sched.h 52 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409141014111412141314141415141614171418141914201421142214231424142514261427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532153315341535153615371538153915401541154215431544154515461547154815491550155115521553155415551556155715581559156015611562156315641565156615671568156915701571157215731574157515761577157815791580158115821583158415851586158715881589159015911592159315941595159615971598159916001601160216031604160516061607160816091610161116121613161416151616161716181619162016211622162316241625162616271628162916301631163216331634163516361637163816391640164116421643164416451646164716481649165016511652165316541655165616571658165916601661166216631664166516661667166816691670167116721673167416751676167716781679168016811682168316841685168616871688168916901691169216931694169516961697169816991700170117021703170417051706
  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. /*
  399. * Cumulative ns of scheduled CPU time for dead threads in the
  400. * group, not including a zombie group leader. (This only differs
  401. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  402. * other than jiffies.)
  403. */
  404. unsigned long long sched_time;
  405. /*
  406. * We don't bother to synchronize most readers of this at all,
  407. * because there is no reader checking a limit that actually needs
  408. * to get both rlim_cur and rlim_max atomically, and either one
  409. * alone is a single word that can safely be read normally.
  410. * getrlimit/setrlimit use task_lock(current->group_leader) to
  411. * protect this instead of the siglock, because they really
  412. * have no need to disable irqs.
  413. */
  414. struct rlimit rlim[RLIM_NLIMITS];
  415. struct list_head cpu_timers[3];
  416. /* keep the process-shared keyrings here so that they do the right
  417. * thing in threads created with CLONE_THREAD */
  418. #ifdef CONFIG_KEYS
  419. struct key *session_keyring; /* keyring inherited over fork */
  420. struct key *process_keyring; /* keyring private to this process */
  421. #endif
  422. #ifdef CONFIG_BSD_PROCESS_ACCT
  423. struct pacct_struct pacct; /* per-process accounting information */
  424. #endif
  425. #ifdef CONFIG_TASKSTATS
  426. struct taskstats *stats;
  427. #endif
  428. };
  429. /* Context switch must be unlocked if interrupts are to be enabled */
  430. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  431. # define __ARCH_WANT_UNLOCKED_CTXSW
  432. #endif
  433. /*
  434. * Bits in flags field of signal_struct.
  435. */
  436. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  437. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  438. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  439. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  440. /*
  441. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  442. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  443. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  444. * values are inverted: lower p->prio value means higher priority.
  445. *
  446. * The MAX_USER_RT_PRIO value allows the actual maximum
  447. * RT priority to be separate from the value exported to
  448. * user-space. This allows kernel threads to set their
  449. * priority to a value higher than any user task. Note:
  450. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  451. */
  452. #define MAX_USER_RT_PRIO 100
  453. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  454. #define MAX_PRIO (MAX_RT_PRIO + 40)
  455. #define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO)
  456. #define rt_task(p) rt_prio((p)->prio)
  457. #define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
  458. #define is_rt_policy(p) ((p) != SCHED_NORMAL && (p) != SCHED_BATCH)
  459. #define has_rt_policy(p) unlikely(is_rt_policy((p)->policy))
  460. /*
  461. * Some day this will be a full-fledged user tracking system..
  462. */
  463. struct user_struct {
  464. atomic_t __count; /* reference count */
  465. atomic_t processes; /* How many processes does this user have? */
  466. atomic_t files; /* How many open files does this user have? */
  467. atomic_t sigpending; /* How many pending signals does this user have? */
  468. #ifdef CONFIG_INOTIFY_USER
  469. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  470. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  471. #endif
  472. /* protected by mq_lock */
  473. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  474. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  475. #ifdef CONFIG_KEYS
  476. struct key *uid_keyring; /* UID specific keyring */
  477. struct key *session_keyring; /* UID's default session keyring */
  478. #endif
  479. /* Hash table maintenance information */
  480. struct list_head uidhash_list;
  481. uid_t uid;
  482. };
  483. extern struct user_struct *find_user(uid_t);
  484. extern struct user_struct root_user;
  485. #define INIT_USER (&root_user)
  486. struct backing_dev_info;
  487. struct reclaim_state;
  488. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  489. struct sched_info {
  490. /* cumulative counters */
  491. unsigned long cpu_time, /* time spent on the cpu */
  492. run_delay, /* time spent waiting on a runqueue */
  493. pcnt; /* # of timeslices run on this cpu */
  494. /* timestamps */
  495. unsigned long last_arrival, /* when we last ran on a cpu */
  496. last_queued; /* when we were last queued to run */
  497. };
  498. #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
  499. #ifdef CONFIG_SCHEDSTATS
  500. extern const struct file_operations proc_schedstat_operations;
  501. #endif /* CONFIG_SCHEDSTATS */
  502. #ifdef CONFIG_TASK_DELAY_ACCT
  503. struct task_delay_info {
  504. spinlock_t lock;
  505. unsigned int flags; /* Private per-task flags */
  506. /* For each stat XXX, add following, aligned appropriately
  507. *
  508. * struct timespec XXX_start, XXX_end;
  509. * u64 XXX_delay;
  510. * u32 XXX_count;
  511. *
  512. * Atomicity of updates to XXX_delay, XXX_count protected by
  513. * single lock above (split into XXX_lock if contention is an issue).
