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>
38 #include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
45 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
46 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
47 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
48 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
49 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
50 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
51 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
52 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
53 #define CLONE_THREAD 0x00010000 /* Same thread group? */
54 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
55 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
56 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
57 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
58 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
59 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
60 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
61 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
62 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
65 * List of flags we want to share for kernel threads,
66 * if only because they are not used by them anyway.
68 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
71 * These are the constant used to fake the fixed-point load-average
72 * counting. Some notes:
73 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
74 * a load-average precision of 10 bits integer + 11 bits fractional
75 * - if you want to count load-averages more often, you need more
76 * precision, or rounding will get you. With 2-second counting freq,
77 * the EXP_n values would be 1981, 2034 and 2043 if still using only
80 extern unsigned long avenrun[]; /* Load averages */
82 #define FSHIFT 11 /* nr of bits of precision */
83 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
84 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
85 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
86 #define EXP_5 2014 /* 1/exp(5sec/5min) */
87 #define EXP_15 2037 /* 1/exp(5sec/15min) */
89 #define CALC_LOAD(load,exp,n) \
91 load += n*(FIXED_1-exp); \
94 extern unsigned long total_forks;
95 extern int nr_threads;
97 DECLARE_PER_CPU(unsigned long, process_counts);
98 extern int nr_processes(void);
99 extern unsigned long nr_running(void);
100 extern unsigned long nr_uninterruptible(void);
101 extern unsigned long nr_iowait(void);
103 #include <linux/time.h>
104 #include <linux/param.h>
105 #include <linux/resource.h>
106 #include <linux/timer.h>
108 #include <asm/processor.h>
110 #define TASK_RUNNING 0
111 #define TASK_INTERRUPTIBLE 1
112 #define TASK_UNINTERRUPTIBLE 2
113 #define TASK_STOPPED 4
114 #define TASK_TRACED 8
115 #define EXIT_ZOMBIE 16
118 #define __set_task_state(tsk, state_value) \
119 do { (tsk)->state = (state_value); } while (0)
120 #define set_task_state(tsk, state_value) \
121 set_mb((tsk)->state, (state_value))
123 #define __set_current_state(state_value) \
124 do { current->state = (state_value); } while (0)
125 #define set_current_state(state_value) \
126 set_mb(current->state, (state_value))
128 /* Task command name length */
129 #define TASK_COMM_LEN 16
132 * Scheduling policies
134 #define SCHED_NORMAL 0
144 #include <linux/spinlock.h>
147 * This serializes "schedule()" and also protects
148 * the run-queue from deletions/modifications (but
149 * _adding_ to the beginning of the run-queue has
152 extern rwlock_t tasklist_lock;
153 extern spinlock_t mmlist_lock;
155 typedef struct task_struct task_t;
157 extern void sched_init(void);
158 extern void sched_init_smp(void);
159 extern void init_idle(task_t *idle, int cpu);
161 extern cpumask_t nohz_cpu_mask;
163 extern void show_state(void);
164 extern void show_regs(struct pt_regs *);
167 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
168 * task), SP is the stack pointer of the first frame that should be shown in the back
169 * trace (or NULL if the entire call-chain of the task should be shown).
