sched.h 49 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590
  1. #ifndef _LINUX_SCHED_H
  2. #define _LINUX_SCHED_H
  3. #include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
  4. /*
  5. * cloning flags:
  6. */
  7. #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
  8. #define CLONE_VM 0x00000100 /* set if VM shared between processes */
  9. #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
  10. #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
  11. #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
  12. #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
  13. #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
  14. #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
  15. #define CLONE_THREAD 0x00010000 /* Same thread group? */
  16. #define CLONE_NEWNS 0x00020000 /* New namespace group? */
  17. #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
  18. #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
  19. #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
  20. #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
  21. #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
  22. #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
  23. #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
  24. #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
  25. /*
  26. * Scheduling policies
  27. */
  28. #define SCHED_NORMAL 0
  29. #define SCHED_FIFO 1
  30. #define SCHED_RR 2
  31. #define SCHED_BATCH 3
  32. #ifdef __KERNEL__
  33. struct sched_param {
  34. int sched_priority;
  35. };
  36. #include <asm/param.h> /* for HZ */
  37. #include <linux/capability.h>
  38. #include <linux/threads.h>
  39. #include <linux/kernel.h>
  40. #include <linux/types.h>
  41. #include <linux/timex.h>
  42. #include <linux/jiffies.h>
  43. #include <linux/rbtree.h>
  44. #include <linux/thread_info.h>
  45. #include <linux/cpumask.h>
  46. #include <linux/errno.h>
  47. #include <linux/nodemask.h>
  48. #include <asm/system.h>
  49. #include <asm/semaphore.h>
  50. #include <asm/page.h>
  51. #include <asm/ptrace.h>
  52. #include <asm/mmu.h>
  53. #include <asm/cputime.h>
  54. #include <linux/smp.h>
  55. #include <linux/sem.h>
  56. #include <linux/signal.h>
  57. #include <linux/securebits.h>
  58. #include <linux/fs_struct.h>
  59. #include <linux/compiler.h>
  60. #include <linux/completion.h>
  61. #include <linux/pid.h>
  62. #include <linux/percpu.h>
  63. #include <linux/topology.h>
  64. #include <linux/seccomp.h>
  65. #include <linux/rcupdate.h>
  66. #include <linux/futex.h>
  67. #include <linux/rtmutex.h>
  68. #include <linux/time.h>
  69. #include <linux/param.h>
  70. #include <linux/resource.h>
  71. #include <linux/timer.h>
  72. #include <linux/hrtimer.h>
  73. #include <asm/processor.h>
  74. struct exec_domain;
  75. struct futex_pi_state;
  76. /*
  77. * List of flags we want to share for kernel threads,
  78. * if only because they are not used by them anyway.
  79. */
  80. #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
  81. /*
  82. * These are the constant used to fake the fixed-point load-average
  83. * counting. Some notes:
  84. * - 11 bit fractions expand to 22 bits by the multiplies: this gives
  85. * a load-average precision of 10 bits integer + 11 bits fractional
  86. * - if you want to count load-averages more often, you need more
  87. * precision, or rounding will get you. With 2-second counting freq,
  88. * the EXP_n values would be 1981, 2034 and 2043 if still using only
  89. * 11 bit fractions.
  90. */
  91. extern unsigned long avenrun[]; /* Load averages */
  92. #define FSHIFT 11 /* nr of bits of precision */
  93. #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
  94. #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
  95. #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
  96. #define EXP_5 2014 /* 1/exp(5sec/5min) */
  97. #define EXP_15 2037 /* 1/exp(5sec/15min) */
  98. #define CALC_LOAD(load,exp,n) \
  99. load *= exp; \
  100. load += n*(FIXED_1-exp); \
  101. load >>= FSHIFT;
  102. extern unsigned long total_forks;
  103. extern int nr_threads;
  104. extern int last_pid;
  105. DECLARE_PER_CPU(unsigned long, process_counts);
  106. extern int nr_processes(void);
  107. extern unsigned long nr_running(void);
  108. extern unsigned long nr_uninterruptible(void);
  109. extern unsigned long nr_active(void);
  110. extern unsigned long nr_iowait(void);
  111. extern unsigned long weighted_cpuload(const int cpu);
  112. /*
  113. * Task state bitmask. NOTE! These bits are also
  114. * encoded in fs/proc/array.c: get_task_state().
  115. *
  116. * We have two separate sets of flags: task->state
  117. * is about runnability, while task->exit_state are
  118. * about the task exiting. Confusing, but this way
  119. * modifying one set can't modify the other one by
  120. * mistake.
  121. */
  122. #define TASK_RUNNING 0
  123. #define TASK_INTERRUPTIBLE 1
  124. #define TASK_UNINTERRUPTIBLE 2
  125. #define TASK_STOPPED 4
  126. #define TASK_TRACED 8
  127. /* in tsk->exit_state */
  128. #define EXIT_ZOMBIE 16
  129. #define EXIT_DEAD 32
  130. /* in tsk->state again */
  131. #define TASK_NONINTERACTIVE 64
  132. #define __set_task_state(tsk, state_value) \
  133. do { (tsk)->state = (state_value); } while (0)
  134. #define set_task_state(tsk, state_value) \
  135. set_mb((tsk)->state, (state_value))
  136. /*
  137. * set_current_state() includes a barrier so that the write of current->state
  138. * is correctly serialised wrt the caller's subsequent test of whether to
  139. * actually sleep:
  140. *
  141. * set_current_state(TASK_UNINTERRUPTIBLE);
  142. * if (do_i_need_to_sleep())
  143. * schedule();
  144. *
  145. * If the caller does not need such serialisation then use __set_current_state()
  146. */
  147. #define __set_current_state(state_value) \
  148. do { current->state = (state_value); } while (0)
  149. #define set_current_state(state_value) \
  150. set_mb(current->state, (state_value))
  151. /* Task command name length */
  152. #define TASK_COMM_LEN 16
  153. #include <linux/spinlock.h>
  154. /*
  155. * This serializes "schedule()" and also protects
  156. * the run-queue from deletions/modifications (but
  157. * _adding_ to the beginning of the run-queue has
  158. * a separate lock).
  159. */
  160. extern rwlock_t tasklist_lock;
  161. extern spinlock_t mmlist_lock;
  162. struct task_struct;
  163. extern void sched_init(void);
  164. extern void sched_init_smp(void);
  165. extern void init_idle(struct task_struct *idle, int cpu);
  166. extern cpumask_t nohz_cpu_mask;
  167. extern void show_state(void);
  168. extern void show_regs(struct pt_regs *);
  169. /*
  170. * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
  171. * task), SP is the stack pointer of the first frame that should be shown in the back
  172. * trace (or NULL if the entire call-chain of the task should be shown).
