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>
43 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
44 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
45 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
46 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
47 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
48 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
49 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
50 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
51 #define CLONE_THREAD 0x00010000 /* Same thread group? */
52 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
53 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
54 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
55 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
56 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
57 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
58 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
59 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
60 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
63 * List of flags we want to share for kernel threads,
64 * if only because they are not used by them anyway.
66 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
69 * These are the constant used to fake the fixed-point load-average
70 * counting. Some notes:
71 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
72 * a load-average precision of 10 bits integer + 11 bits fractional
73 * - if you want to count load-averages more often, you need more
74 * precision, or rounding will get you. With 2-second counting freq,
75 * the EXP_n values would be 1981, 2034 and 2043 if still using only
78 extern unsigned long avenrun
[]; /* Load averages */
80 #define FSHIFT 11 /* nr of bits of precision */
81 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
82 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
83 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
84 #define EXP_5 2014 /* 1/exp(5sec/5min) */
85 #define EXP_15 2037 /* 1/exp(5sec/15min) */
87 #define CALC_LOAD(load,exp,n) \
89 load += n*(FIXED_1-exp); \
92 extern unsigned long total_forks
;
93 extern int nr_threads
;
95 DECLARE_PER_CPU(unsigned long, process_counts
);
96 extern int nr_processes(void);
97 extern unsigned long nr_running(void);
98 extern unsigned long nr_uninterruptible(void);
99 extern unsigned long nr_iowait(void);
101 #include <linux/time.h>
102 #include <linux/param.h>
103 #include <linux/resource.h>
104 #include <linux/timer.h>
106 #include <asm/processor.h>
108 #define TASK_RUNNING 0
109 #define TASK_INTERRUPTIBLE 1
110 #define TASK_UNINTERRUPTIBLE 2
111 #define TASK_STOPPED 4
112 #define TASK_TRACED 8
113 #define EXIT_ZOMBIE 16
116 #define __set_task_state(tsk, state_value) \
117 do { (tsk)->state = (state_value); } while (0)
118 #define set_task_state(tsk, state_value) \
119 set_mb((tsk)->state, (state_value))
121 #define __set_current_state(state_value) \
122 do { current->state = (state_value); } while (0)
123 #define set_current_state(state_value) \
124 set_mb(current->state, (state_value))
126 /* Task command name length */
127 #define TASK_COMM_LEN 16
130 * Scheduling policies
132 #define SCHED_NORMAL 0
142 #include <linux/spinlock.h>
145 * This serializes "schedule()" and also protects
146 * the run-queue from deletions/modifications (but
147 * _adding_ to the beginning of the run-queue has
150 extern rwlock_t tasklist_lock
;
151 extern spinlock_t mmlist_lock
;
153 typedef struct task_struct task_t
;
155 extern void sched_init(void);
156 extern void sched_init_smp(void);
157 extern void init_idle(task_t
*idle
, int cpu
);
159 extern cpumask_t nohz_cpu_mask
;
161 extern void show_state(void);
162 extern void show_regs(struct pt_regs
*);
165 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
166 * task), SP is the stack pointer of the first frame that should be shown in the back
167 * trace (or NULL if the entire call-chain of the task should be shown).
169 extern void show_stack(struct task_struct
*task
, unsigned long *sp
);
171 void io_schedule(void);
172 long io_schedule_timeout(long timeout
);
174 extern void cpu_init (void);
175 extern void trap_init(void);
176 extern void update_process_times(int user
);
177 extern void scheduler_tick(void);
179 /* Attach to any functions which should be ignored in wchan output. */
180 #define __sched __attribute__((__section__(".sched.text")))
181 /* Is this address in the __sched functions? */
182 extern int in_sched_functions(unsigned long addr
);
184 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
185 extern signed long FASTCALL(schedule_timeout(signed long timeout
));
186 asmlinkage
void schedule(void);
190 /* Maximum number of active map areas.. This is a random (large) number */
191 #define DEFAULT_MAX_MAP_COUNT 65536
193 extern int sysctl_max_map_count
;
195 #include <linux/aio.h>
198 arch_get_unmapped_area(struct file
*, unsigned long, unsigned long,
199 unsigned long, unsigned long);
201 arch_get_unmapped_area_topdown(struct file
*filp
, unsigned long addr
,
