sched.h 52 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708
  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. #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
  171. extern int select_nohz_load_balancer(int cpu);
  172. #else
  173. static inline int select_nohz_load_balancer(int cpu)
  174. {
  175. return 0;
  176. }
  177. #endif
  178. /*
  179. * Only dump TASK_* tasks. (0 for all tasks)
  180. */
  181. extern void show_state_filter(unsigned long state_filter);
  182. static inline void show_state(void)
  183. {
  184. show_state_filter(0);
  185. }
  186. extern void show_regs(struct pt_regs *);
  187. /*
  188. * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
  189. * task), SP is the stack pointer of the first frame that should be shown in the back
  190. * trace (or NULL if the entire call-chain of the task should be shown).
  191. */
  192. extern void show_stack(struct task_struct *task, unsigned long *sp);
  193. void io_schedule(void);
  194. long io_schedule_timeout(long timeout);
  195. extern void cpu_init (void);
  196. extern void trap_init(void);
  197. extern void update_process_times(int user);
  198. extern void scheduler_tick(void);
  199. #ifdef CONFIG_DETECT_SOFTLOCKUP
  200. extern void softlockup_tick(void);
  201. extern void spawn_softlockup_task(void);
  202. extern void touch_softlockup_watchdog(void);
  203. extern void touch_all_softlockup_watchdogs(void);
  204. #else
  205. static inline void softlockup_tick(void)
  206. {
  207. }
  208. static inline void spawn_softlockup_task(void)
  209. {
  210. }
  211. static inline void touch_softlockup_watchdog(void)
  212. {
  213. }
  214. static inline void touch_all_softlockup_watchdogs(void)
  215. {
  216. }
  217. #endif
  218. /* Attach to any functions which should be ignored in wchan output. */
  219. #define __sched __attribute__((__section__(".sched.text")))
  220. /* Is this address in the __sched functions? */
  221. extern int in_sched_functions(unsigned long addr);
  222. #define MAX_SCHEDULE_TIMEOUT LONG_MAX
  223. extern signed long FASTCALL(schedule_timeout(signed long timeout));
  224. extern signed long schedule_timeout_interruptible(signed long timeout);
  225. extern signed long schedule_timeout_uninterruptible(signed long timeout);
  226. asmlinkage void schedule(void);
  227. struct nsproxy;
  228. /* Maximum number of active map areas.. This is a random (large) number */
  229. #define DEFAULT_MAX_MAP_COUNT 65536
  230. extern int sysctl_max_map_count;
  231. #include <linux/aio.h>
  232. extern unsigned long
  233. arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
  234. unsigned long, unsigned long);
  235. extern unsigned long
  236. arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
  237. unsigned long len, unsigned long pgoff,
  238. unsigned long flags);
  239. extern void arch_unmap_area(struct mm_struct *, unsigned long);
  240. extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
  241. #if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
  242. /*
  243. * The mm counters are not protected by its page_table_lock,
  244. * so must be incremented atomically.
  245. */
  246. #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
  247. #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
  248. #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
  249. #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
  250. #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
  251. typedef atomic_long_t mm_counter_t;
  252. #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  253. /*
  254. * The mm counters are protected by its page_table_lock,
  255. * so can be incremented directly.
  256. */
  257. #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
  258. #define get_mm_counter(mm, member) ((mm)->_##member)
  259. #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
  260. #define inc_mm_counter(mm, member) (mm)->_##member++
  261. #define dec_mm_counter(mm, member) (mm)->_##member--
  262. typedef unsigned long mm_counter_t;
  263. #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
  264. #define get_mm_rss(mm) \
  265. (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
  266. #define update_hiwater_rss(mm) do { \
  267. unsigned long _rss = get_mm_rss(mm); \
  268. if ((mm)->hiwater_rss < _rss) \
  269. (mm)->hiwater_rss = _rss; \
  270. } while (0)
  271. #define update_hiwater_vm(mm) do { \
  272. if ((mm)->hiwater_vm < (mm)->total_vm) \
  273. (mm)->hiwater_vm = (mm)->total_vm; \
  274. } while (0)
  275. struct mm_struct {
  276. struct vm_area_struct * mmap; /* list of VMAs */
  277. struct rb_root mm_rb;
  278. struct vm_area_struct * mmap_cache; /* last find_vma result */
  279. unsigned long (*get_unmapped_area) (struct file *filp,
  280. unsigned long addr, unsigned long len,
  281. unsigned long pgoff, unsigned long flags);
  282. void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
  283. unsigned long mmap_base; /* base of mmap area */
  284. unsigned long task_size; /* size of task vm space */
  285. unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
  286. unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
  287. pgd_t * pgd;
  288. atomic_t mm_users; /* How many users with user space? */
  289. atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
  290. int map_count; /* number of VMAs */
  291. struct rw_semaphore mmap_sem;
  292. spinlock_t page_table_lock; /* Protects page tables and some counters */
  293. struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
  294. * together off init_mm.mmlist, and are protected
  295. * by mmlist_lock
  296. */
  297. /* Special counters, in some configurations protected by the
  298. * page_table_lock, in other configurations by being atomic.
  299. */
  300. mm_counter_t _file_rss;
  301. mm_counter_t _anon_rss;
  302. unsigned long hiwater_rss; /* High-watermark of RSS usage */
  303. unsigned long hiwater_vm; /* High-water virtual memory usage */
  304. unsigned long total_vm, locked_vm, shared_vm, exec_vm;
  305. unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
  306. unsigned long start_code, end_code, start_data, end_data;
  307. unsigned long start_brk, brk, start_stack;
  308. unsigned long arg_start, arg_end, env_start, env_end;
  309. unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
  310. cpumask_t cpu_vm_mask;
  311. /* Architecture-specific MM context */
  312. mm_context_t context;
  313. /* Swap token stuff */
  314. /*
  315. * Last value of global fault stamp as seen by this process.
  316. * In other words, this value gives an indication of how long
  317. * it has been since this task got the token.
