blob: 8d11d9310db0a0fb14e042f94c1b3335a20a4c63 [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>
Linus Torvalds1da177e2005-04-16 15:20:36 -070076
David Woodhousea3b67142006-04-25 14:54:40 +010077#include <linux/time.h>
78#include <linux/param.h>
79#include <linux/resource.h>
80#include <linux/timer.h>
81#include <linux/hrtimer.h>
82
83#include <asm/processor.h>
H. J. Lu36d57ac2005-09-06 15:16:49 -070084
Linus Torvalds1da177e2005-04-16 15:20:36 -070085struct exec_domain;
86
87/*
Linus Torvalds1da177e2005-04-16 15:20:36 -070088 * List of flags we want to share for kernel threads,
89 * if only because they are not used by them anyway.
90 */
91#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
92
93/*
94 * These are the constant used to fake the fixed-point load-average
95 * counting. Some notes:
96 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
97 * a load-average precision of 10 bits integer + 11 bits fractional
98 * - if you want to count load-averages more often, you need more
99 * precision, or rounding will get you. With 2-second counting freq,
100 * the EXP_n values would be 1981, 2034 and 2043 if still using only
101 * 11 bit fractions.
102 */
103extern unsigned long avenrun[]; /* Load averages */
104
105#define FSHIFT 11 /* nr of bits of precision */
106#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
107#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
108#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
109#define EXP_5 2014 /* 1/exp(5sec/5min) */
110#define EXP_15 2037 /* 1/exp(5sec/15min) */
111
112#define CALC_LOAD(load,exp,n) \
113 load *= exp; \
114 load += n*(FIXED_1-exp); \
115 load >>= FSHIFT;
116
117extern unsigned long total_forks;
118extern int nr_threads;
119extern int last_pid;
120DECLARE_PER_CPU(unsigned long, process_counts);
121extern int nr_processes(void);
122extern unsigned long nr_running(void);
123extern unsigned long nr_uninterruptible(void);
Jack Steinerdb1b1fe2006-03-31 02:31:21 -0800124extern unsigned long nr_active(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700125extern unsigned long nr_iowait(void);
126
Linus Torvalds1da177e2005-04-16 15:20:36 -0700127
Linus Torvalds4a8342d2005-09-29 15:18:21 -0700128/*
129 * Task state bitmask. NOTE! These bits are also
130 * encoded in fs/proc/array.c: get_task_state().
131 *
132 * We have two separate sets of flags: task->state
133 * is about runnability, while task->exit_state are
134 * about the task exiting. Confusing, but this way
135 * modifying one set can't modify the other one by
136 * mistake.
137 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700138#define TASK_RUNNING 0
139#define TASK_INTERRUPTIBLE 1
140#define TASK_UNINTERRUPTIBLE 2
Linus Torvalds4a8342d2005-09-29 15:18:21 -0700141#define TASK_STOPPED 4
142#define TASK_TRACED 8
143/* in tsk->exit_state */
144#define EXIT_ZOMBIE 16
145#define EXIT_DEAD 32
146/* in tsk->state again */
147#define TASK_NONINTERACTIVE 64
Linus Torvalds1da177e2005-04-16 15:20:36 -0700148
149#define __set_task_state(tsk, state_value) \
150 do { (tsk)->state = (state_value); } while (0)
151#define set_task_state(tsk, state_value) \
152 set_mb((tsk)->state, (state_value))
153
Andrew Morton498d0c52005-09-13 01:25:14 -0700154/*
155 * set_current_state() includes a barrier so that the write of current->state
156 * is correctly serialised wrt the caller's subsequent test of whether to
157 * actually sleep:
158 *
159 * set_current_state(TASK_UNINTERRUPTIBLE);
160 * if (do_i_need_to_sleep())
161 * schedule();
162 *
163 * If the caller does not need such serialisation then use __set_current_state()
164 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700165#define __set_current_state(state_value) \
166 do { current->state = (state_value); } while (0)
167#define set_current_state(state_value) \
168 set_mb(current->state, (state_value))
169
170/* Task command name length */
171#define TASK_COMM_LEN 16
172
Linus Torvalds1da177e2005-04-16 15:20:36 -0700173#include <linux/spinlock.h>
174
175/*
176 * This serializes "schedule()" and also protects
177 * the run-queue from deletions/modifications (but
178 * _adding_ to the beginning of the run-queue has
179 * a separate lock).
180 */
181extern rwlock_t tasklist_lock;
182extern spinlock_t mmlist_lock;
183
184typedef struct task_struct task_t;
185
186extern void sched_init(void);
187extern void sched_init_smp(void);
188extern void init_idle(task_t *idle, int cpu);
189
190extern cpumask_t nohz_cpu_mask;
191
192extern void show_state(void);
193extern void show_regs(struct pt_regs *);
194
195/*
196 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
197 * task), SP is the stack pointer of the first frame that should be shown in the back
198 * trace (or NULL if the entire call-chain of the task should be shown).
199 */
200extern void show_stack(struct task_struct *task, unsigned long *sp);
201
202void io_schedule(void);
203long io_schedule_timeout(long timeout);
204
205extern void cpu_init (void);
206extern void trap_init(void);
207extern void update_process_times(int user);
208extern void scheduler_tick(void);
209
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700210#ifdef CONFIG_DETECT_SOFTLOCKUP
Ingo Molnar6687a972006-03-24 03:18:41 -0800211extern void softlockup_tick(void);
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700212extern void spawn_softlockup_task(void);
213extern void touch_softlockup_watchdog(void);
214#else
Ingo Molnar6687a972006-03-24 03:18:41 -0800215static inline void softlockup_tick(void)
Ingo Molnar8446f1d2005-09-06 15:16:27 -0700216{
217}
218static inline void spawn_softlockup_task(void)
219{
220}
221static inline void touch_softlockup_watchdog(void)
222{
223}
224#endif
225
226
Linus Torvalds1da177e2005-04-16 15:20:36 -0700227/* Attach to any functions which should be ignored in wchan output. */
228#define __sched __attribute__((__section__(".sched.text")))
229/* Is this address in the __sched functions? */
230extern int in_sched_functions(unsigned long addr);
231
232#define MAX_SCHEDULE_TIMEOUT LONG_MAX
233extern signed long FASTCALL(schedule_timeout(signed long timeout));
Nishanth Aravamudan64ed93a2005-09-10 00:27:21 -0700234extern signed long schedule_timeout_interruptible(signed long timeout);
235extern signed long schedule_timeout_uninterruptible(signed long timeout);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700236asmlinkage void schedule(void);
237
238struct namespace;
239
240/* Maximum number of active map areas.. This is a random (large) number */
241#define DEFAULT_MAX_MAP_COUNT 65536
242
243extern int sysctl_max_map_count;
244
245#include <linux/aio.h>
246
247extern unsigned long
248arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
249 unsigned long, unsigned long);
250extern unsigned long
251arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
252 unsigned long len, unsigned long pgoff,
253 unsigned long flags);
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700254extern void arch_unmap_area(struct mm_struct *, unsigned long);
255extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700257#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
258/*
259 * The mm counters are not protected by its page_table_lock,
260 * so must be incremented atomically.
