sched.h 52 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710
  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. #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
  26. #define CLONE_NEWIPC 0x08000000 /* New ipcs */
  27. /*
  28. * Scheduling policies
  29. */
  30. #define SCHED_NORMAL 0
  31. #define SCHED_FIFO 1
  32. #define SCHED_RR 2
  33. #define SCHED_BATCH 3
  34. /* SCHED_ISO: reserved but not implemented yet */
  35. #define SCHED_IDLE 5
  36. #ifdef __KERNEL__
  37. struct sched_param {
  38. int sched_priority;
  39. };
  40. #include <asm/param.h> /* for HZ */
  41. #include <linux/capability.h>
  42. #include <linux/threads.h>
  43. #include <linux/kernel.h>
  44. #include <linux/types.h>
  45. #include <linux/timex.h>
  46. #include <linux/jiffies.h>
  47. #include <linux/rbtree.h>
  48. #include <linux/thread_info.h>
  49. #include <linux/cpumask.h>
  50. #include <linux/errno.h>
  51. #include <linux/nodemask.h>
  52. #include <asm/system.h>
  53. #include <asm/semaphore.h>
  54. #include <asm/page.h>
  55. #include <asm/ptrace.h>
  56. #include <asm/mmu.h>
  57. #include <asm/cputime.h>
  58. #include <linux/smp.h>
  59. #include <linux/sem.h>
  60. #include <linux/signal.h>
  61. #include <linux/securebits.h>
  62. #include <linux/fs_struct.h>
  63. #include <linux/compiler.h>
  64. #include <linux/completion.h>
  65. #include <linux/pid.h>
  66. #include <linux/percpu.h>
  67. #include <linux/topology.h>
  68. #include <linux/seccomp.h>
  69. #include <linux/rcupdate.h>
  70. #include <linux/futex.h>
  71. #include <linux/rtmutex.h>
  72. #include <linux/time.h>
  73. #include <linux/param.h>
  74. #include <linux/resource.h>
  75. #include <linux/timer.h>
  76. #include <linux/hrtimer.h>
  77. #include <linux/task_io_accounting.h>
  78. #include <asm/processor.h>
  79. struct exec_domain;
  80. struct futex_pi_state;
  81. struct bio;
  82. /*
  83. * List of flags we want to share for kernel threads,
  84. * if only because they are not used by them anyway.
  85. */
  86. #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
  87. /*
  88. * These are the constant used to fake the fixed-point load-average
  89. * counting. Some notes:
  90. * - 11 bit fractions expand to 22 bits by the multiplies: this gives
  91. * a load-average precision of 10 bits integer + 11 bits fractional
  92. * - if you want to count load-averages more often, you need more
  93. * precision, or rounding will get you. With 2-second counting freq,
  94. * the EXP_n values would be 1981, 2034 and 2043 if still using only
  95. * 11 bit fractions.
  96. */
  97. extern unsigned long avenrun[]; /* Load averages */
  98. #define FSHIFT 11 /* nr of bits of precision */
  99. #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
  100. #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
  101. #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
  102. #define EXP_5 2014 /* 1/exp(5sec/5min) */
  103. #define EXP_15 2037 /* 1/exp(5sec/15min) */
  104. #define CALC_LOAD(load,exp,n) \
  105. load *= exp; \
  106. load += n*(FIXED_1-exp); \
  107. load >>= FSHIFT;
  108. extern unsigned long total_forks;
  109. extern int nr_threads;
  110. DECLARE_PER_CPU(unsigned long, process_counts);
  111. extern int nr_processes(void);
  112. extern unsigned long nr_running(void);
  113. extern unsigned long nr_uninterruptible(void);
  114. extern unsigned long nr_active(void);
  115. extern unsigned long nr_iowait(void);
  116. extern unsigned long weighted_cpuload(const int cpu);
  117. /*
  118. * Task state bitmask. NOTE! These bits are also
  119. * encoded in fs/proc/array.c: get_task_state().
  120. *
  121. * We have two separate sets of flags: task->state
  122. * is about runnability, while task->exit_state are
  123. * about the task exiting. Confusing, but this way
  124. * modifying one set can't modify the other one by
  125. * mistake.
  126. */
  127. #define TASK_RUNNING 0
  128. #define TASK_INTERRUPTIBLE 1
  129. #define TASK_UNINTERRUPTIBLE 2
  130. #define TASK_STOPPED 4
  131. #define TASK_TRACED 8
  132. /* in tsk->exit_state */
  133. #define EXIT_ZOMBIE 16
  134. #define EXIT_DEAD 32
  135. /* in tsk->state again */
  136. #define TASK_NONINTERACTIVE 64
  137. #define TASK_DEAD 128
  138. #define __set_task_state(tsk, state_value) \
  139. do { (tsk)->state = (state_value); } while (0)
  140. #define set_task_state(tsk, state_value) \
  141. set_mb((tsk)->state, (state_value))
  142. /*
  143. * set_current_state() includes a barrier so that the write of current->state
  144. * is correctly serialised wrt the caller's subsequent test of whether to
  145. * actually sleep:
  146. *
  147. * set_current_state(TASK_UNINTERRUPTIBLE);
  148. * if (do_i_need_to_sleep())
  149. * schedule();
  150. *
  151. * If the caller does not need such serialisation then use __set_current_state()
  152. */
  153. #define __set_current_state(state_value) \
  154. do { current->state = (state_value); } while (0)
  155. #define set_current_state(state_value) \
  156. set_mb(current->state, (state_value))
  157. /* Task command name length */
  158. #define TASK_COMM_LEN 16
  159. #include <linux/spinlock.h>
  160. /*
  161. * This serializes "schedule()" and also protects
  162. * the run-queue from deletions/modifications (but
  163. * _adding_ to the beginning of the run-queue has
  164. * a separate lock).
  165. */
  166. extern rwlock_t tasklist_lock;
  167. extern spinlock_t mmlist_lock;
  168. struct task_struct;
  169. extern void sched_init(void);
  170. extern void sched_init_smp(void);
  171. extern void init_idle(struct task_struct *idle, int cpu);
  172. extern void init_idle_bootup_task(struct task_struct *idle);
  173. extern cpumask_t nohz_cpu_mask;
  174. #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
  175. extern int select_nohz_load_balancer(int cpu);
  176. #else
  177. static inline int select_nohz_load_balancer(int cpu)
  178. {
  179. return 0;
  180. }
  181. #endif
  182. /*
  183. * Only dump TASK_* tasks. (0 for all tasks)
  184. */
  185. extern void show_state_filter(unsigned long state_filter);
  186. static inline void show_state(void)
  187. {
  188. show_state_filter(0);
  189. }
  190. extern void show_regs(struct pt_regs *);
  191. /*
  192. * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
  193. * task), SP is the stack pointer of the first frame that should be shown in the back
  194. * trace (or NULL if the entire call-chain of the task should be shown).
