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>
34 extern int exec_shield;
35 extern int exec_shield_randomize;
36 extern int print_fatal_signals;
41 #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
42 #define CLONE_VM 0x00000100 /* set if VM shared between processes */
43 #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
44 #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
45 #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
46 #define CLONE_IDLETASK 0x00001000 /* set if new pid should be 0 (kernel only)*/
47 #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
48 #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
49 #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
50 #define CLONE_THREAD 0x00010000 /* Same thread group? */
51 #define CLONE_NEWNS 0x00020000 /* New namespace group? */
52 #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
53 #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
54 #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
55 #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
56 #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
57 #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
58 #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
59 #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
62 * List of flags we want to share for kernel threads,
63 * if only because they are not used by them anyway.
65 #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
68 * These are the constant used to fake the fixed-point load-average
69 * counting. Some notes:
70 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
71 * a load-average precision of 10 bits integer + 11 bits fractional
72 * - if you want to count load-averages more often, you need more
73 * precision, or rounding will get you. With 2-second counting freq,
74 * the EXP_n values would be 1981, 2034 and 2043 if still using only
77 extern unsigned long avenrun[]; /* Load averages */
79 #define FSHIFT 11 /* nr of bits of precision */
80 #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
81 #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
82 #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
83 #define EXP_5 2014 /* 1/exp(5sec/5min) */
84 #define EXP_15 2037 /* 1/exp(5sec/15min) */
86 #define CALC_LOAD(load,exp,n) \
88 load += n*(FIXED_1-exp); \
91 #define CT_TO_SECS(x) ((x) / HZ)
92 #define CT_TO_USECS(x) (((x) % HZ) * 1000000/HZ)
94 extern int nr_threads;
96 DECLARE_PER_CPU(unsigned long, process_counts);
97 extern int nr_processes(void);
98 extern unsigned long nr_running(void);
99 extern unsigned long nr_uninterruptible(void);
100 extern unsigned long nr_iowait(void);
102 #include <linux/time.h>
103 #include <linux/param.h>
104 #include <linux/resource.h>
105 #include <linux/timer.h>
107 #include <asm/processor.h>
108 #include <linux/vserver/context.h>
110 #define TASK_RUNNING 0
111 #define TASK_INTERRUPTIBLE 1
112 #define TASK_UNINTERRUPTIBLE 2
113 #define TASK_STOPPED 4
114 #define TASK_ZOMBIE 8
116 #define TASK_ONHOLD 32
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))
129 * Scheduling policies
131 #define SCHED_NORMAL 0
141 #include <linux/taskdelays.h>
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(int user_tick, int system);
178 extern unsigned long cache_decay_ticks;
180 /* Attach to any functions which should be ignored in wchan output. */
181 #define __sched __attribute__((__section__(".sched.text")))
182 /* Is this address in the __sched functions? */
183 extern int in_sched_functions(unsigned long addr);
185 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
186 extern signed long FASTCALL(schedule_timeout(signed long timeout));
187 asmlinkage void schedule(void);
191 /* Maximum number of active map areas.. This is a random (large) number */
192 #define DEFAULT_MAX_MAP_COUNT 65536
194 extern int sysctl_max_map_count;
196 #include <linux/aio.h>
199 struct vm_area_struct * mmap; /* list of VMAs */
200 struct rb_root mm_rb;
201 struct vm_area_struct * mmap_cache; /* last find_vma result */
202 unsigned long free_area_cache; /* first hole */
203 unsigned long non_executable_cache; /* last hole top */
204 unsigned long mmap_top; /* top of mmap area */
206 atomic_t mm_users; /* How many users with user space? */
207 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
208 int map_count; /* number of VMAs */
209 struct rw_semaphore mmap_sem;
210 spinlock_t page_table_lock; /* Protects task page tables and mm->rss */
212 struct list_head mmlist; /* List of all active mm's. These are globally strung
213 * together off init_mm.mmlist, and are protected
217 unsigned long start_code, end_code, start_data, end_data;
218 unsigned long start_brk, brk, start_stack;
219 unsigned long arg_start, arg_end, env_start, env_end;
220 unsigned long rss, total_vm, locked_vm;
221 unsigned long def_flags;
223 unsigned long saved_auxv[40]; /* for /proc/PID/auxv */
226 cpumask_t cpu_vm_mask;
228 /* Architecture-specific MM context */
229 mm_context_t context;
230 struct vx_info *mm_vx_info;
232 /* coredumping support */
234 struct completion *core_startup_done, core_done;
237 rwlock_t ioctx_list_lock;
238 struct kioctx *ioctx_list;
240 struct kioctx default_kioctx;
243 extern int mmlist_nr;
245 struct sighand_struct {
247 struct k_sigaction action[_NSIG];
252 * NOTE! "signal_struct" does not have it's own
253 * locking, because a shared signal_struct always
254 * implies a shared sighand_struct, so locking
255 * sighand_struct is always a proper superset of
256 * the locking of signal_struct.
