sched.h 51 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692
  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. #ifdef __KERNEL__
  35. struct sched_param {
  36. int sched_priority;
  37. };
  38. #include <asm/param.h> /* for HZ */
  39. #include <linux/capability.h>
  40. #include <linux/threads.h>
  41. #include <linux/kernel.h>
  42. #include <linux/types.h>
  43. #include <linux/timex.h>
  44. #include <linux/jiffies.h>
  45. #include <linux/rbtree.h>
  46. #include <linux/thread_info.h>
  47. #include <linux/cpumask.h>
  48. #include <linux/errno.h>
  49. #include <linux/nodemask.h>
  50. #include <asm/system.h>
  51. #include <asm/semaphore.h>
  52. #include <asm/page.h>
  53. #include <asm/ptrace.h>
  54. #include <asm/mmu.h>
  55. #include <asm/cputime.h>
  56. #include <linux/smp.h>
  57. #include <linux/sem.h>
  58. #include <linux/signal.h>
  59. #include <linux/securebits.h>
  60. #include <linux/fs_struct.h>
  61. #include <linux/compiler.h>
  62. #include <linux/completion.h>
  63. #include <linux/pid.h>
  64. #include <linux/percpu.h>
  65. #include <linux/topology.h>
  66. #include <linux/seccomp.h>
  67. #include <linux/rcupdate.h>
  68. #include <linux/futex.h>
  69. #include <linux/rtmutex.h>
  70. #include <linux/time.h>
  71. #include <linux/param.h>
  72. #include <linux/resource.h>
  73. #include <linux/timer.h>
  74. #include <linux/hrtimer.h>
  75. #include <linux/task_io_accounting.h>
  76. #include <asm/processor.h>
  77. struct exec_domain;
  78. struct futex_pi_state;
  79. /*
  80. * List of flags we want to share for kernel threads,
  81. * if only because they are not used by them anyway.
  82. */
  83. #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
  84. /*
  85. * These are the constant used to fake the fixed-point load-average
  86. * counting. Some notes:
  87. * - 11 bit fractions expand to 22 bits by the multiplies: this gives
  88. * a load-average precision of 10 bits integer + 11 bits fractional
  89. * - if you want to count load-averages more often, you need more
  90. * precision, or rounding will get you. With 2-second counting freq,
  91. * the EXP_n values would be 1981, 2034 and 2043 if still using only
  92. * 11 bit fractions.
  93. */
  94. extern unsigned long avenrun[]; /* Load averages */
  95. #define FSHIFT 11 /* nr of bits of precision */
  96. #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
  97. #define LOAD_FREQ (5*HZ) /* 5 sec intervals */
  98. #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
  99. #define EXP_5 2014 /* 1/exp(5sec/5min) */
  100. #define EXP_15 2037 /* 1/exp(5sec/15min) */
  101. #define CALC_LOAD(load,exp,n) \
  102. load *= exp; \
  103. load += n*(FIXED_1-exp); \
  104. load >>= FSHIFT;
  105. extern unsigned long total_forks;
  106. extern int nr_threads;
  107. DECLARE_PER_CPU(unsigned long, process_counts);
  108. extern int nr_processes(void);
  109. extern unsigned long nr_running(void);
  110. extern unsigned long nr_uninterruptible(void);
  111. extern unsigned long nr_active(void);
  112. extern unsigned long nr_iowait(void);
  113. extern unsigned long weighted_cpuload(const int cpu);
  114. /*
  115. * Task state bitmask. NOTE! These bits are also
  116. * encoded in fs/proc/array.c: get_task_state().
  117. *
  118. * We have two separate sets of flags: task->state
  119. * is about runnability, while task->exit_state are
  120. * about the task exiting. Confusing, but this way
  121. * modifying one set can't modify the other one by
  122. * mistake.
  123. */
  124. #define TASK_RUNNING 0
  125. #define TASK_INTERRUPTIBLE 1
  126. #define TASK_UNINTERRUPTIBLE 2
  127. #define TASK_STOPPED 4
  128. #define TASK_TRACED 8
  129. /* in tsk->exit_state */
  130. #define EXIT_ZOMBIE 16
  131. #define EXIT_DEAD 32
  132. /* in tsk->state again */
  133. #define TASK_NONINTERACTIVE 64
  134. #define TASK_DEAD 128
  135. #define __set_task_state(tsk, state_value) \
  136. do { (tsk)->state = (state_value); } while (0)
  137. #define set_task_state(tsk, state_value) \
  138. set_mb((tsk)->state, (state_value))
  139. /*
  140. * set_current_state() includes a barrier so that the write of current->state
  141. * is correctly serialised wrt the caller's subsequent test of whether to
  142. * actually sleep:
  143. *
  144. * set_current_state(TASK_UNINTERRUPTIBLE);
  145. * if (do_i_need_to_sleep())
  146. * schedule();
  147. *
  148. * If the caller does not need such serialisation then use __set_current_state()
  149. */
  150. #define __set_current_state(state_value) \
  151. do { current->state = (state_value); } while (0)
  152. #define set_current_state(state_value) \
  153. set_mb(current->state, (state_value))
  154. /* Task command name length */
  155. #define TASK_COMM_LEN 16
  156. #include <linux/spinlock.h>
  157. /*
  158. * This serializes "schedule()" and also protects
  159. * the run-queue from deletions/modifications (but
  160. * _adding_ to the beginning of the run-queue has
  161. * a separate lock).
  162. */
  163. extern rwlock_t tasklist_lock;
  164. extern spinlock_t mmlist_lock;
  165. struct task_struct;
  166. extern void sched_init(void);
  167. extern void sched_init_smp(void);
  168. extern void init_idle(struct task_struct *idle, int cpu);
  169. extern cpumask_t nohz_cpu_mask;
  170. /*
  171. * Only dump TASK_* tasks. (-1 for all tasks)
  172. */
  173. extern void show_state_filter(unsigned long state_filter);
  174. static inline void show_state(void)
  175. {
  176. show_state_filter(-1);
  177. }
  178. extern void show_regs(struct pt_regs *);
  179. /*
  180. * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
  181. * task), SP is the stack pointer of the first frame that should be shown in the back
  182. * trace (or NULL if the entire call-chain of the task should be shown).