  173. */
  174. extern void show_stack(struct task_struct *task, unsigned long *sp);
  175. void io_schedule(void);
  176. 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(void);
  181. #ifdef CONFIG_DETECT_SOFTLOCKUP
  182. extern void softlockup_tick(void);
  183. extern void spawn_softlockup_task(void);
  184. extern void touch_softlockup_watchdog(void);
  185. #else
  186. static inline void softlockup_tick(void)
  187. {
  188. }
  189. static inline void spawn_softlockup_task(void)
  190. {
  191. }
  192. static inline void touch_softlockup_watchdog(void)
  193. {
  194. }
  195. #endif
  196. /* Attach to any functions which should be ignored in wchan output. */
  197. #define __sched __attribute__((__section__(".sched.text")))
  198. /* Is this address in the __sched functions? */
  199. extern int in_sched_functions(unsigned long addr);
  200. #define MAX_SCHEDULE_TIMEOUT LONG_MAX
  201. extern signed long FASTCALL(schedule_timeout(signed long timeout));
  202. extern signed long schedule_timeout_interruptible(signed long timeout);
  203. extern signed long schedule_timeout_uninterruptible(signed long timeout);
  204. asmlinkage void schedule(void);
  205. struct namespace;
  206. /* Maximum number of active map areas.. This is a random (large) number */
  207. #define DEFAULT_MAX_MAP_COUNT 65536
  208. extern int sysctl_max_map_count;
  209. #include <linux/aio.h>
  210. extern unsigned long
  211. arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
  212. unsigned long, unsigned long);
  213. extern unsigned long
  214. arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
  215. unsigned long len, unsigned long pgoff,
  216. unsigned long flags);
  217. extern void arch_unmap_area(struct mm_struct *, unsigned long);
  218. extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
  219. #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
  220. /*
  221. * The mm counters are not protected by its page_table_lock,
  222. * so must be incremented atomically.
  223. */
  224. #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
  225. #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
  226. #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
  227. #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
  228. #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
  229. typedef atomic_long_t mm_counter_t;
  230. #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  231. /*
  232. * The mm counters are protected by its page_table_lock,
  233. * so can be incremented directly.
  234. */
  235. #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
  236. #define get_mm_counter(mm, member) ((mm)->_##member)
  237. #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
  238. #define inc_mm_counter(mm, member) (mm)->_##member++
  239. #define dec_mm_counter(mm, member) (mm)->_##member--
  240. typedef unsigned long mm_counter_t;
  241. #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  242. #define get_mm_rss(mm) \
  243. (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
  244. #define update_hiwater_rss(mm) do { \
  245. unsigned long _rss = get_mm_rss(mm); \
  246. if ((mm)->hiwater_rss < _rss) \
  247. (mm)->hiwater_rss = _rss; \
  248. } while (0)
  249. #define update_hiwater_vm(mm) do { \
  250. if ((mm)->hiwater_vm < (mm)->total_vm) \
  251. (mm)->hiwater_vm = (mm)->total_vm; \
  252. } while (0)
  253. struct mm_struct {
  254. struct vm_area_struct * mmap; /* list of VMAs */
  255. struct rb_root mm_rb;
  256. struct vm_area_struct * mmap_cache; /* last find_vma result */
  257. unsigned long (*get_unmapped_area) (struct file *filp,
  258. unsigned long addr, unsigned long len,
  259. unsigned long pgoff, unsigned long flags);
  260. void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
  261. unsigned long mmap_base; /* base of mmap area */
  262. unsigned long task_size; /* size of task vm space */
  263. unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
  264. unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
  265. pgd_t * pgd;
  266. atomic_t mm_users; /* How many users with user space? */
  267. atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
  268. int map_count; /* number of VMAs */
  269. struct rw_semaphore mmap_sem;
  270. spinlock_t page_table_lock; /* Protects page tables and some counters */
  271. struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
  272. * together off init_mm.mmlist, and are protected
  273. * by mmlist_lock
  274. */
  275. /* Special counters, in some configurations protected by the
  276. * page_table_lock, in other configurations by being atomic.
  277. */
  278. mm_counter_t _file_rss;
  279. mm_counter_t _anon_rss;
  280. unsigned long hiwater_rss; /* High-watermark of RSS usage */
  281. unsigned long hiwater_vm; /* High-water virtual memory usage */
  282. unsigned long total_vm, locked_vm, shared_vm, exec_vm;
  283. unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
  284. unsigned long start_code, end_code, start_data, end_data;
  285. unsigned long start_brk, brk, start_stack;
  286. unsigned long arg_start, arg_end, env_start, env_end;
  287. unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
  288. unsigned dumpable:2;
  289. cpumask_t cpu_vm_mask;
  290. /* Architecture-specific MM context */
  291. mm_context_t context;
  292. /* Token based thrashing protection. */
  293. unsigned long swap_token_time;
  294. char recent_pagein;
  295. /* coredumping support */
  296. int core_waiters;
  297. struct completion *core_startup_done, core_done;
  298. /* aio bits */
  299. rwlock_t ioctx_list_lock;
  300. struct kioctx *ioctx_list;
  301. };
  302. struct sighand_struct {
  303. atomic_t count;
  304. struct k_sigaction action[_NSIG];
  305. spinlock_t siglock;
  306. };
  307. struct pacct_struct {
  308. int ac_flag;
  309. long ac_exitcode;
  310. unsigned long ac_mem;
  311. cputime_t ac_utime, ac_stime;
  312. unsigned long ac_minflt, ac_majflt;
  313. };
  314. /*
  315. * NOTE! "signal_struct" does not have it's own
  316. * locking, because a shared signal_struct always
  317. * implies a shared sighand_struct, so locking
  318. * sighand_struct is always a proper superset of
  319. * the locking of signal_struct.
  320. */
  321. struct signal_struct {
  322. atomic_t count;
  323. atomic_t live;
  324. wait_queue_head_t wait_chldexit; /* for wait4() */
  325. /* current thread group signal load-balancing target: */
  326. struct task_struct *curr_target;
  327. /* shared signal handling: */
  328. struct sigpending shared_pending;
  329. /* thread group exit support */
  330. int group_exit_code;
  331. /* overloaded:
  332. * - notify group_exit_task when ->count is equal to notify_count
  333. * - everyone except group_exit_task is stopped during signal delivery
  334. * of fatal signals, group_exit_task processes the signal.
