sched.h 75 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568256925702571257225732574257525762577257825792580258125822583258425852586258725882589259025912592259325942595259625972598259926002601
  1. #ifndef _LINUX_SCHED_H
  2. #define _LINUX_SCHED_H
  3. /*
  4. * cloning flags:
  5. */
  6. #define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
  7. #define CLONE_VM 0x00000100 /* set if VM shared between processes */
  8. #define CLONE_FS 0x00000200 /* set if fs info shared between processes */
  9. #define CLONE_FILES 0x00000400 /* set if open files shared between processes */
  10. #define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
  11. #define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
  12. #define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
  13. #define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
  14. #define CLONE_THREAD 0x00010000 /* Same thread group? */
  15. #define CLONE_NEWNS 0x00020000 /* New namespace group? */
  16. #define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
  17. #define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
  18. #define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
  19. #define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
  20. #define CLONE_DETACHED 0x00400000 /* Unused, ignored */
  21. #define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
  22. #define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
  23. #define CLONE_STOPPED 0x02000000 /* Start in stopped state */
  24. #define CLONE_NEWUTS 0x04000000 /* New utsname group? */
  25. #define CLONE_NEWIPC 0x08000000 /* New ipcs */
  26. #define CLONE_NEWUSER 0x10000000 /* New user namespace */
  27. #define CLONE_NEWPID 0x20000000 /* New pid namespace */
  28. #define CLONE_NEWNET 0x40000000 /* New network namespace */
  29. #define CLONE_IO 0x80000000 /* Clone io context */
  30. /*
  31. * Scheduling policies
  32. */
  33. #define SCHED_NORMAL 0
  34. #define SCHED_FIFO 1
  35. #define SCHED_RR 2
  36. #define SCHED_BATCH 3
  37. /* SCHED_ISO: reserved but not implemented yet */
  38. #define SCHED_IDLE 5
  39. /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */
  40. #define SCHED_RESET_ON_FORK 0x40000000
  41. #ifdef __KERNEL__
  42. struct sched_param {
  43. int sched_priority;
  44. };
  45. #include <asm/param.h> /* for HZ */
  46. #include <linux/capability.h>
  47. #include <linux/threads.h>
  48. #include <linux/kernel.h>
  49. #include <linux/types.h>
  50. #include <linux/timex.h>
  51. #include <linux/jiffies.h>
  52. #include <linux/rbtree.h>
  53. #include <linux/thread_info.h>
  54. #include <linux/cpumask.h>
  55. #include <linux/errno.h>
  56. #include <linux/nodemask.h>
  57. #include <linux/mm_types.h>
  58. #include <asm/system.h>
  59. #include <asm/page.h>
  60. #include <asm/ptrace.h>
  61. #include <asm/cputime.h>
  62. #include <linux/smp.h>
  63. #include <linux/sem.h>
  64. #include <linux/signal.h>
  65. #include <linux/path.h>
  66. #include <linux/compiler.h>
  67. #include <linux/completion.h>
  68. #include <linux/pid.h>
  69. #include <linux/percpu.h>
  70. #include <linux/topology.h>
  71. #include <linux/proportions.h>
  72. #include <linux/seccomp.h>
  73. #include <linux/rcupdate.h>
  74. #include <linux/rculist.h>
  75. #include <linux/rtmutex.h>
  76. #include <linux/time.h>
  77. #include <linux/param.h>
  78. #include <linux/resource.h>
  79. #include <linux/timer.h>
  80. #include <linux/hrtimer.h>
  81. #include <linux/task_io_accounting.h>
  82. #include <linux/kobject.h>
  83. #include <linux/latencytop.h>
  84. #include <linux/cred.h>
  85. #include <asm/processor.h>
  86. struct exec_domain;
  87. struct futex_pi_state;
  88. struct robust_list_head;
  89. struct bio;
  90. struct fs_struct;
  91. struct bts_context;
  92. struct perf_event_context;
  93. /*
  94. * List of flags we want to share for kernel threads,
  95. * if only because they are not used by them anyway.
  96. */
  97. #define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
  98. /*
  99. * These are the constant used to fake the fixed-point load-average
  100. * counting. Some notes:
  101. * - 11 bit fractions expand to 22 bits by the multiplies: this gives
  102. * a load-average precision of 10 bits integer + 11 bits fractional
  103. * - if you want to count load-averages more often, you need more
  104. * precision, or rounding will get you. With 2-second counting freq,
  105. * the EXP_n values would be 1981, 2034 and 2043 if still using only
  106. * 11 bit fractions.
