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 DECLARE_PER_CPU(struct runqueue, runqueues);
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>
109 #include <linux/vserver/context.h>
111 #define TASK_RUNNING 0
112 #define TASK_INTERRUPTIBLE 1
113 #define TASK_UNINTERRUPTIBLE 2
114 #define TASK_STOPPED 4
115 #define TASK_ZOMBIE 8
117 #define TASK_ONHOLD 32
119 #define __set_task_state(tsk, state_value) \
120 do { (tsk)->state = (state_value); } while (0)
121 #define set_task_state(tsk, state_value) \
122 set_mb((tsk)->state, (state_value))
124 #define __set_current_state(state_value) \
125 do { current->state = (state_value); } while (0)
126 #define set_current_state(state_value) \
127 set_mb(current->state, (state_value))
130 * Scheduling policies
132 #define SCHED_NORMAL 0
142 #include <linux/taskdelays.h>
143 #include <linux/spinlock.h>
146 * This serializes "schedule()" and also protects
147 * the run-queue from deletions/modifications (but
148 * _adding_ to the beginning of the run-queue has
151 extern rwlock_t tasklist_lock;
152 extern spinlock_t mmlist_lock;
154 typedef struct task_struct task_t;
156 extern void sched_init(void);
157 extern void sched_init_smp(void);
158 extern void init_idle(task_t *idle, int cpu);
160 extern cpumask_t nohz_cpu_mask;
162 extern void show_state(void);
163 extern void show_regs(struct pt_regs *);
166 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
167 * task), SP is the stack pointer of the first frame that should be shown in the back
168 * trace (or NULL if the entire call-chain of the task should be shown).
170 extern void show_stack(struct task_struct *task, unsigned long *sp);
172 void io_schedule(void);
173 long io_schedule_timeout(long timeout);
175 extern void cpu_init (void);
176 extern void trap_init(void);
177 extern void update_process_times(int user);
178 extern void scheduler_tick(int user_tick, int system);
179 extern unsigned long cache_decay_ticks;
181 /* Attach to any functions which should be ignored in wchan output. */
182 #define __sched __attribute__((__section__(".sched.text")))
183 /* Is this address in the __sched functions? */
184 extern int in_sched_functions(unsigned long addr);
186 #define MAX_SCHEDULE_TIMEOUT LONG_MAX
187 extern signed long FASTCALL(schedule_timeout(signed long timeout));
188 asmlinkage void schedule(void);
192 /* Maximum number of active map areas.. This is a random (large) number */
193 #define DEFAULT_MAX_MAP_COUNT 65536
195 extern int sysctl_max_map_count;
197 #include <linux/aio.h>
200 arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
201 unsigned long, unsigned long);
204 arch_get_unmapped_exec_area(struct file *, unsigned long, unsigned long,
205 unsigned long, unsigned long);
207 arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
208 unsigned long len, unsigned long pgoff,
209 unsigned long flags);
210 extern void arch_unmap_area(struct vm_area_struct *area);
211 extern void arch_unmap_area_topdown(struct vm_area_struct *area);
215 struct vm_area_struct * mmap; /* list of VMAs */
216 struct rb_root mm_rb;
217 struct vm_area_struct * mmap_cache; /* last find_vma result */
218 unsigned long (*get_unmapped_area) (struct file *filp,
219 unsigned long addr, unsigned long len,
220 unsigned long pgoff, unsigned long flags);
221 unsigned long (*get_unmapped_exec_area) (struct file *filp,
222 unsigned long addr, unsigned long len,
223 unsigned long pgoff, unsigned long flags);
224 void (*unmap_area) (struct vm_area_struct *area);
225 unsigned long mmap_base; /* base of mmap area */
226 unsigned long free_area_cache; /* first hole */
228 atomic_t mm_users; /* How many users with user space? */
229 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
230 int map_count; /* number of VMAs */
231 struct rw_semaphore mmap_sem;
232 spinlock_t page_table_lock; /* Protects task page tables and mm->rss */
234 struct list_head mmlist; /* List of all active mm's. These are globally strung
235 * together off init_mm.mmlist, and are protected
239 unsigned long start_code, end_code, start_data, end_data;
240 unsigned long start_brk, brk, start_stack;
241 unsigned long arg_start, arg_end, env_start, env_end;
242 unsigned long rss, total_vm, locked_vm;
243 unsigned long def_flags;
245 unsigned long saved_auxv[40]; /* for /proc/PID/auxv */
248 cpumask_t cpu_vm_mask;
250 /* Architecture-specific MM context */
251 mm_context_t context;
252 struct vx_info *mm_vx_info;
254 /* Token based thrashing protection. */
255 unsigned long swap_token_time;
258 /* coredumping support */
260 struct completion *core_startup_done, core_done;
263 rwlock_t ioctx_list_lock;
264 struct kioctx *ioctx_list;
266 struct kioctx default_kioctx;
269 extern int mmlist_nr;
271 struct sighand_struct {
273 struct k_sigaction action[_NSIG];
278 * NOTE! "signal_struct" does not have it's own
279 * locking, because a shared signal_struct always
280 * implies a shared sighand_struct, so locking
281 * sighand_struct is always a proper superset of
282 * the locking of signal_struct.
