linux/include/linux/sched.h History
<<
>>
Prefs
   1#ifndef _LINUX_SCHED_H
   2#define _LINUX_SCHED_H
   3
   4#include <asm/param.h>  /* for HZ */
   5
   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
  18#include <asm/system.h>
  19#include <asm/semaphore.h>
  20#include <asm/page.h>
  21#include <asm/ptrace.h>
  22#include <asm/mmu.h>
  23#include <asm/cputime.h>
  24
  25#include <linux/smp.h>
  26#include <linux/sem.h>
  27#include <linux/signal.h>
  28#include <linux/securebits.h>
  29#include <linux/fs_struct.h>
  30#include <linux/compiler.h>
  31#include <linux/completion.h>
  32#include <linux/pid.h>
  33#include <linux/percpu.h>
  34#include <linux/topology.h>
  35
  36struct exec_domain;
  37
  38/*
  39 * cloning flags:
  40 */
  41#define CSIGNAL         0x000000ff      /* signal mask to be sent at exit */
  42#define CLONE_VM        0x00000100      /* set if VM shared between processes */
  43#define CLONE_FS        0x00000200      /* set if fs info shared between processes */
  44#define CLONE_FILES     0x00000400      /* set if open files shared between processes */
  45#define CLONE_SIGHAND   0x00000800      /* set if signal handlers and blocked signals shared */
  46#define CLONE_PTRACE    0x00002000      /* set if we want to let tracing continue on the child too */
  47#define CLONE_VFORK     0x00004000      /* set if the parent wants the child to wake it up on mm_release */
  48#define CLONE_PARENT    0x00008000      /* set if we want to have the same parent as the cloner */
  49#define CLONE_THREAD    0x00010000      /* Same thread group? */
  50#define CLONE_NEWNS     0x00020000      /* New namespace group? */
  51#define CLONE_SYSVSEM   0x00040000      /* share system V SEM_UNDO semantics */
  52#define CLONE_SETTLS    0x00080000      /* create a new TLS for the child */
  53#define CLONE_PARENT_SETTID     0x00100000      /* set the TID in the parent */
  54#define CLONE_CHILD_CLEARTID    0x00200000      /* clear the TID in the child */
  55#define CLONE_DETACHED          0x00400000      /* Unused, ignored */
  56#define CLONE_UNTRACED          0x00800000      /* set if the tracing process can't force CLONE_PTRACE on this clone */
  57#define CLONE_CHILD_SETTID      0x01000000      /* set the TID in the child */
  58#define CLONE_STOPPED           0x02000000      /* Start in stopped state */
  59
  60/*
  61 * List of flags we want to share for kernel threads,
  62 * if only because they are not used by them anyway.
  63 */
  64#define CLONE_KERNEL    (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
  65
  66/*
  67 * These are the constant used to fake the fixed-point load-average
  68 * counting. Some notes:
  69 *  - 11 bit fractions expand to 22 bits by the multiplies: this gives
  70 *    a load-average precision of 10 bits integer + 11 bits fractional
  71 *  - if you want to count load-averages more often, you need more
  72 *    precision, or rounding will get you. With 2-second counting freq,
  73 *    the EXP_n values would be 1981, 2034 and 2043 if still using only
  74 *    11 bit fractions.
  75 */
  76extern unsigned long avenrun[];         /* Load averages */
  77
  78#define FSHIFT          11              /* nr of bits of precision */
  79#define FIXED_1         (1<<FSHIFT)     /* 1.0 as fixed-point */
  80#define LOAD_FREQ       (5*HZ)          /* 5 sec intervals */
  81#define EXP_1           1884            /* 1/exp(5sec/1min) as fixed-point */
  82#define EXP_5           2014            /* 1/exp(5sec/5min) */
  83#define EXP_15          2037            /* 1/exp(5sec/15min) */
  84
  85#define CALC_LOAD(load,exp,n) \
  86        load *= exp; \
  87        load += n*(FIXED_1-exp); \
  88        load >>= FSHIFT;
  89
  90extern unsigned long total_forks;
  91extern int nr_threads;
  92extern int last_pid;
  93DECLARE_PER_CPU(unsigned long, process_counts);
  94extern int nr_processes(void);
  95extern unsigned long nr_running(void);
  96extern unsigned long nr_uninterruptible(void);
  97extern unsigned long nr_iowait(void);
  98
  99#include <linux/time.h>
 100#include <linux/param.h>
 101#include <linux/resource.h>
 102#include <linux/timer.h>
 103
 104#include <asm/processor.h>
 105
 106#define TASK_RUNNING            0
 107#define TASK_INTERRUPTIBLE      1
 108#define TASK_UNINTERRUPTIBLE    2
 109#define TASK_STOPPED            4
 110#define TASK_TRACED             8
 111#define EXIT_ZOMBIE             16
 112#define EXIT_DEAD               32
 113
 114#define __set_task_state(tsk, state_value)              \
 115        do { (tsk)->state = (state_value); } while (0)
 116#define set_task_state(tsk, state_value)                \
 117        set_mb((tsk)->state, (state_value))
 118
 119#define __set_current_state(state_value)                        \
 120        do { current->state = (state_value); } while (0)
 121#define set_current_state(state_value)          \
 122        set_mb(current->state, (state_value))
 123
 124/* Task command name length */
 125#define TASK_COMM_LEN 16
 126
 127/*
 128 * Scheduling policies
 129 */
 130#define SCHED_NORMAL            0
 131#define SCHED_FIFO              1
 132#define SCHED_RR                2
 133
 134struct sched_param {
 135        int sched_priority;
 136};
 137
 138#ifdef __KERNEL__
 139
 140#include <linux/spinlock.h>
 141
 142/*
 143 * This serializes "schedule()" and also protects
 144 * the run-queue from deletions/modifications (but
 145 * _adding_ to the beginning of the run-queue has
 146 * a separate lock).