  183. */
  184. extern void show_stack(struct task_struct *task, unsigned long *sp);
  185. void io_schedule(void);
  186. long io_schedule_timeout(long timeout);
  187. extern void cpu_init (void);
  188. extern void trap_init(void);
  189. extern void update_process_times(int user);
  190. extern void scheduler_tick(void);
  191. #ifdef CONFIG_DETECT_SOFTLOCKUP
  192. extern void softlockup_tick(void);
  193. extern void spawn_softlockup_task(void);
  194. extern void touch_softlockup_watchdog(void);
  195. #else
  196. static inline void softlockup_tick(void)
  197. {
  198. }
  199. static inline void spawn_softlockup_task(void)
  200. {
  201. }
  202. static inline void touch_softlockup_watchdog(void)
  203. {
  204. }
  205. #endif
  206. /* Attach to any functions which should be ignored in wchan output. */
  207. #define __sched __attribute__((__section__(".sched.text")))
  208. /* Is this address in the __sched functions? */
  209. extern int in_sched_functions(unsigned long addr);
  210. #define MAX_SCHEDULE_TIMEOUT LONG_MAX
  211. extern signed long FASTCALL(schedule_timeout(signed long timeout));
  212. extern signed long schedule_timeout_interruptible(signed long timeout);
  213. extern signed long schedule_timeout_uninterruptible(signed long timeout);
  214. asmlinkage void schedule(void);
  215. struct nsproxy;
  216. /* Maximum number of active map areas.. This is a random (large) number */
  217. #define DEFAULT_MAX_MAP_COUNT 65536
  218. extern int sysctl_max_map_count;
  219. #include <linux/aio.h>
  220. extern unsigned long
  221. arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
  222. unsigned long, unsigned long);
  223. extern unsigned long
  224. arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
  225. unsigned long len, unsigned long pgoff,
  226. unsigned long flags);
  227. extern void arch_unmap_area(struct mm_struct *, unsigned long);
  228. extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
  229. #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
  230. /*
  231. * The mm counters are not protected by its page_table_lock,
  232. * so must be incremented atomically.
  233. */
  234. #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
  235. #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
  236. #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
  237. #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
  238. #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
  239. typedef atomic_long_t mm_counter_t;
  240. #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  241. /*
  242. * The mm counters are protected by its page_table_lock,
  243. * so can be incremented directly.
  244. */
  245. #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
  246. #define get_mm_counter(mm, member) ((mm)->_##member)
  247. #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
  248. #define inc_mm_counter(mm, member) (mm)->_##member++
  249. #define dec_mm_counter(mm, member) (mm)->_##member--
  250. typedef unsigned long mm_counter_t;
  251. #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  252. #define get_mm_rss(mm) \
  253. (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
  254. #define update_hiwater_rss(mm) do { \
  255. unsigned long _rss = get_mm_rss(mm); \
  256. if ((mm)->hiwater_rss < _rss) \
  257. (mm)->hiwater_rss = _rss; \
  258. } while (0)
  259. #define update_hiwater_vm(mm) do { \
  260. if ((mm)->hiwater_vm < (mm)->total_vm) \
  261. (mm)->hiwater_vm = (mm)->total_vm; \
  262. } while (0)
  263. struct mm_struct {
  264. struct vm_area_struct * mmap; /* list of VMAs */
  265. struct rb_root mm_rb;
  266. struct vm_area_struct * mmap_cache; /* last find_vma result */
  267. unsigned long (*get_unmapped_area) (struct file *filp,
  268. unsigned long addr, unsigned long len,
  269. unsigned long pgoff, unsigned long flags);
  270. void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
  271. unsigned long mmap_base; /* base of mmap area */
  272. unsigned long task_size; /* size of task vm space */
  273. unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
  274. unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
  275. pgd_t * pgd;
  276. atomic_t mm_users; /* How many users with user space? */
  277. atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
  278. int map_count; /* number of VMAs */
  279. struct rw_semaphore mmap_sem;
  280. spinlock_t page_table_lock; /* Protects page tables and some counters */
  281. struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
  282. * together off init_mm.mmlist, and are protected
  283. * by mmlist_lock
  284. */
  285. /* Special counters, in some configurations protected by the
  286. * page_table_lock, in other configurations by being atomic.
  287. */
  288. mm_counter_t _file_rss;
  289. mm_counter_t _anon_rss;
  290. unsigned long hiwater_rss; /* High-watermark of RSS usage */
  291. unsigned long hiwater_vm; /* High-water virtual memory usage */
  292. unsigned long total_vm, locked_vm, shared_vm, exec_vm;
  293. unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
  294. unsigned long start_code, end_code, start_data, end_data;
  295. unsigned long start_brk, brk, start_stack;
  296. unsigned long arg_start, arg_end, env_start, env_end;
  297. unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
  298. cpumask_t cpu_vm_mask;
  299. /* Architecture-specific MM context */
  300. mm_context_t context;
  301. /* Swap token stuff */
  302. /*
  303. * Last value of global fault stamp as seen by this process.
  304. * In other words, this value gives an indication of how long
  305. * it has been since this task got the token.
  306. * Look at mm/thrash.c
  307. */
  308. unsigned int faultstamp;
  309. unsigned int token_priority;
  310. unsigned int last_interval;
  311. unsigned char dumpable:2;
  312. /* coredumping support */
  313. int core_waiters;
  314. struct completion *core_startup_done, core_done;
  315. /* aio bits */
  316. rwlock_t ioctx_list_lock;
  317. struct kioctx *ioctx_list;
  318. };
  319. struct sighand_struct {
  320. atomic_t count;
  321. struct k_sigaction action[_NSIG];
  322. spinlock_t siglock;
  323. };
  324. struct pacct_struct {
  325. int ac_flag;
  326. long ac_exitcode;
  327. unsigned long ac_mem;
  328. cputime_t ac_utime, ac_stime;
  329. unsigned long ac_minflt, ac_majflt;
  330. };
  331. /*
  332. * NOTE! "signal_struct" does not have it's own
  333. * locking, because a shared signal_struct always
  334. * implies a shared sighand_struct, so locking
  335. * sighand_struct is always a proper superset of
  336. * the locking of signal_struct.
  337. */
  338. struct signal_struct {
  339. atomic_t count;
  340. atomic_t live;
  341. wait_queue_head_t wait_chldexit; /* for wait4() */
  342. /* current thread group signal load-balancing target: */
  343. struct task_struct *curr_target;
  344. /* shared signal handling: */
  345. struct sigpending shared_pending;
  346. /* thread group exit support */
  347. int group_exit_code;
  348. /* overloaded:
  349. * - notify group_exit_task when ->count is equal to notify_count
  350. * - everyone except group_exit_task is stopped during signal delivery
  351. * of fatal signals, group_exit_task processes the signal.