  107. */
  108. extern unsigned long avenrun[]; /* Load averages */
  109. extern void get_avenrun(unsigned long *loads, unsigned long offset, int shift);
  110. #define FSHIFT 11 /* nr of bits of precision */
  111. #define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
  112. #define LOAD_FREQ (5*HZ+1) /* 5 sec intervals */
  113. #define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
  114. #define EXP_5 2014 /* 1/exp(5sec/5min) */
  115. #define EXP_15 2037 /* 1/exp(5sec/15min) */
  116. #define CALC_LOAD(load,exp,n) \
  117. load *= exp; \
  118. load += n*(FIXED_1-exp); \
  119. load >>= FSHIFT;
  120. extern unsigned long total_forks;
  121. extern int nr_threads;
  122. DECLARE_PER_CPU(unsigned long, process_counts);
  123. extern int nr_processes(void);
  124. extern unsigned long nr_running(void);
  125. extern unsigned long nr_uninterruptible(void);
  126. extern unsigned long nr_iowait(void);
  127. extern unsigned long nr_iowait_cpu(void);
  128. extern unsigned long this_cpu_load(void);
  129. extern void calc_global_load(void);
  130. extern unsigned long get_parent_ip(unsigned long addr);
  131. struct seq_file;
  132. struct cfs_rq;
  133. struct task_group;
  134. #ifdef CONFIG_SCHED_DEBUG
  135. extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
  136. extern void proc_sched_set_task(struct task_struct *p);
  137. extern void
  138. print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
  139. #else
  140. static inline void
  141. proc_sched_show_task(struct task_struct *p, struct seq_file *m)
  142. {
  143. }
  144. static inline void proc_sched_set_task(struct task_struct *p)
  145. {
  146. }
  147. static inline void
  148. print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
  149. {
  150. }
  151. #endif
  152. /*
  153. * Task state bitmask. NOTE! These bits are also
  154. * encoded in fs/proc/array.c: get_task_state().
  155. *
  156. * We have two separate sets of flags: task->state
  157. * is about runnability, while task->exit_state are
  158. * about the task exiting. Confusing, but this way
  159. * modifying one set can't modify the other one by
  160. * mistake.
  161. */
  162. #define TASK_RUNNING 0
  163. #define TASK_INTERRUPTIBLE 1
  164. #define TASK_UNINTERRUPTIBLE 2
  165. #define __TASK_STOPPED 4
  166. #define __TASK_TRACED 8
  167. /* in tsk->exit_state */
  168. #define EXIT_ZOMBIE 16
  169. #define EXIT_DEAD 32
  170. /* in tsk->state again */
  171. #define TASK_DEAD 64
  172. #define TASK_WAKEKILL 128
  173. #define TASK_WAKING 256
  174. /* Convenience macros for the sake of set_task_state */
  175. #define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
  176. #define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
  177. #define TASK_TRACED (TASK_WAKEKILL | __TASK_TRACED)
  178. /* Convenience macros for the sake of wake_up */
  179. #define TASK_NORMAL (TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE)
  180. #define TASK_ALL (TASK_NORMAL | __TASK_STOPPED | __TASK_TRACED)
  181. /* get_task_state() */
  182. #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \
  183. TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \
  184. __TASK_TRACED)
  185. #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0)
  186. #define task_is_stopped(task) ((task->state & __TASK_STOPPED) != 0)
  187. #define task_is_stopped_or_traced(task) \
  188. ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
  189. #define task_contributes_to_load(task) \
  190. ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
  191. (task->flags & PF_FREEZING) == 0)
  192. #define __set_task_state(tsk, state_value) \
  193. do { (tsk)->state = (state_value); } while (0)
  194. #define set_task_state(tsk, state_value) \
  195. set_mb((tsk)->state, (state_value))
  196. /*
  197. * set_current_state() includes a barrier so that the write of current->state
  198. * is correctly serialised wrt the caller's subsequent test of whether to
  199. * actually sleep:
  200. *
  201. * set_current_state(TASK_UNINTERRUPTIBLE);
  202. * if (do_i_need_to_sleep())
  203. * schedule();
  204. *
  205. * If the caller does not need such serialisation then use __set_current_state()
  206. */
  207. #define __set_current_state(state_value) \
  208. do { current->state = (state_value); } while (0)
  209. #define set_current_state(state_value) \
  210. set_mb(current->state, (state_value))
  211. /* Task command name length */
  212. #define TASK_COMM_LEN 16
  213. #include <linux/spinlock.h>
  214. /*
  215. * This serializes "schedule()" and also protects
  216. * the run-queue from deletions/modifications (but
  217. * _adding_ to the beginning of the run-queue has
  218. * a separate lock).