 147 */
 148extern rwlock_t tasklist_lock;
 149extern spinlock_t mmlist_lock;
 150
 151typedef struct task_struct task_t;
 152
 153extern void sched_init(void);
 154extern void sched_init_smp(void);
 155extern void init_idle(task_t *idle, int cpu);
 156
 157extern cpumask_t nohz_cpu_mask;
 158
 159extern void show_state(void);
 160extern void show_regs(struct pt_regs *);
 161
 162/*
 163 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
 164 * task), SP is the stack pointer of the first frame that should be shown in the back
 165 * trace (or NULL if the entire call-chain of the task should be shown).
 166 */
 167extern void show_stack(struct task_struct *task, unsigned long *sp);
 168
 169void io_schedule(void);
 170long io_schedule_timeout(long timeout);
 171
 172extern void cpu_init (void);
 173extern void trap_init(void);
 174extern void update_process_times(int user);
 175extern void scheduler_tick(void);
 176extern unsigned long cache_decay_ticks;
 177
 178/* Attach to any functions which should be ignored in wchan output. */
 179#define __sched         __attribute__((__section__(".sched.text")))
 180/* Is this address in the __sched functions? */
 181extern int in_sched_functions(unsigned long addr);
 182
 183#define MAX_SCHEDULE_TIMEOUT    LONG_MAX
 184extern signed long FASTCALL(schedule_timeout(signed long timeout));
 185asmlinkage void schedule(void);
 186
 187struct namespace;
 188
 189/* Maximum number of active map areas.. This is a random (large) number */
 190#define DEFAULT_MAX_MAP_COUNT   65536
 191
 192extern int sysctl_max_map_count;
 193
 194#include <linux/aio.h>
 195
 196extern unsigned long
 197arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
 198                       unsigned long, unsigned long);
 199extern unsigned long
 200arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
 201                          unsigned long len, unsigned long pgoff,
 202                          unsigned long flags);
 203extern void arch_unmap_area(struct vm_area_struct *area);
 204extern void arch_unmap_area_topdown(struct vm_area_struct *area);
 205
 206
 207struct mm_struct {
 208        struct vm_area_struct * mmap;           /* list of VMAs */
 209        struct rb_root mm_rb;
 210        struct vm_area_struct * mmap_cache;     /* last find_vma result */
 211        unsigned long (*get_unmapped_area) (struct file *filp,
 212                                unsigned long addr, unsigned long len,
 213                                unsigned long pgoff, unsigned long flags);
 214        void (*unmap_area) (struct vm_area_struct *area);
 215        unsigned long mmap_base;                /* base of mmap area */
 216        unsigned long free_area_cache;          /* first hole */
 217        pgd_t * pgd;
 218        atomic_t mm_users;                      /* How many users with user space? */
 219        atomic_t mm_count;                      /* How many references to "struct mm_struct" (users count as 1) */
 220        int map_count;                          /* number of VMAs */
 221        struct rw_semaphore mmap_sem;
 222        spinlock_t page_table_lock;             /* Protects page tables, mm->rss, mm->anon_rss */
 223
 224        struct list_head mmlist;                /* List of maybe swapped mm's.  These are globally strung
 225                                                 * together off init_mm.mmlist, and are protected
 226                                                 * by mmlist_lock
 227                                                 */
 228
 229        unsigned long start_code, end_code, start_data, end_data;
 230        unsigned long start_brk, brk, start_stack;
 231        unsigned long arg_start, arg_end, env_start, env_end;
 232        unsigned long rss, anon_rss, total_vm, locked_vm, shared_vm;
 233        unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
 234
 235        unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
 236
 237        unsigned dumpable:1;
 238        cpumask_t cpu_vm_mask;
 239
 240        /* Architecture-specific MM context */
 241        mm_context_t context;
 242
 243        /* Token based thrashing protection. */
 244        unsigned long swap_token_time;
 245        char recent_pagein;
 246
 247        /* coredumping support */
 248        int core_waiters;
 249        struct completion *core_startup_done, core_done;
 250
 251        /* aio bits */
 252        rwlock_t                ioctx_list_lock;
 253        struct kioctx           *ioctx_list;
 254
 255        struct kioctx           default_kioctx;
 256
 257        unsigned long hiwater_rss;      /* High-water RSS usage */
 258        unsigned long hiwater_vm;       /* High-water virtual memory usage */
 259};
 260
 261struct sighand_struct {
 262        atomic_t                count;
 263        struct k_sigaction      action[_NSIG];
 264        spinlock_t              siglock;
 265};
 266
 267/*
 268 * NOTE! "signal_struct" does not have it's own
 269 * locking, because a shared signal_struct always
 270 * implies a shared sighand_struct, so locking
 271 * sighand_struct is always a proper superset of
 272 * the locking of signal_struct.
 273 */
 274struct signal_struct {
 275        atomic_t                count;
 276        atomic_t                live;
 277
 278        wait_queue_head_t       wait_chldexit;  /* for wait4() */
 279
 280        /* current thread group signal load-balancing target: */
 281        task_t                  *curr_target;
 282
 283        /* shared signal handling: */
 284        struct sigpending       shared_pending;
 285
 286        /* thread group exit support */
 287        int                     group_exit_code;
 288        /* overloaded:
 289         * - notify group_exit_task when ->count is equal to notify_count
 290         * - everyone except group_exit_task is stopped during signal delivery
 291         *   of fatal signals, group_exit_task processes the signal.