261 */
Christoph Lameterd3cb4872006-01-06 00:11:20 -0800262#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
263#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
264#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
265#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
266#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
267typedef atomic_long_t mm_counter_t;
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700268
269#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
270/*
271 * The mm counters are protected by its page_table_lock,
272 * so can be incremented directly.
273 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700274#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
275#define get_mm_counter(mm, member) ((mm)->_##member)
276#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
277#define inc_mm_counter(mm, member) (mm)->_##member++
278#define dec_mm_counter(mm, member) (mm)->_##member--
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700279typedef unsigned long mm_counter_t;
Hugh Dickins42946212005-10-29 18:16:05 -0700280
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700281#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
282
283#define get_mm_rss(mm) \
284 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
Hugh Dickins365e9c872005-10-29 18:16:18 -0700285#define update_hiwater_rss(mm) do { \
286 unsigned long _rss = get_mm_rss(mm); \
287 if ((mm)->hiwater_rss < _rss) \
288 (mm)->hiwater_rss = _rss; \
289} while (0)
290#define update_hiwater_vm(mm) do { \
291 if ((mm)->hiwater_vm < (mm)->total_vm) \
292 (mm)->hiwater_vm = (mm)->total_vm; \
293} while (0)
294
Linus Torvalds1da177e2005-04-16 15:20:36 -0700295struct mm_struct {
296 struct vm_area_struct * mmap; /* list of VMAs */
297 struct rb_root mm_rb;
298 struct vm_area_struct * mmap_cache; /* last find_vma result */
299 unsigned long (*get_unmapped_area) (struct file *filp,
300 unsigned long addr, unsigned long len,
301 unsigned long pgoff, unsigned long flags);
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700302 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
Benjamin Herrenschmidt0551fbd2006-02-28 16:59:19 -0800303 unsigned long mmap_base; /* base of mmap area */
304 unsigned long task_size; /* size of task vm space */
305 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
Wolfgang Wander1363c3c2005-06-21 17:14:49 -0700306 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307 pgd_t * pgd;
308 atomic_t mm_users; /* How many users with user space? */
309 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
310 int map_count; /* number of VMAs */
311 struct rw_semaphore mmap_sem;
312 spinlock_t page_table_lock; /* Protects page tables and some counters */
313
314 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
315 * together off init_mm.mmlist, and are protected
316 * by mmlist_lock
317 */
318
Hugh Dickinsf412ac02005-10-29 18:16:41 -0700319 /* Special counters, in some configurations protected by the
320 * page_table_lock, in other configurations by being atomic.
321 */
Hugh Dickins42946212005-10-29 18:16:05 -0700322 mm_counter_t _file_rss;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700323 mm_counter_t _anon_rss;
324
Hugh Dickinsf4499522005-10-29 18:16:19 -0700325 unsigned long hiwater_rss; /* High-watermark of RSS usage */
326 unsigned long hiwater_vm; /* High-water virtual memory usage */
327
328 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
329 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
330 unsigned long start_code, end_code, start_data, end_data;
331 unsigned long start_brk, brk, start_stack;
332 unsigned long arg_start, arg_end, env_start, env_end;
333
H. J. Lu36d57ac2005-09-06 15:16:49 -0700334 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700335
Alan Coxd6e71142005-06-23 00:09:43 -0700336 unsigned dumpable:2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700337 cpumask_t cpu_vm_mask;
338
339 /* Architecture-specific MM context */
340 mm_context_t context;
341
342 /* Token based thrashing protection. */
343 unsigned long swap_token_time;
344 char recent_pagein;
345
346 /* coredumping support */
347 int core_waiters;
348 struct completion *core_startup_done, core_done;
349
350 /* aio bits */
351 rwlock_t ioctx_list_lock;
352 struct kioctx *ioctx_list;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700353};
354
355struct sighand_struct {
356 atomic_t count;
357 struct k_sigaction action[_NSIG];
358 spinlock_t siglock;
359};
360
KaiGai Kohei0e464812006-06-25 05:49:24 -0700361struct pacct_struct {
KaiGai Koheif6ec29a2006-06-25 05:49:25 -0700362 int ac_flag;
363 long ac_exitcode;
KaiGai Kohei0e464812006-06-25 05:49:24 -0700364 unsigned long ac_mem;
KaiGai Kohei77787bf2006-06-25 05:49:26 -0700365 cputime_t ac_utime, ac_stime;
366 unsigned long ac_minflt, ac_majflt;
KaiGai Kohei0e464812006-06-25 05:49:24 -0700367};
368
Linus Torvalds1da177e2005-04-16 15:20:36 -0700369/*
370 * NOTE! "signal_struct" does not have it's own
371 * locking, because a shared signal_struct always
372 * implies a shared sighand_struct, so locking
373 * sighand_struct is always a proper superset of
374 * the locking of signal_struct.
375 */
376struct signal_struct {
377 atomic_t count;
378 atomic_t live;
379
380 wait_queue_head_t wait_chldexit; /* for wait4() */
381
382 /* current thread group signal load-balancing target: */
383 task_t *curr_target;
384
385 /* shared signal handling: */
386 struct sigpending shared_pending;
387
388 /* thread group exit support */
389 int group_exit_code;
390 /* overloaded:
391 * - notify group_exit_task when ->count is equal to notify_count
392 * - everyone except group_exit_task is stopped during signal delivery
393 * of fatal signals, group_exit_task processes the signal.