  219. */
  220. extern rwlock_t tasklist_lock;
  221. extern spinlock_t mmlist_lock;
  222. struct task_struct;
  223. extern void sched_init(void);
  224. extern void sched_init_smp(void);
  225. extern asmlinkage void schedule_tail(struct task_struct *prev);
  226. extern void init_idle(struct task_struct *idle, int cpu);
  227. extern void init_idle_bootup_task(struct task_struct *idle);
  228. extern int runqueue_is_locked(int cpu);
  229. extern void task_rq_unlock_wait(struct task_struct *p);
  230. extern cpumask_var_t nohz_cpu_mask;
  231. #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
  232. extern int select_nohz_load_balancer(int cpu);
  233. extern int get_nohz_load_balancer(void);
  234. #else
  235. static inline int select_nohz_load_balancer(int cpu)
  236. {
  237. return 0;
  238. }
  239. #endif
  240. /*
  241. * Only dump TASK_* tasks. (0 for all tasks)
  242. */
  243. extern void show_state_filter(unsigned long state_filter);
  244. static inline void show_state(void)
  245. {
  246. show_state_filter(0);
  247. }
  248. extern void show_regs(struct pt_regs *);
  249. /*
  250. * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
  251. * task), SP is the stack pointer of the first frame that should be shown in the back
  252. * trace (or NULL if the entire call-chain of the task should be shown).
  253. */
  254. extern void show_stack(struct task_struct *task, unsigned long *sp);
  255. void io_schedule(void);
  256. long io_schedule_timeout(long timeout);
  257. extern void cpu_init (void);
  258. extern void trap_init(void);
  259. extern void update_process_times(int user);
  260. extern void scheduler_tick(void);
  261. extern void sched_show_task(struct task_struct *p);
  262. #ifdef CONFIG_DETECT_SOFTLOCKUP
  263. extern void softlockup_tick(void);
  264. extern void touch_softlockup_watchdog(void);
  265. extern void touch_all_softlockup_watchdogs(void);
  266. extern int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
  267. void __user *buffer,
  268. size_t *lenp, loff_t *ppos);
  269. extern unsigned int softlockup_panic;
  270. extern int softlockup_thresh;
  271. #else
  272. static inline void softlockup_tick(void)
  273. {
  274. }
  275. static inline void touch_softlockup_watchdog(void)
  276. {
  277. }
  278. static inline void touch_all_softlockup_watchdogs(void)
  279. {
  280. }
  281. #endif
  282. #ifdef CONFIG_DETECT_HUNG_TASK
  283. extern unsigned int sysctl_hung_task_panic;
  284. extern unsigned long sysctl_hung_task_check_count;
  285. extern unsigned long sysctl_hung_task_timeout_secs;
  286. extern unsigned long sysctl_hung_task_warnings;
  287. extern int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
  288. void __user *buffer,
  289. size_t *lenp, loff_t *ppos);
  290. #endif
  291. /* Attach to any functions which should be ignored in wchan output. */
  292. #define __sched __attribute__((__section__(".sched.text")))
  293. /* Linker adds these: start and end of __sched functions */
  294. extern char __sched_text_start[], __sched_text_end[];
  295. /* Is this address in the __sched functions? */
  296. extern int in_sched_functions(unsigned long addr);
  297. #define MAX_SCHEDULE_TIMEOUT LONG_MAX
  298. extern signed long schedule_timeout(signed long timeout);
  299. extern signed long schedule_timeout_interruptible(signed long timeout);
  300. extern signed long schedule_timeout_killable(signed long timeout);
  301. extern signed long schedule_timeout_uninterruptible(signed long timeout);
  302. asmlinkage void schedule(void);
  303. extern int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner);
  304. struct nsproxy;
  305. struct user_namespace;
  306. /*
  307. * Default maximum number of active map areas, this limits the number of vmas
  308. * per mm struct. Users can overwrite this number by sysctl but there is a
  309. * problem.