284 struct signal_struct {
287 /* current thread group signal load-balancing target: */
290 /* shared signal handling: */
291 struct sigpending shared_pending;
293 /* thread group exit support */
297 * - notify group_exit_task when ->count is equal to notify_count
298 * - everyone except group_exit_task is stopped during signal delivery
299 * of fatal signals, group_exit_task processes the signal.
301 struct task_struct *group_exit_task;
304 /* thread group stop support, overloads group_exit_code too */
305 int group_stop_count;
307 /* POSIX.1b Interval Timers */
308 struct list_head posix_timers;
310 /* job control IDs */
314 /* boolean value for session group leader */
317 struct tty_struct *tty; /* NULL if no tty */
321 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
322 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL tasks are
323 * in the range MAX_RT_PRIO..MAX_PRIO-1. Priority values
324 * are inverted: lower p->prio value means higher priority.
326 * The MAX_USER_RT_PRIO value allows the actual maximum
327 * RT priority to be separate from the value exported to
328 * user-space. This allows kernel threads to set their
329 * priority to a value higher than any user task. Note:
330 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
333 #define MAX_USER_RT_PRIO 100
334 #define MAX_RT_PRIO MAX_USER_RT_PRIO
336 #define MAX_PRIO (MAX_RT_PRIO + 40)
338 #define rt_task(p) ((p)->prio < MAX_RT_PRIO)
341 * Some day this will be a full-fledged user tracking system..
344 atomic_t __count; /* reference count */
345 atomic_t processes; /* How many processes does this user have? */
346 atomic_t files; /* How many open files does this user have? */
347 atomic_t sigpending; /* How many pending signals does this user have? */
348 /* protected by mq_lock */
349 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
350 unsigned long locked_shm; /* How many pages of mlocked shm ? */
352 /* Hash table maintenance information */
353 struct list_head uidhash_list;
358 extern struct user_struct *find_user(xid_t, uid_t);
360 extern struct user_struct root_user;
361 #define INIT_USER (&root_user)
363 typedef struct prio_array prio_array_t;
364 struct backing_dev_info;
365 struct reclaim_state;
367 /* POSIX.1b interval timer structure. */
369 struct list_head list; /* free/ allocate list */
371 clockid_t it_clock; /* which timer type */
372 timer_t it_id; /* timer id */
373 int it_overrun; /* overrun on pending signal */
374 int it_overrun_last; /* overrun on last delivered signal */
375 int it_requeue_pending; /* waiting to requeue this timer */
376 int it_sigev_notify; /* notify word of sigevent struct */
377 int it_sigev_signo; /* signo word of sigevent struct */
378 sigval_t it_sigev_value; /* value word of sigevent struct */
379 unsigned long it_incr; /* interval specified in jiffies */
380 struct task_struct *it_process; /* process to send signal to */
381 struct timer_list it_timer;
382 struct sigqueue *sigq; /* signal queue entry. */
383 struct list_head abs_timer_entry; /* clock abs_timer_list */
384 struct timespec wall_to_prev; /* wall_to_monotonic used when set */
388 struct io_context; /* See blkdev.h */
389 void exit_io_context(void);
391 #define NGROUPS_SMALL 32
392 #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
396 gid_t small_block[NGROUPS_SMALL];
402 * get_group_info() must be called with the owning task locked (via task_lock())
403 * when task != current. The reason being that the vast majority of callers are
404 * looking at current->group_info, which can not be changed except by the
405 * current task. Changing current->group_info requires the task lock, too.