202 unsigned long len
, unsigned long pgoff
,
203 unsigned long flags
);
204 extern void arch_unmap_area(struct mm_struct
*, unsigned long);
205 extern void arch_unmap_area_topdown(struct mm_struct
*, unsigned long);
207 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
208 #define get_mm_counter(mm, member) ((mm)->_##member)
209 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
210 #define inc_mm_counter(mm, member) (mm)->_##member++
211 #define dec_mm_counter(mm, member) (mm)->_##member--
212 typedef unsigned long mm_counter_t
;
215 struct vm_area_struct
* mmap
; /* list of VMAs */
216 struct rb_root mm_rb
;
217 struct vm_area_struct
* mmap_cache
; /* last find_vma result */
218 unsigned long (*get_unmapped_area
) (struct file
*filp
,
219 unsigned long addr
, unsigned long len
,
220 unsigned long pgoff
, unsigned long flags
);
221 void (*unmap_area
) (struct mm_struct
*mm
, unsigned long addr
);
222 unsigned long mmap_base
; /* base of mmap area */
223 unsigned long cached_hole_size
; /* if non-zero, the largest hole below free_area_cache */
224 unsigned long free_area_cache
; /* first hole of size cached_hole_size or larger */
226 atomic_t mm_users
; /* How many users with user space? */
227 atomic_t mm_count
; /* How many references to "struct mm_struct" (users count as 1) */
228 int map_count
; /* number of VMAs */
229 struct rw_semaphore mmap_sem
;
230 spinlock_t page_table_lock
; /* Protects page tables and some counters */
232 struct list_head mmlist
; /* List of maybe swapped mm's. These are globally strung
233 * together off init_mm.mmlist, and are protected
237 unsigned long start_code
, end_code
, start_data
, end_data
;
238 unsigned long start_brk
, brk
, start_stack
;
239 unsigned long arg_start
, arg_end
, env_start
, env_end
;
240 unsigned long total_vm
, locked_vm
, shared_vm
;
241 unsigned long exec_vm
, stack_vm
, reserved_vm
, def_flags
, nr_ptes
;
243 /* Special counters protected by the page_table_lock */
245 mm_counter_t _anon_rss
;
247 unsigned long saved_auxv
[42]; /* for /proc/PID/auxv */
250 cpumask_t cpu_vm_mask
;
252 /* Architecture-specific MM context */
253 mm_context_t context
;
255 /* Token based thrashing protection. */
256 unsigned long swap_token_time
;
259 /* coredumping support */
261 struct completion
*core_startup_done
, core_done
;
264 rwlock_t ioctx_list_lock
;
265 struct kioctx
*ioctx_list
;
267 struct kioctx default_kioctx
;
269 unsigned long hiwater_rss
; /* High-water RSS usage */
270 unsigned long hiwater_vm
; /* High-water virtual memory usage */
273 struct sighand_struct
{
275 struct k_sigaction action
[_NSIG
];
280 * NOTE! "signal_struct" does not have it's own
281 * locking, because a shared signal_struct always
282 * implies a shared sighand_struct, so locking
283 * sighand_struct is always a proper superset of
284 * the locking of signal_struct.
286 struct signal_struct
{
290 wait_queue_head_t wait_chldexit
; /* for wait4() */
292 /* current thread group signal load-balancing target: */
295 /* shared signal handling: */
296 struct sigpending shared_pending
;
298 /* thread group exit support */
301 * - notify group_exit_task when ->count is equal to notify_count
302 * - everyone except group_exit_task is stopped during signal delivery
303 * of fatal signals, group_exit_task processes the signal.
305 struct task_struct
*group_exit_task
;
308 /* thread group stop support, overloads group_exit_code too */
309 int group_stop_count
;
310 unsigned int flags
; /* see SIGNAL_* flags below */
312 /* POSIX.1b Interval Timers */
313 struct list_head posix_timers
;
315 /* ITIMER_REAL timer for the process */
316 struct timer_list real_timer
;
317 unsigned long it_real_value
, it_real_incr
;
319 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
320 cputime_t it_prof_expires
, it_virt_expires
;
321 cputime_t it_prof_incr
, it_virt_incr
;
323 /* job control IDs */
327 /* boolean value for session group leader */
330 struct tty_struct
*tty
; /* NULL if no tty */
333 * Cumulative resource counters for dead threads in the group,
334 * and for reaped dead child processes forked by this group.
335 * Live threads maintain their own counters and add to these
336 * in __exit_signal, except for the group leader.
338 cputime_t utime
, stime
, cutime
, cstime
;
339 unsigned long nvcsw
, nivcsw
, cnvcsw
, cnivcsw
;
340 unsigned long min_flt
, maj_flt
, cmin_flt
, cmaj_flt
;
343 * Cumulative ns of scheduled CPU time for dead threads in the
344 * group, not including a zombie group leader. (This only differs
345 * from jiffies_to_ns(utime + stime) if sched_clock uses something
346 * other than jiffies.)
348 unsigned long long sched_time
;
351 * We don't bother to synchronize most readers of this at all,
352 * because there is no reader checking a limit that actually needs
353 * to get both rlim_cur and rlim_max atomically, and either one
354 * alone is a single word that can safely be read normally.
355 * getrlimit/setrlimit use task_lock(current->group_leader) to
356 * protect this instead of the siglock, because they really
357 * have no need to disable irqs.