  352. */
  353. struct task_struct *group_exit_task;
  354. int notify_count;
  355. /* thread group stop support, overloads group_exit_code too */
  356. int group_stop_count;
  357. unsigned int flags; /* see SIGNAL_* flags below */
  358. /* POSIX.1b Interval Timers */
  359. struct list_head posix_timers;
  360. /* ITIMER_REAL timer for the process */
  361. struct hrtimer real_timer;
  362. struct task_struct *tsk;
  363. ktime_t it_real_incr;
  364. /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
  365. cputime_t it_prof_expires, it_virt_expires;
  366. cputime_t it_prof_incr, it_virt_incr;
  367. /* job control IDs */
  368. pid_t pgrp;
  369. struct pid *tty_old_pgrp;
  370. union {
  371. pid_t session __deprecated;
  372. pid_t __session;
  373. };
  374. /* boolean value for session group leader */
  375. int leader;
  376. struct tty_struct *tty; /* NULL if no tty */
  377. /*
  378. * Cumulative resource counters for dead threads in the group,
  379. * and for reaped dead child processes forked by this group.
  380. * Live threads maintain their own counters and add to these
  381. * in __exit_signal, except for the group leader.
  382. */
  383. cputime_t utime, stime, cutime, cstime;
  384. unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
  385. unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
  386. /*
  387. * Cumulative ns of scheduled CPU time for dead threads in the
  388. * group, not including a zombie group leader. (This only differs
  389. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  390. * other than jiffies.)
  391. */
  392. unsigned long long sched_time;
  393. /*
  394. * We don't bother to synchronize most readers of this at all,
  395. * because there is no reader checking a limit that actually needs
  396. * to get both rlim_cur and rlim_max atomically, and either one
  397. * alone is a single word that can safely be read normally.
  398. * getrlimit/setrlimit use task_lock(current->group_leader) to
  399. * protect this instead of the siglock, because they really
  400. * have no need to disable irqs.
  401. */
  402. struct rlimit rlim[RLIM_NLIMITS];
  403. struct list_head cpu_timers[3];
  404. /* keep the process-shared keyrings here so that they do the right
  405. * thing in threads created with CLONE_THREAD */
  406. #ifdef CONFIG_KEYS
  407. struct key *session_keyring; /* keyring inherited over fork */
  408. struct key *process_keyring; /* keyring private to this process */
  409. #endif
  410. #ifdef CONFIG_BSD_PROCESS_ACCT
  411. struct pacct_struct pacct; /* per-process accounting information */
  412. #endif
  413. #ifdef CONFIG_TASKSTATS
  414. struct taskstats *stats;
  415. #endif
  416. };
  417. /* Context switch must be unlocked if interrupts are to be enabled */
  418. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  419. # define __ARCH_WANT_UNLOCKED_CTXSW
  420. #endif
  421. /*
  422. * Bits in flags field of signal_struct.
  423. */
  424. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  425. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  426. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  427. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  428. /*
  429. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  430. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  431. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  432. * values are inverted: lower p->prio value means higher priority.
  433. *
  434. * The MAX_USER_RT_PRIO value allows the actual maximum
  435. * RT priority to be separate from the value exported to
  436. * user-space. This allows kernel threads to set their
  437. * priority to a value higher than any user task. Note:
  438. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  439. */
  440. #define MAX_USER_RT_PRIO 100
  441. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  442. #define MAX_PRIO (MAX_RT_PRIO + 40)
  443. #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
  444. #define rt_task(p) rt_prio((p)->prio)
  445. #define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
  446. #define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
  447. #define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
  448. /*
  449. * Some day this will be a full-fledged user tracking system..
  450. */
  451. struct user_struct {
  452. atomic_t __count; /* reference count */
  453. atomic_t processes; /* How many processes does this user have? */
  454. atomic_t files; /* How many open files does this user have? */
  455. atomic_t sigpending; /* How many pending signals does this user have? */
  456. #ifdef CONFIG_INOTIFY_USER
  457. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  458. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  459. #endif
  460. /* protected by mq_lock */
  461. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  462. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  463. #ifdef CONFIG_KEYS
  464. struct key *uid_keyring; /* UID specific keyring */
  465. struct key *session_keyring; /* UID's default session keyring */
  466. #endif
  467. /* Hash table maintenance information */
  468. struct list_head uidhash_list;
  469. uid_t uid;
  470. };
  471. extern struct user_struct *find_user(uid_t);
  472. extern struct user_struct root_user;
  473. #define INIT_USER (&root_user)
  474. struct backing_dev_info;
  475. struct reclaim_state;
  476. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  477. struct sched_info {
  478. /* cumulative counters */
  479. unsigned long cpu_time, /* time spent on the cpu */
  480. run_delay, /* time spent waiting on a runqueue */
  481. pcnt; /* # of timeslices run on this cpu */
  482. /* timestamps */
  483. unsigned long last_arrival, /* when we last ran on a cpu */
  484. last_queued; /* when we were last queued to run */
  485. };
  486. #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
  487. #ifdef CONFIG_SCHEDSTATS
  488. extern const struct file_operations proc_schedstat_operations;
  489. #endif /* CONFIG_SCHEDSTATS */
  490. #ifdef CONFIG_TASK_DELAY_ACCT
  491. struct task_delay_info {
  492. spinlock_t lock;
  493. unsigned int flags; /* Private per-task flags */
  494. /* For each stat XXX, add following, aligned appropriately
  495. *
  496. * struct timespec XXX_start, XXX_end;
  497. * u64 XXX_delay;
  498. * u32 XXX_count;
  499. *
  500. * Atomicity of updates to XXX_delay, XXX_count protected by
  501. * single lock above (split into XXX_lock if contention is an issue).
  502. */
  503. /*
  504. * XXX_count is incremented on every XXX operation, the delay
  505. * associated with the operation is added to XXX_delay.
  506. * XXX_delay contains the accumulated delay time in nanoseconds.
  507. */
  508. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  509. u64 blkio_delay; /* wait for sync block io completion */
  510. u64 swapin_delay; /* wait for swapin block io completion */
  511. u32 blkio_count; /* total count of the number of sync block */
  512. /* io operations performed */
  513. u32 swapin_count; /* total count of the number of swapin block */
  514. /* io operations performed */
  515. };
  516. #endif /* CONFIG_TASK_DELAY_ACCT */
  517. static inline int sched_info_on(void)
  518. {
  519. #ifdef CONFIG_SCHEDSTATS
  520. return 1;
  521. #elif defined(CONFIG_TASK_DELAY_ACCT)
  522. extern int delayacct_on;
  523. return delayacct_on;
  524. #else
  525. return 0;
  526. #endif
  527. }
  528. enum idle_type
  529. {
  530. SCHED_IDLE,
  531. NOT_IDLE,
  532. NEWLY_IDLE,
  533. MAX_IDLE_TYPES
  534. };
  535. /*
  536. * sched-domains (multiprocessor balancing) declarations:
  537. */
  538. #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
  539. #ifdef CONFIG_SMP
  540. #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
  541. #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
  542. #define SD_BALANCE_EXEC 4 /* Balance on exec */
  543. #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
  544. #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
  545. #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
  546. #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
  547. #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
  548. #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
  549. #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
  550. #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
  551. #define BALANCE_FOR_MC_POWER \
  552. (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
  553. #define BALANCE_FOR_PKG_POWER \
  554. ((sched_mc_power_savings || sched_smt_power_savings) ? \
  555. SD_POWERSAVINGS_BALANCE : 0)
  556. #define test_sd_parent(sd, flag) ((sd->parent && \
  557. (sd->parent->flags & flag)) ? 1 : 0)
  558. struct sched_group {
  559. struct sched_group *next; /* Must be a circular list */
  560. cpumask_t cpumask;
  561. /*
  562. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  563. * single CPU. This is read only (except for setup, hotplug CPU).