 292         */
 293        struct task_struct      *group_exit_task;
 294        int                     notify_count;
 295
 296        /* thread group stop support, overloads group_exit_code too */
 297        int                     group_stop_count;
 298        unsigned int            flags; /* see SIGNAL_* flags below */
 299
 300        /* POSIX.1b Interval Timers */
 301        struct list_head posix_timers;
 302
 303        /* job control IDs */
 304        pid_t pgrp;
 305        pid_t tty_old_pgrp;
 306        pid_t session;
 307        /* boolean value for session group leader */
 308        int leader;
 309
 310        struct tty_struct *tty; /* NULL if no tty */
 311
 312        /*
 313         * Cumulative resource counters for dead threads in the group,
 314         * and for reaped dead child processes forked by this group.
 315         * Live threads maintain their own counters and add to these
 316         * in __exit_signal, except for the group leader.
 317         */
 318        cputime_t utime, stime, cutime, cstime;
 319        unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
 320        unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
 321
 322        /*
 323         * We don't bother to synchronize most readers of this at all,
 324         * because there is no reader checking a limit that actually needs
 325         * to get both rlim_cur and rlim_max atomically, and either one
 326         * alone is a single word that can safely be read normally.
 327         * getrlimit/setrlimit use task_lock(current->group_leader) to
 328         * protect this instead of the siglock, because they really
 329         * have no need to disable irqs.
 330         */
 331        struct rlimit rlim[RLIM_NLIMITS];
 332};
 333
 334/*
 335 * Bits in flags field of signal_struct.
 336 */
 337#define SIGNAL_STOP_STOPPED     0x00000001 /* job control stop in effect */
 338#define SIGNAL_STOP_DEQUEUED    0x00000002 /* stop signal dequeued */
 339#define SIGNAL_STOP_CONTINUED   0x00000004 /* SIGCONT since WCONTINUED reap */
 340#define SIGNAL_GROUP_EXIT       0x00000008 /* group exit in progress */
 341
 342
 343/*
 344 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
 345 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
 346 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
 347 * are inverted: lower p->prio value means higher priority.
 348 *
 349 * The MAX_USER_RT_PRIO value allows the actual maximum
 350 * RT priority to be separate from the value exported to
 351 * user-space.  This allows kernel threads to set their
 352 * priority to a value higher than any user task. Note:
 353 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
 354 */
 355
 356#define MAX_USER_RT_PRIO        100
 357#define MAX_RT_PRIO             MAX_USER_RT_PRIO
 358
 359#define MAX_PRIO                (MAX_RT_PRIO + 40)
 360
 361#define rt_task(p)              (unlikely((p)->prio < MAX_RT_PRIO))
 362
 363/*
 364 * Some day this will be a full-fledged user tracking system..
 365 */
 366struct user_struct {
 367        atomic_t __count;       /* reference count */
 368        atomic_t processes;     /* How many processes does this user have? */
 369        atomic_t files;         /* How many open files does this user have? */
 370        atomic_t sigpending;    /* How many pending signals does this user have? */
 371        /* protected by mq_lock */
 372        unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
 373        unsigned long locked_shm; /* How many pages of mlocked shm ? */
 374
 375#ifdef CONFIG_KEYS
 376        struct key *uid_keyring;        /* UID specific keyring */
 377        struct key *session_keyring;    /* UID's default session keyring */
 378#endif
 379
 380        /* Hash table maintenance information */
 381        struct list_head uidhash_list;
 382        uid_t uid;
 383};
 384
 385extern struct user_struct *find_user(uid_t);
 386
 387extern struct user_struct root_user;
 388#define INIT_USER (&root_user)
 389
 390typedef struct prio_array prio_array_t;
 391struct backing_dev_info;
 392struct reclaim_state;
 393
 394#ifdef CONFIG_SCHEDSTATS
 395struct sched_info {
 396        /* cumulative counters */
 397        unsigned long   cpu_time,       /* time spent on the cpu */
 398                        run_delay,      /* time spent waiting on a runqueue */
 399                        pcnt;           /* # of timeslices run on this cpu */
 400
 401        /* timestamps */
 402        unsigned long   last_arrival,   /* when we last ran on a cpu */
 403                        last_queued;    /* when we were last queued to run */
 404};
 405
 406extern struct file_operations proc_schedstat_operations;
 407#endif
 408
 409enum idle_type
 410{
 411        SCHED_IDLE,
 412        NOT_IDLE,
 413        NEWLY_IDLE,
 414        MAX_IDLE_TYPES
 415};
 416
 417/*
 418 * sched-domains (multiprocessor balancing) declarations:
 419 */
 420#ifdef CONFIG_SMP
 421#define SCHED_LOAD_SCALE        128UL   /* increase resolution of load */
 422
 423#define SD_LOAD_BALANCE         1       /* Do load balancing on this domain. */
 424#define SD_BALANCE_NEWIDLE      2       /* Balance when about to become idle */
 425#define SD_BALANCE_EXEC         4       /* Balance on exec */
 426#define SD_WAKE_IDLE            8       /* Wake to idle CPU on task wakeup */
 427#define SD_WAKE_AFFINE          16      /* Wake task to waking CPU */
 428#define SD_WAKE_BALANCE         32      /* Perform balancing at task wakeup */
 429#define SD_SHARE_CPUPOWER       64      /* Domain members share cpu power */
 430
 431struct sched_group {
 432        struct sched_group *next;       /* Must be a circular list */
 433        cpumask_t cpumask;
 434
 435        /*
 436         * CPU power of this group, SCHED_LOAD_SCALE being max power for a
 437         * single CPU. This is read only (except for setup, hotplug CPU).
