4 #include <asm/param.h> /* for HZ */
6 #include <linux/config.h>
7 #include <linux/capability.h>
8 #include <linux/threads.h>
9 #include <linux/kernel.h>
10 #include <linux/types.h>
11 #include <linux/timex.h>
12 #include <linux/jiffies.h>
13 #include <linux/rbtree.h>
14 #include <linux/thread_info.h>
15 #include <linux/cpumask.h>
16 #include <linux/errno.h>
17 #include <linux/nodemask.h>
19 #include <asm/system.h>
20 #include <asm/semaphore.h>
22 #include <asm/ptrace.h>
24 #include <asm/cputime.h>
26 #include <linux/smp.h>
27 #include <linux/sem.h>
28 #include <linux/signal.h>
29 #include <linux/securebits.h>
30 #include <linux/fs_struct.h>
31 #include <linux/compiler.h>
32 #include <linux/completion.h>
33 #include <linux/pid.h>
34 #include <linux/percpu.h>
35 #include <linux/topology.h>
36 #include <linux/seccomp.h>
43 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
44 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
45 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
46 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
47 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
48 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
49 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
50 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
51 #define CLONE_THREAD 0x00010000 /* Same thread group? */
52 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
53 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
54 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
55 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
56 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
57 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
58 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
59 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
60 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
63 * List of flags we want to share for kernel threads,
64 * if only because they are not used by them anyway.
66 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
69 * These are the constant used to fake the fixed-point load-average
70 * counting. Some notes:
71 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
72 * a load-average precision of 10 bits integer + 11 bits fractional
73 * - if you want to count load-averages more often, you need more
74 * precision, or rounding will get you. With 2-second counting freq,
75 * the EXP_n values would be 1981, 2034 and 2043 if still using only
78 extern unsigned long avenrun[]; /* Load averages */
80 #define FSHIFT 11 /* nr of bits of precision */
81 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
82 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
83 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
84 #define EXP_5 2014 /* 1/exp(5sec/5min) */
85 #define EXP_15 2037 /* 1/exp(5sec/15min) */
87 #define CALC_LOAD(load,exp,n) \
89 load += n*(FIXED_1-exp); \
92 extern unsigned long total_forks;
93 extern int nr_threads;
95 DECLARE_PER_CPU(unsigned long, process_counts);
96 extern int nr_processes(void);
97 extern unsigned long nr_running(void);
98 extern unsigned long nr_uninterruptible(void);
99 extern unsigned long nr_iowait(void);
101 #include <linux/time.h>
102 #include <linux/param.h>
103 #include <linux/resource.h>
104 #include <linux/timer.h>
106 #include <asm/processor.h>
108 #define TASK_RUNNING 0
109 #define TASK_INTERRUPTIBLE 1
110 #define TASK_UNINTERRUPTIBLE 2
111 #define TASK_STOPPED 4
112 #define TASK_TRACED 8
113 #define EXIT_ZOMBIE 16
116 #define __set_task_state(tsk, state_value) \
117 do { (tsk)->state = (state_value); } while (0)
118 #define set_task_state(tsk, state_value) \
119 set_mb((tsk)->state, (state_value))
121 #define __set_current_state(state_value) \
122 do { current->state = (state_value); } while (0)
123 #define set_current_state(state_value) \
124 set_mb(current->state, (state_value))
126 /* Task command name length */
127 #define TASK_COMM_LEN 16
130 * Scheduling policies
132 #define SCHED_NORMAL 0
142 #include <linux/spinlock.h>
145 * This serializes "schedule()" and also protects
146 * the run-queue from deletions/modifications (but
147 * _adding_ to the beginning of the run-queue has
150 extern rwlock_t tasklist_lock;
151 extern spinlock_t mmlist_lock;
153 typedef struct task_struct task_t;
155 extern void sched_init(void);
156 extern void sched_init_smp(void);
157 extern void init_idle(task_t *idle, int cpu);
159 extern cpumask_t nohz_cpu_mask;
161 extern void show_state(void);
162 extern void show_regs(struct pt_regs *);
165 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
166 * task), SP is the stack pointer of the first frame that should be shown in the back
167 * trace (or NULL if the entire call-chain of the task should be shown).