  310. *
  311. * When a program's coredump is generated as ELF format, a section is created
  312. * per a vma. In ELF, the number of sections is represented in unsigned short.
  313. * This means the number of sections should be smaller than 65535 at coredump.
  314. * Because the kernel adds some informative sections to a image of program at
  315. * generating coredump, we need some margin. The number of extra sections is
  316. * 1-3 now and depends on arch. We use "5" as safe margin, here.
  317. */
  318. #define MAPCOUNT_ELF_CORE_MARGIN (5)
  319. #define DEFAULT_MAX_MAP_COUNT (USHORT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
  320. extern int sysctl_max_map_count;
  321. #include <linux/aio.h>
  322. extern unsigned long
  323. arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
  324. unsigned long, unsigned long);
  325. extern unsigned long
  326. arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
  327. unsigned long len, unsigned long pgoff,
  328. unsigned long flags);
  329. extern void arch_unmap_area(struct mm_struct *, unsigned long);
  330. extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
  331. #if USE_SPLIT_PTLOCKS
  332. /*
  333. * The mm counters are not protected by its page_table_lock,
  334. * so must be incremented atomically.
  335. */
  336. #define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
  337. #define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
  338. #define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
  339. #define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
  340. #define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
  341. #else /* !USE_SPLIT_PTLOCKS */
  342. /*
  343. * The mm counters are protected by its page_table_lock,
  344. * so can be incremented directly.
  345. */
  346. #define set_mm_counter(mm, member, value) (mm)->_##member = (value)
  347. #define get_mm_counter(mm, member) ((mm)->_##member)
  348. #define add_mm_counter(mm, member, value) (mm)->_##member += (value)
  349. #define inc_mm_counter(mm, member) (mm)->_##member++
  350. #define dec_mm_counter(mm, member) (mm)->_##member--
  351. #endif /* !USE_SPLIT_PTLOCKS */
  352. #define get_mm_rss(mm) \
  353. (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
  354. #define update_hiwater_rss(mm) do { \
  355. unsigned long _rss = get_mm_rss(mm); \
  356. if ((mm)->hiwater_rss < _rss) \
  357. (mm)->hiwater_rss = _rss; \
  358. } while (0)
  359. #define update_hiwater_vm(mm) do { \
  360. if ((mm)->hiwater_vm < (mm)->total_vm) \
  361. (mm)->hiwater_vm = (mm)->total_vm; \
  362. } while (0)
  363. static inline unsigned long get_mm_hiwater_rss(struct mm_struct *mm)
  364. {
  365. return max(mm->hiwater_rss, get_mm_rss(mm));
  366. }
  367. static inline void setmax_mm_hiwater_rss(unsigned long *maxrss,
  368. struct mm_struct *mm)
  369. {
  370. unsigned long hiwater_rss = get_mm_hiwater_rss(mm);
  371. if (*maxrss < hiwater_rss)
  372. *maxrss = hiwater_rss;
  373. }
  374. static inline unsigned long get_mm_hiwater_vm(struct mm_struct *mm)
  375. {
  376. return max(mm->hiwater_vm, mm->total_vm);
  377. }
  378. extern void set_dumpable(struct mm_struct *mm, int value);
  379. extern int get_dumpable(struct mm_struct *mm);
  380. /* mm flags */
  381. /* dumpable bits */
  382. #define MMF_DUMPABLE 0 /* core dump is permitted */
  383. #define MMF_DUMP_SECURELY 1 /* core file is readable only by root */
  384. #define MMF_DUMPABLE_BITS 2
  385. #define MMF_DUMPABLE_MASK ((1 << MMF_DUMPABLE_BITS) - 1)
  386. /* coredump filter bits */
  387. #define MMF_DUMP_ANON_PRIVATE 2