  514. */
  515. /*
  516. * XXX_count is incremented on every XXX operation, the delay
  517. * associated with the operation is added to XXX_delay.
  518. * XXX_delay contains the accumulated delay time in nanoseconds.
  519. */
  520. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  521. u64 blkio_delay; /* wait for sync block io completion */
  522. u64 swapin_delay; /* wait for swapin block io completion */
  523. u32 blkio_count; /* total count of the number of sync block */
  524. /* io operations performed */
  525. u32 swapin_count; /* total count of the number of swapin block */
  526. /* io operations performed */
  527. };
  528. #endif /* CONFIG_TASK_DELAY_ACCT */
  529. static inline int sched_info_on(void)
  530. {
  531. #ifdef CONFIG_SCHEDSTATS
  532. return 1;
  533. #elif defined(CONFIG_TASK_DELAY_ACCT)
  534. extern int delayacct_on;
  535. return delayacct_on;
  536. #else
  537. return 0;
  538. #endif
  539. }
  540. enum idle_type
  541. {
  542. SCHED_IDLE,
  543. NOT_IDLE,
  544. NEWLY_IDLE,
  545. MAX_IDLE_TYPES
  546. };
  547. /*
  548. * sched-domains (multiprocessor balancing) declarations:
  549. */
  550. #define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
  551. #ifdef CONFIG_SMP
  552. #define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
  553. #define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
  554. #define SD_BALANCE_EXEC 4 /* Balance on exec */
  555. #define SD_BALANCE_FORK 8 /* Balance on fork, clone */
  556. #define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
  557. #define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
  558. #define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
  559. #define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
  560. #define SD_POWERSAVINGS_BALANCE 256 /* Balance for power savings */
  561. #define SD_SHARE_PKG_RESOURCES 512 /* Domain members share cpu pkg resources */
  562. #define SD_SERIALIZE 1024 /* Only a single load balancing instance */
  563. #define BALANCE_FOR_MC_POWER \
  564. (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
  565. #define BALANCE_FOR_PKG_POWER \
  566. ((sched_mc_power_savings || sched_smt_power_savings) ? \
  567. SD_POWERSAVINGS_BALANCE : 0)
  568. #define test_sd_parent(sd, flag) ((sd->parent && \
  569. (sd->parent->flags & flag)) ? 1 : 0)
  570. struct sched_group {
  571. struct sched_group *next; /* Must be a circular list */
  572. cpumask_t cpumask;
  573. /*
  574. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  575. * single CPU. This is read only (except for setup, hotplug CPU).
  576. * Note : Never change cpu_power without recompute its reciprocal
  577. */
  578. unsigned int __cpu_power;
  579. /*
  580. * reciprocal value of cpu_power to avoid expensive divides
  581. * (see include/linux/reciprocal_div.h)
  582. */
  583. u32 reciprocal_cpu_power;
  584. };
  585. struct sched_domain {
  586. /* These fields must be setup */
  587. struct sched_domain *parent; /* top domain must be null terminated */
  588. struct sched_domain *child; /* bottom domain must be null terminated */
  589. struct sched_group *groups; /* the balancing groups of the domain */
  590. cpumask_t span; /* span of all CPUs in this domain */
  591. unsigned long min_interval; /* Minimum balance interval ms */
  592. unsigned long max_interval; /* Maximum balance interval ms */
  593. unsigned int busy_factor; /* less balancing by factor if busy */
  594. unsigned int imbalance_pct; /* No balance until over watermark */
  595. unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
  596. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  597. unsigned int busy_idx;
  598. unsigned int idle_idx;
  599. unsigned int newidle_idx;
  600. unsigned int wake_idx;
  601. unsigned int forkexec_idx;
  602. int flags; /* See SD_* */
  603. /* Runtime fields. */
  604. unsigned long last_balance; /* init to jiffies. units in jiffies */
  605. unsigned int balance_interval; /* initialise to 1. units in ms. */
  606. unsigned int nr_balance_failed; /* initialise to 0 */
  607. #ifdef CONFIG_SCHEDSTATS
  608. /* load_balance() stats */
  609. unsigned long lb_cnt[MAX_IDLE_TYPES];
  610. unsigned long lb_failed[MAX_IDLE_TYPES];
  611. unsigned long lb_balanced[MAX_IDLE_TYPES];
  612. unsigned long lb_imbalance[MAX_IDLE_TYPES];
  613. unsigned long lb_gained[MAX_IDLE_TYPES];
  614. unsigned long lb_hot_gained[MAX_IDLE_TYPES];
  615. unsigned long lb_nobusyg[MAX_IDLE_TYPES];
  616. unsigned long lb_nobusyq[MAX_IDLE_TYPES];
  617. /* Active load balancing */
  618. unsigned long alb_cnt;
  619. unsigned long alb_failed;
  620. unsigned long alb_pushed;
  621. /* SD_BALANCE_EXEC stats */
  622. unsigned long sbe_cnt;
  623. unsigned long sbe_balanced;
  624. unsigned long sbe_pushed;
  625. /* SD_BALANCE_FORK stats */
  626. unsigned long sbf_cnt;
  627. unsigned long sbf_balanced;
  628. unsigned long sbf_pushed;
  629. /* try_to_wake_up() stats */
  630. unsigned long ttwu_wake_remote;
  631. unsigned long ttwu_move_affine;
  632. unsigned long ttwu_move_balance;
  633. #endif
  634. };
  635. extern int partition_sched_domains(cpumask_t *partition1,
  636. cpumask_t *partition2);
  637. /*
  638. * Maximum cache size the migration-costs auto-tuning code will
  639. * search from:
  640. */
  641. extern unsigned int max_cache_size;
  642. #endif /* CONFIG_SMP */
  643. struct io_context; /* See blkdev.h */
  644. struct cpuset;
  645. #define NGROUPS_SMALL 32
  646. #define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
  647. struct group_info {
  648. int ngroups;
  649. atomic_t usage;
  650. gid_t small_block[NGROUPS_SMALL];
  651. int nblocks;
  652. gid_t *blocks[0];
  653. };
  654. /*
  655. * get_group_info() must be called with the owning task locked (via task_lock())
  656. * when task != current. The reason being that the vast majority of callers are
  657. * looking at current->group_info, which can not be changed except by the
  658. * current task. Changing current->group_info requires the task lock, too.