169 extern void show_stack(struct task_struct *task, unsigned long *sp);
171 void io_schedule(void);
172 long io_schedule_timeout(long timeout);
174 extern void cpu_init (void);
175 extern void trap_init(void);
176 extern void update_process_times(int user);
177 extern void scheduler_tick(void);
179 /* Attach to any functions which should be ignored in wchan output. */
180 #define __sched __attribute__((__section__(".sched.text")))
181 /* Is this address in the __sched functions? */
182 extern int in_sched_functions(unsigned long addr);
184 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
185 extern signed long FASTCALL(schedule_timeout(signed long timeout));
186 asmlinkage void schedule(void);
190 /* Maximum number of active map areas.. This is a random (large) number */
191 #define DEFAULT_MAX_MAP_COUNT 65536
193 extern int sysctl_max_map_count;
195 #include <linux/aio.h>
198 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
199 unsigned long, unsigned long);
201 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
202 unsigned long len, unsigned long pgoff,
203 unsigned long flags);
204 extern void arch_unmap_area(struct mm_struct *, unsigned long);
205 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
207 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
208 #define get_mm_counter(mm, member) ((mm)->_##member)
209 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
210 #define inc_mm_counter(mm, member) (mm)->_##member++
211 #define dec_mm_counter(mm, member) (mm)->_##member--
212 typedef unsigned long mm_counter_t;
215 struct vm_area_struct * mmap; /* list of VMAs */
216 struct rb_root mm_rb;
217 struct vm_area_struct * mmap_cache; /* last find_vma result */
218 unsigned long (*get_unmapped_area) (struct file *filp,
219 unsigned long addr, unsigned long len,
220 unsigned long pgoff, unsigned long flags);
221 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
222 unsigned long mmap_base; /* base of mmap area */
223 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
224 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
226 atomic_t mm_users; /* How many users with user space? */
227 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
228 int map_count; /* number of VMAs */
229 struct rw_semaphore mmap_sem;
230 spinlock_t page_table_lock; /* Protects page tables and some counters */
232 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
233 * together off init_mm.mmlist, and are protected
237 unsigned long start_code, end_code, start_data, end_data;
238 unsigned long start_brk, brk, start_stack;
239 unsigned long arg_start, arg_end, env_start, env_end;
240 unsigned long total_vm, locked_vm, shared_vm;
241 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
243 /* Special counters protected by the page_table_lock */
245 mm_counter_t _anon_rss;
247 unsigned long saved_auxv[42]; /* for /proc/PID/auxv */
250 cpumask_t cpu_vm_mask;
252 /* Architecture-specific MM context */
253 mm_context_t context;
255 /* Token based thrashing protection. */
256 unsigned long swap_token_time;
259 /* coredumping support */
261 struct completion *core_startup_done, core_done;
264 rwlock_t ioctx_list_lock;
265 struct kioctx *ioctx_list;
267 struct kioctx default_kioctx;
269 unsigned long hiwater_rss; /* High-water RSS usage */
270 unsigned long hiwater_vm; /* High-water virtual memory usage */
273 struct sighand_struct {
275 struct k_sigaction action[_NSIG];
280 * NOTE! "signal_struct" does not have it's own
281 * locking, because a shared signal_struct always
282 * implies a shared sighand_struct, so locking
283 * sighand_struct is always a proper superset of
284 * the locking of signal_struct.
286 struct signal_struct {
290 wait_queue_head_t wait_chldexit; /* for wait4() */
292 /* current thread group signal load-balancing target: */
295 /* shared signal handling: */
296 struct sigpending shared_pending;
298 /* thread group exit support */
301 * - notify group_exit_task when ->count is equal to notify_count
302 * - everyone except group_exit_task is stopped during signal delivery
303 * of fatal signals, group_exit_task processes the signal.
305 struct task_struct *group_exit_task;
308 /* thread group stop support, overloads group_exit_code too */
309 int group_stop_count;
310 unsigned int flags; /* see SIGNAL_* flags below */
312 /* POSIX.1b Interval Timers */
313 struct list_head posix_timers;
315 /* ITIMER_REAL timer for the process */
316 struct timer_list real_timer;
317 unsigned long it_real_value, it_real_incr;
319 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
320 cputime_t it_prof_expires, it_virt_expires;
321 cputime_t it_prof_incr, it_virt_incr;
323 /* job control IDs */
327 /* boolean value for session group leader */
330 struct tty_struct *tty; /* NULL if no tty */
333 * Cumulative resource counters for dead threads in the group,
334 * and for reaped dead child processes forked by this group.
