4 #include <asm/param.h> /* for HZ */
6 #include <linux/config.h>
7 #include <linux/capability.h>
8 #include <linux/threads.h>
9 #include <linux/kernel.h>
10 #include <linux/types.h>
11 #include <linux/timex.h>
12 #include <linux/jiffies.h>
13 #include <linux/rbtree.h>
14 #include <linux/thread_info.h>
15 #include <linux/cpumask.h>
16 #include <linux/errno.h>
17 #include <linux/nodemask.h>
19 #include <asm/system.h>
20 #include <asm/semaphore.h>
22 #include <asm/ptrace.h>
24 #include <asm/cputime.h>
26 #include <linux/smp.h>
27 #include <linux/sem.h>
28 #include <linux/signal.h>
29 #include <linux/securebits.h>
30 #include <linux/fs_struct.h>
31 #include <linux/compiler.h>
32 #include <linux/completion.h>
33 #include <linux/pid.h>
34 #include <linux/percpu.h>
35 #include <linux/topology.h>
36 #include <linux/seccomp.h>
38 #include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
45 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
46 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
47 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
48 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
49 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
50 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
51 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
52 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
53 #define CLONE_THREAD 0x00010000 /* Same thread group? */
54 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
55 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
56 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
57 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
58 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
59 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
60 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
61 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
62 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
65 * List of flags we want to share for kernel threads,
66 * if only because they are not used by them anyway.
68 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
71 * These are the constant used to fake the fixed-point load-average
72 * counting. Some notes:
73 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
74 * a load-average precision of 10 bits integer + 11 bits fractional
75 * - if you want to count load-averages more often, you need more
76 * precision, or rounding will get you. With 2-second counting freq,
77 * the EXP_n values would be 1981, 2034 and 2043 if still using only
80 extern unsigned long avenrun
[]; /* Load averages */
82 #define FSHIFT 11 /* nr of bits of precision */
83 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
84 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
85 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
86 #define EXP_5 2014 /* 1/exp(5sec/5min) */
87 #define EXP_15 2037 /* 1/exp(5sec/15min) */
89 #define CALC_LOAD(load,exp,n) \
91 load += n*(FIXED_1-exp); \
94 extern unsigned long total_forks
;
95 extern int nr_threads
;
97 DECLARE_PER_CPU(unsigned long, process_counts
);
98 extern int nr_processes(void);
99 extern unsigned long nr_running(void);
100 extern unsigned long nr_uninterruptible(void);
101 extern unsigned long nr_iowait(void);
103 #include <linux/time.h>
104 #include <linux/param.h>
105 #include <linux/resource.h>
106 #include <linux/timer.h>
108 #include <asm/processor.h>
110 #define TASK_RUNNING 0
111 #define TASK_INTERRUPTIBLE 1
112 #define TASK_UNINTERRUPTIBLE 2
113 #define TASK_STOPPED 4
114 #define TASK_TRACED 8
115 #define EXIT_ZOMBIE 16
117 #define TASK_NONINTERACTIVE 64
119 #define __set_task_state(tsk, state_value) \
120 do { (tsk)->state = (state_value); } while (0)
121 #define set_task_state(tsk, state_value) \
122 set_mb((tsk)->state, (state_value))
124 #define __set_current_state(state_value) \
125 do { current->state = (state_value); } while (0)
126 #define set_current_state(state_value) \
127 set_mb(current->state, (state_value))
129 /* Task command name length */
130 #define TASK_COMM_LEN 16
133 * Scheduling policies
135 #define SCHED_NORMAL 0
145 #include <linux/spinlock.h>
148 * This serializes "schedule()" and also protects
149 * the run-queue from deletions/modifications (but
150 * _adding_ to the beginning of the run-queue has
153 extern rwlock_t tasklist_lock
;
154 extern spinlock_t mmlist_lock
;
156 typedef struct task_struct task_t
;
158 extern void sched_init(void);
159 extern void sched_init_smp(void);
160 extern void init_idle(task_t
*idle
, int cpu
);
162 extern cpumask_t nohz_cpu_mask
;
164 extern void show_state(void);
165 extern void show_regs(struct pt_regs
*);
168 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
169 * task), SP is the stack pointer of the first frame that should be shown in the back
170 * trace (or NULL if the entire call-chain of the task should be shown).