171 extern void show_stack(struct task_struct *task, unsigned long *sp);
173 void io_schedule(void);
174 long io_schedule_timeout(long timeout);
176 extern void cpu_init (void);
177 extern void trap_init(void);
178 extern void update_process_times(int user);
179 extern void scheduler_tick(void);
181 #ifdef CONFIG_DETECT_SOFTLOCKUP
182 extern void softlockup_tick(struct pt_regs *regs);
183 extern void spawn_softlockup_task(void);
184 extern void touch_softlockup_watchdog(void);
186 static inline void softlockup_tick(struct pt_regs *regs)
189 static inline void spawn_softlockup_task(void)
192 static inline void touch_softlockup_watchdog(void)
198 /* Attach to any functions which should be ignored in wchan output. */
199 #define __sched __attribute__((__section__(".sched.text")))
200 /* Is this address in the __sched functions? */
201 extern int in_sched_functions(unsigned long addr);
203 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
204 extern signed long FASTCALL(schedule_timeout(signed long timeout));
205 asmlinkage void schedule(void);
209 /* Maximum number of active map areas.. This is a random (large) number */
210 #define DEFAULT_MAX_MAP_COUNT 65536
212 extern int sysctl_max_map_count;
214 #include <linux/aio.h>
217 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
218 unsigned long, unsigned long);
220 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
221 unsigned long len, unsigned long pgoff,
222 unsigned long flags);
223 extern void arch_unmap_area(struct mm_struct *, unsigned long);
224 extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
226 #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
227 #define get_mm_counter(mm, member) ((mm)->_##member)
228 #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
229 #define inc_mm_counter(mm, member) (mm)->_##member++
230 #define dec_mm_counter(mm, member) (mm)->_##member--
231 typedef unsigned long mm_counter_t;
234 struct vm_area_struct * mmap; /* list of VMAs */
235 struct rb_root mm_rb;
236 struct vm_area_struct * mmap_cache; /* last find_vma result */
237 unsigned long (*get_unmapped_area) (struct file *filp,
238 unsigned long addr, unsigned long len,
239 unsigned long pgoff, unsigned long flags);
240 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
241 unsigned long mmap_base; /* base of mmap area */
242 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
243 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
245 atomic_t mm_users; /* How many users with user space? */
246 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
247 int map_count; /* number of VMAs */
248 struct rw_semaphore mmap_sem;
249 spinlock_t page_table_lock; /* Protects page tables and some counters */
251 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
252 * together off init_mm.mmlist, and are protected
256 unsigned long start_code, end_code, start_data, end_data;
257 unsigned long start_brk, brk, start_stack;
258 unsigned long arg_start, arg_end, env_start, env_end;
259 unsigned long total_vm, locked_vm, shared_vm;
260 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
262 /* Special counters protected by the page_table_lock */
264 mm_counter_t _anon_rss;
266 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
269 cpumask_t cpu_vm_mask;
271 /* Architecture-specific MM context */
272 mm_context_t context;
274 /* Token based thrashing protection. */
275 unsigned long swap_token_time;
278 /* coredumping support */
280 struct completion *core_startup_done, core_done;
283 rwlock_t ioctx_list_lock;
284 struct kioctx *ioctx_list;
286 struct kioctx default_kioctx;
288 unsigned long hiwater_rss; /* High-water RSS usage */
289 unsigned long hiwater_vm; /* High-water virtual memory usage */
292 struct sighand_struct {
294 struct k_sigaction action[_NSIG];
299 * NOTE! "signal_struct" does not have it's own
300 * locking, because a shared signal_struct always
301 * implies a shared sighand_struct, so locking
302 * sighand_struct is always a proper superset of
303 * the locking of signal_struct.
305 struct signal_struct {
309 wait_queue_head_t wait_chldexit; /* for wait4() */
311 /* current thread group signal load-balancing target: */
314 /* shared signal handling: */
315 struct sigpending shared_pending;
317 /* thread group exit support */
320 * - notify group_exit_task when ->count is equal to notify_count
321 * - everyone except group_exit_task is stopped during signal delivery
322 * of fatal signals, group_exit_task processes the signal.
324 struct task_struct *group_exit_task;
327 /* thread group stop support, overloads group_exit_code too */
328 int group_stop_count;
329 unsigned int flags; /* see SIGNAL_* flags below */
331 /* POSIX.1b Interval Timers */
332 struct list_head posix_timers;
334 /* ITIMER_REAL timer for the process */
335 struct timer_list real_timer;
336 unsigned long it_real_value, it_real_incr;
338 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
339 cputime_t it_prof_expires, it_virt_expires;
340 cputime_t it_prof_incr, it_virt_incr;
342 /* job control IDs */
346 /* boolean value for session group leader */
349 struct tty_struct *tty; /* NULL if no tty */
352 * Cumulative resource counters for dead threads in the group,
353 * and for reaped dead child processes forked by this group.
354 * Live threads maintain their own counters and add to these
355 * in __exit_signal, except for the group leader.
357 cputime_t utime, stime, cutime, cstime;
358 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
359 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
362 * Cumulative ns of scheduled CPU time for dead threads in the
363 * group, not including a zombie group leader. (This only differs
364 * from jiffies_to_ns(utime + stime) if sched_clock uses something
365 * other than jiffies.)