  318. * Look at mm/thrash.c
  319. */
  320. unsigned int faultstamp;
  321. unsigned int token_priority;
  322. unsigned int last_interval;
  323. unsigned char dumpable:2;
  324. /* coredumping support */
  325. int core_waiters;
  326. struct completion *core_startup_done, core_done;
  327. /* aio bits */
  328. rwlock_t ioctx_list_lock;
  329. struct kioctx *ioctx_list;
  330. };
  331. struct sighand_struct {
  332. atomic_t count;
  333. struct k_sigaction action[_NSIG];
  334. spinlock_t siglock;
  335. };
  336. struct pacct_struct {
  337. int ac_flag;
  338. long ac_exitcode;
  339. unsigned long ac_mem;
  340. cputime_t ac_utime, ac_stime;
  341. unsigned long ac_minflt, ac_majflt;
  342. };
  343. /*
  344. * NOTE! "signal_struct" does not have it's own
  345. * locking, because a shared signal_struct always
  346. * implies a shared sighand_struct, so locking
  347. * sighand_struct is always a proper superset of
  348. * the locking of signal_struct.
  349. */
  350. struct signal_struct {
  351. atomic_t count;
  352. atomic_t live;
  353. wait_queue_head_t wait_chldexit; /* for wait4() */
  354. /* current thread group signal load-balancing target: */
  355. struct task_struct *curr_target;
  356. /* shared signal handling: */
  357. struct sigpending shared_pending;
  358. /* thread group exit support */
  359. int group_exit_code;
  360. /* overloaded:
  361. * - notify group_exit_task when ->count is equal to notify_count
  362. * - everyone except group_exit_task is stopped during signal delivery
  363. * of fatal signals, group_exit_task processes the signal.
  364. */
  365. struct task_struct *group_exit_task;
  366. int notify_count;
  367. /* thread group stop support, overloads group_exit_code too */
  368. int group_stop_count;
  369. unsigned int flags; /* see SIGNAL_* flags below */
  370. /* POSIX.1b Interval Timers */
  371. struct list_head posix_timers;
  372. /* ITIMER_REAL timer for the process */
  373. struct hrtimer real_timer;
  374. struct task_struct *tsk;
  375. ktime_t it_real_incr;
  376. /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
  377. cputime_t it_prof_expires, it_virt_expires;
  378. cputime_t it_prof_incr, it_virt_incr;
  379. /* job control IDs */
  380. pid_t pgrp;
  381. struct pid *tty_old_pgrp;
  382. union {
  383. pid_t session __deprecated;
  384. pid_t __session;
  385. };
  386. /* boolean value for session group leader */
  387. int leader;
  388. struct tty_struct *tty; /* NULL if no tty */
  389. /*
  390. * Cumulative resource counters for dead threads in the group,
  391. * and for reaped dead child processes forked by this group.
  392. * Live threads maintain their own counters and add to these
  393. * in __exit_signal, except for the group leader.
  394. */
  395. cputime_t utime, stime, cutime, cstime;
  396. unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
  397. unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
  398. unsigned long inblock, oublock, cinblock, coublock;
  399. /*
  400. * Cumulative ns of scheduled CPU time for dead threads in the
  401. * group, not including a zombie group leader. (This only differs
  402. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  403. * other than jiffies.)
  404. */
  405. unsigned long long sched_time;
  406. /*
  407. * We don't bother to synchronize most readers of this at all,
  408. * because there is no reader checking a limit that actually needs
  409. * to get both rlim_cur and rlim_max atomically, and either one
  410. * alone is a single word that can safely be read normally.
  411. * getrlimit/setrlimit use task_lock(current->group_leader) to
  412. * protect this instead of the siglock, because they really
  413. * have no need to disable irqs.
  414. */
  415. struct rlimit rlim[RLIM_NLIMITS];
  416. struct list_head cpu_timers[3];
  417. /* keep the process-shared keyrings here so that they do the right
  418. * thing in threads created with CLONE_THREAD */
  419. #ifdef CONFIG_KEYS
  420. struct key *session_keyring; /* keyring inherited over fork */
  421. struct key *process_keyring; /* keyring private to this process */
  422. #endif
  423. #ifdef CONFIG_BSD_PROCESS_ACCT
  424. struct pacct_struct pacct; /* per-process accounting information */
  425. #endif
  426. #ifdef CONFIG_TASKSTATS
  427. struct taskstats *stats;
  428. #endif
  429. };
  430. /* Context switch must be unlocked if interrupts are to be enabled */
  431. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  432. # define __ARCH_WANT_UNLOCKED_CTXSW
  433. #endif
  434. /*
  435. * Bits in flags field of signal_struct.
  436. */
  437. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  438. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  439. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  440. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  441. /*
  442. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  443. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  444. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  445. * values are inverted: lower p->prio value means higher priority.
  446. *
  447. * The MAX_USER_RT_PRIO value allows the actual maximum
  448. * RT priority to be separate from the value exported to
  449. * user-space. This allows kernel threads to set their
  450. * priority to a value higher than any user task. Note:
  451. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  452. */
  453. #define MAX_USER_RT_PRIO 100
  454. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  455. #define MAX_PRIO (MAX_RT_PRIO + 40)
  456. #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
  457. #define rt_task(p) rt_prio((p)->prio)
  458. #define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
  459. #define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
  460. #define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
  461. /*
  462. * Some day this will be a full-fledged user tracking system..