335 * Live threads maintain their own counters and add to these
336 * in __exit_signal, except for the group leader.
338 cputime_t utime, stime, cutime, cstime;
339 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
340 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
343 * Cumulative ns of scheduled CPU time for dead threads in the
344 * group, not including a zombie group leader. (This only differs
345 * from jiffies_to_ns(utime + stime) if sched_clock uses something
346 * other than jiffies.)
348 unsigned long long sched_time;
351 * We don't bother to synchronize most readers of this at all,
352 * because there is no reader checking a limit that actually needs
353 * to get both rlim_cur and rlim_max atomically, and either one
354 * alone is a single word that can safely be read normally.
355 * getrlimit/setrlimit use task_lock(current->group_leader) to
356 * protect this instead of the siglock, because they really
357 * have no need to disable irqs.
359 struct rlimit rlim[RLIM_NLIMITS];
361 struct list_head cpu_timers[3];
363 /* keep the process-shared keyrings here so that they do the right
364 * thing in threads created with CLONE_THREAD */
366 struct key *session_keyring; /* keyring inherited over fork */
367 struct key *process_keyring; /* keyring private to this process */
371 /* Context switch must be unlocked if interrupts are to be enabled */
372 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
373 # define __ARCH_WANT_UNLOCKED_CTXSW
377 * Bits in flags field of signal_struct.
379 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
380 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
381 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
382 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
386 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
387 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
388 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
389 * are inverted: lower p->prio value means higher priority.
391 * The MAX_USER_RT_PRIO value allows the actual maximum
392 * RT priority to be separate from the value exported to
393 * user-space. This allows kernel threads to set their
394 * priority to a value higher than any user task. Note:
395 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
398 #define MAX_USER_RT_PRIO 100
399 #define MAX_RT_PRIO MAX_USER_RT_PRIO
401 #define MAX_PRIO (MAX_RT_PRIO + 40)
403 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
406 * Some day this will be a full-fledged user tracking system..
409 atomic_t __count; /* reference count */
410 atomic_t processes; /* How many processes does this user have? */
411 atomic_t files; /* How many open files does this user have? */
412 atomic_t sigpending; /* How many pending signals does this user have? */
413 #ifdef CONFIG_INOTIFY
414 atomic_t inotify_watches; /* How many inotify watches does this user have? */
415 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
417 /* protected by mq_lock */
418 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
419 unsigned long locked_shm; /* How many pages of mlocked shm ? */
422 struct key *uid_keyring; /* UID specific keyring */
423 struct key *session_keyring; /* UID's default session keyring */
426 /* Hash table maintenance information */
427 struct list_head uidhash_list;
431 extern struct user_struct *find_user(uid_t);
433 extern struct user_struct root_user;
434 #define INIT_USER (&root_user)
436 typedef struct prio_array prio_array_t;
437 struct backing_dev_info;
438 struct reclaim_state;
440 #ifdef CONFIG_SCHEDSTATS
442 /* cumulative counters */
443 unsigned long cpu_time, /* time spent on the cpu */
444 run_delay, /* time spent waiting on a runqueue */
445 pcnt; /* # of timeslices run on this cpu */
448 unsigned long last_arrival, /* when we last ran on a cpu */
449 last_queued; /* when we were last queued to run */
452 extern struct file_operations proc_schedstat_operations;
464 * sched-domains (multiprocessor balancing) declarations:
467 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
469 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
470 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
471 #define SD_BALANCE_EXEC 4 /* Balance on exec */
472 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
473 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
474 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
475 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
476 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
479 struct sched_group *next; /* Must be a circular list */
483 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
484 * single CPU. This is read only (except for setup, hotplug CPU).