367 unsigned long long sched_time;
370 * We don't bother to synchronize most readers of this at all,
371 * because there is no reader checking a limit that actually needs
372 * to get both rlim_cur and rlim_max atomically, and either one
373 * alone is a single word that can safely be read normally.
374 * getrlimit/setrlimit use task_lock(current->group_leader) to
375 * protect this instead of the siglock, because they really
376 * have no need to disable irqs.
378 struct rlimit rlim[RLIM_NLIMITS];
380 struct list_head cpu_timers[3];
382 /* keep the process-shared keyrings here so that they do the right
383 * thing in threads created with CLONE_THREAD */
385 struct key *session_keyring; /* keyring inherited over fork */
386 struct key *process_keyring; /* keyring private to this process */
390 /* Context switch must be unlocked if interrupts are to be enabled */
391 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
392 # define __ARCH_WANT_UNLOCKED_CTXSW
396 * Bits in flags field of signal_struct.
398 #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
399 #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
400 #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
401 #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
405 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
406 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
407 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
408 * are inverted: lower p->prio value means higher priority.
410 * The MAX_USER_RT_PRIO value allows the actual maximum
411 * RT priority to be separate from the value exported to
412 * user-space. This allows kernel threads to set their
413 * priority to a value higher than any user task. Note:
414 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
417 #define MAX_USER_RT_PRIO 100
418 #define MAX_RT_PRIO MAX_USER_RT_PRIO
420 #define MAX_PRIO (MAX_RT_PRIO + 40)
422 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
425 * Some day this will be a full-fledged user tracking system..
428 atomic_t __count; /* reference count */
429 atomic_t processes; /* How many processes does this user have? */
430 atomic_t files; /* How many open files does this user have? */
431 atomic_t sigpending; /* How many pending signals does this user have? */
432 #ifdef CONFIG_INOTIFY
433 atomic_t inotify_watches; /* How many inotify watches does this user have? */
434 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
436 /* protected by mq_lock */
437 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
438 unsigned long locked_shm; /* How many pages of mlocked shm ? */
441 struct key *uid_keyring; /* UID specific keyring */
442 struct key *session_keyring; /* UID's default session keyring */
445 /* Hash table maintenance information */
446 struct list_head uidhash_list;
450 extern struct user_struct *find_user(uid_t);
452 extern struct user_struct root_user;
453 #define INIT_USER (&root_user)
455 typedef struct prio_array prio_array_t;
456 struct backing_dev_info;
457 struct reclaim_state;
459 #ifdef CONFIG_SCHEDSTATS
461 /* cumulative counters */
462 unsigned long cpu_time, /* time spent on the cpu */
463 run_delay, /* time spent waiting on a runqueue */
464 pcnt; /* # of timeslices run on this cpu */
467 unsigned long last_arrival, /* when we last ran on a cpu */
468 last_queued; /* when we were last queued to run */
471 extern struct file_operations proc_schedstat_operations;
483 * sched-domains (multiprocessor balancing) declarations:
486 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
488 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
489 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
490 #define SD_BALANCE_EXEC 4 /* Balance on exec */
491 #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
492 #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
493 #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
494 #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
495 #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
498 struct sched_group *next; /* Must be a circular list */
502 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
503 * single CPU. This is read only (except for setup, hotplug CPU).