  463. */
  464. struct user_struct {
  465. atomic_t __count; /* reference count */
  466. atomic_t processes; /* How many processes does this user have? */
  467. atomic_t files; /* How many open files does this user have? */
  468. atomic_t sigpending; /* How many pending signals does this user have? */
  469. #ifdef CONFIG_INOTIFY_USER
  470. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  471. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  472. #endif
  473. /* protected by mq_lock */
  474. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  475. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  476. #ifdef CONFIG_KEYS
  477. struct key *uid_keyring; /* UID specific keyring */
  478. struct key *session_keyring; /* UID's default session keyring */
  479. #endif
  480. /* Hash table maintenance information */
  481. struct list_head uidhash_list;
  482. uid_t uid;
  483. };
  484. extern struct user_struct *find_user(uid_t);
  485. extern struct user_struct root_user;
  486. #define INIT_USER (&root_user)
  487. struct backing_dev_info;
  488. struct reclaim_state;
  489. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  490. struct sched_info {
  491. /* cumulative counters */
  492. unsigned long cpu_time, /* time spent on the cpu */
  493. run_delay, /* time spent waiting on a runqueue */
  494. pcnt; /* # of timeslices run on this cpu */
  495. /* timestamps */
  496. unsigned long last_arrival, /* when we last ran on a cpu */
  497. last_queued; /* when we were last queued to run */
  498. };
  499. #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
  500. #ifdef CONFIG_SCHEDSTATS
  501. extern const struct file_operations proc_schedstat_operations;
  502. #endif /* CONFIG_SCHEDSTATS */
  503. #ifdef CONFIG_TASK_DELAY_ACCT
  504. struct task_delay_info {
  505. spinlock_t lock;
  506. unsigned int flags; /* Private per-task flags */
  507. /* For each stat XXX, add following, aligned appropriately
  508. *
  509. * struct timespec XXX_start, XXX_end;
  510. * u64 XXX_delay;
  511. * u32 XXX_count;
  512. *
  513. * Atomicity of updates to XXX_delay, XXX_count protected by
  514. * single lock above (split into XXX_lock if contention is an issue).
  515. */
  516. /*
  517. * XXX_count is incremented on every XXX operation, the delay
  518. * associated with the operation is added to XXX_delay.
  519. * XXX_delay contains the accumulated delay time in nanoseconds.
  520. */
  521. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  522. u64 blkio_delay; /* wait for sync block io completion */
  523. u64 swapin_delay; /* wait for swapin block io completion */
  524. u32 blkio_count; /* total count of the number of sync block */
  525. /* io operations performed */
  526. u32 swapin_count; /* total count of the number of swapin block */
  527. /* io operations performed */
  528. };
  529. #endif /* CONFIG_TASK_DELAY_ACCT */
  530. static inline int sched_info_on(void)
  531. {
  532. #ifdef CONFIG_SCHEDSTATS
  533. return 1;
  534. #elif defined(CONFIG_TASK_DELAY_ACCT)
  535. extern int delayacct_on;
  536. return delayacct_on;
  537. #else
  538. return 0;
  539. #endif
  540. }
  541. enum idle_type
  542. {
  543. SCHED_IDLE,
  544. NOT_IDLE,
  545. NEWLY_IDLE,
  546. MAX_IDLE_TYPES
  547. };
  548. /*
  549. * sched-domains (multiprocessor balancing) declarations:
  550. */
  551. #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
  552. #ifdef CONFIG_SMP
  553. #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
  554. #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
  555. #define SD_BALANCE_EXEC 4 /* Balance on exec */
  556. #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
  557. #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
  558. #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
  559. #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
  560. #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
  561. #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
  562. #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
  563. #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
  564. #define BALANCE_FOR_MC_POWER \
  565. (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
  566. #define BALANCE_FOR_PKG_POWER \
  567. ((sched_mc_power_savings || sched_smt_power_savings) ? \
  568. SD_POWERSAVINGS_BALANCE : 0)
  569. #define test_sd_parent(sd, flag) ((sd->parent && \
  570. (sd->parent->flags & flag)) ? 1 : 0)
  571. struct sched_group {
  572. struct sched_group *next; /* Must be a circular list */
  573. cpumask_t cpumask;
  574. /*
  575. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  576. * single CPU. This is read only (except for setup, hotplug CPU).
  577. * Note : Never change cpu_power without recompute its reciprocal
  578. */
  579. unsigned int __cpu_power;
  580. /*
  581. * reciprocal value of cpu_power to avoid expensive divides
  582. * (see include/linux/reciprocal_div.h)
  583. */
  584. u32 reciprocal_cpu_power;
  585. };
  586. struct sched_domain {
  587. /* These fields must be setup */
  588. struct sched_domain *parent; /* top domain must be null terminated */
  589. struct sched_domain *child; /* bottom domain must be null terminated */
  590. struct sched_group *groups; /* the balancing groups of the domain */
  591. cpumask_t span; /* span of all CPUs in this domain */
  592. unsigned long min_interval; /* Minimum balance interval ms */
  593. unsigned long max_interval; /* Maximum balance interval ms */
  594. unsigned int busy_factor; /* less balancing by factor if busy */
  595. unsigned int imbalance_pct; /* No balance until over watermark */
  596. unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
  597. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  598. unsigned int busy_idx;
  599. unsigned int idle_idx;
  600. unsigned int newidle_idx;
  601. unsigned int wake_idx;
  602. unsigned int forkexec_idx;
  603. int flags; /* See SD_* */
  604. /* Runtime fields. */
  605. unsigned long last_balance; /* init to jiffies. units in jiffies */
  606. unsigned int balance_interval; /* initialise to 1. units in ms. */
  607. unsigned int nr_balance_failed; /* initialise to 0 */
  608. #ifdef CONFIG_SCHEDSTATS
  609. /* load_balance() stats */
  610. unsigned long lb_cnt[MAX_IDLE_TYPES];
  611. unsigned long lb_failed[MAX_IDLE_TYPES];
  612. unsigned long lb_balanced[MAX_IDLE_TYPES];
  613. unsigned long lb_imbalance[MAX_IDLE_TYPES];
  614. unsigned long lb_gained[MAX_IDLE_TYPES];
  615. unsigned long lb_hot_gained[MAX_IDLE_TYPES];
  616. unsigned long lb_nobusyg[MAX_IDLE_TYPES];
  617. unsigned long lb_nobusyq[MAX_IDLE_TYPES];
  618. /* Active load balancing */
  619. unsigned long alb_cnt;
  620. unsigned long alb_failed;
  621. unsigned long alb_pushed;
  622. /* SD_BALANCE_EXEC stats */
  623. unsigned long sbe_cnt;
  624. unsigned long sbe_balanced;
  625. unsigned long sbe_pushed;
  626. /* SD_BALANCE_FORK stats */
  627. unsigned long sbf_cnt;
  628. unsigned long sbf_balanced;
  629. unsigned long sbf_pushed;
  630. /* try_to_wake_up() stats */
  631. unsigned long ttwu_wake_remote;
  632. unsigned long ttwu_move_affine;
  633. unsigned long ttwu_move_balance;
  634. #endif
  635. };
  636. extern int partition_sched_domains(cpumask_t *partition1,
  637. cpumask_t *partition2);
  638. /*
  639. * Maximum cache size the migration-costs auto-tuning code will
  640. * search from:
  641. */
  642. extern unsigned int max_cache_size;
  643. #endif /* CONFIG_SMP */
  644. struct io_context; /* See blkdev.h */
  645. struct cpuset;
  646. #define NGROUPS_SMALL 32
  647. #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
  648. struct group_info {
  649. int ngroups;
  650. atomic_t usage;
  651. gid_t small_block[NGROUPS_SMALL];
  652. int nblocks;
  653. gid_t *blocks[0];
  654. };
  655. /*
  656. * get_group_info() must be called with the owning task locked (via task_lock())
  657. * when task != current. The reason being that the vast majority of callers are
  658. * looking at current->group_info, which can not be changed except by the
  659. * current task. Changing current->group_info requires the task lock, too.