  659. */
  660. #define get_group_info(group_info) do { \
  661. atomic_inc(&(group_info)->usage); \
  662. } while (0)
  663. #define put_group_info(group_info) do { \
  664. if (atomic_dec_and_test(&(group_info)->usage)) \
  665. groups_free(group_info); \
  666. } while (0)
  667. extern struct group_info *groups_alloc(int gidsetsize);
  668. extern void groups_free(struct group_info *group_info);
  669. extern int set_current_groups(struct group_info *group_info);
  670. extern int groups_search(struct group_info *group_info, gid_t grp);
  671. /* access the groups "array" with this macro */
  672. #define GROUP_AT(gi, i) \
  673. ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
  674. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  675. extern void prefetch_stack(struct task_struct *t);
  676. #else
  677. static inline void prefetch_stack(struct task_struct *t) { }
  678. #endif
  679. struct audit_context; /* See audit.c */
  680. struct mempolicy;
  681. struct pipe_inode_info;
  682. struct uts_namespace;
  683. enum sleep_type {
  684. SLEEP_NORMAL,
  685. SLEEP_NONINTERACTIVE,
  686. SLEEP_INTERACTIVE,
  687. SLEEP_INTERRUPTED,
  688. };
  689. struct prio_array;
  690. struct task_struct {
  691. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  692. struct thread_info *thread_info;
  693. atomic_t usage;
  694. unsigned int flags; /* per process flags, defined below */
  695. unsigned int ptrace;
  696. int lock_depth; /* BKL lock depth */
  697. #ifdef CONFIG_SMP
  698. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  699. int oncpu;
  700. #endif
  701. #endif
  702. int load_weight; /* for niceness load balancing purposes */
  703. int prio, static_prio, normal_prio;
  704. struct list_head run_list;
  705. struct prio_array *array;
  706. unsigned short ioprio;
  707. #ifdef CONFIG_BLK_DEV_IO_TRACE
  708. unsigned int btrace_seq;
  709. #endif
  710. unsigned long sleep_avg;
  711. unsigned long long timestamp, last_ran;
  712. unsigned long long sched_time; /* sched_clock time spent running */
  713. enum sleep_type sleep_type;
  714. unsigned int policy;
  715. cpumask_t cpus_allowed;
  716. unsigned int time_slice, first_time_slice;
  717. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  718. struct sched_info sched_info;
  719. #endif
  720. struct list_head tasks;
  721. /*
  722. * ptrace_list/ptrace_children forms the list of my children
  723. * that were stolen by a ptracer.
  724. */
  725. struct list_head ptrace_children;
  726. struct list_head ptrace_list;
  727. struct mm_struct *mm, *active_mm;
  728. /* task state */
  729. struct linux_binfmt *binfmt;
  730. int exit_state;
  731. int exit_code, exit_signal;
  732. int pdeath_signal; /* The signal sent when the parent dies */
  733. /* ??? */
  734. unsigned int personality;
  735. unsigned did_exec:1;
  736. pid_t pid;
  737. pid_t tgid;
  738. #ifdef CONFIG_CC_STACKPROTECTOR
  739. /* Canary value for the -fstack-protector gcc feature */
  740. unsigned long stack_canary;
  741. #endif
  742. /*
  743. * pointers to (original) parent process, youngest child, younger sibling,
  744. * older sibling, respectively. (p->father can be replaced with
  745. * p->parent->pid)
  746. */
  747. struct task_struct *real_parent; /* real parent process (when being debugged) */
  748. struct task_struct *parent; /* parent process */
  749. /*
  750. * children/sibling forms the list of my children plus the
  751. * tasks I'm ptracing.
