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>
17 #include <asm/system.h>
18 #include <asm/semaphore.h>
20 #include <asm/ptrace.h>
23 #include <linux/smp.h>
24 #include <linux/sem.h>
25 #include <linux/signal.h>
26 #include <linux/securebits.h>
27 #include <linux/fs_struct.h>
28 #include <linux/compiler.h>
29 #include <linux/completion.h>
30 #include <linux/pid.h>
31 #include <linux/percpu.h>
32 #include <linux/topology.h>
33 #include <linux/vs_base.h>
36 extern int exec_shield;
37 extern int exec_shield_randomize;
38 extern int print_fatal_signals;
43 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
44 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
45 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
46 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
47 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
48 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
49 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
50 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
51 #define CLONE_THREAD 0x00010000 /* Same thread group? */
52 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
53 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
54 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
55 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
56 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
57 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
58 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
59 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
60 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
63 * List of flags we want to share for kernel threads,
64 * if only because they are not used by them anyway.
66 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
69 * These are the constant used to fake the fixed-point load-average
70 * counting. Some notes:
71 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
72 * a load-average precision of 10 bits integer + 11 bits fractional
73 * - if you want to count load-averages more often, you need more
74 * precision, or rounding will get you. With 2-second counting freq,
75 * the EXP_n values would be 1981, 2034 and 2043 if still using only
78 extern unsigned long avenrun[]; /* Load averages */
80 #define FSHIFT 11 /* nr of bits of precision */
81 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
82 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
83 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
84 #define EXP_5 2014 /* 1/exp(5sec/5min) */
85 #define EXP_15 2037 /* 1/exp(5sec/15min) */
87 #define CALC_LOAD(load,exp,n) \
89 load += n*(FIXED_1-exp); \
92 #define CT_TO_SECS(x) ((x) / HZ)
93 #define CT_TO_USECS(x) (((x) % HZ) * 1000000/HZ)
95 extern int nr_threads;
97 DECLARE_PER_CPU(unsigned long, process_counts);
98 extern int nr_processes(void);
99 extern unsigned long nr_running(void);
100 extern unsigned long nr_uninterruptible(void);
101 extern unsigned long nr_iowait(void);
103 #include <linux/time.h>
104 #include <linux/param.h>
105 #include <linux/resource.h>
106 #include <linux/timer.h>
108 #include <asm/processor.h>
110 #define TASK_RUNNING 0
111 #define TASK_INTERRUPTIBLE 1
112 #define TASK_UNINTERRUPTIBLE 2
113 #define TASK_STOPPED 4
114 #define TASK_TRACED 8
115 #define EXIT_ZOMBIE 16
117 #define TASK_ONHOLD 64
119 #define __set_task_state(tsk, state_value) \
120 do { (tsk)->state = (state_value); } while (0)
121 #define set_task_state(tsk, state_value) \
122 set_mb((tsk)->state, (state_value))
124 #define __set_current_state(state_value) \
125 do { current->state = (state_value); } while (0)
126 #define set_current_state(state_value) \
127 set_mb(current->state, (state_value))
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(int user_tick, int system);
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);
203 arch_get_unmapped_exec_area(struct file *, unsigned long, unsigned long,
204 unsigned long, unsigned long);
206 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
207 unsigned long len, unsigned long pgoff,
208 unsigned long flags);
209 extern void arch_unmap_area(struct vm_area_struct *area);
210 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
214 struct vm_area_struct * mmap; /* list of VMAs */
215 struct rb_root mm_rb;
216 struct vm_area_struct * mmap_cache; /* last find_vma result */
217 unsigned long (*get_unmapped_area) (struct file *filp,
218 unsigned long addr, unsigned long len,
219 unsigned long pgoff, unsigned long flags);
220 unsigned long (*get_unmapped_exec_area) (struct file *filp,
221 unsigned long addr, unsigned long len,
222 unsigned long pgoff, unsigned long flags);
223 void (*unmap_area) (struct vm_area_struct *area);
224 unsigned long mmap_base; /* base of mmap area */
225 unsigned long free_area_cache; /* first hole */
227 atomic_t mm_users; /* How many users with user space? */
228 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
229 int map_count; /* number of VMAs */
230 struct rw_semaphore mmap_sem;
231 spinlock_t page_table_lock; /* Protects page tables, mm->rss, mm->anon_rss */
233 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
234 * together off init_mm.mmlist, and are protected
238 unsigned long start_code, end_code, start_data, end_data;
239 unsigned long start_brk, brk, start_stack;
240 unsigned long arg_start, arg_end, env_start, env_end;
241 unsigned long rss, anon_rss, total_vm, locked_vm, shared_vm;
242 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
244 unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
247 cpumask_t cpu_vm_mask;
249 /* Architecture-specific MM context */
250 mm_context_t context;
251 struct vx_info *mm_vx_info;
253 /* Token based thrashing protection. */
254 unsigned long swap_token_time;
257 /* coredumping support */
259 struct completion *core_startup_done, core_done;
262 rwlock_t ioctx_list_lock;
263 struct kioctx *ioctx_list;
265 struct kioctx default_kioctx;
268 struct sighand_struct {
270 struct k_sigaction action[_NSIG];
275 * NOTE! "signal_struct" does not have it's own
276 * locking, because a shared signal_struct always
277 * implies a shared sighand_struct, so locking
278 * sighand_struct is always a proper superset of
279 * the locking of signal_struct.