407 #define get_group_info(group_info) do { \
408 atomic_inc(&(group_info)->usage); \
411 #define put_group_info(group_info) do { \
412 if (atomic_dec_and_test(&(group_info)->usage)) \
413 groups_free(group_info); \
416 struct group_info *groups_alloc(int gidsetsize);
417 void groups_free(struct group_info *group_info);
418 int set_current_groups(struct group_info *group_info);
419 /* access the groups "array" with this macro */
420 #define GROUP_AT(gi, i) \
421 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
424 struct audit_context; /* See audit.c */
428 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
429 struct thread_info *thread_info;
431 unsigned long flags; /* per process flags, defined below */
432 unsigned long ptrace;
434 int lock_depth; /* Lock depth */
436 int prio, static_prio;
437 struct list_head run_list;
440 unsigned long sleep_avg;
441 long interactive_credit;
442 unsigned long long timestamp;
445 unsigned long policy;
446 cpumask_t cpus_allowed;
447 unsigned int time_slice, first_time_slice;
449 struct list_head tasks;
451 * ptrace_list/ptrace_children forms the list of my children
452 * that were stolen by a ptracer.
454 struct list_head ptrace_children;
455 struct list_head ptrace_list;
457 struct mm_struct *mm, *active_mm;
460 struct linux_binfmt *binfmt;
461 int exit_code, exit_signal;
462 int pdeath_signal; /* The signal sent when the parent dies */
464 unsigned long personality;
469 * pointers to (original) parent process, youngest child, younger sibling,
470 * older sibling, respectively. (p->father can be replaced with
473 struct task_struct *real_parent; /* real parent process (when being debugged) */
474 struct task_struct *parent; /* parent process */
476 * children/sibling forms the list of my children plus the
477 * tasks I'm ptracing.
479 struct list_head children; /* list of my children */
480 struct list_head sibling; /* linkage in my parent's children list */
481 struct task_struct *group_leader; /* threadgroup leader */
483 /* PID/PID hash table linkage. */
484 struct pid_link pids[PIDTYPE_MAX];
486 wait_queue_head_t wait_chldexit; /* for wait4() */
487 struct completion *vfork_done; /* for vfork() */
488 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
489 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
491 unsigned long rt_priority;
492 unsigned long it_real_value, it_prof_value, it_virt_value;
493 unsigned long it_real_incr, it_prof_incr, it_virt_incr;
494 struct timer_list real_timer;
495 unsigned long utime, stime, cutime, cstime;
496 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw; /* context switch counts */
498 /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
499 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
500 /* process credentials */
501 uid_t uid,euid,suid,fsuid;
502 gid_t gid,egid,sgid,fsgid;
503 struct group_info *group_info;
504 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
505 int keep_capabilities:1;
506 struct user_struct *user;
508 struct rlimit rlim[RLIM_NLIMITS];
509 unsigned short used_math;
511 /* file system info */
512 int link_count, total_link_count;
514 struct sysv_sem sysvsem;
515 /* CPU-specific state of this task */
516 struct thread_struct thread;
517 /* filesystem information */
518 struct fs_struct *fs;
519 /* open file information */
520 struct files_struct *files;
522 struct namespace *namespace;
523 /* signal handlers */
524 struct signal_struct *signal;
525 struct sighand_struct *sighand;
527 sigset_t blocked, real_blocked;
528 struct sigpending pending;
530 unsigned long sas_ss_sp;
532 int (*notifier)(void *priv);
534 sigset_t *notifier_mask;
538 void (*tux_exit)(void);
542 struct audit_context *audit_context;
544 /* vserver context data */
546 struct vx_info *vx_info;
548 /* vserver network data */
550 struct nx_info *nx_info;
552 /* Thread group tracking */
555 /* Protection of (de-)allocation: mm, files, fs, tty */
556 spinlock_t alloc_lock;
557 /* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