258 struct signal_struct {
261 /* current thread group signal load-balancing target: */
264 /* shared signal handling: */
265 struct sigpending shared_pending;
267 /* thread group exit support */
271 * - notify group_exit_task when ->count is equal to notify_count
272 * - everyone except group_exit_task is stopped during signal delivery
273 * of fatal signals, group_exit_task processes the signal.
275 struct task_struct *group_exit_task;
278 /* thread group stop support, overloads group_exit_code too */
279 int group_stop_count;
281 /* POSIX.1b Interval Timers */
282 struct list_head posix_timers;
284 /* job control IDs */
288 /* boolean value for session group leader */
291 struct tty_struct *tty; /* NULL if no tty */
295 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
296 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
297 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
298 * are inverted: lower p->prio value means higher priority.
300 * The MAX_USER_RT_PRIO value allows the actual maximum
301 * RT priority to be separate from the value exported to
302 * user-space. This allows kernel threads to set their
303 * priority to a value higher than any user task. Note:
304 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
307 #define MAX_USER_RT_PRIO 100
308 #define MAX_RT_PRIO MAX_USER_RT_PRIO
310 #define MAX_PRIO (MAX_RT_PRIO + 40)
312 #define rt_task(p) ((p)->prio < MAX_RT_PRIO)
315 * Some day this will be a full-fledged user tracking system..
318 atomic_t __count; /* reference count */
319 atomic_t processes; /* How many processes does this user have? */
320 atomic_t files; /* How many open files does this user have? */
321 atomic_t sigpending; /* How many pending signals does this user have? */
322 /* protected by mq_lock */
323 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
324 unsigned long locked_shm; /* How many pages of mlocked shm ? */
326 /* Hash table maintenance information */
327 struct list_head uidhash_list;
332 extern struct user_struct *find_user(xid_t, uid_t);
334 extern struct user_struct root_user;
335 #define INIT_USER (&root_user)
337 typedef struct prio_array prio_array_t;
338 struct backing_dev_info;
339 struct reclaim_state;
341 /* POSIX.1b interval timer structure. */
343 struct list_head list; /* free/ allocate list */
345 clockid_t it_clock; /* which timer type */
346 timer_t it_id; /* timer id */
347 int it_overrun; /* overrun on pending signal */
348 int it_overrun_last; /* overrun on last delivered signal */
349 int it_requeue_pending; /* waiting to requeue this timer */
350 int it_sigev_notify; /* notify word of sigevent struct */
351 int it_sigev_signo; /* signo word of sigevent struct */
352 sigval_t it_sigev_value; /* value word of sigevent struct */
353 unsigned long it_incr; /* interval specified in jiffies */
354 struct task_struct *it_process; /* process to send signal to */
355 struct timer_list it_timer;
356 struct sigqueue *sigq; /* signal queue entry. */
357 struct list_head abs_timer_entry; /* clock abs_timer_list */
358 struct timespec wall_to_prev; /* wall_to_monotonic used when set */
362 struct io_context; /* See blkdev.h */
363 void exit_io_context(void);
365 #define NGROUPS_SMALL 32
366 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
370 gid_t small_block[NGROUPS_SMALL];
376 * get_group_info() must be called with the owning task locked (via task_lock())
377 * when task != current. The reason being that the vast majority of callers are
378 * looking at current->group_info, which can not be changed except by the
379 * current task. Changing current->group_info requires the task lock, too.
