include/linux/sched.h

   1 #ifndef _LINUX_SCHED_H
   2 #define _LINUX_SCHED_H
   3
   4 /*
   5  * cloning flags:
   6  */
   7 #define CSIGNAL         0x000000ff      /* signal mask to be sent at exit */
   8 #define CLONE_VM        0x00000100      /* set if VM shared between processes */
   9 #define CLONE_FS        0x00000200      /* set if fs info shared between processes */
  10 #define CLONE_FILES     0x00000400      /* set if open files shared between processes */
  11 #define CLONE_SIGHAND   0x00000800      /* set if signal handlers and blocked signals shared */
  12 #define CLONE_PTRACE    0x00002000      /* set if we want to let tracing continue on the child too */
  13 #define CLONE_VFORK     0x00004000      /* set if the parent wants the child to wake it up on mm_release */
  14 #define CLONE_PARENT    0x00008000      /* set if we want to have the same parent as the cloner */
  15 #define CLONE_THREAD    0x00010000      /* Same thread group? */
  16 #define CLONE_NEWNS     0x00020000      /* New namespace group? */
  17 #define CLONE_SYSVSEM   0x00040000      /* share system V SEM_UNDO semantics */
  18 #define CLONE_SETTLS    0x00080000      /* create a new TLS for the child */
  19 #define CLONE_PARENT_SETTID     0x00100000      /* set the TID in the parent */
  20 #define CLONE_CHILD_CLEARTID    0x00200000      /* clear the TID in the child */
  21 #define CLONE_DETACHED          0x00400000      /* Unused, ignored */
  22 #define CLONE_UNTRACED          0x00800000      /* set if the tracing process can't force CLONE_PTRACE on this clone */
  23 #define CLONE_CHILD_SETTID      0x01000000      /* set the TID in the child */
  24 #define CLONE_STOPPED           0x02000000      /* Start in stopped state */
  25 #define CLONE_NEWUTS            0x04000000      /* New utsname group? */
  26 #define CLONE_NEWIPC            0x08000000      /* New ipcs */
  27 #define CLONE_NEWUSER           0x10000000      /* New user namespace */
  28 #define CLONE_NEWPID            0x20000000      /* New pid namespace */
  29 #define CLONE_NEWNET            0x40000000      /* New network namespace */
  30 #define CLONE_IO                0x80000000      /* Clone io context */
  31
  32 /*
  33  * Scheduling policies
  34  */
  35 #define SCHED_NORMAL            0
  36 #define SCHED_FIFO              1
  37 #define SCHED_RR                2
  38 #define SCHED_BATCH             3
  39 /* SCHED_ISO: reserved but not implemented yet */
  40 #define SCHED_IDLE              5
  41
  42 #ifdef __KERNEL__
  43
  44 struct sched_param {
  45         int sched_priority;
  46 };
  47
  48 #include <asm/param.h>  /* for HZ */
  49
  50 #include <linux/capability.h>
  51 #include <linux/threads.h>
  52 #include <linux/kernel.h>
  53 #include <linux/types.h>
  54 #include <linux/timex.h>
  55 #include <linux/jiffies.h>
  56 #include <linux/rbtree.h>
  57 #include <linux/thread_info.h>
  58 #include <linux/cpumask.h>
  59 #include <linux/errno.h>
  60 #include <linux/nodemask.h>
  61 #include <linux/mm_types.h>
  62
  63 #include <asm/system.h>
  64 #include <asm/page.h>
  65 #include <asm/ptrace.h>
  66 #include <asm/cputime.h>
  67
  68 #include <linux/smp.h>
  69 #include <linux/sem.h>
  70 #include <linux/signal.h>
  71 #include <linux/fs_struct.h>
  72 #include <linux/compiler.h>
  73 #include <linux/completion.h>
  74 #include <linux/pid.h>
  75 #include <linux/percpu.h>
  76 #include <linux/topology.h>
  77 #include <linux/proportions.h>
  78 #include <linux/seccomp.h>
  79 #include <linux/rcupdate.h>
  80 #include <linux/rtmutex.h>
  81
  82 #include <linux/time.h>
  83 #include <linux/param.h>
  84 #include <linux/resource.h>
  85 #include <linux/timer.h>
  86 #include <linux/hrtimer.h>
  87 #include <linux/task_io_accounting.h>
  88 #include <linux/kobject.h>
  89 #include <linux/latencytop.h>
  90
  91 #include <asm/processor.h>
  92
  93 struct mem_cgroup;
  94 struct exec_domain;
  95 struct futex_pi_state;
  96 struct robust_list_head;
  97 struct bio;
  98
  99 /*
 100  * List of flags we want to share for kernel threads,
 101  * if only because they are not used by them anyway.
 102  */
 103 #define CLONE_KERNEL    (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
 104
 105 /*
 106  * These are the constant used to fake the fixed-point load-average
 107  * counting. Some notes:
 108  *  - 11 bit fractions expand to 22 bits by the multiplies: this gives
 109  *    a load-average precision of 10 bits integer + 11 bits fractional
 110  *  - if you want to count load-averages more often, you need more
 111  *    precision, or rounding will get you. With 2-second counting freq,
 112  *    the EXP_n values would be 1981, 2034 and 2043 if still using only
 113  *    11 bit fractions.
 114  */
 115 extern unsigned long avenrun[];         /* Load averages */
 116
 117 #define FSHIFT          11              /* nr of bits of precision */
 118 #define FIXED_1         (1<<FSHIFT)     /* 1.0 as fixed-point */
 119 #define LOAD_FREQ       (5*HZ+1)        /* 5 sec intervals */
 120 #define EXP_1           1884            /* 1/exp(5sec/1min) as fixed-point */
 121 #define EXP_5           2014            /* 1/exp(5sec/5min) */
 122 #define EXP_15          2037            /* 1/exp(5sec/15min) */
 123
 124 #define CALC_LOAD(load,exp,n) \
 125         load *= exp; \
 126         load += n*(FIXED_1-exp); \
 127         load >>= FSHIFT;
 128
 129 extern unsigned long total_forks;
 130 extern int nr_threads;
 131 DECLARE_PER_CPU(unsigned long, process_counts);
 132 extern int nr_processes(void);
 133 extern unsigned long nr_running(void);
 134 extern unsigned long nr_uninterruptible(void);
 135 extern unsigned long nr_active(void);
 136 extern unsigned long nr_iowait(void);
 137 extern unsigned long weighted_cpuload(const int cpu);
 138
 139 struct seq_file;
 140 struct cfs_rq;
 141 struct task_group;
 142 #ifdef CONFIG_SCHED_DEBUG
 143 extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
 144 extern void proc_sched_set_task(struct task_struct *p);
 145 extern void
 146 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
 147 #else
 148 static inline void
 149 proc_sched_show_task(struct task_struct *p, struct seq_file *m)
 150 {
 151 }
 152 static inline void proc_sched_set_task(struct task_struct *p)
 153 {
 154 }
 155 static inline void
 156 print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
 157 {
 158 }
 159 #endif
 160
 161 /*
 162  * Task state bitmask. NOTE! These bits are also
 163  * encoded in fs/proc/array.c: get_task_state().
 164  *
 165  * We have two separate sets of flags: task->state
 166  * is about runnability, while task->exit_state are
 167  * about the task exiting. Confusing, but this way
 168  * modifying one set can't modify the other one by
 169  * mistake.
 170  */
 171 #define TASK_RUNNING            0
 172 #define TASK_INTERRUPTIBLE      1
 173 #define TASK_UNINTERRUPTIBLE    2
 174 #define __TASK_STOPPED          4
 175 #define __TASK_TRACED           8
 176 /* in tsk->exit_state */
 177 #define EXIT_ZOMBIE             16
 178 #define EXIT_DEAD               32
 179 /* in tsk->state again */
 180 #define TASK_DEAD               64
 181 #define TASK_WAKEKILL           128
 182
 183 /* Convenience macros for the sake of set_task_state */
 184 #define TASK_KILLABLE           (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
 185 #define TASK_STOPPED            (TASK_WAKEKILL | __TASK_STOPPED)
 186 #define TASK_TRACED             (TASK_WAKEKILL | __TASK_TRACED)
 187
 188 /* Convenience macros for the sake of wake_up */
 189 #define TASK_NORMAL             (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
 190 #define TASK_ALL                (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
 191
 192 /* get_task_state() */
 193 #define TASK_REPORT             (TASK_RUNNING | TASK_INTERRUPTIBLE | \
 194                                  TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
 195                                  __TASK_TRACED)
 196
 197 #define task_is_traced(task)    ((task->state & __TASK_TRACED) != 0)
 198 #define task_is_stopped(task)   ((task->state & __TASK_STOPPED) != 0)
 199 #define task_is_stopped_or_traced(task) \
 200                         ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
 201 #define task_contributes_to_load(task)  \
 202                                 ((task->state & TASK_UNINTERRUPTIBLE) != 0)
 203
 204 #define __set_task_state(tsk, state_value)              \
 205         do { (tsk)->state = (state_value); } while (0)
 206 #define set_task_state(tsk, state_value)                \
 207         set_mb((tsk)->state, (state_value))
 208
 209 /*
 210  * set_current_state() includes a barrier so that the write of current->state
 211  * is correctly serialised wrt the caller's subsequent test of whether to
 212  * actually sleep:
 213  *
 214  *      set_current_state(TASK_UNINTERRUPTIBLE);
 215  *      if (do_i_need_to_sleep())
 216  *              schedule();
 217  *
 218  * If the caller does not need such serialisation then use __set_current_state()
 219  */
 220 #define __set_current_state(state_value)                        \
 221         do { current->state = (state_value); } while (0)
 222 #define set_current_state(state_value)          \
 223         set_mb(current->state, (state_value))
 224
 225 /* Task command name length */
 226 #define TASK_COMM_LEN 16
 227
 228 #include <linux/spinlock.h>
 229
 230 /*
 231  * This serializes "schedule()" and also protects
 232  * the run-queue from deletions/modifications (but
 233  * _adding_ to the beginning of the run-queue has
 234  * a separate lock).
