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>
37 #include <linux/vs_base.h>
44 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
45 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
46 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
47 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
48 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
49 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
50 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
51 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
52 #define CLONE_THREAD 0x00010000 /* Same thread group? */
53 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
54 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
55 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
56 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
57 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
58 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
59 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
60 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
61 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
64 * List of flags we want to share for kernel threads,
65 * if only because they are not used by them anyway.
67 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
70 * These are the constant used to fake the fixed-point load-average
71 * counting. Some notes:
72 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
73 * a load-average precision of 10 bits integer + 11 bits fractional
74 * - if you want to count load-averages more often, you need more
75 * precision, or rounding will get you. With 2-second counting freq,
76 * the EXP_n values would be 1981, 2034 and 2043 if still using only
79 extern unsigned long avenrun[]; /* Load averages */
81 #define FSHIFT 11 /* nr of bits of precision */
82 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
83 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
84 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
85 #define EXP_5 2014 /* 1/exp(5sec/5min) */
86 #define EXP_15 2037 /* 1/exp(5sec/15min) */
88 #define CALC_LOAD(load,exp,n) \
90 load += n*(FIXED_1-exp); \
93 extern unsigned long total_forks;
94 extern int nr_threads;
96 DECLARE_PER_CPU(unsigned long, process_counts);
97 extern int nr_processes(void);
98 extern unsigned long nr_running(void);
99 extern unsigned long nr_uninterruptible(void);
100 extern unsigned long nr_iowait(void);
102 #include <linux/time.h>
103 #include <linux/param.h>
104 #include <linux/resource.h>
105 #include <linux/timer.h>
107 #include <asm/processor.h>
109 #define TASK_RUNNING 0
110 #define TASK_INTERRUPTIBLE 1
111 #define TASK_UNINTERRUPTIBLE 2
112 #define TASK_STOPPED 4
113 #define TASK_TRACED 8
114 #define EXIT_ZOMBIE 16
116 #define TASK_ONHOLD 64
118 #define __set_task_state(tsk, state_value) \
119 do { (tsk)->state = (state_value); } while (0)
120 #define set_task_state(tsk, state_value) \
121 set_mb((tsk)->state, (state_value))
123 #define __set_current_state(state_value) \
124 do { current->state = (state_value); } while (0)
125 #define set_current_state(state_value) \
126 set_mb(current->state, (state_value))
128 /* Task command name length */
129 #define TASK_COMM_LEN 16
132 * Scheduling policies
134 #define SCHED_NORMAL 0
144 #include <linux/spinlock.h>
147 * This serializes "schedule()" and also protects
148 * the run-queue from deletions/modifications (but
149 * _adding_ to the beginning of the run-queue has
152 extern rwlock_t tasklist_lock;
153 extern spinlock_t mmlist_lock;
155 typedef struct task_struct task_t;
157 extern void sched_init(void);
158 extern void sched_init_smp(void);
159 extern void init_idle(task_t *idle, int cpu);
161 extern cpumask_t nohz_cpu_mask;
163 extern void show_state(void);
164 extern void show_regs(struct pt_regs *);
167 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
168 * task), SP is the stack pointer of the first frame that should be shown in the back
169 * trace (or NULL if the entire call-chain of the task should be shown).