172 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
174 void io_schedule(void);
175 long io_schedule_timeout(long timeout
);
177 extern void cpu_init (void);
178 extern void trap_init(void);
179 extern void update_process_times(int user
);
180 extern void scheduler_tick(void);
182 #ifdef CONFIG_DETECT_SOFTLOCKUP
183 extern void softlockup_tick(struct pt_regs
*regs
);
184 extern void spawn_softlockup_task(void);
185 extern void touch_softlockup_watchdog(void);
187 static inline void softlockup_tick(struct pt_regs
*regs
)
190 static inline void spawn_softlockup_task(void)
193 static inline void touch_softlockup_watchdog(void)
199 /* Attach to any functions which should be ignored in wchan output. */
200 #define __sched __attribute__((__section__(".sched.text")))
201 /* Is this address in the __sched functions? */
202 extern int in_sched_functions(unsigned long addr
);
204 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
205 extern signed long FASTCALL(schedule_timeout(signed long timeout
));
206 extern signed long schedule_timeout_interruptible(signed long timeout
);
207 extern signed long schedule_timeout_uninterruptible(signed long timeout
);
208 asmlinkage
void schedule(void);
212 /* Maximum number of active map areas.. This is a random (large) number */
213 #define DEFAULT_MAX_MAP_COUNT 65536
215 extern int sysctl_max_map_count
;
217 #include <linux/aio.h>
220 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
221 unsigned long, unsigned long);
223 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
224 unsigned long len
, unsigned long pgoff
,
225 unsigned long flags
);
226 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
227 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
229 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
230 #define get_mm_counter(mm, member) ((mm)->_##member)
231 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
232 #define inc_mm_counter(mm, member) (mm)->_##member++
233 #define dec_mm_counter(mm, member) (mm)->_##member--
234 typedef unsigned long mm_counter_t
;
237 struct vm_area_struct
* mmap
; /* list of VMAs */
238 struct rb_root mm_rb
;
239 struct vm_area_struct
* mmap_cache
; /* last find_vma result */
240 unsigned long (*get_unmapped_area
) (struct file
*filp
,
241 unsigned long addr
, unsigned long len
,
242 unsigned long pgoff
, unsigned long flags
);
243 void (*unmap_area
) (struct mm_struct
*mm
, unsigned long addr
);
244 unsigned long mmap_base
; /* base of mmap area */
245 unsigned long cached_hole_size
; /* if non-zero, the largest hole below free_area_cache */
246 unsigned long free_area_cache
; /* first hole of size cached_hole_size or larger */
248 atomic_t mm_users
; /* How many users with user space? */
249 atomic_t mm_count
; /* How many references to "struct mm_struct" (users count as 1) */
250 int map_count
; /* number of VMAs */
251 struct rw_semaphore mmap_sem
;
252 spinlock_t page_table_lock
; /* Protects page tables and some counters */
254 struct list_head mmlist
; /* List of maybe swapped mm's. These are globally strung
255 * together off init_mm.mmlist, and are protected
259 unsigned long start_code
, end_code
, start_data
, end_data
;
260 unsigned long start_brk
, brk
, start_stack
;
261 unsigned long arg_start
, arg_end
, env_start
, env_end
;
262 unsigned long total_vm
, locked_vm
, shared_vm
;
263 unsigned long exec_vm
, stack_vm
, reserved_vm
, def_flags
, nr_ptes
;
265 /* Special counters protected by the page_table_lock */
267 mm_counter_t _anon_rss
;
269 unsigned long saved_auxv
[AT_VECTOR_SIZE
]; /* for /proc/PID/auxv */
272 cpumask_t cpu_vm_mask
;
274 /* Architecture-specific MM context */
275 mm_context_t context
;
277 /* Token based thrashing protection. */
278 unsigned long swap_token_time
;
281 /* coredumping support */
283 struct completion
*core_startup_done
, core_done
;
286 rwlock_t ioctx_list_lock
;
287 struct kioctx
*ioctx_list
;
289 struct kioctx default_kioctx
;
291 unsigned long hiwater_rss
; /* High-water RSS usage */
292 unsigned long hiwater_vm
; /* High-water virtual memory usage */
295 struct sighand_struct
{
297 struct k_sigaction action
[_NSIG
];
302 * NOTE! "signal_struct" does not have it's own
303 * locking, because a shared signal_struct always
304 * implies a shared sighand_struct, so locking
305 * sighand_struct is always a proper superset of
306 * the locking of signal_struct.
308 struct signal_struct
{
312 wait_queue_head_t wait_chldexit
; /* for wait4() */
314 /* current thread group signal load-balancing target: */
317 /* shared signal handling: */
318 struct sigpending shared_pending
;
320 /* thread group exit support */
323 * - notify group_exit_task when ->count is equal to notify_count
324 * - everyone except group_exit_task is stopped during signal delivery
325 * of fatal signals, group_exit_task processes the signal.