486 unsigned long cpu_power;
489 struct sched_domain {
490 /* These fields must be setup */
491 struct sched_domain *parent; /* top domain must be null terminated */
492 struct sched_group *groups; /* the balancing groups of the domain */
493 cpumask_t span; /* span of all CPUs in this domain */
494 unsigned long min_interval; /* Minimum balance interval ms */
495 unsigned long max_interval; /* Maximum balance interval ms */
496 unsigned int busy_factor; /* less balancing by factor if busy */
497 unsigned int imbalance_pct; /* No balance until over watermark */
498 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
499 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
500 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
501 unsigned int busy_idx;
502 unsigned int idle_idx;
503 unsigned int newidle_idx;
504 unsigned int wake_idx;
505 unsigned int forkexec_idx;
506 int flags; /* See SD_* */
508 /* Runtime fields. */
509 unsigned long last_balance; /* init to jiffies. units in jiffies */
510 unsigned int balance_interval; /* initialise to 1. units in ms. */
511 unsigned int nr_balance_failed; /* initialise to 0 */
513 #ifdef CONFIG_SCHEDSTATS
514 /* load_balance() stats */
515 unsigned long lb_cnt[MAX_IDLE_TYPES];
516 unsigned long lb_failed[MAX_IDLE_TYPES];
517 unsigned long lb_balanced[MAX_IDLE_TYPES];
518 unsigned long lb_imbalance[MAX_IDLE_TYPES];
519 unsigned long lb_gained[MAX_IDLE_TYPES];
520 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
521 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
522 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
524 /* Active load balancing */
525 unsigned long alb_cnt;
526 unsigned long alb_failed;
527 unsigned long alb_pushed;
529 /* SD_BALANCE_EXEC stats */
530 unsigned long sbe_cnt;
531 unsigned long sbe_balanced;
532 unsigned long sbe_pushed;
534 /* SD_BALANCE_FORK stats */
535 unsigned long sbf_cnt;
536 unsigned long sbf_balanced;
537 unsigned long sbf_pushed;
539 /* try_to_wake_up() stats */
540 unsigned long ttwu_wake_remote;
541 unsigned long ttwu_move_affine;
542 unsigned long ttwu_move_balance;
546 extern void partition_sched_domains(cpumask_t *partition1,
547 cpumask_t *partition2);
548 #ifdef ARCH_HAS_SCHED_DOMAIN
549 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
550 extern cpumask_t cpu_isolated_map;
551 extern void init_sched_build_groups(struct sched_group groups[],
552 cpumask_t span, int (*group_fn)(int cpu));
553 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
554 #endif /* ARCH_HAS_SCHED_DOMAIN */
555 #endif /* CONFIG_SMP */
558 struct io_context; /* See blkdev.h */
559 void exit_io_context(void);
562 #define NGROUPS_SMALL 32
563 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
567 gid_t small_block[NGROUPS_SMALL];
573 * get_group_info() must be called with the owning task locked (via task_lock())
574 * when task != current. The reason being that the vast majority of callers are
575 * looking at current->group_info, which can not be changed except by the
576 * current task. Changing current->group_info requires the task lock, too.
578 #define get_group_info(group_info) do { \
579 atomic_inc(&(group_info)->usage); \
582 #define put_group_info(group_info) do { \
583 if (atomic_dec_and_test(&(group_info)->usage)) \
584 groups_free(group_info); \
587 extern struct group_info *groups_alloc(int gidsetsize);
588 extern void groups_free(struct group_info *group_info);
589 extern int set_current_groups(struct group_info *group_info);
590 extern int groups_search(struct group_info *group_info, gid_t grp);
591 /* access the groups "array" with this macro */
592 #define GROUP_AT(gi, i) \
593 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
596 struct audit_context; /* See audit.c */
600 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
601 struct thread_info *thread_info;
603 unsigned long flags; /* per process flags, defined below */
604 unsigned long ptrace;
606 int lock_depth; /* BKL lock depth */
608 #if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
611 int prio, static_prio;
612 struct list_head run_list;
615 unsigned short ioprio;
617 unsigned long sleep_avg;
618 unsigned long long timestamp, last_ran;
619 unsigned long long sched_time; /* sched_clock time spent running */
622 unsigned long policy;
623 cpumask_t cpus_allowed;
624 unsigned int time_slice, first_time_slice;
626 #ifdef CONFIG_SCHEDSTATS
627 struct sched_info sched_info;
630 struct list_head tasks;
632 * ptrace_list/ptrace_children forms the list of my children
633 * that were stolen by a ptracer.
635 struct list_head ptrace_children;
636 struct list_head ptrace_list;
638 struct mm_struct *mm, *active_mm;
641 struct linux_binfmt *binfmt;
643 int exit_code, exit_signal;
644 int pdeath_signal; /* The signal sent when the parent dies */
646 unsigned long personality;
651 * pointers to (original) parent process, youngest child, younger sibling,
652 * older sibling, respectively. (p->father can be replaced with
655 struct task_struct *real_parent; /* real parent process (when being debugged) */
656 struct task_struct *parent; /* parent process */
658 * children/sibling forms the list of my children plus the
659 * tasks I'm ptracing.