381 #define get_group_info(group_info) do { \
382 atomic_inc(&(group_info)->usage); \
385 #define put_group_info(group_info) do { \
386 if (atomic_dec_and_test(&(group_info)->usage)) \
387 groups_free(group_info); \
390 struct group_info *groups_alloc(int gidsetsize);
391 void groups_free(struct group_info *group_info);
392 int set_current_groups(struct group_info *group_info);
393 /* access the groups "array" with this macro */
394 #define GROUP_AT(gi, i) \
395 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
398 struct audit_context; /* See audit.c */
401 #ifdef CONFIG_CKRM_CPU_SCHEDULE
403 * ckrm_cpu_demand_stat - used to track the cpu demand of a task/class
404 * @run: how much time it has been running since the counter started
405 * @total: total time since the counter started
406 * @last_sleep: the last time it sleeps, last_sleep = 0 when not sleeping
407 * @recalc_interval: how often do we recalculate the cpu_demand
408 * @cpu_demand: moving average of run/total
410 struct ckrm_cpu_demand_stat {
411 unsigned long long run;
412 unsigned long long total;
413 unsigned long long last_sleep;
414 unsigned long long recalc_interval;
415 unsigned long cpu_demand; /*estimated cpu demand */
420 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
421 struct thread_info *thread_info;
423 unsigned long flags; /* per process flags, defined below */
424 unsigned long ptrace;
426 int lock_depth; /* Lock depth */
428 int prio, static_prio;
429 struct list_head run_list;
432 unsigned long sleep_avg;
433 long interactive_credit;
434 unsigned long long timestamp;
437 unsigned long policy;
438 cpumask_t cpus_allowed;
439 unsigned int time_slice, first_time_slice;
441 struct list_head tasks;
442 struct list_head ptrace_children;
443 struct list_head ptrace_list;
445 struct mm_struct *mm, *active_mm;
448 struct linux_binfmt *binfmt;
449 int exit_code, exit_signal;
450 int pdeath_signal; /* The signal sent when the parent dies */
452 unsigned long personality;
457 * pointers to (original) parent process, youngest child, younger sibling,
458 * older sibling, respectively. (p->father can be replaced with
461 struct task_struct *real_parent; /* real parent process (when being debugged) */
462 struct task_struct *parent; /* parent process */
463 struct list_head children; /* list of my children */
464 struct list_head sibling; /* linkage in my parent's children list */
465 struct task_struct *group_leader; /* threadgroup leader */
467 /* PID/PID hash table linkage. */
468 struct pid_link pids[PIDTYPE_MAX];
470 wait_queue_head_t wait_chldexit; /* for wait4() */
471 struct completion *vfork_done; /* for vfork() */
472 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
473 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
475 unsigned long rt_priority;
476 unsigned long it_real_value, it_prof_value, it_virt_value;
477 unsigned long it_real_incr, it_prof_incr, it_virt_incr;
478 struct timer_list real_timer;
479 unsigned long utime, stime, cutime, cstime;
480 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw; /* context switch counts */
482 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
483 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
484 /* process credentials */
485 uid_t uid,euid,suid,fsuid;
486 gid_t gid,egid,sgid,fsgid;
487 struct group_info *group_info;
488 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
489 int keep_capabilities:1;
490 struct user_struct *user;
492 struct rlimit rlim[RLIM_NLIMITS];
493 unsigned short used_math;
495 /* file system info */
496 int link_count, total_link_count;
498 struct sysv_sem sysvsem;
499 /* CPU-specific state of this task */
500 struct thread_struct thread;
501 /* filesystem information */
502 struct fs_struct *fs;
503 /* open file information */
504 struct files_struct *files;
506 struct namespace *namespace;
507 /* signal handlers */
508 struct signal_struct *signal;
509 struct sighand_struct *sighand;
510 sigset_t blocked, real_blocked;
511 struct sigpending pending;
513 unsigned long sas_ss_sp;
515 int (*notifier)(void *priv);
517 sigset_t *notifier_mask;
520 struct audit_context *audit_context;
522 /* vserver context data */
524 struct vx_info *vx_info;
526 /* vserver network data */
528 struct nx_info *nx_info;
530 /* Thread group tracking */
533 /* Protection of (de-)allocation: mm, files, fs, tty */
534 spinlock_t alloc_lock;
535 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
536 spinlock_t proc_lock;
537 /* context-switch lock */
538 spinlock_t switch_lock;
540 /* journalling filesystem info */
544 struct reclaim_state *reclaim_state;
546 struct dentry *proc_dentry;
547 struct backing_dev_info *backing_dev_info;
549 struct io_context *io_context;
551 unsigned long ptrace_message;
552 siginfo_t *last_siginfo; /* For ptrace use. */
555 struct mempolicy *mempolicy;
556 short il_next; /* could be shared with used_math */
560 spinlock_t ckrm_tsklock;
562 #ifdef CONFIG_CKRM_TYPE_TASKCLASS
563 // .. Hubertus should change to CONFIG_CKRM_TYPE_TASKCLASS
564 struct ckrm_task_class *taskclass;
565 struct list_head taskclass_link;
566 #ifdef CONFIG_CKRM_CPU_SCHEDULE
567 struct ckrm_cpu_class *cpu_class;
568 //track cpu demand of this task
569 struct ckrm_cpu_demand_stat demand_stat;
570 #endif //CONFIG_CKRM_CPU_SCHEDULE
