is an unsigned char
877 * so that after 256 times the counter wraps and the behavior turns
878 * lazy again; this to deal with bursty apps that only use FPU for
879 * a short time
880 */
881 unsigned char fpu_counter;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700882 int oomkilladj; /* OOM kill score adjustment (bit shift). */
Paolo 'Blaisorblade' Giarrusso36772092005-05-05 16:16:12 -0700883 char comm[TASK_COMM_LEN]; /* executable name excluding path
884 - access with [gs]et_task_comm (which lock
885 it with task_lock())
886 - initialized normally by flush_old_exec */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700887/* file system info */
888 int link_count, total_link_count;
Alexey Dobriyan3d5b6fc2006-09-29 01:59:40 -0700889#ifdef CONFIG_SYSVIPC
Linus Torvalds1da177e2005-04-16 15:20:36 -0700890/* ipc stuff */
891 struct sysv_sem sysvsem;
Alexey Dobriyan3d5b6fc2006-09-29 01:59:40 -0700892#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700893/* CPU-specific state of this task */
894 struct thread_struct thread;
895/* filesystem information */
896 struct fs_struct *fs;
897/* open file information */
898 struct files_struct *files;
899/* namespace */
900 struct namespace *namespace;
901/* signal handlers */
902 struct signal_struct *signal;
903 struct sighand_struct *sighand;
904
905 sigset_t blocked, real_blocked;
David Woodhouse150256d2006-01-18 17:43:57 -0800906 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700907 struct sigpending pending;
908
909 unsigned long sas_ss_sp;
910 size_t sas_ss_size;
911 int (*notifier)(void *priv);
912 void *notifier_data;
913 sigset_t *notifier_mask;
914
915 void *security;
916 struct audit_context *audit_context;
917 seccomp_t seccomp;
918
919/* Thread group tracking */
920 u32 parent_exec_id;
921 u32 self_exec_id;
922/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
923 spinlock_t alloc_lock;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700924
Ingo Molnarb29739f2006-06-27 02:54:51 -0700925 /* Protection of the PI data structures: */
926 spinlock_t pi_lock;
927
Ingo Molnar23f78d4a2006-06-27 02:54:53 -0700928#ifdef CONFIG_RT_MUTEXES
929 /* PI waiters blocked on a rt_mutex held by this task */
930 struct plist_head pi_waiters;
931 /* Deadlock detection and priority inheritance handling */
932 struct rt_mutex_waiter *pi_blocked_on;
Ingo Molnar23f78d4a2006-06-27 02:54:53 -0700933#endif
934
Ingo Molnar408894e2006-01-09 15:59:20 -0800935#ifdef CONFIG_DEBUG_MUTEXES
936 /* mutex deadlock detection */
937 struct mutex_waiter *blocked_on;
938#endif
Ingo Molnarde30a2b2006-07-03 00:24:42 -0700939#ifdef CONFIG_TRACE_IRQFLAGS
940 unsigned int irq_events;
941 int hardirqs_enabled;
942 unsigned long hardirq_enable_ip;
943 unsigned int hardirq_enable_event;
944 unsigned long hardirq_disable_ip;
945 unsigned int hardirq_disable_event;
946 int softirqs_enabled;
947 unsigned long softirq_disable_ip;
948 unsigned int softirq_disable_event;
949 unsigned long softirq_enable_ip;
950 unsigned int softirq_enable_event;
951 int hardirq_context;
952 int softirq_context;
953#endif
Ingo Molnarfbb9ce952006-07-03 00:24:50 -0700954#ifdef CONFIG_LOCKDEP
955# define MAX_LOCK_DEPTH 30UL
956 u64 curr_chain_key;
957 int lockdep_depth;
958 struct held_lock held_locks[MAX_LOCK_DEPTH];
959 unsigned int lockdep_recursion;
960#endif
Ingo Molnar408894e2006-01-09 15:59:20 -0800961
Linus Torvalds1da177e2005-04-16 15:20:36 -0700962/* journalling filesystem info */
963 void *journal_info;
964
965/* VM state */
966 struct reclaim_state *reclaim_state;
967
Linus Torvalds1da177e2005-04-16 15:20:36 -0700968 struct backing_dev_info *backing_dev_info;
969
970 struct io_context *io_context;
971
972 unsigned long ptrace_message;
973 siginfo_t *last_siginfo; /* For ptrace use. */
974/*
975 * current io wait handle: wait queue entry to use for io waits
976 * If this thread is processing aio, this points at the waitqueue
977 * inside the currently handled kiocb. It may be NULL (i.e. default
978 * to a stack based synchronous wait) if its doing sync IO.