558 spinlock_t proc_lock;
559 /* context-switch lock */
560 spinlock_t switch_lock;
562 /* journalling filesystem info */
566 struct reclaim_state *reclaim_state;
568 struct dentry *proc_dentry;
569 struct backing_dev_info *backing_dev_info;
571 struct io_context *io_context;
573 unsigned long ptrace_message;
574 siginfo_t *last_siginfo; /* For ptrace use. */
577 struct mempolicy *mempolicy;
578 short il_next; /* could be shared with used_math */
582 spinlock_t ckrm_tsklock;
584 #ifdef CONFIG_CKRM_TYPE_TASKCLASS
585 // .. Hubertus should change to CONFIG_CKRM_TYPE_TASKCLASS
586 struct ckrm_task_class *taskclass;
587 struct list_head taskclass_link;
588 #ifdef CONFIG_CKRM_CPU_SCHEDULE
589 struct ckrm_cpu_class *cpu_class;
591 #endif // CONFIG_CKRM_TYPE_TASKCLASS
592 #endif // CONFIG_CKRM
594 struct task_delay_info delays;
597 static inline pid_t process_group(struct task_struct *tsk)
599 return tsk->signal->pgrp;
602 extern void __put_task_struct(struct task_struct *tsk);
603 #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
604 #define put_task_struct(tsk) \
605 do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
610 #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
611 /* Not implemented yet, only for 486*/
612 #define PF_STARTING 0x00000002 /* being created */
613 #define PF_EXITING 0x00000004 /* getting shut down */
614 #define PF_DEAD 0x00000008 /* Dead */
615 #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
616 #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
617 #define PF_DUMPCORE 0x00000200 /* dumped core */
618 #define PF_SIGNALED 0x00000400 /* killed by a signal */
619 #define PF_MEMALLOC 0x00000800 /* Allocating memory */
620 #define PF_MEMDIE 0x00001000 /* Killed for out-of-memory */
621 #define PF_FLUSHER 0x00002000 /* responsible for disk writeback */
623 #define PF_FREEZE 0x00004000 /* this task should be frozen for suspend */
624 #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
625 #define PF_FROZEN 0x00010000 /* frozen for system suspend */
626 #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
627 #define PF_KSWAPD 0x00040000 /* I am kswapd */
628 #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
629 #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
630 #define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
631 #define PF_RELOCEXEC 0x00400000 /* relocate shared libraries */
634 #define PF_MEMIO 0x00400000 /* I am potentially doing I/O for mem */
635 #define PF_IOWAIT 0x00800000 /* I am waiting on disk I/O */
638 #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
640 #define SD_BALANCE_NEWIDLE 1 /* Balance when about to become idle */
641 #define SD_BALANCE_EXEC 2 /* Balance on exec */
642 #define SD_BALANCE_CLONE 4 /* Balance on clone */
643 #define SD_WAKE_IDLE 8 /* Wake to idle CPU on task wakeup */
644 #define SD_WAKE_AFFINE 16 /* Wake task to waking CPU */
645 #define SD_WAKE_BALANCE 32 /* Perform balancing at task wakeup */
646 #define SD_SHARE_CPUPOWER 64 /* Domain members share cpu power */
649 struct sched_group *next; /* Must be a circular list */
653 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
654 * single CPU. This should be read only (except for setup). Although
655 * it will need to be written to at cpu hot(un)plug time, perhaps the
656 * cpucontrol semaphore will provide enough exclusion?
658 unsigned long cpu_power;
661 struct sched_domain {
662 /* These fields must be setup */
663 struct sched_domain *parent; /* top domain must be null terminated */
664 struct sched_group *groups; /* the balancing groups of the domain */
665 cpumask_t span; /* span of all CPUs in this domain */
666 unsigned long min_interval; /* Minimum balance interval ms */
667 unsigned long max_interval; /* Maximum balance interval ms */
668 unsigned int busy_factor; /* less balancing by factor if busy */
669 unsigned int imbalance_pct; /* No balance until over watermark */
670 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