  195. */
  196. extern void show_stack(struct task_struct *task, unsigned long *sp);
  197. void io_schedule(void);
  198. long io_schedule_timeout(long timeout);
  199. extern void cpu_init (void);
  200. extern void trap_init(void);
  201. extern void update_process_times(int user);
  202. extern void scheduler_tick(void);
  203. #ifdef CONFIG_DETECT_SOFTLOCKUP
  204. extern void softlockup_tick(void);
  205. extern void spawn_softlockup_task(void);
  206. extern void touch_softlockup_watchdog(void);
  207. extern void touch_all_softlockup_watchdogs(void);
  208. #else
  209. static inline void softlockup_tick(void)
  210. {
  211. }
  212. static inline void spawn_softlockup_task(void)
  213. {
  214. }
  215. static inline void touch_softlockup_watchdog(void)
  216. {
  217. }
  218. static inline void touch_all_softlockup_watchdogs(void)
  219. {
  220. }
  221. #endif
  222. /* Attach to any functions which should be ignored in wchan output. */
  223. #define __sched __attribute__((__section__(".sched.text")))
  224. /* Is this address in the __sched functions? */
  225. extern int in_sched_functions(unsigned long addr);
  226. #define MAX_SCHEDULE_TIMEOUT LONG_MAX
  227. extern signed long FASTCALL(schedule_timeout(signed long timeout));
  228. extern signed long schedule_timeout_interruptible(signed long timeout);
  229. extern signed long schedule_timeout_uninterruptible(signed long timeout);
  230. asmlinkage void schedule(void);
  231. struct nsproxy;
  232. /* Maximum number of active map areas.. This is a random (large) number */
  233. #define DEFAULT_MAX_MAP_COUNT 65536
  234. extern int sysctl_max_map_count;
  235. #include <linux/aio.h>
  236. extern unsigned long
  237. arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
  238. unsigned long, unsigned long);
  239. extern unsigned long
  240. arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
  241. unsigned long len, unsigned long pgoff,
  242. unsigned long flags);
  243. extern void arch_unmap_area(struct mm_struct *, unsigned long);
  244. extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
  245. #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
  246. /*
  247. * The mm counters are not protected by its page_table_lock,
  248. * so must be incremented atomically.
  249. */
  250. #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
  251. #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
  252. #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
  253. #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
  254. #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
  255. typedef atomic_long_t mm_counter_t;
  256. #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  257. /*
  258. * The mm counters are protected by its page_table_lock,
  259. * so can be incremented directly.
  260. */
  261. #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
  262. #define get_mm_counter(mm, member) ((mm)->_##member)
  263. #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
  264. #define inc_mm_counter(mm, member) (mm)->_##member++
  265. #define dec_mm_counter(mm, member) (mm)->_##member--
  266. typedef unsigned long mm_counter_t;
  267. #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  268. #define get_mm_rss(mm) \
  269. (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
  270. #define update_hiwater_rss(mm) do { \
  271. unsigned long _rss = get_mm_rss(mm); \
  272. if ((mm)->hiwater_rss < _rss) \
  273. (mm)->hiwater_rss = _rss; \
  274. } while (0)
  275. #define update_hiwater_vm(mm) do { \
  276. if ((mm)->hiwater_vm < (mm)->total_vm) \
  277. (mm)->hiwater_vm = (mm)->total_vm; \
  278. } while (0)
  279. struct mm_struct {
  280. struct vm_area_struct * mmap; /* list of VMAs */
  281. struct rb_root mm_rb;
  282. struct vm_area_struct * mmap_cache; /* last find_vma result */
  283. unsigned long (*get_unmapped_area) (struct file *filp,
  284. unsigned long addr, unsigned long len,
  285. unsigned long pgoff, unsigned long flags);
  286. void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
  287. unsigned long mmap_base; /* base of mmap area */
  288. unsigned long task_size; /* size of task vm space */
  289. unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
  290. unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
  291. pgd_t * pgd;
  292. atomic_t mm_users; /* How many users with user space? */
  293. atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
  294. int map_count; /* number of VMAs */
  295. struct rw_semaphore mmap_sem;
  296. spinlock_t page_table_lock; /* Protects page tables and some counters */
  297. struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
  298. * together off init_mm.mmlist, and are protected
  299. * by mmlist_lock
  300. */
  301. /* Special counters, in some configurations protected by the
  302. * page_table_lock, in other configurations by being atomic.
  303. */
  304. mm_counter_t _file_rss;
  305. mm_counter_t _anon_rss;
  306. unsigned long hiwater_rss; /* High-watermark of RSS usage */
  307. unsigned long hiwater_vm; /* High-water virtual memory usage */
  308. unsigned long total_vm, locked_vm, shared_vm, exec_vm;
  309. unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
  310. unsigned long start_code, end_code, start_data, end_data;
  311. unsigned long start_brk, brk, start_stack;
  312. unsigned long arg_start, arg_end, env_start, env_end;
  313. unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
  314. cpumask_t cpu_vm_mask;
  315. /* Architecture-specific MM context */
  316. mm_context_t context;
  317. /* Swap token stuff */
  318. /*
  319. * Last value of global fault stamp as seen by this process.
  320. * In other words, this value gives an indication of how long
  321. * it has been since this task got the token.
  322. * Look at mm/thrash.c
  323. */
  324. unsigned int faultstamp;
  325. unsigned int token_priority;
  326. unsigned int last_interval;
  327. unsigned char dumpable:2;
  328. /* coredumping support */
  329. int core_waiters;
  330. struct completion *core_startup_done, core_done;
  331. /* aio bits */
  332. rwlock_t ioctx_list_lock;
  333. struct kioctx *ioctx_list;
  334. };
  335. struct sighand_struct {
  336. atomic_t count;
  337. struct k_sigaction action[_NSIG];
  338. spinlock_t siglock;
  339. struct list_head signalfd_list;
  340. };
  341. struct pacct_struct {
  342. int ac_flag;
  343. long ac_exitcode;
  344. unsigned long ac_mem;
  345. cputime_t ac_utime, ac_stime;
  346. unsigned long ac_minflt, ac_majflt;
  347. };
  348. /*
  349. * NOTE! "signal_struct" does not have it's own
  350. * locking, because a shared signal_struct always
  351. * implies a shared sighand_struct, so locking
  352. * sighand_struct is always a proper superset of
  353. * the locking of signal_struct.
  354. */
  355. struct signal_struct {
  356. atomic_t count;
  357. atomic_t live;
  358. wait_queue_head_t wait_chldexit; /* for wait4() */
  359. /* current thread group signal load-balancing target: */
  360. struct task_struct *curr_target;
  361. /* shared signal handling: */
  362. struct sigpending shared_pending;
  363. /* thread group exit support */
  364. int group_exit_code;
  365. /* overloaded:
  366. * - notify group_exit_task when ->count is equal to notify_count
  367. * - everyone except group_exit_task is stopped during signal delivery
  368. * of fatal signals, group_exit_task processes the signal.