979 */
980 wait_queue_t *io_wait;
981/* i/o counters(bytes read/written, #syscalls */
982 u64 rchar, wchar, syscr, syscw;
Jay Lan8f0ab512006-09-30 23:28:59 -0700983#if defined(CONFIG_TASK_XACCT)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700984 u64 acct_rss_mem1; /* accumulated rss usage */
985 u64 acct_vm_mem1; /* accumulated virtual memory usage */
Jay Landb5fed262006-09-30 23:29:00 -0700986 cputime_t acct_stimexpd;/* stime since last update */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700987#endif
988#ifdef CONFIG_NUMA
989 struct mempolicy *mempolicy;
990 short il_next;
991#endif
992#ifdef CONFIG_CPUSETS
993 struct cpuset *cpuset;
994 nodemask_t mems_allowed;
995 int cpuset_mems_generation;
Paul Jackson825a46a2006-03-24 03:16:03 -0800996 int cpuset_mem_spread_rotor;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700997#endif
Ingo Molnar0771dfe2006-03-27 01:16:22 -0800998 struct robust_list_head __user *robust_list;
Ingo Molnar34f192c2006-03-27 01:16:24 -0800999#ifdef CONFIG_COMPAT
1000 struct compat_robust_list_head __user *compat_robust_list;
1001#endif
Ingo Molnarc87e2832006-06-27 02:54:58 -07001002 struct list_head pi_state_list;
1003 struct futex_pi_state *pi_state_cache;
Ingo Molnar0771dfe2006-03-27 01:16:22 -08001004
Jens Axboe22e2c502005-06-27 10:55:12 +02001005 atomic_t fs_excl; /* holding fs exclusive resources */
Ingo Molnare56d0902006-01-08 01:01:37 -08001006 struct rcu_head rcu;
Jens Axboeb92ce552006-04-11 13:52:07 +02001007
1008 /*
1009 * cache last used pipe for splice
1010 */
1011 struct pipe_inode_info *splice_pipe;
Shailabh Nagarca74e922006-07-14 00:24:36 -07001012#ifdef CONFIG_TASK_DELAY_ACCT
1013 struct task_delay_info *delays;
1014#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07001015};
1016
1017static inline pid_t process_group(struct task_struct *tsk)
1018{
1019 return tsk->signal->pgrp;
1020}
1021
Eric W. Biederman22c935f2006-10-02 02:17:09 -07001022static inline struct pid *task_pid(struct task_struct *task)
1023{
1024 return task->pids[PIDTYPE_PID].pid;
1025}
1026
1027static inline struct pid *task_tgid(struct task_struct *task)
1028{
1029 return task->group_leader->pids[PIDTYPE_PID].pid;
1030}
1031
1032static inline struct pid *task_pgrp(struct task_struct *task)
1033{
1034 return task->group_leader->pids[PIDTYPE_PGID].pid;
1035}
1036
1037static inline struct pid *task_session(struct task_struct *task)
1038{
1039 return task->group_leader->pids[PIDTYPE_SID].pid;
1040}
1041
Linus Torvalds1da177e2005-04-16 15:20:36 -07001042/**
1043 * pid_alive - check that a task structure is not stale
1044 * @p: Task structure to be checked.
1045 *
1046 * Test if a process is not yet dead (at most zombie state)
1047 * If pid_alive fails, then pointers within the task structure
1048 * can be stale and must not be dereferenced.