171 extern void show_stack(struct task_struct *task, unsigned long *sp);
173 void io_schedule(void);
174 long io_schedule_timeout(long timeout);
176 extern void cpu_init (void);
177 extern void trap_init(void);
178 extern void update_process_times(int user);
179 extern void scheduler_tick(void);
181 /* Attach to any functions which should be ignored in wchan output. */
182 #define __sched __attribute__((__section__(".sched.text")))
183 /* Is this address in the __sched functions? */
184 extern int in_sched_functions(unsigned long addr);
186 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
187 extern signed long FASTCALL(schedule_timeout(signed long timeout));
188 asmlinkage void schedule(void);
192 /* Maximum number of active map areas.. This is a random (large) number */
193 #define DEFAULT_MAX_MAP_COUNT 65536
195 extern int sysctl_max_map_count;
197 #include <linux/aio.h>
200 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
201 unsigned long, unsigned long);
203 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
204 unsigned long len, unsigned long pgoff,
205 unsigned long flags);
206 extern void arch_unmap_area(struct vm_area_struct *area);
207 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
209 #define __set_mm_counter(mm, member, value) (mm)->_##member = (value)
210 #define set_mm_counter(mm, member, value) vx_ ## member ## pages_sub((mm), ((mm)->_##member - value))
211 #define get_mm_counter(mm, member) ((mm)->_##member)
212 #define add_mm_counter(mm, member, value) vx_ ## member ## pages_add((mm), (value))
213 #define inc_mm_counter(mm, member) vx_ ## member ## pages_inc((mm))
214 #define dec_mm_counter(mm, member) vx_ ## member ## pages_dec((mm))
215 typedef unsigned long mm_counter_t;
218 struct vm_area_struct * mmap; /* list of VMAs */
219 struct rb_root mm_rb;
220 struct vm_area_struct * mmap_cache; /* last find_vma result */
221 unsigned long (*get_unmapped_area) (struct file *filp,
222 unsigned long addr, unsigned long len,
223 unsigned long pgoff, unsigned long flags);
224 void (*unmap_area) (struct vm_area_struct *area);
225 unsigned long mmap_base; /* base of mmap area */
226 unsigned long free_area_cache; /* first hole */
228 atomic_t mm_users; /* How many users with user space? */
229 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
230 int map_count; /* number of VMAs */
231 struct rw_semaphore mmap_sem;
232 spinlock_t page_table_lock; /* Protects page tables and some counters */
234 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
235 * together off init_mm.mmlist, and are protected
239 unsigned long start_code, end_code, start_data, end_data;
240 unsigned long start_brk, brk, start_stack;
241 unsigned long arg_start, arg_end, env_start, env_end;
242 unsigned long total_vm, locked_vm, shared_vm;
243 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
245 /* Special counters protected by the page_table_lock */
247 mm_counter_t _anon_rss;
249 unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
252 cpumask_t cpu_vm_mask;
254 /* Architecture-specific MM context */
255 mm_context_t context;
256 struct vx_info *mm_vx_info;
258 /* Token based thrashing protection. */
259 unsigned long swap_token_time;
262 /* coredumping support */
264 struct completion *core_startup_done, core_done;
267 rwlock_t ioctx_list_lock;
268 struct kioctx *ioctx_list;
270 struct kioctx default_kioctx;
272 unsigned long hiwater_rss; /* High-water RSS usage */
273 unsigned long hiwater_vm; /* High-water virtual memory usage */
276 struct sighand_struct {
278 struct k_sigaction action[_NSIG];
283 * NOTE! "signal_struct" does not have it's own
284 * locking, because a shared signal_struct always
285 * implies a shared sighand_struct, so locking
286 * sighand_struct is always a proper superset of
287 * the locking of signal_struct.
289 struct signal_struct {
293 wait_queue_head_t wait_chldexit; /* for wait4() */
295 /* current thread group signal load-balancing target: */
298 /* shared signal handling: */
299 struct sigpending shared_pending;
301 /* thread group exit support */
304 * - notify group_exit_task when ->count is equal to notify_count
305 * - everyone except group_exit_task is stopped during signal delivery
306 * of fatal signals, group_exit_task processes the signal.
308 struct task_struct *group_exit_task;
311 /* thread group stop support, overloads group_exit_code too */
312 int group_stop_count;
313 unsigned int flags; /* see SIGNAL_* flags below */
315 /* POSIX.1b Interval Timers */
316 struct list_head posix_timers;
318 /* ITIMER_REAL timer for the process */
319 struct timer_list real_timer;
320 unsigned long it_real_value, it_real_incr;
322 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
323 cputime_t it_prof_expires, it_virt_expires;
324 cputime_t it_prof_incr, it_virt_incr;
326 /* job control IDs */
330 /* boolean value for session group leader */
333 struct tty_struct *tty; /* NULL if no tty */
336 * Cumulative resource counters for dead threads in the group,
337 * and for reaped dead child processes forked by this group.
338 * Live threads maintain their own counters and add to these
339 * in __exit_signal, except for the group leader.
