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>
40 extern int exec_shield;
41 extern int print_fatal_signals;
46 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
47 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
48 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
49 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
50 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
51 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
52 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
53 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
54 #define CLONE_THREAD 0x00010000 /* Same thread group? */
55 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
56 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
57 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
58 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
59 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
60 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
61 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
62 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
63 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
66 * List of flags we want to share for kernel threads,
67 * if only because they are not used by them anyway.
69 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
72 * These are the constant used to fake the fixed-point load-average
73 * counting. Some notes:
74 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
75 * a load-average precision of 10 bits integer + 11 bits fractional
76 * - if you want to count load-averages more often, you need more
77 * precision, or rounding will get you. With 2-second counting freq,
78 * the EXP_n values would be 1981, 2034 and 2043 if still using only
81 extern unsigned long avenrun[]; /* Load averages */
83 #define FSHIFT 11 /* nr of bits of precision */
84 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
85 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
86 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
87 #define EXP_5 2014 /* 1/exp(5sec/5min) */
88 #define EXP_15 2037 /* 1/exp(5sec/15min) */
90 #define CALC_LOAD(load,exp,n) \
92 load += n*(FIXED_1-exp); \
95 extern unsigned long total_forks;
96 extern int nr_threads;
98 DECLARE_PER_CPU(unsigned long, process_counts);
99 extern int nr_processes(void);
100 extern unsigned long nr_running(void);
101 extern unsigned long nr_uninterruptible(void);
102 extern unsigned long nr_iowait(void);
104 #include <linux/time.h>
105 #include <linux/param.h>
106 #include <linux/resource.h>
107 #include <linux/timer.h>
109 #include <asm/processor.h>
111 #define TASK_RUNNING 0
112 #define TASK_INTERRUPTIBLE 1
113 #define TASK_UNINTERRUPTIBLE 2
114 #define TASK_STOPPED 4
115 #define TASK_TRACED 8
116 #define EXIT_ZOMBIE 16
118 #define TASK_ONHOLD 64
120 #define __set_task_state(tsk, state_value) \
121 do { (tsk)->state = (state_value); } while (0)
122 #define set_task_state(tsk, state_value) \
123 set_mb((tsk)->state, (state_value))
125 #define __set_current_state(state_value) \
126 do { current->state = (state_value); } while (0)
127 #define set_current_state(state_value) \
128 set_mb(current->state, (state_value))
130 /* Task command name length */
131 #define TASK_COMM_LEN 16
134 * Scheduling policies
136 #define SCHED_NORMAL 0
146 #include <linux/spinlock.h>
149 * This serializes "schedule()" and also protects
150 * the run-queue from deletions/modifications (but
151 * _adding_ to the beginning of the run-queue has
154 extern rwlock_t tasklist_lock;
155 extern spinlock_t mmlist_lock;
157 typedef struct task_struct task_t;
159 extern void sched_init(void);
160 extern void sched_init_smp(void);
161 extern void init_idle(task_t *idle, int cpu);
163 extern cpumask_t nohz_cpu_mask;
165 extern void show_state(void);
166 extern void show_regs(struct pt_regs *);
169 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
170 * task), SP is the stack pointer of the first frame that should be shown in the back
171 * trace (or NULL if the entire call-chain of the task should be shown).