  564. */
  565. unsigned long cpu_power;
  566. };
  567. struct sched_domain {
  568. /* These fields must be setup */
  569. struct sched_domain *parent; /* top domain must be null terminated */
  570. struct sched_domain *child; /* bottom domain must be null terminated */
  571. struct sched_group *groups; /* the balancing groups of the domain */
  572. cpumask_t span; /* span of all CPUs in this domain */
  573. unsigned long min_interval; /* Minimum balance interval ms */
  574. unsigned long max_interval; /* Maximum balance interval ms */
  575. unsigned int busy_factor; /* less balancing by factor if busy */
  576. unsigned int imbalance_pct; /* No balance until over watermark */
  577. unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
  578. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  579. unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
  580. unsigned int busy_idx;
  581. unsigned int idle_idx;
  582. unsigned int newidle_idx;
  583. unsigned int wake_idx;
  584. unsigned int forkexec_idx;
  585. int flags; /* See SD_* */
  586. /* Runtime fields. */
  587. unsigned long last_balance; /* init to jiffies. units in jiffies */
  588. unsigned int balance_interval; /* initialise to 1. units in ms. */
  589. unsigned int nr_balance_failed; /* initialise to 0 */
  590. #ifdef CONFIG_SCHEDSTATS
  591. /* load_balance() stats */
  592. unsigned long lb_cnt[MAX_IDLE_TYPES];
  593. unsigned long lb_failed[MAX_IDLE_TYPES];
  594. unsigned long lb_balanced[MAX_IDLE_TYPES];
  595. unsigned long lb_imbalance[MAX_IDLE_TYPES];
  596. unsigned long lb_gained[MAX_IDLE_TYPES];
  597. unsigned long lb_hot_gained[MAX_IDLE_TYPES];
  598. unsigned long lb_nobusyg[MAX_IDLE_TYPES];
  599. unsigned long lb_nobusyq[MAX_IDLE_TYPES];
  600. /* Active load balancing */
  601. unsigned long alb_cnt;
  602. unsigned long alb_failed;
  603. unsigned long alb_pushed;
  604. /* SD_BALANCE_EXEC stats */
  605. unsigned long sbe_cnt;
  606. unsigned long sbe_balanced;
  607. unsigned long sbe_pushed;
  608. /* SD_BALANCE_FORK stats */
  609. unsigned long sbf_cnt;
  610. unsigned long sbf_balanced;
  611. unsigned long sbf_pushed;
  612. /* try_to_wake_up() stats */
  613. unsigned long ttwu_wake_remote;
  614. unsigned long ttwu_move_affine;
  615. unsigned long ttwu_move_balance;
  616. #endif
  617. };
  618. extern int partition_sched_domains(cpumask_t *partition1,
  619. cpumask_t *partition2);
  620. /*
  621. * Maximum cache size the migration-costs auto-tuning code will
  622. * search from:
  623. */
  624. extern unsigned int max_cache_size;
  625. #endif /* CONFIG_SMP */
  626. struct io_context; /* See blkdev.h */
  627. struct cpuset;
  628. #define NGROUPS_SMALL 32
  629. #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
  630. struct group_info {
  631. int ngroups;
  632. atomic_t usage;
  633. gid_t small_block[NGROUPS_SMALL];
  634. int nblocks;
  635. gid_t *blocks[0];
  636. };
  637. /*
  638. * get_group_info() must be called with the owning task locked (via task_lock())
  639. * when task != current. The reason being that the vast majority of callers are
  640. * looking at current->group_info, which can not be changed except by the
  641. * current task. Changing current->group_info requires the task lock, too.
  642. */
  643. #define get_group_info(group_info) do { \
  644. atomic_inc(&(group_info)->usage); \
  645. } while (0)
  646. #define put_group_info(group_info) do { \
  647. if (atomic_dec_and_test(&(group_info)->usage)) \
  648. groups_free(group_info); \
  649. } while (0)
  650. extern struct group_info *groups_alloc(int gidsetsize);
  651. extern void groups_free(struct group_info *group_info);
  652. extern int set_current_groups(struct group_info *group_info);
  653. extern int groups_search(struct group_info *group_info, gid_t grp);
  654. /* access the groups "array" with this macro */
  655. #define GROUP_AT(gi, i) \
  656. ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
  657. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  658. extern void prefetch_stack(struct task_struct *t);
  659. #else
  660. static inline void prefetch_stack(struct task_struct *t) { }
  661. #endif
  662. struct audit_context; /* See audit.c */
  663. struct mempolicy;
  664. struct pipe_inode_info;
  665. struct uts_namespace;
  666. enum sleep_type {
  667. SLEEP_NORMAL,
  668. SLEEP_NONINTERACTIVE,
  669. SLEEP_INTERACTIVE,
  670. SLEEP_INTERRUPTED,
  671. };
  672. struct prio_array;
  673. struct task_struct {
  674. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  675. struct thread_info *thread_info;
  676. atomic_t usage;
  677. unsigned long flags; /* per process flags, defined below */
  678. unsigned long ptrace;
  679. int lock_depth; /* BKL lock depth */
  680. #ifdef CONFIG_SMP
  681. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  682. int oncpu;
  683. #endif
  684. #endif
  685. int load_weight; /* for niceness load balancing purposes */
  686. int prio, static_prio, normal_prio;
  687. struct list_head run_list;
  688. struct prio_array *array;
  689. unsigned short ioprio;
  690. #ifdef CONFIG_BLK_DEV_IO_TRACE
  691. unsigned int btrace_seq;
  692. #endif
  693. unsigned long sleep_avg;
  694. unsigned long long timestamp, last_ran;
  695. unsigned long long sched_time; /* sched_clock time spent running */
  696. enum sleep_type sleep_type;
  697. unsigned long policy;
  698. cpumask_t cpus_allowed;
  699. unsigned int time_slice, first_time_slice;
  700. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  701. struct sched_info sched_info;
  702. #endif
  703. struct list_head tasks;
  704. /*
  705. * ptrace_list/ptrace_children forms the list of my children
  706. * that were stolen by a ptracer.