  752. */
  753. struct list_head children; /* list of my children */
  754. struct list_head sibling; /* linkage in my parent's children list */
  755. struct task_struct *group_leader; /* threadgroup leader */
  756. /* PID/PID hash table linkage. */
  757. struct pid_link pids[PIDTYPE_MAX];
  758. struct list_head thread_group;
  759. struct completion *vfork_done; /* for vfork() */
  760. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  761. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  762. unsigned int rt_priority;
  763. cputime_t utime, stime;
  764. unsigned long nvcsw, nivcsw; /* context switch counts */
  765. struct timespec start_time;
  766. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  767. unsigned long min_flt, maj_flt;
  768. cputime_t it_prof_expires, it_virt_expires;
  769. unsigned long long it_sched_expires;
  770. struct list_head cpu_timers[3];
  771. /* process credentials */
  772. uid_t uid,euid,suid,fsuid;
  773. gid_t gid,egid,sgid,fsgid;
  774. struct group_info *group_info;
  775. kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
  776. unsigned keep_capabilities:1;
  777. struct user_struct *user;
  778. #ifdef CONFIG_KEYS
  779. struct key *request_key_auth; /* assumed request_key authority */
  780. struct key *thread_keyring; /* keyring private to this thread */
  781. unsigned char jit_keyring; /* default keyring to attach requested keys to */
  782. #endif
  783. /*
  784. * fpu_counter contains the number of consecutive context switches
  785. * that the FPU is used. If this is over a threshold, the lazy fpu
  786. * saving becomes unlazy to save the trap. This is an unsigned char
  787. * so that after 256 times the counter wraps and the behavior turns
  788. * lazy again; this to deal with bursty apps that only use FPU for
  789. * a short time
  790. */
  791. unsigned char fpu_counter;
  792. int oomkilladj; /* OOM kill score adjustment (bit shift). */
  793. char comm[TASK_COMM_LEN]; /* executable name excluding path
  794. - access with [gs]et_task_comm (which lock
  795. it with task_lock())
  796. - initialized normally by flush_old_exec */
  797. /* file system info */
  798. int link_count, total_link_count;
  799. #ifdef CONFIG_SYSVIPC
  800. /* ipc stuff */
  801. struct sysv_sem sysvsem;
  802. #endif
  803. /* CPU-specific state of this task */
  804. struct thread_struct thread;
  805. /* filesystem information */
  806. struct fs_struct *fs;
  807. /* open file information */
  808. struct files_struct *files;
  809. /* namespaces */
  810. struct nsproxy *nsproxy;
  811. /* signal handlers */
  812. struct signal_struct *signal;
  813. struct sighand_struct *sighand;
  814. sigset_t blocked, real_blocked;
  815. sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
  816. struct sigpending pending;
  817. unsigned long sas_ss_sp;
  818. size_t sas_ss_size;
  819. int (*notifier)(void *priv);
  820. void *notifier_data;
  821. sigset_t *notifier_mask;
  822. void *security;
  823. struct audit_context *audit_context;
  824. seccomp_t seccomp;
  825. /* Thread group tracking */
  826. u32 parent_exec_id;
  827. u32 self_exec_id;
  828. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
  829. spinlock_t alloc_lock;
  830. /* Protection of the PI data structures: */
  831. spinlock_t pi_lock;
  832. #ifdef CONFIG_RT_MUTEXES
  833. /* PI waiters blocked on a rt_mutex held by this task */
  834. struct plist_head pi_waiters;
  835. /* Deadlock detection and priority inheritance handling */
  836. struct rt_mutex_waiter *pi_blocked_on;
  837. #endif
  838. #ifdef CONFIG_DEBUG_MUTEXES
  839. /* mutex deadlock detection */
  840. struct mutex_waiter *blocked_on;
  841. #endif
  842. #ifdef CONFIG_TRACE_IRQFLAGS
  843. unsigned int irq_events;
  844. int hardirqs_enabled;
  845. unsigned long hardirq_enable_ip;
  846. unsigned int hardirq_enable_event;
  847. unsigned long hardirq_disable_ip;
  848. unsigned int hardirq_disable_event;
  849. int softirqs_enabled;
  850. unsigned long softirq_disable_ip;
  851. unsigned int softirq_disable_event;
  852. unsigned long softirq_enable_ip;
  853. unsigned int softirq_enable_event;
  854. int hardirq_context;
  855. int softirq_context;
  856. #endif
  857. #ifdef CONFIG_LOCKDEP
  858. # define MAX_LOCK_DEPTH 30UL
  859. u64 curr_chain_key;
  860. int lockdep_depth;
  861. struct held_lock held_locks[MAX_LOCK_DEPTH];
  862. unsigned int lockdep_recursion;
  863. #endif
  864. /* journalling filesystem info */
  865. void *journal_info;
  866. /* VM state */
  867. struct reclaim_state *reclaim_state;
  868. struct backing_dev_info *backing_dev_info;
  869. struct io_context *io_context;
  870. unsigned long ptrace_message;
  871. siginfo_t *last_siginfo; /* For ptrace use. */
  872. /*
  873. * current io wait handle: wait queue entry to use for io waits
  874. * If this thread is processing aio, this points at the waitqueue
  875. * inside the currently handled kiocb. It may be NULL (i.e. default
  876. * to a stack based synchronous wait) if its doing sync IO.