173 extern void show_stack(struct task_struct *task, unsigned long *sp);
175 void io_schedule(void);
176 long io_schedule_timeout(long timeout);
178 extern void cpu_init (void);
179 extern void trap_init(void);
180 extern void update_process_times(int user);
181 extern void scheduler_tick(void);
183 /* Attach to any functions which should be ignored in wchan output. */
184 #define __sched __attribute__((__section__(".sched.text")))
185 /* Is this address in the __sched functions? */
186 extern int in_sched_functions(unsigned long addr);
188 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
189 extern signed long FASTCALL(schedule_timeout(signed long timeout));
190 asmlinkage void schedule(void);
194 /* Maximum number of active map areas.. This is a random (large) number */
195 #define DEFAULT_MAX_MAP_COUNT 65536
197 extern int sysctl_max_map_count;
199 #include <linux/aio.h>
202 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
203 unsigned long, unsigned long);
206 arch_get_unmapped_exec_area(struct file *, unsigned long, unsigned long,
207 unsigned long, unsigned long);
209 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
210 unsigned long len, unsigned long pgoff,
211 unsigned long flags);
212 extern void arch_unmap_area(struct vm_area_struct *area);
213 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
215 #define __set_mm_counter(mm, member, value) (mm)->_##member = (value)
216 #define set_mm_counter(mm, member, value) vx_ ## member ## pages_sub((mm), ((mm)->_##member - value))
217 #define get_mm_counter(mm, member) ((mm)->_##member)
218 #define add_mm_counter(mm, member, value) vx_ ## member ## pages_add((mm), (value))
219 #define inc_mm_counter(mm, member) vx_ ## member ## pages_inc((mm))
220 #define dec_mm_counter(mm, member) vx_ ## member ## pages_dec((mm))
221 typedef unsigned long mm_counter_t;
224 struct vm_area_struct * mmap; /* list of VMAs */
225 struct rb_root mm_rb;
226 struct vm_area_struct * mmap_cache; /* last find_vma result */
227 unsigned long (*get_unmapped_area) (struct file *filp,
228 unsigned long addr, unsigned long len,
229 unsigned long pgoff, unsigned long flags);
230 unsigned long (*get_unmapped_exec_area) (struct file *filp,
231 unsigned long addr, unsigned long len,
232 unsigned long pgoff, unsigned long flags);
233 void (*unmap_area) (struct vm_area_struct *area);
234 unsigned long mmap_base; /* base of mmap area */
235 unsigned long free_area_cache; /* first hole */
237 atomic_t mm_users; /* How many users with user space? */
238 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
239 int map_count; /* number of VMAs */
240 struct rw_semaphore mmap_sem;
241 spinlock_t page_table_lock; /* Protects page tables and some counters */
243 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
244 * together off init_mm.mmlist, and are protected
248 unsigned long start_code, end_code, start_data, end_data;
249 unsigned long start_brk, brk, start_stack;
250 unsigned long arg_start, arg_end, env_start, env_end;
251 unsigned long total_vm, locked_vm, shared_vm;
252 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
254 /* Special counters protected by the page_table_lock */
256 mm_counter_t _anon_rss;
258 unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
261 cpumask_t cpu_vm_mask;
263 /* Architecture-specific MM context */
264 mm_context_t context;
265 struct vx_info *mm_vx_info;
267 /* Token based thrashing protection. */
268 unsigned long swap_token_time;
271 /* coredumping support */
273 struct completion *core_startup_done, core_done;
276 rwlock_t ioctx_list_lock;
277 struct kioctx *ioctx_list;
279 struct kioctx default_kioctx;
281 unsigned long hiwater_rss; /* High-water RSS usage */
282 unsigned long hiwater_vm; /* High-water virtual memory usage */
285 struct sighand_struct {
287 struct k_sigaction action[_NSIG];
292 * NOTE! "signal_struct" does not have it's own
293 * locking, because a shared signal_struct always
294 * implies a shared sighand_struct, so locking
295 * sighand_struct is always a proper superset of
296 * the locking of signal_struct.
298 struct signal_struct {
302 wait_queue_head_t wait_chldexit; /* for wait4() */
304 /* current thread group signal load-balancing target: */
307 /* shared signal handling: */
308 struct sigpending shared_pending;
310 /* thread group exit support */
313 * - notify group_exit_task when ->count is equal to notify_count
314 * - everyone except group_exit_task is stopped during signal delivery
315 * of fatal signals, group_exit_task processes the signal.
317 struct task_struct *group_exit_task;
320 /* thread group stop support, overloads group_exit_code too */
321 int group_stop_count;
322 unsigned int flags; /* see SIGNAL_* flags below */
324 /* POSIX.1b Interval Timers */
325 struct list_head posix_timers;
327 /* ITIMER_REAL timer for the process */
328 struct timer_list real_timer;
329 unsigned long it_real_value, it_real_incr;
331 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
332 cputime_t it_prof_expires, it_virt_expires;
333 cputime_t it_prof_incr, it_virt_incr;
335 /* job control IDs */
339 /* boolean value for session group leader */
342 struct tty_struct *tty; /* NULL if no tty */
345 * Cumulative resource counters for dead threads in the group,
346 * and for reaped dead child processes forked by this group.