341 cputime_t utime, stime, cutime, cstime;
342 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
343 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
346 * Cumulative ns of scheduled CPU time for dead threads in the
347 * group, not including a zombie group leader. (This only differs
348 * from jiffies_to_ns(utime + stime) if sched_clock uses something
349 * other than jiffies.)
351 unsigned long long sched_time;
354 * We don't bother to synchronize most readers of this at all,
355 * because there is no reader checking a limit that actually needs
356 * to get both rlim_cur and rlim_max atomically, and either one
357 * alone is a single word that can safely be read normally.
358 * getrlimit/setrlimit use task_lock(current->group_leader) to
359 * protect this instead of the siglock, because they really
360 * have no need to disable irqs.
362 struct rlimit rlim[RLIM_NLIMITS];
364 struct list_head cpu_timers[3];
366 /* keep the process-shared keyrings here so that they do the right
367 * thing in threads created with CLONE_THREAD */
369 struct key *session_keyring; /* keyring inherited over fork */
370 struct key *process_keyring; /* keyring private to this process */
375 * Bits in flags field of signal_struct.
377 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
378 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
379 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
380 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
384 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
385 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
386 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
387 * are inverted: lower p->prio value means higher priority.
389 * The MAX_USER_RT_PRIO value allows the actual maximum
390 * RT priority to be separate from the value exported to
391 * user-space. This allows kernel threads to set their
392 * priority to a value higher than any user task. Note:
393 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
396 #define MAX_USER_RT_PRIO 100
397 #define MAX_RT_PRIO MAX_USER_RT_PRIO
399 #define MAX_PRIO (MAX_RT_PRIO + 40)
401 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
404 * Some day this will be a full-fledged user tracking system..
407 atomic_t __count; /* reference count */
408 atomic_t processes; /* How many processes does this user have? */
409 atomic_t files; /* How many open files does this user have? */
410 atomic_t sigpending; /* How many pending signals does this user have? */
411 /* protected by mq_lock */
412 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
413 unsigned long locked_shm; /* How many pages of mlocked shm ? */
416 struct key *uid_keyring; /* UID specific keyring */
417 struct key *session_keyring; /* UID's default session keyring */
420 /* Hash table maintenance information */
421 struct list_head uidhash_list;
426 extern struct user_struct *find_user(xid_t, uid_t);
428 extern struct user_struct root_user;
429 #define INIT_USER (&root_user)
431 typedef struct prio_array prio_array_t;
432 struct backing_dev_info;
433 struct reclaim_state;
435 #ifdef CONFIG_SCHEDSTATS
437 /* cumulative counters */
438 unsigned long cpu_time, /* time spent on the cpu */
439 run_delay, /* time spent waiting on a runqueue */
440 pcnt; /* # of timeslices run on this cpu */
443 unsigned long last_arrival, /* when we last ran on a cpu */
444 last_queued; /* when we were last queued to run */
447 extern struct file_operations proc_schedstat_operations;
459 * sched-domains (multiprocessor balancing) declarations:
462 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
464 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
465 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
466 #define SD_BALANCE_EXEC 4 /* Balance on exec */
467 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
468 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
469 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
470 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
473 struct sched_group *next; /* Must be a circular list */
477 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
478 * single CPU. This is read only (except for setup, hotplug CPU).