  388. #define MMF_DUMP_ANON_SHARED 3
  389. #define MMF_DUMP_MAPPED_PRIVATE 4
  390. #define MMF_DUMP_MAPPED_SHARED 5
  391. #define MMF_DUMP_ELF_HEADERS 6
  392. #define MMF_DUMP_HUGETLB_PRIVATE 7
  393. #define MMF_DUMP_HUGETLB_SHARED 8
  394. #define MMF_DUMP_FILTER_SHIFT MMF_DUMPABLE_BITS
  395. #define MMF_DUMP_FILTER_BITS 7
  396. #define MMF_DUMP_FILTER_MASK \
  397. (((1 << MMF_DUMP_FILTER_BITS) - 1) << MMF_DUMP_FILTER_SHIFT)
  398. #define MMF_DUMP_FILTER_DEFAULT \
  399. ((1 << MMF_DUMP_ANON_PRIVATE) | (1 << MMF_DUMP_ANON_SHARED) |\
  400. (1 << MMF_DUMP_HUGETLB_PRIVATE) | MMF_DUMP_MASK_DEFAULT_ELF)
  401. #ifdef CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
  402. # define MMF_DUMP_MASK_DEFAULT_ELF (1 << MMF_DUMP_ELF_HEADERS)
  403. #else
  404. # define MMF_DUMP_MASK_DEFAULT_ELF 0
  405. #endif
  406. /* leave room for more dump flags */
  407. #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
  408. #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
  409. struct sighand_struct {
  410. atomic_t count;
  411. struct k_sigaction action[_NSIG];
  412. spinlock_t siglock;
  413. wait_queue_head_t signalfd_wqh;
  414. };
  415. struct pacct_struct {
  416. int ac_flag;
  417. long ac_exitcode;
  418. unsigned long ac_mem;
  419. cputime_t ac_utime, ac_stime;
  420. unsigned long ac_minflt, ac_majflt;
  421. };
  422. struct cpu_itimer {
  423. cputime_t expires;
  424. cputime_t incr;
  425. u32 error;
  426. u32 incr_error;
  427. };
  428. /**
  429. * struct task_cputime - collected CPU time counts
  430. * @utime: time spent in user mode, in &cputime_t units
  431. * @stime: time spent in kernel mode, in &cputime_t units
  432. * @sum_exec_runtime: total time spent on the CPU, in nanoseconds
  433. *
  434. * This structure groups together three kinds of CPU time that are
  435. * tracked for threads and thread groups. Most things considering
  436. * CPU time want to group these counts together and treat all three
  437. * of them in parallel.
  438. */
  439. struct task_cputime {
  440. cputime_t utime;
  441. cputime_t stime;
  442. unsigned long long sum_exec_runtime;
  443. };
  444. /* Alternate field names when used to cache expirations. */
  445. #define prof_exp stime
  446. #define virt_exp utime
  447. #define sched_exp sum_exec_runtime
  448. #define INIT_CPUTIME \
  449. (struct task_cputime) { \
  450. .utime = cputime_zero, \
  451. .stime = cputime_zero, \
  452. .sum_exec_runtime = 0, \
  453. }
  454. /*
  455. * Disable preemption until the scheduler is running.
  456. * Reset by start_kernel()->sched_init()->init_idle().
  457. *
  458. * We include PREEMPT_ACTIVE to avoid cond_resched() from working
  459. * before the scheduler is active -- see should_resched().
  460. */
  461. #define INIT_PREEMPT_COUNT (1 + PREEMPT_ACTIVE)
  462. /**
  463. * struct thread_group_cputimer - thread group interval timer counts
  464. * @cputime: thread group interval timers.
  465. * @running: non-zero when there are timers running and
  466. * @cputime receives updates.
  467. * @lock: lock for fields in this struct.
  468. *
  469. * This structure contains the version of task_cputime, above, that is
  470. * used for thread group CPU timer calculations.
  471. */
  472. struct thread_group_cputimer {
  473. struct task_cputime cputime;
  474. int running;
  475. spinlock_t lock;
  476. };
  477. /*
  478. * NOTE! "signal_struct" does not have it's own
  479. * locking, because a shared signal_struct always
  480. * implies a shared sighand_struct, so locking
  481. * sighand_struct is always a proper superset of
  482. * the locking of signal_struct.