 235  */
 236 extern rwlock_t tasklist_lock;
 237 extern spinlock_t mmlist_lock;
 238
 239 struct task_struct;
 240
 241 extern void sched_init(void);
 242 extern void sched_init_smp(void);
 243 extern asmlinkage void schedule_tail(struct task_struct *prev);
 244 extern void init_idle(struct task_struct *idle, int cpu);
 245 extern void init_idle_bootup_task(struct task_struct *idle);
 246
 247 extern cpumask_t nohz_cpu_mask;
 248 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
 249 extern int select_nohz_load_balancer(int cpu);
 250 #else
 251 static inline int select_nohz_load_balancer(int cpu)
 252 {
 253         return 0;
 254 }
 255 #endif
 256
 257 extern unsigned long rt_needs_cpu(int cpu);
 258
 259 /*
 260  * Only dump TASK_* tasks. (0 for all tasks)
 261  */
 262 extern void show_state_filter(unsigned long state_filter);
 263
 264 static inline void show_state(void)
 265 {
 266         show_state_filter(0);
 267 }
 268
 269 extern void show_regs(struct pt_regs *);
 270
 271 /*
 272  * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
 273  * task), SP is the stack pointer of the first frame that should be shown in the back
 274  * trace (or NULL if the entire call-chain of the task should be shown).
 275  */
 276 extern void show_stack(struct task_struct *task, unsigned long *sp);
 277
 278 void io_schedule(void);
 279 long io_schedule_timeout(long timeout);
 280
 281 extern void cpu_init (void);
 282 extern void trap_init(void);
 283 extern void account_process_tick(struct task_struct *task, int user);
 284 extern void update_process_times(int user);
 285 extern void scheduler_tick(void);
 286 extern void hrtick_resched(void);
 287
 288 extern void sched_show_task(struct task_struct *p);
 289
 290 #ifdef CONFIG_DETECT_SOFTLOCKUP
 291 extern void softlockup_tick(void);
 292 extern void spawn_softlockup_task(void);
 293 extern void touch_softlockup_watchdog(void);
 294 extern void touch_all_softlockup_watchdogs(void);
 295 extern unsigned long  softlockup_thresh;
 296 extern unsigned long sysctl_hung_task_check_count;
 297 extern unsigned long sysctl_hung_task_timeout_secs;
 298 extern unsigned long sysctl_hung_task_warnings;
 299 #else
 300 static inline void softlockup_tick(void)
 301 {
 302 }
 303 static inline void spawn_softlockup_task(void)
 304 {
 305 }
 306 static inline void touch_softlockup_watchdog(void)
 307 {
 308 }
 309 static inline void touch_all_softlockup_watchdogs(void)
 310 {
 311 }
 312 #endif
 313
 314
 315 /* Attach to any functions which should be ignored in wchan output. */
 316 #define __sched         __attribute__((__section__(".sched.text")))
 317
 318 /* Linker adds these: start and end of __sched functions */
 319 extern char __sched_text_start[], __sched_text_end[];
 320
 321 /* Is this address in the __sched functions? */
 322 extern int in_sched_functions(unsigned long addr);
 323
 324 #define MAX_SCHEDULE_TIMEOUT    LONG_MAX
 325 extern signed long schedule_timeout(signed long timeout);
 326 extern signed long schedule_timeout_interruptible(signed long timeout);
 327 extern signed long schedule_timeout_killable(signed long timeout);
 328 extern signed long schedule_timeout_uninterruptible(signed long timeout);
 329 asmlinkage void schedule(void);
 330
 331 struct nsproxy;
 332 struct user_namespace;
 333
 334 /* Maximum number of active map areas.. This is a random (large) number */
 335 #define DEFAULT_MAX_MAP_COUNT   65536
 336
 337 extern int sysctl_max_map_count;
 338
 339 #include <linux/aio.h>
 340
 341 extern unsigned long
 342 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
 343                        unsigned long, unsigned long);
 344 extern unsigned long
 345 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
 346                           unsigned long len, unsigned long pgoff,
 347                           unsigned long flags);
 348 extern void arch_unmap_area(struct mm_struct *, unsigned long);
 349 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
 350
 351 #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
 352 /*
 353  * The mm counters are not protected by its page_table_lock,
 354  * so must be incremented atomically.
 355  */
 356 #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
 357 #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
 358 #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
 359 #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
 360 #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
 361
 362 #else  /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
 363 /*
 364  * The mm counters are protected by its page_table_lock,
 365  * so can be incremented directly.
 366  */
 367 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
 368 #define get_mm_counter(mm, member) ((mm)->_##member)
 369 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
 370 #define inc_mm_counter(mm, member) (mm)->_##member++
 371 #define dec_mm_counter(mm, member) (mm)->_##member--
 372
 373 #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
 374
 375 #define get_mm_rss(mm)                                  \
 376         (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
 377 #define update_hiwater_rss(mm)  do {                    \
 378         unsigned long _rss = get_mm_rss(mm);            \
 379         if ((mm)->hiwater_rss < _rss)                   \
 380                 (mm)->hiwater_rss = _rss;               \
 381 } while (0)
 382 #define update_hiwater_vm(mm)   do {                    \
 383         if ((mm)->hiwater_vm < (mm)->total_vm)          \
 384                 (mm)->hiwater_vm = (mm)->total_vm;      \
 385 } while (0)
 386
 387 extern void set_dumpable(struct mm_struct *mm, int value);
 388 extern int get_dumpable(struct mm_struct *mm);
 389
 390 /* mm flags */
 391 /* dumpable bits */
 392 #define MMF_DUMPABLE      0  /* core dump is permitted */
 393 #define MMF_DUMP_SECURELY 1  /* core file is readable only by root */
 394 #define MMF_DUMPABLE_BITS 2
 395
 396 /* coredump filter bits */
 397 #define MMF_DUMP_ANON_PRIVATE   2
 398 #define MMF_DUMP_ANON_SHARED    3
 399 #define MMF_DUMP_MAPPED_PRIVATE 4
 400 #define MMF_DUMP_MAPPED_SHARED  5
 401 #define MMF_DUMP_ELF_HEADERS    6
 402 #define MMF_DUMP_FILTER_SHIFT   MMF_DUMPABLE_BITS
 403 #define MMF_DUMP_FILTER_BITS    5
 404 #define MMF_DUMP_FILTER_MASK \
 405         (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
 406 #define MMF_DUMP_FILTER_DEFAULT \
 407         ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED))
 408
 409 struct sighand_struct {
 410         atomic_t                count;
 411         struct k_sigaction      action[_NSIG];
 412         spinlock_t              siglock;
 413         wait_queue_head_t       signalfd_wqh;
 414 };
 415
 416 struct pacct_struct {
 417         int                     ac_flag;
 418         long                    ac_exitcode;
 419         unsigned long           ac_mem;
 420         cputime_t               ac_utime, ac_stime;
 421         unsigned long           ac_minflt, ac_majflt;
 422 };
 423
 424 /*
 425  * NOTE! "signal_struct" does not have it's own
 426  * locking, because a shared signal_struct always
 427  * implies a shared sighand_struct, so locking
 428  * sighand_struct is always a proper superset of
 429  * the locking of signal_struct.
 430  */
 431 struct signal_struct {
 432         atomic_t                count;
 433         atomic_t                live;
 434
 435         wait_queue_head_t       wait_chldexit;  /* for wait4() */
 436
 437         /* current thread group signal load-balancing target: */
 438         struct task_struct      *curr_target;
 439
 440         /* shared signal handling: */
 441         struct sigpending       shared_pending;
 442
 443         /* thread group exit support */
 444         int                     group_exit_code;
 445         /* overloaded:
 446          * - notify group_exit_task when ->count is equal to notify_count
 447          * - everyone except group_exit_task is stopped during signal delivery
 448          *   of fatal signals, group_exit_task processes the signal.
 449          */
 450         struct task_struct      *group_exit_task;
 451         int                     notify_count;
 452
 453         /* thread group stop support, overloads group_exit_code too */
 454         int                     group_stop_count;
 455         unsigned int            flags; /* see SIGNAL_* flags below */
 456
 457         /* POSIX.1b Interval Timers */
 458         struct list_head posix_timers;
 459
 460         /* ITIMER_REAL timer for the process */
 461         struct hrtimer real_timer;
 462         struct pid *leader_pid;
 463         ktime_t it_real_incr;
 464
 465         /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
 466         cputime_t it_prof_expires, it_virt_expires;
 467         cputime_t it_prof_incr, it_virt_incr;
 468
 469         /* job control IDs */
 470
 471         /*
 472          * pgrp and session fields are deprecated.
 473          * use the task_session_Xnr and task_pgrp_Xnr routines below
 474          */
 475
 476         union {
 477                 pid_t pgrp __deprecated;
 478                 pid_t __pgrp;
 479         };
 480
 481         struct pid *tty_old_pgrp;
 482
 483         union {
 484                 pid_t session __deprecated;
 485                 pid_t __session;
 486         };
 487
 488         /* boolean value for session group leader */
 489         int leader;
 490
 491         struct tty_struct *tty; /* NULL if no tty */
 492
 493         /*
 494          * Cumulative resource counters for dead threads in the group,
 495          * and for reaped dead child processes forked by this group.
 496          * Live threads maintain their own counters and add to these
 497          * in __exit_signal, except for the group leader.
 498          */
 499         cputime_t utime, stime, cutime, cstime;
 500         cputime_t gtime;
 501         cputime_t cgtime;
 502         unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
 503         unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
 504         unsigned long inblock, oublock, cinblock, coublock;
 505
 506         /*
 507          * Cumulative ns of scheduled CPU time for dead threads in the
 508          * group, not including a zombie group leader.  (This only differs
 509          * from jiffies_to_ns(utime + stime) if sched_clock uses something
 510          * other than jiffies.)
 511          */
 512         unsigned long long sum_sched_runtime;
 513
 514         /*
 515          * We don't bother to synchronize most readers of this at all,
 516          * because there is no reader checking a limit that actually needs
 517          * to get both rlim_cur and rlim_max atomically, and either one
 518          * alone is a single word that can safely be read normally.