661 struct list_head children; /* list of my children */
662 struct list_head sibling; /* linkage in my parent's children list */
663 struct task_struct *group_leader; /* threadgroup leader */
665 /* PID/PID hash table linkage. */
666 struct pid pids[PIDTYPE_MAX];
668 struct completion *vfork_done; /* for vfork() */
669 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
670 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
672 unsigned long rt_priority;
673 cputime_t utime, stime;
674 unsigned long nvcsw, nivcsw; /* context switch counts */
675 struct timespec start_time;
676 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
677 unsigned long min_flt, maj_flt;
679 cputime_t it_prof_expires, it_virt_expires;
680 unsigned long long it_sched_expires;
681 struct list_head cpu_timers[3];
683 /* process credentials */
684 uid_t uid,euid,suid,fsuid;
685 gid_t gid,egid,sgid,fsgid;
686 struct group_info *group_info;
687 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
688 unsigned keep_capabilities:1;
689 struct user_struct *user;
691 struct key *thread_keyring; /* keyring private to this thread */
692 unsigned char jit_keyring; /* default keyring to attach requested keys to */
694 int oomkilladj; /* OOM kill score adjustment (bit shift). */
695 char comm[TASK_COMM_LEN]; /* executable name excluding path
696 - access with [gs]et_task_comm (which lock
698 - initialized normally by flush_old_exec */
699 /* file system info */
700 int link_count, total_link_count;
702 struct sysv_sem sysvsem;
703 /* CPU-specific state of this task */
704 struct thread_struct thread;
705 /* filesystem information */
706 struct fs_struct *fs;
707 /* open file information */
708 struct files_struct *files;
710 struct namespace *namespace;
711 /* signal handlers */
712 struct signal_struct *signal;
713 struct sighand_struct *sighand;
715 sigset_t blocked, real_blocked;
716 struct sigpending pending;
718 unsigned long sas_ss_sp;
720 int (*notifier)(void *priv);
722 sigset_t *notifier_mask;
725 struct audit_context *audit_context;
728 /* Thread group tracking */
731 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
732 spinlock_t alloc_lock;
733 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
734 spinlock_t proc_lock;
736 /* journalling filesystem info */
740 struct reclaim_state *reclaim_state;
742 struct dentry *proc_dentry;
743 struct backing_dev_info *backing_dev_info;
745 struct io_context *io_context;
747 unsigned long ptrace_message;
748 siginfo_t *last_siginfo; /* For ptrace use. */
750 * current io wait handle: wait queue entry to use for io waits
751 * If this thread is processing aio, this points at the waitqueue
752 * inside the currently handled kiocb. It may be NULL (i.e. default
753 * to a stack based synchronous wait) if its doing sync IO.
755 wait_queue_t *io_wait;
756 /* i/o counters(bytes read/written, #syscalls */
757 u64 rchar, wchar, syscr, syscw;
758 #if defined(CONFIG_BSD_PROCESS_ACCT)
759 u64 acct_rss_mem1; /* accumulated rss usage */
760 u64 acct_vm_mem1; /* accumulated virtual memory usage */
761 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
764 struct mempolicy *mempolicy;
767 #ifdef CONFIG_CPUSETS
768 struct cpuset *cpuset;
769 nodemask_t mems_allowed;
770 int cpuset_mems_generation;
772 atomic_t fs_excl; /* holding fs exclusive resources */
775 static inline pid_t process_group(struct task_struct *tsk)
777 return tsk->signal->pgrp;
781 * pid_alive - check that a task structure is not stale
782 * @p: Task structure to be checked.
784 * Test if a process is not yet dead (at most zombie state)
785 * If pid_alive fails, then pointers within the task structure
786 * can be stale and must not be dereferenced.
788 static inline int pid_alive(struct task_struct *p)
790 return p->pids[PIDTYPE_PID].nr != 0;
793 extern void free_task(struct task_struct *tsk);
794 extern void __put_task_struct(struct task_struct *tsk);
795 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
796 #define put_task_struct(tsk) \
797 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
802 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
803 /* Not implemented yet, only for 486*/
804 #define PF_STARTING 0x00000002 /* being created */
805 #define PF_EXITING 0x00000004 /* getting shut down */
806 #define PF_DEAD 0x00000008 /* Dead */
807 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
808 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
809 #define PF_DUMPCORE 0x00000200 /* dumped core */
810 #define PF_SIGNALED 0x00000400 /* killed by a signal */
811 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
812 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
813 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
814 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
815 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
816 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
817 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
818 #define PF_KSWAPD 0x00040000 /* I am kswapd */
819 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
820 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
821 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
822 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
823 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
826 * Only the _current_ task can read/write to tsk->flags, but other
827 * tasks can access tsk->flags in readonly mode for example
828 * with tsk_used_math (like during threaded core dumping).