  369. */
  370. struct task_struct *group_exit_task;
  371. int notify_count;
  372. /* thread group stop support, overloads group_exit_code too */
  373. int group_stop_count;
  374. unsigned int flags; /* see SIGNAL_* flags below */
  375. /* POSIX.1b Interval Timers */
  376. struct list_head posix_timers;
  377. /* ITIMER_REAL timer for the process */
  378. struct hrtimer real_timer;
  379. struct task_struct *tsk;
  380. ktime_t it_real_incr;
  381. /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
  382. cputime_t it_prof_expires, it_virt_expires;
  383. cputime_t it_prof_incr, it_virt_incr;
  384. /* job control IDs */
  385. pid_t pgrp;
  386. struct pid *tty_old_pgrp;
  387. union {
  388. pid_t session __deprecated;
  389. pid_t __session;
  390. };
  391. /* boolean value for session group leader */
  392. int leader;
  393. struct tty_struct *tty; /* NULL if no tty */
  394. /*
  395. * Cumulative resource counters for dead threads in the group,
  396. * and for reaped dead child processes forked by this group.
  397. * Live threads maintain their own counters and add to these
  398. * in __exit_signal, except for the group leader.
  399. */
  400. cputime_t utime, stime, cutime, cstime;
  401. unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
  402. unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
  403. unsigned long inblock, oublock, cinblock, coublock;
  404. /*
  405. * Cumulative ns of scheduled CPU time for dead threads in the
  406. * group, not including a zombie group leader. (This only differs
  407. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  408. * other than jiffies.)
  409. */
  410. unsigned long long sched_time;
  411. /*
  412. * We don't bother to synchronize most readers of this at all,
  413. * because there is no reader checking a limit that actually needs
  414. * to get both rlim_cur and rlim_max atomically, and either one
  415. * alone is a single word that can safely be read normally.
  416. * getrlimit/setrlimit use task_lock(current->group_leader) to
  417. * protect this instead of the siglock, because they really
  418. * have no need to disable irqs.
  419. */
  420. struct rlimit rlim[RLIM_NLIMITS];
  421. struct list_head cpu_timers[3];
  422. /* keep the process-shared keyrings here so that they do the right
  423. * thing in threads created with CLONE_THREAD */
  424. #ifdef CONFIG_KEYS
  425. struct key *session_keyring; /* keyring inherited over fork */
  426. struct key *process_keyring; /* keyring private to this process */
  427. #endif
  428. #ifdef CONFIG_BSD_PROCESS_ACCT
  429. struct pacct_struct pacct; /* per-process accounting information */
  430. #endif
  431. #ifdef CONFIG_TASKSTATS
  432. struct taskstats *stats;
  433. #endif
  434. };
  435. /* Context switch must be unlocked if interrupts are to be enabled */
  436. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  437. # define __ARCH_WANT_UNLOCKED_CTXSW
  438. #endif
  439. /*
  440. * Bits in flags field of signal_struct.
  441. */
  442. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  443. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  444. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  445. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  446. /*
  447. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  448. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  449. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  450. * values are inverted: lower p->prio value means higher priority.
  451. *
  452. * The MAX_USER_RT_PRIO value allows the actual maximum
  453. * RT priority to be separate from the value exported to
  454. * user-space. This allows kernel threads to set their
  455. * priority to a value higher than any user task. Note:
  456. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  457. */
  458. #define MAX_USER_RT_PRIO 100
  459. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  460. #define MAX_PRIO (MAX_RT_PRIO + 40)
  461. #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
  462. #define rt_task(p) rt_prio((p)->prio)
  463. #define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
  464. #define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
  465. #define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
  466. /*
  467. * Some day this will be a full-fledged user tracking system..
  468. */
  469. struct user_struct {
  470. atomic_t __count; /* reference count */
  471. atomic_t processes; /* How many processes does this user have? */
  472. atomic_t files; /* How many open files does this user have? */
  473. atomic_t sigpending; /* How many pending signals does this user have? */
  474. #ifdef CONFIG_INOTIFY_USER
  475. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  476. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  477. #endif
  478. /* protected by mq_lock */
  479. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  480. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  481. #ifdef CONFIG_KEYS
  482. struct key *uid_keyring; /* UID specific keyring */
  483. struct key *session_keyring; /* UID's default session keyring */
  484. #endif
  485. /* Hash table maintenance information */
  486. struct list_head uidhash_list;
  487. uid_t uid;
  488. };
  489. extern struct user_struct *find_user(uid_t);
  490. extern struct user_struct root_user;
  491. #define INIT_USER (&root_user)
  492. struct backing_dev_info;
  493. struct reclaim_state;
  494. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  495. struct sched_info {
  496. /* cumulative counters */
  497. unsigned long cpu_time, /* time spent on the cpu */
  498. run_delay, /* time spent waiting on a runqueue */
  499. pcnt; /* # of timeslices run on this cpu */
  500. /* timestamps */
  501. unsigned long last_arrival, /* when we last ran on a cpu */
  502. last_queued; /* when we were last queued to run */
  503. };
  504. #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
  505. #ifdef CONFIG_SCHEDSTATS
  506. extern const struct file_operations proc_schedstat_operations;
  507. #endif /* CONFIG_SCHEDSTATS */
  508. #ifdef CONFIG_TASK_DELAY_ACCT
  509. struct task_delay_info {
  510. spinlock_t lock;
  511. unsigned int flags; /* Private per-task flags */
  512. /* For each stat XXX, add following, aligned appropriately
  513. *
  514. * struct timespec XXX_start, XXX_end;
  515. * u64 XXX_delay;
  516. * u32 XXX_count;
  517. *
  518. * Atomicity of updates to XXX_delay, XXX_count protected by
  519. * single lock above (split into XXX_lock if contention is an issue).
  520. */
  521. /*
  522. * XXX_count is incremented on every XXX operation, the delay
  523. * associated with the operation is added to XXX_delay.
  524. * XXX_delay contains the accumulated delay time in nanoseconds.