  707. */
  708. struct list_head ptrace_children;
  709. struct list_head ptrace_list;
  710. struct mm_struct *mm, *active_mm;
  711. /* task state */
  712. struct linux_binfmt *binfmt;
  713. long exit_state;
  714. int exit_code, exit_signal;
  715. int pdeath_signal; /* The signal sent when the parent dies */
  716. /* ??? */
  717. unsigned long personality;
  718. unsigned did_exec:1;
  719. pid_t pid;
  720. pid_t tgid;
  721. #ifdef CONFIG_CC_STACKPROTECTOR
  722. /* Canary value for the -fstack-protector gcc feature */
  723. unsigned long stack_canary;
  724. #endif
  725. /*
  726. * pointers to (original) parent process, youngest child, younger sibling,
  727. * older sibling, respectively. (p->father can be replaced with
  728. * p->parent->pid)
  729. */
  730. struct task_struct *real_parent; /* real parent process (when being debugged) */
  731. struct task_struct *parent; /* parent process */
  732. /*
  733. * children/sibling forms the list of my children plus the
  734. * tasks I'm ptracing.
  735. */
  736. struct list_head children; /* list of my children */
  737. struct list_head sibling; /* linkage in my parent's children list */
  738. struct task_struct *group_leader; /* threadgroup leader */
  739. /* PID/PID hash table linkage. */
  740. struct pid_link pids[PIDTYPE_MAX];
  741. struct list_head thread_group;
  742. struct completion *vfork_done; /* for vfork() */
  743. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  744. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  745. unsigned long rt_priority;
  746. cputime_t utime, stime;
  747. unsigned long nvcsw, nivcsw; /* context switch counts */
  748. struct timespec start_time;
  749. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  750. unsigned long min_flt, maj_flt;
  751. cputime_t it_prof_expires, it_virt_expires;
  752. unsigned long long it_sched_expires;
  753. struct list_head cpu_timers[3];
  754. /* process credentials */
  755. uid_t uid,euid,suid,fsuid;
  756. gid_t gid,egid,sgid,fsgid;
  757. struct group_info *group_info;
  758. kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
  759. unsigned keep_capabilities:1;
  760. struct user_struct *user;
  761. #ifdef CONFIG_KEYS
  762. struct key *request_key_auth; /* assumed request_key authority */
  763. struct key *thread_keyring; /* keyring private to this thread */
  764. unsigned char jit_keyring; /* default keyring to attach requested keys to */
  765. #endif
  766. /*
  767. * fpu_counter contains the number of consecutive context switches
  768. * that the FPU is used. If this is over a threshold, the lazy fpu
  769. * saving becomes unlazy to save the trap. This is an unsigned char
  770. * so that after 256 times the counter wraps and the behavior turns
  771. * lazy again; this to deal with bursty apps that only use FPU for
  772. * a short time
  773. */
  774. unsigned char fpu_counter;
  775. int oomkilladj; /* OOM kill score adjustment (bit shift). */
  776. char comm[TASK_COMM_LEN]; /* executable name excluding path
  777. - access with [gs]et_task_comm (which lock
  778. it with task_lock())
  779. - initialized normally by flush_old_exec */
  780. /* file system info */
  781. int link_count, total_link_count;
  782. #ifdef CONFIG_SYSVIPC
  783. /* ipc stuff */
  784. struct sysv_sem sysvsem;
  785. #endif
  786. /* CPU-specific state of this task */
  787. struct thread_struct thread;
  788. /* filesystem information */
  789. struct fs_struct *fs;
  790. /* open file information */
  791. struct files_struct *files;
  792. /* namespaces */
  793. struct nsproxy *nsproxy;
  794. /* signal handlers */
  795. struct signal_struct *signal;
  796. struct sighand_struct *sighand;
  797. sigset_t blocked, real_blocked;
  798. sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
  799. struct sigpending pending;
  800. unsigned long sas_ss_sp;
  801. size_t sas_ss_size;
  802. int (*notifier)(void *priv);
  803. void *notifier_data;
  804. sigset_t *notifier_mask;
  805. void *security;
  806. struct audit_context *audit_context;
  807. seccomp_t seccomp;
  808. /* Thread group tracking */
  809. u32 parent_exec_id;
  810. u32 self_exec_id;
  811. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
  812. spinlock_t alloc_lock;
  813. /* Protection of the PI data structures: */
  814. spinlock_t pi_lock;
  815. #ifdef CONFIG_RT_MUTEXES
  816. /* PI waiters blocked on a rt_mutex held by this task */
  817. struct plist_head pi_waiters;
  818. /* Deadlock detection and priority inheritance handling */
  819. struct rt_mutex_waiter *pi_blocked_on;
  820. #endif
  821. #ifdef CONFIG_DEBUG_MUTEXES
  822. /* mutex deadlock detection */
  823. struct mutex_waiter *blocked_on;
  824. #endif
  825. #ifdef CONFIG_TRACE_IRQFLAGS
  826. unsigned int irq_events;
  827. int hardirqs_enabled;
  828. unsigned long hardirq_enable_ip;
  829. unsigned int hardirq_enable_event;
  830. unsigned long hardirq_disable_ip;
  831. unsigned int hardirq_disable_event;
  832. int softirqs_enabled;
  833. unsigned long softirq_disable_ip;
  834. unsigned int softirq_disable_event;
  835. unsigned long softirq_enable_ip;
  836. unsigned int softirq_enable_event;
  837. int hardirq_context;
  838. int softirq_context;
  839. #endif
  840. #ifdef CONFIG_LOCKDEP
  841. # define MAX_LOCK_DEPTH 30UL
  842. u64 curr_chain_key;
  843. int lockdep_depth;
  844. struct held_lock held_locks[MAX_LOCK_DEPTH];
  845. unsigned int lockdep_recursion;
  846. #endif
  847. /* journalling filesystem info */
  848. void *journal_info;
  849. /* VM state */
  850. struct reclaim_state *reclaim_state;
  851. struct backing_dev_info *backing_dev_info;
  852. struct io_context *io_context;
  853. unsigned long ptrace_message;
  854. siginfo_t *last_siginfo; /* For ptrace use. */
  855. /*
  856. * current io wait handle: wait queue entry to use for io waits
  857. * If this thread is processing aio, this points at the waitqueue
  858. * inside the currently handled kiocb. It may be NULL (i.e. default
  859. * to a stack based synchronous wait) if its doing sync IO.