 438         */
 439        unsigned long cpu_power;
 440};
 441
 442struct sched_domain {
 443        /* These fields must be setup */
 444        struct sched_domain *parent;    /* top domain must be null terminated */
 445        struct sched_group *groups;     /* the balancing groups of the domain */
 446        cpumask_t span;                 /* span of all CPUs in this domain */
 447        unsigned long min_interval;     /* Minimum balance interval ms */
 448        unsigned long max_interval;     /* Maximum balance interval ms */
 449        unsigned int busy_factor;       /* less balancing by factor if busy */
 450        unsigned int imbalance_pct;     /* No balance until over watermark */
 451        unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
 452        unsigned int cache_nice_tries;  /* Leave cache hot tasks for # tries */
 453        unsigned int per_cpu_gain;      /* CPU % gained by adding domain cpus */
 454        int flags;                      /* See SD_* */
 455
 456        /* Runtime fields. */
 457        unsigned long last_balance;     /* init to jiffies. units in jiffies */
 458        unsigned int balance_interval;  /* initialise to 1. units in ms. */
 459        unsigned int nr_balance_failed; /* initialise to 0 */
 460
 461#ifdef CONFIG_SCHEDSTATS
 462        /* load_balance() stats */
 463        unsigned long lb_cnt[MAX_IDLE_TYPES];
 464        unsigned long lb_failed[MAX_IDLE_TYPES];
 465        unsigned long lb_imbalance[MAX_IDLE_TYPES];
 466        unsigned long lb_nobusyg[MAX_IDLE_TYPES];
 467        unsigned long lb_nobusyq[MAX_IDLE_TYPES];
 468
 469        /* sched_balance_exec() stats */
 470        unsigned long sbe_attempts;
 471        unsigned long sbe_pushed;
 472
 473        /* try_to_wake_up() stats */
 474        unsigned long ttwu_wake_affine;
 475        unsigned long ttwu_wake_balance;
 476#endif
 477};
 478
 479#ifdef ARCH_HAS_SCHED_DOMAIN
 480/* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
 481extern cpumask_t cpu_isolated_map;
 482extern void init_sched_build_groups(struct sched_group groups[],
 483                                cpumask_t span, int (*group_fn)(int cpu));
 484extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
 485#endif /* ARCH_HAS_SCHED_DOMAIN */
 486#endif /* CONFIG_SMP */
 487
 488
 489struct io_context;                      /* See blkdev.h */
 490void exit_io_context(void);
 491
 492#define NGROUPS_SMALL           32
 493#define NGROUPS_PER_BLOCK       ((int)(PAGE_SIZE / sizeof(gid_t)))
 494struct group_info {
 495        int ngroups;
 496        atomic_t usage;
 497        gid_t small_block[NGROUPS_SMALL];
 498        int nblocks;
 499        gid_t *blocks[0];
 500};
 501
 502/*
 503 * get_group_info() must be called with the owning task locked (via task_lock())
 504 * when task != current.  The reason being that the vast majority of callers are
 505 * looking at current->group_info, which can not be changed except by the
 506 * current task.  Changing current->group_info requires the task lock, too.
 507 */
 508#define get_group_info(group_info) do { \
 509        atomic_inc(&(group_info)->usage); \
 510} while (0)
 511
 512#define put_group_info(group_info) do { \
 513        if (atomic_dec_and_test(&(group_info)->usage)) \
 514                groups_free(group_info); \
 515} while (0)
 516
 517struct group_info *groups_alloc(int gidsetsize);
 518void groups_free(struct group_info *group_info);
 519int set_current_groups(struct group_info *group_info);
 520/* access the groups "array" with this macro */
 521#define GROUP_AT(gi, i) \
 522    ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
 523
 524
 525struct audit_context;           /* See audit.c */
 526struct mempolicy;
 527
 528struct task_struct {
 529        volatile long state;    /* -1 unrunnable, 0 runnable, >0 stopped */
 530        struct thread_info *thread_info;
 531        atomic_t usage;
 532        unsigned long flags;    /* per process flags, defined below */
 533        unsigned long ptrace;
 534
 535        int lock_depth;         /* Lock depth */
 536
 537        int prio, static_prio;
 538        struct list_head run_list;
 539        prio_array_t *array;
 540
 541        unsigned long sleep_avg;
 542        unsigned long long timestamp, last_ran;
 543        int activated;
 544
 545        unsigned long policy;
 546        cpumask_t cpus_allowed;
 547        unsigned int time_slice, first_time_slice;
 548
 549#ifdef CONFIG_SCHEDSTATS
 550        struct sched_info sched_info;
 551#endif
 552
 553        struct list_head tasks;
 554        /*
 555         * ptrace_list/ptrace_children forms the list of my children
 556         * that were stolen by a ptracer.