  483. */
  484. struct signal_struct {
  485. atomic_t count;
  486. atomic_t live;
  487. wait_queue_head_t wait_chldexit; /* for wait4() */
  488. /* current thread group signal load-balancing target: */
  489. struct task_struct *curr_target;
  490. /* shared signal handling: */
  491. struct sigpending shared_pending;
  492. /* thread group exit support */
  493. int group_exit_code;
  494. /* overloaded:
  495. * - notify group_exit_task when ->count is equal to notify_count
  496. * - everyone except group_exit_task is stopped during signal delivery
  497. * of fatal signals, group_exit_task processes the signal.
  498. */
  499. int notify_count;
  500. struct task_struct *group_exit_task;
  501. /* thread group stop support, overloads group_exit_code too */
  502. int group_stop_count;
  503. unsigned int flags; /* see SIGNAL_* flags below */
  504. /* POSIX.1b Interval Timers */
  505. struct list_head posix_timers;
  506. /* ITIMER_REAL timer for the process */
  507. struct hrtimer real_timer;
  508. struct pid *leader_pid;
  509. ktime_t it_real_incr;
  510. /*
  511. * ITIMER_PROF and ITIMER_VIRTUAL timers for the process, we use
  512. * CPUCLOCK_PROF and CPUCLOCK_VIRT for indexing array as these
  513. * values are defined to 0 and 1 respectively
  514. */
  515. struct cpu_itimer it[2];
  516. /*
  517. * Thread group totals for process CPU timers.
  518. * See thread_group_cputimer(), et al, for details.
  519. */
  520. struct thread_group_cputimer cputimer;
  521. /* Earliest-expiration cache. */
  522. struct task_cputime cputime_expires;
  523. struct list_head cpu_timers[3];
  524. struct pid *tty_old_pgrp;
  525. /* boolean value for session group leader */
  526. int leader;
  527. struct tty_struct *tty; /* NULL if no tty */
  528. /*
  529. * Cumulative resource counters for dead threads in the group,
  530. * and for reaped dead child processes forked by this group.
  531. * Live threads maintain their own counters and add to these
  532. * in __exit_signal, except for the group leader.
  533. */
  534. cputime_t utime, stime, cutime, cstime;
  535. cputime_t gtime;
  536. cputime_t cgtime;
  537. #ifndef CONFIG_VIRT_CPU_ACCOUNTING
  538. cputime_t prev_utime, prev_stime;
  539. #endif
  540. unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
  541. unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
  542. unsigned long inblock, oublock, cinblock, coublock;
  543. unsigned long maxrss, cmaxrss;
  544. struct task_io_accounting ioac;
  545. /*
  546. * Cumulative ns of schedule CPU time fo dead threads in the
  547. * group, not including a zombie group leader, (This only differs
  548. * from jiffies_to_ns(utime + stime) if sched_clock uses something
  549. * other than jiffies.)
  550. */
  551. unsigned long long sum_sched_runtime;
  552. /*
  553. * We don't bother to synchronize most readers of this at all,
  554. * because there is no reader checking a limit that actually needs
  555. * to get both rlim_cur and rlim_max atomically, and either one
  556. * alone is a single word that can safely be read normally.
  557. * getrlimit/setrlimit use task_lock(current->group_leader) to
  558. * protect this instead of the siglock, because they really
  559. * have no need to disable irqs.
  560. */
  561. struct rlimit rlim[RLIM_NLIMITS];
  562. #ifdef CONFIG_BSD_PROCESS_ACCT
  563. struct pacct_struct pacct; /* per-process accounting information */
  564. #endif
  565. #ifdef CONFIG_TASKSTATS
  566. struct taskstats *stats;
  567. #endif
  568. #ifdef CONFIG_AUDIT
  569. unsigned audit_tty;
  570. struct tty_audit_buf *tty_audit_buf;
  571. #endif
  572. int oom_adj; /* OOM kill score adjustment (bit shift) */
  573. };
  574. /* Context switch must be unlocked if interrupts are to be enabled */
  575. #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
  576. # define __ARCH_WANT_UNLOCKED_CTXSW
  577. #endif
  578. /*
  579. * Bits in flags field of signal_struct.
  580. */
  581. #define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
  582. #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
  583. #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
  584. #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
  585. /*
  586. * Pending notifications to parent.