359 struct rlimit rlim
[RLIM_NLIMITS
];
361 struct list_head cpu_timers
[3];
363 /* keep the process-shared keyrings here so that they do the right
364 * thing in threads created with CLONE_THREAD */
366 struct key
*session_keyring
; /* keyring inherited over fork */
367 struct key
*process_keyring
; /* keyring private to this process */
372 * Bits in flags field of signal_struct.
374 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
375 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
376 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
377 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
381 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
382 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
383 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
384 * are inverted: lower p->prio value means higher priority.
386 * The MAX_USER_RT_PRIO value allows the actual maximum
387 * RT priority to be separate from the value exported to
388 * user-space. This allows kernel threads to set their
389 * priority to a value higher than any user task. Note:
390 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
393 #define MAX_USER_RT_PRIO 100
394 #define MAX_RT_PRIO MAX_USER_RT_PRIO
396 #define MAX_PRIO (MAX_RT_PRIO + 40)
398 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
401 * Some day this will be a full-fledged user tracking system..
404 atomic_t __count
; /* reference count */
405 atomic_t processes
; /* How many processes does this user have? */
406 atomic_t files
; /* How many open files does this user have? */
407 atomic_t sigpending
; /* How many pending signals does this user have? */
408 /* protected by mq_lock */
409 unsigned long mq_bytes
; /* How many bytes can be allocated to mqueue? */
410 unsigned long locked_shm
; /* How many pages of mlocked shm ? */
413 struct key
*uid_keyring
; /* UID specific keyring */
414 struct key
*session_keyring
; /* UID's default session keyring */
417 /* Hash table maintenance information */
418 struct list_head uidhash_list
;
422 extern struct user_struct
*find_user(uid_t
);
424 extern struct user_struct root_user
;
425 #define INIT_USER (&root_user)
427 typedef struct prio_array prio_array_t
;
428 struct backing_dev_info
;
429 struct reclaim_state
;
431 #ifdef CONFIG_SCHEDSTATS
433 /* cumulative counters */
434 unsigned long cpu_time
, /* time spent on the cpu */
435 run_delay
, /* time spent waiting on a runqueue */
436 pcnt
; /* # of timeslices run on this cpu */
439 unsigned long last_arrival
, /* when we last ran on a cpu */
440 last_queued
; /* when we were last queued to run */
443 extern struct file_operations proc_schedstat_operations
;
455 * sched-domains (multiprocessor balancing) declarations:
458 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
460 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
461 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
462 #define SD_BALANCE_EXEC 4 /* Balance on exec */
463 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
464 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
465 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
466 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
469 struct sched_group
*next
; /* Must be a circular list */
473 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
474 * single CPU. This is read only (except for setup, hotplug CPU).
476 unsigned long cpu_power
;
479 struct sched_domain
{
480 /* These fields must be setup */
481 struct sched_domain
*parent
; /* top domain must be null terminated */
482 struct sched_group
*groups
; /* the balancing groups of the domain */
483 cpumask_t span
; /* span of all CPUs in this domain */
484 unsigned long min_interval
; /* Minimum balance interval ms */
485 unsigned long max_interval
; /* Maximum balance interval ms */
486 unsigned int busy_factor
; /* less balancing by factor if busy */
487 unsigned int imbalance_pct
; /* No balance until over watermark */
488 unsigned long long cache_hot_time
; /* Task considered cache hot (ns) */
489 unsigned int cache_nice_tries
; /* Leave cache hot tasks for # tries */
490 unsigned int per_cpu_gain
; /* CPU % gained by adding domain cpus */
491 int flags
; /* See SD_* */
493 /* Runtime fields. */
494 unsigned long last_balance
; /* init to jiffies. units in jiffies */
495 unsigned int balance_interval
; /* initialise to 1. units in ms. */
496 unsigned int nr_balance_failed
; /* initialise to 0 */
498 #ifdef CONFIG_SCHEDSTATS
499 /* load_balance() stats */
500 unsigned long lb_cnt
[MAX_IDLE_TYPES
];
501 unsigned long lb_failed
[MAX_IDLE_TYPES
];
502 unsigned long lb_balanced
[MAX_IDLE_TYPES
];
503 unsigned long lb_imbalance
[MAX_IDLE_TYPES
];
504 unsigned long lb_gained
[MAX_IDLE_TYPES
];
505 unsigned long lb_hot_gained
[MAX_IDLE_TYPES
];
506 unsigned long lb_nobusyg
[MAX_IDLE_TYPES
];
507 unsigned long lb_nobusyq
[MAX_IDLE_TYPES
];
509 /* Active load balancing */
510 unsigned long alb_cnt
;
511 unsigned long alb_failed
;
512 unsigned long alb_pushed
;
514 /* sched_balance_exec() stats */
515 unsigned long sbe_attempts
;
516 unsigned long sbe_pushed
;
518 /* try_to_wake_up() stats */
519 unsigned long ttwu_wake_remote
;
520 unsigned long ttwu_move_affine
;
521 unsigned long ttwu_move_balance
;
525 #ifdef ARCH_HAS_SCHED_DOMAIN
526 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
527 extern cpumask_t cpu_isolated_map
;
528 extern void init_sched_build_groups(struct sched_group groups
[],
529 cpumask_t span
, int (*group_fn
)(int cpu
));
530 extern void cpu_attach_domain(struct sched_domain
*sd
, int cpu
);
531 #endif /* ARCH_HAS_SCHED_DOMAIN */
532 #endif /* CONFIG_SMP */
535 struct io_context
; /* See blkdev.h */
536 void exit_io_context(void);
539 #define NGROUPS_SMALL 32
540 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
544 gid_t small_block
[NGROUPS_SMALL
];
550 * get_group_info() must be called with the owning task locked (via task_lock())
551 * when task != current. The reason being that the vast majority of callers are
552 * looking at current->group_info, which can not be changed except by the
553 * current task. Changing current->group_info requires the task lock, too.