 519          * getrlimit/setrlimit use task_lock(current->group_leader) to
 520          * protect this instead of the siglock, because they really
 521          * have no need to disable irqs.
 522          */
 523         struct rlimit rlim[RLIM_NLIMITS];
 524
 525         struct list_head cpu_timers[3];
 526
 527         /* keep the process-shared keyrings here so that they do the right
 528          * thing in threads created with CLONE_THREAD */
 529 #ifdef CONFIG_KEYS
 530         struct key *session_keyring;    /* keyring inherited over fork */
 531         struct key *process_keyring;    /* keyring private to this process */
 532 #endif
 533 #ifdef CONFIG_BSD_PROCESS_ACCT
 534         struct pacct_struct pacct;      /* per-process accounting information */
 535 #endif
 536 #ifdef CONFIG_TASKSTATS
 537         struct taskstats *stats;
 538 #endif
 539 #ifdef CONFIG_AUDIT
 540         unsigned audit_tty;
 541         struct tty_audit_buf *tty_audit_buf;
 542 #endif
 543 };
 544
 545 /* Context switch must be unlocked if interrupts are to be enabled */
 546 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
 547 # define __ARCH_WANT_UNLOCKED_CTXSW
 548 #endif
 549
 550 /*
 551  * Bits in flags field of signal_struct.
 552  */
 553 #define SIGNAL_STOP_STOPPED     0x00000001 /* job control stop in effect */
 554 #define SIGNAL_STOP_DEQUEUED    0x00000002 /* stop signal dequeued */
 555 #define SIGNAL_STOP_CONTINUED   0x00000004 /* SIGCONT since WCONTINUED reap */
 556 #define SIGNAL_GROUP_EXIT       0x00000008 /* group exit in progress */
 557
 558 /* If true, all threads except ->group_exit_task have pending SIGKILL */
 559 static inline int signal_group_exit(const struct signal_struct *sig)
 560 {
 561         return  (sig->flags & SIGNAL_GROUP_EXIT) ||
 562                 (sig->group_exit_task != NULL);
 563 }
 564
 565 /*
 566  * Some day this will be a full-fledged user tracking system..
 567  */
 568 struct user_struct {
 569         atomic_t __count;       /* reference count */
 570         atomic_t processes;     /* How many processes does this user have? */
 571         atomic_t files;         /* How many open files does this user have? */
 572         atomic_t sigpending;    /* How many pending signals does this user have? */
 573 #ifdef CONFIG_INOTIFY_USER
 574         atomic_t inotify_watches; /* How many inotify watches does this user have? */
 575         atomic_t inotify_devs;  /* How many inotify devs does this user have opened? */
 576 #endif
 577 #ifdef CONFIG_POSIX_MQUEUE
 578         /* protected by mq_lock */
 579         unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
 580 #endif
 581         unsigned long locked_shm; /* How many pages of mlocked shm ? */
 582
 583 #ifdef CONFIG_KEYS
 584         struct key *uid_keyring;        /* UID specific keyring */
 585         struct key *session_keyring;    /* UID's default session keyring */
 586 #endif
 587
 588         /* Hash table maintenance information */
 589         struct hlist_node uidhash_node;
 590         uid_t uid;
 591
 592 #ifdef CONFIG_USER_SCHED
 593         struct task_group *tg;
 594 #ifdef CONFIG_SYSFS
 595         struct kobject kobj;
 596         struct work_struct work;
 597 #endif
 598 #endif
 599 };
 600
 601 extern int uids_sysfs_init(void);
 602
 603 extern struct user_struct *find_user(uid_t);
 604
 605 extern struct user_struct root_user;
 606 #define INIT_USER (&root_user)
 607
 608 struct backing_dev_info;
 609 struct reclaim_state;
 610
 611 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
 612 struct sched_info {
 613         /* cumulative counters */
 614         unsigned long pcount;         /* # of times run on this cpu */
 615         unsigned long long cpu_time,  /* time spent on the cpu */
 616                            run_delay; /* time spent waiting on a runqueue */
 617
 618         /* timestamps */
 619         unsigned long long last_arrival,/* when we last ran on a cpu */
 620                            last_queued; /* when we were last queued to run */
 621 #ifdef CONFIG_SCHEDSTATS
 622         /* BKL stats */
 623         unsigned int bkl_count;
 624 #endif
 625 };
 626 #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
 627
 628 #ifdef CONFIG_SCHEDSTATS
 629 extern const struct file_operations proc_schedstat_operations;
 630 #endif /* CONFIG_SCHEDSTATS */
 631
 632 #ifdef CONFIG_TASK_DELAY_ACCT
 633 struct task_delay_info {
 634         spinlock_t      lock;
 635         unsigned int    flags;  /* Private per-task flags */
 636
 637         /* For each stat XXX, add following, aligned appropriately
 638          *
 639          * struct timespec XXX_start, XXX_end;
 640          * u64 XXX_delay;
 641          * u32 XXX_count;
 642          *
 643          * Atomicity of updates to XXX_delay, XXX_count protected by
 644          * single lock above (split into XXX_lock if contention is an issue).
 645          */
 646
 647         /*
 648          * XXX_count is incremented on every XXX operation, the delay
 649          * associated with the operation is added to XXX_delay.
 650          * XXX_delay contains the accumulated delay time in nanoseconds.
 651          */
 652         struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
 653         u64 blkio_delay;        /* wait for sync block io completion */
 654         u64 swapin_delay;       /* wait for swapin block io completion */
 655         u32 blkio_count;        /* total count of the number of sync block */
 656                                 /* io operations performed */
 657         u32 swapin_count;       /* total count of the number of swapin block */
 658                                 /* io operations performed */
 659 };
 660 #endif  /* CONFIG_TASK_DELAY_ACCT */
 661
 662 static inline int sched_info_on(void)
 663 {
 664 #ifdef CONFIG_SCHEDSTATS
 665         return 1;
 666 #elif defined(CONFIG_TASK_DELAY_ACCT)
 667         extern int delayacct_on;
 668         return delayacct_on;
 669 #else
 670         return 0;
 671 #endif
 672 }
 673
 674 enum cpu_idle_type {
 675         CPU_IDLE,
 676         CPU_NOT_IDLE,
 677         CPU_NEWLY_IDLE,
 678         CPU_MAX_IDLE_TYPES
 679 };
 680
 681 /*
 682  * sched-domains (multiprocessor balancing) declarations:
 683  */
 684
 685 /*
 686  * Increase resolution of nice-level calculations:
 687  */
 688 #define SCHED_LOAD_SHIFT        10
 689 #define SCHED_LOAD_SCALE        (1L << SCHED_LOAD_SHIFT)
 690
 691 #define SCHED_LOAD_SCALE_FUZZ   SCHED_LOAD_SCALE
 692
 693 #ifdef CONFIG_SMP
 694 #define SD_LOAD_BALANCE         1       /* Do load balancing on this domain. */
 695 #define SD_BALANCE_NEWIDLE      2       /* Balance when about to become idle */
 696 #define SD_BALANCE_EXEC         4       /* Balance on exec */
 697 #define SD_BALANCE_FORK         8       /* Balance on fork, clone */
 698 #define SD_WAKE_IDLE            16      /* Wake to idle CPU on task wakeup */
 699 #define SD_WAKE_AFFINE          32      /* Wake task to waking CPU */
 700 #define SD_WAKE_BALANCE         64      /* Perform balancing at task wakeup */
 701 #define SD_SHARE_CPUPOWER       128     /* Domain members share cpu power */
 702 #define SD_POWERSAVINGS_BALANCE 256     /* Balance for power savings */
 703 #define SD_SHARE_PKG_RESOURCES  512     /* Domain members share cpu pkg resources */
 704 #define SD_SERIALIZE            1024    /* Only a single load balancing instance */
 705 #define SD_WAKE_IDLE_FAR        2048    /* Gain latency sacrificing cache hit */
 706
 707 #define BALANCE_FOR_MC_POWER    \
 708         (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
 709
 710 #define BALANCE_FOR_PKG_POWER   \
 711         ((sched_mc_power_savings || sched_smt_power_savings) ?  \
 712          SD_POWERSAVINGS_BALANCE : 0)
 713
 714 #define test_sd_parent(sd, flag)        ((sd->parent &&         \
 715                                          (sd->parent->flags & flag)) ? 1 : 0)
 716
 717
 718 struct sched_group {
 719         struct sched_group *next;       /* Must be a circular list */
 720         cpumask_t cpumask;
 721
 722         /*
 723          * CPU power of this group, SCHED_LOAD_SCALE being max power for a
 724          * single CPU. This is read only (except for setup, hotplug CPU).