  335. */
  336. struct task_struct *group_exit_task;
  337. int notify_count;
  338. /* thread group stop support, overloads group_exit_code too */
  339. int group_stop_count;
  340. unsigned int flags; /* see SIGNAL_* flags below */
  341. /* POSIX.1b Interval Timers */
  342. struct list_head posix_timers;
  343. /* ITIMER_REAL timer for the process */
  344. struct hrtimer real_timer;
  345. struct task_struct *tsk;
  346. ktime_t it_real_incr;
  347. /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
  348. cputime_t it_prof_expires, it_virt_expires;
  349. cputime_t it_prof_incr, it_virt_incr;
  350. /* job control IDs */
  351. pid_t pgrp;
  352. pid_t tty_old_pgrp;
  353. pid_t session;
  354. /* boolean value for session group leader */
  355. int leader;
  356. struct tty_struct *tty; /* NULL if no tty */
  357. /*
  358. * Cumulative resource counters for dead threads in the group,
  359. * and for reaped dead child processes forked by this group.
  360. * Live threads maintain their own counters and add to these
  361. * in __exit_signal, except for the group leader.
  362. */
  363. cputime_t utime, stime, cutime, cstime;
  364. unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
  365. unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
  366. /*
  367. * Cumulative ns of scheduled CPU time for dead threads in the
  368. * group, not including a zombie group leader. (This only differs
  369. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  370. * other than jiffies.)
  371. */
  372. unsigned long long sched_time;
  373. /*
  374. * We don't bother to synchronize most readers of this at all,
  375. * because there is no reader checking a limit that actually needs
  376. * to get both rlim_cur and rlim_max atomically, and either one
  377. * alone is a single word that can safely be read normally.
  378. * getrlimit/setrlimit use task_lock(current->group_leader) to
  379. * protect this instead of the siglock, because they really
  380. * have no need to disable irqs.
  381. */
  382. struct rlimit rlim[RLIM_NLIMITS];
  383. struct list_head cpu_timers[3];
  384. /* keep the process-shared keyrings here so that they do the right
  385. * thing in threads created with CLONE_THREAD */
  386. #ifdef CONFIG_KEYS
  387. struct key *session_keyring; /* keyring inherited over fork */
  388. struct key *process_keyring; /* keyring private to this process */
  389. #endif
  390. #ifdef CONFIG_BSD_PROCESS_ACCT
  391. struct pacct_struct pacct; /* per-process accounting information */
  392. #endif
  393. };
  394. /* Context switch must be unlocked if interrupts are to be enabled */
  395. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  396. # define __ARCH_WANT_UNLOCKED_CTXSW
  397. #endif
  398. /*
  399. * Bits in flags field of signal_struct.
  400. */
  401. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  402. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  403. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  404. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  405. /*
  406. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  407. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  408. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  409. * values are inverted: lower p->prio value means higher priority.
  410. *
  411. * The MAX_USER_RT_PRIO value allows the actual maximum
  412. * RT priority to be separate from the value exported to
  413. * user-space. This allows kernel threads to set their
  414. * priority to a value higher than any user task. Note:
  415. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  416. */
  417. #define MAX_USER_RT_PRIO 100
  418. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  419. #define MAX_PRIO (MAX_RT_PRIO + 40)
  420. #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
  421. #define rt_task(p) rt_prio((p)->prio)
  422. #define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
  423. #define has_rt_policy(p) \
  424. unlikely((p)->policy != SCHED_NORMAL && (p)->policy != SCHED_BATCH)
  425. /*
  426. * Some day this will be a full-fledged user tracking system..
  427. */
  428. struct user_struct {
  429. atomic_t __count; /* reference count */
  430. atomic_t processes; /* How many processes does this user have? */
  431. atomic_t files; /* How many open files does this user have? */
  432. atomic_t sigpending; /* How many pending signals does this user have? */
  433. #ifdef CONFIG_INOTIFY_USER
  434. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  435. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  436. #endif
  437. /* protected by mq_lock */
  438. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  439. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  440. #ifdef CONFIG_KEYS
  441. struct key *uid_keyring; /* UID specific keyring */
  442. struct key *session_keyring; /* UID's default session keyring */
  443. #endif
  444. /* Hash table maintenance information */
  445. struct list_head uidhash_list;
  446. uid_t uid;
  447. };
  448. extern struct user_struct *find_user(uid_t);
  449. extern struct user_struct root_user;
  450. #define INIT_USER (&root_user)
  451. struct backing_dev_info;
  452. struct reclaim_state;
  453. #ifdef CONFIG_SCHEDSTATS
  454. struct sched_info {
  455. /* cumulative counters */
  456. unsigned long cpu_time, /* time spent on the cpu */
  457. run_delay, /* time spent waiting on a runqueue */
  458. pcnt; /* # of timeslices run on this cpu */
  459. /* timestamps */
  460. unsigned long last_arrival, /* when we last ran on a cpu */
  461. last_queued; /* when we were last queued to run */
  462. };
  463. extern struct file_operations proc_schedstat_operations;
  464. #endif
  465. #ifdef CONFIG_TASK_DELAY_ACCT
  466. struct task_delay_info {
  467. spinlock_t lock;
  468. unsigned int flags; /* Private per-task flags */
  469. /* For each stat XXX, add following, aligned appropriately
  470. *
  471. * struct timespec XXX_start, XXX_end;
  472. * u64 XXX_delay;
  473. * u32 XXX_count;
  474. *
  475. * Atomicity of updates to XXX_delay, XXX_count protected by
  476. * single lock above (split into XXX_lock if contention is an issue).
  477. */
  478. /*
  479. * XXX_count is incremented on every XXX operation, the delay
  480. * associated with the operation is added to XXX_delay.
  481. * XXX_delay contains the accumulated delay time in nanoseconds.
