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>
18 #include <asm/system.h>
19 #include <asm/semaphore.h>
21 #include <asm/ptrace.h>
23 #include <asm/cputime.h>
25 #include <linux/smp.h>
26 #include <linux/sem.h>
27 #include <linux/signal.h>
28 #include <linux/securebits.h>
29 #include <linux/fs_struct.h>
30 #include <linux/compiler.h>
31 #include <linux/completion.h>
32 #include <linux/pid.h>
33 #include <linux/percpu.h>
34 #include <linux/topology.h>
35 #include <linux/vs_base.h>
42 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
43 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
44 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
45 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
46 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
47 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
48 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
49 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
50 #define CLONE_THREAD 0x00010000 /* Same thread group? */
51 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
52 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
53 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
54 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
55 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
56 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
57 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
58 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
59 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
62 * List of flags we want to share for kernel threads,
63 * if only because they are not used by them anyway.
65 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
68 * These are the constant used to fake the fixed-point load-average
69 * counting. Some notes:
70 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
71 * a load-average precision of 10 bits integer + 11 bits fractional
72 * - if you want to count load-averages more often, you need more
73 * precision, or rounding will get you. With 2-second counting freq,
74 * the EXP_n values would be 1981, 2034 and 2043 if still using only
77 extern unsigned long avenrun[]; /* Load averages */
79 #define FSHIFT 11 /* nr of bits of precision */
80 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
81 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
82 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
83 #define EXP_5 2014 /* 1/exp(5sec/5min) */
84 #define EXP_15 2037 /* 1/exp(5sec/15min) */
86 #define CALC_LOAD(load,exp,n) \
88 load += n*(FIXED_1-exp); \
91 extern unsigned long total_forks;
92 extern int nr_threads;
94 DECLARE_PER_CPU(unsigned long, process_counts);
95 extern int nr_processes(void);
96 extern unsigned long nr_running(void);
97 extern unsigned long nr_uninterruptible(void);
98 extern unsigned long nr_iowait(void);
100 #include <linux/time.h>
101 #include <linux/param.h>
102 #include <linux/resource.h>
103 #include <linux/timer.h>
105 #include <asm/processor.h>
107 #define TASK_RUNNING 0
108 #define TASK_INTERRUPTIBLE 1
109 #define TASK_UNINTERRUPTIBLE 2
110 #define TASK_STOPPED 4
111 #define TASK_TRACED 8
112 #define EXIT_ZOMBIE 16
114 #define TASK_ONHOLD 64
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);
178 extern unsigned long cache_decay_ticks;
180 /* Attach to any functions which should be ignored in wchan output. */
181 #define __sched __attribute__((__section__(".sched.text")))
182 /* Is this address in the __sched functions? */
183 extern int in_sched_functions(unsigned long addr);
185 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
186 extern signed long FASTCALL(schedule_timeout(signed long timeout));
187 asmlinkage void schedule(void);
191 /* Maximum number of active map areas.. This is a random (large) number */
192 #define DEFAULT_MAX_MAP_COUNT 65536
194 extern int sysctl_max_map_count;
196 #include <linux/aio.h>
199 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
200 unsigned long, unsigned long);
202 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
203 unsigned long len, unsigned long pgoff,
204 unsigned long flags);
205 extern void arch_unmap_area(struct vm_area_struct *area);
206 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
210 struct vm_area_struct * mmap; /* list of VMAs */
211 struct rb_root mm_rb;
212 struct vm_area_struct * mmap_cache; /* last find_vma result */
213 unsigned long (*get_unmapped_area) (struct file *filp,
214 unsigned long addr, unsigned long len,
215 unsigned long pgoff, unsigned long flags);
216 void (*unmap_area) (struct vm_area_struct *area);
217 unsigned long mmap_base; /* base of mmap area */
218 unsigned long free_area_cache; /* first hole */
220 atomic_t mm_users; /* How many users with user space? */
221 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
222 int map_count; /* number of VMAs */
223 struct rw_semaphore mmap_sem;
224 spinlock_t page_table_lock; /* Protects page tables, mm->rss, mm->anon_rss */
226 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
227 * together off init_mm.mmlist, and are protected
231 unsigned long start_code, end_code, start_data, end_data;
232 unsigned long start_brk, brk, start_stack;
233 unsigned long arg_start, arg_end, env_start, env_end;
234 unsigned long rss, anon_rss, total_vm, locked_vm, shared_vm;
235 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
237 unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
240 cpumask_t cpu_vm_mask;
242 /* Architecture-specific MM context */
243 mm_context_t context;
244 struct vx_info *mm_vx_info;
246 /* Token based thrashing protection. */
247 unsigned long swap_token_time;
250 /* coredumping support */
252 struct completion *core_startup_done, core_done;
255 rwlock_t ioctx_list_lock;
256 struct kioctx *ioctx_list;
258 struct kioctx default_kioctx;
260 unsigned long hiwater_rss; /* High-water RSS usage */
261 unsigned long hiwater_vm; /* High-water virtual memory usage */
264 struct sighand_struct {
266 struct k_sigaction action[_NSIG];
271 * NOTE! "signal_struct" does not have it's own
272 * locking, because a shared signal_struct always
273 * implies a shared sighand_struct, so locking
274 * sighand_struct is always a proper superset of
275 * the locking of signal_struct.
