4 #include <asm/param.h> /* for HZ */
6 #include <linux/config.h>
7 #include <linux/capability.h>
8 #include <linux/threads.h>
9 #include <linux/kernel.h>
10 #include <linux/types.h>
11 #include <linux/timex.h>
12 #include <linux/jiffies.h>
13 #include <linux/rbtree.h>
14 #include <linux/thread_info.h>
15 #include <linux/cpumask.h>
17 #include <asm/system.h>
18 #include <asm/semaphore.h>
20 #include <asm/ptrace.h>
23 #include <linux/smp.h>
24 #include <linux/sem.h>
25 #include <linux/signal.h>
26 #include <linux/securebits.h>
27 #include <linux/fs_struct.h>
28 #include <linux/compiler.h>
29 #include <linux/completion.h>
30 #include <linux/pid.h>
31 #include <linux/percpu.h>
32 #include <linux/topology.h>
33 #include <linux/vs_base.h>
34 #include <linux/taskdelays.h>
37 extern int exec_shield;
38 extern int exec_shield_randomize;
39 extern int print_fatal_signals;
44 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
45 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
46 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
47 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
48 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
49 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
50 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
51 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
52 #define CLONE_THREAD 0x00010000 /* Same thread group? */
53 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
54 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
55 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
56 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
57 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
58 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
59 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
60 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
61 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
64 * List of flags we want to share for kernel threads,
65 * if only because they are not used by them anyway.
67 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
70 * These are the constant used to fake the fixed-point load-average
71 * counting. Some notes:
72 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
73 * a load-average precision of 10 bits integer + 11 bits fractional
74 * - if you want to count load-averages more often, you need more
75 * precision, or rounding will get you. With 2-second counting freq,
76 * the EXP_n values would be 1981, 2034 and 2043 if still using only
79 extern unsigned long avenrun[]; /* Load averages */
81 #define FSHIFT 11 /* nr of bits of precision */
82 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
83 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
84 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
85 #define EXP_5 2014 /* 1/exp(5sec/5min) */
86 #define EXP_15 2037 /* 1/exp(5sec/15min) */
88 #define CALC_LOAD(load,exp,n) \
90 load += n*(FIXED_1-exp); \
93 #define CT_TO_SECS(x) ((x) / HZ)
94 #define CT_TO_USECS(x) (((x) % HZ) * 1000000/HZ)
96 extern int nr_threads;
98 DECLARE_PER_CPU(unsigned long, process_counts);
99 // DECLARE_PER_CPU(struct runqueue, runqueues); -- removed after ckrm cpu v7 merge
100 extern int nr_processes(void);
101 extern unsigned long nr_running(void);
102 extern unsigned long nr_uninterruptible(void);
103 extern unsigned long nr_iowait(void);
105 #include <linux/time.h>
106 #include <linux/param.h>
107 #include <linux/resource.h>
108 #include <linux/timer.h>
110 #include <asm/processor.h>
112 #define TASK_RUNNING 0
113 #define TASK_INTERRUPTIBLE 1
114 #define TASK_UNINTERRUPTIBLE 2
115 #define TASK_STOPPED 4
116 #define TASK_TRACED 8
117 #define EXIT_ZOMBIE 16
119 #define TASK_ONHOLD 64
121 #define __set_task_state(tsk, state_value) \
122 do { (tsk)->state = (state_value); } while (0)
123 #define set_task_state(tsk, state_value) \
124 set_mb((tsk)->state, (state_value))
126 #define __set_current_state(state_value) \
127 do { current->state = (state_value); } while (0)
128 #define set_current_state(state_value) \
129 set_mb(current->state, (state_value))
132 * Scheduling policies
134 #define SCHED_NORMAL 0
144 #include <linux/taskdelays.h>
145 #include <linux/spinlock.h>
148 * This serializes "schedule()" and also protects
149 * the run-queue from deletions/modifications (but
150 * _adding_ to the beginning of the run-queue has
153 extern rwlock_t tasklist_lock;
154 extern spinlock_t mmlist_lock;
156 typedef struct task_struct task_t;
158 extern void sched_init(void);
159 extern void sched_init_smp(void);
160 extern void init_idle(task_t *idle, int cpu);
162 extern cpumask_t nohz_cpu_mask;
164 extern void show_state(void);
165 extern void show_regs(struct pt_regs *);
168 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
169 * task), SP is the stack pointer of the first frame that should be shown in the back
170 * trace (or NULL if the entire call-chain of the task should be shown).