  482. */
  483. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  484. u64 blkio_delay; /* wait for sync block io completion */
  485. u64 swapin_delay; /* wait for swapin block io completion */
  486. u32 blkio_count; /* total count of the number of sync block */
  487. /* io operations performed */
  488. u32 swapin_count; /* total count of the number of swapin block */
  489. /* io operations performed */
  490. };
  491. #endif
  492. enum idle_type
  493. {
  494. SCHED_IDLE,
  495. NOT_IDLE,
  496. NEWLY_IDLE,
  497. MAX_IDLE_TYPES
  498. };
  499. /*
  500. * sched-domains (multiprocessor balancing) declarations:
  501. */
  502. #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
  503. #ifdef CONFIG_SMP
  504. #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
  505. #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
  506. #define SD_BALANCE_EXEC 4 /* Balance on exec */
  507. #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
  508. #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
  509. #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
  510. #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
  511. #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
  512. #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
  513. #define BALANCE_FOR_POWER ((sched_mc_power_savings || sched_smt_power_savings) \
  514. ? SD_POWERSAVINGS_BALANCE : 0)
  515. struct sched_group {
  516. struct sched_group *next; /* Must be a circular list */
  517. cpumask_t cpumask;
  518. /*
  519. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  520. * single CPU. This is read only (except for setup, hotplug CPU).
  521. */
  522. unsigned long cpu_power;
  523. };
  524. struct sched_domain {
  525. /* These fields must be setup */
  526. struct sched_domain *parent; /* top domain must be null terminated */
  527. struct sched_group *groups; /* the balancing groups of the domain */
  528. cpumask_t span; /* span of all CPUs in this domain */
  529. unsigned long min_interval; /* Minimum balance interval ms */
  530. unsigned long max_interval; /* Maximum balance interval ms */
  531. unsigned int busy_factor; /* less balancing by factor if busy */
  532. unsigned int imbalance_pct; /* No balance until over watermark */
  533. unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
  534. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  535. unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
  536. unsigned int busy_idx;
  537. unsigned int idle_idx;
  538. unsigned int newidle_idx;
  539. unsigned int wake_idx;
  540. unsigned int forkexec_idx;
  541. int flags; /* See SD_* */
  542. /* Runtime fields. */
  543. unsigned long last_balance; /* init to jiffies. units in jiffies */
  544. unsigned int balance_interval; /* initialise to 1. units in ms. */
  545. unsigned int nr_balance_failed; /* initialise to 0 */
  546. #ifdef CONFIG_SCHEDSTATS
  547. /* load_balance() stats */
  548. unsigned long lb_cnt[MAX_IDLE_TYPES];
  549. unsigned long lb_failed[MAX_IDLE_TYPES];
  550. unsigned long lb_balanced[MAX_IDLE_TYPES];
  551. unsigned long lb_imbalance[MAX_IDLE_TYPES];
  552. unsigned long lb_gained[MAX_IDLE_TYPES];
  553. unsigned long lb_hot_gained[MAX_IDLE_TYPES];
  554. unsigned long lb_nobusyg[MAX_IDLE_TYPES];
  555. unsigned long lb_nobusyq[MAX_IDLE_TYPES];
  556. /* Active load balancing */
  557. unsigned long alb_cnt;
  558. unsigned long alb_failed;
  559. unsigned long alb_pushed;
  560. /* SD_BALANCE_EXEC stats */
  561. unsigned long sbe_cnt;
  562. unsigned long sbe_balanced;
  563. unsigned long sbe_pushed;
  564. /* SD_BALANCE_FORK stats */
  565. unsigned long sbf_cnt;
  566. unsigned long sbf_balanced;
  567. unsigned long sbf_pushed;
  568. /* try_to_wake_up() stats */
  569. unsigned long ttwu_wake_remote;
  570. unsigned long ttwu_move_affine;
  571. unsigned long ttwu_move_balance;
  572. #endif
  573. };
  574. extern int partition_sched_domains(cpumask_t *partition1,
  575. cpumask_t *partition2);
  576. /*
  577. * Maximum cache size the migration-costs auto-tuning code will
  578. * search from:
  579. */
  580. extern unsigned int max_cache_size;
  581. #endif /* CONFIG_SMP */
  582. struct io_context; /* See blkdev.h */
  583. void exit_io_context(void);
  584. struct cpuset;
  585. #define NGROUPS_SMALL 32
  586. #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
  587. struct group_info {
  588. int ngroups;
  589. atomic_t usage;
  590. gid_t small_block[NGROUPS_SMALL];
  591. int nblocks;
  592. gid_t *blocks[0];
  593. };
  594. /*
  595. * get_group_info() must be called with the owning task locked (via task_lock())
  596. * when task != current. The reason being that the vast majority of callers are
  597. * looking at current->group_info, which can not be changed except by the
  598. * current task. Changing current->group_info requires the task lock, too.
  599. */
  600. #define get_group_info(group_info) do { \
  601. atomic_inc(&(group_info)->usage); \
  602. } while (0)
  603. #define put_group_info(group_info) do { \
  604. if (atomic_dec_and_test(&(group_info)->usage)) \
  605. groups_free(group_info); \
  606. } while (0)
  607. extern struct group_info *groups_alloc(int gidsetsize);
  608. extern void groups_free(struct group_info *group_info);
  609. extern int set_current_groups(struct group_info *group_info);
  610. extern int groups_search(struct group_info *group_info, gid_t grp);
  611. /* access the groups "array" with this macro */
  612. #define GROUP_AT(gi, i) \
  613. ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
  614. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  615. extern void prefetch_stack(struct task_struct *t);
  616. #else
  617. static inline void prefetch_stack(struct task_struct *t) { }
  618. #endif
  619. struct audit_context; /* See audit.c */
  620. struct mempolicy;
  621. struct pipe_inode_info;
  622. enum sleep_type {
  623. SLEEP_NORMAL,
  624. SLEEP_NONINTERACTIVE,
  625. SLEEP_INTERACTIVE,
  626. SLEEP_INTERRUPTED,
  627. };
  628. struct prio_array;
  629. struct task_struct {
  630. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  631. struct thread_info *thread_info;
  632. atomic_t usage;
  633. unsigned long flags; /* per process flags, defined below */
  634. unsigned long ptrace;
  635. int lock_depth; /* BKL lock depth */
  636. #ifdef CONFIG_SMP
  637. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  638. int oncpu;
  639. #endif
  640. #endif
  641. int load_weight; /* for niceness load balancing purposes */
  642. int prio, static_prio, normal_prio;
  643. struct list_head run_list;
  644. struct prio_array *array;
  645. unsigned short ioprio;
  646. unsigned int btrace_seq;
  647. unsigned long sleep_avg;
  648. unsigned long long timestamp, last_ran;
  649. unsigned long long sched_time; /* sched_clock time spent running */
  650. enum sleep_type sleep_type;
  651. unsigned long policy;
  652. cpumask_t cpus_allowed;
  653. unsigned int time_slice, first_time_slice;
  654. #ifdef CONFIG_SCHEDSTATS
  655. struct sched_info sched_info;
  656. #endif
  657. struct list_head tasks;
  658. /*
  659. * ptrace_list/ptrace_children forms the list of my children
  660. * that were stolen by a ptracer.