394 */
395 struct task_struct *group_exit_task;
396 int notify_count;
397
398 /* thread group stop support, overloads group_exit_code too */
399 int group_stop_count;
400 unsigned int flags; /* see SIGNAL_* flags below */
401
402 /* POSIX.1b Interval Timers */
403 struct list_head posix_timers;
404
405 /* ITIMER_REAL timer for the process */
Thomas Gleixner2ff678b2006-01-09 20:52:34 -0800406 struct hrtimer real_timer;
Roman Zippel05cfb612006-03-26 01:38:12 -0800407 struct task_struct *tsk;
Thomas Gleixner2ff678b2006-01-09 20:52:34 -0800408 ktime_t it_real_incr;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700409
410 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
411 cputime_t it_prof_expires, it_virt_expires;
412 cputime_t it_prof_incr, it_virt_incr;
413
414 /* job control IDs */
415 pid_t pgrp;
416 pid_t tty_old_pgrp;
417 pid_t session;
418 /* boolean value for session group leader */
419 int leader;
420
421 struct tty_struct *tty; /* NULL if no tty */
422
423 /*
424 * Cumulative resource counters for dead threads in the group,
425 * and for reaped dead child processes forked by this group.
426 * Live threads maintain their own counters and add to these
427 * in __exit_signal, except for the group leader.
428 */
429 cputime_t utime, stime, cutime, cstime;
430 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
431 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
432
433 /*
434 * Cumulative ns of scheduled CPU time for dead threads in the
435 * group, not including a zombie group leader. (This only differs
436 * from jiffies_to_ns(utime + stime) if sched_clock uses something
437 * other than jiffies.)
438 */
439 unsigned long long sched_time;
440
441 /*
442 * We don't bother to synchronize most readers of this at all,
443 * because there is no reader checking a limit that actually needs
444 * to get both rlim_cur and rlim_max atomically, and either one
445 * alone is a single word that can safely be read normally.
446 * getrlimit/setrlimit use task_lock(current->group_leader) to
447 * protect this instead of the siglock, because they really
448 * have no need to disable irqs.
449 */
450 struct rlimit rlim[RLIM_NLIMITS];
451
452 struct list_head cpu_timers[3];
453
454 /* keep the process-shared keyrings here so that they do the right
455 * thing in threads created with CLONE_THREAD */
456#ifdef CONFIG_KEYS
457 struct key *session_keyring; /* keyring inherited over fork */
458 struct key *process_keyring; /* keyring private to this process */
459#endif
KaiGai Kohei0e464812006-06-25 05:49:24 -0700460#ifdef CONFIG_BSD_PROCESS_ACCT
461 struct pacct_struct pacct; /* per-process accounting information */
462#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700463};
464
Nick Piggin4866cde2005-06-25 14:57:23 -0700465/* Context switch must be unlocked if interrupts are to be enabled */
466#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
467# define __ARCH_WANT_UNLOCKED_CTXSW
468#endif
469
Linus Torvalds1da177e2005-04-16 15:20:36 -0700470/*
471 * Bits in flags field of signal_struct.
472 */
473#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
474#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
475#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
476#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
477
478
479/*
480 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
Ingo Molnarb0a94992006-01-14 13:20:41 -0800481 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
482 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
483 * values are inverted: lower p->prio value means higher priority.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700484 *
485 * The MAX_USER_RT_PRIO value allows the actual maximum
486 * RT priority to be separate from the value exported to
487 * user-space. This allows kernel threads to set their
488 * priority to a value higher than any user task. Note:
489 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
490 */
491
492#define MAX_USER_RT_PRIO 100
493#define MAX_RT_PRIO MAX_USER_RT_PRIO
494
495#define MAX_PRIO (MAX_RT_PRIO + 40)
496
497#define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
Con Kolivasd425b272006-03-31 02:31:29 -0800498#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700499
500/*
501 * Some day this will be a full-fledged user tracking system..
502 */
503struct user_struct {
504 atomic_t __count; /* reference count */
505 atomic_t processes; /* How many processes does this user have? */
506 atomic_t files; /* How many open files does this user have? */
507 atomic_t sigpending; /* How many pending signals does this user have? */
Amy Griffis2d9048e2006-06-01 13:10:59 -0700508#ifdef CONFIG_INOTIFY_USER
Robert Love0eeca282005-07-12 17:06:03 -0400509 atomic_t inotify_watches; /* How many inotify watches does this user have? */
510 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
511#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700512 /* protected by mq_lock */
513 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
514 unsigned long locked_shm; /* How many pages of mlocked shm ? */
515
516#ifdef CONFIG_KEYS
517 struct key *uid_keyring; /* UID specific keyring */
518 struct key *session_keyring; /* UID's default session keyring */
519#endif
520
521 /* Hash table maintenance information */
522 struct list_head uidhash_list;
523 uid_t uid;
524};
525
526extern struct user_struct *find_user(uid_t);
527
528extern struct user_struct root_user;
529#define INIT_USER (&root_user)
530
531typedef struct prio_array prio_array_t;
532struct backing_dev_info;
533struct reclaim_state;
534
535#ifdef CONFIG_SCHEDSTATS
536struct sched_info {
537 /* cumulative counters */
538 unsigned long cpu_time, /* time spent on the cpu */
539 run_delay, /* time spent waiting on a runqueue */
540 pcnt; /* # of timeslices run on this cpu */
541
542 /* timestamps */
543 unsigned long last_arrival, /* when we last ran on a cpu */
544 last_queued; /* when we were last queued to run */
545};
546
547extern struct file_operations proc_schedstat_operations;
548#endif
549
550enum idle_type
551{
552 SCHED_IDLE,
553 NOT_IDLE,
554 NEWLY_IDLE,
555 MAX_IDLE_TYPES
556};
557
558/*
559 * sched-domains (multiprocessor balancing) declarations:
560 */
561#ifdef CONFIG_SMP
562#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
563
564#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
565#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
566#define SD_BALANCE_EXEC 4 /* Balance on exec */
Nick Piggin147cbb42005-06-25 14:57:19 -0700567#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
568#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
569#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
570#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
571#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700572
573struct sched_group {
574 struct sched_group *next; /* Must be a circular list */
575 cpumask_t cpumask;
576
577 /*
578 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
579 * single CPU. This is read only (except for setup, hotplug CPU).