1049 */
1050static inline int pid_alive(struct task_struct *p)
1051{
Eric W. Biederman92476d72006-03-31 02:31:42 -08001052 return p->pids[PIDTYPE_PID].pid != NULL;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001053}
1054
Sukadev Bhattiproluf400e192006-09-29 02:00:07 -07001055/**
1056 * is_init - check if a task structure is the first user space
1057 * task the kernel created.
1058 * @p: Task structure to be checked.
1059 */
1060static inline int is_init(struct task_struct *tsk)
1061{
1062 return tsk->pid == 1;
1063}
1064
Linus Torvalds1da177e2005-04-16 15:20:36 -07001065extern void free_task(struct task_struct *tsk);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001066#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
Ingo Molnare56d0902006-01-08 01:01:37 -08001067
Andrew Morton158d9eb2006-03-31 02:31:34 -08001068extern void __put_task_struct(struct task_struct *t);
Ingo Molnare56d0902006-01-08 01:01:37 -08001069
1070static inline void put_task_struct(struct task_struct *t)
1071{
1072 if (atomic_dec_and_test(&t->usage))
Eric W. Biederman8c7904a2006-03-31 02:31:37 -08001073 __put_task_struct(t);
Ingo Molnare56d0902006-01-08 01:01:37 -08001074}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001075
1076/*
1077 * Per process flags
1078 */
1079#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
1080 /* Not implemented yet, only for 486*/
1081#define PF_STARTING 0x00000002 /* being created */
1082#define PF_EXITING 0x00000004 /* getting shut down */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001083#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
1084#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
1085#define PF_DUMPCORE 0x00000200 /* dumped core */
1086#define PF_SIGNALED 0x00000400 /* killed by a signal */
1087#define PF_MEMALLOC 0x00000800 /* Allocating memory */
1088#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
1089#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
1090#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
1091#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
1092#define PF_FROZEN 0x00010000 /* frozen for system suspend */
1093#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
1094#define PF_KSWAPD 0x00040000 /* I am kswapd */
1095#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
1096#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
Jens Axboeb31dc662006-06-13 08:26:10 +02001097#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
1098#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
1099#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
1100#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
1101#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
Paul Jacksonc61afb12006-03-24 03:16:08 -08001102#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
Thomas Gleixner61a87122006-06-27 02:54:56 -07001103#define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001104
1105/*
1106 * Only the _current_ task can read/write to tsk->flags, but other
1107 * tasks can access tsk->flags in readonly mode for example
1108 * with tsk_used_math (like during threaded core dumping).
1109 * There is however an exception to this rule during ptrace
1110 * or during fork: the ptracer task is allowed to write to the
1111 * child->flags of its traced child (same goes for fork, the parent
1112 * can write to the child->flags), because we're guaranteed the
1113 * child is not running and in turn not changing child->flags
1114 * at the same time the parent does it.