  661. */
  662. struct list_head ptrace_children;
  663. struct list_head ptrace_list;
  664. struct mm_struct *mm, *active_mm;
  665. /* task state */
  666. struct linux_binfmt *binfmt;
  667. long exit_state;
  668. int exit_code, exit_signal;
  669. int pdeath_signal; /* The signal sent when the parent dies */
  670. /* ??? */
  671. unsigned long personality;
  672. unsigned did_exec:1;
  673. pid_t pid;
  674. pid_t tgid;
  675. /*
  676. * pointers to (original) parent process, youngest child, younger sibling,
  677. * older sibling, respectively. (p->father can be replaced with
  678. * p->parent->pid)
  679. */
  680. struct task_struct *real_parent; /* real parent process (when being debugged) */
  681. struct task_struct *parent; /* parent process */
  682. /*
  683. * children/sibling forms the list of my children plus the
  684. * tasks I'm ptracing.
  685. */
  686. struct list_head children; /* list of my children */
  687. struct list_head sibling; /* linkage in my parent's children list */
  688. struct task_struct *group_leader; /* threadgroup leader */
  689. /* PID/PID hash table linkage. */
  690. struct pid_link pids[PIDTYPE_MAX];
  691. struct list_head thread_group;
  692. struct completion *vfork_done; /* for vfork() */
  693. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  694. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  695. unsigned long rt_priority;
  696. cputime_t utime, stime;
  697. unsigned long nvcsw, nivcsw; /* context switch counts */
  698. struct timespec start_time;
  699. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  700. unsigned long min_flt, maj_flt;
  701. cputime_t it_prof_expires, it_virt_expires;
  702. unsigned long long it_sched_expires;
  703. struct list_head cpu_timers[3];
  704. /* process credentials */
  705. uid_t uid,euid,suid,fsuid;
  706. gid_t gid,egid,sgid,fsgid;
  707. struct group_info *group_info;
  708. kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
  709. unsigned keep_capabilities:1;
  710. struct user_struct *user;
  711. #ifdef CONFIG_KEYS
  712. struct key *request_key_auth; /* assumed request_key authority */
  713. struct key *thread_keyring; /* keyring private to this thread */
  714. unsigned char jit_keyring; /* default keyring to attach requested keys to */
  715. #endif
  716. int oomkilladj; /* OOM kill score adjustment (bit shift). */
  717. char comm[TASK_COMM_LEN]; /* executable name excluding path
  718. - access with [gs]et_task_comm (which lock
  719. it with task_lock())
  720. - initialized normally by flush_old_exec */
  721. /* file system info */
  722. int link_count, total_link_count;
  723. /* ipc stuff */
  724. struct sysv_sem sysvsem;
  725. /* CPU-specific state of this task */
  726. struct thread_struct thread;
  727. /* filesystem information */
  728. struct fs_struct *fs;
  729. /* open file information */
  730. struct files_struct *files;
  731. /* namespace */
  732. struct namespace *namespace;
  733. /* signal handlers */
  734. struct signal_struct *signal;
  735. struct sighand_struct *sighand;
  736. sigset_t blocked, real_blocked;
  737. sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
  738. struct sigpending pending;
  739. unsigned long sas_ss_sp;
  740. size_t sas_ss_size;
  741. int (*notifier)(void *priv);
  742. void *notifier_data;
  743. sigset_t *notifier_mask;
  744. void *security;
  745. struct audit_context *audit_context;
  746. seccomp_t seccomp;
  747. /* Thread group tracking */
  748. u32 parent_exec_id;
  749. u32 self_exec_id;
  750. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
  751. spinlock_t alloc_lock;
  752. /* Protection of the PI data structures: */
  753. spinlock_t pi_lock;
  754. #ifdef CONFIG_RT_MUTEXES
  755. /* PI waiters blocked on a rt_mutex held by this task */
  756. struct plist_head pi_waiters;
  757. /* Deadlock detection and priority inheritance handling */
  758. struct rt_mutex_waiter *pi_blocked_on;
  759. #endif
  760. #ifdef CONFIG_DEBUG_MUTEXES
  761. /* mutex deadlock detection */
  762. struct mutex_waiter *blocked_on;
  763. #endif
  764. #ifdef CONFIG_TRACE_IRQFLAGS
  765. unsigned int irq_events;
  766. int hardirqs_enabled;
  767. unsigned long hardirq_enable_ip;
  768. unsigned int hardirq_enable_event;
  769. unsigned long hardirq_disable_ip;
  770. unsigned int hardirq_disable_event;
  771. int softirqs_enabled;
  772. unsigned long softirq_disable_ip;
  773. unsigned int softirq_disable_event;
  774. unsigned long softirq_enable_ip;
  775. unsigned int softirq_enable_event;
  776. int hardirq_context;
  777. int softirq_context;
  778. #endif
  779. #ifdef CONFIG_LOCKDEP
  780. # define MAX_LOCK_DEPTH 30UL
  781. u64 curr_chain_key;
  782. int lockdep_depth;
  783. struct held_lock held_locks[MAX_LOCK_DEPTH];
  784. unsigned int lockdep_recursion;
  785. #endif
  786. /* journalling filesystem info */
  787. void *journal_info;
  788. /* VM state */
  789. struct reclaim_state *reclaim_state;
  790. struct backing_dev_info *backing_dev_info;
  791. struct io_context *io_context;
  792. unsigned long ptrace_message;
  793. siginfo_t *last_siginfo; /* For ptrace use. */
  794. /*
  795. * current io wait handle: wait queue entry to use for io waits
  796. * If this thread is processing aio, this points at the waitqueue
  797. * inside the currently handled kiocb. It may be NULL (i.e. default
  798. * to a stack based synchronous wait) if its doing sync IO.