580 */
581 unsigned long cpu_power;
582};
583
584struct sched_domain {
585 /* These fields must be setup */
586 struct sched_domain *parent; /* top domain must be null terminated */
587 struct sched_group *groups; /* the balancing groups of the domain */
588 cpumask_t span; /* span of all CPUs in this domain */
589 unsigned long min_interval; /* Minimum balance interval ms */
590 unsigned long max_interval; /* Maximum balance interval ms */
591 unsigned int busy_factor; /* less balancing by factor if busy */
592 unsigned int imbalance_pct; /* No balance until over watermark */
593 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
594 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
595 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
Nick Piggin78979862005-06-25 14:57:13 -0700596 unsigned int busy_idx;
597 unsigned int idle_idx;
598 unsigned int newidle_idx;
599 unsigned int wake_idx;
Nick Piggin147cbb42005-06-25 14:57:19 -0700600 unsigned int forkexec_idx;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700601 int flags; /* See SD_* */
602
603 /* Runtime fields. */
604 unsigned long last_balance; /* init to jiffies. units in jiffies */
605 unsigned int balance_interval; /* initialise to 1. units in ms. */
606 unsigned int nr_balance_failed; /* initialise to 0 */
607
608#ifdef CONFIG_SCHEDSTATS
609 /* load_balance() stats */
610 unsigned long lb_cnt[MAX_IDLE_TYPES];
611 unsigned long lb_failed[MAX_IDLE_TYPES];
612 unsigned long lb_balanced[MAX_IDLE_TYPES];
613 unsigned long lb_imbalance[MAX_IDLE_TYPES];
614 unsigned long lb_gained[MAX_IDLE_TYPES];
615 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
616 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
617 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
618
619 /* Active load balancing */
620 unsigned long alb_cnt;
621 unsigned long alb_failed;
622 unsigned long alb_pushed;
623
Nick Piggin68767a02005-06-25 14:57:20 -0700624 /* SD_BALANCE_EXEC stats */
625 unsigned long sbe_cnt;
626 unsigned long sbe_balanced;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700627 unsigned long sbe_pushed;
628
Nick Piggin68767a02005-06-25 14:57:20 -0700629 /* SD_BALANCE_FORK stats */
630 unsigned long sbf_cnt;
631 unsigned long sbf_balanced;
632 unsigned long sbf_pushed;
633
Linus Torvalds1da177e2005-04-16 15:20:36 -0700634 /* try_to_wake_up() stats */
635 unsigned long ttwu_wake_remote;
636 unsigned long ttwu_move_affine;
637 unsigned long ttwu_move_balance;
638#endif
639};
640
Dinakar Guniguntala1a20ff22005-06-25 14:57:33 -0700641extern void partition_sched_domains(cpumask_t *partition1,
642 cpumask_t *partition2);
akpm@osdl.org198e2f12006-01-12 01:05:30 -0800643
644/*
645 * Maximum cache size the migration-costs auto-tuning code will
646 * search from:
647 */
648extern unsigned int max_cache_size;
649
650#endif /* CONFIG_SMP */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700651
652
653struct io_context; /* See blkdev.h */
654void exit_io_context(void);
655struct cpuset;
656
657#define NGROUPS_SMALL 32
658#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
659struct group_info {
660 int ngroups;
661 atomic_t usage;
662 gid_t small_block[NGROUPS_SMALL];
663 int nblocks;
664 gid_t *blocks[0];
665};
666
667/*
668 * get_group_info() must be called with the owning task locked (via task_lock())
669 * when task != current. The reason being that the vast majority of callers are
670 * looking at current->group_info, which can not be changed except by the
671 * current task. Changing current->group_info requires the task lock, too.
672 */
673#define get_group_info(group_info) do { \
674 atomic_inc(&(group_info)->usage); \
675} while (0)
676
677#define put_group_info(group_info) do { \
678 if (atomic_dec_and_test(&(group_info)->usage)) \
679 groups_free(group_info); \
680} while (0)
681
David Howells3e301482005-06-23 22:00:56 -0700682extern struct group_info *groups_alloc(int gidsetsize);
683extern void groups_free(struct group_info *group_info);
684extern int set_current_groups(struct group_info *group_info);
685extern int groups_search(struct group_info *group_info, gid_t grp);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700686/* access the groups "array" with this macro */
687#define GROUP_AT(gi, i) \
688 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
689
Chen, Kenneth W383f2832005-09-09 13:02:02 -0700690#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
691extern void prefetch_stack(struct task_struct*);
692#else
693static inline void prefetch_stack(struct task_struct *t) { }
694#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700695
696struct audit_context; /* See audit.c */
697struct mempolicy;
Jens Axboeb92ce552006-04-11 13:52:07 +0200698struct pipe_inode_info;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700699
Con Kolivas3dee3862006-03-31 02:31:23 -0800700enum sleep_type {
701 SLEEP_NORMAL,
702 SLEEP_NONINTERACTIVE,
703 SLEEP_INTERACTIVE,
704 SLEEP_INTERRUPTED,
705};
706
Linus Torvalds1da177e2005-04-16 15:20:36 -0700707struct task_struct {
708 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
709 struct thread_info *thread_info;
710 atomic_t usage;
711 unsigned long flags; /* per process flags, defined below */
712 unsigned long ptrace;
713
Paolo 'Blaisorblade' Giarrusso36772092005-05-05 16:16:12 -0700714 int lock_depth; /* BKL lock depth */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700715
Chen, Kenneth Wd6077cb2006-02-14 13:53:10 -0800716#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
Nick Piggin4866cde2005-06-25 14:57:23 -0700717 int oncpu;
718#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700719 int prio, static_prio;
720 struct list_head run_list;
721 prio_array_t *array;
722
Jens Axboe22e2c502005-06-27 10:55:12 +0200723 unsigned short ioprio;
Jens Axboe2056a782006-03-23 20:00:26 +0100724 unsigned int btrace_seq;
Jens Axboe22e2c502005-06-27 10:55:12 +0200725
Linus Torvalds1da177e2005-04-16 15:20:36 -0700726 unsigned long sleep_avg;
727 unsigned long long timestamp, last_ran;
728 unsigned long long sched_time; /* sched_clock time spent running */
Con Kolivas3dee3862006-03-31 02:31:23 -0800729 enum sleep_type sleep_type;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700730
731 unsigned long policy;
732 cpumask_t cpus_allowed;
733 unsigned int time_slice, first_time_slice;
734
735#ifdef CONFIG_SCHEDSTATS
736 struct sched_info sched_info;
737#endif
738
739 struct list_head tasks;
740 /*
741 * ptrace_list/ptrace_children forms the list of my children
742 * that were stolen by a ptracer.
743 */
744 struct list_head ptrace_children;
745 struct list_head ptrace_list;
746
747 struct mm_struct *mm, *active_mm;
748
749/* task state */
750 struct linux_binfmt *binfmt;
751 long exit_state;
752 int exit_code, exit_signal;
753 int pdeath_signal; /* The signal sent when the parent dies */
754 /* ??? */
755 unsigned long personality;
756 unsigned did_exec:1;
757 pid_t pid;
758 pid_t tgid;
759 /*
760 * pointers to (original) parent process, youngest child, younger sibling,
761 * older sibling, respectively. (p->father can be replaced with
762 * p->parent->pid)
763 */
764 struct task_struct *real_parent; /* real parent process (when being debugged) */
765 struct task_struct *parent; /* parent process */
766 /*
767 * children/sibling forms the list of my children plus the
768 * tasks I'm ptracing.