505 unsigned long cpu_power;
508 struct sched_domain {
509 /* These fields must be setup */
510 struct sched_domain *parent; /* top domain must be null terminated */
511 struct sched_group *groups; /* the balancing groups of the domain */
512 cpumask_t span; /* span of all CPUs in this domain */
513 unsigned long min_interval; /* Minimum balance interval ms */
514 unsigned long max_interval; /* Maximum balance interval ms */
515 unsigned int busy_factor; /* less balancing by factor if busy */
516 unsigned int imbalance_pct; /* No balance until over watermark */
517 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
518 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
519 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
520 unsigned int busy_idx;
521 unsigned int idle_idx;
522 unsigned int newidle_idx;
523 unsigned int wake_idx;
524 unsigned int forkexec_idx;
525 int flags; /* See SD_* */
527 /* Runtime fields. */
528 unsigned long last_balance; /* init to jiffies. units in jiffies */
529 unsigned int balance_interval; /* initialise to 1. units in ms. */
530 unsigned int nr_balance_failed; /* initialise to 0 */
532 #ifdef CONFIG_SCHEDSTATS
533 /* load_balance() stats */
534 unsigned long lb_cnt[MAX_IDLE_TYPES];
535 unsigned long lb_failed[MAX_IDLE_TYPES];
536 unsigned long lb_balanced[MAX_IDLE_TYPES];
537 unsigned long lb_imbalance[MAX_IDLE_TYPES];
538 unsigned long lb_gained[MAX_IDLE_TYPES];
539 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
540 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
541 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
543 /* Active load balancing */
544 unsigned long alb_cnt;
545 unsigned long alb_failed;
546 unsigned long alb_pushed;
548 /* SD_BALANCE_EXEC stats */
549 unsigned long sbe_cnt;
550 unsigned long sbe_balanced;
551 unsigned long sbe_pushed;
553 /* SD_BALANCE_FORK stats */
554 unsigned long sbf_cnt;
555 unsigned long sbf_balanced;
556 unsigned long sbf_pushed;
558 /* try_to_wake_up() stats */
559 unsigned long ttwu_wake_remote;
560 unsigned long ttwu_move_affine;
561 unsigned long ttwu_move_balance;
565 extern void partition_sched_domains(cpumask_t *partition1,
566 cpumask_t *partition2);
567 #endif /* CONFIG_SMP */
570 struct io_context; /* See blkdev.h */
571 void exit_io_context(void);
574 #define NGROUPS_SMALL 32
575 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
579 gid_t small_block[NGROUPS_SMALL];
585 * get_group_info() must be called with the owning task locked (via task_lock())
586 * when task != current. The reason being that the vast majority of callers are
587 * looking at current->group_info, which can not be changed except by the
588 * current task. Changing current->group_info requires the task lock, too.
590 #define get_group_info(group_info) do { \
591 atomic_inc(&(group_info)->usage); \
594 #define put_group_info(group_info) do { \
595 if (atomic_dec_and_test(&(group_info)->usage)) \
596 groups_free(group_info); \
599 extern struct group_info *groups_alloc(int gidsetsize);
600 extern void groups_free(struct group_info *group_info);
601 extern int set_current_groups(struct group_info *group_info);
602 extern int groups_search(struct group_info *group_info, gid_t grp);
603 /* access the groups "array" with this macro */
604 #define GROUP_AT(gi, i) \
605 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
608 struct audit_context; /* See audit.c */
612 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
613 struct thread_info *thread_info;
615 unsigned long flags; /* per process flags, defined below */
616 unsigned long ptrace;
618 int lock_depth; /* BKL lock depth */
620 #if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
623 int prio, static_prio;
624 struct list_head run_list;
627 unsigned short ioprio;
629 unsigned long sleep_avg;
630 unsigned long long timestamp, last_ran;
631 unsigned long long sched_time; /* sched_clock time spent running */
634 unsigned long policy;
635 cpumask_t cpus_allowed;
636 unsigned int time_slice, first_time_slice;
638 #ifdef CONFIG_SCHEDSTATS
639 struct sched_info sched_info;
642 struct list_head tasks;
644 * ptrace_list/ptrace_children forms the list of my children
645 * that were stolen by a ptracer.
647 struct list_head ptrace_children;
648 struct list_head ptrace_list;
650 struct mm_struct *mm, *active_mm;
653 struct linux_binfmt *binfmt;
655 int exit_code, exit_signal;
656 int pdeath_signal; /* The signal sent when the parent dies */
658 unsigned long personality;
663 * pointers to (original) parent process, youngest child, younger sibling,
664 * older sibling, respectively. (p->father can be replaced with
667 struct task_struct *real_parent; /* real parent process (when being debugged) */
668 struct task_struct *parent; /* parent process */
670 * children/sibling forms the list of my children plus the
671 * tasks I'm ptracing.