1115 */
1116#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1117#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1118#define clear_used_math() clear_stopped_child_used_math(current)
1119#define set_used_math() set_stopped_child_used_math(current)
1120#define conditional_stopped_child_used_math(condition, child) \
1121 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1122#define conditional_used_math(condition) \
1123 conditional_stopped_child_used_math(condition, current)
1124#define copy_to_stopped_child_used_math(child) \
1125 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1126/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1127#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1128#define used_math() tsk_used_math(current)
1129
1130#ifdef CONFIG_SMP
Ingo Molnar36c8b582006-07-03 00:25:41 -07001131extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001132#else
Ingo Molnar36c8b582006-07-03 00:25:41 -07001133static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001134{
Paul Jackson4098f992005-10-30 15:03:21 -08001135 if (!cpu_isset(0, new_mask))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001136 return -EINVAL;
1137 return 0;
1138}
1139#endif
1140
1141extern unsigned long long sched_clock(void);
Ingo Molnar36c8b582006-07-03 00:25:41 -07001142extern unsigned long long
1143current_sched_time(const struct task_struct *current_task);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001144
1145/* sched_exec is called by processes performing an exec */
1146#ifdef CONFIG_SMP
1147extern void sched_exec(void);
1148#else
1149#define sched_exec() {}
1150#endif
1151
1152#ifdef CONFIG_HOTPLUG_CPU
1153extern void idle_task_exit(void);
1154#else
1155static inline void idle_task_exit(void) {}
1156#endif
1157
1158extern void sched_idle_next(void);
Ingo Molnarb29739f2006-06-27 02:54:51 -07001159
1160#ifdef CONFIG_RT_MUTEXES
Ingo Molnar36c8b582006-07-03 00:25:41 -07001161extern int rt_mutex_getprio(struct task_struct *p);
1162extern void rt_mutex_setprio(struct task_struct *p, int prio);
1163extern void rt_mutex_adjust_pi(struct task_struct *p);
Ingo Molnarb29739f2006-06-27 02:54:51 -07001164#else
Ingo Molnar36c8b582006-07-03 00:25:41 -07001165static inline int rt_mutex_getprio(struct task_struct *p)
Ingo Molnarb29739f2006-06-27 02:54:51 -07001166{
1167 return p->normal_prio;
1168}
Thomas Gleixner95e02ca2006-06-27 02:55:02 -07001169# define rt_mutex_adjust_pi(p) do { } while (0)
Ingo Molnarb29739f2006-06-27 02:54:51 -07001170#endif
1171
Ingo Molnar36c8b582006-07-03 00:25:41 -07001172extern void set_user_nice(struct task_struct *p, long nice);
1173extern int task_prio(const struct task_struct *p);
1174extern int task_nice(const struct task_struct *p);
1175extern int can_nice(const struct task_struct *p, const int nice);
1176extern int task_curr(const struct task_struct *p);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001177extern int idle_cpu(int cpu);
1178extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
Ingo Molnar36c8b582006-07-03 00:25:41 -07001179extern struct task_struct *idle_task(int cpu);
1180extern struct task_struct *curr_task(int cpu);
1181extern void set_curr_task(int cpu, struct task_struct *p);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001182
1183void yield(void);
1184
1185/*
1186 * The default (Linux) execution domain.
1187 */
1188extern struct exec_domain default_exec_domain;
1189
1190union thread_union {
1191 struct thread_info thread_info;
1192 unsigned long stack[THREAD_SIZE/sizeof(long)];
1193};
1194
1195#ifndef __HAVE_ARCH_KSTACK_END
1196static inline int kstack_end(void *addr)
1197{
1198 /* Reliable end of stack detection:
1199 * Some APM bios versions misalign the stack
1200 */
1201 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1202}
1203#endif
1204
1205extern union thread_union init_thread_union;
1206extern struct task_struct init_task;
1207
1208extern struct mm_struct init_mm;
1209
1210#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1211extern struct task_struct *find_task_by_pid_type(int type, int pid);
1212extern void set_special_pids(pid_t session, pid_t pgrp);
1213extern void __set_special_pids(pid_t session, pid_t pgrp);
1214
1215/* per-UID process charging. */
1216extern struct user_struct * alloc_uid(uid_t);
1217static inline struct user_struct *get_uid(struct user_struct *u)
1218{
1219 atomic_inc(&u->__count);
1220 return u;
1221}
1222extern void free_uid(struct user_struct *);
1223extern void switch_uid(struct user_struct *);
1224
1225#include <asm/current.h>
1226
Atsushi Nemoto3171a032006-09-29 02:00:32 -07001227extern void do_timer(unsigned long ticks);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001228