347 * Live threads maintain their own counters and add to these
348 * in __exit_signal, except for the group leader.
350 cputime_t utime, stime, cutime, cstime;
351 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
352 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
355 * Cumulative ns of scheduled CPU time for dead threads in the
356 * group, not including a zombie group leader. (This only differs
357 * from jiffies_to_ns(utime + stime) if sched_clock uses something
358 * other than jiffies.)
360 unsigned long long sched_time;
363 * We don't bother to synchronize most readers of this at all,
364 * because there is no reader checking a limit that actually needs
365 * to get both rlim_cur and rlim_max atomically, and either one
366 * alone is a single word that can safely be read normally.
367 * getrlimit/setrlimit use task_lock(current->group_leader) to
368 * protect this instead of the siglock, because they really
369 * have no need to disable irqs.
371 struct rlimit rlim[RLIM_NLIMITS];
373 struct list_head cpu_timers[3];
375 /* keep the process-shared keyrings here so that they do the right
376 * thing in threads created with CLONE_THREAD */
378 struct key *session_keyring; /* keyring inherited over fork */
379 struct key *process_keyring; /* keyring private to this process */
384 * Bits in flags field of signal_struct.
386 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
387 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
388 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
389 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
393 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
394 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
395 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
396 * are inverted: lower p->prio value means higher priority.
398 * The MAX_USER_RT_PRIO value allows the actual maximum
399 * RT priority to be separate from the value exported to
400 * user-space. This allows kernel threads to set their
401 * priority to a value higher than any user task. Note:
402 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
405 #define MAX_USER_RT_PRIO 100
406 #define MAX_RT_PRIO MAX_USER_RT_PRIO
408 #define MAX_PRIO (MAX_RT_PRIO + 40)
410 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
413 * Some day this will be a full-fledged user tracking system..
416 atomic_t __count; /* reference count */
417 atomic_t processes; /* How many processes does this user have? */
418 atomic_t files; /* How many open files does this user have? */
419 atomic_t sigpending; /* How many pending signals does this user have? */
420 /* protected by mq_lock */
421 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
422 unsigned long locked_shm; /* How many pages of mlocked shm ? */
425 struct key *uid_keyring; /* UID specific keyring */
426 struct key *session_keyring; /* UID's default session keyring */
429 /* Hash table maintenance information */
430 struct list_head uidhash_list;
435 extern struct user_struct *find_user(xid_t, uid_t);
437 extern struct user_struct root_user;
438 #define INIT_USER (&root_user)
440 typedef struct prio_array prio_array_t;
441 struct backing_dev_info;
442 struct reclaim_state;
444 #ifdef CONFIG_SCHEDSTATS
446 /* cumulative counters */
447 unsigned long cpu_time, /* time spent on the cpu */
448 run_delay, /* time spent waiting on a runqueue */
449 pcnt; /* # of timeslices run on this cpu */
452 unsigned long last_arrival, /* when we last ran on a cpu */
453 last_queued; /* when we were last queued to run */
456 extern struct file_operations proc_schedstat_operations;
468 * sched-domains (multiprocessor balancing) declarations:
471 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
473 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
474 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
475 #define SD_BALANCE_EXEC 4 /* Balance on exec */
476 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
477 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
478 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
479 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
482 struct sched_group *next; /* Must be a circular list */
486 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
487 * single CPU. This is read only (except for setup, hotplug CPU).