  660. */
  661. #define get_group_info(group_info) do { \
  662. atomic_inc(&(group_info)->usage); \
  663. } while (0)
  664. #define put_group_info(group_info) do { \
  665. if (atomic_dec_and_test(&(group_info)->usage)) \
  666. groups_free(group_info); \
  667. } while (0)
  668. extern struct group_info *groups_alloc(int gidsetsize);
  669. extern void groups_free(struct group_info *group_info);
  670. extern int set_current_groups(struct group_info *group_info);
  671. extern int groups_search(struct group_info *group_info, gid_t grp);
  672. /* access the groups "array" with this macro */
  673. #define GROUP_AT(gi, i) \
  674. ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
  675. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  676. extern void prefetch_stack(struct task_struct *t);
  677. #else
  678. static inline void prefetch_stack(struct task_struct *t) { }
  679. #endif
  680. struct audit_context; /* See audit.c */
  681. struct mempolicy;
  682. struct pipe_inode_info;
  683. struct uts_namespace;
  684. enum sleep_type {
  685. SLEEP_NORMAL,
  686. SLEEP_NONINTERACTIVE,
  687. SLEEP_INTERACTIVE,
  688. SLEEP_INTERRUPTED,
  689. };
  690. struct prio_array;
  691. struct task_struct {
  692. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  693. void *stack;
  694. atomic_t usage;
  695. unsigned int flags; /* per process flags, defined below */
  696. unsigned int ptrace;
  697. int lock_depth; /* BKL lock depth */
  698. #ifdef CONFIG_SMP
  699. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  700. int oncpu;
  701. #endif
  702. #endif
  703. int load_weight; /* for niceness load balancing purposes */
  704. int prio, static_prio, normal_prio;
  705. struct list_head run_list;
  706. struct prio_array *array;
  707. unsigned short ioprio;
  708. #ifdef CONFIG_BLK_DEV_IO_TRACE
  709. unsigned int btrace_seq;
  710. #endif
  711. unsigned long sleep_avg;
  712. unsigned long long timestamp, last_ran;
  713. unsigned long long sched_time; /* sched_clock time spent running */
  714. enum sleep_type sleep_type;
  715. unsigned int policy;
  716. cpumask_t cpus_allowed;
  717. unsigned int time_slice, first_time_slice;
  718. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  719. struct sched_info sched_info;
  720. #endif
  721. struct list_head tasks;
  722. /*
  723. * ptrace_list/ptrace_children forms the list of my children
  724. * that were stolen by a ptracer.
  725. */
  726. struct list_head ptrace_children;
  727. struct list_head ptrace_list;
  728. struct mm_struct *mm, *active_mm;
  729. /* task state */
  730. struct linux_binfmt *binfmt;
  731. int exit_state;
  732. int exit_code, exit_signal;
  733. int pdeath_signal; /* The signal sent when the parent dies */
  734. /* ??? */
  735. unsigned int personality;
  736. unsigned did_exec:1;
  737. pid_t pid;
  738. pid_t tgid;
  739. #ifdef CONFIG_CC_STACKPROTECTOR
  740. /* Canary value for the -fstack-protector gcc feature */
  741. unsigned long stack_canary;
  742. #endif
  743. /*
  744. * pointers to (original) parent process, youngest child, younger sibling,
  745. * older sibling, respectively. (p->father can be replaced with
  746. * p->parent->pid)
  747. */
  748. struct task_struct *real_parent; /* real parent process (when being debugged) */
  749. struct task_struct *parent; /* parent process */
  750. /*
  751. * children/sibling forms the list of my children plus the
  752. * tasks I'm ptracing.