1229extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1230extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1231extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1232 unsigned long clone_flags));
1233#ifdef CONFIG_SMP
1234 extern void kick_process(struct task_struct *tsk);
1235#else
1236 static inline void kick_process(struct task_struct *tsk) { }
1237#endif
Ingo Molnar36c8b582006-07-03 00:25:41 -07001238extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
1239extern void FASTCALL(sched_exit(struct task_struct * p));
Linus Torvalds1da177e2005-04-16 15:20:36 -07001240
1241extern int in_group_p(gid_t);
1242extern int in_egroup_p(gid_t);
1243
1244extern void proc_caches_init(void);
1245extern void flush_signals(struct task_struct *);
1246extern void flush_signal_handlers(struct task_struct *, int force_default);
1247extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1248
1249static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1250{
1251 unsigned long flags;
1252 int ret;
1253
1254 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1255 ret = dequeue_signal(tsk, mask, info);
1256 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1257
1258 return ret;
1259}
1260
1261extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1262 sigset_t *mask);
1263extern void unblock_all_signals(void);
1264extern void release_task(struct task_struct * p);
1265extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1266extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1267extern int force_sigsegv(int, struct task_struct *);
1268extern int force_sig_info(int, struct siginfo *, struct task_struct *);
Eric W. Biedermanc4b92fc2006-10-02 02:17:10 -07001269extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1270extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1271extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1272extern int kill_pgrp(struct pid *pid, int sig, int priv);
1273extern int kill_pid(struct pid *pid, int sig, int priv);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001274extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1275extern int kill_pg_info(int, struct siginfo *, pid_t);
1276extern int kill_proc_info(int, struct siginfo *, pid_t);
David Quigley8f95dc52006-06-30 01:55:47 -07001277extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t, u32);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001278extern void do_notify_parent(struct task_struct *, int);
1279extern void force_sig(int, struct task_struct *);
1280extern void force_sig_specific(int, struct task_struct *);
1281extern int send_sig(int, struct task_struct *, int);
1282extern void zap_other_threads(struct task_struct *p);
1283extern int kill_pg(pid_t, int, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001284extern int kill_proc(pid_t, int, int);
1285extern struct sigqueue *sigqueue_alloc(void);
1286extern void sigqueue_free(struct sigqueue *);
1287extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1288extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
Oleg Nesterov9ac95f22006-02-09 22:41:50 +03001289extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001290extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1291
1292/* These can be the second arg to send_sig_info/send_group_sig_info. */
1293#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1294#define SEND_SIG_PRIV ((struct siginfo *) 1)
1295#define SEND_SIG_FORCED ((struct siginfo *) 2)
1296
Oleg Nesterov621d3122005-10-30 15:03:45 -08001297static inline int is_si_special(const struct siginfo *info)
1298{
1299 return info <= SEND_SIG_FORCED;
1300}
1301
Linus Torvalds1da177e2005-04-16 15:20:36 -07001302/* True if we are on the alternate signal stack. */
1303
1304static inline int on_sig_stack(unsigned long sp)
1305{
1306 return (sp - current->sas_ss_sp < current->sas_ss_size);
1307}
1308
1309static inline int sas_ss_flags(unsigned long sp)
1310{
1311 return (current->sas_ss_size == 0 ? SS_DISABLE
1312 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1313}
1314
Linus Torvalds1da177e2005-04-16 15:20:36 -07001315/*
1316 * Routines for handling mm_structs
1317 */
1318extern struct mm_struct * mm_alloc(void);
1319
1320/* mmdrop drops the mm and the page tables */
1321extern void FASTCALL(__mmdrop(struct mm_struct *));
1322static inline void mmdrop(struct mm_struct * mm)
1323{
1324 if (atomic_dec_and_test(&mm->mm_count))
1325 __mmdrop(mm);
1326}
1327
1328/* mmput gets rid of the mappings and all user-space */
1329extern void mmput(struct mm_struct *);
1330/* Grab a reference to a task's mm, if it is not already going away */