  799. */
  800. wait_queue_t *io_wait;
  801. /* i/o counters(bytes read/written, #syscalls */
  802. u64 rchar, wchar, syscr, syscw;
  803. #if defined(CONFIG_BSD_PROCESS_ACCT)
  804. u64 acct_rss_mem1; /* accumulated rss usage */
  805. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  806. clock_t acct_stimexpd; /* clock_t-converted stime since last update */
  807. #endif
  808. #ifdef CONFIG_NUMA
  809. struct mempolicy *mempolicy;
  810. short il_next;
  811. #endif
  812. #ifdef CONFIG_CPUSETS
  813. struct cpuset *cpuset;
  814. nodemask_t mems_allowed;
  815. int cpuset_mems_generation;
  816. int cpuset_mem_spread_rotor;
  817. #endif
  818. struct robust_list_head __user *robust_list;
  819. #ifdef CONFIG_COMPAT
  820. struct compat_robust_list_head __user *compat_robust_list;
  821. #endif
  822. struct list_head pi_state_list;
  823. struct futex_pi_state *pi_state_cache;
  824. atomic_t fs_excl; /* holding fs exclusive resources */
  825. struct rcu_head rcu;
  826. /*
  827. * cache last used pipe for splice
  828. */
  829. struct pipe_inode_info *splice_pipe;
  830. #ifdef CONFIG_TASK_DELAY_ACCT
  831. struct task_delay_info *delays;
  832. #endif
  833. };
  834. static inline pid_t process_group(struct task_struct *tsk)
  835. {
  836. return tsk->signal->pgrp;
  837. }
  838. /**
  839. * pid_alive - check that a task structure is not stale
  840. * @p: Task structure to be checked.
  841. *
  842. * Test if a process is not yet dead (at most zombie state)
  843. * If pid_alive fails, then pointers within the task structure
  844. * can be stale and must not be dereferenced.
  845. */
  846. static inline int pid_alive(struct task_struct *p)
  847. {
  848. return p->pids[PIDTYPE_PID].pid != NULL;
  849. }
  850. extern void free_task(struct task_struct *tsk);
  851. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  852. extern void __put_task_struct(struct task_struct *t);
  853. static inline void put_task_struct(struct task_struct *t)
  854. {
  855. if (atomic_dec_and_test(&t->usage))
  856. __put_task_struct(t);
  857. }
  858. /*
  859. * Per process flags
  860. */
  861. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  862. /* Not implemented yet, only for 486*/
  863. #define PF_STARTING 0x00000002 /* being created */
  864. #define PF_EXITING 0x00000004 /* getting shut down */
  865. #define PF_DEAD 0x00000008 /* Dead */
  866. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  867. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  868. #define PF_DUMPCORE 0x00000200 /* dumped core */
  869. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  870. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  871. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  872. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  873. #define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
  874. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  875. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  876. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  877. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  878. #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
  879. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  880. #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
  881. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  882. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  883. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  884. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  885. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  886. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  887. /*
  888. * Only the _current_ task can read/write to tsk->flags, but other
  889. * tasks can access tsk->flags in readonly mode for example
  890. * with tsk_used_math (like during threaded core dumping).
  891. * There is however an exception to this rule during ptrace
  892. * or during fork: the ptracer task is allowed to write to the
  893. * child->flags of its traced child (same goes for fork, the parent
  894. * can write to the child->flags), because we're guaranteed the
  895. * child is not running and in turn not changing child->flags
  896. * at the same time the parent does it.
  897. */
  898. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  899. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  900. #define clear_used_math() clear_stopped_child_used_math(current)
  901. #define set_used_math() set_stopped_child_used_math(current)
  902. #define conditional_stopped_child_used_math(condition, child) \
  903. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  904. #define conditional_used_math(condition) \
  905. conditional_stopped_child_used_math(condition, current)
  906. #define copy_to_stopped_child_used_math(child) \
  907. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  908. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  909. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  910. #define used_math() tsk_used_math(current)
  911. #ifdef CONFIG_SMP
  912. extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
  913. #else
  914. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  915. {
  916. if (!cpu_isset(0, new_mask))
  917. return -EINVAL;
  918. return 0;
  919. }
  920. #endif
  921. extern unsigned long long sched_clock(void);
  922. extern unsigned long long
  923. current_sched_time(const struct task_struct *current_task);
  924. /* sched_exec is called by processes performing an exec */
  925. #ifdef CONFIG_SMP
  926. extern void sched_exec(void);
  927. #else
  928. #define sched_exec() {}
  929. #endif
  930. #ifdef CONFIG_HOTPLUG_CPU
  931. extern void idle_task_exit(void);
  932. #else
  933. static inline void idle_task_exit(void) {}
  934. #endif
  935. extern void sched_idle_next(void);
  936. #ifdef CONFIG_RT_MUTEXES
  937. extern int rt_mutex_getprio(struct task_struct *p);
  938. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  939. extern void rt_mutex_adjust_pi(struct task_struct *p);
  940. #else
  941. static inline int rt_mutex_getprio(struct task_struct *p)
  942. {
  943. return p->normal_prio;
  944. }
  945. # define rt_mutex_adjust_pi(p) do { } while (0)
  946. #endif
  947. extern void set_user_nice(struct task_struct *p, long nice);
  948. extern int task_prio(const struct task_struct *p);
  949. extern int task_nice(const struct task_struct *p);
  950. extern int can_nice(const struct task_struct *p, const int nice);
  951. extern int task_curr(const struct task_struct *p);
  952. extern int idle_cpu(int cpu);
  953. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  954. extern struct task_struct *idle_task(int cpu);
  955. extern struct task_struct *curr_task(int cpu);
  956. extern void set_curr_task(int cpu, struct task_struct *p);
  957. void yield(void);
  958. /*
  959. * The default (Linux) execution domain.