 557         */
 558        struct list_head ptrace_children;
 559        struct list_head ptrace_list;
 560
 561        struct mm_struct *mm, *active_mm;
 562
 563/* task state */
 564        struct linux_binfmt *binfmt;
 565        long exit_state;
 566        int exit_code, exit_signal;
 567        int pdeath_signal;  /*  The signal sent when the parent dies  */
 568        /* ??? */
 569        unsigned long personality;
 570        unsigned did_exec:1;
 571        pid_t pid;
 572        pid_t tgid;
 573        /* 
 574         * pointers to (original) parent process, youngest child, younger sibling,
 575         * older sibling, respectively.  (p->father can be replaced with 
 576         * p->parent->pid)
 577         */
 578        struct task_struct *real_parent; /* real parent process (when being debugged) */
 579        struct task_struct *parent;     /* parent process */
 580        /*
 581         * children/sibling forms the list of my children plus the
 582         * tasks I'm ptracing.
 583         */
 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 */
 587
 588        /* PID/PID hash table linkage. */
 589        struct pid pids[PIDTYPE_MAX];
 590
 591        struct completion *vfork_done;          /* for vfork() */
 592        int __user *set_child_tid;              /* CLONE_CHILD_SETTID */
 593        int __user *clear_child_tid;            /* CLONE_CHILD_CLEARTID */
 594
 595        unsigned long rt_priority;
 596        unsigned long it_real_value, it_real_incr;
 597        cputime_t it_virt_value, it_virt_incr;
 598        cputime_t it_prof_value, it_prof_incr;
 599        struct timer_list real_timer;
 600        cputime_t 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;
 612#ifdef CONFIG_KEYS
 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 */
 616#endif
 617        int oomkilladj; /* OOM kill score adjustment (bit shift). */
 618        char comm[TASK_COMM_LEN];
 619/* file system info */
 620        int link_count, total_link_count;
 621/* ipc stuff */
 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;
 629/* namespace */
 630        struct namespace *namespace;
 631/* signal handlers */
 632        struct signal_struct *signal;
 633        struct sighand_struct *sighand;
 634
 635        sigset_t blocked, real_blocked;
 636        struct sigpending pending;
 637
 638        unsigned long sas_ss_sp;
 639        size_t sas_ss_size;
 640        int (*notifier)(void *priv);
 641        void *notifier_data;
 642        sigset_t *notifier_mask;
 643        
 644        void *security;
 645        struct audit_context *audit_context;
 646
 647/* Thread group tracking */
 648        u32 parent_exec_id;
 649        u32 self_exec_id;
 650/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
 651        spinlock_t alloc_lock;
 652/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
 653        spinlock_t proc_lock;
 654/* context-switch lock */
 655        spinlock_t switch_lock;
 656
 657/* journalling filesystem info */
 658        void *journal_info;
 659
 660/* VM state */
 661        struct reclaim_state *reclaim_state;
 662
 663        struct dentry *proc_dentry;
 664        struct backing_dev_info *backing_dev_info;
 665
 666        struct io_context *io_context;
 667
 668        unsigned long ptrace_message;
 669        siginfo_t *last_siginfo; /* For ptrace use.  */
 670/*
 671 * current io wait handle: wait queue entry to use for io waits
 672 * If this thread is processing aio, this points at the waitqueue
 673 * inside the currently handled kiocb. It may be NULL (i.e. default
 674 * to a stack based synchronous wait) if its doing sync IO.
 675 */
 676        wait_queue_t *io_wait;
 677/* i/o counters(bytes read/written, #syscalls */
 678        u64 rchar, wchar, syscr, syscw;
 679#if defined(CONFIG_BSD_PROCESS_ACCT)
 680        u64 acct_rss_mem1;      /* accumulated rss usage */
 681        u64 acct_vm_mem1;       /* accumulated virtual memory usage */
 682        clock_t acct_stimexpd;  /* clock_t-converted stime since last update */
 683#endif
 684#ifdef CONFIG_NUMA
 685        struct mempolicy *mempolicy;
 686        short il_next;
 687#endif
 688};
 689
 690static inline pid_t process_group(struct task_struct *tsk)
 691{
 692        return tsk->signal->pgrp;
 693}
 694
 695/**
 696 * pid_alive - check that a task structure is not stale
 697 * @p: Task structure to be checked.
 698 *
 699 * Test if a process is not yet dead (at most zombie state)
 700 * If pid_alive fails, then pointers within the task structure
 701 * can be stale and must not be dereferenced.
 702 */
 703static inline int pid_alive(struct task_struct *p)
 704{
 705        return p->pids[PIDTYPE_PID].nr != 0;
 706}
 707
 708extern void free_task(struct task_struct *tsk);
 709extern void __put_task_struct(struct task_struct *tsk);
 710#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
 711#define put_task_struct(tsk) \
 712do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
 713
 714/*
 715 * Per process flags
 716 */
 717#define PF_ALIGNWARN    0x00000001      /* Print alignment warning msgs */
 718                                        /* Not implemented yet, only for 486*/
 719#define PF_STARTING     0x00000002      /* being created */
 720#define PF_EXITING      0x00000004      /* getting shut down */
 721#define PF_DEAD         0x00000008      /* Dead */
 722#define PF_FORKNOEXEC   0x00000040      /* forked but didn't exec */
 723#define PF_SUPERPRIV    0x00000100      /* used super-user privileges */
 724#define PF_DUMPCORE     0x00000200      /* dumped core */
 725#define PF_SIGNALED     0x00000400      /* killed by a signal */
 726#define PF_MEMALLOC     0x00000800      /* Allocating memory */
 727#define PF_FLUSHER      0x00001000      /* responsible for disk writeback */
 728#define PF_USED_MATH    0x00002000      /* if unset the fpu must be initialized before use */
 729#define PF_FREEZE       0x00004000      /* this task is being frozen for suspend now */
 730#define PF_NOFREEZE     0x00008000      /* this thread should not be frozen */
 731#define PF_FROZEN       0x00010000      /* frozen for system suspend */
 732#define PF_FSTRANS      0x00020000      /* inside a filesystem transaction */
 733#define PF_KSWAPD       0x00040000      /* I am kswapd */
 734#define PF_SWAPOFF      0x00080000      /* I am in swapoff */
 735#define PF_LESS_THROTTLE 0x00100000     /* Throttle me less: I clean memory */
 736#define PF_SYNCWRITE    0x00200000      /* I am doing a sync write */
 737#define PF_BORROWED_MM  0x00400000      /* I am a kthread doing use_mm */
 738
 739/*
 740 * Only the _current_ task can read/write to tsk->flags, but other
 741 * tasks can access tsk->flags in readonly mode for example
 742 * with tsk_used_math (like during threaded core dumping).