172 extern void show_stack(struct task_struct *task, unsigned long *sp);
174 void io_schedule(void);
175 long io_schedule_timeout(long timeout);
177 extern void cpu_init (void);
178 extern void trap_init(void);
179 extern void update_process_times(int user);
180 extern void scheduler_tick(int user_tick, int system);
181 extern unsigned long cache_decay_ticks;
183 /* Attach to any functions which should be ignored in wchan output. */
184 #define __sched __attribute__((__section__(".sched.text")))
185 /* Is this address in the __sched functions? */
186 extern int in_sched_functions(unsigned long addr);
188 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
189 extern signed long FASTCALL(schedule_timeout(signed long timeout));
190 asmlinkage void schedule(void);
194 /* Maximum number of active map areas.. This is a random (large) number */
195 #define DEFAULT_MAX_MAP_COUNT 65536
197 extern int sysctl_max_map_count;
199 #include <linux/aio.h>
202 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
203 unsigned long, unsigned long);
206 arch_get_unmapped_exec_area(struct file *, unsigned long, unsigned long,
207 unsigned long, unsigned long);
209 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
210 unsigned long len, unsigned long pgoff,
211 unsigned long flags);
212 extern void arch_unmap_area(struct vm_area_struct *area);
213 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
217 struct vm_area_struct * mmap; /* list of VMAs */
218 struct rb_root mm_rb;
219 struct vm_area_struct * mmap_cache; /* last find_vma result */
220 unsigned long (*get_unmapped_area) (struct file *filp,
221 unsigned long addr, unsigned long len,
222 unsigned long pgoff, unsigned long flags);
223 unsigned long (*get_unmapped_exec_area) (struct file *filp,
224 unsigned long addr, unsigned long len,
225 unsigned long pgoff, unsigned long flags);
226 void (*unmap_area) (struct vm_area_struct *area);
227 unsigned long mmap_base; /* base of mmap area */
228 unsigned long free_area_cache; /* first hole */
230 atomic_t mm_users; /* How many users with user space? */
231 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
232 int map_count; /* number of VMAs */
233 struct rw_semaphore mmap_sem;
234 spinlock_t page_table_lock; /* Protects page tables, mm->rss, mm->anon_rss */
236 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
237 * together off init_mm.mmlist, and are protected
241 unsigned long start_code, end_code, start_data, end_data;
242 unsigned long start_brk, brk, start_stack;
243 unsigned long arg_start, arg_end, env_start, env_end;
244 unsigned long rss, anon_rss, total_vm, locked_vm, shared_vm;
245 unsigned long exec_vm, stack_vm, reserved_vm, def_flags, nr_ptes;
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;
254 struct vx_info *mm_vx_info;
256 /* Token based thrashing protection. */
257 unsigned long swap_token_time;
260 /* coredumping support */
262 struct completion *core_startup_done, core_done;
265 rwlock_t ioctx_list_lock;
266 struct kioctx *ioctx_list;
268 struct kioctx default_kioctx;
269 #ifdef CONFIG_CKRM_RES_MEM
270 struct ckrm_mem_res *memclass;
271 struct list_head tasklist; /* tasks sharing this address space */
272 spinlock_t peertask_lock; /* protect tasklist above */
276 struct sighand_struct {
278 struct k_sigaction action[_NSIG];
283 * NOTE! "signal_struct" does not have it's own
284 * locking, because a shared signal_struct always
285 * implies a shared sighand_struct, so locking
286 * sighand_struct is always a proper superset of
287 * the locking of signal_struct.
289 struct signal_struct {
293 /* current thread group signal load-balancing target: */
296 /* shared signal handling: */
297 struct sigpending shared_pending;
299 /* thread group exit support */
303 * - notify group_exit_task when ->count is equal to notify_count
304 * - everyone except group_exit_task is stopped during signal delivery
305 * of fatal signals, group_exit_task processes the signal.
307 struct task_struct *group_exit_task;
310 /* thread group stop support, overloads group_exit_code too */
311 int group_stop_count;
312 /* 1 if group stopped since last SIGCONT, -1 if SIGCONT since report */
315 /* POSIX.1b Interval Timers */
316 struct list_head posix_timers;
318 /* job control IDs */
322 /* boolean value for session group leader */
325 struct tty_struct *tty; /* NULL if no tty */
328 * Cumulative resource counters for dead threads in the group,
329 * and for reaped dead child processes forked by this group.
330 * Live threads maintain their own counters and add to these
331 * in __exit_signal, except for the group leader.
333 unsigned long utime, stime, cutime, cstime;
334 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
335 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
338 * We don't bother to synchronize most readers of this at all,
339 * because there is no reader checking a limit that actually needs
340 * to get both rlim_cur and rlim_max atomically, and either one
341 * alone is a single word that can safely be read normally.
342 * getrlimit/setrlimit use task_lock(current->group_leader) to
343 * protect this instead of the siglock, because they really
344 * have no need to disable irqs.
346 struct rlimit rlim[RLIM_NLIMITS];
350 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
351 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
352 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
353 * are inverted: lower p->prio value means higher priority.