  877. */
  878. wait_queue_t *io_wait;
  879. #ifdef CONFIG_TASK_XACCT
  880. /* i/o counters(bytes read/written, #syscalls */
  881. u64 rchar, wchar, syscr, syscw;
  882. #endif
  883. struct task_io_accounting ioac;
  884. #if defined(CONFIG_TASK_XACCT)
  885. u64 acct_rss_mem1; /* accumulated rss usage */
  886. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  887. cputime_t acct_stimexpd;/* stime since last update */
  888. #endif
  889. #ifdef CONFIG_NUMA
  890. struct mempolicy *mempolicy;
  891. short il_next;
  892. #endif
  893. #ifdef CONFIG_CPUSETS
  894. struct cpuset *cpuset;
  895. nodemask_t mems_allowed;
  896. int cpuset_mems_generation;
  897. int cpuset_mem_spread_rotor;
  898. #endif
  899. struct robust_list_head __user *robust_list;
  900. #ifdef CONFIG_COMPAT
  901. struct compat_robust_list_head __user *compat_robust_list;
  902. #endif
  903. struct list_head pi_state_list;
  904. struct futex_pi_state *pi_state_cache;
  905. atomic_t fs_excl; /* holding fs exclusive resources */
  906. struct rcu_head rcu;
  907. /*
  908. * cache last used pipe for splice
  909. */
  910. struct pipe_inode_info *splice_pipe;
  911. #ifdef CONFIG_TASK_DELAY_ACCT
  912. struct task_delay_info *delays;
  913. #endif
  914. #ifdef CONFIG_FAULT_INJECTION
  915. int make_it_fail;
  916. #endif
  917. };
  918. static inline pid_t process_group(struct task_struct *tsk)
  919. {
  920. return tsk->signal->pgrp;
  921. }
  922. static inline pid_t signal_session(struct signal_struct *sig)
  923. {
  924. return sig->__session;
  925. }
  926. static inline pid_t process_session(struct task_struct *tsk)
  927. {
  928. return signal_session(tsk->signal);
  929. }
  930. static inline void set_signal_session(struct signal_struct *sig, pid_t session)
  931. {
  932. sig->__session = session;
  933. }
  934. static inline struct pid *task_pid(struct task_struct *task)
  935. {
  936. return task->pids[PIDTYPE_PID].pid;
  937. }
  938. static inline struct pid *task_tgid(struct task_struct *task)
  939. {
  940. return task->group_leader->pids[PIDTYPE_PID].pid;
  941. }
  942. static inline struct pid *task_pgrp(struct task_struct *task)
  943. {
  944. return task->group_leader->pids[PIDTYPE_PGID].pid;
  945. }
  946. static inline struct pid *task_session(struct task_struct *task)
  947. {
  948. return task->group_leader->pids[PIDTYPE_SID].pid;
  949. }
  950. /**
  951. * pid_alive - check that a task structure is not stale
  952. * @p: Task structure to be checked.
  953. *
  954. * Test if a process is not yet dead (at most zombie state)
  955. * If pid_alive fails, then pointers within the task structure
  956. * can be stale and must not be dereferenced.
  957. */
  958. static inline int pid_alive(struct task_struct *p)
  959. {
  960. return p->pids[PIDTYPE_PID].pid != NULL;
  961. }
  962. /**
  963. * is_init - check if a task structure is init
  964. * @tsk: Task structure to be checked.
  965. *
  966. * Check if a task structure is the first user space task the kernel created.
  967. */
  968. static inline int is_init(struct task_struct *tsk)
  969. {
  970. return tsk->pid == 1;
  971. }
  972. extern struct pid *cad_pid;
  973. extern void free_task(struct task_struct *tsk);
  974. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  975. extern void __put_task_struct(struct task_struct *t);
  976. static inline void put_task_struct(struct task_struct *t)
  977. {
  978. if (atomic_dec_and_test(&t->usage))
  979. __put_task_struct(t);
  980. }
  981. /*
  982. * Per process flags
  983. */
  984. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  985. /* Not implemented yet, only for 486*/
  986. #define PF_STARTING 0x00000002 /* being created */
  987. #define PF_EXITING 0x00000004 /* getting shut down */
  988. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  989. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  990. #define PF_DUMPCORE 0x00000200 /* dumped core */
  991. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  992. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  993. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  994. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  995. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  996. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  997. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  998. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  999. #define PF_SWAPOFF 0x00080000 /* I am in swapoff */
  1000. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  1001. #define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
  1002. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  1003. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  1004. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  1005. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  1006. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  1007. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  1008. /*
  1009. * Only the _current_ task can read/write to tsk->flags, but other
  1010. * tasks can access tsk->flags in readonly mode for example
  1011. * with tsk_used_math (like during threaded core dumping).
  1012. * There is however an exception to this rule during ptrace
  1013. * or during fork: the ptracer task is allowed to write to the
  1014. * child->flags of its traced child (same goes for fork, the parent
  1015. * can write to the child->flags), because we're guaranteed the
  1016. * child is not running and in turn not changing child->flags
  1017. * at the same time the parent does it.