571 #endif // CONFIG_CKRM_TYPE_TASKCLASS
572 #endif // CONFIG_CKRM
574 struct task_delay_info delays;
577 static inline pid_t process_group(struct task_struct *tsk)
579 return tsk->signal->pgrp;
582 extern void __put_task_struct(struct task_struct *tsk);
583 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
584 #define put_task_struct(tsk) \
585 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
590 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
591 /* Not implemented yet, only for 486*/
592 #define PF_STARTING 0x00000002 /* being created */
593 #define PF_EXITING 0x00000004 /* getting shut down */
594 #define PF_DEAD 0x00000008 /* Dead */
595 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
596 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
597 #define PF_DUMPCORE 0x00000200 /* dumped core */
598 #define PF_SIGNALED 0x00000400 /* killed by a signal */
599 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
600 #define PF_MEMDIE 0x00001000 /* Killed for out-of-memory */
601 #define PF_FLUSHER 0x00002000 /* responsible for disk writeback */
603 #define PF_FREEZE 0x00004000 /* this task should be frozen for suspend */
604 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
605 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
606 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
607 #define PF_KSWAPD 0x00040000 /* I am kswapd */
608 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
609 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
610 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
611 #define PF_RELOCEXEC 0x00400000 /* relocate shared libraries */
614 #define PF_MEMIO 0x00400000 /* I am potentially doing I/O for mem */
615 #define PF_IOWAIT 0x00800000 /* I am waiting on disk I/O */
618 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
620 #define SD_BALANCE_NEWIDLE 1 /* Balance when about to become idle */
621 #define SD_BALANCE_EXEC 2 /* Balance on exec */
622 #define SD_BALANCE_CLONE 4 /* Balance on clone */
623 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
624 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
625 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
626 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
629 struct sched_group *next; /* Must be a circular list */
633 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
634 * single CPU. This should be read only (except for setup). Although
635 * it will need to be written to at cpu hot(un)plug time, perhaps the
636 * cpucontrol semaphore will provide enough exclusion?
638 unsigned long cpu_power;
641 struct sched_domain {
642 /* These fields must be setup */
643 struct sched_domain *parent; /* top domain must be null terminated */
644 struct sched_group *groups; /* the balancing groups of the domain */
645 cpumask_t span; /* span of all CPUs in this domain */
646 unsigned long min_interval; /* Minimum balance interval ms */
647 unsigned long max_interval; /* Maximum balance interval ms */
648 unsigned int busy_factor; /* less balancing by factor if busy */
649 unsigned int imbalance_pct; /* No balance until over watermark */
650 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
651 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
652 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
653 int flags; /* See SD_* */
655 /* Runtime fields. */
656 unsigned long last_balance; /* init to jiffies. units in jiffies */
657 unsigned int balance_interval; /* initialise to 1. units in ms. */
658 unsigned int nr_balance_failed; /* initialise to 0 */
661 /* Common values for SMT siblings */
662 #define SD_SIBLING_INIT (struct sched_domain) { \
663 .span = CPU_MASK_NONE, \
669 .imbalance_pct = 110, \
670 .cache_hot_time = 0, \
671 .cache_nice_tries = 0, \
672 .per_cpu_gain = 15, \
673 .flags = SD_BALANCE_NEWIDLE \
678 | SD_SHARE_CPUPOWER, \
679 .last_balance = jiffies, \
680 .balance_interval = 1, \
681 .nr_balance_failed = 0, \
684 /* Common values for CPUs */
685 #define SD_CPU_INIT (struct sched_domain) { \
686 .span = CPU_MASK_NONE, \
692 .imbalance_pct = 125, \
693 .cache_hot_time = (5*1000000/2), \
694 .cache_nice_tries = 1, \
695 .per_cpu_gain = 100, \
696 .flags = SD_BALANCE_NEWIDLE \
701 .last_balance = jiffies, \
702 .balance_interval = 1, \
703 .nr_balance_failed = 0, \
707 /* Common values for NUMA nodes */
708 #define SD_NODE_INIT (struct sched_domain) { \
709 .span = CPU_MASK_NONE, \
713 .max_interval = 32, \
715 .imbalance_pct = 125, \
716 .cache_hot_time = (10*1000000), \
717 .cache_nice_tries = 1, \
718 .per_cpu_gain = 100, \
719 .flags = SD_BALANCE_EXEC \
722 .last_balance = jiffies, \
723 .balance_interval = 1, \
724 .nr_balance_failed = 0, \
728 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
730 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
732 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
738 extern unsigned long long sched_clock(void);
741 extern void sched_balance_exec(void);
743 #define sched_balance_exec() {}
746 extern void sched_idle_next(void);
747 extern void set_user_nice(task_t *p, long nice);
748 extern int task_prio(const task_t *p);
749 extern int task_nice(const task_t *p);
750 extern int task_curr(const task_t *p);
751 extern int idle_cpu(int cpu);
756 * The default (Linux) execution domain.