1331extern struct mm_struct *get_task_mm(struct task_struct *task);
1332/* Remove the current tasks stale references to the old mm_struct */
1333extern void mm_release(struct task_struct *, struct mm_struct *);
1334
1335extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1336extern void flush_thread(void);
1337extern void exit_thread(void);
1338
Linus Torvalds1da177e2005-04-16 15:20:36 -07001339extern void exit_files(struct task_struct *);
Oleg Nesterov6b3934e2006-03-28 16:11:16 -08001340extern void __cleanup_signal(struct signal_struct *);
Oleg Nesterova7e53282006-03-28 16:11:27 -08001341extern void __cleanup_sighand(struct sighand_struct *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001342extern void exit_itimers(struct signal_struct *);
1343
1344extern NORET_TYPE void do_group_exit(int);
1345
Linus Torvalds1da177e2005-04-16 15:20:36 -07001346extern void daemonize(const char *, ...);
1347extern int allow_signal(int);
1348extern int disallow_signal(int);
Ingo Molnar36c8b582006-07-03 00:25:41 -07001349extern struct task_struct *child_reaper;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001350
1351extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1352extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
Ingo Molnar36c8b582006-07-03 00:25:41 -07001353struct task_struct *fork_idle(int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001354
1355extern void set_task_comm(struct task_struct *tsk, char *from);
1356extern void get_task_comm(char *to, struct task_struct *tsk);
1357
1358#ifdef CONFIG_SMP
Ingo Molnar36c8b582006-07-03 00:25:41 -07001359extern void wait_task_inactive(struct task_struct * p);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001360#else
1361#define wait_task_inactive(p) do { } while (0)
1362#endif
1363
1364#define remove_parent(p) list_del_init(&(p)->sibling)
Oleg Nesterov8fafabd2006-03-28 16:11:05 -08001365#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001366
Eric W. Biederman5e85d4a2006-04-18 22:20:16 -07001367#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001368
1369#define for_each_process(p) \
1370 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1371
1372/*
1373 * Careful: do_each_thread/while_each_thread is a double loop so
1374 * 'break' will not work as expected - use goto instead.
1375 */
1376#define do_each_thread(g, t) \
1377 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1378
1379#define while_each_thread(g, t) \
1380 while ((t = next_thread(t)) != g)
1381
Eric W. Biedermande12a782006-04-10 17:16:49 -06001382/* de_thread depends on thread_group_leader not being a pid based check */
1383#define thread_group_leader(p) (p == p->group_leader)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001384
Eric W. Biederman0804ef42006-10-02 02:17:04 -07001385/* Do to the insanities of de_thread it is possible for a process
1386 * to have the pid of the thread group leader without actually being
1387 * the thread group leader. For iteration through the pids in proc
1388 * all we care about is that we have a task with the appropriate
1389 * pid, we don't actually care if we have the right task.
1390 */
1391static inline int has_group_leader_pid(struct task_struct *p)
1392{
1393 return p->pid == p->tgid;
1394}
1395
Ingo Molnar36c8b582006-07-03 00:25:41 -07001396static inline struct task_struct *next_thread(const struct task_struct *p)
Oleg Nesterov47e65322006-03-28 16:11:25 -08001397{
1398 return list_entry(rcu_dereference(p->thread_group.next),
Ingo Molnar36c8b582006-07-03 00:25:41 -07001399 struct task_struct, thread_group);
Oleg Nesterov47e65322006-03-28 16:11:25 -08001400}
1401
Ingo Molnar36c8b582006-07-03 00:25:41 -07001402static inline int thread_group_empty(struct task_struct *p)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001403{
Oleg Nesterov47e65322006-03-28 16:11:25 -08001404 return list_empty(&p->thread_group);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001405}
1406
1407#define delay_group_leader(p) \
1408 (thread_group_leader(p) && !thread_group_empty(p))
1409
Linus Torvalds1da177e2005-04-16 15:20:36 -07001410/*
Eric W. Biederman260ea102006-06-23 02:05:18 -07001411 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
Jens Axboe22e2c502005-06-27 10:55:12 +02001412 * subscriptions and synchronises with wait4(). Also used in procfs. Also
Paul Jackson053199e2005-10-30 15:02:30 -08001413 * pins the final release of task.io_context. Also protects ->cpuset.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001414 *
1415 * Nests both inside and outside of read_lock(&tasklist_lock).
1416 * It must not be nested with write_lock_irq(&tasklist_lock),
1417 * neither inside nor outside.