281 struct signal_struct {
285 /* current thread group signal load-balancing target: */
288 /* shared signal handling: */
289 struct sigpending shared_pending;
291 /* thread group exit support */
295 * - notify group_exit_task when ->count is equal to notify_count
296 * - everyone except group_exit_task is stopped during signal delivery
297 * of fatal signals, group_exit_task processes the signal.
299 struct task_struct *group_exit_task;
302 /* thread group stop support, overloads group_exit_code too */
303 int group_stop_count;
304 /* 1 if group stopped since last SIGCONT, -1 if SIGCONT since report */
307 /* POSIX.1b Interval Timers */
308 struct list_head posix_timers;
310 /* job control IDs */
314 /* boolean value for session group leader */
317 struct tty_struct *tty; /* NULL if no tty */
320 * Cumulative resource counters for dead threads in the group,
321 * and for reaped dead child processes forked by this group.
322 * Live threads maintain their own counters and add to these
323 * in __exit_signal, except for the group leader.
325 unsigned long utime, stime, cutime, cstime;
326 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
327 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
330 * We don't bother to synchronize most readers of this at all,
331 * because there is no reader checking a limit that actually needs
332 * to get both rlim_cur and rlim_max atomically, and either one
333 * alone is a single word that can safely be read normally.
334 * getrlimit/setrlimit use task_lock(current->group_leader) to
335 * protect this instead of the siglock, because they really
336 * have no need to disable irqs.
338 struct rlimit rlim[RLIM_NLIMITS];
342 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
343 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
344 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
345 * are inverted: lower p->prio value means higher priority.
347 * The MAX_USER_RT_PRIO value allows the actual maximum
348 * RT priority to be separate from the value exported to
349 * user-space. This allows kernel threads to set their
350 * priority to a value higher than any user task. Note:
351 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
354 #define MAX_USER_RT_PRIO 100
355 #define MAX_RT_PRIO MAX_USER_RT_PRIO
357 #define MAX_PRIO (MAX_RT_PRIO + 40)
359 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
362 * Some day this will be a full-fledged user tracking system..
365 atomic_t __count; /* reference count */
366 atomic_t processes; /* How many processes does this user have? */
367 atomic_t files; /* How many open files does this user have? */
368 atomic_t sigpending; /* How many pending signals does this user have? */
369 /* protected by mq_lock */
370 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
371 unsigned long locked_shm; /* How many pages of mlocked shm ? */
374 struct key *uid_keyring; /* UID specific keyring */
375 struct key *session_keyring; /* UID's default session keyring */
378 /* Hash table maintenance information */
379 struct list_head uidhash_list;
384 extern struct user_struct *find_user(xid_t, uid_t);
386 extern struct user_struct root_user;
387 #define INIT_USER (&root_user)
389 typedef struct prio_array prio_array_t;
390 struct backing_dev_info;
391 struct reclaim_state;
393 #ifdef CONFIG_SCHEDSTATS
395 /* cumulative counters */
396 unsigned long cpu_time, /* time spent on the cpu */
397 run_delay, /* time spent waiting on a runqueue */
398 pcnt; /* # of timeslices run on this cpu */
401 unsigned long last_arrival, /* when we last ran on a cpu */
402 last_queued; /* when we were last queued to run */
405 extern struct file_operations proc_schedstat_operations;
417 * sched-domains (multiprocessor balancing) declarations:
420 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
422 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
423 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
424 #define SD_BALANCE_EXEC 4 /* Balance on exec */
425 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
426 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
427 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
428 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
431 struct sched_group *next; /* Must be a circular list */
435 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
436 * single CPU. This is read only (except for setup, hotplug CPU).