 743 * There is however an exception to this rule during ptrace
 744 * or during fork: the ptracer task is allowed to write to the
 745 * child->flags of its traced child (same goes for fork, the parent
 746 * can write to the child->flags), because we're guaranteed the
 747 * child is not running and in turn not changing child->flags
 748 * at the same time the parent does it.
 749 */
 750#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
 751#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
 752#define clear_used_math() clear_stopped_child_used_math(current)
 753#define set_used_math() set_stopped_child_used_math(current)
 754#define conditional_stopped_child_used_math(condition, child) \
 755        do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
 756#define conditional_used_math(condition) \
 757        conditional_stopped_child_used_math(condition, current)
 758#define copy_to_stopped_child_used_math(child) \
 759        do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
 760/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
 761#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
 762#define used_math() tsk_used_math(current)
 763
 764#ifdef CONFIG_SMP
 765extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
 766#else
 767static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
 768{
 769        if (!cpus_intersects(new_mask, cpu_online_map))
 770                return -EINVAL;
 771        return 0;
 772}
 773#endif
 774
 775extern unsigned long long sched_clock(void);
 776
 777/* sched_exec is called by processes performing an exec */
 778#ifdef CONFIG_SMP
 779extern void sched_exec(void);
 780#else
 781#define sched_exec()   {}
 782#endif
 783
 784#ifdef CONFIG_HOTPLUG_CPU
 785extern void idle_task_exit(void);
 786#else
 787static inline void idle_task_exit(void) {}
 788#endif
 789
 790extern void sched_idle_next(void);
 791extern void set_user_nice(task_t *p, long nice);
 792extern int task_prio(const task_t *p);
 793extern int task_nice(const task_t *p);
 794extern int task_curr(const task_t *p);
 795extern int idle_cpu(int cpu);
 796extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
 797extern task_t *idle_task(int cpu);
 798
 799void yield(void);
 800
 801/*
 802 * The default (Linux) execution domain.
 803 */
 804extern struct exec_domain       default_exec_domain;
 805
 806union thread_union {
 807        struct thread_info thread_info;
 808        unsigned long stack[THREAD_SIZE/sizeof(long)];
 809};
 810
 811#ifndef __HAVE_ARCH_KSTACK_END
 812static inline int kstack_end(void *addr)
 813{
 814        /* Reliable end of stack detection:
 815         * Some APM bios versions misalign the stack
 816         */
 817        return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
 818}
 819#endif
 820
 821extern union thread_union init_thread_union;
 822extern struct task_struct init_task;
 823
 824extern struct   mm_struct init_mm;
 825
 826#define find_task_by_pid(nr)    find_task_by_pid_type(PIDTYPE_PID, nr)
 827extern struct task_struct *find_task_by_pid_type(int type, int pid);
 828extern void set_special_pids(pid_t session, pid_t pgrp);
 829extern void __set_special_pids(pid_t session, pid_t pgrp);
 830
 831/* per-UID process charging. */
 832extern struct user_struct * alloc_uid(uid_t);
 833static inline struct user_struct *get_uid(struct user_struct *u)
 834{
 835        atomic_inc(&u->__count);
 836        return u;
 837}
 838extern void free_uid(struct user_struct *);
 839extern void switch_uid(struct user_struct *);
 840
 841#include <asm/current.h>
 842
 843extern void do_timer(struct pt_regs *);
 844
 845extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
 846extern int FASTCALL(wake_up_process(struct task_struct * tsk));
 847extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
 848                                                unsigned long clone_flags));
 849#ifdef CONFIG_SMP
 850 extern void kick_process(struct task_struct *tsk);
 851#else
 852 static inline void kick_process(struct task_struct *tsk) { }
 853#endif
 854extern void FASTCALL(sched_fork(task_t * p));
 855extern void FASTCALL(sched_exit(task_t * p));
 856
 857extern int in_group_p(gid_t);
 858extern int in_egroup_p(gid_t);
 859
 860extern void proc_caches_init(void);
 861extern void flush_signals(struct task_struct *);
 862extern void flush_signal_handlers(struct task_struct *, int force_default);
 863extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
 864
 865static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
 866{
 867        unsigned long flags;
 868        int ret;
 869
 870        spin_lock_irqsave(&tsk->sighand->siglock, flags);
 871        ret = dequeue_signal(tsk, mask, info);
 872        spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
 873
 874        return ret;
 875}       
 876
 877extern void block_all_signals(int (*notifier)(void *priv), void *priv,
 878                              sigset_t *mask);
 879extern void unblock_all_signals(void);
 880extern void release_task(struct task_struct * p);
 881extern int send_sig_info(int, struct siginfo *, struct task_struct *);
 882extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