  525. */
  526. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  527. u64 blkio_delay; /* wait for sync block io completion */
  528. u64 swapin_delay; /* wait for swapin block io completion */
  529. u32 blkio_count; /* total count of the number of sync block */
  530. /* io operations performed */
  531. u32 swapin_count; /* total count of the number of swapin block */
  532. /* io operations performed */
  533. };
  534. #endif /* CONFIG_TASK_DELAY_ACCT */
  535. static inline int sched_info_on(void)
  536. {
  537. #ifdef CONFIG_SCHEDSTATS
  538. return 1;
  539. #elif defined(CONFIG_TASK_DELAY_ACCT)
  540. extern int delayacct_on;
  541. return delayacct_on;
  542. #else
  543. return 0;
  544. #endif
  545. }
  546. enum cpu_idle_type {
  547. CPU_IDLE,
  548. CPU_NOT_IDLE,
  549. CPU_NEWLY_IDLE,
  550. CPU_MAX_IDLE_TYPES
  551. };
  552. /*
  553. * sched-domains (multiprocessor balancing) declarations:
  554. */
  555. #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
  556. #ifdef CONFIG_SMP
  557. #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
  558. #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
  559. #define SD_BALANCE_EXEC 4 /* Balance on exec */
  560. #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
  561. #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
  562. #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
  563. #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
  564. #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
  565. #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
  566. #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
  567. #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
  568. #define BALANCE_FOR_MC_POWER \
  569. (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
  570. #define BALANCE_FOR_PKG_POWER \
  571. ((sched_mc_power_savings || sched_smt_power_savings) ? \
  572. SD_POWERSAVINGS_BALANCE : 0)
  573. #define test_sd_parent(sd, flag) ((sd->parent && \
  574. (sd->parent->flags & flag)) ? 1 : 0)
  575. struct sched_group {
  576. struct sched_group *next; /* Must be a circular list */
  577. cpumask_t cpumask;
  578. /*
  579. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  580. * single CPU. This is read only (except for setup, hotplug CPU).
  581. * Note : Never change cpu_power without recompute its reciprocal
  582. */
  583. unsigned int __cpu_power;
  584. /*
  585. * reciprocal value of cpu_power to avoid expensive divides
  586. * (see include/linux/reciprocal_div.h)
  587. */
  588. u32 reciprocal_cpu_power;
  589. };
  590. struct sched_domain {
  591. /* These fields must be setup */
  592. struct sched_domain *parent; /* top domain must be null terminated */
  593. struct sched_domain *child; /* bottom domain must be null terminated */
  594. struct sched_group *groups; /* the balancing groups of the domain */
  595. cpumask_t span; /* span of all CPUs in this domain */
  596. unsigned long min_interval; /* Minimum balance interval ms */
  597. unsigned long max_interval; /* Maximum balance interval ms */
  598. unsigned int busy_factor; /* less balancing by factor if busy */
  599. unsigned int imbalance_pct; /* No balance until over watermark */
  600. unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
  601. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  602. unsigned int busy_idx;
  603. unsigned int idle_idx;
  604. unsigned int newidle_idx;
  605. unsigned int wake_idx;
  606. unsigned int forkexec_idx;
  607. int flags; /* See SD_* */
  608. /* Runtime fields. */
  609. unsigned long last_balance; /* init to jiffies. units in jiffies */
  610. unsigned int balance_interval; /* initialise to 1. units in ms. */
  611. unsigned int nr_balance_failed; /* initialise to 0 */
  612. #ifdef CONFIG_SCHEDSTATS
  613. /* load_balance() stats */
  614. unsigned long lb_cnt[CPU_MAX_IDLE_TYPES];
  615. unsigned long lb_failed[CPU_MAX_IDLE_TYPES];
  616. unsigned long lb_balanced[CPU_MAX_IDLE_TYPES];
  617. unsigned long lb_imbalance[CPU_MAX_IDLE_TYPES];
  618. unsigned long lb_gained[CPU_MAX_IDLE_TYPES];
  619. unsigned long lb_hot_gained[CPU_MAX_IDLE_TYPES];
  620. unsigned long lb_nobusyg[CPU_MAX_IDLE_TYPES];
  621. unsigned long lb_nobusyq[CPU_MAX_IDLE_TYPES];
  622. /* Active load balancing */
  623. unsigned long alb_cnt;
  624. unsigned long alb_failed;
  625. unsigned long alb_pushed;
  626. /* SD_BALANCE_EXEC stats */
  627. unsigned long sbe_cnt;
  628. unsigned long sbe_balanced;
  629. unsigned long sbe_pushed;
  630. /* SD_BALANCE_FORK stats */
  631. unsigned long sbf_cnt;
  632. unsigned long sbf_balanced;
  633. unsigned long sbf_pushed;
  634. /* try_to_wake_up() stats */
  635. unsigned long ttwu_wake_remote;
  636. unsigned long ttwu_move_affine;
  637. unsigned long ttwu_move_balance;
  638. #endif
  639. };
  640. extern int partition_sched_domains(cpumask_t *partition1,
  641. cpumask_t *partition2);
  642. #endif /* CONFIG_SMP */
  643. struct io_context; /* See blkdev.h */
  644. struct cpuset;
  645. #define NGROUPS_SMALL 32
  646. #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
  647. struct group_info {
  648. int ngroups;
  649. atomic_t usage;
  650. gid_t small_block[NGROUPS_SMALL];
  651. int nblocks;
  652. gid_t *blocks[0];
  653. };
  654. /*
  655. * get_group_info() must be called with the owning task locked (via task_lock())
  656. * when task != current. The reason being that the vast majority of callers are
  657. * looking at current->group_info, which can not be changed except by the
  658. * current task. Changing current->group_info requires the task lock, too.
  659. */
  660. #define get_group_info(group_info) do { \
  661. atomic_inc(&(group_info)->usage); \
  662. } while (0)
  663. #define put_group_info(group_info) do { \
  664. if (atomic_dec_and_test(&(group_info)->usage)) \
  665. groups_free(group_info); \
  666. } while (0)
  667. extern struct group_info *groups_alloc(int gidsetsize);
  668. extern void groups_free(struct group_info *group_info);
  669. extern int set_current_groups(struct group_info *group_info);
  670. extern int groups_search(struct group_info *group_info, gid_t grp);
  671. /* access the groups "array" with this macro */
  672. #define GROUP_AT(gi, i) \
  673. ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
  674. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  675. extern void prefetch_stack(struct task_struct *t);
  676. #else
  677. static inline void prefetch_stack(struct task_struct *t) { }
  678. #endif
  679. struct audit_context; /* See audit.c */
  680. struct mempolicy;
  681. struct pipe_inode_info;
  682. struct uts_namespace;
  683. enum sleep_type {
  684. SLEEP_NORMAL,
  685. SLEEP_NONINTERACTIVE,
  686. SLEEP_INTERACTIVE,
  687. SLEEP_INTERRUPTED,
  688. };
  689. struct prio_array;
  690. struct task_struct {
  691. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  692. void *stack;
  693. atomic_t usage;
  694. unsigned int flags; /* per process flags, defined below */
  695. unsigned int ptrace;
  696. int lock_depth; /* BKL lock depth */
  697. #ifdef CONFIG_SMP
  698. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  699. int oncpu;
  700. #endif
  701. #endif
  702. int load_weight; /* for niceness load balancing purposes */
  703. int prio, static_prio, normal_prio;
  704. struct list_head run_list;
  705. struct prio_array *array;
  706. unsigned short ioprio;
  707. #ifdef CONFIG_BLK_DEV_IO_TRACE
  708. unsigned int btrace_seq;
  709. #endif
  710. unsigned long sleep_avg;
  711. unsigned long long timestamp, last_ran;
  712. unsigned long long sched_time; /* sched_clock time spent running */
  713. enum sleep_type sleep_type;
  714. unsigned int policy;
  715. cpumask_t cpus_allowed;
  716. unsigned int time_slice, first_time_slice;
  717. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  718. struct sched_info sched_info;
  719. #endif
  720. struct list_head tasks;
  721. /*
  722. * ptrace_list/ptrace_children forms the list of my children
  723. * that were stolen by a ptracer.