277 struct signal_struct {
281 wait_queue_head_t wait_chldexit; /* for wait4() */
283 /* current thread group signal load-balancing target: */
286 /* shared signal handling: */
287 struct sigpending shared_pending;
289 /* thread group exit support */
292 * - notify group_exit_task when ->count is equal to notify_count
293 * - everyone except group_exit_task is stopped during signal delivery
294 * of fatal signals, group_exit_task processes the signal.
296 struct task_struct *group_exit_task;
299 /* thread group stop support, overloads group_exit_code too */
300 int group_stop_count;
301 unsigned int flags; /* see SIGNAL_* flags below */
303 /* POSIX.1b Interval Timers */
304 struct list_head posix_timers;
306 /* job control IDs */
310 /* boolean value for session group leader */
313 struct tty_struct *tty; /* NULL if no tty */
316 * Cumulative resource counters for dead threads in the group,
317 * and for reaped dead child processes forked by this group.
318 * Live threads maintain their own counters and add to these
319 * in __exit_signal, except for the group leader.
321 cputime_t utime, stime, cutime, cstime;
322 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
323 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
326 * We don't bother to synchronize most readers of this at all,
327 * because there is no reader checking a limit that actually needs
328 * to get both rlim_cur and rlim_max atomically, and either one
329 * alone is a single word that can safely be read normally.
330 * getrlimit/setrlimit use task_lock(current->group_leader) to
331 * protect this instead of the siglock, because they really
332 * have no need to disable irqs.
334 struct rlimit rlim[RLIM_NLIMITS];
338 * Bits in flags field of signal_struct.
340 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
341 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
342 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
343 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
347 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
348 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
349 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
350 * are inverted: lower p->prio value means higher priority.
352 * The MAX_USER_RT_PRIO value allows the actual maximum
353 * RT priority to be separate from the value exported to
354 * user-space. This allows kernel threads to set their
355 * priority to a value higher than any user task. Note:
356 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
359 #define MAX_USER_RT_PRIO 100
360 #define MAX_RT_PRIO MAX_USER_RT_PRIO
362 #define MAX_PRIO (MAX_RT_PRIO + 40)
364 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
367 * Some day this will be a full-fledged user tracking system..
370 atomic_t __count; /* reference count */
371 atomic_t processes; /* How many processes does this user have? */
372 atomic_t files; /* How many open files does this user have? */
373 atomic_t sigpending; /* How many pending signals does this user have? */
374 /* protected by mq_lock */
375 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
376 unsigned long locked_shm; /* How many pages of mlocked shm ? */
379 struct key *uid_keyring; /* UID specific keyring */
380 struct key *session_keyring; /* UID's default session keyring */
383 /* Hash table maintenance information */
384 struct list_head uidhash_list;
389 extern struct user_struct *find_user(xid_t, uid_t);
391 extern struct user_struct root_user;
392 #define INIT_USER (&root_user)
394 typedef struct prio_array prio_array_t;
395 struct backing_dev_info;
396 struct reclaim_state;
398 #ifdef CONFIG_SCHEDSTATS
400 /* cumulative counters */
401 unsigned long cpu_time, /* time spent on the cpu */
402 run_delay, /* time spent waiting on a runqueue */
403 pcnt; /* # of timeslices run on this cpu */
406 unsigned long last_arrival, /* when we last ran on a cpu */
407 last_queued; /* when we were last queued to run */
410 extern struct file_operations proc_schedstat_operations;
422 * sched-domains (multiprocessor balancing) declarations:
425 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
427 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
428 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
429 #define SD_BALANCE_EXEC 4 /* Balance on exec */
430 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
431 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
432 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
433 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
436 struct sched_group *next; /* Must be a circular list */
440 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
441 * single CPU. This is read only (except for setup, hotplug CPU).