480 unsigned long cpu_power;
483 struct sched_domain {
484 /* These fields must be setup */
485 struct sched_domain *parent; /* top domain must be null terminated */
486 struct sched_group *groups; /* the balancing groups of the domain */
487 cpumask_t span; /* span of all CPUs in this domain */
488 unsigned long min_interval; /* Minimum balance interval ms */
489 unsigned long max_interval; /* Maximum balance interval ms */
490 unsigned int busy_factor; /* less balancing by factor if busy */
491 unsigned int imbalance_pct; /* No balance until over watermark */
492 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
493 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
494 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
495 int flags; /* See SD_* */
497 /* Runtime fields. */
498 unsigned long last_balance; /* init to jiffies. units in jiffies */
499 unsigned int balance_interval; /* initialise to 1. units in ms. */
500 unsigned int nr_balance_failed; /* initialise to 0 */
502 #ifdef CONFIG_SCHEDSTATS
503 /* load_balance() stats */
504 unsigned long lb_cnt[MAX_IDLE_TYPES];
505 unsigned long lb_failed[MAX_IDLE_TYPES];
506 unsigned long lb_balanced[MAX_IDLE_TYPES];
507 unsigned long lb_imbalance[MAX_IDLE_TYPES];
508 unsigned long lb_gained[MAX_IDLE_TYPES];
509 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
510 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
511 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
513 /* Active load balancing */
514 unsigned long alb_cnt;
515 unsigned long alb_failed;
516 unsigned long alb_pushed;
518 /* sched_balance_exec() stats */
519 unsigned long sbe_attempts;
520 unsigned long sbe_pushed;
522 /* try_to_wake_up() stats */
523 unsigned long ttwu_wake_remote;
524 unsigned long ttwu_move_affine;
525 unsigned long ttwu_move_balance;
529 #ifdef ARCH_HAS_SCHED_DOMAIN
530 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
531 extern cpumask_t cpu_isolated_map;
532 extern void init_sched_build_groups(struct sched_group groups[],
533 cpumask_t span, int (*group_fn)(int cpu));
534 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
535 #endif /* ARCH_HAS_SCHED_DOMAIN */
536 #endif /* CONFIG_SMP */
539 struct io_context; /* See blkdev.h */
540 void exit_io_context(void);
543 #define NGROUPS_SMALL 32
544 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
548 gid_t small_block[NGROUPS_SMALL];
554 * get_group_info() must be called with the owning task locked (via task_lock())
555 * when task != current. The reason being that the vast majority of callers are
556 * looking at current->group_info, which can not be changed except by the
557 * current task. Changing current->group_info requires the task lock, too.
559 #define get_group_info(group_info) do { \
560 atomic_inc(&(group_info)->usage); \
563 #define put_group_info(group_info) do { \
564 if (atomic_dec_and_test(&(group_info)->usage)) \
565 groups_free(group_info); \
568 struct group_info *groups_alloc(int gidsetsize);
569 void groups_free(struct group_info *group_info);
570 int set_current_groups(struct group_info *group_info);
571 /* access the groups "array" with this macro */
572 #define GROUP_AT(gi, i) \
573 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
576 struct audit_context; /* See audit.c */
580 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
581 struct thread_info *thread_info;
583 unsigned long flags; /* per process flags, defined below */
584 unsigned long ptrace;
586 int lock_depth; /* BKL lock depth */
588 int prio, static_prio;
589 struct list_head run_list;
592 unsigned long sleep_avg;
593 unsigned long long timestamp, last_ran;
594 unsigned long long sched_time; /* sched_clock time spent running */
597 unsigned long policy;
598 cpumask_t cpus_allowed;
599 unsigned int time_slice, first_time_slice;
601 #ifdef CONFIG_SCHEDSTATS
602 struct sched_info sched_info;
605 struct list_head tasks;
607 * ptrace_list/ptrace_children forms the list of my children
608 * that were stolen by a ptracer.
610 struct list_head ptrace_children;
611 struct list_head ptrace_list;
613 struct mm_struct *mm, *active_mm;
616 struct linux_binfmt *binfmt;
618 int exit_code, exit_signal;
619 int pdeath_signal; /* The signal sent when the parent dies */
621 unsigned long personality;
626 * pointers to (original) parent process, youngest child, younger sibling,
627 * older sibling, respectively. (p->father can be replaced with
630 struct task_struct *real_parent; /* real parent process (when being debugged) */
631 struct task_struct *parent; /* parent process */
633 * children/sibling forms the list of my children plus the
634 * tasks I'm ptracing.