355 * The MAX_USER_RT_PRIO value allows the actual maximum
356 * RT priority to be separate from the value exported to
357 * user-space. This allows kernel threads to set their
358 * priority to a value higher than any user task. Note:
359 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
362 #define MAX_USER_RT_PRIO 100
363 #define MAX_RT_PRIO MAX_USER_RT_PRIO
365 #define MAX_PRIO (MAX_RT_PRIO + 40)
367 #define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
370 * Some day this will be a full-fledged user tracking system..
373 atomic_t __count; /* reference count */
374 atomic_t processes; /* How many processes does this user have? */
375 atomic_t files; /* How many open files does this user have? */
376 atomic_t sigpending; /* How many pending signals does this user have? */
377 /* protected by mq_lock */
378 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
379 unsigned long locked_shm; /* How many pages of mlocked shm ? */
382 struct key *uid_keyring; /* UID specific keyring */
383 struct key *session_keyring; /* UID's default session keyring */
386 /* Hash table maintenance information */
387 struct list_head uidhash_list;
392 extern struct user_struct *find_user(xid_t, uid_t);
394 extern struct user_struct root_user;
395 #define INIT_USER (&root_user)
397 typedef struct prio_array prio_array_t;
398 struct backing_dev_info;
399 struct reclaim_state;
401 #ifdef CONFIG_SCHEDSTATS
403 /* cumulative counters */
404 unsigned long cpu_time, /* time spent on the cpu */
405 run_delay, /* time spent waiting on a runqueue */
406 pcnt; /* # of timeslices run on this cpu */
409 unsigned long last_arrival, /* when we last ran on a cpu */
410 last_queued; /* when we were last queued to run */
413 extern struct file_operations proc_schedstat_operations;
425 * sched-domains (multiprocessor balancing) declarations:
428 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
430 #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
431 #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
432 #define SD_BALANCE_EXEC 4 /* Balance on exec */
433 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
434 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
435 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
436 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
439 struct sched_group *next; /* Must be a circular list */
443 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
444 * single CPU. This is read only (except for setup, hotplug CPU).
446 unsigned long cpu_power;
449 struct sched_domain {
450 /* These fields must be setup */
451 struct sched_domain *parent; /* top domain must be null terminated */
452 struct sched_group *groups; /* the balancing groups of the domain */
453 cpumask_t span; /* span of all CPUs in this domain */
454 unsigned long min_interval; /* Minimum balance interval ms */
455 unsigned long max_interval; /* Maximum balance interval ms */
456 unsigned int busy_factor; /* less balancing by factor if busy */
457 unsigned int imbalance_pct; /* No balance until over watermark */
458 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
459 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
460 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
461 int flags; /* See SD_* */
463 /* Runtime fields. */
464 unsigned long last_balance; /* init to jiffies. units in jiffies */
465 unsigned int balance_interval; /* initialise to 1. units in ms. */
466 unsigned int nr_balance_failed; /* initialise to 0 */
468 #ifdef CONFIG_SCHEDSTATS
469 /* load_balance() stats */
470 unsigned long lb_cnt[MAX_IDLE_TYPES];
471 unsigned long lb_failed[MAX_IDLE_TYPES];
472 unsigned long lb_imbalance[MAX_IDLE_TYPES];
473 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
474 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
476 /* sched_balance_exec() stats */
477 unsigned long sbe_attempts;
478 unsigned long sbe_pushed;
480 /* try_to_wake_up() stats */
481 unsigned long ttwu_wake_affine;
482 unsigned long ttwu_wake_balance;
486 #ifdef ARCH_HAS_SCHED_DOMAIN
487 /* Useful helpers that arch setup code may use. Defined in kernel/sched.c */
488 extern cpumask_t cpu_isolated_map;
489 extern void init_sched_build_groups(struct sched_group groups[],
490 cpumask_t span, int (*group_fn)(int cpu));
491 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
492 #endif /* ARCH_HAS_SCHED_DOMAIN */
493 #endif /* CONFIG_SMP */
496 struct io_context; /* See blkdev.h */
497 void exit_io_context(void);
499 #define NGROUPS_SMALL 32
500 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
504 gid_t small_block[NGROUPS_SMALL];
510 * get_group_info() must be called with the owning task locked (via task_lock())
511 * when task != current. The reason being that the vast majority of callers are
512 * looking at current->group_info, which can not be changed except by the
513 * current task. Changing current->group_info requires the task lock, too.