  753. */
  754. struct list_head children; /* list of my children */
  755. struct list_head sibling; /* linkage in my parent's children list */
  756. struct task_struct *group_leader; /* threadgroup leader */
  757. /* PID/PID hash table linkage. */
  758. struct pid_link pids[PIDTYPE_MAX];
  759. struct list_head thread_group;
  760. struct completion *vfork_done; /* for vfork() */
  761. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  762. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  763. unsigned int rt_priority;
  764. cputime_t utime, stime;
  765. unsigned long nvcsw, nivcsw; /* context switch counts */
  766. struct timespec start_time;
  767. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  768. unsigned long min_flt, maj_flt;
  769. cputime_t it_prof_expires, it_virt_expires;
  770. unsigned long long it_sched_expires;
  771. struct list_head cpu_timers[3];
  772. /* process credentials */
  773. uid_t uid,euid,suid,fsuid;
  774. gid_t gid,egid,sgid,fsgid;
  775. struct group_info *group_info;
  776. kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
  777. unsigned keep_capabilities:1;
  778. struct user_struct *user;
  779. #ifdef CONFIG_KEYS
  780. struct key *request_key_auth; /* assumed request_key authority */
  781. struct key *thread_keyring; /* keyring private to this thread */
  782. unsigned char jit_keyring; /* default keyring to attach requested keys to */
  783. #endif
  784. /*
  785. * fpu_counter contains the number of consecutive context switches
  786. * that the FPU is used. If this is over a threshold, the lazy fpu
  787. * saving becomes unlazy to save the trap. This is an unsigned char
  788. * so that after 256 times the counter wraps and the behavior turns
  789. * lazy again; this to deal with bursty apps that only use FPU for
  790. * a short time
  791. */
  792. unsigned char fpu_counter;
  793. int oomkilladj; /* OOM kill score adjustment (bit shift). */
  794. char comm[TASK_COMM_LEN]; /* executable name excluding path
  795. - access with [gs]et_task_comm (which lock
  796. it with task_lock())
  797. - initialized normally by flush_old_exec */
  798. /* file system info */
  799. int link_count, total_link_count;
  800. #ifdef CONFIG_SYSVIPC
  801. /* ipc stuff */
  802. struct sysv_sem sysvsem;
  803. #endif
  804. /* CPU-specific state of this task */
  805. struct thread_struct thread;
  806. /* filesystem information */
  807. struct fs_struct *fs;
  808. /* open file information */
  809. struct files_struct *files;
  810. /* namespaces */
  811. struct nsproxy *nsproxy;
  812. /* signal handlers */
  813. struct signal_struct *signal;
  814. struct sighand_struct *sighand;
  815. sigset_t blocked, real_blocked;
  816. sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
  817. struct sigpending pending;
  818. unsigned long sas_ss_sp;
  819. size_t sas_ss_size;
  820. int (*notifier)(void *priv);
  821. void *notifier_data;
  822. sigset_t *notifier_mask;
  823. void *security;
  824. struct audit_context *audit_context;
  825. seccomp_t seccomp;
  826. /* Thread group tracking */
  827. u32 parent_exec_id;
  828. u32 self_exec_id;
  829. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
  830. spinlock_t alloc_lock;
  831. /* Protection of the PI data structures: */
  832. spinlock_t pi_lock;
  833. #ifdef CONFIG_RT_MUTEXES
  834. /* PI waiters blocked on a rt_mutex held by this task */
  835. struct plist_head pi_waiters;
  836. /* Deadlock detection and priority inheritance handling */
  837. struct rt_mutex_waiter *pi_blocked_on;
  838. #endif
  839. #ifdef CONFIG_DEBUG_MUTEXES
  840. /* mutex deadlock detection */
  841. struct mutex_waiter *blocked_on;
  842. #endif
  843. #ifdef CONFIG_TRACE_IRQFLAGS
  844. unsigned int irq_events;
  845. int hardirqs_enabled;
  846. unsigned long hardirq_enable_ip;
  847. unsigned int hardirq_enable_event;
  848. unsigned long hardirq_disable_ip;
  849. unsigned int hardirq_disable_event;
  850. int softirqs_enabled;
  851. unsigned long softirq_disable_ip;
  852. unsigned int softirq_disable_event;
  853. unsigned long softirq_enable_ip;
  854. unsigned int softirq_enable_event;
  855. int hardirq_context;
  856. int softirq_context;
  857. #endif
  858. #ifdef CONFIG_LOCKDEP
  859. # define MAX_LOCK_DEPTH 30UL
  860. u64 curr_chain_key;
  861. int lockdep_depth;
  862. struct held_lock held_locks[MAX_LOCK_DEPTH];
  863. unsigned int lockdep_recursion;
  864. #endif
  865. /* journalling filesystem info */
  866. void *journal_info;
  867. /* VM state */
  868. struct reclaim_state *reclaim_state;
  869. struct backing_dev_info *backing_dev_info;
  870. struct io_context *io_context;
  871. unsigned long ptrace_message;
  872. siginfo_t *last_siginfo; /* For ptrace use. */
  873. /*
  874. * current io wait handle: wait queue entry to use for io waits
  875. * If this thread is processing aio, this points at the waitqueue
  876. * inside the currently handled kiocb. It may be NULL (i.e. default
  877. * to a stack based synchronous wait) if its doing sync IO.
  878. */
  879. wait_queue_t *io_wait;
  880. #ifdef CONFIG_TASK_XACCT
  881. /* i/o counters(bytes read/written, #syscalls */
  882. u64 rchar, wchar, syscr, syscw;
  883. #endif
  884. struct task_io_accounting ioac;
  885. #if defined(CONFIG_TASK_XACCT)
  886. u64 acct_rss_mem1; /* accumulated rss usage */
  887. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  888. cputime_t acct_stimexpd;/* stime since last update */
  889. #endif
  890. #ifdef CONFIG_NUMA
  891. struct mempolicy *mempolicy;
  892. short il_next;
  893. #endif
  894. #ifdef CONFIG_CPUSETS
  895. struct cpuset *cpuset;
  896. nodemask_t mems_allowed;
  897. int cpuset_mems_generation;
  898. int cpuset_mem_spread_rotor;
  899. #endif
  900. struct robust_list_head __user *robust_list;
  901. #ifdef CONFIG_COMPAT
  902. struct compat_robust_list_head __user *compat_robust_list;
  903. #endif
  904. struct list_head pi_state_list;
  905. struct futex_pi_state *pi_state_cache;
  906. atomic_t fs_excl; /* holding fs exclusive resources */
  907. struct rcu_head rcu;
  908. /*
  909. * cache last used pipe for splice
  910. */
  911. struct pipe_inode_info *splice_pipe;
  912. #ifdef CONFIG_TASK_DELAY_ACCT
  913. struct task_delay_info *delays;
  914. #endif
  915. #ifdef CONFIG_FAULT_INJECTION
  916. int make_it_fail;
  917. #endif
  918. };
  919. static inline pid_t process_group(struct task_struct *tsk)
  920. {
  921. return tsk->signal->pgrp;
  922. }
  923. static inline pid_t signal_session(struct signal_struct *sig)
  924. {
  925. return sig->__session;
  926. }
  927. static inline pid_t process_session(struct task_struct *tsk)
  928. {
  929. return signal_session(tsk->signal);
  930. }
  931. static inline void set_signal_session(struct signal_struct *sig, pid_t session)
  932. {
  933. sig->__session = session;
  934. }
  935. static inline struct pid *task_pid(struct task_struct *task)
  936. {
  937. return task->pids[PIDTYPE_PID].pid;
  938. }
  939. static inline struct pid *task_tgid(struct task_struct *task)
  940. {
  941. return task->group_leader->pids[PIDTYPE_PID].pid;
  942. }
  943. static inline struct pid *task_pgrp(struct task_struct *task)
  944. {
  945. return task->group_leader->pids[PIDTYPE_PGID].pid;
  946. }
  947. static inline struct pid *task_session(struct task_struct *task)
  948. {
  949. return task->group_leader->pids[PIDTYPE_SID].pid;
  950. }
  951. /**
  952. * pid_alive - check that a task structure is not stale
  953. * @p: Task structure to be checked.