438 unsigned long cpu_power;
441 struct sched_domain {
442 /* These fields must be setup */
443 struct sched_domain *parent; /* top domain must be null terminated */
444 struct sched_group *groups; /* the balancing groups of the domain */
445 cpumask_t span; /* span of all CPUs in this domain */
446 unsigned long min_interval; /* Minimum balance interval ms */
447 unsigned long max_interval; /* Maximum balance interval ms */
448 unsigned int busy_factor; /* less balancing by factor if busy */
449 unsigned int imbalance_pct; /* No balance until over watermark */
450 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
451 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
452 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
453 int flags; /* See SD_* */
455 /* Runtime fields. */
456 unsigned long last_balance; /* init to jiffies. units in jiffies */
457 unsigned int balance_interval; /* initialise to 1. units in ms. */
458 unsigned int nr_balance_failed; /* initialise to 0 */
460 #ifdef CONFIG_SCHEDSTATS
461 /* load_balance() stats */
462 unsigned long lb_cnt[MAX_IDLE_TYPES];
463 unsigned long lb_failed[MAX_IDLE_TYPES];
464 unsigned long lb_imbalance[MAX_IDLE_TYPES];
465 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
466 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
468 /* sched_balance_exec() stats */
469 unsigned long sbe_attempts;
470 unsigned long sbe_pushed;
472 /* try_to_wake_up() stats */
473 unsigned long ttwu_wake_affine;
474 unsigned long ttwu_wake_balance;
478 #ifdef ARCH_HAS_SCHED_DOMAIN
479 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
480 extern cpumask_t cpu_isolated_map;
481 extern void init_sched_build_groups(struct sched_group groups[],
482 cpumask_t span, int (*group_fn)(int cpu));
483 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
484 #endif /* ARCH_HAS_SCHED_DOMAIN */
485 #endif /* CONFIG_SMP */
488 struct io_context; /* See blkdev.h */
489 void exit_io_context(void);
491 #define NGROUPS_SMALL 32
492 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
496 gid_t small_block[NGROUPS_SMALL];
502 * get_group_info() must be called with the owning task locked (via task_lock())
503 * when task != current. The reason being that the vast majority of callers are
504 * looking at current->group_info, which can not be changed except by the
505 * current task. Changing current->group_info requires the task lock, too.
507 #define get_group_info(group_info) do { \
508 atomic_inc(&(group_info)->usage); \
511 #define put_group_info(group_info) do { \
512 if (atomic_dec_and_test(&(group_info)->usage)) \
513 groups_free(group_info); \
516 struct group_info *groups_alloc(int gidsetsize);
517 void groups_free(struct group_info *group_info);
518 int set_current_groups(struct group_info *group_info);
519 /* access the groups "array" with this macro */
520 #define GROUP_AT(gi, i) \
521 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
524 struct audit_context; /* See audit.c */
528 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
529 struct thread_info *thread_info;
531 unsigned long flags; /* per process flags, defined below */
532 unsigned long ptrace;
534 int lock_depth; /* Lock depth */
536 int prio, static_prio;
537 struct list_head run_list;
540 unsigned long sleep_avg;
541 long interactive_credit;
542 unsigned long long timestamp, last_ran;
545 unsigned long policy;
546 cpumask_t cpus_allowed;
547 unsigned int time_slice, first_time_slice;
549 #ifdef CONFIG_SCHEDSTATS
550 struct sched_info sched_info;
553 struct list_head tasks;
555 * ptrace_list/ptrace_children forms the list of my children
556 * that were stolen by a ptracer.
558 struct list_head ptrace_children;
559 struct list_head ptrace_list;
561 struct mm_struct *mm, *active_mm;
564 struct linux_binfmt *binfmt;
566 int exit_code, exit_signal;
567 int pdeath_signal; /* The signal sent when the parent dies */
569 unsigned long personality;
574 * pointers to (original) parent process, youngest child, younger sibling,
575 * older sibling, respectively. (p->father can be replaced with
578 struct task_struct *real_parent; /* real parent process (when being debugged) */
579 struct task_struct *parent; /* parent process */
581 * children/sibling forms the list of my children plus the
582 * tasks I'm ptracing.