758 extern struct exec_domain default_exec_domain;
761 struct thread_info thread_info;
762 unsigned long stack[THREAD_SIZE/sizeof(long)];
765 #ifndef __HAVE_ARCH_KSTACK_END
766 static inline int kstack_end(void *addr)
768 /* Reliable end of stack detection:
769 * Some APM bios versions misalign the stack
771 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
775 extern union thread_union init_thread_union;
776 extern struct task_struct init_task;
778 extern struct mm_struct init_mm;
780 extern struct task_struct *find_task_by_pid(int pid);
781 extern void set_special_pids(pid_t session, pid_t pgrp);
782 extern void __set_special_pids(pid_t session, pid_t pgrp);
784 /* per-UID process charging. */
785 extern struct user_struct * alloc_uid(xid_t, uid_t);
786 static inline struct user_struct *get_uid(struct user_struct *u)
788 atomic_inc(&u->__count);
791 extern void free_uid(struct user_struct *);
792 extern void switch_uid(struct user_struct *);
794 #include <asm/current.h>
796 extern unsigned long itimer_ticks;
797 extern unsigned long itimer_next;
798 extern void do_timer(struct pt_regs *);
800 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
801 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
802 extern void FASTCALL(wake_up_forked_process(struct task_struct * tsk));
804 extern void kick_process(struct task_struct *tsk);
805 extern void FASTCALL(wake_up_forked_thread(struct task_struct * tsk));
807 static inline void kick_process(struct task_struct *tsk) { }
808 static inline void wake_up_forked_thread(struct task_struct * tsk)
810 wake_up_forked_process(tsk);
813 extern void FASTCALL(sched_fork(task_t * p));
814 extern void FASTCALL(sched_exit(task_t * p));
816 extern int in_group_p(gid_t);
817 extern int in_egroup_p(gid_t);
819 extern void proc_caches_init(void);
820 extern void flush_signals(struct task_struct *);
821 extern void flush_signal_handlers(struct task_struct *, int force_default);
822 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
824 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
829 spin_lock_irqsave(&tsk->sighand->siglock, flags);
830 ret = dequeue_signal(tsk, mask, info);
831 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
836 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
838 extern void unblock_all_signals(void);
839 extern void release_task(struct task_struct * p);
840 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
841 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
842 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
843 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
844 extern int kill_pg_info(int, struct siginfo *, pid_t);
845 extern int kill_sl_info(int, struct siginfo *, pid_t);
846 extern int kill_proc_info(int, struct siginfo *, pid_t);
847 extern void notify_parent(struct task_struct *, int);
848 extern void do_notify_parent(struct task_struct *, int);
849 extern void force_sig(int, struct task_struct *);
850 extern void force_sig_specific(int, struct task_struct *);
851 extern int send_sig(int, struct task_struct *, int);
852 extern void zap_other_threads(struct task_struct *p);
853 extern int kill_pg(pid_t, int, int);
854 extern int kill_sl(pid_t, int, int);
855 extern int kill_proc(pid_t, int, int);
856 extern struct sigqueue *sigqueue_alloc(void);
857 extern void sigqueue_free(struct sigqueue *);
858 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
859 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
860 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
861 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
863 /* These can be the second arg to send_sig_info/send_group_sig_info. */
864 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
865 #define SEND_SIG_PRIV ((struct siginfo *) 1)
866 #define SEND_SIG_FORCED ((struct siginfo *) 2)
868 /* True if we are on the alternate signal stack. */
870 static inline int on_sig_stack(unsigned long sp)
872 return (sp - current->sas_ss_sp < current->sas_ss_size);