671 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
672 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
673 int flags; /* See SD_* */
675 /* Runtime fields. */
676 unsigned long last_balance; /* init to jiffies. units in jiffies */
677 unsigned int balance_interval; /* initialise to 1. units in ms. */
678 unsigned int nr_balance_failed; /* initialise to 0 */
681 /* Common values for SMT siblings */
682 #define SD_SIBLING_INIT (struct sched_domain) { \
683 .span = CPU_MASK_NONE, \
689 .imbalance_pct = 110, \
690 .cache_hot_time = 0, \
691 .cache_nice_tries = 0, \
692 .per_cpu_gain = 15, \
693 .flags = SD_BALANCE_NEWIDLE \
698 | SD_SHARE_CPUPOWER, \
699 .last_balance = jiffies, \
700 .balance_interval = 1, \
701 .nr_balance_failed = 0, \
704 /* Common values for CPUs */
705 #define SD_CPU_INIT (struct sched_domain) { \
706 .span = CPU_MASK_NONE, \
712 .imbalance_pct = 125, \
713 .cache_hot_time = (5*1000000/2), \
714 .cache_nice_tries = 1, \
715 .per_cpu_gain = 100, \
716 .flags = SD_BALANCE_NEWIDLE \
721 .last_balance = jiffies, \
722 .balance_interval = 1, \
723 .nr_balance_failed = 0, \
727 /* Common values for NUMA nodes */
728 #define SD_NODE_INIT (struct sched_domain) { \
729 .span = CPU_MASK_NONE, \
733 .max_interval = 32, \
735 .imbalance_pct = 125, \
736 .cache_hot_time = (10*1000000), \
737 .cache_nice_tries = 1, \
738 .per_cpu_gain = 100, \
739 .flags = SD_BALANCE_EXEC \
742 .last_balance = jiffies, \
743 .balance_interval = 1, \
744 .nr_balance_failed = 0, \
748 extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
750 extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
752 static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
758 extern unsigned long long sched_clock(void);
761 extern void sched_balance_exec(void);
763 #define sched_balance_exec() {}
766 extern void sched_idle_next(void);
767 extern void set_user_nice(task_t *p, long nice);
768 extern int task_prio(const task_t *p);
769 extern int task_nice(const task_t *p);
770 extern int task_curr(const task_t *p);
771 extern int idle_cpu(int cpu);
776 * These are the runqueue data structures:
778 typedef struct runqueue runqueue_t;
780 #ifdef CONFIG_CKRM_CPU_SCHEDULE
781 #include <linux/ckrm_classqueue.h>
784 #ifdef CONFIG_CKRM_CPU_SCHEDULE
787 * if belong to different class, compare class priority
788 * otherwise compare task priority
790 #define TASK_PREEMPTS_CURR(p, rq) \
791 (((p)->cpu_class != (rq)->curr->cpu_class) && ((rq)->curr != (rq)->idle))? class_preempts_curr((p),(rq)->curr) : ((p)->prio < (rq)->curr->prio)
793 #define BITMAP_SIZE ((((MAX_PRIO+1+7)/8)+sizeof(long)-1)/sizeof(long))
795 unsigned int nr_active;
796 unsigned long bitmap[BITMAP_SIZE];
797 struct list_head queue[MAX_PRIO];
799 #define rq_active(p,rq) (rq->active)
800 #define rq_expired(p,rq) (rq->expired)
801 #define ckrm_rebalance_tick(j,this_cpu) do {} while (0)
802 #define TASK_PREEMPTS_CURR(p, rq) \
803 ((p)->prio < (rq)->curr->prio)
807 * This is the main, per-CPU runqueue data structure.
809 * Locking rule: those places that want to lock multiple runqueues
810 * (such as the load balancing or the thread migration code), lock
811 * acquire operations must be ordered by ascending &runqueue.
817 * nr_running and cpu_load should be in the same cacheline because
818 * remote CPUs use both these fields when doing load calculation.
820 unsigned long nr_running;
821 #if defined(CONFIG_SMP)
822 unsigned long cpu_load;
824 unsigned long long nr_switches, nr_preempt;
825 unsigned long expired_timestamp, nr_uninterruptible;
826 unsigned long long timestamp_last_tick;
828 struct mm_struct *prev_mm;
829 #ifdef CONFIG_CKRM_CPU_SCHEDULE
830 unsigned long ckrm_cpu_load;
831 struct classqueue_struct classqueue;
833 prio_array_t *active, *expired, arrays[2];
835 int best_expired_prio;
839 struct sched_domain *sd;
841 /* For active balancing */
845 task_t *migration_thread;
846 struct list_head migration_queue;
848 struct list_head hold_queue;
853 * The default (Linux) execution domain.