636 struct list_head children; /* list of my children */
637 struct list_head sibling; /* linkage in my parent's children list */
638 struct task_struct *group_leader; /* threadgroup leader */
640 /* PID/PID hash table linkage. */
641 struct pid pids[PIDTYPE_MAX];
643 struct completion *vfork_done; /* for vfork() */
644 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
645 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
647 unsigned long rt_priority;
648 cputime_t utime, stime;
649 unsigned long nvcsw, nivcsw; /* context switch counts */
650 struct timespec start_time;
651 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
652 unsigned long min_flt, maj_flt;
654 cputime_t it_prof_expires, it_virt_expires;
655 unsigned long long it_sched_expires;
656 struct list_head cpu_timers[3];
658 /* process credentials */
659 uid_t uid,euid,suid,fsuid;
660 gid_t gid,egid,sgid,fsgid;
661 struct group_info *group_info;
662 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
663 unsigned keep_capabilities:1;
664 struct user_struct *user;
666 struct key *thread_keyring; /* keyring private to this thread */
668 int oomkilladj; /* OOM kill score adjustment (bit shift). */
669 char comm[TASK_COMM_LEN]; /* executable name excluding path
670 - access with [gs]et_task_comm (which lock
672 - initialized normally by flush_old_exec */
673 /* file system info */
674 int link_count, total_link_count;
676 struct sysv_sem sysvsem;
677 /* CPU-specific state of this task */
678 struct thread_struct thread;
679 /* filesystem information */
680 struct fs_struct *fs;
681 /* open file information */
682 struct files_struct *files;
684 struct namespace *namespace;
685 /* signal handlers */
686 struct signal_struct *signal;
687 struct sighand_struct *sighand;
689 sigset_t blocked, real_blocked;
690 struct sigpending pending;
692 unsigned long sas_ss_sp;
694 int (*notifier)(void *priv);
696 sigset_t *notifier_mask;
699 struct audit_context *audit_context;
702 /* vserver context data */
704 struct vx_info *vx_info;
706 /* vserver network data */
708 struct nx_info *nx_info;
710 /* Thread group tracking */
713 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
714 spinlock_t alloc_lock;
715 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
716 spinlock_t proc_lock;
717 /* context-switch lock */
718 spinlock_t switch_lock;
720 /* journalling filesystem info */
724 struct reclaim_state *reclaim_state;
726 struct dentry *proc_dentry;
727 struct backing_dev_info *backing_dev_info;
729 struct io_context *io_context;
731 unsigned long ptrace_message;
732 siginfo_t *last_siginfo; /* For ptrace use. */
734 * current io wait handle: wait queue entry to use for io waits
735 * If this thread is processing aio, this points at the waitqueue
736 * inside the currently handled kiocb. It may be NULL (i.e. default
737 * to a stack based synchronous wait) if its doing sync IO.
739 wait_queue_t *io_wait;
740 /* i/o counters(bytes read/written, #syscalls */
741 u64 rchar, wchar, syscr, syscw;
742 #if defined(CONFIG_BSD_PROCESS_ACCT)
743 u64 acct_rss_mem1; /* accumulated rss usage */
744 u64 acct_vm_mem1; /* accumulated virtual memory usage */
745 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
748 struct mempolicy *mempolicy;
751 #ifdef CONFIG_CPUSETS
752 struct cpuset *cpuset;
753 nodemask_t mems_allowed;
754 int cpuset_mems_generation;
758 static inline pid_t process_group(struct task_struct *tsk)
760 return tsk->signal->pgrp;
764 * pid_alive - check that a task structure is not stale
765 * @p: Task structure to be checked.
767 * Test if a process is not yet dead (at most zombie state)
768 * If pid_alive fails, then pointers within the task structure
769 * can be stale and must not be dereferenced.
771 static inline int pid_alive(struct task_struct *p)
773 return p->pids[PIDTYPE_PID].nr != 0;
776 extern void free_task(struct task_struct *tsk);
777 extern void __put_task_struct(struct task_struct *tsk);
778 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
779 #define put_task_struct(tsk) \
780 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
785 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
786 /* Not implemented yet, only for 486*/
787 #define PF_STARTING 0x00000002 /* being created */
788 #define PF_EXITING 0x00000004 /* getting shut down */
789 #define PF_DEAD 0x00000008 /* Dead */
790 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
791 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
792 #define PF_DUMPCORE 0x00000200 /* dumped core */
793 #define PF_SIGNALED 0x00000400 /* killed by a signal */
794 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
795 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
796 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
797 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
798 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
799 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
800 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
801 #define PF_KSWAPD 0x00040000 /* I am kswapd */
802 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
803 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
804 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
805 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
806 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
809 * Only the _current_ task can read/write to tsk->flags, but other
810 * tasks can access tsk->flags in readonly mode for example
811 * with tsk_used_math (like during threaded core dumping).