875 static inline int sas_ss_flags(unsigned long sp)
877 return (current->sas_ss_size == 0 ? SS_DISABLE
878 : on_sig_stack(sp) ? SS_ONSTACK : 0);
882 #ifdef CONFIG_SECURITY
883 /* code is in security.c */
884 extern int capable(int cap);
886 static inline int capable(int cap)
888 if (cap_raised(current->cap_effective, cap)) {
889 current->flags |= PF_SUPERPRIV;
898 * Routines for handling mm_structs
900 extern struct mm_struct * mm_alloc(void);
902 /* mmdrop drops the mm and the page tables */
903 extern void FASTCALL(__mmdrop(struct mm_struct *));
904 static inline void mmdrop(struct mm_struct * mm)
906 if (atomic_dec_and_test(&mm->mm_count))
910 /* mmput gets rid of the mappings and all user-space */
911 extern void mmput(struct mm_struct *);
912 /* Grab a reference to the mm if its not already going away */
913 extern struct mm_struct *mmgrab(struct mm_struct *);
914 /* Remove the current tasks stale references to the old mm_struct */
915 extern void mm_release(struct task_struct *, struct mm_struct *);
917 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
918 extern void flush_thread(void);
919 extern void exit_thread(void);
921 extern void exit_mm(struct task_struct *);
922 extern void exit_files(struct task_struct *);
923 extern void exit_signal(struct task_struct *);
924 extern void __exit_signal(struct task_struct *);
925 extern void exit_sighand(struct task_struct *);
926 extern void __exit_sighand(struct task_struct *);
927 extern void exit_itimers(struct signal_struct *);
929 extern NORET_TYPE void do_group_exit(int);
931 extern void reparent_to_init(void);
932 extern void daemonize(const char *, ...);
933 extern int allow_signal(int);
934 extern int disallow_signal(int);
935 extern task_t *child_reaper;
937 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
938 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
939 extern struct task_struct * copy_process(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
942 extern void wait_task_inactive(task_t * p);
944 #define wait_task_inactive(p) do { } while (0)
947 #define remove_parent(p) list_del_init(&(p)->sibling)
948 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
950 #define REMOVE_LINKS(p) do { \
951 if (thread_group_leader(p)) \
952 list_del_init(&(p)->tasks); \
956 #define SET_LINKS(p) do { \
957 if (thread_group_leader(p)) \
958 list_add_tail(&(p)->tasks,&init_task.tasks); \
959 add_parent(p, (p)->parent); \
962 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
963 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
965 #define for_each_process(p) \
966 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
969 * Careful: do_each_thread/while_each_thread is a double loop so
970 * 'break' will not work as expected - use goto instead.
972 #define do_each_thread(g, t) \
973 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
975 #define while_each_thread(g, t) \
976 while ((t = next_thread(t)) != g)
978 extern task_t * FASTCALL(next_thread(const task_t *p));
980 #define thread_group_leader(p) (p->pid == p->tgid)
982 static inline int thread_group_empty(task_t *p)
984 struct pid *pid = p->pids[PIDTYPE_TGID].pidptr;
986 return pid->task_list.next->next == &pid->task_list;
989 #define delay_group_leader(p) \
990 (thread_group_leader(p) && !thread_group_empty(p))
992 extern void unhash_process(struct task_struct *p);
995 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info and synchronises with
998 * Nests both inside and outside of read_lock(&tasklist_lock).
999 * It must not be nested with write_lock_irq(&tasklist_lock),
1000 * neither inside nor outside.
1002 static inline void task_lock(struct task_struct *p)
1004 spin_lock(&p->alloc_lock);
1007 static inline void task_unlock(struct task_struct *p)
1009 spin_unlock(&p->alloc_lock);
1013 * get_task_mm - acquire a reference to the task's mm
1015 * Returns %NULL if the task has no mm. User must release
1016 * the mm via mmput() after use.