  724. */
  725. struct list_head ptrace_children;
  726. struct list_head ptrace_list;
  727. struct mm_struct *mm, *active_mm;
  728. /* task state */
  729. struct linux_binfmt *binfmt;
  730. int exit_state;
  731. int exit_code, exit_signal;
  732. int pdeath_signal; /* The signal sent when the parent dies */
  733. /* ??? */
  734. unsigned int personality;
  735. unsigned did_exec:1;
  736. pid_t pid;
  737. pid_t tgid;
  738. #ifdef CONFIG_CC_STACKPROTECTOR
  739. /* Canary value for the -fstack-protector gcc feature */
  740. unsigned long stack_canary;
  741. #endif
  742. /*
  743. * pointers to (original) parent process, youngest child, younger sibling,
  744. * older sibling, respectively. (p->father can be replaced with
  745. * p->parent->pid)
  746. */
  747. struct task_struct *real_parent; /* real parent process (when being debugged) */
  748. struct task_struct *parent; /* parent process */
  749. /*
  750. * children/sibling forms the list of my children plus the
  751. * tasks I'm ptracing.
  752. */
  753. struct list_head children; /* list of my children */
  754. struct list_head sibling; /* linkage in my parent's children list */
  755. struct task_struct *group_leader; /* threadgroup leader */
  756. /* PID/PID hash table linkage. */
  757. struct pid_link pids[PIDTYPE_MAX];
  758. struct list_head thread_group;
  759. struct completion *vfork_done; /* for vfork() */
  760. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  761. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  762. unsigned int rt_priority;
  763. cputime_t utime, stime;
  764. unsigned long nvcsw, nivcsw; /* context switch counts */
  765. struct timespec start_time;
  766. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  767. unsigned long min_flt, maj_flt;
  768. cputime_t it_prof_expires, it_virt_expires;
  769. unsigned long long it_sched_expires;
  770. struct list_head cpu_timers[3];
  771. /* process credentials */
  772. uid_t uid,euid,suid,fsuid;
  773. gid_t gid,egid,sgid,fsgid;
  774. struct group_info *group_info;
  775. kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
  776. unsigned keep_capabilities:1;
  777. struct user_struct *user;
  778. #ifdef CONFIG_KEYS
  779. struct key *request_key_auth; /* assumed request_key authority */
  780. struct key *thread_keyring; /* keyring private to this thread */
  781. unsigned char jit_keyring; /* default keyring to attach requested keys to */
  782. #endif
  783. /*
  784. * fpu_counter contains the number of consecutive context switches
  785. * that the FPU is used. If this is over a threshold, the lazy fpu
  786. * saving becomes unlazy to save the trap. This is an unsigned char
  787. * so that after 256 times the counter wraps and the behavior turns
  788. * lazy again; this to deal with bursty apps that only use FPU for
  789. * a short time
  790. */
  791. unsigned char fpu_counter;
  792. int oomkilladj; /* OOM kill score adjustment (bit shift). */
  793. char comm[TASK_COMM_LEN]; /* executable name excluding path
  794. - access with [gs]et_task_comm (which lock
  795. it with task_lock())
  796. - initialized normally by flush_old_exec */
  797. /* file system info */
  798. int link_count, total_link_count;
  799. #ifdef CONFIG_SYSVIPC
  800. /* ipc stuff */
  801. struct sysv_sem sysvsem;
  802. #endif
  803. /* CPU-specific state of this task */
  804. struct thread_struct thread;
  805. /* filesystem information */
  806. struct fs_struct *fs;
  807. /* open file information */
  808. struct files_struct *files;
  809. /* namespaces */
  810. struct nsproxy *nsproxy;
  811. /* signal handlers */
  812. struct signal_struct *signal;
  813. struct sighand_struct *sighand;
  814. sigset_t blocked, real_blocked;
  815. sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
  816. struct sigpending pending;
  817. unsigned long sas_ss_sp;
  818. size_t sas_ss_size;
  819. int (*notifier)(void *priv);
  820. void *notifier_data;
  821. sigset_t *notifier_mask;
  822. void *security;
  823. struct audit_context *audit_context;
  824. seccomp_t seccomp;
  825. /* Thread group tracking */
  826. u32 parent_exec_id;
  827. u32 self_exec_id;
  828. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
  829. spinlock_t alloc_lock;
  830. /* Protection of the PI data structures: */
  831. spinlock_t pi_lock;
  832. #ifdef CONFIG_RT_MUTEXES
  833. /* PI waiters blocked on a rt_mutex held by this task */
  834. struct plist_head pi_waiters;
  835. /* Deadlock detection and priority inheritance handling */
  836. struct rt_mutex_waiter *pi_blocked_on;
  837. #endif
  838. #ifdef CONFIG_DEBUG_MUTEXES
  839. /* mutex deadlock detection */
  840. struct mutex_waiter *blocked_on;
  841. #endif
  842. #ifdef CONFIG_TRACE_IRQFLAGS
  843. unsigned int irq_events;
  844. int hardirqs_enabled;
  845. unsigned long hardirq_enable_ip;
  846. unsigned int hardirq_enable_event;
  847. unsigned long hardirq_disable_ip;
  848. unsigned int hardirq_disable_event;
  849. int softirqs_enabled;
  850. unsigned long softirq_disable_ip;
  851. unsigned int softirq_disable_event;
  852. unsigned long softirq_enable_ip;
  853. unsigned int softirq_enable_event;
  854. int hardirq_context;
  855. int softirq_context;
  856. #endif
  857. #ifdef CONFIG_LOCKDEP
  858. # define MAX_LOCK_DEPTH 30UL
  859. u64 curr_chain_key;
  860. int lockdep_depth;
  861. struct held_lock held_locks[MAX_LOCK_DEPTH];
  862. unsigned int lockdep_recursion;
  863. #endif
  864. /* journalling filesystem info */
  865. void *journal_info;
  866. /* stacked block device info */
  867. struct bio *bio_list, **bio_tail;
  868. /* VM state */
  869. struct reclaim_state *reclaim_state;
  870. struct backing_dev_info *backing_dev_info;
  871. struct io_context *io_context;
  872. unsigned long ptrace_message;
  873. siginfo_t *last_siginfo; /* For ptrace use. */
  874. /*
  875. * current io wait handle: wait queue entry to use for io waits
  876. * If this thread is processing aio, this points at the waitqueue
  877. * inside the currently handled kiocb. It may be NULL (i.e. default
  878. * to a stack based synchronous wait) if its doing sync IO.