327 struct task_struct
*group_exit_task
;
330 /* thread group stop support, overloads group_exit_code too */
331 int group_stop_count
;
332 unsigned int flags
; /* see SIGNAL_* flags below */
334 /* POSIX.1b Interval Timers */
335 struct list_head posix_timers
;
337 /* ITIMER_REAL timer for the process */
338 struct timer_list real_timer
;
339 unsigned long it_real_value
, it_real_incr
;
341 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
342 cputime_t it_prof_expires
, it_virt_expires
;
343 cputime_t it_prof_incr
, it_virt_incr
;
345 /* job control IDs */
349 /* boolean value for session group leader */
352 struct tty_struct
*tty
; /* NULL if no tty */
355 * Cumulative resource counters for dead threads in the group,
356 * and for reaped dead child processes forked by this group.
357 * Live threads maintain their own counters and add to these
358 * in __exit_signal, except for the group leader.
360 cputime_t utime
, stime
, cutime
, cstime
;
361 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
362 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
365 * Cumulative ns of scheduled CPU time for dead threads in the
366 * group, not including a zombie group leader. (This only differs
367 * from jiffies_to_ns(utime + stime) if sched_clock uses something
368 * other than jiffies.)
370 unsigned long long sched_time
;
373 * We don't bother to synchronize most readers of this at all,
374 * because there is no reader checking a limit that actually needs
375 * to get both rlim_cur and rlim_max atomically, and either one
376 * alone is a single word that can safely be read normally.
377 * getrlimit/setrlimit use task_lock(current->group_leader) to
378 * protect this instead of the siglock, because they really
379 * have no need to disable irqs.
381 struct rlimit rlim
[RLIM_NLIMITS
];
383 struct list_head cpu_timers
[3];
385 /* keep the process-shared keyrings here so that they do the right
386 * thing in threads created with CLONE_THREAD */
388 struct key
*session_keyring
; /* keyring inherited over fork */
389 struct key
*process_keyring
; /* keyring private to this process */
393 /* Context switch must be unlocked if interrupts are to be enabled */
394 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
395 # define __ARCH_WANT_UNLOCKED_CTXSW
399 * Bits in flags field of signal_struct.
401 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
402 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
403 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
404 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
408 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
409 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
410 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
411 * are inverted: lower p->prio value means higher priority.
413 * The MAX_USER_RT_PRIO value allows the actual maximum
414 * RT priority to be separate from the value exported to
415 * user-space. This allows kernel threads to set their
416 * priority to a value higher than any user task. Note:
417 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
420 #define MAX_USER_RT_PRIO 100
421 #define MAX_RT_PRIO MAX_USER_RT_PRIO
423 #define MAX_PRIO (MAX_RT_PRIO + 40)
425 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
428 * Some day this will be a full-fledged user tracking system..
431 atomic_t __count
; /* reference count */
432 atomic_t processes
; /* How many processes does this user have? */
433 atomic_t files
; /* How many open files does this user have? */
434 atomic_t sigpending
; /* How many pending signals does this user have? */
435 #ifdef CONFIG_INOTIFY
436 atomic_t inotify_watches
; /* How many inotify watches does this user have? */
437 atomic_t inotify_devs
; /* How many inotify devs does this user have opened? */
439 /* protected by mq_lock */
440 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
441 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
444 struct key
*uid_keyring
; /* UID specific keyring */
445 struct key
*session_keyring
; /* UID's default session keyring */
448 /* Hash table maintenance information */
449 struct list_head uidhash_list
;
453 extern struct user_struct
*find_user(uid_t
);
455 extern struct user_struct root_user
;
456 #define INIT_USER (&root_user)
458 typedef struct prio_array prio_array_t
;
459 struct backing_dev_info
;
460 struct reclaim_state
;
462 #ifdef CONFIG_SCHEDSTATS
464 /* cumulative counters */
465 unsigned long cpu_time
, /* time spent on the cpu */
466 run_delay
, /* time spent waiting on a runqueue */
467 pcnt
; /* # of timeslices run on this cpu */
470 unsigned long last_arrival
, /* when we last ran on a cpu */
471 last_queued
; /* when we were last queued to run */
474 extern struct file_operations proc_schedstat_operations
;
486 * sched-domains (multiprocessor balancing) declarations:
489 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
491 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
492 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
493 #define SD_BALANCE_EXEC 4 /* Balance on exec */
494 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
495 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
496 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
497 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
498 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
501 struct sched_group
*next
; /* Must be a circular list */
505 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
506 * single CPU. This is read only (except for setup, hotplug CPU).