584 struct list_head children; /* list of my children */
585 struct list_head sibling; /* linkage in my parent's children list */
586 struct task_struct *group_leader; /* threadgroup leader */
588 /* PID/PID hash table linkage. */
589 struct pid pids[PIDTYPE_MAX];
591 wait_queue_head_t wait_chldexit; /* for wait4() */
592 struct completion *vfork_done; /* for vfork() */
593 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
594 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
596 unsigned long rt_priority;
597 unsigned long it_real_value, it_prof_value, it_virt_value;
598 unsigned long it_real_incr, it_prof_incr, it_virt_incr;
599 struct timer_list real_timer;
600 unsigned long utime, stime;
601 unsigned long nvcsw, nivcsw; /* context switch counts */
602 struct timespec start_time;
603 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
604 unsigned long min_flt, maj_flt;
605 /* process credentials */
606 uid_t uid,euid,suid,fsuid;
607 gid_t gid,egid,sgid,fsgid;
608 struct group_info *group_info;
609 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
610 unsigned keep_capabilities:1;
611 struct user_struct *user;
613 struct key *session_keyring; /* keyring inherited over fork */
614 struct key *process_keyring; /* keyring private to this process (CLONE_THREAD) */
615 struct key *thread_keyring; /* keyring private to this thread */
617 unsigned short used_math;
619 /* file system info */
620 int link_count, total_link_count;
622 struct sysv_sem sysvsem;
623 /* CPU-specific state of this task */
624 struct thread_struct thread;
625 /* filesystem information */
626 struct fs_struct *fs;
627 /* open file information */
628 struct files_struct *files;
630 struct namespace *namespace;
631 /* signal handlers */
632 struct signal_struct *signal;
633 struct sighand_struct *sighand;
634 sigset_t blocked, real_blocked;
635 struct sigpending pending;
637 unsigned long sas_ss_sp;
639 int (*notifier)(void *priv);
641 sigset_t *notifier_mask;
645 void (*tux_exit)(void);
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;
682 unsigned long ptrace_message;
683 siginfo_t *last_siginfo; /* For ptrace use. */
685 * current io wait handle: wait queue entry to use for io waits
686 * If this thread is processing aio, this points at the waitqueue
687 * inside the currently handled kiocb. It may be NULL (i.e. default
688 * to a stack based synchronous wait) if its doing sync IO.
690 wait_queue_t *io_wait;
692 struct mempolicy *mempolicy;
693 short il_next; /* could be shared with used_math */
697 static inline pid_t process_group(struct task_struct *tsk)
699 return tsk->signal->pgrp;
703 * pid_alive - check that a task structure is not stale
704 * @p: Task structure to be checked.
706 * Test if a process is not yet dead (at most zombie state)
707 * If pid_alive fails, then pointers within the task structure
708 * can be stale and must not be dereferenced.
710 static inline int pid_alive(struct task_struct *p)
712 return p->pids[PIDTYPE_PID].nr != 0;
715 extern void free_task(struct task_struct *tsk);
716 extern void __put_task_struct(struct task_struct *tsk);
717 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
718 #define put_task_struct(tsk) \
719 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
724 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
725 /* Not implemented yet, only for 486*/
726 #define PF_STARTING 0x00000002 /* being created */
727 #define PF_EXITING 0x00000004 /* getting shut down */
728 #define PF_DEAD 0x00000008 /* Dead */
729 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
730 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
731 #define PF_DUMPCORE 0x00000200 /* dumped core */
732 #define PF_SIGNALED 0x00000400 /* killed by a signal */
733 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
734 #define PF_MEMDIE 0x00001000 /* Killed for out-of-memory */
735 #define PF_FLUSHER 0x00002000 /* responsible for disk writeback */
737 #define PF_FREEZE 0x00004000 /* this task should be frozen for suspend */
738 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
739 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
740 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
741 #define PF_KSWAPD 0x00040000 /* I am kswapd */
742 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
743 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
744 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
745 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
746 #define PF_RELOCEXEC 0x00800000 /* relocate shared libraries */
749 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
751 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
757 extern unsigned long long sched_clock(void);
759 /* sched_exec is called by processes performing an exec */
761 extern void sched_exec(void);
763 #define sched_exec() {}
766 extern void sched_idle_next(void);
767 extern void set_user_nice(task_t *p, long nice);
768 extern int task_prio(const task_t *p);
769 extern int task_nice(const task_t *p);
770 extern int task_curr(const task_t *p);
771 extern int idle_cpu(int cpu);
776 * The default (Linux) execution domain.