515 #define get_group_info(group_info) do { \
516 atomic_inc(&(group_info)->usage); \
519 #define put_group_info(group_info) do { \
520 if (atomic_dec_and_test(&(group_info)->usage)) \
521 groups_free(group_info); \
524 struct group_info *groups_alloc(int gidsetsize);
525 void groups_free(struct group_info *group_info);
526 int set_current_groups(struct group_info *group_info);
527 /* access the groups "array" with this macro */
528 #define GROUP_AT(gi, i) \
529 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
532 struct audit_context; /* See audit.c */
535 #ifdef CONFIG_CKRM_CPU_SCHEDULE
537 * ckrm_cpu_demand_stat - used to track the cpu demand of a task/class
538 * @run: how much time it has been running since the counter started
539 * @total: total time since the counter started
540 * @last_sleep: the last time it sleeps, last_sleep = 0 when not sleeping
541 * @recalc_interval: how often do we recalculate the cpu_demand
542 * @cpu_demand: moving average of run/total
544 struct ckrm_cpu_demand_stat {
545 unsigned long long run;
546 unsigned long long total;
547 unsigned long long last_sleep;
548 unsigned long long recalc_interval;
549 unsigned long cpu_demand; /*estimated cpu demand */
555 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
556 struct thread_info *thread_info;
558 unsigned long flags; /* per process flags, defined below */
559 unsigned long ptrace;
561 int lock_depth; /* Lock depth */
563 int prio, static_prio;
564 struct list_head run_list;
567 unsigned long sleep_avg;
568 long interactive_credit;
569 unsigned long long timestamp, last_ran;
572 unsigned long policy;
573 cpumask_t cpus_allowed;
574 unsigned int time_slice, first_time_slice;
576 #ifdef CONFIG_SCHEDSTATS
577 struct sched_info sched_info;
580 struct list_head tasks;
582 * ptrace_list/ptrace_children forms the list of my children
583 * that were stolen by a ptracer.
585 struct list_head ptrace_children;
586 struct list_head ptrace_list;
588 struct mm_struct *mm, *active_mm;
591 struct linux_binfmt *binfmt;
593 int exit_code, exit_signal;
594 int pdeath_signal; /* The signal sent when the parent dies */
596 unsigned long personality;
601 * pointers to (original) parent process, youngest child, younger sibling,
602 * older sibling, respectively. (p->father can be replaced with
605 struct task_struct *real_parent; /* real parent process (when being debugged) */
606 struct task_struct *parent; /* parent process */
608 * children/sibling forms the list of my children plus the
609 * tasks I'm ptracing.
611 struct list_head children; /* list of my children */
612 struct list_head sibling; /* linkage in my parent's children list */
613 struct task_struct *group_leader; /* threadgroup leader */
615 /* PID/PID hash table linkage. */
616 struct pid pids[PIDTYPE_MAX];
618 wait_queue_head_t wait_chldexit; /* for wait4() */
619 struct completion *vfork_done; /* for vfork() */
620 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
621 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
623 unsigned long rt_priority;
624 unsigned long it_real_value, it_prof_value, it_virt_value;
625 unsigned long it_real_incr, it_prof_incr, it_virt_incr;
626 struct timer_list real_timer;
627 unsigned long utime, stime;
628 unsigned long nvcsw, nivcsw; /* context switch counts */
629 struct timespec start_time;
630 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
631 unsigned long min_flt, maj_flt;
632 /* process credentials */
633 uid_t uid,euid,suid,fsuid;
634 gid_t gid,egid,sgid,fsgid;
635 struct group_info *group_info;
636 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
637 unsigned keep_capabilities:1;
638 struct user_struct *user;
640 struct key *session_keyring; /* keyring inherited over fork */
641 struct key *process_keyring; /* keyring private to this process (CLONE_THREAD) */
642 struct key *thread_keyring; /* keyring private to this thread */
644 unsigned short used_math;
646 /* file system info */
647 int link_count, total_link_count;
649 struct sysv_sem sysvsem;
650 /* CPU-specific state of this task */
651 struct thread_struct thread;
652 /* filesystem information */
653 struct fs_struct *fs;
654 /* open file information */
655 struct files_struct *files;
657 struct namespace *namespace;
658 /* signal handlers */
659 struct signal_struct *signal;
660 struct sighand_struct *sighand;
661 sigset_t blocked, real_blocked;
662 struct sigpending pending;
664 unsigned long sas_ss_sp;
666 int (*notifier)(void *priv);
668 sigset_t *notifier_mask;
672 void (*tux_exit)(void);
676 struct audit_context *audit_context;
678 /* vserver context data */
680 struct vx_info *vx_info;
682 /* vserver network data */
684 struct nx_info *nx_info;
686 /* Thread group tracking */
689 /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
690 spinlock_t alloc_lock;
691 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
692 spinlock_t proc_lock;
693 /* context-switch lock */
694 spinlock_t switch_lock;
696 /* journalling filesystem info */
700 struct reclaim_state *reclaim_state;
702 struct dentry *proc_dentry;
703 struct backing_dev_info *backing_dev_info;
705 struct io_context *io_context;
709 unsigned long ptrace_message;
710 siginfo_t *last_siginfo; /* For ptrace use. */
712 * current io wait handle: wait queue entry to use for io waits
713 * If this thread is processing aio, this points at the waitqueue
714 * inside the currently handled kiocb. It may be NULL (i.e. default
715 * to a stack based synchronous wait) if its doing sync IO.