  1018. */
  1019. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  1020. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  1021. #define clear_used_math() clear_stopped_child_used_math(current)
  1022. #define set_used_math() set_stopped_child_used_math(current)
  1023. #define conditional_stopped_child_used_math(condition, child) \
  1024. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  1025. #define conditional_used_math(condition) \
  1026. conditional_stopped_child_used_math(condition, current)
  1027. #define copy_to_stopped_child_used_math(child) \
  1028. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  1029. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  1030. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  1031. #define used_math() tsk_used_math(current)
  1032. #ifdef CONFIG_SMP
  1033. extern int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask);
  1034. #else
  1035. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  1036. {
  1037. if (!cpu_isset(0, new_mask))
  1038. return -EINVAL;
  1039. return 0;
  1040. }
  1041. #endif
  1042. extern unsigned long long sched_clock(void);
  1043. extern unsigned long long
  1044. current_sched_time(const struct task_struct *current_task);
  1045. /* sched_exec is called by processes performing an exec */
  1046. #ifdef CONFIG_SMP
  1047. extern void sched_exec(void);
  1048. #else
  1049. #define sched_exec() {}
  1050. #endif
  1051. #ifdef CONFIG_HOTPLUG_CPU
  1052. extern void idle_task_exit(void);
  1053. #else
  1054. static inline void idle_task_exit(void) {}
  1055. #endif
  1056. extern void sched_idle_next(void);
  1057. #ifdef CONFIG_RT_MUTEXES
  1058. extern int rt_mutex_getprio(struct task_struct *p);
  1059. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  1060. extern void rt_mutex_adjust_pi(struct task_struct *p);
  1061. #else
  1062. static inline int rt_mutex_getprio(struct task_struct *p)
  1063. {
  1064. return p->normal_prio;
  1065. }
  1066. # define rt_mutex_adjust_pi(p) do { } while (0)
  1067. #endif
  1068. extern void set_user_nice(struct task_struct *p, long nice);
  1069. extern int task_prio(const struct task_struct *p);
  1070. extern int task_nice(const struct task_struct *p);
  1071. extern int can_nice(const struct task_struct *p, const int nice);
  1072. extern int task_curr(const struct task_struct *p);
  1073. extern int idle_cpu(int cpu);
  1074. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  1075. extern struct task_struct *idle_task(int cpu);
  1076. extern struct task_struct *curr_task(int cpu);
  1077. extern void set_curr_task(int cpu, struct task_struct *p);
  1078. void yield(void);
  1079. /*
  1080. * The default (Linux) execution domain.
  1081. */
  1082. extern struct exec_domain default_exec_domain;
  1083. union thread_union {
  1084. struct thread_info thread_info;
  1085. unsigned long stack[THREAD_SIZE/sizeof(long)];
  1086. };
  1087. #ifndef __HAVE_ARCH_KSTACK_END
  1088. static inline int kstack_end(void *addr)
  1089. {
  1090. /* Reliable end of stack detection:
  1091. * Some APM bios versions misalign the stack
  1092. */
  1093. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  1094. }
  1095. #endif
  1096. extern union thread_union init_thread_union;
  1097. extern struct task_struct init_task;
  1098. extern struct mm_struct init_mm;
  1099. #define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
  1100. extern struct task_struct *find_task_by_pid_type(int type, int pid);
  1101. extern void __set_special_pids(pid_t session, pid_t pgrp);
  1102. /* per-UID process charging. */
  1103. extern struct user_struct * alloc_uid(uid_t);
  1104. static inline struct user_struct *get_uid(struct user_struct *u)
  1105. {
  1106. atomic_inc(&u->__count);
  1107. return u;
  1108. }
  1109. extern void free_uid(struct user_struct *);
  1110. extern void switch_uid(struct user_struct *);
  1111. #include <asm/current.h>
  1112. extern void do_timer(unsigned long ticks);
  1113. extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
  1114. extern int FASTCALL(wake_up_process(struct task_struct * tsk));
  1115. extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
  1116. unsigned long clone_flags));
  1117. #ifdef CONFIG_SMP
  1118. extern void kick_process(struct task_struct *tsk);
  1119. #else
  1120. static inline void kick_process(struct task_struct *tsk) { }
  1121. #endif
  1122. extern void FASTCALL(sched_fork(struct task_struct * p, int clone_flags));
  1123. extern void FASTCALL(sched_exit(struct task_struct * p));
  1124. extern int in_group_p(gid_t);
  1125. extern int in_egroup_p(gid_t);
  1126. extern void proc_caches_init(void);
  1127. extern void flush_signals(struct task_struct *);
  1128. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1129. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1130. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1131. {
  1132. unsigned long flags;
  1133. int ret;
  1134. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1135. ret = dequeue_signal(tsk, mask, info);
  1136. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1137. return ret;
  1138. }
  1139. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1140. sigset_t *mask);
  1141. extern void unblock_all_signals(void);
  1142. extern void release_task(struct task_struct * p);
  1143. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1144. extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
  1145. extern int force_sigsegv(int, struct task_struct *);
  1146. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1147. extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1148. extern int kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1149. extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