  879. */
  880. wait_queue_t *io_wait;
  881. #ifdef CONFIG_TASK_XACCT
  882. /* i/o counters(bytes read/written, #syscalls */
  883. u64 rchar, wchar, syscr, syscw;
  884. #endif
  885. struct task_io_accounting ioac;
  886. #if defined(CONFIG_TASK_XACCT)
  887. u64 acct_rss_mem1; /* accumulated rss usage */
  888. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  889. cputime_t acct_stimexpd;/* stime since last update */
  890. #endif
  891. #ifdef CONFIG_NUMA
  892. struct mempolicy *mempolicy;
  893. short il_next;
  894. #endif
  895. #ifdef CONFIG_CPUSETS
  896. struct cpuset *cpuset;
  897. nodemask_t mems_allowed;
  898. int cpuset_mems_generation;
  899. int cpuset_mem_spread_rotor;
  900. #endif
  901. struct robust_list_head __user *robust_list;
  902. #ifdef CONFIG_COMPAT
  903. struct compat_robust_list_head __user *compat_robust_list;
  904. #endif
  905. struct list_head pi_state_list;
  906. struct futex_pi_state *pi_state_cache;
  907. atomic_t fs_excl; /* holding fs exclusive resources */
  908. struct rcu_head rcu;
  909. /*
  910. * cache last used pipe for splice
  911. */
  912. struct pipe_inode_info *splice_pipe;
  913. #ifdef CONFIG_TASK_DELAY_ACCT
  914. struct task_delay_info *delays;
  915. #endif
  916. #ifdef CONFIG_FAULT_INJECTION
  917. int make_it_fail;
  918. #endif
  919. };
  920. static inline pid_t process_group(struct task_struct *tsk)
  921. {
  922. return tsk->signal->pgrp;
  923. }
  924. static inline pid_t signal_session(struct signal_struct *sig)
  925. {
  926. return sig->__session;
  927. }
  928. static inline pid_t process_session(struct task_struct *tsk)
  929. {
  930. return signal_session(tsk->signal);
  931. }
  932. static inline void set_signal_session(struct signal_struct *sig, pid_t session)
  933. {
  934. sig->__session = session;
  935. }
  936. static inline struct pid *task_pid(struct task_struct *task)
  937. {
  938. return task->pids[PIDTYPE_PID].pid;
  939. }
  940. static inline struct pid *task_tgid(struct task_struct *task)
  941. {
  942. return task->group_leader->pids[PIDTYPE_PID].pid;
  943. }
  944. static inline struct pid *task_pgrp(struct task_struct *task)
  945. {
  946. return task->group_leader->pids[PIDTYPE_PGID].pid;
  947. }
  948. static inline struct pid *task_session(struct task_struct *task)
  949. {
  950. return task->group_leader->pids[PIDTYPE_SID].pid;
  951. }
  952. /**
  953. * pid_alive - check that a task structure is not stale
  954. * @p: Task structure to be checked.
  955. *
  956. * Test if a process is not yet dead (at most zombie state)
  957. * If pid_alive fails, then pointers within the task structure
  958. * can be stale and must not be dereferenced.
  959. */
  960. static inline int pid_alive(struct task_struct *p)
  961. {
  962. return p->pids[PIDTYPE_PID].pid != NULL;
  963. }
  964. /**
  965. * is_init - check if a task structure is init
  966. * @tsk: Task structure to be checked.
  967. *
  968. * Check if a task structure is the first user space task the kernel created.
  969. */
  970. static inline int is_init(struct task_struct *tsk)
  971. {
  972. return tsk->pid == 1;
  973. }
  974. extern struct pid *cad_pid;
  975. extern void free_task(struct task_struct *tsk);
  976. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  977. extern void __put_task_struct(struct task_struct *t);
  978. static inline void put_task_struct(struct task_struct *t)
  979. {
  980. if (atomic_dec_and_test(&t->usage))
  981. __put_task_struct(t);
  982. }
  983. /*
  984. * Per process flags
  985. */
  986. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  987. /* Not implemented yet, only for 486*/
  988. #define PF_STARTING 0x00000002 /* being created */
  989. #define PF_EXITING 0x00000004 /* getting shut down */
  990. #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
  991. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  992. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  993. #define PF_DUMPCORE 0x00000200 /* dumped core */
  994. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  995. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  996. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  997. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  998. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  999. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  1000. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  1001. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  1002. #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
  1003. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  1004. #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
  1005. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  1006. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  1007. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  1008. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  1009. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  1010. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  1011. #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
  1012. /*
  1013. * Only the _current_ task can read/write to tsk->flags, but other
  1014. * tasks can access tsk->flags in readonly mode for example
  1015. * with tsk_used_math (like during threaded core dumping).
  1016. * There is however an exception to this rule during ptrace
  1017. * or during fork: the ptracer task is allowed to write to the
  1018. * child->flags of its traced child (same goes for fork, the parent
  1019. * can write to the child->flags), because we're guaranteed the
  1020. * child is not running and in turn not changing child->flags
  1021. * at the same time the parent does it.
  1022. */
  1023. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  1024. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  1025. #define clear_used_math() clear_stopped_child_used_math(current)
  1026. #define set_used_math() set_stopped_child_used_math(current)
  1027. #define conditional_stopped_child_used_math(condition, child) \
  1028. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  1029. #define conditional_used_math(condition) \
  1030. conditional_stopped_child_used_math(condition, current)
  1031. #define copy_to_stopped_child_used_math(child) \
  1032. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  1033. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  1034. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  1035. #define used_math() tsk_used_math(current)
  1036. #ifdef CONFIG_SMP
  1037. extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
  1038. #else
  1039. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  1040. {
  1041. if (!cpu_isset(0, new_mask))
  1042. return -EINVAL;
  1043. return 0;
  1044. }
  1045. #endif
  1046. extern unsigned long long sched_clock(void);
  1047. extern unsigned long long
  1048. current_sched_time(const struct task_struct *current_task);
  1049. /* sched_exec is called by processes performing an exec */
  1050. #ifdef CONFIG_SMP
  1051. extern void sched_exec(void);
  1052. #else
  1053. #define sched_exec() {}
  1054. #endif
  1055. #ifdef CONFIG_HOTPLUG_CPU
  1056. extern void idle_task_exit(void);
  1057. #else
  1058. static inline void idle_task_exit(void) {}
  1059. #endif
  1060. extern void sched_idle_next(void);
  1061. #ifdef CONFIG_RT_MUTEXES
  1062. extern int rt_mutex_getprio(struct task_struct *p);
  1063. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  1064. extern void rt_mutex_adjust_pi(struct task_struct *p);
  1065. #else
  1066. static inline int rt_mutex_getprio(struct task_struct *p)
  1067. {
  1068. return p->normal_prio;
  1069. }
  1070. # define rt_mutex_adjust_pi(p) do { } while (0)
  1071. #endif
  1072. extern void set_user_nice(struct task_struct *p, long nice);
  1073. extern int task_prio(const struct task_struct *p);
  1074. extern int task_nice(const struct task_struct *p);
  1075. extern int can_nice(const struct task_struct *p, const int nice);
  1076. extern int task_curr(const struct task_struct *p);
  1077. extern int idle_cpu(int cpu);
  1078. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  1079. extern struct task_struct *idle_task(int cpu);
  1080. extern struct task_struct *curr_task(int cpu);
  1081. extern void set_curr_task(int cpu, struct task_struct *p);
  1082. void yield(void);
  1083. /*
  1084. * The default (Linux) execution domain.