717 wait_queue_t *io_wait;
719 struct mempolicy *mempolicy;
720 short il_next; /* could be shared with used_math */
723 spinlock_t ckrm_tsklock;
725 #ifdef CONFIG_CKRM_TYPE_TASKCLASS
726 struct ckrm_task_class *taskclass;
727 struct list_head taskclass_link;
728 #ifdef CONFIG_CKRM_CPU_SCHEDULE
729 struct ckrm_cpu_class *cpu_class;
730 /* track cpu demand of this task */
731 struct ckrm_cpu_demand_stat demand_stat;
732 #endif /* CONFIG_CKRM_CPU_SCHEDULE */
733 #endif /* CONFIG_CKRM_TYPE_TASKCLASS */
734 #ifdef CONFIG_CKRM_RES_MEM
735 struct list_head mm_peers; /* list of tasks using same mm_struct */
737 #endif /* CONFIG_CKRM */
738 #ifdef CONFIG_DELAY_ACCT
739 struct task_delay_info delays;
743 static inline pid_t process_group(struct task_struct *tsk)
745 return tsk->signal->pgrp;
749 * pid_alive - check that a task structure is not stale
750 * @p: Task structure to be checked.
752 * Test if a process is not yet dead (at most zombie state)
753 * If pid_alive fails, then pointers within the task structure
754 * can be stale and must not be dereferenced.
756 static inline int pid_alive(struct task_struct *p)
758 return p->pids[PIDTYPE_PID].nr != 0;
761 extern void free_task(struct task_struct *tsk);
762 extern void __put_task_struct(struct task_struct *tsk);
763 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
764 #define put_task_struct(tsk) \
765 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
770 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
771 /* Not implemented yet, only for 486*/
772 #define PF_STARTING 0x00000002 /* being created */
773 #define PF_EXITING 0x00000004 /* getting shut down */
774 #define PF_DEAD 0x00000008 /* Dead */
775 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
776 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
777 #define PF_DUMPCORE 0x00000200 /* dumped core */
778 #define PF_SIGNALED 0x00000400 /* killed by a signal */
779 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
780 #define PF_MEMDIE 0x00001000 /* Killed for out-of-memory */
781 #define PF_FLUSHER 0x00002000 /* responsible for disk writeback */
783 #define PF_FREEZE 0x00004000 /* this task should be frozen for suspend */
784 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
785 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
786 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
787 #define PF_KSWAPD 0x00040000 /* I am kswapd */
788 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
789 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
790 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
791 #define PF_BORROWED_MM 0x00400000 /* I am a kthread doing use_mm */
792 #define PF_RELOCEXEC 0x00800000 /* relocate shared libraries */
793 #define PF_MEMIO 0x01000000 /* I am potentially doing I/O for mem */
794 #define PF_IOWAIT 0x02000000 /* I am waiting on disk I/O */
797 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
799 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
805 extern unsigned long long sched_clock(void);
807 /* sched_exec is called by processes performing an exec */
809 extern void sched_exec(void);
811 #define sched_exec() {}
814 extern void sched_idle_next(void);
815 extern void set_user_nice(task_t *p, long nice);
816 extern int task_prio(const task_t *p);
817 extern int task_nice(const task_t *p);
818 extern int task_curr(const task_t *p);
819 extern int idle_cpu(int cpu);
824 * The default (Linux) execution domain.
826 extern struct exec_domain default_exec_domain;
829 struct thread_info thread_info;
830 unsigned long stack[THREAD_SIZE/sizeof(long)];
833 #ifndef __HAVE_ARCH_KSTACK_END
834 static inline int kstack_end(void *addr)
836 /* Reliable end of stack detection:
837 * Some APM bios versions misalign the stack
839 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
843 extern union thread_union init_thread_union;
844 extern struct task_struct init_task;
846 extern struct mm_struct init_mm;
849 #define find_task_by_real_pid(nr) \
850 find_task_by_pid_type(PIDTYPE_PID, nr)
851 #define find_task_by_pid(nr) \
852 find_task_by_pid_type(PIDTYPE_PID, \
855 extern struct task_struct *find_task_by_pid_type(int type, int pid);
856 extern void set_special_pids(pid_t session, pid_t pgrp);
857 extern void __set_special_pids(pid_t session, pid_t pgrp);
859 /* per-UID process charging. */
860 extern struct user_struct * alloc_uid(xid_t, uid_t);
861 static inline struct user_struct *get_uid(struct user_struct *u)
863 atomic_inc(&u->__count);
867 extern void free_uid(struct user_struct *);
868 extern void switch_uid(struct user_struct *);
870 #include <asm/current.h>
872 extern void do_timer(struct pt_regs *);
874 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
875 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
876 extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
877 unsigned long clone_flags));
879 extern void kick_process(struct task_struct *tsk);
881 static inline void kick_process(struct task_struct *tsk) { }
883 extern void FASTCALL(sched_fork(task_t * p));
884 extern void FASTCALL(sched_exit(task_t * p));
886 extern int in_group_p(gid_t);
887 extern int in_egroup_p(gid_t);
889 extern void proc_caches_init(void);