508 unsigned long cpu_power
;
511 struct sched_domain
{
512 /* These fields must be setup */
513 struct sched_domain
*parent
; /* top domain must be null terminated */
514 struct sched_group
*groups
; /* the balancing groups of the domain */
515 cpumask_t span
; /* span of all CPUs in this domain */
516 unsigned long min_interval
; /* Minimum balance interval ms */
517 unsigned long max_interval
; /* Maximum balance interval ms */
518 unsigned int busy_factor
; /* less balancing by factor if busy */
519 unsigned int imbalance_pct
; /* No balance until over watermark */
520 unsigned long long cache_hot_time
; /* Task considered cache hot (ns) */
521 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
522 unsigned int per_cpu_gain
; /* CPU % gained by adding domain cpus */
523 unsigned int busy_idx
;
524 unsigned int idle_idx
;
525 unsigned int newidle_idx
;
526 unsigned int wake_idx
;
527 unsigned int forkexec_idx
;
528 int flags
; /* See SD_* */
530 /* Runtime fields. */
531 unsigned long last_balance
; /* init to jiffies. units in jiffies */
532 unsigned int balance_interval
; /* initialise to 1. units in ms. */
533 unsigned int nr_balance_failed
; /* initialise to 0 */
535 #ifdef CONFIG_SCHEDSTATS
536 /* load_balance() stats */
537 unsigned long lb_cnt
[MAX_IDLE_TYPES
];
538 unsigned long lb_failed
[MAX_IDLE_TYPES
];
539 unsigned long lb_balanced
[MAX_IDLE_TYPES
];
540 unsigned long lb_imbalance
[MAX_IDLE_TYPES
];
541 unsigned long lb_gained
[MAX_IDLE_TYPES
];
542 unsigned long lb_hot_gained
[MAX_IDLE_TYPES
];
543 unsigned long lb_nobusyg
[MAX_IDLE_TYPES
];
544 unsigned long lb_nobusyq
[MAX_IDLE_TYPES
];
546 /* Active load balancing */
547 unsigned long alb_cnt
;
548 unsigned long alb_failed
;
549 unsigned long alb_pushed
;
551 /* SD_BALANCE_EXEC stats */
552 unsigned long sbe_cnt
;
553 unsigned long sbe_balanced
;
554 unsigned long sbe_pushed
;
556 /* SD_BALANCE_FORK stats */
557 unsigned long sbf_cnt
;
558 unsigned long sbf_balanced
;
559 unsigned long sbf_pushed
;
561 /* try_to_wake_up() stats */
562 unsigned long ttwu_wake_remote
;
563 unsigned long ttwu_move_affine
;
564 unsigned long ttwu_move_balance
;
568 extern void partition_sched_domains(cpumask_t
*partition1
,
569 cpumask_t
*partition2
);
570 #endif /* CONFIG_SMP */
573 struct io_context
; /* See blkdev.h */
574 void exit_io_context(void);
577 #define NGROUPS_SMALL 32
578 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
582 gid_t small_block
[NGROUPS_SMALL
];
588 * get_group_info() must be called with the owning task locked (via task_lock())
589 * when task != current. The reason being that the vast majority of callers are
590 * looking at current->group_info, which can not be changed except by the
591 * current task. Changing current->group_info requires the task lock, too.
593 #define get_group_info(group_info) do { \
594 atomic_inc(&(group_info)->usage); \
597 #define put_group_info(group_info) do { \
598 if (atomic_dec_and_test(&(group_info)->usage)) \
599 groups_free(group_info); \
602 extern struct group_info
*groups_alloc(int gidsetsize
);
603 extern void groups_free(struct group_info
*group_info
);
604 extern int set_current_groups(struct group_info
*group_info
);
605 extern int groups_search(struct group_info
*group_info
, gid_t grp
);
606 /* access the groups "array" with this macro */
607 #define GROUP_AT(gi, i) \
608 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
610 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
611 extern void prefetch_stack(struct task_struct
*);
613 static inline void prefetch_stack(struct task_struct
*t
) { }
616 struct audit_context
; /* See audit.c */
620 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
621 struct thread_info
*thread_info
;
623 unsigned long flags
; /* per process flags, defined below */
624 unsigned long ptrace
;
626 int lock_depth
; /* BKL lock depth */
628 #if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
631 int prio
, static_prio
;
632 struct list_head run_list
;
635 unsigned short ioprio
;
637 unsigned long sleep_avg
;
638 unsigned long long timestamp
, last_ran
;
639 unsigned long long sched_time
; /* sched_clock time spent running */
642 unsigned long policy
;
643 cpumask_t cpus_allowed
;
644 unsigned int time_slice
, first_time_slice
;
646 #ifdef CONFIG_SCHEDSTATS
647 struct sched_info sched_info
;
650 struct list_head tasks
;
652 * ptrace_list/ptrace_children forms the list of my children
653 * that were stolen by a ptracer.
655 struct list_head ptrace_children
;
656 struct list_head ptrace_list
;
658 struct mm_struct
*mm
, *active_mm
;
661 struct linux_binfmt
*binfmt
;
663 int exit_code
, exit_signal
;
664 int pdeath_signal
; /* The signal sent when the parent dies */
666 unsigned long personality
;
671 * pointers to (original) parent process, youngest child, younger sibling,
672 * older sibling, respectively. (p->father can be replaced with
675 struct task_struct
*real_parent
; /* real parent process (when being debugged) */
676 struct task_struct
*parent
; /* parent process */
678 * children/sibling forms the list of my children plus the
679 * tasks I'm ptracing.