555 #define get_group_info(group_info) do { \
556 atomic_inc(&(group_info)->usage); \
559 #define put_group_info(group_info) do { \
560 if (atomic_dec_and_test(&(group_info)->usage)) \
561 groups_free(group_info); \
564 struct group_info
*groups_alloc(int gidsetsize
);
565 void groups_free(struct group_info
*group_info
);
566 int set_current_groups(struct group_info
*group_info
);
567 /* access the groups "array" with this macro */
568 #define GROUP_AT(gi, i) \
569 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
572 struct audit_context
; /* See audit.c */
576 volatile long state
; /* -1 unrunnable, 0 runnable, >0 stopped */
577 struct thread_info
*thread_info
;
579 unsigned long flags
; /* per process flags, defined below */
580 unsigned long ptrace
;
582 int lock_depth
; /* BKL lock depth */
584 int prio
, static_prio
;
585 struct list_head run_list
;
588 unsigned long sleep_avg
;
589 unsigned long long timestamp
, last_ran
;
590 unsigned long long sched_time
; /* sched_clock time spent running */
593 unsigned long policy
;
594 cpumask_t cpus_allowed
;
595 unsigned int time_slice
, first_time_slice
;
597 #ifdef CONFIG_SCHEDSTATS
598 struct sched_info sched_info
;
601 struct list_head tasks
;
603 * ptrace_list/ptrace_children forms the list of my children
604 * that were stolen by a ptracer.
606 struct list_head ptrace_children
;
607 struct list_head ptrace_list
;
609 struct mm_struct
*mm
, *active_mm
;
612 struct linux_binfmt
*binfmt
;
614 int exit_code
, exit_signal
;
615 int pdeath_signal
; /* The signal sent when the parent dies */
617 unsigned long personality
;
622 * pointers to (original) parent process, youngest child, younger sibling,
623 * older sibling, respectively. (p->father can be replaced with
626 struct task_struct
*real_parent
; /* real parent process (when being debugged) */
627 struct task_struct
*parent
; /* parent process */
629 * children/sibling forms the list of my children plus the
630 * tasks I'm ptracing.
632 struct list_head children
; /* list of my children */
633 struct list_head sibling
; /* linkage in my parent's children list */
634 struct task_struct
*group_leader
; /* threadgroup leader */
636 /* PID/PID hash table linkage. */
637 struct pid pids
[PIDTYPE_MAX
];
639 struct completion
*vfork_done
; /* for vfork() */
640 int __user
*set_child_tid
; /* CLONE_CHILD_SETTID */
641 int __user
*clear_child_tid
; /* CLONE_CHILD_CLEARTID */
643 unsigned long rt_priority
;
644 cputime_t utime
, stime
;
645 unsigned long nvcsw
, nivcsw
; /* context switch counts */
646 struct timespec start_time
;
647 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
648 unsigned long min_flt
, maj_flt
;
650 cputime_t it_prof_expires
, it_virt_expires
;
651 unsigned long long it_sched_expires
;
652 struct list_head cpu_timers
[3];
654 /* process credentials */
655 uid_t uid
,euid
,suid
,fsuid
;
656 gid_t gid
,egid
,sgid
,fsgid
;
657 struct group_info
*group_info
;
658 kernel_cap_t cap_effective
, cap_inheritable
, cap_permitted
;
659 unsigned keep_capabilities
:1;
660 struct user_struct
*user
;
662 struct key
*thread_keyring
; /* keyring private to this thread */
664 int oomkilladj
; /* OOM kill score adjustment (bit shift). */
665 char comm
[TASK_COMM_LEN
]; /* executable name excluding path
666 - access with [gs]et_task_comm (which lock
668 - initialized normally by flush_old_exec */
669 /* file system info */
670 int link_count
, total_link_count
;
672 struct sysv_sem sysvsem
;
673 /* CPU-specific state of this task */
674 struct thread_struct thread
;
675 /* filesystem information */
676 struct fs_struct
*fs
;
677 /* open file information */
678 struct files_struct
*files
;
680 struct namespace *namespace;
681 /* signal handlers */
682 struct signal_struct
*signal
;
683 struct sighand_struct
*sighand
;
685 sigset_t blocked
, real_blocked
;
686 struct sigpending pending
;
688 unsigned long sas_ss_sp
;
690 int (*notifier
)(void *priv
);