489 unsigned long cpu_power;
492 struct sched_domain {
493 /* These fields must be setup */
494 struct sched_domain *parent; /* top domain must be null terminated */
495 struct sched_group *groups; /* the balancing groups of the domain */
496 cpumask_t span; /* span of all CPUs in this domain */
497 unsigned long min_interval; /* Minimum balance interval ms */
498 unsigned long max_interval; /* Maximum balance interval ms */
499 unsigned int busy_factor; /* less balancing by factor if busy */
500 unsigned int imbalance_pct; /* No balance until over watermark */
501 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
502 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
503 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
504 int flags; /* See SD_* */
506 /* Runtime fields. */
507 unsigned long last_balance; /* init to jiffies. units in jiffies */
508 unsigned int balance_interval; /* initialise to 1. units in ms. */
509 unsigned int nr_balance_failed; /* initialise to 0 */
511 #ifdef CONFIG_SCHEDSTATS
512 /* load_balance() stats */
513 unsigned long lb_cnt[MAX_IDLE_TYPES];
514 unsigned long lb_failed[MAX_IDLE_TYPES];
515 unsigned long lb_balanced[MAX_IDLE_TYPES];
516 unsigned long lb_imbalance[MAX_IDLE_TYPES];
517 unsigned long lb_gained[MAX_IDLE_TYPES];
518 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
519 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
520 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
522 /* Active load balancing */
523 unsigned long alb_cnt;
524 unsigned long alb_failed;
525 unsigned long alb_pushed;
527 /* sched_balance_exec() stats */
528 unsigned long sbe_attempts;
529 unsigned long sbe_pushed;
531 /* try_to_wake_up() stats */
532 unsigned long ttwu_wake_remote;
533 unsigned long ttwu_move_affine;
534 unsigned long ttwu_move_balance;
538 #ifdef ARCH_HAS_SCHED_DOMAIN
539 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
540 extern cpumask_t cpu_isolated_map;
541 extern void init_sched_build_groups(struct sched_group groups[],
542 cpumask_t span, int (*group_fn)(int cpu));
543 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
544 #endif /* ARCH_HAS_SCHED_DOMAIN */
545 #endif /* CONFIG_SMP */
548 struct io_context; /* See blkdev.h */
549 void exit_io_context(void);
552 #define NGROUPS_SMALL 32
553 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
557 gid_t small_block[NGROUPS_SMALL];
563 * get_group_info() must be called with the owning task locked (via task_lock())
564 * when task != current. The reason being that the vast majority of callers are
565 * looking at current->group_info, which can not be changed except by the
566 * current task. Changing current->group_info requires the task lock, too.
568 #define get_group_info(group_info) do { \
569 atomic_inc(&(group_info)->usage); \
572 #define put_group_info(group_info) do { \
573 if (atomic_dec_and_test(&(group_info)->usage)) \
574 groups_free(group_info); \
577 struct group_info *groups_alloc(int gidsetsize);
578 void groups_free(struct group_info *group_info);
579 int set_current_groups(struct group_info *group_info);
580 /* access the groups "array" with this macro */
581 #define GROUP_AT(gi, i) \
582 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
585 struct audit_context; /* See audit.c */
589 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
590 struct thread_info *thread_info;
592 unsigned long flags; /* per process flags, defined below */
593 unsigned long ptrace;
595 int lock_depth; /* BKL lock depth */
597 int prio, static_prio;
598 struct list_head run_list;
601 unsigned long sleep_avg;
602 unsigned long long timestamp, last_ran;
603 unsigned long long sched_time; /* sched_clock time spent running */
606 unsigned long policy;
607 cpumask_t cpus_allowed;
608 unsigned int time_slice, first_time_slice;
610 #ifdef CONFIG_SCHEDSTATS
611 struct sched_info sched_info;
614 struct list_head tasks;
616 * ptrace_list/ptrace_children forms the list of my children
617 * that were stolen by a ptracer.
619 struct list_head ptrace_children;
620 struct list_head ptrace_list;
622 struct mm_struct *mm, *active_mm;
625 struct linux_binfmt *binfmt;
627 int exit_code, exit_signal;
628 int pdeath_signal; /* The signal sent when the parent dies */
630 unsigned long personality;
635 * pointers to (original) parent process, youngest child, younger sibling,
636 * older sibling, respectively. (p->father can be replaced with
639 struct task_struct *real_parent; /* real parent process (when being debugged) */
640 struct task_struct *parent; /* parent process */
642 * children/sibling forms the list of my children plus the
643 * tasks I'm ptracing.