 725          * Note : Never change cpu_power without recompute its reciprocal
 726          */
 727         unsigned int __cpu_power;
 728         /*
 729          * reciprocal value of cpu_power to avoid expensive divides
 730          * (see include/linux/reciprocal_div.h)
 731          */
 732         u32 reciprocal_cpu_power;
 733 };
 734
 735 enum sched_domain_level {
 736         SD_LV_NONE = 0,
 737         SD_LV_SIBLING,
 738         SD_LV_MC,
 739         SD_LV_CPU,
 740         SD_LV_NODE,
 741         SD_LV_ALLNODES,
 742         SD_LV_MAX
 743 };
 744
 745 struct sched_domain_attr {
 746         int relax_domain_level;
 747 };
 748
 749 #define SD_ATTR_INIT    (struct sched_domain_attr) {    \
 750         .relax_domain_level = -1,                       \
 751 }
 752
 753 struct sched_domain {
 754         /* These fields must be setup */
 755         struct sched_domain *parent;    /* top domain must be null terminated */
 756         struct sched_domain *child;     /* bottom domain must be null terminated */
 757         struct sched_group *groups;     /* the balancing groups of the domain */
 758         cpumask_t span;                 /* span of all CPUs in this domain */
 759         int first_cpu;                  /* cache of the first cpu in this domain */
 760         unsigned long min_interval;     /* Minimum balance interval ms */
 761         unsigned long max_interval;     /* Maximum balance interval ms */
 762         unsigned int busy_factor;       /* less balancing by factor if busy */
 763         unsigned int imbalance_pct;     /* No balance until over watermark */
 764         unsigned int cache_nice_tries;  /* Leave cache hot tasks for # tries */
 765         unsigned int busy_idx;
 766         unsigned int idle_idx;
 767         unsigned int newidle_idx;
 768         unsigned int wake_idx;
 769         unsigned int forkexec_idx;
 770         int flags;                      /* See SD_* */
 771         enum sched_domain_level level;
 772
 773         /* Runtime fields. */
 774         unsigned long last_balance;     /* init to jiffies. units in jiffies */
 775         unsigned int balance_interval;  /* initialise to 1. units in ms. */
 776         unsigned int nr_balance_failed; /* initialise to 0 */
 777
 778 #ifdef CONFIG_SCHEDSTATS
 779         /* load_balance() stats */
 780         unsigned int lb_count[CPU_MAX_IDLE_TYPES];
 781         unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
 782         unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
 783         unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
 784         unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
 785         unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
 786         unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
 787         unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
 788
 789         /* Active load balancing */
 790         unsigned int alb_count;
 791         unsigned int alb_failed;
 792         unsigned int alb_pushed;
 793
 794         /* SD_BALANCE_EXEC stats */
 795         unsigned int sbe_count;
 796         unsigned int sbe_balanced;
 797         unsigned int sbe_pushed;
 798
 799         /* SD_BALANCE_FORK stats */
 800         unsigned int sbf_count;
 801         unsigned int sbf_balanced;
 802         unsigned int sbf_pushed;
 803
 804         /* try_to_wake_up() stats */
 805         unsigned int ttwu_wake_remote;
 806         unsigned int ttwu_move_affine;
 807         unsigned int ttwu_move_balance;
 808 #endif
 809 };
 810
 811 extern void partition_sched_domains(int ndoms_new, cpumask_t *doms_new,
 812                                     struct sched_domain_attr *dattr_new);
 813 extern int arch_reinit_sched_domains(void);
 814
 815 #endif  /* CONFIG_SMP */
 816
 817 /*
 818  * A runqueue laden with a single nice 0 task scores a weighted_cpuload of
 819  * SCHED_LOAD_SCALE. This function returns 1 if any cpu is laden with a
 820  * task of nice 0 or enough lower priority tasks to bring up the
 821  * weighted_cpuload
 822  */
 823 static inline int above_background_load(void)
 824 {
 825         unsigned long cpu;
 826
 827         for_each_online_cpu(cpu) {
 828                 if (weighted_cpuload(cpu) >= SCHED_LOAD_SCALE)
 829                         return 1;
 830         }
 831         return 0;
 832 }
 833
 834 struct io_context;                      /* See blkdev.h */
 835 #define NGROUPS_SMALL           32
 836 #define NGROUPS_PER_BLOCK       ((unsigned int)(PAGE_SIZE / sizeof(gid_t)))
 837 struct group_info {
 838         int ngroups;
 839         atomic_t usage;
 840         gid_t small_block[NGROUPS_SMALL];
 841         int nblocks;
 842         gid_t *blocks[0];
 843 };
 844
 845 /*
 846  * get_group_info() must be called with the owning task locked (via task_lock())
 847  * when task != current.  The reason being that the vast majority of callers are
 848  * looking at current->group_info, which can not be changed except by the
 849  * current task.  Changing current->group_info requires the task lock, too.
 850  */
 851 #define get_group_info(group_info) do { \
 852         atomic_inc(&(group_info)->usage); \
 853 } while (0)
 854
 855 #define put_group_info(group_info) do { \
 856         if (atomic_dec_and_test(&(group_info)->usage)) \
 857                 groups_free(group_info); \
 858 } while (0)
 859
 860 extern struct group_info *groups_alloc(int gidsetsize);
 861 extern void groups_free(struct group_info *group_info);
 862 extern int set_current_groups(struct group_info *group_info);
 863 extern int groups_search(struct group_info *group_info, gid_t grp);
 864 /* access the groups "array" with this macro */
 865 #define GROUP_AT(gi, i) \
 866     ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
 867
 868 #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
 869 extern void prefetch_stack(struct task_struct *t);
 870 #else
 871 static inline void prefetch_stack(struct task_struct *t) { }
 872 #endif
 873
 874 struct audit_context;           /* See audit.c */
 875 struct mempolicy;
 876 struct pipe_inode_info;
 877 struct uts_namespace;
 878
 879 struct rq;
 880 struct sched_domain;
 881
 882 struct sched_class {
 883         const struct sched_class *next;
 884
 885         void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
 886         void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
 887         void (*yield_task) (struct rq *rq);
 888         int  (*select_task_rq)(struct task_struct *p, int sync);
 889
 890         void (*check_preempt_curr) (struct rq *rq, struct task_struct *p);
 891
 892         struct task_struct * (*pick_next_task) (struct rq *rq);
 893         void (*put_prev_task) (struct rq *rq, struct task_struct *p);
 894
 895 #ifdef CONFIG_SMP
 896         unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
 897                         struct rq *busiest, unsigned long max_load_move,
 898                         struct sched_domain *sd, enum cpu_idle_type idle,
 899                         int *all_pinned, int *this_best_prio);
 900
 901         int (*move_one_task) (struct rq *this_rq, int this_cpu,
 902                               struct rq *busiest, struct sched_domain *sd,
 903                               enum cpu_idle_type idle);
 904         void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
 905         void (*post_schedule) (struct rq *this_rq);
 906         void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
 907 #endif
 908
 909         void (*set_curr_task) (struct rq *rq);
 910         void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
 911         void (*task_new) (struct rq *rq, struct task_struct *p);
 912         void (*set_cpus_allowed)(struct task_struct *p,
 913                                  const cpumask_t *newmask);
 914
 915         void (*join_domain)(struct rq *rq);
 916         void (*leave_domain)(struct rq *rq);
 917
 918         void (*switched_from) (struct rq *this_rq, struct task_struct *task,
 919                                int running);
 920         void (*switched_to) (struct rq *this_rq, struct task_struct *task,
 921                              int running);
 922         void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
 923                              int oldprio, int running);
 924
 925 #ifdef CONFIG_FAIR_GROUP_SCHED
 926         void (*moved_group) (struct task_struct *p);
 927 #endif
 928 };
 929
 930 struct load_weight {
 931         unsigned long weight, inv_weight;
 932 };
 933
 934 /*
 935  * CFS stats for a schedulable entity (task, task-group etc)
 936  *
 937  * Current field usage histogram:
 938  *
 939  *     4 se->block_start
 940  *     4 se->run_node
 941  *     4 se->sleep_start
 942  *     6 se->load.weight
 943  */
 944 struct sched_entity {
 945         struct load_weight      load;           /* for load-balancing */
 946         struct rb_node          run_node;
 947         struct list_head        group_node;
 948         unsigned int            on_rq;
 949
 950         u64                     exec_start;
 951         u64                     sum_exec_runtime;
 952         u64                     vruntime;
 953         u64                     prev_sum_exec_runtime;
 954
 955         u64                     last_wakeup;
 956         u64                     avg_overlap;
 957
 958 #ifdef CONFIG_SCHEDSTATS
 959         u64                     wait_start;
 960         u64                     wait_max;
 961         u64                     wait_count;
 962         u64                     wait_sum;
 963
 964         u64                     sleep_start;
 965         u64                     sleep_max;
 966         s64                     sum_sleep_runtime;
 967
 968         u64                     block_start;
 969         u64                     block_max;
 970         u64                     exec_max;
 971         u64                     slice_max;
 972
 973         u64                     nr_migrations;
 974         u64                     nr_migrations_cold;
 975         u64                     nr_failed_migrations_affine;
 976         u64                     nr_failed_migrations_running;
 977         u64                     nr_failed_migrations_hot;
 978         u64                     nr_forced_migrations;
 979         u64                     nr_forced2_migrations;
 980
 981         u64                     nr_wakeups;
 982         u64                     nr_wakeups_sync;
 983         u64                     nr_wakeups_migrate;
 984         u64                     nr_wakeups_local;
 985         u64                     nr_wakeups_remote;
 986         u64                     nr_wakeups_affine;
 987         u64                     nr_wakeups_affine_attempts;
 988         u64                     nr_wakeups_passive;
 989         u64                     nr_wakeups_idle;
 990 #endif
 991
 992 #ifdef CONFIG_FAIR_GROUP_SCHED
 993         struct sched_entity     *parent;
 994         /* rq on which this entity is (to be) queued: */
 995         struct cfs_rq           *cfs_rq;
 996         /* rq "owned" by this entity/group: */
 997         struct cfs_rq           *my_q;
 998 #endif
 999 };
1000
1001 struct sched_rt_entity {
1002         struct list_head run_list;
1003         unsigned int time_slice;
1004         unsigned long timeout;
1005         int nr_cpus_allowed;
1006
1007         struct sched_rt_entity *back;
1008 #ifdef CONFIG_RT_GROUP_SCHED
1009         struct sched_rt_entity  *parent;
1010         /* rq on which this entity is (to be) queued: */
1011         struct rt_rq            *rt_rq;
1012         /* rq "owned" by this entity/group: */
1013         struct rt_rq            *my_q;
1014 #endif
1015 };
1016
1017 struct task_struct {
1018         volatile long state;    /* -1 unrunnable, 0 runnable, >0 stopped */
1019         void *stack;
1020         atomic_t usage;
1021         unsigned int flags;     /* per process flags, defined below */
1022         unsigned int ptrace;
1023
1024         int lock_depth;         /* BKL lock depth */
1025
1026 #ifdef CONFIG_SMP
1027 #ifdef __ARCH_WANT_UNLOCKED_CTXSW
1028         int oncpu;
1029 #endif
1030 #endif
1031
1032         int prio, static_prio, normal_prio;
1033         const struct sched_class *sched_class;
1034         struct sched_entity se;
1035         struct sched_rt_entity rt;
1036
1037 #ifdef CONFIG_PREEMPT_NOTIFIERS
1038         /* list of struct preempt_notifier: */
1039         struct hlist_head preempt_notifiers;
1040 #endif
1041
1042         /*
1043          * fpu_counter contains the number of consecutive context switches
1044          * that the FPU is used. If this is over a threshold, the lazy fpu
1045          * saving becomes unlazy to save the trap. This is an unsigned char
1046          * so that after 256 times the counter wraps and the behavior turns
1047          * lazy again; this to deal with bursty apps that only use FPU for
1048          * a short time
1049          */
1050         unsigned char fpu_counter;
1051         s8 oomkilladj; /* OOM kill score adjustment (bit shift). */
1052 #ifdef CONFIG_BLK_DEV_IO_TRACE
1053         unsigned int btrace_seq;
1054 #endif
1055
1056         unsigned int policy;
1057         cpumask_t cpus_allowed;
1058
1059 #ifdef CONFIG_PREEMPT_RCU
1060         int rcu_read_lock_nesting;
1061         int rcu_flipctr_idx;
1062 #endif /* #ifdef CONFIG_PREEMPT_RCU */
1063
1064 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
1065         struct sched_info sched_info;
1066 #endif
1067
1068         struct list_head tasks;
1069         /*
1070          * ptrace_list/ptrace_children forms the list of my children
1071          * that were stolen by a ptracer.