812 * There is however an exception to this rule during ptrace
813 * or during fork: the ptracer task is allowed to write to the
814 * child->flags of its traced child (same goes for fork, the parent
815 * can write to the child->flags), because we're guaranteed the
816 * child is not running and in turn not changing child->flags
817 * at the same time the parent does it.
819 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
820 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
821 #define clear_used_math() clear_stopped_child_used_math(current)
822 #define set_used_math() set_stopped_child_used_math(current)
823 #define conditional_stopped_child_used_math(condition, child) \
824 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
825 #define conditional_used_math(condition) \
826 conditional_stopped_child_used_math(condition, current)
827 #define copy_to_stopped_child_used_math(child) \
828 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
829 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
830 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
831 #define used_math() tsk_used_math(current)
834 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
836 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
838 if (!cpus_intersects(new_mask, cpu_online_map))
844 extern unsigned long long sched_clock(void);
845 extern unsigned long long current_sched_time(const task_t *current_task);
847 /* sched_exec is called by processes performing an exec */
849 extern void sched_exec(void);
851 #define sched_exec() {}
854 #ifdef CONFIG_HOTPLUG_CPU
855 extern void idle_task_exit(void);
857 static inline void idle_task_exit(void) {}
860 extern void sched_idle_next(void);
861 extern void set_user_nice(task_t *p, long nice);
862 extern int task_prio(const task_t *p);
863 extern int task_nice(const task_t *p);
864 extern int can_nice(const task_t *p, const int nice);
865 extern int task_curr(const task_t *p);
866 extern int idle_cpu(int cpu);
867 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
868 extern task_t *idle_task(int cpu);
873 * The default (Linux) execution domain.
875 extern struct exec_domain default_exec_domain;
878 struct thread_info thread_info;
879 unsigned long stack[THREAD_SIZE/sizeof(long)];
882 #ifndef __HAVE_ARCH_KSTACK_END
883 static inline int kstack_end(void *addr)
885 /* Reliable end of stack detection:
886 * Some APM bios versions misalign the stack
888 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
892 extern union thread_union init_thread_union;
893 extern struct task_struct init_task;
895 extern struct mm_struct init_mm;
898 #define find_task_by_real_pid(nr) \
899 find_task_by_pid_type(PIDTYPE_PID, nr)
900 #define find_task_by_pid(nr) \
901 find_task_by_pid_type(PIDTYPE_PID, \
904 extern struct task_struct *find_task_by_pid_type(int type, int pid);
905 extern void set_special_pids(pid_t session, pid_t pgrp);
906 extern void __set_special_pids(pid_t session, pid_t pgrp);
908 /* per-UID process charging. */
909 extern struct user_struct * alloc_uid(xid_t, uid_t);
910 static inline struct user_struct *get_uid(struct user_struct *u)
912 atomic_inc(&u->__count);
915 extern void free_uid(struct user_struct *);
916 extern void switch_uid(struct user_struct *);
918 #include <asm/current.h>
920 extern void do_timer(struct pt_regs *);
922 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
923 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
924 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
925 unsigned long clone_flags));
927 extern void kick_process(struct task_struct *tsk);
929 static inline void kick_process(struct task_struct *tsk) { }
931 extern void FASTCALL(sched_fork(task_t * p));
932 extern void FASTCALL(sched_exit(task_t * p));
934 extern int in_group_p(gid_t);
935 extern int in_egroup_p(gid_t);
937 extern void proc_caches_init(void);
938 extern void flush_signals(struct task_struct *);
939 extern void flush_signal_handlers(struct task_struct *, int force_default);
940 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
942 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
947 spin_lock_irqsave(&tsk->sighand->siglock, flags);
948 ret = dequeue_signal(tsk, mask, info);
949 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
954 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
956 extern void unblock_all_signals(void);
957 extern void release_task(struct task_struct * p);
958 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
959 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