443 unsigned long cpu_power;
446 struct sched_domain {
447 /* These fields must be setup */
448 struct sched_domain *parent; /* top domain must be null terminated */
449 struct sched_group *groups; /* the balancing groups of the domain */
450 cpumask_t span; /* span of all CPUs in this domain */
451 unsigned long min_interval; /* Minimum balance interval ms */
452 unsigned long max_interval; /* Maximum balance interval ms */
453 unsigned int busy_factor; /* less balancing by factor if busy */
454 unsigned int imbalance_pct; /* No balance until over watermark */
455 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
456 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
457 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
458 int flags; /* See SD_* */
460 /* Runtime fields. */
461 unsigned long last_balance; /* init to jiffies. units in jiffies */
462 unsigned int balance_interval; /* initialise to 1. units in ms. */
463 unsigned int nr_balance_failed; /* initialise to 0 */
465 #ifdef CONFIG_SCHEDSTATS
466 /* load_balance() stats */
467 unsigned long lb_cnt[MAX_IDLE_TYPES];
468 unsigned long lb_failed[MAX_IDLE_TYPES];
469 unsigned long lb_imbalance[MAX_IDLE_TYPES];
470 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
471 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
473 /* sched_balance_exec() stats */
474 unsigned long sbe_attempts;
475 unsigned long sbe_pushed;
477 /* try_to_wake_up() stats */
478 unsigned long ttwu_wake_affine;
479 unsigned long ttwu_wake_balance;
483 #ifdef ARCH_HAS_SCHED_DOMAIN
484 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
485 extern cpumask_t cpu_isolated_map;
486 extern void init_sched_build_groups(struct sched_group groups[],
487 cpumask_t span, int (*group_fn)(int cpu));
488 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
489 #endif /* ARCH_HAS_SCHED_DOMAIN */
490 #endif /* CONFIG_SMP */
493 struct io_context; /* See blkdev.h */
494 void exit_io_context(void);
496 #define NGROUPS_SMALL 32
497 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
501 gid_t small_block[NGROUPS_SMALL];
507 * get_group_info() must be called with the owning task locked (via task_lock())
508 * when task != current. The reason being that the vast majority of callers are
509 * looking at current->group_info, which can not be changed except by the
510 * current task. Changing current->group_info requires the task lock, too.
512 #define get_group_info(group_info) do { \
513 atomic_inc(&(group_info)->usage); \
516 #define put_group_info(group_info) do { \
517 if (atomic_dec_and_test(&(group_info)->usage)) \
518 groups_free(group_info); \
521 struct group_info *groups_alloc(int gidsetsize);
522 void groups_free(struct group_info *group_info);
523 int set_current_groups(struct group_info *group_info);
524 /* access the groups "array" with this macro */
525 #define GROUP_AT(gi, i) \
526 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
529 struct audit_context; /* See audit.c */
533 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
534 struct thread_info *thread_info;
536 unsigned long flags; /* per process flags, defined below */
537 unsigned long ptrace;
539 int lock_depth; /* Lock depth */
541 int prio, static_prio;
542 struct list_head run_list;
545 unsigned long sleep_avg;
546 unsigned long long timestamp, last_ran;
549 unsigned long policy;
550 cpumask_t cpus_allowed;
551 unsigned int time_slice, first_time_slice;
553 #ifdef CONFIG_SCHEDSTATS
554 struct sched_info sched_info;
557 struct list_head tasks;
559 * ptrace_list/ptrace_children forms the list of my children
560 * that were stolen by a ptracer.
562 struct list_head ptrace_children;
563 struct list_head ptrace_list;
565 struct mm_struct *mm, *active_mm;
568 struct linux_binfmt *binfmt;
570 int exit_code, exit_signal;
571 int pdeath_signal; /* The signal sent when the parent dies */
573 unsigned long personality;
578 * pointers to (original) parent process, youngest child, younger sibling,
579 * older sibling, respectively. (p->father can be replaced with
582 struct task_struct *real_parent; /* real parent process (when being debugged) */
583 struct task_struct *parent; /* parent process */
585 * children/sibling forms the list of my children plus the
586 * tasks I'm ptracing.