681 struct list_head children
; /* list of my children */
682 struct list_head sibling
; /* linkage in my parent's children list */
683 struct task_struct
*group_leader
; /* threadgroup leader */
685 /* PID/PID hash table linkage. */
686 struct pid pids
[PIDTYPE_MAX
];
688 struct completion
*vfork_done
; /* for vfork() */
689 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
690 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
692 unsigned long rt_priority
;
693 cputime_t utime
, stime
;
694 unsigned long nvcsw
, nivcsw
; /* context switch counts */
695 struct timespec start_time
;
696 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
697 unsigned long min_flt
, maj_flt
;
699 cputime_t it_prof_expires
, it_virt_expires
;
700 unsigned long long it_sched_expires
;
701 struct list_head cpu_timers
[3];
703 /* process credentials */
704 uid_t uid
,euid
,suid
,fsuid
;
705 gid_t gid
,egid
,sgid
,fsgid
;
706 struct group_info
*group_info
;
707 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
;
708 unsigned keep_capabilities
:1;
709 struct user_struct
*user
;
711 struct key
*thread_keyring
; /* keyring private to this thread */
712 unsigned char jit_keyring
; /* default keyring to attach requested keys to */
714 int oomkilladj
; /* OOM kill score adjustment (bit shift). */
715 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
716 - access with [gs]et_task_comm (which lock
718 - initialized normally by flush_old_exec */
719 /* file system info */
720 int link_count
, total_link_count
;
722 struct sysv_sem sysvsem
;
723 /* CPU-specific state of this task */
724 struct thread_struct thread
;
725 /* filesystem information */
726 struct fs_struct
*fs
;
727 /* open file information */
728 struct files_struct
*files
;
730 struct namespace *namespace;
731 /* signal handlers */
732 struct signal_struct
*signal
;
733 struct sighand_struct
*sighand
;
735 sigset_t blocked
, real_blocked
;
736 struct sigpending pending
;
738 unsigned long sas_ss_sp
;
740 int (*notifier
)(void *priv
);
742 sigset_t
*notifier_mask
;
745 struct audit_context
*audit_context
;
748 /* Thread group tracking */
751 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
752 spinlock_t alloc_lock
;
753 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
754 spinlock_t proc_lock
;
756 /* journalling filesystem info */
760 struct reclaim_state
*reclaim_state
;
762 struct dentry
*proc_dentry
;
763 struct backing_dev_info
*backing_dev_info
;
765 struct io_context
*io_context
;
767 unsigned long ptrace_message
;
768 siginfo_t
*last_siginfo
; /* For ptrace use. */
770 * current io wait handle: wait queue entry to use for io waits
771 * If this thread is processing aio, this points at the waitqueue
772 * inside the currently handled kiocb. It may be NULL (i.e. default
773 * to a stack based synchronous wait) if its doing sync IO.
775 wait_queue_t
*io_wait
;
776 /* i/o counters(bytes read/written, #syscalls */
777 u64 rchar
, wchar
, syscr
, syscw
;
778 #if defined(CONFIG_BSD_PROCESS_ACCT)
779 u64 acct_rss_mem1
; /* accumulated rss usage */
780 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
781 clock_t acct_stimexpd
; /* clock_t-converted stime since last update */
784 struct mempolicy
*mempolicy
;
787 #ifdef CONFIG_CPUSETS
788 short cpuset_sem_nest_depth
;
789 struct cpuset
*cpuset
;
790 nodemask_t mems_allowed
;
791 int cpuset_mems_generation
;
793 atomic_t fs_excl
; /* holding fs exclusive resources */
796 static inline pid_t
process_group(struct task_struct
*tsk
)
798 return tsk
->signal
->pgrp
;
802 * pid_alive - check that a task structure is not stale
803 * @p: Task structure to be checked.
805 * Test if a process is not yet dead (at most zombie state)
806 * If pid_alive fails, then pointers within the task structure
807 * can be stale and must not be dereferenced.