960 extern int force_sigsegv(int, struct task_struct *);
961 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
962 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
963 extern int kill_pg_info(int, struct siginfo *, pid_t);
964 extern int kill_proc_info(int, struct siginfo *, pid_t);
965 extern void do_notify_parent(struct task_struct *, int);
966 extern void force_sig(int, struct task_struct *);
967 extern void force_sig_specific(int, struct task_struct *);
968 extern int send_sig(int, struct task_struct *, int);
969 extern void zap_other_threads(struct task_struct *p);
970 extern int kill_pg(pid_t, int, int);
971 extern int kill_sl(pid_t, int, int);
972 extern int kill_proc(pid_t, int, int);
973 extern struct sigqueue *sigqueue_alloc(void);
974 extern void sigqueue_free(struct sigqueue *);
975 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
976 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
977 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
978 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
980 /* These can be the second arg to send_sig_info/send_group_sig_info. */
981 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
982 #define SEND_SIG_PRIV ((struct siginfo *) 1)
983 #define SEND_SIG_FORCED ((struct siginfo *) 2)
985 /* True if we are on the alternate signal stack. */
987 static inline int on_sig_stack(unsigned long sp)
989 return (sp - current->sas_ss_sp < current->sas_ss_size);
992 static inline int sas_ss_flags(unsigned long sp)
994 return (current->sas_ss_size == 0 ? SS_DISABLE
995 : on_sig_stack(sp) ? SS_ONSTACK : 0);
999 #ifdef CONFIG_SECURITY
1000 /* code is in security.c */
1001 extern int capable(int cap);
1002 extern int vx_capable(int cap, int ccap);
1004 static inline int capable(int cap)
1006 if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
1008 if (cap_raised(current->cap_effective, cap)) {
1009 current->flags |= PF_SUPERPRIV;
1015 static inline int vx_capable(int cap, int ccap)
1017 if (cap_raised(current->cap_effective, cap) &&
1019 current->flags |= PF_SUPERPRIV;
1027 * Routines for handling mm_structs
1029 extern struct mm_struct * mm_alloc(void);
1031 /* mmdrop drops the mm and the page tables */
1032 extern void FASTCALL(__mmdrop(struct mm_struct *));
1033 static inline void mmdrop(struct mm_struct * mm)
1035 if (atomic_dec_and_test(&mm->mm_count))
1039 /* mmput gets rid of the mappings and all user-space */
1040 extern void mmput(struct mm_struct *);
1041 /* Grab a reference to a task's mm, if it is not already going away */
1042 extern struct mm_struct *get_task_mm(struct task_struct *task);
1043 /* Remove the current tasks stale references to the old mm_struct */
1044 extern void mm_release(struct task_struct *, struct mm_struct *);
1046 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1047 extern void flush_thread(void);
1048 extern void exit_thread(void);
1050 extern void exit_files(struct task_struct *);
1051 extern void exit_signal(struct task_struct *);
1052 extern void __exit_signal(struct task_struct *);
1053 extern void exit_sighand(struct task_struct *);
1054 extern void __exit_sighand(struct task_struct *);
1055 extern void exit_itimers(struct signal_struct *);
1057 extern NORET_TYPE void do_group_exit(int);
1059 extern void daemonize(const char *, ...);
1060 extern int allow_signal(int);
1061 extern int disallow_signal(int);
1062 extern task_t *child_reaper;
1064 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1065 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1066 task_t *fork_idle(int);
1068 extern void set_task_comm(struct task_struct *tsk, char *from);
1069 extern void get_task_comm(char *to, struct task_struct *tsk);
1072 extern void wait_task_inactive(task_t * p);
1074 #define wait_task_inactive(p) do { } while (0)
1077 #define remove_parent(p) list_del_init(&(p)->sibling)
1078 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1080 #define REMOVE_LINKS(p) do { \
1081 if (thread_group_leader(p)) \
1082 list_del_init(&(p)->tasks); \
1086 #define SET_LINKS(p) do { \
1087 if (thread_group_leader(p)) \
1088 list_add_tail(&(p)->tasks,&init_task.tasks); \
1089 add_parent(p, (p)->parent); \
1092 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1093 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1095 #define for_each_process(p) \
1096 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1099 * Careful: do_each_thread/while_each_thread is a double loop so
1100 * 'break' will not work as expected - use goto instead.