1418 */
1419static inline void task_lock(struct task_struct *p)
1420{
1421 spin_lock(&p->alloc_lock);
1422}
1423
1424static inline void task_unlock(struct task_struct *p)
1425{
1426 spin_unlock(&p->alloc_lock);
1427}
1428
Oleg Nesterovf63ee722006-03-28 16:11:13 -08001429extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1430 unsigned long *flags);
1431
1432static inline void unlock_task_sighand(struct task_struct *tsk,
1433 unsigned long *flags)
1434{
1435 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1436}
1437
Al Virof0373602005-11-13 16:06:57 -08001438#ifndef __HAVE_THREAD_FUNCTIONS
1439
Al Viroa1261f52005-11-13 16:06:55 -08001440#define task_thread_info(task) (task)->thread_info
Al Viro9fc65872006-01-12 01:05:34 -08001441#define task_stack_page(task) ((void*)((task)->thread_info))
Al Viroa1261f52005-11-13 16:06:55 -08001442
Al Viro10ebffd2005-11-13 16:06:56 -08001443static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1444{
1445 *task_thread_info(p) = *task_thread_info(org);
1446 task_thread_info(p)->task = p;
1447}
1448
1449static inline unsigned long *end_of_stack(struct task_struct *p)
1450{
1451 return (unsigned long *)(p->thread_info + 1);
1452}
1453
Al Virof0373602005-11-13 16:06:57 -08001454#endif
1455
Linus Torvalds1da177e2005-04-16 15:20:36 -07001456/* set thread flags in other task's structures
1457 * - see asm/thread_info.h for TIF_xxxx flags available
1458 */
1459static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1460{
Al Viroa1261f52005-11-13 16:06:55 -08001461 set_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001462}
1463
1464static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1465{
Al Viroa1261f52005-11-13 16:06:55 -08001466 clear_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001467}
1468
1469static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1470{
Al Viroa1261f52005-11-13 16:06:55 -08001471 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001472}
1473
1474static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1475{
Al Viroa1261f52005-11-13 16:06:55 -08001476 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001477}
1478
1479static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1480{
Al Viroa1261f52005-11-13 16:06:55 -08001481 return test_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001482}
1483
1484static inline void set_tsk_need_resched(struct task_struct *tsk)
1485{
1486 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1487}
1488
1489static inline void clear_tsk_need_resched(struct task_struct *tsk)
1490{
1491 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1492}
1493
1494static inline int signal_pending(struct task_struct *p)
1495{
1496 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1497}
1498
1499static inline int need_resched(void)
1500{
1501 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1502}
1503
1504/*
1505 * cond_resched() and cond_resched_lock(): latency reduction via
1506 * explicit rescheduling in places that are safe. The return
1507 * value indicates whether a reschedule was done in fact.
1508 * cond_resched_lock() will drop the spinlock before scheduling,
1509 * cond_resched_softirq() will enable bhs before scheduling.
1510 */
1511extern int cond_resched(void);
1512extern int cond_resched_lock(spinlock_t * lock);
1513extern int cond_resched_softirq(void);
1514
1515/*
1516 * Does a critical section need to be broken due to another
1517 * task waiting?:
1518 */
1519#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1520# define need_lockbreak(lock) ((lock)->break_lock)
1521#else
1522# define need_lockbreak(lock) 0
1523#endif
1524
1525/*
1526 * Does a critical section need to be broken due to another
1527 * task waiting or preemption being signalled:
1528 */
1529static inline int lock_need_resched(spinlock_t *lock)
1530{
1531 if (need_lockbreak(lock) || need_resched())
1532 return 1;
1533 return 0;
1534}
1535
1536/* Reevaluate whether the task has signals pending delivery.
1537 This is required every time the blocked sigset_t changes.
1538 callers must hold sighand->siglock. */
1539
1540extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1541extern void recalc_sigpending(void);
1542
1543extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1544
1545/*
1546 * Wrappers for p->thread_info->cpu access. No-op on UP.