  860. */
  861. wait_queue_t *io_wait;
  862. #ifdef CONFIG_TASK_XACCT
  863. /* i/o counters(bytes read/written, #syscalls */
  864. u64 rchar, wchar, syscr, syscw;
  865. #endif
  866. struct task_io_accounting ioac;
  867. #if defined(CONFIG_TASK_XACCT)
  868. u64 acct_rss_mem1; /* accumulated rss usage */
  869. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  870. cputime_t acct_stimexpd;/* stime since last update */
  871. #endif
  872. #ifdef CONFIG_NUMA
  873. struct mempolicy *mempolicy;
  874. short il_next;
  875. #endif
  876. #ifdef CONFIG_CPUSETS
  877. struct cpuset *cpuset;
  878. nodemask_t mems_allowed;
  879. int cpuset_mems_generation;
  880. int cpuset_mem_spread_rotor;
  881. #endif
  882. struct robust_list_head __user *robust_list;
  883. #ifdef CONFIG_COMPAT
  884. struct compat_robust_list_head __user *compat_robust_list;
  885. #endif
  886. struct list_head pi_state_list;
  887. struct futex_pi_state *pi_state_cache;
  888. atomic_t fs_excl; /* holding fs exclusive resources */
  889. struct rcu_head rcu;
  890. /*
  891. * cache last used pipe for splice
  892. */
  893. struct pipe_inode_info *splice_pipe;
  894. #ifdef CONFIG_TASK_DELAY_ACCT
  895. struct task_delay_info *delays;
  896. #endif
  897. #ifdef CONFIG_FAULT_INJECTION
  898. int make_it_fail;
  899. #endif
  900. };
  901. static inline pid_t process_group(struct task_struct *tsk)
  902. {
  903. return tsk->signal->pgrp;
  904. }
  905. static inline pid_t signal_session(struct signal_struct *sig)
  906. {
  907. return sig->__session;
  908. }
  909. static inline pid_t process_session(struct task_struct *tsk)
  910. {
  911. return signal_session(tsk->signal);
  912. }
  913. static inline void set_signal_session(struct signal_struct *sig, pid_t session)
  914. {
  915. sig->__session = session;
  916. }
  917. static inline struct pid *task_pid(struct task_struct *task)
  918. {
  919. return task->pids[PIDTYPE_PID].pid;
  920. }
  921. static inline struct pid *task_tgid(struct task_struct *task)
  922. {
  923. return task->group_leader->pids[PIDTYPE_PID].pid;
  924. }
  925. static inline struct pid *task_pgrp(struct task_struct *task)
  926. {
  927. return task->group_leader->pids[PIDTYPE_PGID].pid;
  928. }
  929. static inline struct pid *task_session(struct task_struct *task)
  930. {
  931. return task->group_leader->pids[PIDTYPE_SID].pid;
  932. }
  933. /**
  934. * pid_alive - check that a task structure is not stale
  935. * @p: Task structure to be checked.
  936. *
  937. * Test if a process is not yet dead (at most zombie state)
  938. * If pid_alive fails, then pointers within the task structure
  939. * can be stale and must not be dereferenced.
  940. */
  941. static inline int pid_alive(struct task_struct *p)
  942. {
  943. return p->pids[PIDTYPE_PID].pid != NULL;
  944. }
  945. /**
  946. * is_init - check if a task structure is init
  947. * @tsk: Task structure to be checked.
  948. *
  949. * Check if a task structure is the first user space task the kernel created.
  950. */
  951. static inline int is_init(struct task_struct *tsk)
  952. {
  953. return tsk->pid == 1;
  954. }
  955. extern struct pid *cad_pid;
  956. extern void free_task(struct task_struct *tsk);
  957. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  958. extern void __put_task_struct(struct task_struct *t);
  959. static inline void put_task_struct(struct task_struct *t)
  960. {
  961. if (atomic_dec_and_test(&t->usage))
  962. __put_task_struct(t);
  963. }
  964. /*
  965. * Per process flags
  966. */
  967. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  968. /* Not implemented yet, only for 486*/
  969. #define PF_STARTING 0x00000002 /* being created */
  970. #define PF_EXITING 0x00000004 /* getting shut down */
  971. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  972. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  973. #define PF_DUMPCORE 0x00000200 /* dumped core */
  974. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  975. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  976. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  977. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  978. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  979. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  980. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  981. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  982. #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
  983. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  984. #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
  985. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  986. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  987. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  988. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  989. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  990. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  991. /*
  992. * Only the _current_ task can read/write to tsk->flags, but other
  993. * tasks can access tsk->flags in readonly mode for example
  994. * with tsk_used_math (like during threaded core dumping).
  995. * There is however an exception to this rule during ptrace
  996. * or during fork: the ptracer task is allowed to write to the
  997. * child->flags of its traced child (same goes for fork, the parent
  998. * can write to the child->flags), because we're guaranteed the
  999. * child is not running and in turn not changing child->flags
  1000. * at the same time the parent does it.
  1001. */
  1002. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  1003. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  1004. #define clear_used_math() clear_stopped_child_used_math(current)
  1005. #define set_used_math() set_stopped_child_used_math(current)
  1006. #define conditional_stopped_child_used_math(condition, child) \
  1007. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  1008. #define conditional_used_math(condition) \
  1009. conditional_stopped_child_used_math(condition, current)
  1010. #define copy_to_stopped_child_used_math(child) \
  1011. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  1012. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  1013. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  1014. #define used_math() tsk_used_math(current)
  1015. #ifdef CONFIG_SMP
  1016. extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
  1017. #else
  1018. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  1019. {
  1020. if (!cpu_isset(0, new_mask))
  1021. return -EINVAL;
  1022. return 0;
  1023. }
  1024. #endif
  1025. extern unsigned long long sched_clock(void);
  1026. extern unsigned long long
  1027. current_sched_time(const struct task_struct *current_task);
  1028. /* sched_exec is called by processes performing an exec */
  1029. #ifdef CONFIG_SMP
  1030. extern void sched_exec(void);
  1031. #else
  1032. #define sched_exec() {}
  1033. #endif
  1034. #ifdef CONFIG_HOTPLUG_CPU
  1035. extern void idle_task_exit(void);
  1036. #else
  1037. static inline void idle_task_exit(void) {}
  1038. #endif
  1039. extern void sched_idle_next(void);
  1040. #ifdef CONFIG_RT_MUTEXES
  1041. extern int rt_mutex_getprio(struct task_struct *p);
  1042. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  1043. extern void rt_mutex_adjust_pi(struct task_struct *p);
  1044. #else
  1045. static inline int rt_mutex_getprio(struct task_struct *p)
  1046. {
  1047. return p->normal_prio;
  1048. }
  1049. # define rt_mutex_adjust_pi(p) do { } while (0)
  1050. #endif
  1051. extern void set_user_nice(struct task_struct *p, long nice);
  1052. extern int task_prio(const struct task_struct *p);
  1053. extern int task_nice(const struct task_struct *p);
  1054. extern int can_nice(const struct task_struct *p, const int nice);
  1055. extern int task_curr(const struct task_struct *p);
  1056. extern int idle_cpu(int cpu);
  1057. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  1058. extern struct task_struct *idle_task(int cpu);
  1059. extern struct task_struct *curr_task(int cpu);
  1060. extern void set_curr_task(int cpu, struct task_struct *p);
  1061. void yield(void);
  1062. /*
  1063. * The default (Linux) execution domain.