809 static inline int pid_alive(struct task_struct
*p
)
811 return p
->pids
[PIDTYPE_PID
].nr
!= 0;
814 extern void free_task(struct task_struct
*tsk
);
815 extern void __put_task_struct(struct task_struct
*tsk
);
816 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
817 #define put_task_struct(tsk) \
818 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
823 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
824 /* Not implemented yet, only for 486*/
825 #define PF_STARTING 0x00000002 /* being created */
826 #define PF_EXITING 0x00000004 /* getting shut down */
827 #define PF_DEAD 0x00000008 /* Dead */
828 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
829 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
830 #define PF_DUMPCORE 0x00000200 /* dumped core */
831 #define PF_SIGNALED 0x00000400 /* killed by a signal */
832 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
833 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
834 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
835 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
836 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
837 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
838 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
839 #define PF_KSWAPD 0x00040000 /* I am kswapd */
840 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
841 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
842 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
843 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
844 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
847 * Only the _current_ task can read/write to tsk->flags, but other
848 * tasks can access tsk->flags in readonly mode for example
849 * with tsk_used_math (like during threaded core dumping).
850 * There is however an exception to this rule during ptrace
851 * or during fork: the ptracer task is allowed to write to the
852 * child->flags of its traced child (same goes for fork, the parent
853 * can write to the child->flags), because we're guaranteed the
854 * child is not running and in turn not changing child->flags
855 * at the same time the parent does it.
857 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
858 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
859 #define clear_used_math() clear_stopped_child_used_math(current)
860 #define set_used_math() set_stopped_child_used_math(current)
861 #define conditional_stopped_child_used_math(condition, child) \
862 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
863 #define conditional_used_math(condition) \
864 conditional_stopped_child_used_math(condition, current)
865 #define copy_to_stopped_child_used_math(child) \
866 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
867 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
868 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
869 #define used_math() tsk_used_math(current)
872 extern int set_cpus_allowed(task_t
*p
, cpumask_t new_mask
);
874 static inline int set_cpus_allowed(task_t
*p
, cpumask_t new_mask
)
876 if (!cpus_intersects(new_mask
, cpu_online_map
))
882 extern unsigned long long sched_clock(void);
883 extern unsigned long long current_sched_time(const task_t
*current_task
);
885 /* sched_exec is called by processes performing an exec */
887 extern void sched_exec(void);
889 #define sched_exec() {}
892 #ifdef CONFIG_HOTPLUG_CPU
893 extern void idle_task_exit(void);
895 static inline void idle_task_exit(void) {}
898 extern void sched_idle_next(void);
899 extern void set_user_nice(task_t
*p
, long nice
);
900 extern int task_prio(const task_t
*p
);
901 extern int task_nice(const task_t
*p
);
902 extern int can_nice(const task_t
*p
, const int nice
);
903 extern int task_curr(const task_t
*p
);
904 extern int idle_cpu(int cpu
);
905 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
906 extern task_t
*idle_task(int cpu
);
911 * The default (Linux) execution domain.
913 extern struct exec_domain default_exec_domain
;
916 struct thread_info thread_info
;
917 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
920 #ifndef __HAVE_ARCH_KSTACK_END
921 static inline int kstack_end(void *addr
)
923 /* Reliable end of stack detection:
924 * Some APM bios versions misalign the stack
926 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
930 extern union thread_union init_thread_union
;
931 extern struct task_struct init_task
;
933 extern struct mm_struct init_mm
;
935 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
936 extern struct task_struct
*find_task_by_pid_type(int type
, int pid
);
937 extern void set_special_pids(pid_t session
, pid_t pgrp
);
938 extern void __set_special_pids(pid_t session
, pid_t pgrp
);
940 /* per-UID process charging. */
941 extern struct user_struct
* alloc_uid(uid_t
);
942 static inline struct user_struct
*get_uid(struct user_struct
*u
)
944 atomic_inc(&u
->__count
);
947 extern void free_uid(struct user_struct
*);
948 extern void switch_uid(struct user_struct
*);
950 #include <asm/current.h>
952 extern void do_timer(struct pt_regs
*);
954 extern int FASTCALL(wake_up_state(struct task_struct
* tsk
, unsigned int state
));
955 extern int FASTCALL(wake_up_process(struct task_struct
* tsk
));
956 extern void FASTCALL(wake_up_new_task(struct task_struct
* tsk
,
957 unsigned long clone_flags
));
959 extern void kick_process(struct task_struct
*tsk
);
961 static inline void kick_process(struct task_struct
*tsk
) { }
963 extern void FASTCALL(sched_fork(task_t
* p
, int clone_flags
));
964 extern void FASTCALL(sched_exit(task_t
* p
));
966 extern int in_group_p(gid_t
);
967 extern int in_egroup_p(gid_t
);
969 extern void proc_caches_init(void);
970 extern void flush_signals(struct task_struct
*);
971 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