673 struct list_head children; /* list of my children */
674 struct list_head sibling; /* linkage in my parent's children list */
675 struct task_struct *group_leader; /* threadgroup leader */
677 /* PID/PID hash table linkage. */
678 struct pid pids[PIDTYPE_MAX];
680 struct completion *vfork_done; /* for vfork() */
681 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
682 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
684 unsigned long rt_priority;
685 cputime_t utime, stime;
686 unsigned long nvcsw, nivcsw; /* context switch counts */
687 struct timespec start_time;
688 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
689 unsigned long min_flt, maj_flt;
691 cputime_t it_prof_expires, it_virt_expires;
692 unsigned long long it_sched_expires;
693 struct list_head cpu_timers[3];
695 /* process credentials */
696 uid_t uid,euid,suid,fsuid;
697 gid_t gid,egid,sgid,fsgid;
698 struct group_info *group_info;
699 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
700 unsigned keep_capabilities:1;
701 struct user_struct *user;
703 struct key *thread_keyring; /* keyring private to this thread */
704 unsigned char jit_keyring; /* default keyring to attach requested keys to */
706 int oomkilladj; /* OOM kill score adjustment (bit shift). */
707 char comm[TASK_COMM_LEN]; /* executable name excluding path
708 - access with [gs]et_task_comm (which lock
710 - initialized normally by flush_old_exec */
711 /* file system info */
712 int link_count, total_link_count;
714 struct sysv_sem sysvsem;
715 /* CPU-specific state of this task */
716 struct thread_struct thread;
717 /* filesystem information */
718 struct fs_struct *fs;
719 /* open file information */
720 struct files_struct *files;
722 struct namespace *namespace;
723 /* signal handlers */
724 struct signal_struct *signal;
725 struct sighand_struct *sighand;
727 sigset_t blocked, real_blocked;
728 struct sigpending pending;
730 unsigned long sas_ss_sp;
732 int (*notifier)(void *priv);
734 sigset_t *notifier_mask;
737 struct audit_context *audit_context;
740 /* Thread group tracking */
743 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
744 spinlock_t alloc_lock;
745 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
746 spinlock_t proc_lock;
748 /* journalling filesystem info */
752 struct reclaim_state *reclaim_state;
754 struct dentry *proc_dentry;
755 struct backing_dev_info *backing_dev_info;
757 struct io_context *io_context;
759 unsigned long ptrace_message;
760 siginfo_t *last_siginfo; /* For ptrace use. */
762 * current io wait handle: wait queue entry to use for io waits
763 * If this thread is processing aio, this points at the waitqueue
764 * inside the currently handled kiocb. It may be NULL (i.e. default
765 * to a stack based synchronous wait) if its doing sync IO.
767 wait_queue_t *io_wait;
768 /* i/o counters(bytes read/written, #syscalls */
769 u64 rchar, wchar, syscr, syscw;
770 #if defined(CONFIG_BSD_PROCESS_ACCT)
771 u64 acct_rss_mem1; /* accumulated rss usage */
772 u64 acct_vm_mem1; /* accumulated virtual memory usage */
773 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
776 struct mempolicy *mempolicy;
779 #ifdef CONFIG_CPUSETS
780 struct cpuset *cpuset;
781 nodemask_t mems_allowed;
782 int cpuset_mems_generation;
784 atomic_t fs_excl; /* holding fs exclusive resources */
787 static inline pid_t process_group(struct task_struct *tsk)
789 return tsk->signal->pgrp;
793 * pid_alive - check that a task structure is not stale
794 * @p: Task structure to be checked.
796 * Test if a process is not yet dead (at most zombie state)
797 * If pid_alive fails, then pointers within the task structure
798 * can be stale and must not be dereferenced.
800 static inline int pid_alive(struct task_struct *p)
802 return p->pids[PIDTYPE_PID].nr != 0;
805 extern void free_task(struct task_struct *tsk);
806 extern void __put_task_struct(struct task_struct *tsk);
807 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
808 #define put_task_struct(tsk) \
809 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
814 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
815 /* Not implemented yet, only for 486*/
816 #define PF_STARTING 0x00000002 /* being created */
817 #define PF_EXITING 0x00000004 /* getting shut down */
818 #define PF_DEAD 0x00000008 /* Dead */
819 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
820 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
821 #define PF_DUMPCORE 0x00000200 /* dumped core */
822 #define PF_SIGNALED 0x00000400 /* killed by a signal */
823 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
824 #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
825 #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
826 #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
827 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
828 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
829 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
830 #define PF_KSWAPD 0x00040000 /* I am kswapd */
831 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
832 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
833 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
834 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
835 #define PF_RANDOMIZE 0x00800000 /* randomize virtual address space */
838 * Only the _current_ task can read/write to tsk->flags, but other
839 * tasks can access tsk->flags in readonly mode for example
840 * with tsk_used_math (like during threaded core dumping).