692 sigset_t
*notifier_mask
;
695 struct audit_context
*audit_context
;
698 /* Thread group tracking */
701 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
702 spinlock_t alloc_lock
;
703 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
704 spinlock_t proc_lock
;
705 /* context-switch lock */
706 spinlock_t switch_lock
;
708 /* journalling filesystem info */
712 struct reclaim_state
*reclaim_state
;
714 struct dentry
*proc_dentry
;
715 struct backing_dev_info
*backing_dev_info
;
717 struct io_context
*io_context
;
719 unsigned long ptrace_message
;
720 siginfo_t
*last_siginfo
; /* For ptrace use. */
722 * current io wait handle: wait queue entry to use for io waits
723 * If this thread is processing aio, this points at the waitqueue
724 * inside the currently handled kiocb. It may be NULL (i.e. default
725 * to a stack based synchronous wait) if its doing sync IO.
727 wait_queue_t
*io_wait
;
728 /* i/o counters(bytes read/written, #syscalls */
729 u64 rchar
, wchar
, syscr
, syscw
;
730 #if defined(CONFIG_BSD_PROCESS_ACCT)
731 u64 acct_rss_mem1
; /* accumulated rss usage */
732 u64 acct_vm_mem1
; /* accumulated virtual memory usage */
733 clock_t acct_stimexpd
; /* clock_t-converted stime since last update */
736 struct mempolicy
*mempolicy
;
739 #ifdef CONFIG_CPUSETS
740 struct cpuset
*cpuset
;
741 nodemask_t mems_allowed
;
742 int cpuset_mems_generation
;
746 static inline pid_t
process_group(struct task_struct
*tsk
)
748 return tsk
->signal
->pgrp
;
752 * pid_alive - check that a task structure is not stale
753 * @p: Task structure to be checked.
755 * Test if a process is not yet dead (at most zombie state)
756 * If pid_alive fails, then pointers within the task structure
757 * can be stale and must not be dereferenced.
759 static inline int pid_alive(struct task_struct
*p
)
761 return p
->pids
[PIDTYPE_PID
].nr
!= 0;
764 extern void free_task(struct task_struct
*tsk
);
765 extern void __put_task_struct(struct task_struct
*tsk
);
766 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
767 #define put_task_struct(tsk) \
768 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
773 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
774 /* Not implemented yet, only for 486*/
775 #define PF_STARTING 0x00000002 /* being created */
776 #define PF_EXITING 0x00000004 /* getting shut down */
777 #define PF_DEAD 0x00000008 /* Dead */
778 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
779 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
780 #define PF_DUMPCORE 0x00000200 /* dumped core */
781 #define PF_SIGNALED 0x00000400 /* killed by a signal */
782 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
783 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
784 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
785 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
786 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
787 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
788 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
789 #define PF_KSWAPD 0x00040000 /* I am kswapd */
790 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
791 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
792 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
793 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
794 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
797 * Only the _current_ task can read/write to tsk->flags, but other
798 * tasks can access tsk->flags in readonly mode for example
799 * with tsk_used_math (like during threaded core dumping).
800 * There is however an exception to this rule during ptrace
801 * or during fork: the ptracer task is allowed to write to the
802 * child->flags of its traced child (same goes for fork, the parent
803 * can write to the child->flags), because we're guaranteed the
804 * child is not running and in turn not changing child->flags
805 * at the same time the parent does it.