972 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
974 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
979 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
980 ret
= dequeue_signal(tsk
, mask
, info
);
981 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
986 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
988 extern void unblock_all_signals(void);
989 extern void release_task(struct task_struct
* p
);
990 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
991 extern int send_group_sig_info(int, struct siginfo
*, struct task_struct
*);
992 extern int force_sigsegv(int, struct task_struct
*);
993 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
994 extern int __kill_pg_info(int sig
, struct siginfo
*info
, pid_t pgrp
);
995 extern int kill_pg_info(int, struct siginfo
*, pid_t
);
996 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
997 extern void do_notify_parent(struct task_struct
*, int);
998 extern void force_sig(int, struct task_struct
*);
999 extern void force_sig_specific(int, struct task_struct
*);
1000 extern int send_sig(int, struct task_struct
*, int);
1001 extern void zap_other_threads(struct task_struct
*p
);
1002 extern int kill_pg(pid_t
, int, int);
1003 extern int kill_sl(pid_t
, int, int);
1004 extern int kill_proc(pid_t
, int, int);
1005 extern struct sigqueue
*sigqueue_alloc(void);
1006 extern void sigqueue_free(struct sigqueue
*);
1007 extern int send_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1008 extern int send_group_sigqueue(int, struct sigqueue
*, struct task_struct
*);
1009 extern int do_sigaction(int, const struct k_sigaction
*, struct k_sigaction
*);
1010 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
1012 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1013 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1014 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1015 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1017 /* True if we are on the alternate signal stack. */
1019 static inline int on_sig_stack(unsigned long sp
)
1021 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
1024 static inline int sas_ss_flags(unsigned long sp
)
1026 return (current
->sas_ss_size
== 0 ? SS_DISABLE
1027 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
1031 #ifdef CONFIG_SECURITY
1032 /* code is in security.c */
1033 extern int capable(int cap
);
1035 static inline int capable(int cap
)
1037 if (cap_raised(current
->cap_effective
, cap
)) {
1038 current
->flags
|= PF_SUPERPRIV
;
1046 * Routines for handling mm_structs
1048 extern struct mm_struct
* mm_alloc(void);
1050 /* mmdrop drops the mm and the page tables */
1051 extern void FASTCALL(__mmdrop(struct mm_struct
*));
1052 static inline void mmdrop(struct mm_struct
* mm
)
1054 if (atomic_dec_and_test(&mm
->mm_count
))
1058 /* mmput gets rid of the mappings and all user-space */
1059 extern void mmput(struct mm_struct
*);
1060 /* Grab a reference to a task's mm, if it is not already going away */
1061 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1062 /* Remove the current tasks stale references to the old mm_struct */
1063 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1065 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1066 extern void flush_thread(void);
1067 extern void exit_thread(void);
1069 extern void exit_files(struct task_struct
*);
1070 extern void exit_signal(struct task_struct
*);
1071 extern void __exit_signal(struct task_struct
*);
1072 extern void exit_sighand(struct task_struct
*);
1073 extern void __exit_sighand(struct task_struct
*);
1074 extern void exit_itimers(struct signal_struct
*);
1076 extern NORET_TYPE
void do_group_exit(int);
1078 extern void daemonize(const char *, ...);
1079 extern int allow_signal(int);
1080 extern int disallow_signal(int);
1081 extern task_t
*child_reaper
;
1083 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1084 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1085 task_t
*fork_idle(int);
1087 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1088 extern void get_task_comm(char *to
, struct task_struct
*tsk
);
1091 extern void wait_task_inactive(task_t
* p
);
1093 #define wait_task_inactive(p) do { } while (0)
1096 #define remove_parent(p) list_del_init(&(p)->sibling)
1097 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1099 #define REMOVE_LINKS(p) do { \
1100 if (thread_group_leader(p)) \
1101 list_del_init(&(p)->tasks); \
1105 #define SET_LINKS(p) do { \
1106 if (thread_group_leader(p)) \
1107 list_add_tail(&(p)->tasks,&init_task.tasks); \
1108 add_parent(p, (p)->parent); \
1111 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1112 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1114 #define for_each_process(p) \
1115 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1118 * Careful: do_each_thread/while_each_thread is a double loop so
1119 * 'break' will not work as expected - use goto instead.
1121 #define do_each_thread(g, t) \
1122 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1124 #define while_each_thread(g, t) \
1125 while ((t = next_thread(t)) != g)
1127 extern task_t
* FASTCALL(next_thread(const task_t
*p
));
1129 #define thread_group_leader(p) (p->pid == p->tgid)
1131 static inline int thread_group_empty(task_t
*p
)
1133 return list_empty(&p
->pids
[PIDTYPE_TGID
].pid_list
);
1136 #define delay_group_leader(p) \
1137 (thread_group_leader(p) && !thread_group_empty(p))
1139 extern void unhash_process(struct task_struct
*p
);
1142 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1143 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1144 * pins the final release of task.io_context.
1146 * Nests both inside and outside of read_lock(&tasklist_lock).
1147 * It must not be nested with write_lock_irq(&tasklist_lock),
1148 * neither inside nor outside.