769 */
770 struct list_head children; /* list of my children */
771 struct list_head sibling; /* linkage in my parent's children list */
772 struct task_struct *group_leader; /* threadgroup leader */
773
774 /* PID/PID hash table linkage. */
Eric W. Biederman92476d72006-03-31 02:31:42 -0800775 struct pid_link pids[PIDTYPE_MAX];
Oleg Nesterov47e65322006-03-28 16:11:25 -0800776 struct list_head thread_group;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700777
778 struct completion *vfork_done; /* for vfork() */
779 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
780 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
781
782 unsigned long rt_priority;
783 cputime_t utime, stime;
784 unsigned long nvcsw, nivcsw; /* context switch counts */
785 struct timespec start_time;
786/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
787 unsigned long min_flt, maj_flt;
788
789 cputime_t it_prof_expires, it_virt_expires;
790 unsigned long long it_sched_expires;
791 struct list_head cpu_timers[3];
792
793/* process credentials */
794 uid_t uid,euid,suid,fsuid;
795 gid_t gid,egid,sgid,fsgid;
796 struct group_info *group_info;
797 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
798 unsigned keep_capabilities:1;
799 struct user_struct *user;
800#ifdef CONFIG_KEYS
David Howellsb5f545c2006-01-08 01:02:47 -0800801 struct key *request_key_auth; /* assumed request_key authority */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700802 struct key *thread_keyring; /* keyring private to this thread */
David Howells3e301482005-06-23 22:00:56 -0700803 unsigned char jit_keyring; /* default keyring to attach requested keys to */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700804#endif
805 int oomkilladj; /* OOM kill score adjustment (bit shift). */
Paolo 'Blaisorblade' Giarrusso36772092005-05-05 16:16:12 -0700806 char comm[TASK_COMM_LEN]; /* executable name excluding path
807 - access with [gs]et_task_comm (which lock
808 it with task_lock())
809 - initialized normally by flush_old_exec */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700810/* file system info */
811 int link_count, total_link_count;
812/* ipc stuff */
813 struct sysv_sem sysvsem;
814/* CPU-specific state of this task */
815 struct thread_struct thread;
816/* filesystem information */
817 struct fs_struct *fs;
818/* open file information */
819 struct files_struct *files;
820/* namespace */
821 struct namespace *namespace;
822/* signal handlers */
823 struct signal_struct *signal;
824 struct sighand_struct *sighand;
825
826 sigset_t blocked, real_blocked;
David Woodhouse150256d2006-01-18 17:43:57 -0800827 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700828 struct sigpending pending;
829
830 unsigned long sas_ss_sp;
831 size_t sas_ss_size;
832 int (*notifier)(void *priv);
833 void *notifier_data;
834 sigset_t *notifier_mask;
835
836 void *security;
837 struct audit_context *audit_context;
838 seccomp_t seccomp;
839
840/* Thread group tracking */
841 u32 parent_exec_id;
842 u32 self_exec_id;
843/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
844 spinlock_t alloc_lock;
845/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
846 spinlock_t proc_lock;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700847
Ingo Molnar408894e2006-01-09 15:59:20 -0800848#ifdef CONFIG_DEBUG_MUTEXES
849 /* mutex deadlock detection */
850 struct mutex_waiter *blocked_on;
851#endif
852
Linus Torvalds1da177e2005-04-16 15:20:36 -0700853/* journalling filesystem info */
854 void *journal_info;
855
856/* VM state */
857 struct reclaim_state *reclaim_state;
858
859 struct dentry *proc_dentry;
860 struct backing_dev_info *backing_dev_info;
861
862 struct io_context *io_context;
863
864 unsigned long ptrace_message;
865 siginfo_t *last_siginfo; /* For ptrace use. */
866/*
867 * current io wait handle: wait queue entry to use for io waits
868 * If this thread is processing aio, this points at the waitqueue
869 * inside the currently handled kiocb. It may be NULL (i.e. default
870 * to a stack based synchronous wait) if its doing sync IO.
871 */
872 wait_queue_t *io_wait;
873/* i/o counters(bytes read/written, #syscalls */
874 u64 rchar, wchar, syscr, syscw;
875#if defined(CONFIG_BSD_PROCESS_ACCT)
876 u64 acct_rss_mem1; /* accumulated rss usage */
877 u64 acct_vm_mem1; /* accumulated virtual memory usage */
878 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
879#endif
880#ifdef CONFIG_NUMA
881 struct mempolicy *mempolicy;
882 short il_next;
883#endif
884#ifdef CONFIG_CPUSETS
885 struct cpuset *cpuset;
886 nodemask_t mems_allowed;
887 int cpuset_mems_generation;
Paul Jackson825a46a2006-03-24 03:16:03 -0800888 int cpuset_mem_spread_rotor;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700889#endif
Ingo Molnar0771dfe2006-03-27 01:16:22 -0800890 struct robust_list_head __user *robust_list;
Ingo Molnar34f192c2006-03-27 01:16:24 -0800891#ifdef CONFIG_COMPAT
892 struct compat_robust_list_head __user *compat_robust_list;
893#endif
Ingo Molnar0771dfe2006-03-27 01:16:22 -0800894
Jens Axboe22e2c502005-06-27 10:55:12 +0200895 atomic_t fs_excl; /* holding fs exclusive resources */
Ingo Molnare56d0902006-01-08 01:01:37 -0800896 struct rcu_head rcu;
Jens Axboeb92ce552006-04-11 13:52:07 +0200897
898 /*
899 * cache last used pipe for splice
900 */
901 struct pipe_inode_info *splice_pipe;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700902};
903
904static inline pid_t process_group(struct task_struct *tsk)
905{
906 return tsk->signal->pgrp;
907}
908
909/**
910 * pid_alive - check that a task structure is not stale
911 * @p: Task structure to be checked.
912 *
913 * Test if a process is not yet dead (at most zombie state)
914 * If pid_alive fails, then pointers within the task structure
915 * can be stale and must not be dereferenced.