  954. *
  955. * Test if a process is not yet dead (at most zombie state)
  956. * If pid_alive fails, then pointers within the task structure
  957. * can be stale and must not be dereferenced.
  958. */
  959. static inline int pid_alive(struct task_struct *p)
  960. {
  961. return p->pids[PIDTYPE_PID].pid != NULL;
  962. }
  963. /**
  964. * is_init - check if a task structure is init
  965. * @tsk: Task structure to be checked.
  966. *
  967. * Check if a task structure is the first user space task the kernel created.
  968. */
  969. static inline int is_init(struct task_struct *tsk)
  970. {
  971. return tsk->pid == 1;
  972. }
  973. extern struct pid *cad_pid;
  974. extern void free_task(struct task_struct *tsk);
  975. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  976. extern void __put_task_struct(struct task_struct *t);
  977. static inline void put_task_struct(struct task_struct *t)
  978. {
  979. if (atomic_dec_and_test(&t->usage))
  980. __put_task_struct(t);
  981. }
  982. /*
  983. * Per process flags
  984. */
  985. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  986. /* Not implemented yet, only for 486*/
  987. #define PF_STARTING 0x00000002 /* being created */
  988. #define PF_EXITING 0x00000004 /* getting shut down */
  989. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  990. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  991. #define PF_DUMPCORE 0x00000200 /* dumped core */
  992. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  993. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  994. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  995. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  996. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  997. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  998. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  999. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  1000. #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
  1001. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  1002. #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
  1003. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  1004. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  1005. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  1006. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  1007. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  1008. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  1009. /*
  1010. * Only the _current_ task can read/write to tsk->flags, but other
  1011. * tasks can access tsk->flags in readonly mode for example
  1012. * with tsk_used_math (like during threaded core dumping).
  1013. * There is however an exception to this rule during ptrace
  1014. * or during fork: the ptracer task is allowed to write to the
  1015. * child->flags of its traced child (same goes for fork, the parent
  1016. * can write to the child->flags), because we're guaranteed the
  1017. * child is not running and in turn not changing child->flags
  1018. * at the same time the parent does it.
  1019. */
  1020. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  1021. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  1022. #define clear_used_math() clear_stopped_child_used_math(current)
  1023. #define set_used_math() set_stopped_child_used_math(current)
  1024. #define conditional_stopped_child_used_math(condition, child) \
  1025. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  1026. #define conditional_used_math(condition) \
  1027. conditional_stopped_child_used_math(condition, current)
  1028. #define copy_to_stopped_child_used_math(child) \
  1029. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  1030. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  1031. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  1032. #define used_math() tsk_used_math(current)
  1033. #ifdef CONFIG_SMP
  1034. extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
  1035. #else
  1036. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  1037. {
  1038. if (!cpu_isset(0, new_mask))
  1039. return -EINVAL;
  1040. return 0;
  1041. }
  1042. #endif
  1043. extern unsigned long long sched_clock(void);
  1044. extern unsigned long long
  1045. current_sched_time(const struct task_struct *current_task);
  1046. /* sched_exec is called by processes performing an exec */
  1047. #ifdef CONFIG_SMP
  1048. extern void sched_exec(void);
  1049. #else
  1050. #define sched_exec() {}
  1051. #endif
  1052. #ifdef CONFIG_HOTPLUG_CPU
  1053. extern void idle_task_exit(void);
  1054. #else
  1055. static inline void idle_task_exit(void) {}
  1056. #endif
  1057. extern void sched_idle_next(void);
  1058. #ifdef CONFIG_RT_MUTEXES
  1059. extern int rt_mutex_getprio(struct task_struct *p);
  1060. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  1061. extern void rt_mutex_adjust_pi(struct task_struct *p);
  1062. #else
  1063. static inline int rt_mutex_getprio(struct task_struct *p)
  1064. {
  1065. return p->normal_prio;
  1066. }
  1067. # define rt_mutex_adjust_pi(p) do { } while (0)
  1068. #endif
  1069. extern void set_user_nice(struct task_struct *p, long nice);
  1070. extern int task_prio(const struct task_struct *p);
  1071. extern int task_nice(const struct task_struct *p);
  1072. extern int can_nice(const struct task_struct *p, const int nice);
  1073. extern int task_curr(const struct task_struct *p);
  1074. extern int idle_cpu(int cpu);
  1075. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  1076. extern struct task_struct *idle_task(int cpu);
  1077. extern struct task_struct *curr_task(int cpu);
  1078. extern void set_curr_task(int cpu, struct task_struct *p);
  1079. void yield(void);
  1080. /*
  1081. * The default (Linux) execution domain.