855 extern struct exec_domain default_exec_domain;
858 struct thread_info thread_info;
859 unsigned long stack[THREAD_SIZE/sizeof(long)];
862 #ifndef __HAVE_ARCH_KSTACK_END
863 static inline int kstack_end(void *addr)
865 /* Reliable end of stack detection:
866 * Some APM bios versions misalign the stack
868 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
872 extern union thread_union init_thread_union;
873 extern struct task_struct init_task;
875 extern struct mm_struct init_mm;
877 extern struct task_struct *find_task_by_pid(int pid);
878 extern void set_special_pids(pid_t session, pid_t pgrp);
879 extern void __set_special_pids(pid_t session, pid_t pgrp);
881 /* per-UID process charging. */
882 extern struct user_struct * alloc_uid(xid_t, uid_t);
883 static inline struct user_struct *get_uid(struct user_struct *u)
885 atomic_inc(&u->__count);
888 extern void free_uid(struct user_struct *);
889 extern void switch_uid(struct user_struct *);
891 #include <asm/current.h>
893 extern unsigned long itimer_ticks;
894 extern unsigned long itimer_next;
895 extern void do_timer(struct pt_regs *);
897 extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
898 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
899 extern void FASTCALL(wake_up_forked_process(struct task_struct * tsk));
901 extern void kick_process(struct task_struct *tsk);
902 extern void FASTCALL(wake_up_forked_thread(struct task_struct * tsk));
904 static inline void kick_process(struct task_struct *tsk) { }
905 static inline void wake_up_forked_thread(struct task_struct * tsk)
907 wake_up_forked_process(tsk);
910 extern void FASTCALL(sched_fork(task_t * p));
911 extern void FASTCALL(sched_exit(task_t * p));
913 extern int in_group_p(gid_t);
914 extern int in_egroup_p(gid_t);
916 extern void proc_caches_init(void);
917 extern void flush_signals(struct task_struct *);
918 extern void flush_signal_handlers(struct task_struct *, int force_default);
919 extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
921 static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
926 spin_lock_irqsave(&tsk->sighand->siglock, flags);
927 ret = dequeue_signal(tsk, mask, info);
928 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
933 extern void block_all_signals(int (*notifier)(void *priv), void *priv,
935 extern void unblock_all_signals(void);
936 extern void release_task(struct task_struct * p);
937 extern int send_sig_info(int, struct siginfo *, struct task_struct *);
938 extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
939 extern int force_sig_info(int, struct siginfo *, struct task_struct *);
940 extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
941 extern int kill_pg_info(int, struct siginfo *, pid_t);
942 extern int kill_sl_info(int, struct siginfo *, pid_t);
943 extern int kill_proc_info(int, struct siginfo *, pid_t);
944 extern void notify_parent(struct task_struct *, int);
945 extern void do_notify_parent(struct task_struct *, int);
946 extern void force_sig(int, struct task_struct *);
947 extern void force_sig_specific(int, struct task_struct *);
948 extern int send_sig(int, struct task_struct *, int);
949 extern void zap_other_threads(struct task_struct *p);
950 extern int kill_pg(pid_t, int, int);
951 extern int kill_sl(pid_t, int, int);
952 extern int kill_proc(pid_t, int, int);
953 extern struct sigqueue *sigqueue_alloc(void);
954 extern void sigqueue_free(struct sigqueue *);
955 extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
956 extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
957 extern int do_sigaction(int, const struct k_sigaction *, struct k_sigaction *);
958 extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
960 /* These can be the second arg to send_sig_info/send_group_sig_info. */
961 #define SEND_SIG_NOINFO ((struct siginfo *) 0)
962 #define SEND_SIG_PRIV ((struct siginfo *) 1)
963 #define SEND_SIG_FORCED ((struct siginfo *) 2)
965 /* True if we are on the alternate signal stack. */
967 static inline int on_sig_stack(unsigned long sp)
969 return (sp - current->sas_ss_sp < current->sas_ss_size);
972 static inline int sas_ss_flags(unsigned long sp)
974 return (current->sas_ss_size == 0 ? SS_DISABLE
975 : on_sig_stack(sp) ? SS_ONSTACK : 0);
979 #ifdef CONFIG_SECURITY
980 /* code is in security.c */
981 extern int capable(int cap);
983 static inline int capable(int cap)
985 if (cap_raised(current->cap_effective, cap)) {
986 current->flags |= PF_SUPERPRIV;
994 * Routines for handling mm_structs
996 extern struct mm_struct * mm_alloc(void);
998 /* mmdrop drops the mm and the page tables */
999 extern void FASTCALL(__mmdrop(struct mm_struct *));
1000 static inline void mmdrop(struct mm_struct * mm)
1002 if (atomic_dec_and_test(&mm->mm_count))
1006 /* mmput gets rid of the mappings and all user-space */
1007 extern void mmput(struct mm_struct *);
1008 /* Grab a reference to the mm if its not already going away */