588 struct list_head children; /* list of my children */
589 struct list_head sibling; /* linkage in my parent's children list */
590 struct task_struct *group_leader; /* threadgroup leader */
592 /* PID/PID hash table linkage. */
593 struct pid pids[PIDTYPE_MAX];
595 struct completion *vfork_done; /* for vfork() */
596 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
597 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
599 unsigned long rt_priority;
600 unsigned long it_real_value, it_real_incr;
601 cputime_t it_virt_value, it_virt_incr;
602 cputime_t it_prof_value, it_prof_incr;
603 struct timer_list real_timer;
604 cputime_t utime, stime;
605 unsigned long nvcsw, nivcsw; /* context switch counts */
606 struct timespec start_time;
607 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
608 unsigned long min_flt, maj_flt;
609 /* process credentials */
610 uid_t uid,euid,suid,fsuid;
611 gid_t gid,egid,sgid,fsgid;
612 struct group_info *group_info;
613 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
614 unsigned keep_capabilities:1;
615 struct user_struct *user;
617 struct key *session_keyring; /* keyring inherited over fork */
618 struct key *process_keyring; /* keyring private to this process (CLONE_THREAD) */
619 struct key *thread_keyring; /* keyring private to this thread */
621 int oomkilladj; /* OOM kill score adjustment (bit shift). */
622 char comm[TASK_COMM_LEN];
623 /* file system info */
624 int link_count, total_link_count;
626 struct sysv_sem sysvsem;
627 /* CPU-specific state of this task */
628 struct thread_struct thread;
629 /* filesystem information */
630 struct fs_struct *fs;
631 /* open file information */
632 struct files_struct *files;
634 struct namespace *namespace;
635 /* signal handlers */
636 struct signal_struct *signal;
637 struct sighand_struct *sighand;
639 sigset_t blocked, real_blocked;
640 struct sigpending pending;
642 unsigned long sas_ss_sp;
644 int (*notifier)(void *priv);
646 sigset_t *notifier_mask;
649 struct audit_context *audit_context;
651 /* vserver context data */
653 struct vx_info *vx_info;
655 /* vserver network data */
657 struct nx_info *nx_info;
659 /* Thread group tracking */
662 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
663 spinlock_t alloc_lock;
664 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
665 spinlock_t proc_lock;
666 /* context-switch lock */
667 spinlock_t switch_lock;
669 /* journalling filesystem info */
673 struct reclaim_state *reclaim_state;
675 struct dentry *proc_dentry;
676 struct backing_dev_info *backing_dev_info;
678 struct io_context *io_context;
680 unsigned long ptrace_message;
681 siginfo_t *last_siginfo; /* For ptrace use. */
683 * current io wait handle: wait queue entry to use for io waits
684 * If this thread is processing aio, this points at the waitqueue
685 * inside the currently handled kiocb. It may be NULL (i.e. default
686 * to a stack based synchronous wait) if its doing sync IO.
688 wait_queue_t *io_wait;
689 /* i/o counters(bytes read/written, #syscalls */
690 u64 rchar, wchar, syscr, syscw;
691 #if defined(CONFIG_BSD_PROCESS_ACCT)
692 u64 acct_rss_mem1; /* accumulated rss usage */
693 u64 acct_vm_mem1; /* accumulated virtual memory usage */
694 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
697 struct mempolicy *mempolicy;
702 static inline pid_t process_group(struct task_struct *tsk)
704 return tsk->signal->pgrp;
708 * pid_alive - check that a task structure is not stale
709 * @p: Task structure to be checked.
711 * Test if a process is not yet dead (at most zombie state)
712 * If pid_alive fails, then pointers within the task structure
713 * can be stale and must not be dereferenced.