 883extern int force_sigsegv(int, struct task_struct *);
 884extern int force_sig_info(int, struct siginfo *, struct task_struct *);
 885extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
 886extern int kill_pg_info(int, struct siginfo *, pid_t);
 887extern int kill_proc_info(int, struct siginfo *, pid_t);
 888extern void do_notify_parent(struct task_struct *, int);
 889extern void force_sig(int, struct task_struct *);
 890extern void force_sig_specific(int, struct task_struct *);
 891extern int send_sig(int, struct task_struct *, int);
 892extern void zap_other_threads(struct task_struct *p);
 893extern int kill_pg(pid_t, int, int);
 894extern int kill_sl(pid_t, int, int);
 895extern int kill_proc(pid_t, int, int);
 896extern struct sigqueue *sigqueue_alloc(void);
 897extern void sigqueue_free(struct sigqueue *);
 898extern int send_sigqueue(int, struct sigqueue *,  struct task_struct *);
 899extern int send_group_sigqueue(int, struct sigqueue *,  struct task_struct *);
 900extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
 901extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
 902
 903/* These can be the second arg to send_sig_info/send_group_sig_info.  */
 904#define SEND_SIG_NOINFO ((struct siginfo *) 0)
 905#define SEND_SIG_PRIV   ((struct siginfo *) 1)
 906#define SEND_SIG_FORCED ((struct siginfo *) 2)
 907
 908/* True if we are on the alternate signal stack.  */
 909
 910static inline int on_sig_stack(unsigned long sp)
 911{
 912        return (sp - current->sas_ss_sp < current->sas_ss_size);
 913}
 914
 915static inline int sas_ss_flags(unsigned long sp)
 916{
 917        return (current->sas_ss_size == 0 ? SS_DISABLE
 918                : on_sig_stack(sp) ? SS_ONSTACK : 0);
 919}
 920
 921
 922#ifdef CONFIG_SECURITY
 923/* code is in security.c */
 924extern int capable(int cap);
 925#else
 926static inline int capable(int cap)
 927{
 928        if (cap_raised(current->cap_effective, cap)) {
 929                current->flags |= PF_SUPERPRIV;
 930                return 1;
 931        }
 932        return 0;
 933}
 934#endif
 935
 936/*
 937 * Routines for handling mm_structs
 938 */
 939extern struct mm_struct * mm_alloc(void);
 940
 941/* mmdrop drops the mm and the page tables */
 942extern void FASTCALL(__mmdrop(struct mm_struct *));
 943static inline void mmdrop(struct mm_struct * mm)
 944{
 945        if (atomic_dec_and_test(&mm->mm_count))
 946                __mmdrop(mm);
 947}
 948
 949/* mmput gets rid of the mappings and all user-space */
 950extern void mmput(struct mm_struct *);
 951/* Grab a reference to a task's mm, if it is not already going away */
 952extern struct mm_struct *get_task_mm(struct task_struct *task);
 953/* Remove the current tasks stale references to the old mm_struct */
 954extern void mm_release(struct task_struct *, struct mm_struct *);
 955
 956extern int  copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
 957extern void flush_thread(void);
 958extern void exit_thread(void);
 959
 960extern void exit_mm(struct task_struct *);
 961extern void exit_files(struct task_struct *);
 962extern void exit_signal(struct task_struct *);
 963extern void __exit_signal(struct task_struct *);
 964extern void exit_sighand(struct task_struct *);
 965extern void __exit_sighand(struct task_struct *);
 966extern void exit_itimers(struct signal_struct *);
 967
 968extern NORET_TYPE void do_group_exit(int);
 969
 970extern void reparent_to_init(void);
 971extern void daemonize(const char *, ...);
 972extern int allow_signal(int);
 973extern int disallow_signal(int);
 974extern task_t *child_reaper;
 975
 976extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
 977extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
 978task_t *fork_idle(int);
 979
 980extern void set_task_comm(struct task_struct *tsk, char *from);
 981extern void get_task_comm(char *to, struct task_struct *tsk);
 982
 983#ifdef CONFIG_SMP
 984extern void wait_task_inactive(task_t * p);
 985#else
 986#define wait_task_inactive(p)   do { } while (0)
 987#endif
 988
 989#define remove_parent(p)        list_del_init(&(p)->sibling)
 990#define add_parent(p, parent)   list_add_tail(&(p)->sibling,&(parent)->children)
 991
 992#define REMOVE_LINKS(p) do {                                    \
 993        if (thread_group_leader(p))                             \
 994                list_del_init(&(p)->tasks);                     \
 995        remove_parent(p);                                       \
 996        } while (0)
 997
 998#define SET_LINKS(p) do {                                       \
 999        if (thread_group_leader(p))                             \
1000                list_add_tail(&(p)->tasks,&init_task.tasks);    \
1001        add_parent(p, (p)->parent);                             \
1002        } while (0)
1003
1004#define next_task(p)    list_entry((p)->tasks.next, struct task_struct, tasks)
1005#define prev_task(p)    list_entry((p)->tasks.prev, struct task_struct, tasks)
1006
1007#define for_each_process(p) \
1008        for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1009
1010/*
1011 * Careful: do_each_thread/while_each_thread is a double loop so
1012 *          'break' will not work as expected - use goto instead.