1547 */
1548#ifdef CONFIG_SMP
1549
1550static inline unsigned int task_cpu(const struct task_struct *p)
1551{
Al Viroa1261f52005-11-13 16:06:55 -08001552 return task_thread_info(p)->cpu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001553}
1554
1555static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1556{
Al Viroa1261f52005-11-13 16:06:55 -08001557 task_thread_info(p)->cpu = cpu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001558}
1559
1560#else
1561
1562static inline unsigned int task_cpu(const struct task_struct *p)
1563{
1564 return 0;
1565}
1566
1567static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1568{
1569}
1570
1571#endif /* CONFIG_SMP */
1572
1573#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1574extern void arch_pick_mmap_layout(struct mm_struct *mm);
1575#else
1576static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1577{
1578 mm->mmap_base = TASK_UNMAPPED_BASE;
1579 mm->get_unmapped_area = arch_get_unmapped_area;
1580 mm->unmap_area = arch_unmap_area;
1581}
1582#endif
1583
1584extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1585extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1586
Siddha, Suresh B5c45bf22006-06-27 02:54:42 -07001587#include <linux/sysdev.h>
1588extern int sched_mc_power_savings, sched_smt_power_savings;
1589extern struct sysdev_attribute attr_sched_mc_power_savings, attr_sched_smt_power_savings;
1590extern int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls);
1591
Linus Torvalds1da177e2005-04-16 15:20:36 -07001592extern void normalize_rt_tasks(void);
1593
Linus Torvalds1da177e2005-04-16 15:20:36 -07001594#ifdef CONFIG_PM
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001595/*
1596 * Check if a process has been frozen
1597 */
1598static inline int frozen(struct task_struct *p)
1599{
1600 return p->flags & PF_FROZEN;
1601}
1602
1603/*
1604 * Check if there is a request to freeze a process
1605 */
1606static inline int freezing(struct task_struct *p)
1607{
1608 return p->flags & PF_FREEZE;
1609}
1610
1611/*
1612 * Request that a process be frozen
1613 * FIXME: SMP problem. We may not modify other process' flags!
1614 */
1615static inline void freeze(struct task_struct *p)
1616{
1617 p->flags |= PF_FREEZE;
1618}
1619
1620/*
Rafael J. Wysockia7ef7872006-08-05 12:13:42 -07001621 * Sometimes we may need to cancel the previous 'freeze' request
1622 */
1623static inline void do_not_freeze(struct task_struct *p)
1624{
1625 p->flags &= ~PF_FREEZE;
1626}
1627
1628/*
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001629 * Wake up a frozen process
1630 */
1631static inline int thaw_process(struct task_struct *p)
1632{
1633 if (frozen(p)) {
1634 p->flags &= ~PF_FROZEN;
1635 wake_up_process(p);
1636 return 1;
1637 }
1638 return 0;
1639}
1640
1641/*
1642 * freezing is complete, mark process as frozen
1643 */
1644static inline void frozen_process(struct task_struct *p)
1645{
1646 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1647}
1648
1649extern void refrigerator(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001650extern int freeze_processes(void);
1651extern void thaw_processes(void);
1652
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001653static inline int try_to_freeze(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001654{
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001655 if (freezing(current)) {
1656 refrigerator();
Linus Torvalds1da177e2005-04-16 15:20:36 -07001657 return 1;
1658 } else
1659 return 0;
1660}
1661#else
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001662static inline int frozen(struct task_struct *p) { return 0; }
1663static inline int freezing(struct task_struct *p) { return 0; }
1664static inline void freeze(struct task_struct *p) { BUG(); }
1665static inline int thaw_process(struct task_struct *p) { return 1; }
1666static inline void frozen_process(struct task_struct *p) { BUG(); }
1667
1668static inline void refrigerator(void) {}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001669static inline int freeze_processes(void) { BUG(); return 0; }
1670static inline void thaw_processes(void) {}
1671
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001672static inline int try_to_freeze(void) { return 0; }
1673
Linus Torvalds1da177e2005-04-16 15:20:36 -07001674#endif /* CONFIG_PM */
1675#endif /* __KERNEL__ */
1676
1677#endif