715 static inline int pid_alive(struct task_struct *p)
717 return p->pids[PIDTYPE_PID].nr != 0;
720 extern void free_task(struct task_struct *tsk);
721 extern void __put_task_struct(struct task_struct *tsk);
722 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
723 #define put_task_struct(tsk) \
724 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
729 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
730 /* Not implemented yet, only for 486*/
731 #define PF_STARTING 0x00000002 /* being created */
732 #define PF_EXITING 0x00000004 /* getting shut down */
733 #define PF_DEAD 0x00000008 /* Dead */
734 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
735 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
736 #define PF_DUMPCORE 0x00000200 /* dumped core */
737 #define PF_SIGNALED 0x00000400 /* killed by a signal */
738 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
739 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
740 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
741 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
742 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
743 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
744 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
745 #define PF_KSWAPD 0x00040000 /* I am kswapd */
746 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
747 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
748 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
749 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
752 * Only the _current_ task can read/write to tsk->flags, but other
753 * tasks can access tsk->flags in readonly mode for example
754 * with tsk_used_math (like during threaded core dumping).
755 * There is however an exception to this rule during ptrace
756 * or during fork: the ptracer task is allowed to write to the
757 * child->flags of its traced child (same goes for fork, the parent
758 * can write to the child->flags), because we're guaranteed the
759 * child is not running and in turn not changing child->flags
760 * at the same time the parent does it.
762 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
763 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
764 #define clear_used_math() clear_stopped_child_used_math(current)
765 #define set_used_math() set_stopped_child_used_math(current)
766 #define conditional_stopped_child_used_math(condition, child) \
767 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
768 #define conditional_used_math(condition) \
769 conditional_stopped_child_used_math(condition, current)
770 #define copy_to_stopped_child_used_math(child) \
771 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
772 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
773 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
774 #define used_math() tsk_used_math(current)
777 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
779 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
781 if (!cpus_intersects(new_mask, cpu_online_map))
787 extern unsigned long long sched_clock(void);
789 /* sched_exec is called by processes performing an exec */
791 extern void sched_exec(void);
793 #define sched_exec() {}
796 #ifdef CONFIG_HOTPLUG_CPU
797 extern void idle_task_exit(void);
799 static inline void idle_task_exit(void) {}
802 extern void sched_idle_next(void);
803 extern void set_user_nice(task_t *p, long nice);
804 extern int task_prio(const task_t *p);
805 extern int task_nice(const task_t *p);
806 extern int task_curr(const task_t *p);
807 extern int idle_cpu(int cpu);
808 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
809 extern task_t *idle_task(int cpu);
814 * The default (Linux) execution domain.
816 extern struct exec_domain default_exec_domain;
819 struct thread_info thread_info;
820 unsigned long stack[THREAD_SIZE/sizeof(long)];
823 #ifndef __HAVE_ARCH_KSTACK_END
824 static inline int kstack_end(void *addr)
826 /* Reliable end of stack detection:
827 * Some APM bios versions misalign the stack
829 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
833 extern union thread_union init_thread_union;
834 extern struct task_struct init_task;
836 extern struct mm_struct init_mm;
839 #define find_task_by_real_pid(nr) \
840 find_task_by_pid_type(PIDTYPE_PID, nr)
841 #define find_task_by_pid(nr) \
842 find_task_by_pid_type(PIDTYPE_PID, \
845 extern struct task_struct *find_task_by_pid_type(int type, int pid);
846 extern void set_special_pids(pid_t session, pid_t pgrp);
847 extern void __set_special_pids(pid_t session, pid_t pgrp);
849 /* per-UID process charging. */
850 extern struct user_struct * alloc_uid(xid_t, uid_t);
851 static inline struct user_struct *get_uid(struct user_struct *u)
853 atomic_inc(&u->__count);
856 extern void free_uid(struct user_struct *);
857 extern void switch_uid(struct user_struct *);
859 #include <asm/current.h>
861 extern void do_timer(struct pt_regs *);
863 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
864 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
865 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
866 unsigned long clone_flags));
868 extern void kick_process(struct task_struct *tsk);
870 static inline void kick_process(struct task_struct *tsk) { }
872 extern void FASTCALL(sched_fork(task_t * p));
873 extern void FASTCALL(sched_exit(task_t * p));