916 */
917static inline int pid_alive(struct task_struct *p)
918{
Eric W. Biederman92476d72006-03-31 02:31:42 -0800919 return p->pids[PIDTYPE_PID].pid != NULL;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700920}
921
922extern void free_task(struct task_struct *tsk);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700923#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
Ingo Molnare56d0902006-01-08 01:01:37 -0800924
Andrew Morton158d9eb2006-03-31 02:31:34 -0800925extern void __put_task_struct(struct task_struct *t);
Ingo Molnare56d0902006-01-08 01:01:37 -0800926
927static inline void put_task_struct(struct task_struct *t)
928{
929 if (atomic_dec_and_test(&t->usage))
Eric W. Biederman8c7904a2006-03-31 02:31:37 -0800930 __put_task_struct(t);
Ingo Molnare56d0902006-01-08 01:01:37 -0800931}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700932
933/*
934 * Per process flags
935 */
936#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
937 /* Not implemented yet, only for 486*/
938#define PF_STARTING 0x00000002 /* being created */
939#define PF_EXITING 0x00000004 /* getting shut down */
940#define PF_DEAD 0x00000008 /* Dead */
941#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
942#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
943#define PF_DUMPCORE 0x00000200 /* dumped core */
944#define PF_SIGNALED 0x00000400 /* killed by a signal */
945#define PF_MEMALLOC 0x00000800 /* Allocating memory */
946#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
947#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
948#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
949#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
950#define PF_FROZEN 0x00010000 /* frozen for system suspend */
951#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
952#define PF_KSWAPD 0x00040000 /* I am kswapd */
953#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
954#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
Jens Axboeb31dc662006-06-13 08:26:10 +0200955#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
956#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
957#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
958#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
959#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
Paul Jacksonc61afb12006-03-24 03:16:08 -0800960#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700961
962/*
963 * Only the _current_ task can read/write to tsk->flags, but other
964 * tasks can access tsk->flags in readonly mode for example
965 * with tsk_used_math (like during threaded core dumping).
966 * There is however an exception to this rule during ptrace
967 * or during fork: the ptracer task is allowed to write to the
968 * child->flags of its traced child (same goes for fork, the parent
969 * can write to the child->flags), because we're guaranteed the
970 * child is not running and in turn not changing child->flags
971 * at the same time the parent does it.
972 */
973#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
974#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
975#define clear_used_math() clear_stopped_child_used_math(current)
976#define set_used_math() set_stopped_child_used_math(current)
977#define conditional_stopped_child_used_math(condition, child) \
978 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
979#define conditional_used_math(condition) \
980 conditional_stopped_child_used_math(condition, current)
981#define copy_to_stopped_child_used_math(child) \
982 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
983/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
984#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
985#define used_math() tsk_used_math(current)
986
987#ifdef CONFIG_SMP
988extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
989#else
990static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
991{
Paul Jackson4098f992005-10-30 15:03:21 -0800992 if (!cpu_isset(0, new_mask))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700993 return -EINVAL;
994 return 0;
995}
996#endif
997
998extern unsigned long long sched_clock(void);
999extern unsigned long long current_sched_time(const task_t *current_task);
1000
1001/* sched_exec is called by processes performing an exec */
1002#ifdef CONFIG_SMP
1003extern void sched_exec(void);
1004#else
1005#define sched_exec() {}
1006#endif
1007
1008#ifdef CONFIG_HOTPLUG_CPU
1009extern void idle_task_exit(void);
1010#else
1011static inline void idle_task_exit(void) {}
1012#endif
1013
1014extern void sched_idle_next(void);
1015extern void set_user_nice(task_t *p, long nice);
1016extern int task_prio(const task_t *p);
1017extern int task_nice(const task_t *p);
Matt Mackalle43379f2005-05-01 08:59:00 -07001018extern int can_nice(const task_t *p, const int nice);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001019extern int task_curr(const task_t *p);
1020extern int idle_cpu(int cpu);
1021extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1022extern task_t *idle_task(int cpu);
Keith Owensa2a97982005-09-11 17:19:06 +10001023extern task_t *curr_task(int cpu);
1024extern void set_curr_task(int cpu, task_t *p);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001025
1026void yield(void);
1027
1028/*
1029 * The default (Linux) execution domain.
1030 */
1031extern struct exec_domain default_exec_domain;
1032
1033union thread_union {
1034 struct thread_info thread_info;
1035 unsigned long stack[THREAD_SIZE/sizeof(long)];
1036};
1037
1038#ifndef __HAVE_ARCH_KSTACK_END
1039static inline int kstack_end(void *addr)
1040{
1041 /* Reliable end of stack detection:
1042 * Some APM bios versions misalign the stack
1043 */
1044 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1045}
1046#endif
1047
1048extern union thread_union init_thread_union;
1049extern struct task_struct init_task;
1050
1051extern struct mm_struct init_mm;
1052
1053#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1054extern struct task_struct *find_task_by_pid_type(int type, int pid);
1055extern void set_special_pids(pid_t session, pid_t pgrp);
1056extern void __set_special_pids(pid_t session, pid_t pgrp);
1057
1058/* per-UID process charging. */
1059extern struct user_struct * alloc_uid(uid_t);
1060static inline struct user_struct *get_uid(struct user_struct *u)
1061{
1062 atomic_inc(&u->__count);
1063 return u;
1064}
1065extern void free_uid(struct user_struct *);
1066extern void switch_uid(struct user_struct *);
1067
1068#include <asm/current.h>
1069
1070extern void do_timer(struct pt_regs *);
1071
1072extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1073extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1074extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1075 unsigned long clone_flags));
1076#ifdef CONFIG_SMP
1077 extern void kick_process(struct task_struct *tsk);
1078#else
1079 static inline void kick_process(struct task_struct *tsk) { }
1080#endif
Nick Piggin476d1392005-06-25 14:57:29 -07001081extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
Linus Torvalds1da177e2005-04-16 15:20:36 -07001082extern void FASTCALL(sched_exit(task_t * p));
1083