778 extern struct exec_domain default_exec_domain;
781 struct thread_info thread_info;
782 unsigned long stack[THREAD_SIZE/sizeof(long)];
785 #ifndef __HAVE_ARCH_KSTACK_END
786 static inline int kstack_end(void *addr)
788 /* Reliable end of stack detection:
789 * Some APM bios versions misalign the stack
791 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
795 extern union thread_union init_thread_union;
796 extern struct task_struct init_task;
798 extern struct mm_struct init_mm;
801 #define find_task_by_real_pid(nr) \
802 find_task_by_pid_type(PIDTYPE_PID, nr)
803 #define find_task_by_pid(nr) \
804 find_task_by_pid_type(PIDTYPE_PID, \
807 extern struct task_struct *find_task_by_pid_type(int type, int pid);
808 extern void set_special_pids(pid_t session, pid_t pgrp);
809 extern void __set_special_pids(pid_t session, pid_t pgrp);
811 /* per-UID process charging. */
812 extern struct user_struct * alloc_uid(xid_t, uid_t);
813 static inline struct user_struct *get_uid(struct user_struct *u)
815 atomic_inc(&u->__count);
819 extern void free_uid(struct user_struct *);
820 extern void switch_uid(struct user_struct *);
822 #include <asm/current.h>
824 extern void do_timer(struct pt_regs *);
826 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
827 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
828 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
829 unsigned long clone_flags));
831 extern void kick_process(struct task_struct *tsk);
833 static inline void kick_process(struct task_struct *tsk) { }
835 extern void FASTCALL(sched_fork(task_t * p));
836 extern void FASTCALL(sched_exit(task_t * p));
838 extern int in_group_p(gid_t);
839 extern int in_egroup_p(gid_t);
841 extern void proc_caches_init(void);
842 extern void flush_signals(struct task_struct *);
843 extern void flush_signal_handlers(struct task_struct *, int force_default);
844 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
846 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
851 spin_lock_irqsave(&tsk->sighand->siglock, flags);
852 ret = dequeue_signal(tsk, mask, info);
853 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
858 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
860 extern void unblock_all_signals(void);
861 extern void release_task(struct task_struct * p);
862 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
863 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
864 extern int force_sigsegv(int, struct task_struct *);
865 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
866 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
867 extern int kill_pg_info(int, struct siginfo *, pid_t);
868 extern int kill_proc_info(int, struct siginfo *, pid_t);
869 extern void do_notify_parent(struct task_struct *, int);
870 extern void force_sig(int, struct task_struct *);
871 extern void force_sig_specific(int, struct task_struct *);
872 extern int send_sig(int, struct task_struct *, int);
873 extern void zap_other_threads(struct task_struct *p);
874 extern int kill_pg(pid_t, int, int);
875 extern int kill_sl(pid_t, int, int);
876 extern int kill_proc(pid_t, int, int);
877 extern struct sigqueue *sigqueue_alloc(void);
878 extern void sigqueue_free(struct sigqueue *);
879 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
880 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
881 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
882 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
884 /* These can be the second arg to send_sig_info/send_group_sig_info. */
885 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
886 #define SEND_SIG_PRIV ((struct siginfo *) 1)
887 #define SEND_SIG_FORCED ((struct siginfo *) 2)
889 /* True if we are on the alternate signal stack. */
891 static inline int on_sig_stack(unsigned long sp)
893 return (sp - current->sas_ss_sp < current->sas_ss_size);
896 static inline int sas_ss_flags(unsigned long sp)
898 return (current->sas_ss_size == 0 ? SS_DISABLE
899 : on_sig_stack(sp) ? SS_ONSTACK : 0);
903 #ifdef CONFIG_SECURITY
904 /* code is in security.c */
905 extern int capable(int cap);
906 extern int vx_capable(int cap, int ccap);
908 static inline int capable(int cap)
910 if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
912 if (cap_raised(current->cap_effective, cap)) {
913 current->flags |= PF_SUPERPRIV;
919 static inline int vx_capable(int cap, int ccap)
921 if (cap_raised(current->cap_effective, cap) &&
923 current->flags |= PF_SUPERPRIV;
932 * Routines for handling mm_structs
934 extern struct mm_struct * mm_alloc(void);
936 /* mmdrop drops the mm and the page tables */
937 extern void FASTCALL(__mmdrop(struct mm_struct *));
938 static inline void mmdrop(struct mm_struct * mm)
940 if (atomic_dec_and_test(&mm->mm_count))