807 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
808 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
809 #define clear_used_math() clear_stopped_child_used_math(current)
810 #define set_used_math() set_stopped_child_used_math(current)
811 #define conditional_stopped_child_used_math(condition, child) \
812 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
813 #define conditional_used_math(condition) \
814 conditional_stopped_child_used_math(condition, current)
815 #define copy_to_stopped_child_used_math(child) \
816 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
817 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
818 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
819 #define used_math() tsk_used_math(current)
822 extern int set_cpus_allowed(task_t
*p
, cpumask_t new_mask
);
824 static inline int set_cpus_allowed(task_t
*p
, cpumask_t new_mask
)
826 if (!cpus_intersects(new_mask
, cpu_online_map
))
832 extern unsigned long long sched_clock(void);
833 extern unsigned long long current_sched_time(const task_t
*current_task
);
835 /* sched_exec is called by processes performing an exec */
837 extern void sched_exec(void);
839 #define sched_exec() {}
842 #ifdef CONFIG_HOTPLUG_CPU
843 extern void idle_task_exit(void);
845 static inline void idle_task_exit(void) {}
848 extern void sched_idle_next(void);
849 extern void set_user_nice(task_t
*p
, long nice
);
850 extern int task_prio(const task_t
*p
);
851 extern int task_nice(const task_t
*p
);
852 extern int can_nice(const task_t
*p
, const int nice
);
853 extern int task_curr(const task_t
*p
);
854 extern int idle_cpu(int cpu
);
855 extern int sched_setscheduler(struct task_struct
*, int, struct sched_param
*);
856 extern task_t
*idle_task(int cpu
);
861 * The default (Linux) execution domain.
863 extern struct exec_domain default_exec_domain
;
866 struct thread_info thread_info
;
867 unsigned long stack
[THREAD_SIZE
/sizeof(long)];
870 #ifndef __HAVE_ARCH_KSTACK_END
871 static inline int kstack_end(void *addr
)
873 /* Reliable end of stack detection:
874 * Some APM bios versions misalign the stack
876 return !(((unsigned long)addr
+sizeof(void*)-1) & (THREAD_SIZE
-sizeof(void*)));
880 extern union thread_union init_thread_union
;
881 extern struct task_struct init_task
;
883 extern struct mm_struct init_mm
;
885 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
886 extern struct task_struct
*find_task_by_pid_type(int type
, int pid
);
887 extern void set_special_pids(pid_t session
, pid_t pgrp
);
888 extern void __set_special_pids(pid_t session
, pid_t pgrp
);
890 /* per-UID process charging. */
891 extern struct user_struct
* alloc_uid(uid_t
);
892 static inline struct user_struct
*get_uid(struct user_struct
*u
)
894 atomic_inc(&u
->__count
);
897 extern void free_uid(struct user_struct
*);
898 extern void switch_uid(struct user_struct
*);
900 #include <asm/current.h>
902 extern void do_timer(struct pt_regs
*);
904 extern int FASTCALL(wake_up_state(struct task_struct
* tsk
, unsigned int state
));
905 extern int FASTCALL(wake_up_process(struct task_struct
* tsk
));
906 extern void FASTCALL(wake_up_new_task(struct task_struct
* tsk
,
907 unsigned long clone_flags
));
909 extern void kick_process(struct task_struct
*tsk
);
911 static inline void kick_process(struct task_struct
*tsk
) { }
913 extern void FASTCALL(sched_fork(task_t
* p
));
914 extern void FASTCALL(sched_exit(task_t
* p
));
916 extern int in_group_p(gid_t
);
917 extern int in_egroup_p(gid_t
);
919 extern void proc_caches_init(void);
920 extern void flush_signals(struct task_struct
*);
921 extern void flush_signal_handlers(struct task_struct
*, int force_default
);
922 extern int dequeue_signal(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
);
924 static inline int dequeue_signal_lock(struct task_struct
*tsk
, sigset_t
*mask
, siginfo_t
*info
)
929 spin_lock_irqsave(&tsk
->sighand
->siglock
, flags
);
930 ret
= dequeue_signal(tsk
, mask
, info
);
931 spin_unlock_irqrestore(&tsk
->sighand
->siglock
, flags
);
936 extern void block_all_signals(int (*notifier
)(void *priv
), void *priv
,
938 extern void unblock_all_signals(void);
939 extern void release_task(struct task_struct
* p
);
940 extern int send_sig_info(int, struct siginfo
*, struct task_struct
*);
941 extern int send_group_sig_info(int, struct siginfo
*, struct task_struct
*);
942 extern int force_sigsegv(int, struct task_struct
*);
943 extern int force_sig_info(int, struct siginfo
*, struct task_struct
*);
944 extern int __kill_pg_info(int sig
, struct siginfo
*info
, pid_t pgrp
);
945 extern int kill_pg_info(int, struct siginfo
*, pid_t
);
946 extern int kill_proc_info(int, struct siginfo
*, pid_t
);
947 extern void do_notify_parent(struct task_struct
*, int);
948 extern void force_sig(int, struct task_struct
*);