1072          */
1073         struct list_head ptrace_children;
1074         struct list_head ptrace_list;
1075
1076         struct mm_struct *mm, *active_mm;
1077
1078 /* task state */
1079         struct linux_binfmt *binfmt;
1080         int exit_state;
1081         int exit_code, exit_signal;
1082         int pdeath_signal;  /*  The signal sent when the parent dies  */
1083         /* ??? */
1084         unsigned int personality;
1085         unsigned did_exec:1;
1086         pid_t pid;
1087         pid_t tgid;
1088
1089 #ifdef CONFIG_CC_STACKPROTECTOR
1090         /* Canary value for the -fstack-protector gcc feature */
1091         unsigned long stack_canary;
1092 #endif
1093         /*
1094          * pointers to (original) parent process, youngest child, younger sibling,
1095          * older sibling, respectively.  (p->father can be replaced with
1096          * p->parent->pid)
1097          */
1098         struct task_struct *real_parent; /* real parent process (when being debugged) */
1099         struct task_struct *parent;     /* parent process */
1100         /*
1101          * children/sibling forms the list of my children plus the
1102          * tasks I'm ptracing.
1103          */
1104         struct list_head children;      /* list of my children */
1105         struct list_head sibling;       /* linkage in my parent's children list */
1106         struct task_struct *group_leader;       /* threadgroup leader */
1107
1108         /* PID/PID hash table linkage. */
1109         struct pid_link pids[PIDTYPE_MAX];
1110         struct list_head thread_group;
1111
1112         struct completion *vfork_done;          /* for vfork() */
1113         int __user *set_child_tid;              /* CLONE_CHILD_SETTID */
1114         int __user *clear_child_tid;            /* CLONE_CHILD_CLEARTID */
1115
1116         unsigned int rt_priority;
1117         cputime_t utime, stime, utimescaled, stimescaled;
1118         cputime_t gtime;
1119         cputime_t prev_utime, prev_stime;
1120         unsigned long nvcsw, nivcsw; /* context switch counts */
1121         struct timespec start_time;             /* monotonic time */
1122         struct timespec real_start_time;        /* boot based time */
1123 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
1124         unsigned long min_flt, maj_flt;
1125
1126         cputime_t it_prof_expires, it_virt_expires;
1127         unsigned long long it_sched_expires;
1128         struct list_head cpu_timers[3];
1129
1130 /* process credentials */
1131         uid_t uid,euid,suid,fsuid;
1132         gid_t gid,egid,sgid,fsgid;
1133         struct group_info *group_info;
1134         kernel_cap_t   cap_effective, cap_inheritable, cap_permitted, cap_bset;
1135         unsigned securebits;
1136         struct user_struct *user;
1137 #ifdef CONFIG_KEYS
1138         struct key *request_key_auth;   /* assumed request_key authority */
1139         struct key *thread_keyring;     /* keyring private to this thread */
1140         unsigned char jit_keyring;      /* default keyring to attach requested keys to */
1141 #endif
1142         char comm[TASK_COMM_LEN]; /* executable name excluding path
1143                                      - access with [gs]et_task_comm (which lock
1144                                        it with task_lock())
1145                                      - initialized normally by flush_old_exec */
1146 /* file system info */
1147         int link_count, total_link_count;
1148 #ifdef CONFIG_SYSVIPC
1149 /* ipc stuff */
1150         struct sysv_sem sysvsem;
1151 #endif
1152 #ifdef CONFIG_DETECT_SOFTLOCKUP
1153 /* hung task detection */
1154         unsigned long last_switch_timestamp;
1155         unsigned long last_switch_count;
1156 #endif
1157 /* CPU-specific state of this task */
1158         struct thread_struct thread;
1159 /* filesystem information */
1160         struct fs_struct *fs;
1161 /* open file information */
1162         struct files_struct *files;
1163 /* namespaces */
1164         struct nsproxy *nsproxy;
1165 /* signal handlers */
1166         struct signal_struct *signal;
1167         struct sighand_struct *sighand;
1168
1169         sigset_t blocked, real_blocked;
1170         sigset_t saved_sigmask;         /* To be restored with TIF_RESTORE_SIGMASK */
1171         struct sigpending pending;
1172
1173         unsigned long sas_ss_sp;
1174         size_t sas_ss_size;
1175         int (*notifier)(void *priv);
1176         void *notifier_data;
1177         sigset_t *notifier_mask;
1178 #ifdef CONFIG_SECURITY
1179         void *security;
1180 #endif
1181         struct audit_context *audit_context;
1182 #ifdef CONFIG_AUDITSYSCALL
1183         uid_t loginuid;
1184         unsigned int sessionid;
1185 #endif
1186         seccomp_t seccomp;
1187
1188 /* Thread group tracking */
1189         u32 parent_exec_id;
1190         u32 self_exec_id;
1191 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
1192         spinlock_t alloc_lock;
1193
1194         /* Protection of the PI data structures: */
1195         spinlock_t pi_lock;
1196
1197 #ifdef CONFIG_RT_MUTEXES
1198         /* PI waiters blocked on a rt_mutex held by this task */
1199         struct plist_head pi_waiters;
1200         /* Deadlock detection and priority inheritance handling */
1201         struct rt_mutex_waiter *pi_blocked_on;
1202 #endif
1203
1204 #ifdef CONFIG_DEBUG_MUTEXES
1205         /* mutex deadlock detection */
1206         struct mutex_waiter *blocked_on;
1207 #endif
1208 #ifdef CONFIG_TRACE_IRQFLAGS
1209         unsigned int irq_events;
1210         int hardirqs_enabled;
1211         unsigned long hardirq_enable_ip;
1212         unsigned int hardirq_enable_event;
1213         unsigned long hardirq_disable_ip;
1214         unsigned int hardirq_disable_event;
1215         int softirqs_enabled;
1216         unsigned long softirq_disable_ip;
1217         unsigned int softirq_disable_event;
1218         unsigned long softirq_enable_ip;
1219         unsigned int softirq_enable_event;
1220         int hardirq_context;
1221         int softirq_context;
1222 #endif
1223 #ifdef CONFIG_LOCKDEP
1224 # define MAX_LOCK_DEPTH 48UL
1225         u64 curr_chain_key;
1226         int lockdep_depth;
1227         struct held_lock held_locks[MAX_LOCK_DEPTH];
1228         unsigned int lockdep_recursion;
1229 #endif
1230
1231 /* journalling filesystem info */
1232         void *journal_info;
1233
1234 /* stacked block device info */
1235         struct bio *bio_list, **bio_tail;
1236
1237 /* VM state */
1238         struct reclaim_state *reclaim_state;
1239
1240         struct backing_dev_info *backing_dev_info;
1241
1242         struct io_context *io_context;
1243
1244         unsigned long ptrace_message;
1245         siginfo_t *last_siginfo; /* For ptrace use.  */
1246 #ifdef CONFIG_TASK_XACCT
1247 /* i/o counters(bytes read/written, #syscalls */
1248         u64 rchar, wchar, syscr, syscw;
1249 #endif
1250         struct task_io_accounting ioac;
1251 #if defined(CONFIG_TASK_XACCT)
1252         u64 acct_rss_mem1;      /* accumulated rss usage */
1253         u64 acct_vm_mem1;       /* accumulated virtual memory usage */
1254         cputime_t acct_stimexpd;/* stime since last update */
1255 #endif
1256 #ifdef CONFIG_NUMA
1257         struct mempolicy *mempolicy;
1258         short il_next;
1259 #endif
1260 #ifdef CONFIG_CPUSETS
1261         nodemask_t mems_allowed;
1262         int cpuset_mems_generation;
1263         int cpuset_mem_spread_rotor;
1264 #endif
1265 #ifdef CONFIG_CGROUPS
1266         /* Control Group info protected by css_set_lock */
1267         struct css_set *cgroups;
1268         /* cg_list protected by css_set_lock and tsk->alloc_lock */
1269         struct list_head cg_list;
1270 #endif
1271 #ifdef CONFIG_FUTEX
1272         struct robust_list_head __user *robust_list;
1273 #ifdef CONFIG_COMPAT
1274         struct compat_robust_list_head __user *compat_robust_list;
1275 #endif
1276         struct list_head pi_state_list;
1277         struct futex_pi_state *pi_state_cache;
1278 #endif
1279         atomic_t fs_excl;       /* holding fs exclusive resources */
1280         struct rcu_head rcu;
1281
1282         /*
1283          * cache last used pipe for splice
1284          */
1285         struct pipe_inode_info *splice_pipe;
1286 #ifdef  CONFIG_TASK_DELAY_ACCT
1287         struct task_delay_info *delays;
1288 #endif
1289 #ifdef CONFIG_FAULT_INJECTION
1290         int make_it_fail;
1291 #endif
1292         struct prop_local_single dirties;
1293 #ifdef CONFIG_LATENCYTOP
1294         int latency_record_count;
1295         struct latency_record latency_record[LT_SAVECOUNT];
1296 #endif
1297 };
1298
1299 /*
1300  * Priority of a process goes from 0..MAX_PRIO-1, valid RT
1301  * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
1302  * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
1303  * values are inverted: lower p->prio value means higher priority.