875 extern int in_group_p(gid_t);
876 extern int in_egroup_p(gid_t);
878 extern void proc_caches_init(void);
879 extern void flush_signals(struct task_struct *);
880 extern void flush_signal_handlers(struct task_struct *, int force_default);
881 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
883 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
888 spin_lock_irqsave(&tsk->sighand->siglock, flags);
889 ret = dequeue_signal(tsk, mask, info);
890 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
895 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
897 extern void unblock_all_signals(void);
898 extern void release_task(struct task_struct * p);
899 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
900 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
901 extern int force_sigsegv(int, struct task_struct *);
902 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
903 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
904 extern int kill_pg_info(int, struct siginfo *, pid_t);
905 extern int kill_proc_info(int, struct siginfo *, pid_t);
906 extern void do_notify_parent(struct task_struct *, int);
907 extern void force_sig(int, struct task_struct *);
908 extern void force_sig_specific(int, struct task_struct *);
909 extern int send_sig(int, struct task_struct *, int);
910 extern void zap_other_threads(struct task_struct *p);
911 extern int kill_pg(pid_t, int, int);
912 extern int kill_sl(pid_t, int, int);
913 extern int kill_proc(pid_t, int, int);
914 extern struct sigqueue *sigqueue_alloc(void);
915 extern void sigqueue_free(struct sigqueue *);
916 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
917 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
918 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
919 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
921 /* These can be the second arg to send_sig_info/send_group_sig_info. */
922 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
923 #define SEND_SIG_PRIV ((struct siginfo *) 1)
924 #define SEND_SIG_FORCED ((struct siginfo *) 2)
926 /* True if we are on the alternate signal stack. */
928 static inline int on_sig_stack(unsigned long sp)
930 return (sp - current->sas_ss_sp < current->sas_ss_size);
933 static inline int sas_ss_flags(unsigned long sp)
935 return (current->sas_ss_size == 0 ? SS_DISABLE
936 : on_sig_stack(sp) ? SS_ONSTACK : 0);
940 #ifdef CONFIG_SECURITY
941 /* code is in security.c */
942 extern int capable(int cap);
943 extern int vx_capable(int cap, int ccap);
945 static inline int capable(int cap)
947 if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
949 if (cap_raised(current->cap_effective, cap)) {
950 current->flags |= PF_SUPERPRIV;
956 static inline int vx_capable(int cap, int ccap)
958 if (cap_raised(current->cap_effective, cap) &&
960 current->flags |= PF_SUPERPRIV;
968 * Routines for handling mm_structs
970 extern struct mm_struct * mm_alloc(void);
972 /* mmdrop drops the mm and the page tables */
973 extern void FASTCALL(__mmdrop(struct mm_struct *));
974 static inline void mmdrop(struct mm_struct * mm)
976 if (atomic_dec_and_test(&mm->mm_count))
980 /* mmput gets rid of the mappings and all user-space */
981 extern void mmput(struct mm_struct *);
982 /* Grab a reference to a task's mm, if it is not already going away */
983 extern struct mm_struct *get_task_mm(struct task_struct *task);
984 /* Remove the current tasks stale references to the old mm_struct */
985 extern void mm_release(struct task_struct *, struct mm_struct *);
987 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
988 extern void flush_thread(void);
989 extern void exit_thread(void);
991 extern void exit_mm(struct task_struct *);
992 extern void exit_files(struct task_struct *);
993 extern void exit_signal(struct task_struct *);
994 extern void __exit_signal(struct task_struct *);
995 extern void exit_sighand(struct task_struct *);
996 extern void __exit_sighand(struct task_struct *);
997 extern void exit_itimers(struct signal_struct *);
999 extern NORET_TYPE void do_group_exit(int);
1001 extern void reparent_to_init(void);
1002 extern void daemonize(const char *, ...);
1003 extern int allow_signal(int);
1004 extern int disallow_signal(int);
1005 extern task_t *child_reaper;
1007 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1008 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1009 task_t *fork_idle(int);
1011 extern void set_task_comm(struct task_struct *tsk, char *from);
1012 extern void get_task_comm(char *to, struct task_struct *tsk);
1015 extern void wait_task_inactive(task_t * p);
1017 #define wait_task_inactive(p) do { } while (0)
1020 #define remove_parent(p) list_del_init(&(p)->sibling)
1021 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1023 #define REMOVE_LINKS(p) do { \
1024 if (thread_group_leader(p)) \
1025 list_del_init(&(p)->tasks); \
1029 #define SET_LINKS(p) do { \
1030 if (thread_group_leader(p)) \
1031 list_add_tail(&(p)->tasks,&init_task.tasks); \
1032 add_parent(p, (p)->parent); \
1035 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1036 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1038 #define for_each_process(p) \
1039 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1042 * Careful: do_each_thread/while_each_thread is a double loop so
1043 * 'break' will not work as expected - use goto instead.