  1085. */
  1086. extern struct exec_domain default_exec_domain;
  1087. union thread_union {
  1088. struct thread_info thread_info;
  1089. unsigned long stack[THREAD_SIZE/sizeof(long)];
  1090. };
  1091. #ifndef __HAVE_ARCH_KSTACK_END
  1092. static inline int kstack_end(void *addr)
  1093. {
  1094. /* Reliable end of stack detection:
  1095. * Some APM bios versions misalign the stack
  1096. */
  1097. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  1098. }
  1099. #endif
  1100. extern union thread_union init_thread_union;
  1101. extern struct task_struct init_task;
  1102. extern struct mm_struct init_mm;
  1103. #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
  1104. extern struct task_struct *find_task_by_pid_type(int type, int pid);
  1105. extern void __set_special_pids(pid_t session, pid_t pgrp);
  1106. /* per-UID process charging. */
  1107. extern struct user_struct * alloc_uid(uid_t);
  1108. static inline struct user_struct *get_uid(struct user_struct *u)
  1109. {
  1110. atomic_inc(&u->__count);
  1111. return u;
  1112. }
  1113. extern void free_uid(struct user_struct *);
  1114. extern void switch_uid(struct user_struct *);
  1115. #include <asm/current.h>
  1116. extern void do_timer(unsigned long ticks);
  1117. extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
  1118. extern int FASTCALL(wake_up_process(struct task_struct * tsk));
  1119. extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
  1120. unsigned long clone_flags));
  1121. #ifdef CONFIG_SMP
  1122. extern void kick_process(struct task_struct *tsk);
  1123. #else
  1124. static inline void kick_process(struct task_struct *tsk) { }
  1125. #endif
  1126. extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
  1127. extern void FASTCALL(sched_exit(struct task_struct * p));
  1128. extern int in_group_p(gid_t);
  1129. extern int in_egroup_p(gid_t);
  1130. extern void proc_caches_init(void);
  1131. extern void flush_signals(struct task_struct *);
  1132. extern void ignore_signals(struct task_struct *);
  1133. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1134. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1135. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1136. {
  1137. unsigned long flags;
  1138. int ret;
  1139. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1140. ret = dequeue_signal(tsk, mask, info);
  1141. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1142. return ret;
  1143. }
  1144. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1145. sigset_t *mask);
  1146. extern void unblock_all_signals(void);
  1147. extern void release_task(struct task_struct * p);
  1148. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1149. extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
  1150. extern int force_sigsegv(int, struct task_struct *);
  1151. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1152. extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1153. extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1154. extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
  1155. extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
  1156. extern int kill_pgrp(struct pid *pid, int sig, int priv);
  1157. extern int kill_pid(struct pid *pid, int sig, int priv);
  1158. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1159. extern void do_notify_parent(struct task_struct *, int);
  1160. extern void force_sig(int, struct task_struct *);
  1161. extern void force_sig_specific(int, struct task_struct *);
  1162. extern int send_sig(int, struct task_struct *, int);
  1163. extern void zap_other_threads(struct task_struct *p);
  1164. extern int kill_proc(pid_t, int, int);
  1165. extern struct sigqueue *sigqueue_alloc(void);
  1166. extern void sigqueue_free(struct sigqueue *);
  1167. extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
  1168. extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
  1169. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1170. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1171. static inline int kill_cad_pid(int sig, int priv)
  1172. {
  1173. return kill_pid(cad_pid, sig, priv);
  1174. }
  1175. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1176. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1177. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1178. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1179. static inline int is_si_special(const struct siginfo *info)
  1180. {
  1181. return info <= SEND_SIG_FORCED;
  1182. }
  1183. /* True if we are on the alternate signal stack. */
  1184. static inline int on_sig_stack(unsigned long sp)
  1185. {
  1186. return (sp - current->sas_ss_sp < current->sas_ss_size);
  1187. }
  1188. static inline int sas_ss_flags(unsigned long sp)
  1189. {
  1190. return (current->sas_ss_size == 0 ? SS_DISABLE
  1191. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1192. }
  1193. /*
  1194. * Routines for handling mm_structs
  1195. */
  1196. extern struct mm_struct * mm_alloc(void);
  1197. /* mmdrop drops the mm and the page tables */
  1198. extern void FASTCALL(__mmdrop(struct mm_struct *));
  1199. static inline void mmdrop(struct mm_struct * mm)
  1200. {
  1201. if (atomic_dec_and_test(&mm->mm_count))
  1202. __mmdrop(mm);
  1203. }
  1204. /* mmput gets rid of the mappings and all user-space */
  1205. extern void mmput(struct mm_struct *);
  1206. /* Grab a reference to a task's mm, if it is not already going away */
  1207. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1208. /* Remove the current tasks stale references to the old mm_struct */
  1209. extern void mm_release(struct task_struct *, struct mm_struct *);
  1210. extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
  1211. extern void flush_thread(void);
  1212. extern void exit_thread(void);
  1213. extern void exit_files(struct task_struct *);
  1214. extern void __cleanup_signal(struct signal_struct *);
  1215. extern void __cleanup_sighand(struct sighand_struct *);
  1216. extern void exit_itimers(struct signal_struct *);
  1217. extern NORET_TYPE void do_group_exit(int);
  1218. extern void daemonize(const char *, ...);
  1219. extern int allow_signal(int);
  1220. extern int disallow_signal(int);
  1221. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1222. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1223. struct task_struct *fork_idle(int);
  1224. extern void set_task_comm(struct task_struct *tsk, char *from);
  1225. extern void get_task_comm(char *to, struct task_struct *tsk);
  1226. #ifdef CONFIG_SMP
  1227. extern void wait_task_inactive(struct task_struct * p);
  1228. #else
  1229. #define wait_task_inactive(p) do { } while (0)
  1230. #endif
  1231. #define remove_parent(p) list_del_init(&(p)->sibling)
  1232. #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
  1233. #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
  1234. #define for_each_process(p) \
  1235. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1236. /*
  1237. * Careful: do_each_thread/while_each_thread is a double loop so
  1238. * 'break' will not work as expected - use goto instead.