645 struct list_head children; /* list of my children */
646 struct list_head sibling; /* linkage in my parent's children list */
647 struct task_struct *group_leader; /* threadgroup leader */
649 /* PID/PID hash table linkage. */
650 struct pid pids[PIDTYPE_MAX];
652 struct completion *vfork_done; /* for vfork() */
653 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
654 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
656 unsigned long rt_priority;
657 cputime_t utime, stime;
658 unsigned long nvcsw, nivcsw; /* context switch counts */
659 struct timespec start_time;
660 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
661 unsigned long min_flt, maj_flt;
663 cputime_t it_prof_expires, it_virt_expires;
664 unsigned long long it_sched_expires;
665 struct list_head cpu_timers[3];
667 /* process credentials */
668 uid_t uid,euid,suid,fsuid;
669 gid_t gid,egid,sgid,fsgid;
670 struct group_info *group_info;
671 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
672 unsigned keep_capabilities:1;
673 struct user_struct *user;
675 struct key *thread_keyring; /* keyring private to this thread */
677 int oomkilladj; /* OOM kill score adjustment (bit shift). */
678 char comm[TASK_COMM_LEN]; /* executable name excluding path
679 - access with [gs]et_task_comm (which lock
681 - initialized normally by flush_old_exec */
682 /* file system info */
683 int link_count, total_link_count;
685 struct sysv_sem sysvsem;
686 /* CPU-specific state of this task */
687 struct thread_struct thread;
688 /* filesystem information */
689 struct fs_struct *fs;
690 /* open file information */
691 struct files_struct *files;
693 struct namespace *namespace;
694 /* signal handlers */
695 struct signal_struct *signal;
696 struct sighand_struct *sighand;
698 sigset_t blocked, real_blocked;
699 struct sigpending pending;
701 unsigned long sas_ss_sp;
703 int (*notifier)(void *priv);
705 sigset_t *notifier_mask;
709 void (*tux_exit)(void);
713 struct audit_context *audit_context;
716 /* vserver context data */
718 struct vx_info *vx_info;
720 /* vserver network data */
722 struct nx_info *nx_info;
724 /* Thread group tracking */
727 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
728 spinlock_t alloc_lock;
729 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
730 spinlock_t proc_lock;
731 /* context-switch lock */
732 spinlock_t switch_lock;
734 /* journalling filesystem info */
738 struct reclaim_state *reclaim_state;
740 struct dentry *proc_dentry;
741 struct backing_dev_info *backing_dev_info;
743 struct io_context *io_context;
747 unsigned long ptrace_message;
748 siginfo_t *last_siginfo; /* For ptrace use. */
750 * current io wait handle: wait queue entry to use for io waits
751 * If this thread is processing aio, this points at the waitqueue
752 * inside the currently handled kiocb. It may be NULL (i.e. default
753 * to a stack based synchronous wait) if its doing sync IO.
755 wait_queue_t *io_wait;
756 /* i/o counters(bytes read/written, #syscalls */
757 u64 rchar, wchar, syscr, syscw;
758 #if defined(CONFIG_BSD_PROCESS_ACCT)
759 u64 acct_rss_mem1; /* accumulated rss usage */
760 u64 acct_vm_mem1; /* accumulated virtual memory usage */
761 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
764 struct mempolicy *mempolicy;
767 #ifdef CONFIG_CPUSETS
768 struct cpuset *cpuset;
769 nodemask_t mems_allowed;
770 int cpuset_mems_generation;
774 static inline pid_t process_group(struct task_struct *tsk)
776 return tsk->signal->pgrp;
780 * pid_alive - check that a task structure is not stale
781 * @p: Task structure to be checked.
783 * Test if a process is not yet dead (at most zombie state)
784 * If pid_alive fails, then pointers within the task structure
785 * can be stale and must not be dereferenced.
787 static inline int pid_alive(struct task_struct *p)
789 return p->pids[PIDTYPE_PID].nr != 0;
792 extern void free_task(struct task_struct *tsk);
793 extern void __put_task_struct(struct task_struct *tsk);
794 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
795 #define put_task_struct(tsk) \
796 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
801 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
802 /* Not implemented yet, only for 486*/
803 #define PF_STARTING 0x00000002 /* being created */
804 #define PF_EXITING 0x00000004 /* getting shut down */
805 #define PF_DEAD 0x00000008 /* Dead */
806 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
807 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
808 #define PF_DUMPCORE 0x00000200 /* dumped core */
809 #define PF_SIGNALED 0x00000400 /* killed by a signal */
810 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
811 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
812 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
813 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
814 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
815 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
816 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
817 #define PF_KSWAPD 0x00040000 /* I am kswapd */
818 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
819 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
820 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
821 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
822 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
825 * Only the _current_ task can read/write to tsk->flags, but other
826 * tasks can access tsk->flags in readonly mode for example
827 * with tsk_used_math (like during threaded core dumping).