  960. */
  961. extern struct exec_domain default_exec_domain;
  962. union thread_union {
  963. struct thread_info thread_info;
  964. unsigned long stack[THREAD_SIZE/sizeof(long)];
  965. };
  966. #ifndef __HAVE_ARCH_KSTACK_END
  967. static inline int kstack_end(void *addr)
  968. {
  969. /* Reliable end of stack detection:
  970. * Some APM bios versions misalign the stack
  971. */
  972. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  973. }
  974. #endif
  975. extern union thread_union init_thread_union;
  976. extern struct task_struct init_task;
  977. extern struct mm_struct init_mm;
  978. #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
  979. extern struct task_struct *find_task_by_pid_type(int type, int pid);
  980. extern void set_special_pids(pid_t session, pid_t pgrp);
  981. extern void __set_special_pids(pid_t session, pid_t pgrp);
  982. /* per-UID process charging. */
  983. extern struct user_struct * alloc_uid(uid_t);
  984. static inline struct user_struct *get_uid(struct user_struct *u)
  985. {
  986. atomic_inc(&u->__count);
  987. return u;
  988. }
  989. extern void free_uid(struct user_struct *);
  990. extern void switch_uid(struct user_struct *);
  991. #include <asm/current.h>
  992. extern void do_timer(struct pt_regs *);
  993. extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
  994. extern int FASTCALL(wake_up_process(struct task_struct * tsk));
  995. extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
  996. unsigned long clone_flags));
  997. #ifdef CONFIG_SMP
  998. extern void kick_process(struct task_struct *tsk);
  999. #else
  1000. static inline void kick_process(struct task_struct *tsk) { }
  1001. #endif
  1002. extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
  1003. extern void FASTCALL(sched_exit(struct task_struct * p));
  1004. extern int in_group_p(gid_t);
  1005. extern int in_egroup_p(gid_t);
  1006. extern void proc_caches_init(void);
  1007. extern void flush_signals(struct task_struct *);
  1008. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1009. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1010. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1011. {
  1012. unsigned long flags;
  1013. int ret;
  1014. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1015. ret = dequeue_signal(tsk, mask, info);
  1016. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1017. return ret;
  1018. }
  1019. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1020. sigset_t *mask);
  1021. extern void unblock_all_signals(void);
  1022. extern void release_task(struct task_struct * p);
  1023. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1024. extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
  1025. extern int force_sigsegv(int, struct task_struct *);
  1026. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1027. extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
  1028. extern int kill_pg_info(int, struct siginfo *, pid_t);
  1029. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1030. extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t, u32);
  1031. extern void do_notify_parent(struct task_struct *, int);
  1032. extern void force_sig(int, struct task_struct *);
  1033. extern void force_sig_specific(int, struct task_struct *);
  1034. extern int send_sig(int, struct task_struct *, int);
  1035. extern void zap_other_threads(struct task_struct *p);
  1036. extern int kill_pg(pid_t, int, int);
  1037. extern int kill_proc(pid_t, int, int);
  1038. extern struct sigqueue *sigqueue_alloc(void);
  1039. extern void sigqueue_free(struct sigqueue *);
  1040. extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
  1041. extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
  1042. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1043. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1044. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1045. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1046. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1047. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1048. static inline int is_si_special(const struct siginfo *info)
  1049. {
  1050. return info <= SEND_SIG_FORCED;
  1051. }
  1052. /* True if we are on the alternate signal stack. */
  1053. static inline int on_sig_stack(unsigned long sp)
  1054. {
  1055. return (sp - current->sas_ss_sp < current->sas_ss_size);
  1056. }
  1057. static inline int sas_ss_flags(unsigned long sp)
  1058. {
  1059. return (current->sas_ss_size == 0 ? SS_DISABLE
  1060. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1061. }
  1062. /*
  1063. * Routines for handling mm_structs
  1064. */
  1065. extern struct mm_struct * mm_alloc(void);
  1066. /* mmdrop drops the mm and the page tables */
  1067. extern void FASTCALL(__mmdrop(struct mm_struct *));
  1068. static inline void mmdrop(struct mm_struct * mm)
  1069. {
  1070. if (atomic_dec_and_test(&mm->mm_count))
  1071. __mmdrop(mm);
  1072. }
  1073. /* mmput gets rid of the mappings and all user-space */
  1074. extern void mmput(struct mm_struct *);
  1075. /* Grab a reference to a task's mm, if it is not already going away */
  1076. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1077. /* Remove the current tasks stale references to the old mm_struct */
  1078. extern void mm_release(struct task_struct *, struct mm_struct *);
  1079. extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
  1080. extern void flush_thread(void);
  1081. extern void exit_thread(void);
  1082. extern void exit_files(struct task_struct *);
  1083. extern void __cleanup_signal(struct signal_struct *);
  1084. extern void __cleanup_sighand(struct sighand_struct *);
  1085. extern void exit_itimers(struct signal_struct *);
  1086. extern NORET_TYPE void do_group_exit(int);
  1087. extern void daemonize(const char *, ...);
  1088. extern int allow_signal(int);
  1089. extern int disallow_signal(int);
  1090. extern struct task_struct *child_reaper;
  1091. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1092. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1093. struct task_struct *fork_idle(int);
  1094. extern void set_task_comm(struct task_struct *tsk, char *from);
  1095. extern void get_task_comm(char *to, struct task_struct *tsk);
  1096. #ifdef CONFIG_SMP
  1097. extern void wait_task_inactive(struct task_struct * p);
  1098. #else
  1099. #define wait_task_inactive(p) do { } while (0)
  1100. #endif
  1101. #define remove_parent(p) list_del_init(&(p)->sibling)
  1102. #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
  1103. #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
  1104. #define for_each_process(p) \
  1105. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1106. /*
  1107. * Careful: do_each_thread/while_each_thread is a double loop so
  1108. * 'break' will not work as expected - use goto instead.
  1109. */
  1110. #define do_each_thread(g, t) \
  1111. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1112. #define while_each_thread(g, t) \
  1113. while ((t = next_thread(t)) != g)
  1114. /* de_thread depends on thread_group_leader not being a pid based check */
  1115. #define thread_group_leader(p) (p == p->group_leader)
  1116. static inline struct task_struct *next_thread(const struct task_struct *p)
  1117. {
  1118. return list_entry(rcu_dereference(p->thread_group.next),
  1119. struct task_struct, thread_group);
  1120. }
  1121. static inline int thread_group_empty(struct task_struct *p)
  1122. {
  1123. return list_empty(&p->thread_group);
  1124. }
  1125. #define delay_group_leader(p) \
  1126. (thread_group_leader(p) && !thread_group_empty(p))
  1127. /*
  1128. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1129. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1130. * pins the final release of task.io_context. Also protects ->cpuset.
  1131. *
  1132. * Nests both inside and outside of read_lock(&tasklist_lock).
  1133. * It must not be nested with write_lock_irq(&tasklist_lock),
  1134. * neither inside nor outside.