829 * There is however an exception to this rule during ptrace
830 * or during fork: the ptracer task is allowed to write to the
831 * child->flags of its traced child (same goes for fork, the parent
832 * can write to the child->flags), because we're guaranteed the
833 * child is not running and in turn not changing child->flags
834 * at the same time the parent does it.
836 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
837 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
838 #define clear_used_math() clear_stopped_child_used_math(current)
839 #define set_used_math() set_stopped_child_used_math(current)
840 #define conditional_stopped_child_used_math(condition, child) \
841 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
842 #define conditional_used_math(condition) \
843 conditional_stopped_child_used_math(condition, current)
844 #define copy_to_stopped_child_used_math(child) \
845 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
846 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
847 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
848 #define used_math() tsk_used_math(current)
851 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
853 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
855 if (!cpus_intersects(new_mask, cpu_online_map))
861 extern unsigned long long sched_clock(void);
862 extern unsigned long long current_sched_time(const task_t *current_task);
864 /* sched_exec is called by processes performing an exec */
866 extern void sched_exec(void);
868 #define sched_exec() {}
871 #ifdef CONFIG_HOTPLUG_CPU
872 extern void idle_task_exit(void);
874 static inline void idle_task_exit(void) {}
877 extern void sched_idle_next(void);
878 extern void set_user_nice(task_t *p, long nice);
879 extern int task_prio(const task_t *p);
880 extern int task_nice(const task_t *p);
881 extern int can_nice(const task_t *p, const int nice);
882 extern int task_curr(const task_t *p);
883 extern int idle_cpu(int cpu);
884 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
885 extern task_t *idle_task(int cpu);
890 * The default (Linux) execution domain.
892 extern struct exec_domain default_exec_domain;
895 struct thread_info thread_info;
896 unsigned long stack[THREAD_SIZE/sizeof(long)];
899 #ifndef __HAVE_ARCH_KSTACK_END
900 static inline int kstack_end(void *addr)
902 /* Reliable end of stack detection:
903 * Some APM bios versions misalign the stack
905 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
909 extern union thread_union init_thread_union;
910 extern struct task_struct init_task;
912 extern struct mm_struct init_mm;
914 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
915 extern struct task_struct *find_task_by_pid_type(int type, int pid);
916 extern void set_special_pids(pid_t session, pid_t pgrp);
917 extern void __set_special_pids(pid_t session, pid_t pgrp);
919 /* per-UID process charging. */
920 extern struct user_struct * alloc_uid(uid_t);
921 static inline struct user_struct *get_uid(struct user_struct *u)
923 atomic_inc(&u->__count);
926 extern void free_uid(struct user_struct *);
927 extern void switch_uid(struct user_struct *);
929 #include <asm/current.h>
931 extern void do_timer(struct pt_regs *);
933 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
934 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
935 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
936 unsigned long clone_flags));
938 extern void kick_process(struct task_struct *tsk);
940 static inline void kick_process(struct task_struct *tsk) { }
942 extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
943 extern void FASTCALL(sched_exit(task_t * p));
945 extern int in_group_p(gid_t);
946 extern int in_egroup_p(gid_t);
948 extern void proc_caches_init(void);
949 extern void flush_signals(struct task_struct *);
950 extern void flush_signal_handlers(struct task_struct *, int force_default);
951 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
953 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
958 spin_lock_irqsave(&tsk->sighand->siglock, flags);
959 ret = dequeue_signal(tsk, mask, info);
960 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
965 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
967 extern void unblock_all_signals(void);
968 extern void release_task(struct task_struct * p);
969 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
970 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
971 extern int force_sigsegv(int, struct task_struct *);
972 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
973 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
974 extern int kill_pg_info(int, struct siginfo *, pid_t);
975 extern int kill_proc_info(int, struct siginfo *, pid_t);