890 extern void flush_signals(struct task_struct *);
891 extern void flush_signal_handlers(struct task_struct *, int force_default);
892 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
894 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
899 spin_lock_irqsave(&tsk->sighand->siglock, flags);
900 ret = dequeue_signal(tsk, mask, info);
901 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
906 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
908 extern void unblock_all_signals(void);
909 extern void release_task(struct task_struct * p);
910 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
911 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
912 extern int force_sigsegv(int, struct task_struct *);
913 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
914 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
915 extern int kill_pg_info(int, struct siginfo *, pid_t);
916 extern int kill_proc_info(int, struct siginfo *, pid_t);
917 extern void do_notify_parent(struct task_struct *, int);
918 extern void force_sig(int, struct task_struct *);
919 extern void force_sig_specific(int, struct task_struct *);
920 extern int send_sig(int, struct task_struct *, int);
921 extern void zap_other_threads(struct task_struct *p);
922 extern int kill_pg(pid_t, int, int);
923 extern int kill_sl(pid_t, int, int);
924 extern int kill_proc(pid_t, int, int);
925 extern struct sigqueue *sigqueue_alloc(void);
926 extern void sigqueue_free(struct sigqueue *);
927 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
928 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
929 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
930 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
932 /* These can be the second arg to send_sig_info/send_group_sig_info. */
933 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
934 #define SEND_SIG_PRIV ((struct siginfo *) 1)
935 #define SEND_SIG_FORCED ((struct siginfo *) 2)
937 /* True if we are on the alternate signal stack. */
939 static inline int on_sig_stack(unsigned long sp)
941 return (sp - current->sas_ss_sp < current->sas_ss_size);
944 static inline int sas_ss_flags(unsigned long sp)
946 return (current->sas_ss_size == 0 ? SS_DISABLE
947 : on_sig_stack(sp) ? SS_ONSTACK : 0);
951 #ifdef CONFIG_SECURITY
952 /* code is in security.c */
953 extern int capable(int cap);
954 extern int vx_capable(int cap, int ccap);
956 static inline int capable(int cap)
958 if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
960 if (cap_raised(current->cap_effective, cap)) {
961 current->flags |= PF_SUPERPRIV;
967 static inline int vx_capable(int cap, int ccap)
969 if (cap_raised(current->cap_effective, cap) &&
971 current->flags |= PF_SUPERPRIV;
980 * Routines for handling mm_structs
982 extern struct mm_struct * mm_alloc(void);
984 /* mmdrop drops the mm and the page tables */
985 extern void FASTCALL(__mmdrop(struct mm_struct *));
986 static inline void mmdrop(struct mm_struct * mm)
988 if (atomic_dec_and_test(&mm->mm_count))
992 /* mmput gets rid of the mappings and all user-space */
993 extern void mmput(struct mm_struct *);
994 /* Grab a reference to a task's mm, if it is not already going away */
995 extern struct mm_struct *get_task_mm(struct task_struct *task);
996 /* Remove the current tasks stale references to the old mm_struct */
997 extern void mm_release(struct task_struct *, struct mm_struct *);
999 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1000 extern void flush_thread(void);
1001 extern void exit_thread(void);
1003 extern void exit_mm(struct task_struct *);
1004 extern void exit_files(struct task_struct *);
1005 extern void exit_signal(struct task_struct *);
1006 extern void __exit_signal(struct task_struct *);
1007 extern void exit_sighand(struct task_struct *);
1008 extern void __exit_sighand(struct task_struct *);
1009 extern void exit_itimers(struct signal_struct *);
1011 extern NORET_TYPE void do_group_exit(int);
1013 extern void reparent_to_init(void);
1014 extern void daemonize(const char *, ...);
1015 extern int allow_signal(int);
1016 extern int disallow_signal(int);
1017 extern task_t *child_reaper;
1019 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1020 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1021 task_t *fork_idle(int);
1023 extern void set_task_comm(struct task_struct *tsk, char *from);
1024 extern void get_task_comm(char *to, struct task_struct *tsk);
1027 extern void wait_task_inactive(task_t * p);
1029 #define wait_task_inactive(p) do { } while (0)
1032 #define remove_parent(p) list_del_init(&(p)->sibling)
1033 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1035 #define REMOVE_LINKS(p) do { \
1036 if (thread_group_leader(p)) \
1037 list_del_init(&(p)->tasks); \
1041 #define SET_LINKS(p) do { \
1042 if (thread_group_leader(p)) \
1043 list_add_tail(&(p)->tasks,&init_task.tasks); \
1044 add_parent(p, (p)->parent); \
1047 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1048 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1050 #define for_each_process(p) \
1051 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1054 * Careful: do_each_thread/while_each_thread is a double loop so
1055 * 'break' will not work as expected - use goto instead.