949 extern void force_sig_specific(int, struct task_struct
*);
950 extern int send_sig(int, struct task_struct
*, int);
951 extern void zap_other_threads(struct task_struct
*p
);
952 extern int kill_pg(pid_t
, int, int);
953 extern int kill_sl(pid_t
, int, int);
954 extern int kill_proc(pid_t
, int, int);
955 extern struct sigqueue
*sigqueue_alloc(void);
956 extern void sigqueue_free(struct sigqueue
*);
957 extern int send_sigqueue(int, struct sigqueue
*, struct task_struct
*);
958 extern int send_group_sigqueue(int, struct sigqueue
*, struct task_struct
*);
959 extern int do_sigaction(int, const struct k_sigaction
*, struct k_sigaction
*);
960 extern int do_sigaltstack(const stack_t __user
*, stack_t __user
*, unsigned long);
962 /* These can be the second arg to send_sig_info/send_group_sig_info. */
963 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
964 #define SEND_SIG_PRIV ((struct siginfo *) 1)
965 #define SEND_SIG_FORCED ((struct siginfo *) 2)
967 /* True if we are on the alternate signal stack. */
969 static inline int on_sig_stack(unsigned long sp
)
971 return (sp
- current
->sas_ss_sp
< current
->sas_ss_size
);
974 static inline int sas_ss_flags(unsigned long sp
)
976 return (current
->sas_ss_size
== 0 ? SS_DISABLE
977 : on_sig_stack(sp
) ? SS_ONSTACK
: 0);
981 #ifdef CONFIG_SECURITY
982 /* code is in security.c */
983 extern int capable(int cap
);
985 static inline int capable(int cap
)
987 if (cap_raised(current
->cap_effective
, cap
)) {
988 current
->flags
|= PF_SUPERPRIV
;
996 * Routines for handling mm_structs
998 extern struct mm_struct
* mm_alloc(void);
1000 /* mmdrop drops the mm and the page tables */
1001 extern void FASTCALL(__mmdrop(struct mm_struct
*));
1002 static inline void mmdrop(struct mm_struct
* mm
)
1004 if (atomic_dec_and_test(&mm
->mm_count
))
1008 /* mmput gets rid of the mappings and all user-space */
1009 extern void mmput(struct mm_struct
*);
1010 /* Grab a reference to a task's mm, if it is not already going away */
1011 extern struct mm_struct
*get_task_mm(struct task_struct
*task
);
1012 /* Remove the current tasks stale references to the old mm_struct */
1013 extern void mm_release(struct task_struct
*, struct mm_struct
*);
1015 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct
*, struct pt_regs
*);
1016 extern void flush_thread(void);
1017 extern void exit_thread(void);
1019 extern void exit_files(struct task_struct
*);
1020 extern void exit_signal(struct task_struct
*);
1021 extern void __exit_signal(struct task_struct
*);
1022 extern void exit_sighand(struct task_struct
*);
1023 extern void __exit_sighand(struct task_struct
*);
1024 extern void exit_itimers(struct signal_struct
*);
1026 extern NORET_TYPE
void do_group_exit(int);
1028 extern void daemonize(const char *, ...);
1029 extern int allow_signal(int);
1030 extern int disallow_signal(int);
1031 extern task_t
*child_reaper
;
1033 extern int do_execve(char *, char __user
* __user
*, char __user
* __user
*, struct pt_regs
*);
1034 extern long do_fork(unsigned long, unsigned long, struct pt_regs
*, unsigned long, int __user
*, int __user
*);
1035 task_t
*fork_idle(int);
1037 extern void set_task_comm(struct task_struct
*tsk
, char *from
);
1038 extern void get_task_comm(char *to
, struct task_struct
*tsk
);
1041 extern void wait_task_inactive(task_t
* p
);
1043 #define wait_task_inactive(p) do { } while (0)
1046 #define remove_parent(p) list_del_init(&(p)->sibling)
1047 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1049 #define REMOVE_LINKS(p) do { \
1050 if (thread_group_leader(p)) \
1051 list_del_init(&(p)->tasks); \
1055 #define SET_LINKS(p) do { \
1056 if (thread_group_leader(p)) \
1057 list_add_tail(&(p)->tasks,&init_task.tasks); \
1058 add_parent(p, (p)->parent); \
1061 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1062 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1064 #define for_each_process(p) \
1065 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1068 * Careful: do_each_thread/while_each_thread is a double loop so
1069 * 'break' will not work as expected - use goto instead.
1071 #define do_each_thread(g, t) \
1072 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1074 #define while_each_thread(g, t) \
1075 while ((t = next_thread(t)) != g)
1077 extern task_t
* FASTCALL(next_thread(const task_t
*p
));
1079 #define thread_group_leader(p) (p->pid == p->tgid)
1081 static inline int thread_group_empty(task_t
*p
)
1083 return list_empty(&p
->pids
[PIDTYPE_TGID
].pid_list
);
1086 #define delay_group_leader(p) \
1087 (thread_group_leader(p) && !thread_group_empty(p))
1089 extern void unhash_process(struct task_struct
*p
);
1092 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1093 * subscriptions and synchronises with wait4(). Also used in procfs.
1095 * Nests both inside and outside of read_lock(&tasklist_lock).