  1082. */
  1083. extern struct exec_domain default_exec_domain;
  1084. union thread_union {
  1085. struct thread_info thread_info;
  1086. unsigned long stack[THREAD_SIZE/sizeof(long)];
  1087. };
  1088. #ifndef __HAVE_ARCH_KSTACK_END
  1089. static inline int kstack_end(void *addr)
  1090. {
  1091. /* Reliable end of stack detection:
  1092. * Some APM bios versions misalign the stack
  1093. */
  1094. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  1095. }
  1096. #endif
  1097. extern union thread_union init_thread_union;
  1098. extern struct task_struct init_task;
  1099. extern struct mm_struct init_mm;
  1100. #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
  1101. extern struct task_struct *find_task_by_pid_type(int type, int pid);
  1102. extern void __set_special_pids(pid_t session, pid_t pgrp);
  1103. /* per-UID process charging. */
  1104. extern struct user_struct * alloc_uid(uid_t);
  1105. static inline struct user_struct *get_uid(struct user_struct *u)
  1106. {
  1107. atomic_inc(&u->__count);
  1108. return u;
  1109. }
  1110. extern void free_uid(struct user_struct *);
  1111. extern void switch_uid(struct user_struct *);
  1112. #include <asm/current.h>
  1113. extern void do_timer(unsigned long ticks);
  1114. extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
  1115. extern int FASTCALL(wake_up_process(struct task_struct * tsk));
  1116. extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
  1117. unsigned long clone_flags));
  1118. #ifdef CONFIG_SMP
  1119. extern void kick_process(struct task_struct *tsk);
  1120. #else
  1121. static inline void kick_process(struct task_struct *tsk) { }
  1122. #endif
  1123. extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
  1124. extern void FASTCALL(sched_exit(struct task_struct * p));
  1125. extern int in_group_p(gid_t);
  1126. extern int in_egroup_p(gid_t);
  1127. extern void proc_caches_init(void);
  1128. extern void flush_signals(struct task_struct *);
  1129. extern void ignore_signals(struct task_struct *);
  1130. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1131. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1132. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1133. {
  1134. unsigned long flags;
  1135. int ret;
  1136. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1137. ret = dequeue_signal(tsk, mask, info);
  1138. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1139. return ret;
  1140. }
  1141. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1142. sigset_t *mask);
  1143. extern void unblock_all_signals(void);
  1144. extern void release_task(struct task_struct * p);
  1145. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1146. extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
  1147. extern int force_sigsegv(int, struct task_struct *);
  1148. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1149. extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1150. extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1151. extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
  1152. extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
  1153. extern int kill_pgrp(struct pid *pid, int sig, int priv);
  1154. extern int kill_pid(struct pid *pid, int sig, int priv);
  1155. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1156. extern void do_notify_parent(struct task_struct *, int);
  1157. extern void force_sig(int, struct task_struct *);
  1158. extern void force_sig_specific(int, struct task_struct *);
  1159. extern int send_sig(int, struct task_struct *, int);
  1160. extern void zap_other_threads(struct task_struct *p);
  1161. extern int kill_proc(pid_t, int, int);
  1162. extern struct sigqueue *sigqueue_alloc(void);
  1163. extern void sigqueue_free(struct sigqueue *);
  1164. extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
  1165. extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
  1166. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1167. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1168. static inline int kill_cad_pid(int sig, int priv)
  1169. {
  1170. return kill_pid(cad_pid, sig, priv);
  1171. }
  1172. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1173. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1174. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1175. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1176. static inline int is_si_special(const struct siginfo *info)
  1177. {
  1178. return info <= SEND_SIG_FORCED;
  1179. }
  1180. /* True if we are on the alternate signal stack. */
  1181. static inline int on_sig_stack(unsigned long sp)
  1182. {
  1183. return (sp - current->sas_ss_sp < current->sas_ss_size);
  1184. }
  1185. static inline int sas_ss_flags(unsigned long sp)
  1186. {
  1187. return (current->sas_ss_size == 0 ? SS_DISABLE
  1188. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1189. }
  1190. /*
  1191. * Routines for handling mm_structs
  1192. */
  1193. extern struct mm_struct * mm_alloc(void);
  1194. /* mmdrop drops the mm and the page tables */
  1195. extern void FASTCALL(__mmdrop(struct mm_struct *));
  1196. static inline void mmdrop(struct mm_struct * mm)
  1197. {
  1198. if (atomic_dec_and_test(&mm->mm_count))
  1199. __mmdrop(mm);
  1200. }
  1201. /* mmput gets rid of the mappings and all user-space */
  1202. extern void mmput(struct mm_struct *);
  1203. /* Grab a reference to a task's mm, if it is not already going away */
  1204. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1205. /* Remove the current tasks stale references to the old mm_struct */
  1206. extern void mm_release(struct task_struct *, struct mm_struct *);
  1207. extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
  1208. extern void flush_thread(void);
  1209. extern void exit_thread(void);
  1210. extern void exit_files(struct task_struct *);
  1211. extern void __cleanup_signal(struct signal_struct *);
  1212. extern void __cleanup_sighand(struct sighand_struct *);
  1213. extern void exit_itimers(struct signal_struct *);
  1214. extern NORET_TYPE void do_group_exit(int);
  1215. extern void daemonize(const char *, ...);
  1216. extern int allow_signal(int);
  1217. extern int disallow_signal(int);
  1218. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1219. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1220. struct task_struct *fork_idle(int);
  1221. extern void set_task_comm(struct task_struct *tsk, char *from);
  1222. extern void get_task_comm(char *to, struct task_struct *tsk);
  1223. #ifdef CONFIG_SMP
  1224. extern void wait_task_inactive(struct task_struct * p);
  1225. #else
  1226. #define wait_task_inactive(p) do { } while (0)
  1227. #endif
  1228. #define remove_parent(p) list_del_init(&(p)->sibling)
  1229. #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
  1230. #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
  1231. #define for_each_process(p) \
  1232. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1233. /*
  1234. * Careful: do_each_thread/while_each_thread is a double loop so
  1235. * 'break' will not work as expected - use goto instead.
  1236. */
  1237. #define do_each_thread(g, t) \
  1238. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1239. #define while_each_thread(g, t) \
  1240. while ((t = next_thread(t)) != g)
  1241. /* de_thread depends on thread_group_leader not being a pid based check */
  1242. #define thread_group_leader(p) (p == p->group_leader)
  1243. /* Do to the insanities of de_thread it is possible for a process
  1244. * to have the pid of the thread group leader without actually being
  1245. * the thread group leader. For iteration through the pids in proc
  1246. * all we care about is that we have a task with the appropriate
  1247. * pid, we don't actually care if we have the right task.