1009 extern struct mm_struct *mmgrab(struct mm_struct *);
1010 /* Remove the current tasks stale references to the old mm_struct */
1011 extern void mm_release(struct task_struct *, struct mm_struct *);
1013 extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1014 extern void flush_thread(void);
1015 extern void exit_thread(void);
1017 extern void exit_mm(struct task_struct *);
1018 extern void exit_files(struct task_struct *);
1019 extern void exit_signal(struct task_struct *);
1020 extern void __exit_signal(struct task_struct *);
1021 extern void exit_sighand(struct task_struct *);
1022 extern void __exit_sighand(struct task_struct *);
1023 extern void exit_itimers(struct signal_struct *);
1025 extern NORET_TYPE void do_group_exit(int);
1027 extern void reparent_to_init(void);
1028 extern void daemonize(const char *, ...);
1029 extern int allow_signal(int);
1030 extern int disallow_signal(int);
1031 extern task_t *child_reaper;
1033 extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1034 extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1035 extern struct task_struct * copy_process(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1038 extern void wait_task_inactive(task_t * p);
1040 #define wait_task_inactive(p) do { } while (0)
1043 #define remove_parent(p) list_del_init(&(p)->sibling)
1044 #define add_parent(p, parent) list_add_tail(&(p)->sibling,&(parent)->children)
1046 #define REMOVE_LINKS(p) do { \
1047 if (thread_group_leader(p)) \
1048 list_del_init(&(p)->tasks); \
1052 #define SET_LINKS(p) do { \
1053 if (thread_group_leader(p)) \
1054 list_add_tail(&(p)->tasks,&init_task.tasks); \
1055 add_parent(p, (p)->parent); \
1058 #define next_task(p) list_entry((p)->tasks.next, struct task_struct, tasks)
1059 #define prev_task(p) list_entry((p)->tasks.prev, struct task_struct, tasks)
1061 #define for_each_process(p) \
1062 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1065 * Careful: do_each_thread/while_each_thread is a double loop so
1066 * 'break' will not work as expected - use goto instead.
1068 #define do_each_thread(g, t) \
1069 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1071 #define while_each_thread(g, t) \
1072 while ((t = next_thread(t)) != g)
1074 extern task_t * FASTCALL(next_thread(const task_t *p));
1076 #define thread_group_leader(p) (p->pid == p->tgid)
1078 static inline int thread_group_empty(task_t *p)
1080 struct pid *pid = p->pids[PIDTYPE_TGID].pidptr;
1082 return pid->task_list.next->next == &pid->task_list;
1085 #define delay_group_leader(p) \
1086 (thread_group_leader(p) && !thread_group_empty(p))
1088 extern void unhash_process(struct task_struct *p);
1091 * Protects ->fs, ->files, ->mm, ->ptrace, ->group_info and synchronises with
1094 * Nests both inside and outside of read_lock(&tasklist_lock).
1095 * It must not be nested with write_lock_irq(&tasklist_lock),
1096 * neither inside nor outside.
1098 static inline void task_lock(struct task_struct *p)
1100 spin_lock(&p->alloc_lock);
1103 static inline void task_unlock(struct task_struct *p)
1105 spin_unlock(&p->alloc_lock);
1109 * get_task_mm - acquire a reference to the task's mm
1111 * Returns %NULL if the task has no mm. User must release
1112 * the mm via mmput() after use.
1114 static inline struct mm_struct * get_task_mm(struct task_struct * task)
1116 struct mm_struct * mm;
1127 /* set thread flags in other task's structures
1128 * - see asm/thread_info.h for TIF_xxxx flags available
1130 static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1132 set_ti_thread_flag(tsk->thread_info,flag);
1135 static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1137 clear_ti_thread_flag(tsk->thread_info,flag);
1140 static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1142 return test_and_set_ti_thread_flag(tsk->thread_info,flag);
1145 static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1147 return test_and_clear_ti_thread_flag(tsk->thread_info,flag);
1150 static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1152 return test_ti_thread_flag(tsk->thread_info,flag);
1155 static inline void set_tsk_need_resched(struct task_struct *tsk)
1157 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1160 static inline void clear_tsk_need_resched(struct task_struct *tsk)
1162 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1165 static inline int signal_pending(struct task_struct *p)
1167 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1170 static inline int need_resched(void)
1172 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1175 extern void __cond_resched(void);
1177 static inline void cond_resched(void)
1179 #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
1180 __might_sleep(__FILE__, __LINE__, 0);
1186 * cond_resched_lock() - if a reschedule is pending, drop the given lock,
1187 * call schedule, and on return reacquire the lock.