  1135. */
  1136. static inline void task_lock(struct task_struct *p)
  1137. {
  1138. spin_lock(&p->alloc_lock);
  1139. }
  1140. static inline void task_unlock(struct task_struct *p)
  1141. {
  1142. spin_unlock(&p->alloc_lock);
  1143. }
  1144. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1145. unsigned long *flags);
  1146. static inline void unlock_task_sighand(struct task_struct *tsk,
  1147. unsigned long *flags)
  1148. {
  1149. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1150. }
  1151. #ifndef __HAVE_THREAD_FUNCTIONS
  1152. #define task_thread_info(task) (task)->thread_info
  1153. #define task_stack_page(task) ((void*)((task)->thread_info))
  1154. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1155. {
  1156. *task_thread_info(p) = *task_thread_info(org);
  1157. task_thread_info(p)->task = p;
  1158. }
  1159. static inline unsigned long *end_of_stack(struct task_struct *p)
  1160. {
  1161. return (unsigned long *)(p->thread_info + 1);
  1162. }
  1163. #endif
  1164. /* set thread flags in other task's structures
  1165. * - see asm/thread_info.h for TIF_xxxx flags available
  1166. */
  1167. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1168. {
  1169. set_ti_thread_flag(task_thread_info(tsk), flag);
  1170. }
  1171. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1172. {
  1173. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1174. }
  1175. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1176. {
  1177. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1178. }
  1179. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1180. {
  1181. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1182. }
  1183. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1184. {
  1185. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1186. }
  1187. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1188. {
  1189. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1190. }
  1191. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1192. {
  1193. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1194. }
  1195. static inline int signal_pending(struct task_struct *p)
  1196. {
  1197. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1198. }
  1199. static inline int need_resched(void)
  1200. {
  1201. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  1202. }
  1203. /*
  1204. * cond_resched() and cond_resched_lock(): latency reduction via
  1205. * explicit rescheduling in places that are safe. The return
  1206. * value indicates whether a reschedule was done in fact.
  1207. * cond_resched_lock() will drop the spinlock before scheduling,
  1208. * cond_resched_softirq() will enable bhs before scheduling.
  1209. */
  1210. extern int cond_resched(void);
  1211. extern int cond_resched_lock(spinlock_t * lock);
  1212. extern int cond_resched_softirq(void);
  1213. /*
  1214. * Does a critical section need to be broken due to another
  1215. * task waiting?:
  1216. */
  1217. #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
  1218. # define need_lockbreak(lock) ((lock)->break_lock)
  1219. #else
  1220. # define need_lockbreak(lock) 0
  1221. #endif
  1222. /*
  1223. * Does a critical section need to be broken due to another
  1224. * task waiting or preemption being signalled:
  1225. */
  1226. static inline int lock_need_resched(spinlock_t *lock)
  1227. {
  1228. if (need_lockbreak(lock) || need_resched())
  1229. return 1;
  1230. return 0;
  1231. }
  1232. /* Reevaluate whether the task has signals pending delivery.
  1233. This is required every time the blocked sigset_t changes.
  1234. callers must hold sighand->siglock. */
  1235. extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
  1236. extern void recalc_sigpending(void);
  1237. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  1238. /*
  1239. * Wrappers for p->thread_info->cpu access. No-op on UP.
  1240. */
  1241. #ifdef CONFIG_SMP
  1242. static inline unsigned int task_cpu(const struct task_struct *p)
  1243. {
  1244. return task_thread_info(p)->cpu;
  1245. }
  1246. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1247. {
  1248. task_thread_info(p)->cpu = cpu;
  1249. }
  1250. #else
  1251. static inline unsigned int task_cpu(const struct task_struct *p)
  1252. {
  1253. return 0;
  1254. }
  1255. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1256. {
  1257. }
  1258. #endif /* CONFIG_SMP */
  1259. #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
  1260. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  1261. #else
  1262. static inline void arch_pick_mmap_layout(struct mm_struct *mm)
  1263. {
  1264. mm->mmap_base = TASK_UNMAPPED_BASE;
  1265. mm->get_unmapped_area = arch_get_unmapped_area;
  1266. mm->unmap_area = arch_unmap_area;
  1267. }
  1268. #endif
  1269. extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
  1270. extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
  1271. #include <linux/sysdev.h>
  1272. extern int sched_mc_power_savings, sched_smt_power_savings;
  1273. extern struct sysdev_attribute attr_sched_mc_power_savings, attr_sched_smt_power_savings;
  1274. extern int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls);
  1275. extern void normalize_rt_tasks(void);
  1276. #ifdef CONFIG_PM
  1277. /*
  1278. * Check if a process has been frozen
  1279. */
  1280. static inline int frozen(struct task_struct *p)
  1281. {
  1282. return p->flags & PF_FROZEN;
  1283. }
  1284. /*
  1285. * Check if there is a request to freeze a process
  1286. */
  1287. static inline int freezing(struct task_struct *p)
  1288. {
  1289. return p->flags & PF_FREEZE;
  1290. }
  1291. /*
  1292. * Request that a process be frozen
  1293. * FIXME: SMP problem. We may not modify other process' flags!
  1294. */
  1295. static inline void freeze(struct task_struct *p)
  1296. {
  1297. p->flags |= PF_FREEZE;
  1298. }
  1299. /*
  1300. * Wake up a frozen process
  1301. */
  1302. static inline int thaw_process(struct task_struct *p)
  1303. {
  1304. if (frozen(p)) {
  1305. p->flags &= ~PF_FROZEN;
  1306. wake_up_process(p);
  1307. return 1;
  1308. }
  1309. return 0;
  1310. }
  1311. /*
  1312. * freezing is complete, mark process as frozen
  1313. */
  1314. static inline void frozen_process(struct task_struct *p)
  1315. {
  1316. p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
  1317. }
  1318. extern void refrigerator(void);
  1319. extern int freeze_processes(void);
  1320. extern void thaw_processes(void);
  1321. static inline int try_to_freeze(void)
  1322. {
  1323. if (freezing(current)) {
  1324. refrigerator();
  1325. return 1;
  1326. } else
  1327. return 0;
  1328. }
  1329. #else
  1330. static inline int frozen(struct task_struct *p) { return 0; }
  1331. static inline int freezing(struct task_struct *p) { return 0; }
  1332. static inline void freeze(struct task_struct *p) { BUG(); }
  1333. static inline int thaw_process(struct task_struct *p) { return 1; }
  1334. static inline void frozen_process(struct task_struct *p) { BUG(); }
  1335. static inline void refrigerator(void) {}
  1336. static inline int freeze_processes(void) { BUG(); return 0; }
  1337. static inline void thaw_processes(void) {}
  1338. static inline int try_to_freeze(void) { return 0; }
  1339. #endif /* CONFIG_PM */
  1340. #endif /* __KERNEL__ */
  1341. #endif