1096 * It must not be nested with write_lock_irq(&tasklist_lock),
1097 * neither inside nor outside.
1099 static inline void task_lock(struct task_struct
*p
)
1101 spin_lock(&p
->alloc_lock
);
1104 static inline void task_unlock(struct task_struct
*p
)
1106 spin_unlock(&p
->alloc_lock
);
1109 /* set thread flags in other task's structures
1110 * - see asm/thread_info.h for TIF_xxxx flags available
1112 static inline void set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1114 set_ti_thread_flag(tsk
->thread_info
,flag
);
1117 static inline void clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1119 clear_ti_thread_flag(tsk
->thread_info
,flag
);
1122 static inline int test_and_set_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1124 return test_and_set_ti_thread_flag(tsk
->thread_info
,flag
);
1127 static inline int test_and_clear_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1129 return test_and_clear_ti_thread_flag(tsk
->thread_info
,flag
);
1132 static inline int test_tsk_thread_flag(struct task_struct
*tsk
, int flag
)
1134 return test_ti_thread_flag(tsk
->thread_info
,flag
);
1137 static inline void set_tsk_need_resched(struct task_struct
*tsk
)
1139 set_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1142 static inline void clear_tsk_need_resched(struct task_struct
*tsk
)
1144 clear_tsk_thread_flag(tsk
,TIF_NEED_RESCHED
);
1147 static inline int signal_pending(struct task_struct
*p
)
1149 return unlikely(test_tsk_thread_flag(p
,TIF_SIGPENDING
));
1152 static inline int need_resched(void)
1154 return unlikely(test_thread_flag(TIF_NEED_RESCHED
));
1158 * cond_resched() and cond_resched_lock(): latency reduction via
1159 * explicit rescheduling in places that are safe. The return
1160 * value indicates whether a reschedule was done in fact.
1161 * cond_resched_lock() will drop the spinlock before scheduling,
1162 * cond_resched_softirq() will enable bhs before scheduling.
1164 extern int cond_resched(void);
1165 extern int cond_resched_lock(spinlock_t
* lock
);
1166 extern int cond_resched_softirq(void);
1169 * Does a critical section need to be broken due to another
1172 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1173 # define need_lockbreak(lock) ((lock)->break_lock)
1175 # define need_lockbreak(lock) 0
1179 * Does a critical section need to be broken due to another
1180 * task waiting or preemption being signalled:
1182 static inline int lock_need_resched(spinlock_t
*lock
)
1184 if (need_lockbreak(lock
) || need_resched())
1189 /* Reevaluate whether the task has signals pending delivery.
1190 This is required every time the blocked sigset_t changes.
1191 callers must hold sighand->siglock. */
1193 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct
*t
));
1194 extern void recalc_sigpending(void);
1196 extern void signal_wake_up(struct task_struct
*t
, int resume_stopped
);
1199 * Wrappers for p->thread_info->cpu access. No-op on UP.
1203 static inline unsigned int task_cpu(const struct task_struct
*p
)
1205 return p
->thread_info
->cpu
;
1208 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1210 p
->thread_info
->cpu
= cpu
;
1215 static inline unsigned int task_cpu(const struct task_struct
*p
)
1220 static inline void set_task_cpu(struct task_struct
*p
, unsigned int cpu
)
1224 #endif /* CONFIG_SMP */
1226 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1227 extern void arch_pick_mmap_layout(struct mm_struct
*mm
);
1229 static inline void arch_pick_mmap_layout(struct mm_struct
*mm
)
1231 mm
->mmap_base
= TASK_UNMAPPED_BASE
;
1232 mm
->get_unmapped_area
= arch_get_unmapped_area
;
1233 mm
->unmap_area
= arch_unmap_area
;
1237 extern long sched_setaffinity(pid_t pid
, cpumask_t new_mask
);
1238 extern long sched_getaffinity(pid_t pid
, cpumask_t
*mask
);
1240 #ifdef CONFIG_MAGIC_SYSRQ
1242 extern void normalize_rt_tasks(void);
1248 * Checks whether we need to enter the refrigerator
1249 * and returns 1 if we did so.
1252 extern void refrigerator(unsigned long);
1253 extern int freeze_processes(void);
1254 extern void thaw_processes(void);
1256 static inline int try_to_freeze(unsigned long refrigerator_flags
)
1258 if (unlikely(current
->flags
& PF_FREEZE
)) {
1259 refrigerator(refrigerator_flags
);
1265 static inline void refrigerator(unsigned long flag
) {}
1266 static inline int freeze_processes(void) { BUG(); return 0; }
1267 static inline void thaw_processes(void) {}
1269 static inline int try_to_freeze(unsigned long refrigerator_flags
)
1273 #endif /* CONFIG_PM */
1274 #endif /* __KERNEL__ */