828 * There is however an exception to this rule during ptrace
829 * or during fork: the ptracer task is allowed to write to the
830 * child->flags of its traced child (same goes for fork, the parent
831 * can write to the child->flags), because we're guaranteed the
832 * child is not running and in turn not changing child->flags
833 * at the same time the parent does it.
835 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
836 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
837 #define clear_used_math() clear_stopped_child_used_math(current)
838 #define set_used_math() set_stopped_child_used_math(current)
839 #define conditional_stopped_child_used_math(condition, child) \
840 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
841 #define conditional_used_math(condition) \
842 conditional_stopped_child_used_math(condition, current)
843 #define copy_to_stopped_child_used_math(child) \
844 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
845 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
846 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
847 #define used_math() tsk_used_math(current)
850 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
852 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
854 if (!cpus_intersects(new_mask, cpu_online_map))
860 extern unsigned long long sched_clock(void);
861 extern unsigned long long current_sched_time(const task_t *current_task);
863 /* sched_exec is called by processes performing an exec */
865 extern void sched_exec(void);
867 #define sched_exec() {}
870 #ifdef CONFIG_HOTPLUG_CPU
871 extern void idle_task_exit(void);
873 static inline void idle_task_exit(void) {}
876 extern void sched_idle_next(void);
877 extern void set_user_nice(task_t *p, long nice);
878 extern int task_prio(const task_t *p);
879 extern int task_nice(const task_t *p);
880 extern int can_nice(const task_t *p, const int nice);
881 extern int task_curr(const task_t *p);
882 extern int idle_cpu(int cpu);
883 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
884 extern task_t *idle_task(int cpu);
889 * The default (Linux) execution domain.
891 extern struct exec_domain default_exec_domain;
894 struct thread_info thread_info;
895 unsigned long stack[THREAD_SIZE/sizeof(long)];
898 #ifndef __HAVE_ARCH_KSTACK_END
899 static inline int kstack_end(void *addr)
901 /* Reliable end of stack detection:
902 * Some APM bios versions misalign the stack
904 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
908 extern union thread_union init_thread_union;
909 extern struct task_struct init_task;
911 extern struct mm_struct init_mm;
914 #define find_task_by_real_pid(nr) \
915 find_task_by_pid_type(PIDTYPE_PID, nr)
916 #define find_task_by_pid(nr) \
917 find_task_by_pid_type(PIDTYPE_PID, \
920 extern struct task_struct *find_task_by_pid_type(int type, int pid);
921 extern void set_special_pids(pid_t session, pid_t pgrp);
922 extern void __set_special_pids(pid_t session, pid_t pgrp);
924 /* per-UID process charging. */
925 extern struct user_struct * alloc_uid(xid_t, uid_t);
926 static inline struct user_struct *get_uid(struct user_struct *u)
928 atomic_inc(&u->__count);
931 extern void free_uid(struct user_struct *);
932 extern void switch_uid(struct user_struct *);
934 #include <asm/current.h>
936 extern void do_timer(struct pt_regs *);
938 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
939 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
940 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
941 unsigned long clone_flags));
943 extern void kick_process(struct task_struct *tsk);
945 static inline void kick_process(struct task_struct *tsk) { }
947 extern void FASTCALL(sched_fork(task_t * p));
948 extern void FASTCALL(sched_exit(task_t * p));
950 extern int in_group_p(gid_t);
951 extern int in_egroup_p(gid_t);
953 extern void proc_caches_init(void);
954 extern void flush_signals(struct task_struct *);
955 extern void flush_signal_handlers(struct task_struct *, int force_default);
956 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
958 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
963 spin_lock_irqsave(&tsk->sighand->siglock, flags);
964 ret = dequeue_signal(tsk, mask, info);
965 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
970 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
972 extern void unblock_all_signals(void);
973 extern void release_task(struct task_struct * p);
974 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
975 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