1057 #define do_each_thread(g, t) \
1058 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1060 #define while_each_thread(g, t) \
1061 while ((t = next_thread(t)) != g)
1063 extern task_t * FASTCALL(next_thread(const task_t *p));
1065 #define thread_group_leader(p) (p->pid == p->tgid)
1067 static inline int thread_group_empty(task_t *p)
1069 return list_empty(&p->pids[PIDTYPE_TGID].pid_list);
1072 #define delay_group_leader(p) \
1073 (thread_group_leader(p) && !thread_group_empty(p))
1075 extern void unhash_process(struct task_struct *p);
1078 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info, ->comm, keyring
1079 * subscriptions and synchronises with wait4(). Also used in procfs.
1081 * Nests both inside and outside of read_lock(&tasklist_lock).
1082 * It must not be nested with write_lock_irq(&tasklist_lock),
1083 * neither inside nor outside.
1085 static inline void task_lock(struct task_struct *p)
1087 spin_lock(&p->alloc_lock);
1090 static inline void task_unlock(struct task_struct *p)
1092 spin_unlock(&p->alloc_lock);
1095 /* set thread flags in other task's structures
1096 * - see asm/thread_info.h for TIF_xxxx flags available
1098 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1100 set_ti_thread_flag(tsk->thread_info,flag);
1103 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1105 clear_ti_thread_flag(tsk->thread_info,flag);
1108 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1110 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1113 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1115 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1118 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1120 return test_ti_thread_flag(tsk->thread_info,flag);
1123 static inline void set_tsk_need_resched(struct task_struct *tsk)
1125 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1128 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1130 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1133 static inline int signal_pending(struct task_struct *p)
1135 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1138 static inline int need_resched(void)
1140 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1143 extern void __cond_resched(void);
1144 static inline void cond_resched(void)
1151 * cond_resched_lock() - if a reschedule is pending, drop the given lock,
1152 * call schedule, and on return reacquire the lock.
1154 * This works OK both with and without CONFIG_PREEMPT. We do strange low-level
1155 * operations here to prevent schedule() from being called twice (once via
1156 * spin_unlock(), once by hand).
1158 static inline void cond_resched_lock(spinlock_t * lock)
1160 if (need_resched()) {
1161 _raw_spin_unlock(lock);
1162 preempt_enable_no_resched();
1168 /* Reevaluate whether the task has signals pending delivery.
1169 This is required every time the blocked sigset_t changes.
1170 callers must hold sighand->siglock. */
1172 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1173 extern void recalc_sigpending(void);
1175 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1178 * Wrappers for p->thread_info->cpu access. No-op on UP.
1182 static inline unsigned int task_cpu(const struct task_struct *p)
1184 return p->thread_info->cpu;
1187 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1189 p->thread_info->cpu = cpu;
1194 static inline unsigned int task_cpu(const struct task_struct *p)
1199 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1203 #endif /* CONFIG_SMP */
1205 /* API for registering delay info */
1206 #ifdef CONFIG_DELAY_ACCT
1208 #define test_delay_flag(tsk,flg) ((tsk)->flags & (flg))
1209 #define set_delay_flag(tsk,flg) ((tsk)->flags |= (flg))
1210 #define clear_delay_flag(tsk,flg) ((tsk)->flags &= ~(flg))
1212 #define def_delay_var(var) unsigned long long var
1213 #define get_delay(tsk,field) ((tsk)->delays.field)
1215 #define start_delay(var) ((var) = sched_clock())
1216 #define start_delay_set(var,flg) (set_delay_flag(current,flg),(var) = sched_clock())
1218 #define inc_delay(tsk,field) (((tsk)->delays.field)++)
1220 /* because of hardware timer drifts in SMPs and task continue on different cpu
1221 * then where the start_ts was taken there is a possibility that
1222 * end_ts < start_ts by some usecs. In this case we ignore the diff
1223 * and add nothing to the total.