976 extern void do_notify_parent(struct task_struct *, int);
977 extern void force_sig(int, struct task_struct *);
978 extern void force_sig_specific(int, struct task_struct *);
979 extern int send_sig(int, struct task_struct *, int);
980 extern void zap_other_threads(struct task_struct *p);
981 extern int kill_pg(pid_t, int, int);
982 extern int kill_sl(pid_t, int, int);
983 extern int kill_proc(pid_t, int, int);
984 extern struct sigqueue *sigqueue_alloc(void);
985 extern void sigqueue_free(struct sigqueue *);
986 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
987 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
988 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
989 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
991 /* These can be the second arg to send_sig_info/send_group_sig_info. */
992 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
993 #define SEND_SIG_PRIV ((struct siginfo *) 1)
994 #define SEND_SIG_FORCED ((struct siginfo *) 2)
996 /* True if we are on the alternate signal stack. */
998 static inline int on_sig_stack(unsigned long sp)
1000 return (sp - current->sas_ss_sp < current->sas_ss_size);
1003 static inline int sas_ss_flags(unsigned long sp)
1005 return (current->sas_ss_size == 0 ? SS_DISABLE
1006 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1010 #ifdef CONFIG_SECURITY
1011 /* code is in security.c */
1012 extern int capable(int cap);
1014 static inline int capable(int cap)
1016 if (cap_raised(current->cap_effective, cap)) {
1017 current->flags |= PF_SUPERPRIV;
1025 * Routines for handling mm_structs
1027 extern struct mm_struct * mm_alloc(void);
1029 /* mmdrop drops the mm and the page tables */
1030 extern void FASTCALL(__mmdrop(struct mm_struct *));
1031 static inline void mmdrop(struct mm_struct * mm)
1033 if (atomic_dec_and_test(&mm->mm_count))
1037 /* mmput gets rid of the mappings and all user-space */
1038 extern void mmput(struct mm_struct *);
1039 /* Grab a reference to a task's mm, if it is not already going away */
1040 extern struct mm_struct *get_task_mm(struct task_struct *task);
1041 /* Remove the current tasks stale references to the old mm_struct */
1042 extern void mm_release(struct task_struct *, struct mm_struct *);
1044 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1045 extern void flush_thread(void);
1046 extern void exit_thread(void);
1048 extern void exit_files(struct task_struct *);
1049 extern void exit_signal(struct task_struct *);
1050 extern void __exit_signal(struct task_struct *);
1051 extern void exit_sighand(struct task_struct *);
1052 extern void __exit_sighand(struct task_struct *);
1053 extern void exit_itimers(struct signal_struct *);
1055 extern NORET_TYPE void do_group_exit(int);
1057 extern void daemonize(const char *, ...);
1058 extern int allow_signal(int);
1059 extern int disallow_signal(int);
1060 extern task_t *child_reaper;
1062 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1063 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1064 task_t *fork_idle(int);
1066 extern void set_task_comm(struct task_struct *tsk, char *from);
1067 extern void get_task_comm(char *to, struct task_struct *tsk);
1070 extern void wait_task_inactive(task_t * p);
1072 #define wait_task_inactive(p) do { } while (0)
1075 #define remove_parent(p) list_del_init(&(p)->sibling)
1076 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1078 #define REMOVE_LINKS(p) do { \
1079 if (thread_group_leader(p)) \
1080 list_del_init(&(p)->tasks); \
1084 #define SET_LINKS(p) do { \
1085 if (thread_group_leader(p)) \
1086 list_add_tail(&(p)->tasks,&init_task.tasks); \
1087 add_parent(p, (p)->parent); \
1090 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1091 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1093 #define for_each_process(p) \
1094 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1097 * Careful: do_each_thread/while_each_thread is a double loop so
1098 * 'break' will not work as expected - use goto instead.
1100 #define do_each_thread(g, t) \
1101 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1103 #define while_each_thread(g, t) \
1104 while ((t = next_thread(t)) != g)
1106 extern task_t * FASTCALL(next_thread(const task_t *p));
1108 #define thread_group_leader(p) (p->pid == p->tgid)
1110 static inline int thread_group_empty(task_t *p)
1112 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1115 #define delay_group_leader(p) \
1116 (thread_group_leader(p) && !thread_group_empty(p))
1118 extern void unhash_process(struct task_struct *p);
1121 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1122 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1123 * pins the final release of task.io_context.
1125 * Nests both inside and outside of read_lock(&tasklist_lock).