  1064. */
  1065. extern struct exec_domain default_exec_domain;
  1066. union thread_union {
  1067. struct thread_info thread_info;
  1068. unsigned long stack[THREAD_SIZE/sizeof(long)];
  1069. };
  1070. #ifndef __HAVE_ARCH_KSTACK_END
  1071. static inline int kstack_end(void *addr)
  1072. {
  1073. /* Reliable end of stack detection:
  1074. * Some APM bios versions misalign the stack
  1075. */
  1076. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  1077. }
  1078. #endif
  1079. extern union thread_union init_thread_union;
  1080. extern struct task_struct init_task;
  1081. extern struct mm_struct init_mm;
  1082. #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
  1083. extern struct task_struct *find_task_by_pid_type(int type, int pid);
  1084. extern void __set_special_pids(pid_t session, pid_t pgrp);
  1085. /* per-UID process charging. */
  1086. extern struct user_struct * alloc_uid(uid_t);
  1087. static inline struct user_struct *get_uid(struct user_struct *u)
  1088. {
  1089. atomic_inc(&u->__count);
  1090. return u;
  1091. }
  1092. extern void free_uid(struct user_struct *);
  1093. extern void switch_uid(struct user_struct *);
  1094. #include <asm/current.h>
  1095. extern void do_timer(unsigned long ticks);
  1096. extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
  1097. extern int FASTCALL(wake_up_process(struct task_struct * tsk));
  1098. extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
  1099. unsigned long clone_flags));
  1100. #ifdef CONFIG_SMP
  1101. extern void kick_process(struct task_struct *tsk);
  1102. #else
  1103. static inline void kick_process(struct task_struct *tsk) { }
  1104. #endif
  1105. extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
  1106. extern void FASTCALL(sched_exit(struct task_struct * p));
  1107. extern int in_group_p(gid_t);
  1108. extern int in_egroup_p(gid_t);
  1109. extern void proc_caches_init(void);
  1110. extern void flush_signals(struct task_struct *);
  1111. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1112. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1113. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1114. {
  1115. unsigned long flags;
  1116. int ret;
  1117. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1118. ret = dequeue_signal(tsk, mask, info);
  1119. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1120. return ret;
  1121. }
  1122. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1123. sigset_t *mask);
  1124. extern void unblock_all_signals(void);
  1125. extern void release_task(struct task_struct * p);
  1126. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1127. extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
  1128. extern int force_sigsegv(int, struct task_struct *);
  1129. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1130. extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1131. extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1132. extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
  1133. extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
  1134. extern int kill_pgrp(struct pid *pid, int sig, int priv);
  1135. extern int kill_pid(struct pid *pid, int sig, int priv);
  1136. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1137. extern void do_notify_parent(struct task_struct *, int);
  1138. extern void force_sig(int, struct task_struct *);
  1139. extern void force_sig_specific(int, struct task_struct *);
  1140. extern int send_sig(int, struct task_struct *, int);
  1141. extern void zap_other_threads(struct task_struct *p);
  1142. extern int kill_proc(pid_t, int, int);
  1143. extern struct sigqueue *sigqueue_alloc(void);
  1144. extern void sigqueue_free(struct sigqueue *);
  1145. extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
  1146. extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
  1147. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1148. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1149. static inline int kill_cad_pid(int sig, int priv)
  1150. {
  1151. return kill_pid(cad_pid, sig, priv);
  1152. }
  1153. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1154. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1155. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1156. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1157. static inline int is_si_special(const struct siginfo *info)
  1158. {
  1159. return info <= SEND_SIG_FORCED;
  1160. }
  1161. /* True if we are on the alternate signal stack. */
  1162. static inline int on_sig_stack(unsigned long sp)
  1163. {
  1164. return (sp - current->sas_ss_sp < current->sas_ss_size);
  1165. }
  1166. static inline int sas_ss_flags(unsigned long sp)
  1167. {
  1168. return (current->sas_ss_size == 0 ? SS_DISABLE
  1169. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1170. }
  1171. /*
  1172. * Routines for handling mm_structs
  1173. */
  1174. extern struct mm_struct * mm_alloc(void);
  1175. /* mmdrop drops the mm and the page tables */
  1176. extern void FASTCALL(__mmdrop(struct mm_struct *));
  1177. static inline void mmdrop(struct mm_struct * mm)
  1178. {
  1179. if (atomic_dec_and_test(&mm->mm_count))
  1180. __mmdrop(mm);
  1181. }
  1182. /* mmput gets rid of the mappings and all user-space */
  1183. extern void mmput(struct mm_struct *);
  1184. /* Grab a reference to a task's mm, if it is not already going away */
  1185. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1186. /* Remove the current tasks stale references to the old mm_struct */
  1187. extern void mm_release(struct task_struct *, struct mm_struct *);
  1188. extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
  1189. extern void flush_thread(void);
  1190. extern void exit_thread(void);
  1191. extern void exit_files(struct task_struct *);
  1192. extern void __cleanup_signal(struct signal_struct *);
  1193. extern void __cleanup_sighand(struct sighand_struct *);
  1194. extern void exit_itimers(struct signal_struct *);
  1195. extern NORET_TYPE void do_group_exit(int);
  1196. extern void daemonize(const char *, ...);
  1197. extern int allow_signal(int);
  1198. extern int disallow_signal(int);
  1199. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1200. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1201. struct task_struct *fork_idle(int);
  1202. extern void set_task_comm(struct task_struct *tsk, char *from);
  1203. extern void get_task_comm(char *to, struct task_struct *tsk);
  1204. #ifdef CONFIG_SMP
  1205. extern void wait_task_inactive(struct task_struct * p);
  1206. #else
  1207. #define wait_task_inactive(p) do { } while (0)
  1208. #endif
  1209. #define remove_parent(p) list_del_init(&(p)->sibling)
  1210. #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
  1211. #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
  1212. #define for_each_process(p) \
  1213. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1214. /*
  1215. * Careful: do_each_thread/while_each_thread is a double loop so
  1216. * 'break' will not work as expected - use goto instead.