1304  *
1305  * The MAX_USER_RT_PRIO value allows the actual maximum
1306  * RT priority to be separate from the value exported to
1307  * user-space.  This allows kernel threads to set their
1308  * priority to a value higher than any user task. Note:
1309  * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
1310  */
1311
1312 #define MAX_USER_RT_PRIO        100
1313 #define MAX_RT_PRIO             MAX_USER_RT_PRIO
1314
1315 #define MAX_PRIO                (MAX_RT_PRIO + 40)
1316 #define DEFAULT_PRIO            (MAX_RT_PRIO + 20)
1317
1318 static inline int rt_prio(int prio)
1319 {
1320         if (unlikely(prio < MAX_RT_PRIO))
1321                 return 1;
1322         return 0;
1323 }
1324
1325 static inline int rt_task(struct task_struct *p)
1326 {
1327         return rt_prio(p->prio);
1328 }
1329
1330 static inline void set_task_session(struct task_struct *tsk, pid_t session)
1331 {
1332         tsk->signal->__session = session;
1333 }
1334
1335 static inline void set_task_pgrp(struct task_struct *tsk, pid_t pgrp)
1336 {
1337         tsk->signal->__pgrp = pgrp;
1338 }
1339
1340 static inline struct pid *task_pid(struct task_struct *task)
1341 {
1342         return task->pids[PIDTYPE_PID].pid;
1343 }
1344
1345 static inline struct pid *task_tgid(struct task_struct *task)
1346 {
1347         return task->group_leader->pids[PIDTYPE_PID].pid;
1348 }
1349
1350 static inline struct pid *task_pgrp(struct task_struct *task)
1351 {
1352         return task->group_leader->pids[PIDTYPE_PGID].pid;
1353 }
1354
1355 static inline struct pid *task_session(struct task_struct *task)
1356 {
1357         return task->group_leader->pids[PIDTYPE_SID].pid;
1358 }
1359
1360 struct pid_namespace;
1361
1362 /*
1363  * the helpers to get the task's different pids as they are seen
1364  * from various namespaces
1365  *
1366  * task_xid_nr()     : global id, i.e. the id seen from the init namespace;
1367  * task_xid_vnr()    : virtual id, i.e. the id seen from the pid namespace of
1368  *                     current.
1369  * task_xid_nr_ns()  : id seen from the ns specified;
1370  *
1371  * set_task_vxid()   : assigns a virtual id to a task;
1372  *
1373  * see also pid_nr() etc in include/linux/pid.h
1374  */
1375
1376 static inline pid_t task_pid_nr(struct task_struct *tsk)
1377 {
1378         return tsk->pid;
1379 }
1380
1381 pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1382
1383 static inline pid_t task_pid_vnr(struct task_struct *tsk)
1384 {
1385         return pid_vnr(task_pid(tsk));
1386 }
1387
1388
1389 static inline pid_t task_tgid_nr(struct task_struct *tsk)
1390 {
1391         return tsk->tgid;
1392 }
1393
1394 pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1395
1396 static inline pid_t task_tgid_vnr(struct task_struct *tsk)
1397 {
1398         return pid_vnr(task_tgid(tsk));
1399 }
1400
1401
1402 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
1403 {
1404         return tsk->signal->__pgrp;
1405 }
1406
1407 pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1408
1409 static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
1410 {
1411         return pid_vnr(task_pgrp(tsk));
1412 }
1413
1414
1415 static inline pid_t task_session_nr(struct task_struct *tsk)
1416 {
1417         return tsk->signal->__session;
1418 }
1419
1420 pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
1421
1422 static inline pid_t task_session_vnr(struct task_struct *tsk)
1423 {
1424         return pid_vnr(task_session(tsk));
1425 }
1426
1427
1428 /**
1429  * pid_alive - check that a task structure is not stale
1430  * @p: Task structure to be checked.
1431  *
1432  * Test if a process is not yet dead (at most zombie state)
1433  * If pid_alive fails, then pointers within the task structure
1434  * can be stale and must not be dereferenced.
1435  */
1436 static inline int pid_alive(struct task_struct *p)
1437 {
1438         return p->pids[PIDTYPE_PID].pid != NULL;
1439 }
1440
1441 /**
1442  * is_global_init - check if a task structure is init
1443  * @tsk: Task structure to be checked.
1444  *
1445  * Check if a task structure is the first user space task the kernel created.
1446  */
1447 static inline int is_global_init(struct task_struct *tsk)
1448 {
1449         return tsk->pid == 1;
1450 }
1451
1452 /*
1453  * is_container_init:
1454  * check whether in the task is init in its own pid namespace.
1455  */
1456 extern int is_container_init(struct task_struct *tsk);
1457
1458 extern struct pid *cad_pid;
1459
1460 extern void free_task(struct task_struct *tsk);
1461 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
1462
1463 extern void __put_task_struct(struct task_struct *t);
1464
1465 static inline void put_task_struct(struct task_struct *t)
1466 {
1467         if (atomic_dec_and_test(&t->usage))
1468                 __put_task_struct(t);
1469 }
1470
1471 /*
1472  * Per process flags
1473  */
1474 #define PF_ALIGNWARN    0x00000001      /* Print alignment warning msgs */
1475                                         /* Not implemented yet, only for 486*/
1476 #define PF_STARTING     0x00000002      /* being created */
1477 #define PF_EXITING      0x00000004      /* getting shut down */
1478 #define PF_EXITPIDONE   0x00000008      /* pi exit done on shut down */
1479 #define PF_VCPU         0x00000010      /* I'm a virtual CPU */
1480 #define PF_FORKNOEXEC   0x00000040      /* forked but didn't exec */
1481 #define PF_SUPERPRIV    0x00000100      /* used super-user privileges */
1482 #define PF_DUMPCORE     0x00000200      /* dumped core */
1483 #define PF_SIGNALED     0x00000400      /* killed by a signal */
1484 #define PF_MEMALLOC     0x00000800      /* Allocating memory */
1485 #define PF_FLUSHER      0x00001000      /* responsible for disk writeback */
1486 #define PF_USED_MATH    0x00002000      /* if unset the fpu must be initialized before use */
1487 #define PF_NOFREEZE     0x00008000      /* this thread should not be frozen */
1488 #define PF_FROZEN       0x00010000      /* frozen for system suspend */
1489 #define PF_FSTRANS      0x00020000      /* inside a filesystem transaction */
1490 #define PF_KSWAPD       0x00040000      /* I am kswapd */
1491 #define PF_SWAPOFF      0x00080000      /* I am in swapoff */
1492 #define PF_LESS_THROTTLE 0x00100000     /* Throttle me less: I clean memory */
1493 #define PF_BORROWED_MM  0x00200000      /* I am a kthread doing use_mm */
1494 #define PF_RANDOMIZE    0x00400000      /* randomize virtual address space */
1495 #define PF_SWAPWRITE    0x00800000      /* Allowed to write to swap */
1496 #define PF_SPREAD_PAGE  0x01000000      /* Spread page cache over cpuset */
1497 #define PF_SPREAD_SLAB  0x02000000      /* Spread some slab caches over cpuset */
1498 #define PF_MEMPOLICY    0x10000000      /* Non-default NUMA mempolicy */
1499 #define PF_MUTEX_TESTER 0x20000000      /* Thread belongs to the rt mutex tester */
1500 #define PF_FREEZER_SKIP 0x40000000      /* Freezer should not count it as freezeable */
1501
1502 /*
1503  * Only the _current_ task can read/write to tsk->flags, but other
1504  * tasks can access tsk->flags in readonly mode for example
1505  * with tsk_used_math (like during threaded core dumping).
1506  * There is however an exception to this rule during ptrace
1507  * or during fork: the ptracer task is allowed to write to the
1508  * child->flags of its traced child (same goes for fork, the parent
1509  * can write to the child->flags), because we're guaranteed the
1510  * child is not running and in turn not changing child->flags
1511  * at the same time the parent does it.