1126 * It must not be nested with write_lock_irq(&tasklist_lock),
1127 * neither inside nor outside.
1129 static inline void task_lock(struct task_struct *p)
1131 spin_lock(&p->alloc_lock);
1134 static inline void task_unlock(struct task_struct *p)
1136 spin_unlock(&p->alloc_lock);
1139 /* set thread flags in other task's structures
1140 * - see asm/thread_info.h for TIF_xxxx flags available
1142 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1144 set_ti_thread_flag(tsk->thread_info,flag);
1147 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1149 clear_ti_thread_flag(tsk->thread_info,flag);
1152 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1154 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1157 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1159 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1162 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1164 return test_ti_thread_flag(tsk->thread_info,flag);
1167 static inline void set_tsk_need_resched(struct task_struct *tsk)
1169 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1172 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1174 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1177 static inline int signal_pending(struct task_struct *p)
1179 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1182 static inline int need_resched(void)
1184 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1188 * cond_resched() and cond_resched_lock(): latency reduction via
1189 * explicit rescheduling in places that are safe. The return
1190 * value indicates whether a reschedule was done in fact.
1191 * cond_resched_lock() will drop the spinlock before scheduling,
1192 * cond_resched_softirq() will enable bhs before scheduling.
1194 extern int cond_resched(void);
1195 extern int cond_resched_lock(spinlock_t * lock);
1196 extern int cond_resched_softirq(void);
1199 * Does a critical section need to be broken due to another
1202 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1203 # define need_lockbreak(lock) ((lock)->break_lock)
1205 # define need_lockbreak(lock) 0
1209 * Does a critical section need to be broken due to another
1210 * task waiting or preemption being signalled:
1212 static inline int lock_need_resched(spinlock_t *lock)
1214 if (need_lockbreak(lock) || need_resched())
1219 /* Reevaluate whether the task has signals pending delivery.
1220 This is required every time the blocked sigset_t changes.
1221 callers must hold sighand->siglock. */
1223 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1224 extern void recalc_sigpending(void);
1226 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1229 * Wrappers for p->thread_info->cpu access. No-op on UP.
1233 static inline unsigned int task_cpu(const struct task_struct *p)
1235 return p->thread_info->cpu;
1238 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1240 p->thread_info->cpu = cpu;
1245 static inline unsigned int task_cpu(const struct task_struct *p)
1250 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1254 #endif /* CONFIG_SMP */
1256 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1257 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1259 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1261 mm->mmap_base = TASK_UNMAPPED_BASE;
1262 mm->get_unmapped_area = arch_get_unmapped_area;
1263 mm->unmap_area = arch_unmap_area;
1267 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1268 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1270 #ifdef CONFIG_MAGIC_SYSRQ
1272 extern void normalize_rt_tasks(void);
1278 * Check if a process has been frozen
1280 static inline int frozen(struct task_struct *p)
1282 return p->flags & PF_FROZEN;
1286 * Check if there is a request to freeze a process
1288 static inline int freezing(struct task_struct *p)
1290 return p->flags & PF_FREEZE;
1294 * Request that a process be frozen
1295 * FIXME: SMP problem. We may not modify other process' flags!
1297 static inline void freeze(struct task_struct *p)
1299 p->flags |= PF_FREEZE;
1303 * Wake up a frozen process
1305 static inline int thaw_process(struct task_struct *p)
1308 p->flags &= ~PF_FROZEN;
1316 * freezing is complete, mark process as frozen
1318 static inline void frozen_process(struct task_struct *p)
1320 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1323 extern void refrigerator(void);
1324 extern int freeze_processes(void);
1325 extern void thaw_processes(void);
1327 static inline int try_to_freeze(void)
1329 if (freezing(current)) {
1336 static inline int frozen(struct task_struct *p) { return 0; }
1337 static inline int freezing(struct task_struct *p) { return 0; }
1338 static inline void freeze(struct task_struct *p) { BUG(); }
1339 static inline int thaw_process(struct task_struct *p) { return 1; }
1340 static inline void frozen_process(struct task_struct *p) { BUG(); }
1342 static inline void refrigerator(void) {}
1343 static inline int freeze_processes(void) { BUG(); return 0; }
1344 static inline void thaw_processes(void) {}
1346 static inline int try_to_freeze(void) { return 0; }
1348 #endif /* CONFIG_PM */
1349 #endif /* __KERNEL__ */