1084extern int in_group_p(gid_t);
1085extern int in_egroup_p(gid_t);
1086
1087extern void proc_caches_init(void);
1088extern void flush_signals(struct task_struct *);
1089extern void flush_signal_handlers(struct task_struct *, int force_default);
1090extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1091
1092static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1093{
1094 unsigned long flags;
1095 int ret;
1096
1097 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1098 ret = dequeue_signal(tsk, mask, info);
1099 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1100
1101 return ret;
1102}
1103
1104extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1105 sigset_t *mask);
1106extern void unblock_all_signals(void);
1107extern void release_task(struct task_struct * p);
1108extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1109extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1110extern int force_sigsegv(int, struct task_struct *);
1111extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1112extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1113extern int kill_pg_info(int, struct siginfo *, pid_t);
1114extern int kill_proc_info(int, struct siginfo *, pid_t);
Harald Welte46113832005-10-10 19:44:29 +02001115extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001116extern void do_notify_parent(struct task_struct *, int);
1117extern void force_sig(int, struct task_struct *);
1118extern void force_sig_specific(int, struct task_struct *);
1119extern int send_sig(int, struct task_struct *, int);
1120extern void zap_other_threads(struct task_struct *p);
1121extern int kill_pg(pid_t, int, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001122extern int kill_proc(pid_t, int, int);
1123extern struct sigqueue *sigqueue_alloc(void);
1124extern void sigqueue_free(struct sigqueue *);
1125extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1126extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
Oleg Nesterov9ac95f22006-02-09 22:41:50 +03001127extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001128extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1129
1130/* These can be the second arg to send_sig_info/send_group_sig_info. */
1131#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1132#define SEND_SIG_PRIV ((struct siginfo *) 1)
1133#define SEND_SIG_FORCED ((struct siginfo *) 2)
1134
Oleg Nesterov621d3122005-10-30 15:03:45 -08001135static inline int is_si_special(const struct siginfo *info)
1136{
1137 return info <= SEND_SIG_FORCED;
1138}
1139
Linus Torvalds1da177e2005-04-16 15:20:36 -07001140/* True if we are on the alternate signal stack. */
1141
1142static inline int on_sig_stack(unsigned long sp)
1143{
1144 return (sp - current->sas_ss_sp < current->sas_ss_size);
1145}
1146
1147static inline int sas_ss_flags(unsigned long sp)
1148{
1149 return (current->sas_ss_size == 0 ? SS_DISABLE
1150 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1151}
1152
Linus Torvalds1da177e2005-04-16 15:20:36 -07001153/*
1154 * Routines for handling mm_structs
1155 */
1156extern struct mm_struct * mm_alloc(void);
1157
1158/* mmdrop drops the mm and the page tables */
1159extern void FASTCALL(__mmdrop(struct mm_struct *));
1160static inline void mmdrop(struct mm_struct * mm)
1161{
1162 if (atomic_dec_and_test(&mm->mm_count))
1163 __mmdrop(mm);
1164}
1165
1166/* mmput gets rid of the mappings and all user-space */
1167extern void mmput(struct mm_struct *);
1168/* Grab a reference to a task's mm, if it is not already going away */
1169extern struct mm_struct *get_task_mm(struct task_struct *task);
1170/* Remove the current tasks stale references to the old mm_struct */
1171extern void mm_release(struct task_struct *, struct mm_struct *);
1172
1173extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1174extern void flush_thread(void);
1175extern void exit_thread(void);
1176
Linus Torvalds1da177e2005-04-16 15:20:36 -07001177extern void exit_files(struct task_struct *);
Oleg Nesterov6b3934e2006-03-28 16:11:16 -08001178extern void __cleanup_signal(struct signal_struct *);
Oleg Nesterova7e53282006-03-28 16:11:27 -08001179extern void __cleanup_sighand(struct sighand_struct *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001180extern void exit_itimers(struct signal_struct *);
1181
1182extern NORET_TYPE void do_group_exit(int);
1183
Linus Torvalds1da177e2005-04-16 15:20:36 -07001184extern void daemonize(const char *, ...);
1185extern int allow_signal(int);
1186extern int disallow_signal(int);
1187extern task_t *child_reaper;
1188
1189extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1190extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1191task_t *fork_idle(int);
1192
1193extern void set_task_comm(struct task_struct *tsk, char *from);
1194extern void get_task_comm(char *to, struct task_struct *tsk);
1195
1196#ifdef CONFIG_SMP
1197extern void wait_task_inactive(task_t * p);
1198#else
1199#define wait_task_inactive(p) do { } while (0)
1200#endif
1201
1202#define remove_parent(p) list_del_init(&(p)->sibling)
Oleg Nesterov8fafabd2006-03-28 16:11:05 -08001203#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001204
Eric W. Biederman5e85d4a2006-04-18 22:20:16 -07001205#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001206
1207#define for_each_process(p) \
1208 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1209
1210/*
1211 * Careful: do_each_thread/while_each_thread is a double loop so
1212 * 'break' will not work as expected - use goto instead.
1213 */
1214#define do_each_thread(g, t) \
1215 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1216
1217#define while_each_thread(g, t) \
1218 while ((t = next_thread(t)) != g)
1219
Eric W. Biedermande12a782006-04-10 17:16:49 -06001220/* de_thread depends on thread_group_leader not being a pid based check */
1221#define thread_group_leader(p) (p == p->group_leader)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001222
Keith Owensa9cdf412006-04-10 22:54:07 -07001223static inline task_t *next_thread(const task_t *p)
Oleg Nesterov47e65322006-03-28 16:11:25 -08001224{
1225 return list_entry(rcu_dereference(p->thread_group.next),
1226 task_t, thread_group);
1227}
1228
Linus Torvalds1da177e2005-04-16 15:20:36 -07001229static inline int thread_group_empty(task_t *p)
1230{
Oleg Nesterov47e65322006-03-28 16:11:25 -08001231 return list_empty(&p->thread_group);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001232}
1233
1234#define delay_group_leader(p) \
1235 (thread_group_leader(p) && !thread_group_empty(p))
1236
Linus Torvalds1da177e2005-04-16 15:20:36 -07001237/*
Eric W. Biederman260ea102006-06-23 02:05:18 -07001238 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
Jens Axboe22e2c502005-06-27 10:55:12 +02001239 * subscriptions and synchronises with wait4(). Also used in procfs. Also
Paul Jackson053199e2005-10-30 15:02:30 -08001240 * pins the final release of task.io_context. Also protects ->cpuset.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001241 *
1242 * Nests both inside and outside of read_lock(&tasklist_lock).
1243 * It must not be nested with write_lock_irq(&tasklist_lock),
1244 * neither inside nor outside.