  1150. extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
  1151. extern int kill_pgrp(struct pid *pid, int sig, int priv);
  1152. extern int kill_pid(struct pid *pid, int sig, int priv);
  1153. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1154. extern void do_notify_parent(struct task_struct *, int);
  1155. extern void force_sig(int, struct task_struct *);
  1156. extern void force_sig_specific(int, struct task_struct *);
  1157. extern int send_sig(int, struct task_struct *, int);
  1158. extern void zap_other_threads(struct task_struct *p);
  1159. extern int kill_proc(pid_t, int, int);
  1160. extern struct sigqueue *sigqueue_alloc(void);
  1161. extern void sigqueue_free(struct sigqueue *);
  1162. extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
  1163. extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
  1164. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1165. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1166. static inline int kill_cad_pid(int sig, int priv)
  1167. {
  1168. return kill_pid(cad_pid, sig, priv);
  1169. }
  1170. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1171. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1172. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1173. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1174. static inline int is_si_special(const struct siginfo *info)
  1175. {
  1176. return info <= SEND_SIG_FORCED;
  1177. }
  1178. /* True if we are on the alternate signal stack. */
  1179. static inline int on_sig_stack(unsigned long sp)
  1180. {
  1181. return (sp - current->sas_ss_sp < current->sas_ss_size);
  1182. }
  1183. static inline int sas_ss_flags(unsigned long sp)
  1184. {
  1185. return (current->sas_ss_size == 0 ? SS_DISABLE
  1186. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1187. }
  1188. /*
  1189. * Routines for handling mm_structs
  1190. */
  1191. extern struct mm_struct * mm_alloc(void);
  1192. /* mmdrop drops the mm and the page tables */
  1193. extern void FASTCALL(__mmdrop(struct mm_struct *));
  1194. static inline void mmdrop(struct mm_struct * mm)
  1195. {
  1196. if (atomic_dec_and_test(&mm->mm_count))
  1197. __mmdrop(mm);
  1198. }
  1199. /* mmput gets rid of the mappings and all user-space */
  1200. extern void mmput(struct mm_struct *);
  1201. /* Grab a reference to a task's mm, if it is not already going away */
  1202. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1203. /* Remove the current tasks stale references to the old mm_struct */
  1204. extern void mm_release(struct task_struct *, struct mm_struct *);
  1205. extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
  1206. extern void flush_thread(void);
  1207. extern void exit_thread(void);
  1208. extern void exit_files(struct task_struct *);
  1209. extern void __cleanup_signal(struct signal_struct *);
  1210. extern void __cleanup_sighand(struct sighand_struct *);
  1211. extern void exit_itimers(struct signal_struct *);
  1212. extern NORET_TYPE void do_group_exit(int);
  1213. extern void daemonize(const char *, ...);
  1214. extern int allow_signal(int);
  1215. extern int disallow_signal(int);
  1216. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1217. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1218. struct task_struct *fork_idle(int);
  1219. extern void set_task_comm(struct task_struct *tsk, char *from);
  1220. extern void get_task_comm(char *to, struct task_struct *tsk);
  1221. #ifdef CONFIG_SMP
  1222. extern void wait_task_inactive(struct task_struct * p);
  1223. #else
  1224. #define wait_task_inactive(p) do { } while (0)
  1225. #endif
  1226. #define remove_parent(p) list_del_init(&(p)->sibling)
  1227. #define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
  1228. #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
  1229. #define for_each_process(p) \
  1230. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1231. /*
  1232. * Careful: do_each_thread/while_each_thread is a double loop so
  1233. * 'break' will not work as expected - use goto instead.
  1234. */
  1235. #define do_each_thread(g, t) \
  1236. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1237. #define while_each_thread(g, t) \
  1238. while ((t = next_thread(t)) != g)
  1239. /* de_thread depends on thread_group_leader not being a pid based check */
  1240. #define thread_group_leader(p) (p == p->group_leader)
  1241. /* Do to the insanities of de_thread it is possible for a process
  1242. * to have the pid of the thread group leader without actually being
  1243. * the thread group leader. For iteration through the pids in proc
  1244. * all we care about is that we have a task with the appropriate
  1245. * pid, we don't actually care if we have the right task.
  1246. */
  1247. static inline int has_group_leader_pid(struct task_struct *p)
  1248. {
  1249. return p->pid == p->tgid;
  1250. }
  1251. static inline struct task_struct *next_thread(const struct task_struct *p)
  1252. {
  1253. return list_entry(rcu_dereference(p->thread_group.next),
  1254. struct task_struct, thread_group);
  1255. }
  1256. static inline int thread_group_empty(struct task_struct *p)
  1257. {
  1258. return list_empty(&p->thread_group);
  1259. }
  1260. #define delay_group_leader(p) \
  1261. (thread_group_leader(p) && !thread_group_empty(p))
  1262. /*
  1263. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1264. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1265. * pins the final release of task.io_context. Also protects ->cpuset.
  1266. *
  1267. * Nests both inside and outside of read_lock(&tasklist_lock).
  1268. * It must not be nested with write_lock_irq(&tasklist_lock),
  1269. * neither inside nor outside.