1013 */
1014#define do_each_thread(g, t) \
1015        for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1016
1017#define while_each_thread(g, t) \
1018        while ((t = next_thread(t)) != g)
1019
1020extern task_t * FASTCALL(next_thread(const task_t *p));
1021
1022#define thread_group_leader(p)  (p->pid == p->tgid)
1023
1024static inline int thread_group_empty(task_t *p)
1025{
1026        return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1027}
1028
1029#define delay_group_leader(p) \
1030                (thread_group_leader(p) && !thread_group_empty(p))
1031
1032extern void unhash_process(struct task_struct *p);
1033
1034/*
1035 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1036 * subscriptions and synchronises with wait4().  Also used in procfs.
1037 *
1038 * Nests both inside and outside of read_lock(&tasklist_lock).
1039 * It must not be nested with write_lock_irq(&tasklist_lock),
1040 * neither inside nor outside.
1041 */
1042static inline void task_lock(struct task_struct *p)
1043{
1044        spin_lock(&p->alloc_lock);
1045}
1046
1047static inline void task_unlock(struct task_struct *p)
1048{
1049        spin_unlock(&p->alloc_lock);
1050}
1051
1052/* set thread flags in other task's structures
1053 * - see asm/thread_info.h for TIF_xxxx flags available
1054 */
1055static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1056{
1057        set_ti_thread_flag(tsk->thread_info,flag);
1058}
1059
1060static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1061{
1062        clear_ti_thread_flag(tsk->thread_info,flag);
1063}
1064
1065static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1066{
1067        return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1068}
1069
1070static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1071{
1072        return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1073}
1074
1075static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1076{
1077        return test_ti_thread_flag(tsk->thread_info,flag);
1078}
1079
1080static inline void set_tsk_need_resched(struct task_struct *tsk)
1081{
1082        set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1083}
1084
1085static inline void clear_tsk_need_resched(struct task_struct *tsk)
1086{
1087        clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1088}
1089
1090static inline int signal_pending(struct task_struct *p)
1091{
1092        return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1093}
1094  
1095static inline int need_resched(void)
1096{
1097        return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1098}
1099
1100/*
1101 * cond_resched() and cond_resched_lock(): latency reduction via
1102 * explicit rescheduling in places that are safe. The return
1103 * value indicates whether a reschedule was done in fact.
1104 * cond_resched_lock() will drop the spinlock before scheduling,
1105 * cond_resched_softirq() will enable bhs before scheduling.
1106 */
1107extern int cond_resched(void);
1108extern int cond_resched_lock(spinlock_t * lock);
1109extern int cond_resched_softirq(void);
1110
1111/*
1112 * Does a critical section need to be broken due to another
1113 * task waiting?:
1114 */
1115#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1116# define need_lockbreak(lock) ((lock)->break_lock)
1117#else
1118# define need_lockbreak(lock) 0
1119#endif
1120
1121/*
1122 * Does a critical section need to be broken due to another
1123 * task waiting or preemption being signalled:
1124 */
1125static inline int lock_need_resched(spinlock_t *lock)
1126{
1127        if (need_lockbreak(lock) || need_resched())
1128                return 1;
1129        return 0;
1130}
1131
1132/* Reevaluate whether the task has signals pending delivery.
1133   This is required every time the blocked sigset_t changes.
1134   callers must hold sighand->siglock.  */
1135
1136extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1137extern void recalc_sigpending(void);
1138
1139extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1140
1141/*
1142 * Wrappers for p->thread_info->cpu access. No-op on UP.
1143 */
1144#ifdef CONFIG_SMP
1145
1146static inline unsigned int task_cpu(const struct task_struct *p)
1147{
1148        return p->thread_info->cpu;
1149}
1150
1151static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1152{
1153        p->thread_info->cpu = cpu;
1154}
1155
1156#else
1157
1158static inline unsigned int task_cpu(const struct task_struct *p)
1159{
1160        return 0;
1161}
1162
1163static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1164{
1165}
1166
1167#endif /* CONFIG_SMP */
1168
1169#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1170extern void arch_pick_mmap_layout(struct mm_struct *mm);
1171#else
1172static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1173{
1174        mm->mmap_base = TASK_UNMAPPED_BASE;
1175        mm->get_unmapped_area = arch_get_unmapped_area;
1176        mm->unmap_area = arch_unmap_area;
1177}
1178#endif
1179
1180extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1181extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1182
1183#ifdef CONFIG_MAGIC_SYSRQ
1184
1185extern void normalize_rt_tasks(void);
1186
1187#endif
1188
1189/* try_to_freeze
1190 *
1191 * Checks whether we need to enter the refrigerator
1192 * and returns 1 if we did so.
1193 */
1194#ifdef CONFIG_PM
1195extern void refrigerator(unsigned long);
1196extern int freeze_processes(void);
1197extern void thaw_processes(void);
1198
1199static inline int try_to_freeze(unsigned long refrigerator_flags)
1200{
1201        if (unlikely(current->flags & PF_FREEZE)) {
1202                refrigerator(refrigerator_flags);
1203                return 1;
1204        } else
1205                return 0;
1206}
1207#else
1208static inline void refrigerator(unsigned long flag) {}
1209static inline int freeze_processes(void) { BUG(); return 0; }
1210static inline void thaw_processes(void) {}
1211
1212static inline int try_to_freeze(unsigned long refrigerator_flags)
1213{
1214        return 0;
1215}
1216#endif /* CONFIG_PM */
1217#endif /* __KERNEL__ */
1218
1219#endif
1220
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.