  1248. */
  1249. static inline int has_group_leader_pid(struct task_struct *p)
  1250. {
  1251. return p->pid == p->tgid;
  1252. }
  1253. static inline struct task_struct *next_thread(const struct task_struct *p)
  1254. {
  1255. return list_entry(rcu_dereference(p->thread_group.next),
  1256. struct task_struct, thread_group);
  1257. }
  1258. static inline int thread_group_empty(struct task_struct *p)
  1259. {
  1260. return list_empty(&p->thread_group);
  1261. }
  1262. #define delay_group_leader(p) \
  1263. (thread_group_leader(p) && !thread_group_empty(p))
  1264. /*
  1265. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1266. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1267. * pins the final release of task.io_context. Also protects ->cpuset.
  1268. *
  1269. * Nests both inside and outside of read_lock(&tasklist_lock).
  1270. * It must not be nested with write_lock_irq(&tasklist_lock),
  1271. * neither inside nor outside.
  1272. */
  1273. static inline void task_lock(struct task_struct *p)
  1274. {
  1275. spin_lock(&p->alloc_lock);
  1276. }
  1277. static inline void task_unlock(struct task_struct *p)
  1278. {
  1279. spin_unlock(&p->alloc_lock);
  1280. }
  1281. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1282. unsigned long *flags);
  1283. static inline void unlock_task_sighand(struct task_struct *tsk,
  1284. unsigned long *flags)
  1285. {
  1286. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1287. }
  1288. #ifndef __HAVE_THREAD_FUNCTIONS
  1289. #define task_thread_info(task) ((struct thread_info *)(task)->stack)
  1290. #define task_stack_page(task) ((task)->stack)
  1291. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1292. {
  1293. *task_thread_info(p) = *task_thread_info(org);
  1294. task_thread_info(p)->task = p;
  1295. }
  1296. static inline unsigned long *end_of_stack(struct task_struct *p)
  1297. {
  1298. return (unsigned long *)(task_thread_info(p) + 1);
  1299. }
  1300. #endif
  1301. /* set thread flags in other task's structures
  1302. * - see asm/thread_info.h for TIF_xxxx flags available
  1303. */
  1304. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1305. {
  1306. set_ti_thread_flag(task_thread_info(tsk), flag);
  1307. }
  1308. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1309. {
  1310. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1311. }
  1312. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1313. {
  1314. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1315. }
  1316. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1317. {
  1318. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1319. }
  1320. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1321. {
  1322. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1323. }
  1324. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1325. {
  1326. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1327. }
  1328. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1329. {
  1330. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1331. }
  1332. static inline int signal_pending(struct task_struct *p)
  1333. {
  1334. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1335. }
  1336. static inline int need_resched(void)
  1337. {
  1338. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  1339. }
  1340. /*
  1341. * cond_resched() and cond_resched_lock(): latency reduction via
  1342. * explicit rescheduling in places that are safe. The return
  1343. * value indicates whether a reschedule was done in fact.
  1344. * cond_resched_lock() will drop the spinlock before scheduling,
  1345. * cond_resched_softirq() will enable bhs before scheduling.
  1346. */
  1347. extern int cond_resched(void);
  1348. extern int cond_resched_lock(spinlock_t * lock);
  1349. extern int cond_resched_softirq(void);
  1350. /*
  1351. * Does a critical section need to be broken due to another
  1352. * task waiting?:
  1353. */
  1354. #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
  1355. # define need_lockbreak(lock) ((lock)->break_lock)
  1356. #else
  1357. # define need_lockbreak(lock) 0
  1358. #endif
  1359. /*
  1360. * Does a critical section need to be broken due to another
  1361. * task waiting or preemption being signalled:
  1362. */
  1363. static inline int lock_need_resched(spinlock_t *lock)
  1364. {
  1365. if (need_lockbreak(lock) || need_resched())
  1366. return 1;
  1367. return 0;
  1368. }
  1369. /* Reevaluate whether the task has signals pending delivery.
  1370. This is required every time the blocked sigset_t changes.
  1371. callers must hold sighand->siglock. */
  1372. extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
  1373. extern void recalc_sigpending(void);
  1374. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  1375. /*
  1376. * Wrappers for p->thread_info->cpu access. No-op on UP.
  1377. */
  1378. #ifdef CONFIG_SMP
  1379. static inline unsigned int task_cpu(const struct task_struct *p)
  1380. {
  1381. return task_thread_info(p)->cpu;
  1382. }
  1383. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1384. {
  1385. task_thread_info(p)->cpu = cpu;
  1386. }
  1387. #else
  1388. static inline unsigned int task_cpu(const struct task_struct *p)
  1389. {
  1390. return 0;
  1391. }
  1392. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1393. {
  1394. }
  1395. #endif /* CONFIG_SMP */
  1396. #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
  1397. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  1398. #else
  1399. static inline void arch_pick_mmap_layout(struct mm_struct *mm)
  1400. {
  1401. mm->mmap_base = TASK_UNMAPPED_BASE;
  1402. mm->get_unmapped_area = arch_get_unmapped_area;
  1403. mm->unmap_area = arch_unmap_area;
  1404. }
  1405. #endif
  1406. extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
  1407. extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
  1408. extern int sched_mc_power_savings, sched_smt_power_savings;
  1409. extern void normalize_rt_tasks(void);
  1410. #ifdef CONFIG_TASK_XACCT
  1411. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1412. {
  1413. tsk->rchar += amt;
  1414. }
  1415. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1416. {
  1417. tsk->wchar += amt;
  1418. }
  1419. static inline void inc_syscr(struct task_struct *tsk)
  1420. {
  1421. tsk->syscr++;
  1422. }
  1423. static inline void inc_syscw(struct task_struct *tsk)
  1424. {
  1425. tsk->syscw++;
  1426. }
  1427. #else
  1428. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1429. {
  1430. }
  1431. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1432. {
  1433. }
  1434. static inline void inc_syscr(struct task_struct *tsk)
  1435. {
  1436. }
  1437. static inline void inc_syscw(struct task_struct *tsk)
  1438. {
  1439. }
  1440. #endif
  1441. #endif /* __KERNEL__ */
  1442. #endif