  1270. */
  1271. static inline void task_lock(struct task_struct *p)
  1272. {
  1273. spin_lock(&p->alloc_lock);
  1274. }
  1275. static inline void task_unlock(struct task_struct *p)
  1276. {
  1277. spin_unlock(&p->alloc_lock);
  1278. }
  1279. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1280. unsigned long *flags);
  1281. static inline void unlock_task_sighand(struct task_struct *tsk,
  1282. unsigned long *flags)
  1283. {
  1284. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1285. }
  1286. #ifndef __HAVE_THREAD_FUNCTIONS
  1287. #define task_thread_info(task) (task)->thread_info
  1288. #define task_stack_page(task) ((void*)((task)->thread_info))
  1289. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1290. {
  1291. *task_thread_info(p) = *task_thread_info(org);
  1292. task_thread_info(p)->task = p;
  1293. }
  1294. static inline unsigned long *end_of_stack(struct task_struct *p)
  1295. {
  1296. return (unsigned long *)(p->thread_info + 1);
  1297. }
  1298. #endif
  1299. /* set thread flags in other task's structures
  1300. * - see asm/thread_info.h for TIF_xxxx flags available
  1301. */
  1302. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1303. {
  1304. set_ti_thread_flag(task_thread_info(tsk), flag);
  1305. }
  1306. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1307. {
  1308. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1309. }
  1310. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1311. {
  1312. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1313. }
  1314. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1315. {
  1316. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1317. }
  1318. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1319. {
  1320. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1321. }
  1322. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1323. {
  1324. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1325. }
  1326. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1327. {
  1328. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1329. }
  1330. static inline int signal_pending(struct task_struct *p)
  1331. {
  1332. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1333. }
  1334. static inline int need_resched(void)
  1335. {
  1336. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  1337. }
  1338. /*
  1339. * cond_resched() and cond_resched_lock(): latency reduction via
  1340. * explicit rescheduling in places that are safe. The return
  1341. * value indicates whether a reschedule was done in fact.
  1342. * cond_resched_lock() will drop the spinlock before scheduling,
  1343. * cond_resched_softirq() will enable bhs before scheduling.
  1344. */
  1345. extern int cond_resched(void);
  1346. extern int cond_resched_lock(spinlock_t * lock);
  1347. extern int cond_resched_softirq(void);
  1348. /*
  1349. * Does a critical section need to be broken due to another
  1350. * task waiting?:
  1351. */
  1352. #if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
  1353. # define need_lockbreak(lock) ((lock)->break_lock)
  1354. #else
  1355. # define need_lockbreak(lock) 0
  1356. #endif
  1357. /*
  1358. * Does a critical section need to be broken due to another
  1359. * task waiting or preemption being signalled:
  1360. */
  1361. static inline int lock_need_resched(spinlock_t *lock)
  1362. {
  1363. if (need_lockbreak(lock) || need_resched())
  1364. return 1;
  1365. return 0;
  1366. }
  1367. /* Reevaluate whether the task has signals pending delivery.
  1368. This is required every time the blocked sigset_t changes.
  1369. callers must hold sighand->siglock. */
  1370. extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
  1371. extern void recalc_sigpending(void);
  1372. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  1373. /*
  1374. * Wrappers for p->thread_info->cpu access. No-op on UP.
  1375. */
  1376. #ifdef CONFIG_SMP
  1377. static inline unsigned int task_cpu(const struct task_struct *p)
  1378. {
  1379. return task_thread_info(p)->cpu;
  1380. }
  1381. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1382. {
  1383. task_thread_info(p)->cpu = cpu;
  1384. }
  1385. #else
  1386. static inline unsigned int task_cpu(const struct task_struct *p)
  1387. {
  1388. return 0;
  1389. }
  1390. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  1391. {
  1392. }
  1393. #endif /* CONFIG_SMP */
  1394. #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
  1395. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  1396. #else
  1397. static inline void arch_pick_mmap_layout(struct mm_struct *mm)
  1398. {
  1399. mm->mmap_base = TASK_UNMAPPED_BASE;
  1400. mm->get_unmapped_area = arch_get_unmapped_area;
  1401. mm->unmap_area = arch_unmap_area;
  1402. }
  1403. #endif
  1404. extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
  1405. extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
  1406. extern int sched_mc_power_savings, sched_smt_power_savings;
  1407. extern void normalize_rt_tasks(void);
  1408. #ifdef CONFIG_TASK_XACCT
  1409. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1410. {
  1411. tsk->rchar += amt;
  1412. }
  1413. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1414. {
  1415. tsk->wchar += amt;
  1416. }
  1417. static inline void inc_syscr(struct task_struct *tsk)
  1418. {
  1419. tsk->syscr++;
  1420. }
  1421. static inline void inc_syscw(struct task_struct *tsk)
  1422. {
  1423. tsk->syscw++;
  1424. }
  1425. #else
  1426. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  1427. {
  1428. }
  1429. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  1430. {
  1431. }
  1432. static inline void inc_syscr(struct task_struct *tsk)
  1433. {
  1434. }
  1435. static inline void inc_syscw(struct task_struct *tsk)
  1436. {
  1437. }
  1438. #endif
  1439. #endif /* __KERNEL__ */
  1440. #endif