944 /* mmput gets rid of the mappings and all user-space */
945 extern void mmput(struct mm_struct *);
946 /* Grab a reference to a task's mm, if it is not already going away */
947 extern struct mm_struct *get_task_mm(struct task_struct *task);
948 /* Remove the current tasks stale references to the old mm_struct */
949 extern void mm_release(struct task_struct *, struct mm_struct *);
951 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
952 extern void flush_thread(void);
953 extern void exit_thread(void);
955 extern void exit_mm(struct task_struct *);
956 extern void exit_files(struct task_struct *);
957 extern void exit_signal(struct task_struct *);
958 extern void __exit_signal(struct task_struct *);
959 extern void exit_sighand(struct task_struct *);
960 extern void __exit_sighand(struct task_struct *);
961 extern void exit_itimers(struct signal_struct *);
963 extern NORET_TYPE void do_group_exit(int);
965 extern void reparent_to_init(void);
966 extern void daemonize(const char *, ...);
967 extern int allow_signal(int);
968 extern int disallow_signal(int);
969 extern task_t *child_reaper;
971 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
972 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
973 task_t *fork_idle(int);
975 extern void set_task_comm(struct task_struct *tsk, char *from);
976 extern void get_task_comm(char *to, struct task_struct *tsk);
979 extern void wait_task_inactive(task_t * p);
981 #define wait_task_inactive(p) do { } while (0)
984 #define remove_parent(p) list_del_init(&(p)->sibling)
985 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
987 #define REMOVE_LINKS(p) do { \
988 if (thread_group_leader(p)) \
989 list_del_init(&(p)->tasks); \
993 #define SET_LINKS(p) do { \
994 if (thread_group_leader(p)) \
995 list_add_tail(&(p)->tasks,&init_task.tasks); \
996 add_parent(p, (p)->parent); \
999 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1000 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1002 #define for_each_process(p) \
1003 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1006 * Careful: do_each_thread/while_each_thread is a double loop so
1007 * 'break' will not work as expected - use goto instead.
1009 #define do_each_thread(g, t) \
1010 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1012 #define while_each_thread(g, t) \
1013 while ((t = next_thread(t)) != g)
1015 extern task_t * FASTCALL(next_thread(const task_t *p));
1017 #define thread_group_leader(p) (p->pid == p->tgid)
1019 static inline int thread_group_empty(task_t *p)
1021 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1024 #define delay_group_leader(p) \
1025 (thread_group_leader(p) && !thread_group_empty(p))
1027 extern void unhash_process(struct task_struct *p);
1030 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1031 * subscriptions and synchronises with wait4(). Also used in procfs.
1033 * Nests both inside and outside of read_lock(&tasklist_lock).
1034 * It must not be nested with write_lock_irq(&tasklist_lock),
1035 * neither inside nor outside.
1037 static inline void task_lock(struct task_struct *p)
1039 spin_lock(&p->alloc_lock);
1042 static inline void task_unlock(struct task_struct *p)
1044 spin_unlock(&p->alloc_lock);
1047 /* set thread flags in other task's structures
1048 * - see asm/thread_info.h for TIF_xxxx flags available
1050 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1052 set_ti_thread_flag(tsk->thread_info,flag);
1055 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1057 clear_ti_thread_flag(tsk->thread_info,flag);
1060 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1062 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1065 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1067 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1070 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1072 return test_ti_thread_flag(tsk->thread_info,flag);
1075 static inline void set_tsk_need_resched(struct task_struct *tsk)
1077 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1080 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1082 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1085 static inline int signal_pending(struct task_struct *p)
1087 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1090 static inline int need_resched(void)
1092 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1095 extern void __cond_resched(void);
1096 static inline void cond_resched(void)
1103 * cond_resched_lock() - if a reschedule is pending, drop the given lock,
1104 * call schedule, and on return reacquire the lock.
1106 * This works OK both with and without CONFIG_PREEMPT. We do strange low-level
1107 * operations here to prevent schedule() from being called twice (once via
1108 * spin_unlock(), once by hand).