  587. */
  588. #define SIGNAL_CLD_STOPPED 0x00000010
  589. #define SIGNAL_CLD_CONTINUED 0x00000020
  590. #define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED)
  591. #define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
  592. /* If true, all threads except ->group_exit_task have pending SIGKILL */
  593. static inline int signal_group_exit(const struct signal_struct *sig)
  594. {
  595. return (sig->flags & SIGNAL_GROUP_EXIT) ||
  596. (sig->group_exit_task != NULL);
  597. }
  598. /*
  599. * Some day this will be a full-fledged user tracking system..
  600. */
  601. struct user_struct {
  602. atomic_t __count; /* reference count */
  603. atomic_t processes; /* How many processes does this user have? */
  604. atomic_t files; /* How many open files does this user have? */
  605. atomic_t sigpending; /* How many pending signals does this user have? */
  606. #ifdef CONFIG_INOTIFY_USER
  607. atomic_t inotify_watches; /* How many inotify watches does this user have? */
  608. atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
  609. #endif
  610. #ifdef CONFIG_EPOLL
  611. atomic_t epoll_watches; /* The number of file descriptors currently watched */
  612. #endif
  613. #ifdef CONFIG_POSIX_MQUEUE
  614. /* protected by mq_lock */
  615. unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
  616. #endif
  617. unsigned long locked_shm; /* How many pages of mlocked shm ? */
  618. #ifdef CONFIG_KEYS
  619. struct key *uid_keyring; /* UID specific keyring */
  620. struct key *session_keyring; /* UID's default session keyring */
  621. #endif
  622. /* Hash table maintenance information */
  623. struct hlist_node uidhash_node;
  624. uid_t uid;
  625. struct user_namespace *user_ns;
  626. #ifdef CONFIG_USER_SCHED
  627. struct task_group *tg;
  628. #ifdef CONFIG_SYSFS
  629. struct kobject kobj;
  630. struct delayed_work work;
  631. #endif
  632. #endif
  633. #ifdef CONFIG_PERF_EVENTS
  634. atomic_long_t locked_vm;
  635. #endif
  636. };
  637. extern int uids_sysfs_init(void);
  638. extern struct user_struct *find_user(uid_t);
  639. extern struct user_struct root_user;
  640. #define INIT_USER (&root_user)
  641. struct backing_dev_info;
  642. struct reclaim_state;
  643. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  644. struct sched_info {
  645. /* cumulative counters */
  646. unsigned long pcount; /* # of times run on this cpu */
  647. unsigned long long run_delay; /* time spent waiting on a runqueue */
  648. /* timestamps */
  649. unsigned long long last_arrival,/* when we last ran on a cpu */
  650. last_queued; /* when we were last queued to run */
  651. #ifdef CONFIG_SCHEDSTATS
  652. /* BKL stats */
  653. unsigned int bkl_count;
  654. #endif
  655. };
  656. #endif /* defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT) */
  657. #ifdef CONFIG_TASK_DELAY_ACCT
  658. struct task_delay_info {
  659. spinlock_t lock;
  660. unsigned int flags; /* Private per-task flags */
  661. /* For each stat XXX, add following, aligned appropriately
  662. *
  663. * struct timespec XXX_start, XXX_end;
  664. * u64 XXX_delay;
  665. * u32 XXX_count;
  666. *
  667. * Atomicity of updates to XXX_delay, XXX_count protected by
  668. * single lock above (split into XXX_lock if contention is an issue).
  669. */
  670. /*
  671. * XXX_count is incremented on every XXX operation, the delay
  672. * associated with the operation is added to XXX_delay.
  673. * XXX_delay contains the accumulated delay time in nanoseconds.