1045 #define do_each_thread(g, t) \
1046 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1048 #define while_each_thread(g, t) \
1049 while ((t = next_thread(t)) != g)
1051 extern task_t * FASTCALL(next_thread(const task_t *p));
1053 #define thread_group_leader(p) (p->pid == p->tgid)
1055 static inline int thread_group_empty(task_t *p)
1057 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1060 #define delay_group_leader(p) \
1061 (thread_group_leader(p) && !thread_group_empty(p))
1063 extern void unhash_process(struct task_struct *p);
1066 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1067 * subscriptions and synchronises with wait4(). Also used in procfs.
1069 * Nests both inside and outside of read_lock(&tasklist_lock).
1070 * It must not be nested with write_lock_irq(&tasklist_lock),
1071 * neither inside nor outside.
1073 static inline void task_lock(struct task_struct *p)
1075 spin_lock(&p->alloc_lock);
1078 static inline void task_unlock(struct task_struct *p)
1080 spin_unlock(&p->alloc_lock);
1083 /* set thread flags in other task's structures
1084 * - see asm/thread_info.h for TIF_xxxx flags available
1086 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1088 set_ti_thread_flag(tsk->thread_info,flag);
1091 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1093 clear_ti_thread_flag(tsk->thread_info,flag);
1096 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1098 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1101 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1103 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1106 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1108 return test_ti_thread_flag(tsk->thread_info,flag);
1111 static inline void set_tsk_need_resched(struct task_struct *tsk)
1113 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1116 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1118 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1121 static inline int signal_pending(struct task_struct *p)
1123 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1126 static inline int need_resched(void)
1128 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1132 * cond_resched() and cond_resched_lock(): latency reduction via
1133 * explicit rescheduling in places that are safe. The return
1134 * value indicates whether a reschedule was done in fact.
1135 * cond_resched_lock() will drop the spinlock before scheduling,
1136 * cond_resched_softirq() will enable bhs before scheduling.
1138 extern int cond_resched(void);
1139 extern int cond_resched_lock(spinlock_t * lock);
1140 extern int cond_resched_softirq(void);
1143 * Does a critical section need to be broken due to another
1146 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1147 # define need_lockbreak(lock) ((lock)->break_lock)
1149 # define need_lockbreak(lock) 0
1153 * Does a critical section need to be broken due to another
1154 * task waiting or preemption being signalled:
1156 static inline int lock_need_resched(spinlock_t *lock)
1158 if (need_lockbreak(lock) || need_resched())
1163 /* Reevaluate whether the task has signals pending delivery.
1164 This is required every time the blocked sigset_t changes.
1165 callers must hold sighand->siglock. */
1167 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1168 extern void recalc_sigpending(void);
1170 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1173 * Wrappers for p->thread_info->cpu access. No-op on UP.
1177 static inline unsigned int task_cpu(const struct task_struct *p)
1179 return p->thread_info->cpu;
1182 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1184 p->thread_info->cpu = cpu;
1189 static inline unsigned int task_cpu(const struct task_struct *p)
1194 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1198 #endif /* CONFIG_SMP */
1200 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1201 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1203 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1205 mm->mmap_base = TASK_UNMAPPED_BASE;
1206 mm->get_unmapped_area = arch_get_unmapped_area;
1207 mm->unmap_area = arch_unmap_area;
1211 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1212 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1214 #ifdef CONFIG_MAGIC_SYSRQ
1216 extern void normalize_rt_tasks(void);
1222 * Checks whether we need to enter the refrigerator
1223 * and returns 1 if we did so.
1226 extern void refrigerator(unsigned long);
1227 extern int freeze_processes(void);
1228 extern void thaw_processes(void);
1230 static inline int try_to_freeze(unsigned long refrigerator_flags)
1232 if (unlikely(current->flags & PF_FREEZE)) {
1233 refrigerator(refrigerator_flags);
1239 static inline void refrigerator(unsigned long flag) {}
1240 static inline int freeze_processes(void) { BUG(); return 0; }
1241 static inline void thaw_processes(void) {}
1243 static inline int try_to_freeze(unsigned long refrigerator_flags)
1247 #endif /* CONFIG_PM */
1248 #endif /* __KERNEL__ */