1226 #define test_ts_integrity(start_ts,end_ts) (likely((end_ts) > (start_ts)))
1228 #define test_ts_integrity(start_ts,end_ts) (1)
1231 #define add_delay_ts(tsk,field,start_ts,end_ts) \
1232 do { if (test_ts_integrity(start_ts,end_ts)) (tsk)->delays.field += ((end_ts)-(start_ts)); } while (0)
1234 #define add_delay_clear(tsk,field,start_ts,flg) \
1236 unsigned long long now = sched_clock();\
1237 add_delay_ts(tsk,field,start_ts,now); \
1238 clear_delay_flag(tsk,flg); \
1241 static inline void add_io_delay(unsigned long long dstart)
1243 struct task_struct * tsk = current;
1244 unsigned long long now = sched_clock();
1245 unsigned long long val;
1247 if (test_ts_integrity(dstart,now))
1251 if (test_delay_flag(tsk,PF_MEMIO)) {
1252 tsk->delays.mem_iowait_total += val;
1253 tsk->delays.num_memwaits++;
1255 tsk->delays.iowait_total += val;
1256 tsk->delays.num_iowaits++;
1258 clear_delay_flag(tsk,PF_IOWAIT);
1261 inline static void init_delays(struct task_struct *tsk)
1263 memset((void*)&tsk->delays,0,sizeof(tsk->delays));
1268 #define test_delay_flag(tsk,flg) (0)
1269 #define set_delay_flag(tsk,flg) do { } while (0)
1270 #define clear_delay_flag(tsk,flg) do { } while (0)
1272 #define def_delay_var(var)
1273 #define get_delay(tsk,field) (0)
1275 #define start_delay(var) do { } while (0)
1276 #define start_delay_set(var,flg) do { } while (0)
1278 #define inc_delay(tsk,field) do { } while (0)
1279 #define add_delay_ts(tsk,field,start_ts,now) do { } while (0)
1280 #define add_delay_clear(tsk,field,start_ts,flg) do { } while (0)
1281 #define add_io_delay(dstart) do { } while (0)
1282 #define init_delays(tsk) do { } while (0)
1287 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1288 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1290 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1292 mm->mmap_base = TASK_UNMAPPED_BASE;
1293 mm->get_unmapped_area = arch_get_unmapped_area;
1294 mm->unmap_area = arch_unmap_area;
1298 extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1299 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1301 #ifdef CONFIG_MAGIC_SYSRQ
1303 extern void normalize_rt_tasks(void);
1307 /* API for registering delay info */
1308 #ifdef CONFIG_DELAY_ACCT
1310 #define test_delay_flag(tsk,flg) ((tsk)->flags & (flg))
1311 #define set_delay_flag(tsk,flg) ((tsk)->flags |= (flg))
1312 #define clear_delay_flag(tsk,flg) ((tsk)->flags &= ~(flg))
1314 #define def_delay_var(var) unsigned long long var
1315 #define get_delay(tsk,field) ((tsk)->delays.field)
1317 #define start_delay(var) ((var) = sched_clock())
1318 #define start_delay_set(var,flg) (set_delay_flag(current,flg),(var) = sched_clock())
1320 #define inc_delay(tsk,field) (((tsk)->delays.field)++)
1322 /* because of hardware timer drifts in SMPs and task continue on different cpu
1323 * then where the start_ts was taken there is a possibility that
1324 * end_ts < start_ts by some usecs. In this case we ignore the diff
1325 * and add nothing to the total.
1328 #define test_ts_integrity(start_ts,end_ts) (likely((end_ts) > (start_ts)))
1330 #define test_ts_integrity(start_ts,end_ts) (1)
1333 #define add_delay_ts(tsk,field,start_ts,end_ts) \
1334 do { if (test_ts_integrity(start_ts,end_ts)) (tsk)->delays.field += ((end_ts)-(start_ts)); } while (0)
1336 #define add_delay_clear(tsk,field,start_ts,flg) \
1338 unsigned long long now = sched_clock();\
1339 add_delay_ts(tsk,field,start_ts,now); \
1340 clear_delay_flag(tsk,flg); \
1343 static inline void add_io_delay(unsigned long long dstart)
1345 struct task_struct * tsk = current;
1346 unsigned long long now = sched_clock();
1347 unsigned long long val;
1349 if (test_ts_integrity(dstart,now))
1353 if (test_delay_flag(tsk,PF_MEMIO)) {
1354 tsk->delays.mem_iowait_total += val;
1355 tsk->delays.num_memwaits++;
1357 tsk->delays.iowait_total += val;
1358 tsk->delays.num_iowaits++;
1360 clear_delay_flag(tsk,PF_IOWAIT);
1363 inline static void init_delays(struct task_struct *tsk)
1365 memset((void*)&tsk->delays,0,sizeof(tsk->delays));
1370 #define test_delay_flag(tsk,flg) (0)
1371 #define set_delay_flag(tsk,flg) do { } while (0)
1372 #define clear_delay_flag(tsk,flg) do { } while (0)
1374 #define def_delay_var(var)
1375 #define get_delay(tsk,field) (0)
1377 #define start_delay(var) do { } while (0)
1378 #define start_delay_set(var,flg) do { } while (0)
1380 #define inc_delay(tsk,field) do { } while (0)
1381 #define add_delay_ts(tsk,field,start_ts,now) do { } while (0)
1382 #define add_delay_clear(tsk,field,start_ts,flg) do { } while (0)
1383 #define add_io_delay(dstart) do { } while (0)
1384 #define init_delays(tsk) do { } while (0)
1387 #endif /* __KERNEL__ */