1512  */
1513 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
1514 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
1515 #define clear_used_math() clear_stopped_child_used_math(current)
1516 #define set_used_math() set_stopped_child_used_math(current)
1517 #define conditional_stopped_child_used_math(condition, child) \
1518         do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
1519 #define conditional_used_math(condition) \
1520         conditional_stopped_child_used_math(condition, current)
1521 #define copy_to_stopped_child_used_math(child) \
1522         do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
1523 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
1524 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
1525 #define used_math() tsk_used_math(current)
1526
1527 #ifdef CONFIG_SMP
1528 extern int set_cpus_allowed_ptr(struct task_struct *p,
1529                                 const cpumask_t *new_mask);
1530 #else
1531 static inline int set_cpus_allowed_ptr(struct task_struct *p,
1532                                        const cpumask_t *new_mask)
1533 {
1534         if (!cpu_isset(0, *new_mask))
1535                 return -EINVAL;
1536         return 0;
1537 }
1538 #endif
1539 static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
1540 {
1541         return set_cpus_allowed_ptr(p, &new_mask);
1542 }
1543
1544 extern unsigned long long sched_clock(void);
1545
1546 /*
1547  * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
1548  * clock constructed from sched_clock():
1549  */
1550 extern unsigned long long cpu_clock(int cpu);
1551
1552 extern unsigned long long
1553 task_sched_runtime(struct task_struct *task);
1554
1555 /* sched_exec is called by processes performing an exec */
1556 #ifdef CONFIG_SMP
1557 extern void sched_exec(void);
1558 #else
1559 #define sched_exec()   {}
1560 #endif
1561
1562 extern void sched_clock_idle_sleep_event(void);
1563 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
1564
1565 #ifdef CONFIG_HOTPLUG_CPU
1566 extern void idle_task_exit(void);
1567 #else
1568 static inline void idle_task_exit(void) {}
1569 #endif
1570
1571 extern void sched_idle_next(void);
1572
1573 #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
1574 extern void wake_up_idle_cpu(int cpu);
1575 #else
1576 static inline void wake_up_idle_cpu(int cpu) { }
1577 #endif
1578
1579 #ifdef CONFIG_SCHED_DEBUG
1580 extern unsigned int sysctl_sched_latency;
1581 extern unsigned int sysctl_sched_min_granularity;
1582 extern unsigned int sysctl_sched_wakeup_granularity;
1583 extern unsigned int sysctl_sched_child_runs_first;
1584 extern unsigned int sysctl_sched_features;
1585 extern unsigned int sysctl_sched_migration_cost;
1586 extern unsigned int sysctl_sched_nr_migrate;
1587
1588 int sched_nr_latency_handler(struct ctl_table *table, int write,
1589                 struct file *file, void __user *buffer, size_t *length,
1590                 loff_t *ppos);
1591 #endif
1592 extern unsigned int sysctl_sched_rt_period;
1593 extern int sysctl_sched_rt_runtime;
1594
1595 int sched_rt_handler(struct ctl_table *table, int write,
1596                 struct file *filp, void __user *buffer, size_t *lenp,
1597                 loff_t *ppos);
1598
1599 extern unsigned int sysctl_sched_compat_yield;
1600
1601 #ifdef CONFIG_RT_MUTEXES
1602 extern int rt_mutex_getprio(struct task_struct *p);
1603 extern void rt_mutex_setprio(struct task_struct *p, int prio);
1604 extern void rt_mutex_adjust_pi(struct task_struct *p);
1605 #else
1606 static inline int rt_mutex_getprio(struct task_struct *p)
1607 {
1608         return p->normal_prio;
1609 }
1610 # define rt_mutex_adjust_pi(p)          do { } while (0)
1611 #endif
1612
1613 extern void set_user_nice(struct task_struct *p, long nice);
1614 extern int task_prio(const struct task_struct *p);
1615 extern int task_nice(const struct task_struct *p);
1616 extern int can_nice(const struct task_struct *p, const int nice);
1617 extern int task_curr(const struct task_struct *p);
1618 extern int idle_cpu(int cpu);
1619 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1620 extern struct task_struct *idle_task(int cpu);
1621 extern struct task_struct *curr_task(int cpu);
1622 extern void set_curr_task(int cpu, struct task_struct *p);
1623
1624 void yield(void);
1625
1626 /*
1627  * The default (Linux) execution domain.
1628  */
1629 extern struct exec_domain       default_exec_domain;
1630
1631 union thread_union {
1632         struct thread_info thread_info;
1633         unsigned long stack[THREAD_SIZE/sizeof(long)];
1634 };
1635
1636 #ifndef __HAVE_ARCH_KSTACK_END
1637 static inline int kstack_end(void *addr)
1638 {
1639         /* Reliable end of stack detection:
1640          * Some APM bios versions misalign the stack
1641          */
1642         return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1643 }
1644 #endif
1645
1646 extern union thread_union init_thread_union;
1647 extern struct task_struct init_task;
1648
1649 extern struct   mm_struct init_mm;
1650
1651 extern struct pid_namespace init_pid_ns;
1652
1653 /*
1654  * find a task by one of its numerical ids
1655  *
1656  * find_task_by_pid_type_ns():
1657  *      it is the most generic call - it finds a task by all id,
1658  *      type and namespace specified
1659  * find_task_by_pid_ns():
1660  *      finds a task by its pid in the specified namespace
1661  * find_task_by_vpid():
1662  *      finds a task by its virtual pid
1663  * find_task_by_pid():
1664  *      finds a task by its global pid
1665  *
1666  * see also find_pid() etc in include/linux/pid.h
1667  */
1668
1669 extern struct task_struct *find_task_by_pid_type_ns(int type, int pid,
1670                 struct pid_namespace *ns);
1671
1672 extern struct task_struct *find_task_by_pid(pid_t nr);
1673 extern struct task_struct *find_task_by_vpid(pid_t nr);
1674 extern struct task_struct *find_task_by_pid_ns(pid_t nr,
1675                 struct pid_namespace *ns);
1676
1677 extern void __set_special_pids(struct pid *pid);
1678
1679 /* per-UID process charging. */
1680 extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
1681 static inline struct user_struct *get_uid(struct user_struct *u)
1682 {
1683         atomic_inc(&u->__count);
1684         return u;
1685 }
1686 extern void free_uid(struct user_struct *);
1687 extern void switch_uid(struct user_struct *);
1688 extern void release_uids(struct user_namespace *ns);
1689
1690 #include <asm/current.h>
1691
1692 extern void do_timer(unsigned long ticks);
1693
1694 extern int wake_up_state(struct task_struct *tsk, unsigned int state);
1695 extern int wake_up_process(struct task_struct *tsk);
1696 extern void wake_up_new_task(struct task_struct *tsk,
1697                                 unsigned long clone_flags);
1698 #ifdef CONFIG_SMP
1699  extern void kick_process(struct task_struct *tsk);
1700 #else
1701  static inline void kick_process(struct task_struct *tsk) { }
1702 #endif
1703 extern void sched_fork(struct task_struct *p, int clone_flags);
1704 extern void sched_dead(struct task_struct *p);
1705
1706 extern int in_group_p(gid_t);
1707 extern int in_egroup_p(gid_t);
1708
1709 extern void proc_caches_init(void);
1710 extern void flush_signals(struct task_struct *);
1711 extern void ignore_signals(struct task_struct *);
1712 extern void flush_signal_handlers(struct task_struct *, int force_default);
1713 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1714
1715 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1716 {
1717         unsigned long flags;
1718         int ret;
1719
1720         spin_lock_irqsave(&tsk->sighand->siglock, flags);
1721         ret = dequeue_signal(tsk, mask, info);
1722         spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1723
1724         return ret;
1725 }
1726
1727 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1728                               sigset_t *mask);
1729 extern void unblock_all_signals(void);
1730 extern void release_task(struct task_struct * p);
1731 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1732 extern int force_sigsegv(int, struct task_struct *);
1733 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1734 extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
1735 extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
1736 extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
1737 extern int kill_pgrp(struct pid *pid, int sig, int priv);
1738 extern int kill_pid(struct pid *pid, int sig, int priv);
1739 extern int kill_proc_info(int, struct siginfo *, pid_t);
1740 extern void do_notify_parent(struct task_struct *, int);
1741 extern void force_sig(int, struct task_struct *);
1742 extern void force_sig_specific(int, struct task_struct *);
1743 extern int send_sig(int, struct task_struct *, int);
1744 extern void zap_other_threads(struct task_struct *p);
1745 extern int kill_proc(pid_t, int, int);
1746 extern struct sigqueue *sigqueue_alloc(void);
1747 extern void sigqueue_free(struct sigqueue *);
1748 extern int send_sigqueue(int, struct sigqueue *,  struct task_struct *);
1749 extern int send_group_sigqueue(int, struct sigqueue *,  struct task_struct *);
1750 extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1751 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1752
1753 static inline int kill_cad_pid(int sig, int priv)
1754 {
1755         return kill_pid(cad_pid, sig, priv);
1756 }
1757
1758 /* These can be the second arg to send_sig_info/send_group_sig_info.  */
1759 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1760 #define SEND_SIG_PRIV   ((struct siginfo *) 1)
1761 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1762
1763 static inline int is_si_special(const struct siginfo *info)
1764 {
1765         return info <= SEND_SIG_FORCED;
1766 }
1767
1768 /* True if we are on the alternate signal stack.  */
1769
1770 static inline int on_sig_stack(unsigned long sp)
1771 {
1772         return (sp - current->sas_ss_sp < current->sas_ss_size);
1773 }
1774
1775 static inline int sas_ss_flags(unsigned long sp)
1776 {
1777         return (current->sas_ss_size == 0 ? SS_DISABLE
1778                 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1779 }
1780
1781 /*
1782  * Routines for handling mm_structs
1783  */
1784 extern struct mm_struct * mm_alloc(void);
1785
1786 /* mmdrop drops the mm and the page tables */
1787 extern void __mmdrop(struct mm_struct *);
1788 static inline void mmdrop(struct mm_struct * mm)
1789 {
1790         if (unlikely(atomic_dec_and_test(&mm->mm_count)))
1791                 __mmdrop(mm);
1792 }
1793
1794 /* mmput gets rid of the mappings and all user-space */
1795 extern void mmput(struct mm_struct *);
1796 /* Grab a reference to a task's mm, if it is not already going away */
1797 extern struct mm_struct *get_task_mm(struct task_struct *task);
1798 /* Remove the current tasks stale references to the old mm_struct */
1799 extern void mm_release(struct task_struct *, struct mm_struct *);
1800 /* Allocate a new mm structure and copy contents from tsk->mm */
1801 extern struct mm_struct *dup_mm(struct task_struct *tsk);
1802
1803 extern int  copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1804 extern void flush_thread(void);
1805 extern void exit_thread(void);
1806
1807 extern void exit_files(struct task_struct *);
1808 extern void __cleanup_signal(struct signal_struct *);
1809 extern void __cleanup_sighand(struct sighand_struct *);
1810 extern void exit_itimers(struct signal_struct *);
1811
1812 extern NORET_TYPE void do_group_exit(int);
1813
1814 extern void daemonize(const char *, ...);
1815 extern int allow_signal(int);
1816 extern int disallow_signal(int);
1817
1818 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1819 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1820 struct task_struct *fork_idle(int);
1821
1822 extern void set_task_comm(struct task_struct *tsk, char *from);
1823 extern char *get_task_comm(char *to, struct task_struct *tsk);
1824
1825 #ifdef CONFIG_SMP
1826 extern void wait_task_inactive(struct task_struct * p);
1827 #else
1828 #define wait_task_inactive(p)   do { } while (0)
1829 #endif
1830
1831 #define remove_parent(p)        list_del_init(&(p)->sibling)
1832 #define add_parent(p)           list_add_tail(&(p)->sibling,&(p)->parent->children)
1833
1834 #define next_task(p)    list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1835
1836 #define for_each_process(p) \
1837         for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1838
1839 /*
1840  * Careful: do_each_thread/while_each_thread is a double loop so
1841  *          'break' will not work as expected - use goto instead.