841 * There is however an exception to this rule during ptrace
842 * or during fork: the ptracer task is allowed to write to the
843 * child->flags of its traced child (same goes for fork, the parent
844 * can write to the child->flags), because we're guaranteed the
845 * child is not running and in turn not changing child->flags
846 * at the same time the parent does it.
848 #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
849 #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
850 #define clear_used_math() clear_stopped_child_used_math(current)
851 #define set_used_math() set_stopped_child_used_math(current)
852 #define conditional_stopped_child_used_math(condition, child) \
853 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
854 #define conditional_used_math(condition) \
855 conditional_stopped_child_used_math(condition, current)
856 #define copy_to_stopped_child_used_math(child) \
857 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
858 /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
859 #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
860 #define used_math() tsk_used_math(current)
863 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
865 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
867 if (!cpus_intersects(new_mask, cpu_online_map))
873 extern unsigned long long sched_clock(void);
874 extern unsigned long long current_sched_time(const task_t *current_task);
876 /* sched_exec is called by processes performing an exec */
878 extern void sched_exec(void);
880 #define sched_exec() {}
883 #ifdef CONFIG_HOTPLUG_CPU
884 extern void idle_task_exit(void);
886 static inline void idle_task_exit(void) {}
889 extern void sched_idle_next(void);
890 extern void set_user_nice(task_t *p, long nice);
891 extern int task_prio(const task_t *p);
892 extern int task_nice(const task_t *p);
893 extern int can_nice(const task_t *p, const int nice);
894 extern int task_curr(const task_t *p);
895 extern int idle_cpu(int cpu);
896 extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
897 extern task_t *idle_task(int cpu);
902 * The default (Linux) execution domain.
904 extern struct exec_domain default_exec_domain;
907 struct thread_info thread_info;
908 unsigned long stack[THREAD_SIZE/sizeof(long)];
911 #ifndef __HAVE_ARCH_KSTACK_END
912 static inline int kstack_end(void *addr)
914 /* Reliable end of stack detection:
915 * Some APM bios versions misalign the stack
917 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
921 extern union thread_union init_thread_union;
922 extern struct task_struct init_task;
924 extern struct mm_struct init_mm;
926 #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
927 extern struct task_struct *find_task_by_pid_type(int type, int pid);
928 extern void set_special_pids(pid_t session, pid_t pgrp);
929 extern void __set_special_pids(pid_t session, pid_t pgrp);
931 /* per-UID process charging. */
932 extern struct user_struct * alloc_uid(uid_t);
933 static inline struct user_struct *get_uid(struct user_struct *u)
935 atomic_inc(&u->__count);
938 extern void free_uid(struct user_struct *);
939 extern void switch_uid(struct user_struct *);
941 #include <asm/current.h>
943 extern void do_timer(struct pt_regs *);
945 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
946 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
947 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
948 unsigned long clone_flags));
950 extern void kick_process(struct task_struct *tsk);
952 static inline void kick_process(struct task_struct *tsk) { }
954 extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
955 extern void FASTCALL(sched_exit(task_t * p));
957 extern int in_group_p(gid_t);
958 extern int in_egroup_p(gid_t);
960 extern void proc_caches_init(void);
961 extern void flush_signals(struct task_struct *);
962 extern void flush_signal_handlers(struct task_struct *, int force_default);
963 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
965 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
970 spin_lock_irqsave(&tsk->sighand->siglock, flags);
971 ret = dequeue_signal(tsk, mask, info);
972 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
977 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
979 extern void unblock_all_signals(void);
980 extern void release_task(struct task_struct * p);
981 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
982 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
983 extern int force_sigsegv(int, struct task_struct *);
984 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
985 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
986 extern int kill_pg_info(int, struct siginfo *, pid_t);
987 extern int kill_proc_info(int, struct siginfo *, pid_t);
988 extern void do_notify_parent(struct task_struct *, int);