1189 * This works OK both with and without CONFIG_PREEMPT. We do strange low-level
1190 * operations here to prevent schedule() from being called twice (once via
1191 * spin_unlock(), once by hand).
1193 extern void __cond_resched_lock(spinlock_t * lock);
1195 static inline void cond_resched_lock(spinlock_t * lock)
1197 #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
1198 __might_sleep(__FILE__, __LINE__, 1);
1200 __cond_resched_lock(lock);
1203 /* Reevaluate whether the task has signals pending delivery.
1204 This is required every time the blocked sigset_t changes.
1205 callers must hold sighand->siglock. */
1207 extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1208 extern void recalc_sigpending(void);
1210 extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1213 * Wrappers for p->thread_info->cpu access. No-op on UP.
1217 static inline unsigned int task_cpu(const struct task_struct *p)
1219 return p->thread_info->cpu;
1222 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1224 p->thread_info->cpu = cpu;
1229 static inline unsigned int task_cpu(const struct task_struct *p)
1234 static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1238 #endif /* CONFIG_SMP */
1241 /* API for registering delay info */
1242 #ifdef CONFIG_DELAY_ACCT
1244 #define test_delay_flag(tsk,flg) ((tsk)->flags & (flg))
1245 #define set_delay_flag(tsk,flg) ((tsk)->flags |= (flg))
1246 #define clear_delay_flag(tsk,flg) ((tsk)->flags &= ~(flg))
1248 #define def_delay_var(var) unsigned long long var
1249 #define get_delay(tsk,field) ((tsk)->delays.field)
1250 #define delay_value(x) (((unsigned long)(x))/1000)
1252 #define start_delay(var) ((var) = sched_clock())
1253 #define start_delay_set(var,flg) (set_delay_flag(current,flg),(var) = sched_clock())
1255 #define inc_delay(tsk,field) (((tsk)->delays.field)++)
1256 #define add_delay_ts(tsk,field,start_ts,end_ts) ((tsk)->delays.field += delay_value((end_ts)-(start_ts)))
1257 #define add_delay_clear(tsk,field,start_ts,flg) (add_delay_ts(tsk,field,start_ts,sched_clock()),clear_delay_flag(tsk,flg))
1259 static inline void add_io_delay(unsigned long dstart)
1261 struct task_struct * tsk = current;
1262 unsigned long val = delay_value(sched_clock()-dstart);
1263 if (test_delay_flag(tsk,PF_MEMIO)) {
1264 tsk->delays.mem_iowait_total += val;
1265 tsk->delays.num_memwaits++;
1267 tsk->delays.iowait_total += val;
1268 tsk->delays.num_iowaits++;
1270 clear_delay_flag(tsk,PF_IOWAIT);
1273 inline static void init_delays(struct task_struct *tsk)
1275 memset((void*)&tsk->delays,0,sizeof(tsk->delays));
1280 #define test_delay_flag(tsk,flg) (0)
1281 #define set_delay_flag(tsk,flg) do { } while (0)
1282 #define clear_delay_flag(tsk,flg) do { } while (0)
1284 #define def_delay_var(var)
1285 #define get_delay(tsk,field) (0)
1287 #define start_delay(var) do { } while (0)
1288 #define start_delay_set(var,flg) do { } while (0)
1290 #define inc_delay(tsk,field) do { } while (0)
1291 #define add_delay_ts(tsk,field,start_ts,now) do { } while (0)
1292 #define add_delay_clear(tsk,field,start_ts,flg) do { } while (0)
1293 #define add_io_delay(dstart) do { } while (0)
1294 #define init_delays(tsk) do { } while (0)
1299 #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1300 extern void arch_pick_mmap_layout(struct mm_struct *mm);
1302 static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1304 mm->mmap_base = TASK_UNMAPPED_BASE;
1305 mm->get_unmapped_area = arch_get_unmapped_area;
1306 mm->unmap_area = arch_unmap_area;
1310 #endif /* __KERNEL__ */