  1217. */
  1218. #define do_each_thread(g, t) \
  1219. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1220. #define while_each_thread(g, t) \
  1221. while ((t = next_thread(t)) != g)
  1222. /* de_thread depends on thread_group_leader not being a pid based check */
  1223. #define thread_group_leader(p) (p == p->group_leader)
  1224. /* Do to the insanities of de_thread it is possible for a process
  1225. * to have the pid of the thread group leader without actually being
  1226. * the thread group leader. For iteration through the pids in proc
  1227. * all we care about is that we have a task with the appropriate
  1228. * pid, we don't actually care if we have the right task.
  1229. */
  1230. static inline int has_group_leader_pid(struct task_struct *p)
  1231. {
  1232. return p->pid == p->tgid;
  1233. }
  1234. static inline struct task_struct *next_thread(const struct task_struct *p)
  1235. {
  1236. return list_entry(rcu_dereference(p->thread_group.next),
  1237. struct task_struct, thread_group);
  1238. }
  1239. static inline int thread_group_empty(struct task_struct *p)
  1240. {
  1241. return list_empty(&p->thread_group);
  1242. }
  1243. #define delay_group_leader(p) \
  1244. (thread_group_leader(p) && !thread_group_empty(p))
  1245. /*
  1246. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1247. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1248. * pins the final release of task.io_context. Also protects ->cpuset.
  1249. *
  1250. * Nests both inside and outside of read_lock(&tasklist_lock).
  1251. * It must not be nested with write_lock_irq(&tasklist_lock),
  1252. * neither inside nor outside.
  1253. */
  1254. static inline void task_lock(struct task_struct *p)
  1255. {
  1256. spin_lock(&p->alloc_lock);
  1257. }
  1258. static inline void task_unlock(struct task_struct *p)
  1259. {
  1260. spin_unlock(&p->alloc_lock);
  1261. }
  1262. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1263. unsigned long *flags);
  1264. static inline void unlock_task_sighand(struct task_struct *tsk,
  1265. unsigned long *flags)
  1266. {
  1267. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1268. }
  1269. #ifndef __HAVE_THREAD_FUNCTIONS
  1270. #define task_thread_info(task) (task)->thread_info
  1271. #define task_stack_page(task) ((void*)((task)->thread_info))
  1272. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1273. {
  1274. *task_thread_info(p) = *task_thread_info(org);
  1275. task_thread_info(p)->task = p;
  1276. }
  1277. static inline unsigned long *end_of_stack(struct task_struct *p)
  1278. {
  1279. return (unsigned long *)(p->thread_info + 1);
  1280. }
  1281. #endif
  1282. /* set thread flags in other task's structures
  1283. * - see asm/thread_info.h for TIF_xxxx flags available
  1284. */
  1285. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1286. {
  1287. set_ti_thread_flag(task_thread_info(tsk), flag);
  1288. }
  1289. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1290. {
  1291. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1292. }
  1293. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1294. {
  1295. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1296. }
  1297. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1298. {
  1299. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1300. }
  1301. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1302. {
  1303. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1304. }
  1305. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1306. {
  1307. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1308. }
  1309. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1310. {
  1311. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1312. }
  1313. static inline int signal_pending(struct task_struct *p)
  1314. {
  1315. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1316. }
  1317. static inline int need_resched(void)
  1318. {
  1319. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  1320. }
  1321. /*
  1322. * cond_resched() and cond_resched_lock(): latency reduction via
  1323. * explicit rescheduling in places that are safe. The return
  1324. * value indicates whether a reschedule was done in fact.
  1325. * cond_resched_lock() will drop the spinlock before scheduling,
  1326. * cond_resched_softirq() will enable bhs before scheduling.
  1327. */
  1328. extern int cond_resched(void);
  1329. extern int cond_resched_lock(spinlock_t * lock);
  1330. extern int cond_resched_softirq(void);
  1331. /*
  1332. * Does a critical section need to be broken due to another
  1333. * task waiting?:
  1334. */
  1335. #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
  1336. # define need_lockbreak(lock) ((lock)->break_lock)
  1337. #else
  1338. # define need_lockbreak(lock) 0
  1339. #endif
  1340. /*
  1341. * Does a critical section need to be broken due to another
  1342. * task waiting or preemption being signalled:
  1343. */
  1344. static inline int lock_need_resched(spinlock_t *lock)
  1345. {
  1346. if (need_lockbreak(lock) || need_resched())
  1347. return 1;
  1348. return 0;
  1349. }
  1350. /* Reevaluate whether the task has signals pending delivery.
  1351. This is required every time the blocked sigset_t changes.
  1352. callers must hold sighand->siglock. */
  1353. extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
  1354. extern void recalc_sigpending(void);
  1355. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  1356. /*
  1357. * Wrappers for p->thread_info->cpu access. No-op on UP.
  1358. */
  1359. #ifdef CONFIG_SMP
  1360. static inline unsigned int task_cpu(const struct task_struct *p)
  1361. {
  1362. return task_thread_info(p)->cpu;
  1363. }
  1364. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1365. {
  1366. task_thread_info(p)->cpu = cpu;
  1367. }
  1368. #else
  1369. static inline unsigned int task_cpu(const struct task_struct *p)
  1370. {
  1371. return 0;
  1372. }
  1373. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1374. {
  1375. }
  1376. #endif /* CONFIG_SMP */
  1377. #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
  1378. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  1379. #else
  1380. static inline void arch_pick_mmap_layout(struct mm_struct *mm)
  1381. {
  1382. mm->mmap_base = TASK_UNMAPPED_BASE;
  1383. mm->get_unmapped_area = arch_get_unmapped_area;
  1384. mm->unmap_area = arch_unmap_area;
  1385. }
  1386. #endif
  1387. extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
  1388. extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
  1389. #include <linux/sysdev.h>
  1390. extern int sched_mc_power_savings, sched_smt_power_savings;
  1391. extern struct sysdev_attribute attr_sched_mc_power_savings, attr_sched_smt_power_savings;
  1392. extern int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls);
  1393. extern void normalize_rt_tasks(void);
  1394. #ifdef CONFIG_TASK_XACCT
  1395. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1396. {
  1397. tsk->rchar += amt;
  1398. }
  1399. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1400. {
  1401. tsk->wchar += amt;
  1402. }
  1403. static inline void inc_syscr(struct task_struct *tsk)
  1404. {
  1405. tsk->syscr++;
  1406. }
  1407. static inline void inc_syscw(struct task_struct *tsk)
  1408. {
  1409. tsk->syscw++;
  1410. }
  1411. #else
  1412. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1413. {
  1414. }
  1415. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1416. {
  1417. }
  1418. static inline void inc_syscr(struct task_struct *tsk)
  1419. {
  1420. }
  1421. static inline void inc_syscw(struct task_struct *tsk)
  1422. {
  1423. }
  1424. #endif
  1425. #endif /* __KERNEL__ */
  1426. #endif