989 extern void force_sig(int, struct task_struct *);
990 extern void force_sig_specific(int, struct task_struct *);
991 extern int send_sig(int, struct task_struct *, int);
992 extern void zap_other_threads(struct task_struct *p);
993 extern int kill_pg(pid_t, int, int);
994 extern int kill_sl(pid_t, int, int);
995 extern int kill_proc(pid_t, int, int);
996 extern struct sigqueue *sigqueue_alloc(void);
997 extern void sigqueue_free(struct sigqueue *);
998 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
999 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
1000 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
1001 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1003 /* These can be the second arg to send_sig_info/send_group_sig_info. */
1004 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
1005 #define SEND_SIG_PRIV ((struct siginfo *) 1)
1006 #define SEND_SIG_FORCED ((struct siginfo *) 2)
1008 /* True if we are on the alternate signal stack. */
1010 static inline int on_sig_stack(unsigned long sp)
1012 return (sp - current->sas_ss_sp < current->sas_ss_size);
1015 static inline int sas_ss_flags(unsigned long sp)
1017 return (current->sas_ss_size == 0 ? SS_DISABLE
1018 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1022 #ifdef CONFIG_SECURITY
1023 /* code is in security.c */
1024 extern int capable(int cap);
1026 static inline int capable(int cap)
1028 if (cap_raised(current->cap_effective, cap)) {
1029 current->flags |= PF_SUPERPRIV;
1037 * Routines for handling mm_structs
1039 extern struct mm_struct * mm_alloc(void);
1041 /* mmdrop drops the mm and the page tables */
1042 extern void FASTCALL(__mmdrop(struct mm_struct *));
1043 static inline void mmdrop(struct mm_struct * mm)
1045 if (atomic_dec_and_test(&mm->mm_count))
1049 /* mmput gets rid of the mappings and all user-space */
1050 extern void mmput(struct mm_struct *);
1051 /* Grab a reference to a task's mm, if it is not already going away */
1052 extern struct mm_struct *get_task_mm(struct task_struct *task);
1053 /* Remove the current tasks stale references to the old mm_struct */
1054 extern void mm_release(struct task_struct *, struct mm_struct *);
1056 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1057 extern void flush_thread(void);
1058 extern void exit_thread(void);
1060 extern void exit_files(struct task_struct *);
1061 extern void exit_signal(struct task_struct *);
1062 extern void __exit_signal(struct task_struct *);
1063 extern void exit_sighand(struct task_struct *);
1064 extern void __exit_sighand(struct task_struct *);
1065 extern void exit_itimers(struct signal_struct *);
1067 extern NORET_TYPE void do_group_exit(int);
1069 extern void daemonize(const char *, ...);
1070 extern int allow_signal(int);
1071 extern int disallow_signal(int);
1072 extern task_t *child_reaper;
1074 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1075 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1076 task_t *fork_idle(int);
1078 extern void set_task_comm(struct task_struct *tsk, char *from);
1079 extern void get_task_comm(char *to, struct task_struct *tsk);
1082 extern void wait_task_inactive(task_t * p);
1084 #define wait_task_inactive(p) do { } while (0)
1087 #define remove_parent(p) list_del_init(&(p)->sibling)
1088 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1090 #define REMOVE_LINKS(p) do { \
1091 if (thread_group_leader(p)) \
1092 list_del_init(&(p)->tasks); \
1096 #define SET_LINKS(p) do { \
1097 if (thread_group_leader(p)) \
1098 list_add_tail(&(p)->tasks,&init_task.tasks); \
1099 add_parent(p, (p)->parent); \
1102 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1103 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1105 #define for_each_process(p) \
1106 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1109 * Careful: do_each_thread/while_each_thread is a double loop so
1110 * 'break' will not work as expected - use goto instead.
1112 #define do_each_thread(g, t) \
1113 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1115 #define while_each_thread(g, t) \
1116 while ((t = next_thread(t)) != g)
1118 extern task_t * FASTCALL(next_thread(const task_t *p));
1120 #define thread_group_leader(p) (p->pid == p->tgid)
1122 static inline int thread_group_empty(task_t *p)
1124 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1127 #define delay_group_leader(p) \
1128 (thread_group_leader(p) && !thread_group_empty(p))
1130 extern void unhash_process(struct task_struct *p);
1133 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1134 * subscriptions and synchronises with wait4(). Also used in procfs. Also
1135 * pins the final release of task.io_context.
1137 * Nests both inside and outside of read_lock(&tasklist_lock).
1138 * It must not be nested with write_lock_irq(&tasklist_lock),
1139 * neither inside nor outside.