976 extern int force_sigsegv(int, struct task_struct *);
977 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
978 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
979 extern int kill_pg_info(int, struct siginfo *, pid_t);
980 extern int kill_proc_info(int, struct siginfo *, pid_t);
981 extern void do_notify_parent(struct task_struct *, int);
982 extern void force_sig(int, struct task_struct *);
983 extern void force_sig_specific(int, struct task_struct *);
984 extern int send_sig(int, struct task_struct *, int);
985 extern void zap_other_threads(struct task_struct *p);
986 extern int kill_pg(pid_t, int, int);
987 extern int kill_sl(pid_t, int, int);
988 extern int kill_proc(pid_t, int, int);
989 extern struct sigqueue *sigqueue_alloc(void);
990 extern void sigqueue_free(struct sigqueue *);
991 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
992 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
993 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
994 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
996 /* These can be the second arg to send_sig_info/send_group_sig_info. */
997 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
998 #define SEND_SIG_PRIV ((struct siginfo *) 1)
999 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1001 /* True if we are on the alternate signal stack. */
1003 static inline int on_sig_stack(unsigned long sp)
1005 return (sp - current->sas_ss_sp < current->sas_ss_size);
1008 static inline int sas_ss_flags(unsigned long sp)
1010 return (current->sas_ss_size == 0 ? SS_DISABLE
1011 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1015 #ifdef CONFIG_SECURITY
1016 /* code is in security.c */
1017 extern int capable(int cap);
1018 extern int vx_capable(int cap, int ccap);
1020 static inline int capable(int cap)
1022 if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
1024 if (cap_raised(current->cap_effective, cap)) {
1025 current->flags |= PF_SUPERPRIV;
1031 static inline int vx_capable(int cap, int ccap)
1033 if (cap_raised(current->cap_effective, cap) &&
1035 current->flags |= PF_SUPERPRIV;
1043 * Routines for handling mm_structs
1045 extern struct mm_struct * mm_alloc(void);
1047 /* mmdrop drops the mm and the page tables */
1048 extern void FASTCALL(__mmdrop(struct mm_struct *));
1049 static inline void mmdrop(struct mm_struct * mm)
1051 if (atomic_dec_and_test(&mm->mm_count))
1055 /* mmput gets rid of the mappings and all user-space */
1056 extern void mmput(struct mm_struct *);
1057 /* Grab a reference to a task's mm, if it is not already going away */
1058 extern struct mm_struct *get_task_mm(struct task_struct *task);
1059 /* Remove the current tasks stale references to the old mm_struct */
1060 extern void mm_release(struct task_struct *, struct mm_struct *);
1062 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1063 extern void flush_thread(void);
1064 extern void exit_thread(void);
1066 extern void exit_files(struct task_struct *);
1067 extern void exit_signal(struct task_struct *);
1068 extern void __exit_signal(struct task_struct *);
1069 extern void exit_sighand(struct task_struct *);
1070 extern void __exit_sighand(struct task_struct *);
1071 extern void exit_itimers(struct signal_struct *);
1073 extern NORET_TYPE void do_group_exit(int);
1075 extern void daemonize(const char *, ...);
1076 extern int allow_signal(int);
1077 extern int disallow_signal(int);
1078 extern task_t *child_reaper;
1080 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1081 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1082 task_t *fork_idle(int);
1084 extern void set_task_comm(struct task_struct *tsk, char *from);
1085 extern void get_task_comm(char *to, struct task_struct *tsk);
1088 extern void wait_task_inactive(task_t * p);
1090 #define wait_task_inactive(p) do { } while (0)
1093 #define remove_parent(p) list_del_init(&(p)->sibling)
1094 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1096 #define REMOVE_LINKS(p) do { \
1097 if (thread_group_leader(p)) \
1098 list_del_init(&(p)->tasks); \
1102 #define SET_LINKS(p) do { \
1103 if (thread_group_leader(p)) \
1104 list_add_tail(&(p)->tasks,&init_task.tasks); \
1105 add_parent(p, (p)->parent); \
1108 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1109 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1111 #define for_each_process(p) \
1112 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1115 * Careful: do_each_thread/while_each_thread is a double loop so
1116 * 'break' will not work as expected - use goto instead.