1110 static inline void cond_resched_lock(spinlock_t * lock)
1112 if (need_resched()) {
1113 _raw_spin_unlock(lock);
1114 preempt_enable_no_resched();
1120 /* Reevaluate whether the task has signals pending delivery.
1121 This is required every time the blocked sigset_t changes.
1122 callers must hold sighand->siglock. */
1124 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1125 extern void recalc_sigpending(void);
1127 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1130 * Wrappers for p->thread_info->cpu access. No-op on UP.
1134 static inline unsigned int task_cpu(const struct task_struct *p)
1136 return p->thread_info->cpu;
1139 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1141 p->thread_info->cpu = cpu;
1146 static inline unsigned int task_cpu(const struct task_struct *p)
1151 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1155 #endif /* CONFIG_SMP */
1157 /* API for registering delay info */
1158 #ifdef CONFIG_DELAY_ACCT
1160 #define test_delay_flag(tsk,flg) ((tsk)->flags & (flg))
1161 #define set_delay_flag(tsk,flg) ((tsk)->flags |= (flg))
1162 #define clear_delay_flag(tsk,flg) ((tsk)->flags &= ~(flg))
1164 #define def_delay_var(var) unsigned long long var
1165 #define get_delay(tsk,field) ((tsk)->delays.field)
1167 #define start_delay(var) ((var) = sched_clock())
1168 #define start_delay_set(var,flg) (set_delay_flag(current,flg),(var) = sched_clock())
1170 #define inc_delay(tsk,field) (((tsk)->delays.field)++)
1172 /* because of hardware timer drifts in SMPs and task continue on different cpu
1173 * then where the start_ts was taken there is a possibility that
1174 * end_ts < start_ts by some usecs. In this case we ignore the diff
1175 * and add nothing to the total.
1178 #define test_ts_integrity(start_ts,end_ts) (likely((end_ts) > (start_ts)))
1180 #define test_ts_integrity(start_ts,end_ts) (1)
1183 #define add_delay_ts(tsk,field,start_ts,end_ts) \
1184 do { if (test_ts_integrity(start_ts,end_ts)) (tsk)->delays.field += ((end_ts)-(start_ts)); } while (0)
1186 #define add_delay_clear(tsk,field,start_ts,flg) \
1188 unsigned long long now = sched_clock();\
1189 add_delay_ts(tsk,field,start_ts,now); \
1190 clear_delay_flag(tsk,flg); \
1193 static inline void add_io_delay(unsigned long long dstart)
1195 struct task_struct * tsk = current;
1196 unsigned long long now = sched_clock();
1197 unsigned long long val;
1199 if (test_ts_integrity(dstart,now))
1203 if (test_delay_flag(tsk,PF_MEMIO)) {
1204 tsk->delays.mem_iowait_total += val;
1205 tsk->delays.num_memwaits++;
1207 tsk->delays.iowait_total += val;
1208 tsk->delays.num_iowaits++;
1210 clear_delay_flag(tsk,PF_IOWAIT);
1213 inline static void init_delays(struct task_struct *tsk)
1215 memset((void*)&tsk->delays,0,sizeof(tsk->delays));
1220 #define test_delay_flag(tsk,flg) (0)
1221 #define set_delay_flag(tsk,flg) do { } while (0)
1222 #define clear_delay_flag(tsk,flg) do { } while (0)
1224 #define def_delay_var(var)
1225 #define get_delay(tsk,field) (0)
1227 #define start_delay(var) do { } while (0)
1228 #define start_delay_set(var,flg) do { } while (0)
1230 #define inc_delay(tsk,field) do { } while (0)
1231 #define add_delay_ts(tsk,field,start_ts,now) do { } while (0)
1232 #define add_delay_clear(tsk,field,start_ts,flg) do { } while (0)
1233 #define add_io_delay(dstart) do { } while (0)
1234 #define init_delays(tsk) do { } while (0)
1239 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1240 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1242 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1244 mm->mmap_base = TASK_UNMAPPED_BASE;
1245 mm->get_unmapped_area = arch_get_unmapped_area;
1246 mm->unmap_area = arch_unmap_area;
1250 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1251 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1253 #ifdef CONFIG_MAGIC_SYSRQ
1255 extern void normalize_rt_tasks(void);
1259 #endif /* __KERNEL__ */