1102 #define do_each_thread(g, t) \
1103 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1105 #define while_each_thread(g, t) \
1106 while ((t = next_thread(t)) != g)
1108 extern task_t * FASTCALL(next_thread(const task_t *p));
1110 #define thread_group_leader(p) (p->pid == p->tgid)
1112 static inline int thread_group_empty(task_t *p)
1114 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1117 #define delay_group_leader(p) \
1118 (thread_group_leader(p) && !thread_group_empty(p))
1120 extern void unhash_process(struct task_struct *p);
1123 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1124 * subscriptions and synchronises with wait4(). Also used in procfs.
1126 * Nests both inside and outside of read_lock(&tasklist_lock).
1127 * It must not be nested with write_lock_irq(&tasklist_lock),
1128 * neither inside nor outside.
1130 static inline void task_lock(struct task_struct *p)
1132 spin_lock(&p->alloc_lock);
1135 static inline void task_unlock(struct task_struct *p)
1137 spin_unlock(&p->alloc_lock);
1140 /* set thread flags in other task's structures
1141 * - see asm/thread_info.h for TIF_xxxx flags available
1143 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1145 set_ti_thread_flag(tsk->thread_info,flag);
1148 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1150 clear_ti_thread_flag(tsk->thread_info,flag);
1153 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1155 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1158 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1160 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1163 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1165 return test_ti_thread_flag(tsk->thread_info,flag);
1168 static inline void set_tsk_need_resched(struct task_struct *tsk)
1170 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1173 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1175 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1178 static inline int signal_pending(struct task_struct *p)
1180 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1183 static inline int need_resched(void)
1185 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1189 * cond_resched() and cond_resched_lock(): latency reduction via
1190 * explicit rescheduling in places that are safe. The return
1191 * value indicates whether a reschedule was done in fact.
1192 * cond_resched_lock() will drop the spinlock before scheduling,
1193 * cond_resched_softirq() will enable bhs before scheduling.
1195 extern int cond_resched(void);
1196 extern int cond_resched_lock(spinlock_t * lock);
1197 extern int cond_resched_softirq(void);
1200 * Does a critical section need to be broken due to another
1203 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1204 # define need_lockbreak(lock) ((lock)->break_lock)
1206 # define need_lockbreak(lock) 0
1210 * Does a critical section need to be broken due to another
1211 * task waiting or preemption being signalled:
1213 static inline int lock_need_resched(spinlock_t *lock)
1215 if (need_lockbreak(lock) || need_resched())
1220 /* Reevaluate whether the task has signals pending delivery.
1221 This is required every time the blocked sigset_t changes.
1222 callers must hold sighand->siglock. */
1224 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1225 extern void recalc_sigpending(void);
1227 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1230 * Wrappers for p->thread_info->cpu access. No-op on UP.
1234 static inline unsigned int task_cpu(const struct task_struct *p)
1236 return p->thread_info->cpu;
1239 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1241 p->thread_info->cpu = cpu;
1246 static inline unsigned int task_cpu(const struct task_struct *p)
1251 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1255 #endif /* CONFIG_SMP */
1257 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1258 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1260 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1262 mm->mmap_base = TASK_UNMAPPED_BASE;
1263 mm->get_unmapped_area = arch_get_unmapped_area;
1264 mm->unmap_area = arch_unmap_area;
1268 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1269 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1271 #ifdef CONFIG_MAGIC_SYSRQ
1273 extern void normalize_rt_tasks(void);
1279 * Checks whether we need to enter the refrigerator
1280 * and returns 1 if we did so.
1283 extern void refrigerator(unsigned long);
1284 extern int freeze_processes(void);
1285 extern void thaw_processes(void);
1287 static inline int try_to_freeze(unsigned long refrigerator_flags)
1289 if (unlikely(current->flags & PF_FREEZE)) {
1290 refrigerator(refrigerator_flags);
1296 static inline void refrigerator(unsigned long flag) {}
1297 static inline int freeze_processes(void) { BUG(); return 0; }
1298 static inline void thaw_processes(void) {}
1300 static inline int try_to_freeze(unsigned long refrigerator_flags)
1304 #endif /* CONFIG_PM */
1305 #endif /* __KERNEL__ */