  1239. */
  1240. #define do_each_thread(g, t) \
  1241. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1242. #define while_each_thread(g, t) \
  1243. while ((t = next_thread(t)) != g)
  1244. /* de_thread depends on thread_group_leader not being a pid based check */
  1245. #define thread_group_leader(p) (p == p->group_leader)
  1246. /* Do to the insanities of de_thread it is possible for a process
  1247. * to have the pid of the thread group leader without actually being
  1248. * the thread group leader. For iteration through the pids in proc
  1249. * all we care about is that we have a task with the appropriate
  1250. * pid, we don't actually care if we have the right task.
  1251. */
  1252. static inline int has_group_leader_pid(struct task_struct *p)
  1253. {
  1254. return p->pid == p->tgid;
  1255. }
  1256. static inline struct task_struct *next_thread(const struct task_struct *p)
  1257. {
  1258. return list_entry(rcu_dereference(p->thread_group.next),
  1259. struct task_struct, thread_group);
  1260. }
  1261. static inline int thread_group_empty(struct task_struct *p)
  1262. {
  1263. return list_empty(&p->thread_group);
  1264. }
  1265. #define delay_group_leader(p) \
  1266. (thread_group_leader(p) && !thread_group_empty(p))
  1267. /*
  1268. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1269. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1270. * pins the final release of task.io_context. Also protects ->cpuset.
  1271. *
  1272. * Nests both inside and outside of read_lock(&tasklist_lock).
  1273. * It must not be nested with write_lock_irq(&tasklist_lock),
  1274. * neither inside nor outside.
  1275. */
  1276. static inline void task_lock(struct task_struct *p)
  1277. {
  1278. spin_lock(&p->alloc_lock);
  1279. }
  1280. static inline void task_unlock(struct task_struct *p)
  1281. {
  1282. spin_unlock(&p->alloc_lock);
  1283. }
  1284. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1285. unsigned long *flags);
  1286. static inline void unlock_task_sighand(struct task_struct *tsk,
  1287. unsigned long *flags)
  1288. {
  1289. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1290. }
  1291. #ifndef __HAVE_THREAD_FUNCTIONS
  1292. #define task_thread_info(task) ((struct thread_info *)(task)->stack)
  1293. #define task_stack_page(task) ((task)->stack)
  1294. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1295. {
  1296. *task_thread_info(p) = *task_thread_info(org);
  1297. task_thread_info(p)->task = p;
  1298. }
  1299. static inline unsigned long *end_of_stack(struct task_struct *p)
  1300. {
  1301. return (unsigned long *)(task_thread_info(p) + 1);
  1302. }
  1303. #endif
  1304. /* set thread flags in other task's structures
  1305. * - see asm/thread_info.h for TIF_xxxx flags available
  1306. */
  1307. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1308. {
  1309. set_ti_thread_flag(task_thread_info(tsk), flag);
  1310. }
  1311. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1312. {
  1313. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1314. }
  1315. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1316. {
  1317. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1318. }
  1319. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1320. {
  1321. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1322. }
  1323. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1324. {
  1325. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1326. }
  1327. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1328. {
  1329. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1330. }
  1331. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1332. {
  1333. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1334. }
  1335. static inline int signal_pending(struct task_struct *p)
  1336. {
  1337. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1338. }
  1339. static inline int need_resched(void)
  1340. {
  1341. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  1342. }
  1343. /*
  1344. * cond_resched() and cond_resched_lock(): latency reduction via
  1345. * explicit rescheduling in places that are safe. The return
  1346. * value indicates whether a reschedule was done in fact.
  1347. * cond_resched_lock() will drop the spinlock before scheduling,
  1348. * cond_resched_softirq() will enable bhs before scheduling.
  1349. */
  1350. extern int cond_resched(void);
  1351. extern int cond_resched_lock(spinlock_t * lock);
  1352. extern int cond_resched_softirq(void);
  1353. /*
  1354. * Does a critical section need to be broken due to another
  1355. * task waiting?:
  1356. */
  1357. #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
  1358. # define need_lockbreak(lock) ((lock)->break_lock)
  1359. #else
  1360. # define need_lockbreak(lock) 0
  1361. #endif
  1362. /*
  1363. * Does a critical section need to be broken due to another
  1364. * task waiting or preemption being signalled:
  1365. */
  1366. static inline int lock_need_resched(spinlock_t *lock)
  1367. {
  1368. if (need_lockbreak(lock) || need_resched())
  1369. return 1;
  1370. return 0;
  1371. }
  1372. /*
  1373. * Reevaluate whether the task has signals pending delivery.
  1374. * Wake the task if so.
  1375. * This is required every time the blocked sigset_t changes.
  1376. * callers must hold sighand->siglock.
  1377. */
  1378. extern void recalc_sigpending_and_wake(struct task_struct *t);
  1379. extern void recalc_sigpending(void);
  1380. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  1381. /*
  1382. * Wrappers for p->thread_info->cpu access. No-op on UP.
  1383. */
  1384. #ifdef CONFIG_SMP
  1385. static inline unsigned int task_cpu(const struct task_struct *p)
  1386. {
  1387. return task_thread_info(p)->cpu;
  1388. }
  1389. extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
  1390. #else
  1391. static inline unsigned int task_cpu(const struct task_struct *p)
  1392. {
  1393. return 0;
  1394. }
  1395. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1396. {
  1397. }
  1398. #endif /* CONFIG_SMP */
  1399. #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
  1400. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  1401. #else
  1402. static inline void arch_pick_mmap_layout(struct mm_struct *mm)
  1403. {
  1404. mm->mmap_base = TASK_UNMAPPED_BASE;
  1405. mm->get_unmapped_area = arch_get_unmapped_area;
  1406. mm->unmap_area = arch_unmap_area;
  1407. }
  1408. #endif
  1409. extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
  1410. extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
  1411. extern int sched_mc_power_savings, sched_smt_power_savings;
  1412. extern void normalize_rt_tasks(void);
  1413. #ifdef CONFIG_TASK_XACCT
  1414. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1415. {
  1416. tsk->rchar += amt;
  1417. }
  1418. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1419. {
  1420. tsk->wchar += amt;
  1421. }
  1422. static inline void inc_syscr(struct task_struct *tsk)
  1423. {
  1424. tsk->syscr++;
  1425. }
  1426. static inline void inc_syscw(struct task_struct *tsk)
  1427. {
  1428. tsk->syscw++;
  1429. }
  1430. #else
  1431. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1432. {
  1433. }
  1434. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1435. {
  1436. }
  1437. static inline void inc_syscr(struct task_struct *tsk)
  1438. {
  1439. }
  1440. static inline void inc_syscw(struct task_struct *tsk)
  1441. {
  1442. }
  1443. #endif
  1444. #endif /* __KERNEL__ */
  1445. #endif