1150 static inline void task_lock(struct task_struct
*p
)
1152 spin_lock(&p
->alloc_lock
);
1155 static inline void task_unlock(struct task_struct
*p
)
1157 spin_unlock(&p
->alloc_lock
);
1160 /* set thread flags in other task's structures
1161 * - see asm/thread_info.h for TIF_xxxx flags available
1163 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1165 set_ti_thread_flag(tsk
->thread_info
,flag
);
1168 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1170 clear_ti_thread_flag(tsk
->thread_info
,flag
);
1173 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1175 return test_and_set_ti_thread_flag(tsk
->thread_info
,flag
);
1178 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1180 return test_and_clear_ti_thread_flag(tsk
->thread_info
,flag
);
1183 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1185 return test_ti_thread_flag(tsk
->thread_info
,flag
);
1188 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
1190 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1193 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
1195 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1198 static inline int signal_pending(struct task_struct
*p
)
1200 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
1203 static inline int need_resched(void)
1205 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
1209 * cond_resched() and cond_resched_lock(): latency reduction via
1210 * explicit rescheduling in places that are safe. The return
1211 * value indicates whether a reschedule was done in fact.
1212 * cond_resched_lock() will drop the spinlock before scheduling,
1213 * cond_resched_softirq() will enable bhs before scheduling.
1215 extern int cond_resched(void);
1216 extern int cond_resched_lock(spinlock_t
* lock
);
1217 extern int cond_resched_softirq(void);
1220 * Does a critical section need to be broken due to another
1223 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1224 # define need_lockbreak(lock) ((lock)->break_lock)
1226 # define need_lockbreak(lock) 0
1230 * Does a critical section need to be broken due to another
1231 * task waiting or preemption being signalled:
1233 static inline int lock_need_resched(spinlock_t
*lock
)
1235 if (need_lockbreak(lock
) || need_resched())
1240 /* Reevaluate whether the task has signals pending delivery.
1241 This is required every time the blocked sigset_t changes.
1242 callers must hold sighand->siglock. */
1244 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct
*t
));
1245 extern void recalc_sigpending(void);
1247 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
1250 * Wrappers for p->thread_info->cpu access. No-op on UP.
1254 static inline unsigned int task_cpu(const struct task_struct
*p
)
1256 return p
->thread_info
->cpu
;
1259 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1261 p
->thread_info
->cpu
= cpu
;
1266 static inline unsigned int task_cpu(const struct task_struct
*p
)
1271 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1275 #endif /* CONFIG_SMP */
1277 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1278 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
1280 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
)
1282 mm
->mmap_base
= TASK_UNMAPPED_BASE
;
1283 mm
->get_unmapped_area
= arch_get_unmapped_area
;
1284 mm
->unmap_area
= arch_unmap_area
;
1288 extern long sched_setaffinity(pid_t pid
, cpumask_t new_mask
);
1289 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
1291 #ifdef CONFIG_MAGIC_SYSRQ
1293 extern void normalize_rt_tasks(void);
1299 * Check if a process has been frozen
1301 static inline int frozen(struct task_struct
*p
)
1303 return p
->flags
& PF_FROZEN
;
1307 * Check if there is a request to freeze a process
1309 static inline int freezing(struct task_struct
*p
)
1311 return p
->flags
& PF_FREEZE
;
1315 * Request that a process be frozen
1316 * FIXME: SMP problem. We may not modify other process' flags!
1318 static inline void freeze(struct task_struct
*p
)
1320 p
->flags
|= PF_FREEZE
;
1324 * Wake up a frozen process
1326 static inline int thaw_process(struct task_struct
*p
)
1329 p
->flags
&= ~PF_FROZEN
;
1337 * freezing is complete, mark process as frozen
1339 static inline void frozen_process(struct task_struct
*p
)
1341 p
->flags
= (p
->flags
& ~PF_FREEZE
) | PF_FROZEN
;
1344 extern void refrigerator(void);
1345 extern int freeze_processes(void);
1346 extern void thaw_processes(void);
1348 static inline int try_to_freeze(void)
1350 if (freezing(current
)) {
1357 static inline int frozen(struct task_struct
*p
) { return 0; }
1358 static inline int freezing(struct task_struct
*p
) { return 0; }
1359 static inline void freeze(struct task_struct
*p
) { BUG(); }
1360 static inline int thaw_process(struct task_struct
*p
) { return 1; }
1361 static inline void frozen_process(struct task_struct
*p
) { BUG(); }
1363 static inline void refrigerator(void) {}
1364 static inline int freeze_processes(void) { BUG(); return 0; }
1365 static inline void thaw_processes(void) {}
1367 static inline int try_to_freeze(void) { return 0; }
1369 #endif /* CONFIG_PM */
1370 #endif /* __KERNEL__ */