1118 #define do_each_thread(g, t) \
1119 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1121 #define while_each_thread(g, t) \
1122 while ((t = next_thread(t)) != g)
1124 extern task_t * FASTCALL(next_thread(const task_t *p));
1126 #define thread_group_leader(p) (p->pid == p->tgid)
1128 static inline int thread_group_empty(task_t *p)
1130 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1133 #define delay_group_leader(p) \
1134 (thread_group_leader(p) && !thread_group_empty(p))
1136 extern void unhash_process(struct task_struct *p);
1139 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1140 * subscriptions and synchronises with wait4(). Also used in procfs.
1142 * Nests both inside and outside of read_lock(&tasklist_lock).
1143 * It must not be nested with write_lock_irq(&tasklist_lock),
1144 * neither inside nor outside.
1146 static inline void task_lock(struct task_struct *p)
1148 spin_lock(&p->alloc_lock);
1151 static inline void task_unlock(struct task_struct *p)
1153 spin_unlock(&p->alloc_lock);
1156 /* set thread flags in other task's structures
1157 * - see asm/thread_info.h for TIF_xxxx flags available
1159 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1161 set_ti_thread_flag(tsk->thread_info,flag);
1164 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1166 clear_ti_thread_flag(tsk->thread_info,flag);
1169 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1171 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1174 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1176 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1179 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1181 return test_ti_thread_flag(tsk->thread_info,flag);
1184 static inline void set_tsk_need_resched(struct task_struct *tsk)
1186 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1189 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1191 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1194 static inline int signal_pending(struct task_struct *p)
1196 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1199 static inline int need_resched(void)
1201 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1205 * cond_resched() and cond_resched_lock(): latency reduction via
1206 * explicit rescheduling in places that are safe. The return
1207 * value indicates whether a reschedule was done in fact.
1208 * cond_resched_lock() will drop the spinlock before scheduling,
1209 * cond_resched_softirq() will enable bhs before scheduling.
1211 extern int cond_resched(void);
1212 extern int cond_resched_lock(spinlock_t * lock);
1213 extern int cond_resched_softirq(void);
1216 * Does a critical section need to be broken due to another
1219 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1220 # define need_lockbreak(lock) ((lock)->break_lock)
1222 # define need_lockbreak(lock) 0
1226 * Does a critical section need to be broken due to another
1227 * task waiting or preemption being signalled:
1229 static inline int lock_need_resched(spinlock_t *lock)
1231 if (need_lockbreak(lock) || need_resched())
1236 /* Reevaluate whether the task has signals pending delivery.
1237 This is required every time the blocked sigset_t changes.
1238 callers must hold sighand->siglock. */
1240 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1241 extern void recalc_sigpending(void);
1243 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1246 * Wrappers for p->thread_info->cpu access. No-op on UP.
1250 static inline unsigned int task_cpu(const struct task_struct *p)
1252 return p->thread_info->cpu;
1255 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1257 p->thread_info->cpu = cpu;
1262 static inline unsigned int task_cpu(const struct task_struct *p)
1267 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1271 #endif /* CONFIG_SMP */
1273 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1274 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1276 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1278 mm->mmap_base = TASK_UNMAPPED_BASE;
1279 mm->get_unmapped_area = arch_get_unmapped_area;
1280 mm->unmap_area = arch_unmap_area;
1284 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1285 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1287 #ifdef CONFIG_MAGIC_SYSRQ
1289 extern void normalize_rt_tasks(void);
1295 * Checks whether we need to enter the refrigerator
1296 * and returns 1 if we did so.
1299 extern void refrigerator(unsigned long);
1300 extern int freeze_processes(void);
1301 extern void thaw_processes(void);
1303 static inline int try_to_freeze(unsigned long refrigerator_flags)
1305 if (unlikely(current->flags & PF_FREEZE)) {
1306 refrigerator(refrigerator_flags);
1312 static inline void refrigerator(unsigned long flag) {}
1313 static inline int freeze_processes(void) { BUG(); return 0; }
1314 static inline void thaw_processes(void) {}
1316 static inline int try_to_freeze(unsigned long refrigerator_flags)
1320 #endif /* CONFIG_PM */
1321 #endif /* __KERNEL__ */