1018 static inline struct mm_struct * get_task_mm(struct task_struct * task)
1020 struct mm_struct * mm;
1031 /* set thread flags in other task's structures
1032 * - see asm/thread_info.h for TIF_xxxx flags available
1034 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1036 set_ti_thread_flag(tsk->thread_info,flag);
1039 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1041 clear_ti_thread_flag(tsk->thread_info,flag);
1044 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1046 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1049 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1051 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1054 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1056 return test_ti_thread_flag(tsk->thread_info,flag);
1059 static inline void set_tsk_need_resched(struct task_struct *tsk)
1061 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1064 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1066 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1069 static inline int signal_pending(struct task_struct *p)
1071 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1074 static inline int need_resched(void)
1076 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1079 extern void __cond_resched(void);
1081 static inline void cond_resched(void)
1083 #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
1084 __might_sleep(__FILE__, __LINE__, 0);
1090 * cond_resched_lock() - if a reschedule is pending, drop the given lock,
1091 * call schedule, and on return reacquire the lock.
1093 * This works OK both with and without CONFIG_PREEMPT. We do strange low-level
1094 * operations here to prevent schedule() from being called twice (once via
1095 * spin_unlock(), once by hand).
1097 extern void __cond_resched_lock(spinlock_t * lock);
1099 static inline void cond_resched_lock(spinlock_t * lock)
1101 #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
1102 __might_sleep(__FILE__, __LINE__, 1);
1104 __cond_resched_lock(lock);
1107 /* Reevaluate whether the task has signals pending delivery.
1108 This is required every time the blocked sigset_t changes.
1109 callers must hold sighand->siglock. */
1111 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1112 extern void recalc_sigpending(void);
1114 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1117 * Wrappers for p->thread_info->cpu access. No-op on UP.
1121 static inline unsigned int task_cpu(const struct task_struct *p)
1123 return p->thread_info->cpu;
1126 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1128 p->thread_info->cpu = cpu;
1133 static inline unsigned int task_cpu(const struct task_struct *p)
1138 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1142 #endif /* CONFIG_SMP */
1145 /* API for registering delay info */
1146 #ifdef CONFIG_DELAY_ACCT
1148 #define test_delay_flag(tsk,flg) ((tsk)->flags & (flg))
1149 #define set_delay_flag(tsk,flg) ((tsk)->flags |= (flg))
1150 #define clear_delay_flag(tsk,flg) ((tsk)->flags &= ~(flg))
1152 #define def_delay_var(var) unsigned long long var
1153 #define get_delay(tsk,field) ((tsk)->delays.field)
1154 #define delay_value(x) (((unsigned long)(x))/1000)
1156 #define start_delay(var) ((var) = sched_clock())
1157 #define start_delay_set(var,flg) (set_delay_flag(current,flg),(var) = sched_clock())
1159 #define inc_delay(tsk,field) (((tsk)->delays.field)++)
1160 #define add_delay_ts(tsk,field,start_ts,end_ts) ((tsk)->delays.field += delay_value((end_ts)-(start_ts)))
1161 #define add_delay_clear(tsk,field,start_ts,flg) (add_delay_ts(tsk,field,start_ts,sched_clock()),clear_delay_flag(tsk,flg))
1163 static inline void add_io_delay(unsigned long dstart)
1165 struct task_struct * tsk = current;
1166 unsigned long val = delay_value(sched_clock()-dstart);
1167 if (test_delay_flag(tsk,PF_MEMIO)) {
1168 tsk->delays.mem_iowait_total += val;
1169 tsk->delays.num_memwaits++;
1171 tsk->delays.iowait_total += val;
1172 tsk->delays.num_iowaits++;
1174 clear_delay_flag(tsk,PF_IOWAIT);
1177 inline static void init_delays(struct task_struct *tsk)
1179 memset((void*)&tsk->delays,0,sizeof(tsk->delays));
1184 #define test_delay_flag(tsk,flg) (0)
1185 #define set_delay_flag(tsk,flg) do { } while (0)
1186 #define clear_delay_flag(tsk,flg) do { } while (0)
1188 #define def_delay_var(var)
1189 #define get_delay(tsk,field) (0)
1191 #define start_delay(var) do { } while (0)
1192 #define start_delay_set(var,flg) do { } while (0)
1194 #define inc_delay(tsk,field) do { } while (0)
1195 #define add_delay_ts(tsk,field,start_ts,now) do { } while (0)
1196 #define add_delay_clear(tsk,field,start_ts,flg) do { } while (0)
1197 #define add_io_delay(dstart) do { } while (0)
1198 #define init_delays(tsk) do { } while (0)
1203 #endif /* __KERNEL__ */