1141 static inline void task_lock(struct task_struct *p)
1143 spin_lock(&p->alloc_lock);
1146 static inline void task_unlock(struct task_struct *p)
1148 spin_unlock(&p->alloc_lock);
1151 /* set thread flags in other task's structures
1152 * - see asm/thread_info.h for TIF_xxxx flags available
1154 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1156 set_ti_thread_flag(tsk->thread_info,flag);
1159 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1161 clear_ti_thread_flag(tsk->thread_info,flag);
1164 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1166 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1169 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1171 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1174 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1176 return test_ti_thread_flag(tsk->thread_info,flag);
1179 static inline void set_tsk_need_resched(struct task_struct *tsk)
1181 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1184 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1186 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1189 static inline int signal_pending(struct task_struct *p)
1191 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1194 static inline int need_resched(void)
1196 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1200 * cond_resched() and cond_resched_lock(): latency reduction via
1201 * explicit rescheduling in places that are safe. The return
1202 * value indicates whether a reschedule was done in fact.
1203 * cond_resched_lock() will drop the spinlock before scheduling,
1204 * cond_resched_softirq() will enable bhs before scheduling.
1206 extern int cond_resched(void);
1207 extern int cond_resched_lock(spinlock_t * lock);
1208 extern int cond_resched_softirq(void);
1211 * Does a critical section need to be broken due to another
1214 #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1215 # define need_lockbreak(lock) ((lock)->break_lock)
1217 # define need_lockbreak(lock) 0
1221 * Does a critical section need to be broken due to another
1222 * task waiting or preemption being signalled:
1224 static inline int lock_need_resched(spinlock_t *lock)
1226 if (need_lockbreak(lock) || need_resched())
1231 /* Reevaluate whether the task has signals pending delivery.
1232 This is required every time the blocked sigset_t changes.
1233 callers must hold sighand->siglock. */
1235 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1236 extern void recalc_sigpending(void);
1238 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1241 * Wrappers for p->thread_info->cpu access. No-op on UP.
1245 static inline unsigned int task_cpu(const struct task_struct *p)
1247 return p->thread_info->cpu;
1250 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1252 p->thread_info->cpu = cpu;
1257 static inline unsigned int task_cpu(const struct task_struct *p)
1262 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1266 #endif /* CONFIG_SMP */
1268 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1269 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1271 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1273 mm->mmap_base = TASK_UNMAPPED_BASE;
1274 mm->get_unmapped_area = arch_get_unmapped_area;
1275 mm->unmap_area = arch_unmap_area;
1279 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1280 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1282 #ifdef CONFIG_MAGIC_SYSRQ
1284 extern void normalize_rt_tasks(void);
1290 * Check if a process has been frozen
1292 static inline int frozen(struct task_struct *p)
1294 return p->flags & PF_FROZEN;
1298 * Check if there is a request to freeze a process
1300 static inline int freezing(struct task_struct *p)
1302 return p->flags & PF_FREEZE;
1306 * Request that a process be frozen
1307 * FIXME: SMP problem. We may not modify other process' flags!
1309 static inline void freeze(struct task_struct *p)
1311 p->flags |= PF_FREEZE;
1315 * Wake up a frozen process
1317 static inline int thaw_process(struct task_struct *p)
1320 p->flags &= ~PF_FROZEN;
1328 * freezing is complete, mark process as frozen
1330 static inline void frozen_process(struct task_struct *p)
1332 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1335 extern void refrigerator(void);
1336 extern int freeze_processes(void);
1337 extern void thaw_processes(void);
1339 static inline int try_to_freeze(void)
1341 if (freezing(current)) {
1348 static inline int frozen(struct task_struct *p) { return 0; }
1349 static inline int freezing(struct task_struct *p) { return 0; }
1350 static inline void freeze(struct task_struct *p) { BUG(); }
1351 static inline int thaw_process(struct task_struct *p) { return 1; }
1352 static inline void frozen_process(struct task_struct *p) { BUG(); }
1354 static inline void refrigerator(void) {}
1355 static inline int freeze_processes(void) { BUG(); return 0; }
1356 static inline void thaw_processes(void) {}
1358 static inline int try_to_freeze(void) { return 0; }
1360 #endif /* CONFIG_PM */
1361 #endif /* __KERNEL__ */