1245 */
1246static inline void task_lock(struct task_struct *p)
1247{
1248 spin_lock(&p->alloc_lock);
1249}
1250
1251static inline void task_unlock(struct task_struct *p)
1252{
1253 spin_unlock(&p->alloc_lock);
1254}
1255
Oleg Nesterovf63ee722006-03-28 16:11:13 -08001256extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1257 unsigned long *flags);
1258
1259static inline void unlock_task_sighand(struct task_struct *tsk,
1260 unsigned long *flags)
1261{
1262 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1263}
1264
Al Virof0373602005-11-13 16:06:57 -08001265#ifndef __HAVE_THREAD_FUNCTIONS
1266
Al Viroa1261f52005-11-13 16:06:55 -08001267#define task_thread_info(task) (task)->thread_info
Al Viro9fc65872006-01-12 01:05:34 -08001268#define task_stack_page(task) ((void*)((task)->thread_info))
Al Viroa1261f52005-11-13 16:06:55 -08001269
Al Viro10ebffd2005-11-13 16:06:56 -08001270static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1271{
1272 *task_thread_info(p) = *task_thread_info(org);
1273 task_thread_info(p)->task = p;
1274}
1275
1276static inline unsigned long *end_of_stack(struct task_struct *p)
1277{
1278 return (unsigned long *)(p->thread_info + 1);
1279}
1280
Al Virof0373602005-11-13 16:06:57 -08001281#endif
1282
Linus Torvalds1da177e2005-04-16 15:20:36 -07001283/* set thread flags in other task's structures
1284 * - see asm/thread_info.h for TIF_xxxx flags available
1285 */
1286static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1287{
Al Viroa1261f52005-11-13 16:06:55 -08001288 set_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001289}
1290
1291static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1292{
Al Viroa1261f52005-11-13 16:06:55 -08001293 clear_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001294}
1295
1296static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1297{
Al Viroa1261f52005-11-13 16:06:55 -08001298 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001299}
1300
1301static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1302{
Al Viroa1261f52005-11-13 16:06:55 -08001303 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001304}
1305
1306static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1307{
Al Viroa1261f52005-11-13 16:06:55 -08001308 return test_ti_thread_flag(task_thread_info(tsk), flag);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001309}
1310
1311static inline void set_tsk_need_resched(struct task_struct *tsk)
1312{
1313 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1314}
1315
1316static inline void clear_tsk_need_resched(struct task_struct *tsk)
1317{
1318 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1319}
1320
1321static inline int signal_pending(struct task_struct *p)
1322{
1323 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1324}
1325
1326static inline int need_resched(void)
1327{
1328 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1329}
1330
1331/*
1332 * cond_resched() and cond_resched_lock(): latency reduction via
1333 * explicit rescheduling in places that are safe. The return
1334 * value indicates whether a reschedule was done in fact.
1335 * cond_resched_lock() will drop the spinlock before scheduling,
1336 * cond_resched_softirq() will enable bhs before scheduling.
1337 */
1338extern int cond_resched(void);
1339extern int cond_resched_lock(spinlock_t * lock);
1340extern int cond_resched_softirq(void);
1341
1342/*
1343 * Does a critical section need to be broken due to another
1344 * task waiting?:
1345 */
1346#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1347# define need_lockbreak(lock) ((lock)->break_lock)
1348#else
1349# define need_lockbreak(lock) 0
1350#endif
1351
1352/*
1353 * Does a critical section need to be broken due to another
1354 * task waiting or preemption being signalled:
1355 */
1356static inline int lock_need_resched(spinlock_t *lock)
1357{
1358 if (need_lockbreak(lock) || need_resched())
1359 return 1;
1360 return 0;
1361}
1362
1363/* Reevaluate whether the task has signals pending delivery.
1364 This is required every time the blocked sigset_t changes.
1365 callers must hold sighand->siglock. */
1366
1367extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1368extern void recalc_sigpending(void);
1369
1370extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1371
1372/*
1373 * Wrappers for p->thread_info->cpu access. No-op on UP.
1374 */
1375#ifdef CONFIG_SMP
1376
1377static inline unsigned int task_cpu(const struct task_struct *p)
1378{
Al Viroa1261f52005-11-13 16:06:55 -08001379 return task_thread_info(p)->cpu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001380}
1381
1382static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1383{
Al Viroa1261f52005-11-13 16:06:55 -08001384 task_thread_info(p)->cpu = cpu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001385}
1386
1387#else
1388
1389static inline unsigned int task_cpu(const struct task_struct *p)
1390{
1391 return 0;
1392}
1393
1394static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1395{
1396}
1397
1398#endif /* CONFIG_SMP */
1399
1400#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1401extern void arch_pick_mmap_layout(struct mm_struct *mm);
1402#else
1403static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1404{
1405 mm->mmap_base = TASK_UNMAPPED_BASE;
1406 mm->get_unmapped_area = arch_get_unmapped_area;
1407 mm->unmap_area = arch_unmap_area;
1408}
1409#endif
1410
1411extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1412extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1413
Linus Torvalds1da177e2005-04-16 15:20:36 -07001414extern void normalize_rt_tasks(void);
1415
Linus Torvalds1da177e2005-04-16 15:20:36 -07001416#ifdef CONFIG_PM
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001417/*
1418 * Check if a process has been frozen
1419 */
1420static inline int frozen(struct task_struct *p)
1421{
1422 return p->flags & PF_FROZEN;
1423}
1424
1425/*
1426 * Check if there is a request to freeze a process
1427 */
1428static inline int freezing(struct task_struct *p)
1429{
1430 return p->flags & PF_FREEZE;
1431}
1432
1433/*
1434 * Request that a process be frozen
1435 * FIXME: SMP problem. We may not modify other process' flags!
1436 */
1437static inline void freeze(struct task_struct *p)
1438{
1439 p->flags |= PF_FREEZE;
1440}
1441
1442/*
1443 * Wake up a frozen process
1444 */
1445static inline int thaw_process(struct task_struct *p)
1446{
1447 if (frozen(p)) {
1448 p->flags &= ~PF_FROZEN;
1449 wake_up_process(p);
1450 return 1;
1451 }
1452 return 0;
1453}
1454
1455/*
1456 * freezing is complete, mark process as frozen
1457 */
1458static inline void frozen_process(struct task_struct *p)
1459{
1460 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1461}
1462
1463extern void refrigerator(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001464extern int freeze_processes(void);
1465extern void thaw_processes(void);
1466
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001467static inline int try_to_freeze(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001468{
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001469 if (freezing(current)) {
1470 refrigerator();
Linus Torvalds1da177e2005-04-16 15:20:36 -07001471 return 1;
1472 } else
1473 return 0;
1474}
1475#else
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001476static inline int frozen(struct task_struct *p) { return 0; }
1477static inline int freezing(struct task_struct *p) { return 0; }
1478static inline void freeze(struct task_struct *p) { BUG(); }
1479static inline int thaw_process(struct task_struct *p) { return 1; }
1480static inline void frozen_process(struct task_struct *p) { BUG(); }
1481
1482static inline void refrigerator(void) {}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001483static inline int freeze_processes(void) { BUG(); return 0; }
1484static inline void thaw_processes(void) {}
1485
Christoph Lameter3e1d1d22005-06-24 23:13:50 -07001486static inline int try_to_freeze(void) { return 0; }
1487
Linus Torvalds1da177e2005-04-16 15:20:36 -07001488#endif /* CONFIG_PM */
1489#endif /* __KERNEL__ */
1490
1491#endif