1842  */
1843 #define do_each_thread(g, t) \
1844         for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1845
1846 #define while_each_thread(g, t) \
1847         while ((t = next_thread(t)) != g)
1848
1849 /* de_thread depends on thread_group_leader not being a pid based check */
1850 #define thread_group_leader(p)  (p == p->group_leader)
1851
1852 /* Do to the insanities of de_thread it is possible for a process
1853  * to have the pid of the thread group leader without actually being
1854  * the thread group leader.  For iteration through the pids in proc
1855  * all we care about is that we have a task with the appropriate
1856  * pid, we don't actually care if we have the right task.
1857  */
1858 static inline int has_group_leader_pid(struct task_struct *p)
1859 {
1860         return p->pid == p->tgid;
1861 }
1862
1863 static inline
1864 int same_thread_group(struct task_struct *p1, struct task_struct *p2)
1865 {
1866         return p1->tgid == p2->tgid;
1867 }
1868
1869 static inline struct task_struct *next_thread(const struct task_struct *p)
1870 {
1871         return list_entry(rcu_dereference(p->thread_group.next),
1872                           struct task_struct, thread_group);
1873 }
1874
1875 static inline int thread_group_empty(struct task_struct *p)
1876 {
1877         return list_empty(&p->thread_group);
1878 }
1879
1880 #define delay_group_leader(p) \
1881                 (thread_group_leader(p) && !thread_group_empty(p))
1882
1883 /*
1884  * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
1885  * subscriptions and synchronises with wait4().  Also used in procfs.  Also
1886  * pins the final release of task.io_context.  Also protects ->cpuset and
1887  * ->cgroup.subsys[].
1888  *
1889  * Nests both inside and outside of read_lock(&tasklist_lock).
1890  * It must not be nested with write_lock_irq(&tasklist_lock),
1891  * neither inside nor outside.
1892  */
1893 static inline void task_lock(struct task_struct *p)
1894 {
1895         spin_lock(&p->alloc_lock);
1896 }
1897
1898 static inline void task_unlock(struct task_struct *p)
1899 {
1900         spin_unlock(&p->alloc_lock);
1901 }
1902
1903 extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1904                                                         unsigned long *flags);
1905
1906 static inline void unlock_task_sighand(struct task_struct *tsk,
1907                                                 unsigned long *flags)
1908 {
1909         spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1910 }
1911
1912 #ifndef __HAVE_THREAD_FUNCTIONS
1913
1914 #define task_thread_info(task)  ((struct thread_info *)(task)->stack)
1915 #define task_stack_page(task)   ((task)->stack)
1916
1917 static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1918 {
1919         *task_thread_info(p) = *task_thread_info(org);
1920         task_thread_info(p)->task = p;
1921 }
1922
1923 static inline unsigned long *end_of_stack(struct task_struct *p)
1924 {
1925         return (unsigned long *)(task_thread_info(p) + 1);
1926 }
1927
1928 #endif
1929
1930 extern void thread_info_cache_init(void);
1931
1932 /* set thread flags in other task's structures
1933  * - see asm/thread_info.h for TIF_xxxx flags available
1934  */
1935 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1936 {
1937         set_ti_thread_flag(task_thread_info(tsk), flag);
1938 }
1939
1940 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1941 {
1942         clear_ti_thread_flag(task_thread_info(tsk), flag);
1943 }
1944
1945 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1946 {
1947         return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1948 }
1949
1950 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1951 {
1952         return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1953 }
1954
1955 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1956 {
1957         return test_ti_thread_flag(task_thread_info(tsk), flag);
1958 }
1959
1960 static inline void set_tsk_need_resched(struct task_struct *tsk)
1961 {
1962         set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1963 }
1964
1965 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1966 {
1967         clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1968 }
1969
1970 static inline int signal_pending(struct task_struct *p)
1971 {
1972         return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1973 }
1974
1975 extern int __fatal_signal_pending(struct task_struct *p);
1976
1977 static inline int fatal_signal_pending(struct task_struct *p)
1978 {
1979         return signal_pending(p) && __fatal_signal_pending(p);
1980 }
1981
1982 static inline int need_resched(void)
1983 {
1984         return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1985 }
1986
1987 /*
1988  * cond_resched() and cond_resched_lock(): latency reduction via
1989  * explicit rescheduling in places that are safe. The return
1990  * value indicates whether a reschedule was done in fact.
1991  * cond_resched_lock() will drop the spinlock before scheduling,
1992  * cond_resched_softirq() will enable bhs before scheduling.
1993  */
1994 #ifdef CONFIG_PREEMPT
1995 static inline int cond_resched(void)
1996 {
1997         return 0;
1998 }
1999 #else
2000 extern int _cond_resched(void);
2001 static inline int cond_resched(void)
2002 {
2003         return _cond_resched();
2004 }
2005 #endif
2006 extern int cond_resched_lock(spinlock_t * lock);
2007 extern int cond_resched_softirq(void);
2008
2009 /*
2010  * Does a critical section need to be broken due to another
2011  * task waiting?: (technically does not depend on CONFIG_PREEMPT,
2012  * but a general need for low latency)
2013  */
2014 static inline int spin_needbreak(spinlock_t *lock)
2015 {
2016 #ifdef CONFIG_PREEMPT
2017         return spin_is_contended(lock);
2018 #else
2019         return 0;
2020 #endif
2021 }
2022
2023 /*
2024  * Reevaluate whether the task has signals pending delivery.
2025  * Wake the task if so.
2026  * This is required every time the blocked sigset_t changes.
2027  * callers must hold sighand->siglock.
2028  */
2029 extern void recalc_sigpending_and_wake(struct task_struct *t);
2030 extern void recalc_sigpending(void);
2031
2032 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
2033
2034 /*
2035  * Wrappers for p->thread_info->cpu access. No-op on UP.
2036  */
2037 #ifdef CONFIG_SMP
2038
2039 static inline unsigned int task_cpu(const struct task_struct *p)
2040 {
2041         return task_thread_info(p)->cpu;
2042 }
2043
2044 extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
2045
2046 #else
2047
2048 static inline unsigned int task_cpu(const struct task_struct *p)
2049 {
2050         return 0;
2051 }
2052
2053 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
2054 {
2055 }
2056
2057 #endif /* CONFIG_SMP */
2058
2059 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
2060 extern void arch_pick_mmap_layout(struct mm_struct *mm);
2061 #else
2062 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
2063 {
2064         mm->mmap_base = TASK_UNMAPPED_BASE;
2065         mm->get_unmapped_area = arch_get_unmapped_area;
2066         mm->unmap_area = arch_unmap_area;
2067 }
2068 #endif
2069
2070 extern long sched_setaffinity(pid_t pid, const cpumask_t *new_mask);
2071 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
2072
2073 extern int sched_mc_power_savings, sched_smt_power_savings;
2074
2075 extern void normalize_rt_tasks(void);
2076
2077 #ifdef CONFIG_GROUP_SCHED
2078
2079 extern struct task_group init_task_group;
2080 #ifdef CONFIG_USER_SCHED
2081 extern struct task_group root_task_group;
2082 #endif
2083
2084 extern struct task_group *sched_create_group(struct task_group *parent);
2085 extern void sched_destroy_group(struct task_group *tg);
2086 extern void sched_move_task(struct task_struct *tsk);
2087 #ifdef CONFIG_FAIR_GROUP_SCHED
2088 extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
2089 extern unsigned long sched_group_shares(struct task_group *tg);
2090 #endif
2091 #ifdef CONFIG_RT_GROUP_SCHED
2092 extern int sched_group_set_rt_runtime(struct task_group *tg,
2093                                       long rt_runtime_us);
2094 extern long sched_group_rt_runtime(struct task_group *tg);
2095 extern int sched_group_set_rt_period(struct task_group *tg,
2096                                       long rt_period_us);
2097 extern long sched_group_rt_period(struct task_group *tg);
2098 #endif
2099 #endif
2100
2101 #ifdef CONFIG_TASK_XACCT
2102 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2103 {
2104         tsk->rchar += amt;
2105 }
2106
2107 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2108 {
2109         tsk->wchar += amt;
2110 }
2111
2112 static inline void inc_syscr(struct task_struct *tsk)
2113 {
2114         tsk->syscr++;
2115 }
2116
2117 static inline void inc_syscw(struct task_struct *tsk)
2118 {
2119         tsk->syscw++;
2120 }
2121 #else
2122 static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
2123 {
2124 }
2125
2126 static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
2127 {
2128 }
2129
2130 static inline void inc_syscr(struct task_struct *tsk)
2131 {
2132 }
2133
2134 static inline void inc_syscw(struct task_struct *tsk)
2135 {
2136 }
2137 #endif
2138
2139 #ifdef CONFIG_SMP
2140 void migration_init(void);
2141 #else
2142 static inline void migration_init(void)
2143 {
2144 }
2145 #endif
2146
2147 #ifndef TASK_SIZE_OF
2148 #define TASK_SIZE_OF(tsk)       TASK_SIZE
2149 #endif
2150
2151 #ifdef CONFIG_MM_OWNER
2152 extern void mm_update_next_owner(struct mm_struct *mm);
2153 extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
2154 #else
2155 static inline void mm_update_next_owner(struct mm_struct *mm)
2156 {
2157 }
2158
2159 static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
2160 {
2161 }
2162 #endif /* CONFIG_MM_OWNER */
2163
2164 #endif /* __KERNEL__ */
2165
2166 #endif