  674. */
  675. struct timespec blkio_start, blkio_end; /* Shared by blkio, swapin */
  676. u64 blkio_delay; /* wait for sync block io completion */
  677. u64 swapin_delay; /* wait for swapin block io completion */
  678. u32 blkio_count; /* total count of the number of sync block */
  679. /* io operations performed */
  680. u32 swapin_count; /* total count of the number of swapin block */
  681. /* io operations performed */
  682. struct timespec freepages_start, freepages_end;
  683. u64 freepages_delay; /* wait for memory reclaim */
  684. u32 freepages_count; /* total count of memory reclaim */
  685. };
  686. #endif /* CONFIG_TASK_DELAY_ACCT */
  687. static inline int sched_info_on(void)
  688. {
  689. #ifdef CONFIG_SCHEDSTATS
  690. return 1;
  691. #elif defined(CONFIG_TASK_DELAY_ACCT)
  692. extern int delayacct_on;
  693. return delayacct_on;
  694. #else
  695. return 0;
  696. #endif
  697. }
  698. enum cpu_idle_type {
  699. CPU_IDLE,
  700. CPU_NOT_IDLE,
  701. CPU_NEWLY_IDLE,
  702. CPU_MAX_IDLE_TYPES
  703. };
  704. /*
  705. * sched-domains (multiprocessor balancing) declarations:
  706. */
  707. /*
  708. * Increase resolution of nice-level calculations:
  709. */
  710. #define SCHED_LOAD_SHIFT 10
  711. #define SCHED_LOAD_SCALE (1L << SCHED_LOAD_SHIFT)
  712. #define SCHED_LOAD_SCALE_FUZZ SCHED_LOAD_SCALE
  713. #ifdef CONFIG_SMP
  714. #define SD_LOAD_BALANCE 0x0001 /* Do load balancing on this domain. */
  715. #define SD_BALANCE_NEWIDLE 0x0002 /* Balance when about to become idle */
  716. #define SD_BALANCE_EXEC 0x0004 /* Balance on exec */
  717. #define SD_BALANCE_FORK 0x0008 /* Balance on fork, clone */
  718. #define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
  719. #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
  720. #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
  721. #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
  722. #define SD_POWERSAVINGS_BALANCE 0x0100 /* Balance for power savings */
  723. #define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
  724. #define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
  725. #define SD_PREFER_SIBLING 0x1000 /* Prefer to place tasks in a sibling domain */
  726. enum powersavings_balance_level {
  727. POWERSAVINGS_BALANCE_NONE = 0, /* No power saving load balance */
  728. POWERSAVINGS_BALANCE_BASIC, /* Fill one thread/core/package
  729. * first for long running threads
  730. */
  731. POWERSAVINGS_BALANCE_WAKEUP, /* Also bias task wakeups to semi-idle
  732. * cpu package for power savings
  733. */
  734. MAX_POWERSAVINGS_BALANCE_LEVELS
  735. };
  736. extern int sched_mc_power_savings, sched_smt_power_savings;
  737. static inline int sd_balance_for_mc_power(void)
  738. {
  739. if (sched_smt_power_savings)
  740. return SD_POWERSAVINGS_BALANCE;
  741. return SD_PREFER_SIBLING;
  742. }
  743. static inline int sd_balance_for_package_power(void)
  744. {
  745. if (sched_mc_power_savings | sched_smt_power_savings)
  746. return SD_POWERSAVINGS_BALANCE;
  747. return SD_PREFER_SIBLING;
  748. }
  749. /*
  750. * Optimise SD flags for power savings:
  751. * SD_BALANCE_NEWIDLE helps agressive task consolidation and power savings.
  752. * Keep default SD flags if sched_{smt,mc}_power_saving=0
  753. */
  754. static inline int sd_power_saving_flags(void)
  755. {
  756. if (sched_mc_power_savings | sched_smt_power_savings)
  757. return SD_BALANCE_NEWIDLE;
  758. return 0;
  759. }
  760. struct sched_group {
  761. struct sched_group *next; /* Must be a circular list */
  762. /*
  763. * CPU power of this group, SCHED_LOAD_SCALE being max power for a
  764. * single CPU.
  765. */
  766. unsigned int cpu_power;
  767. /*
  768. * The CPUs this group covers.
  769. *
  770. * NOTE: this field is variable length. (Allocated dynamically
  771. * by attaching extra space to the end of the structure,
  772. * depending on how many CPUs the kernel has booted up with)
  773. *
  774. * It is also be embedded into static data structures at build
  775. * time. (See 'struct static_sched_group' in kernel/sched.c)
  776. */
  777. unsigned long cpumask[0];
  778. };
  779. static inline struct cpumask *sched_group_cpus(struct sched_group *sg)
  780. {
  781. return to_cpumask(sg->cpumask);
  782. }
  783. enum sched_domain_level {
  784. SD_LV_NONE = 0,
  785. SD_LV_SIBLING,
  786. SD_LV_MC,
  787. SD_LV_CPU,
  788. SD_LV_NODE,
  789. SD_LV_ALLNODES,
  790. SD_LV_MAX
  791. };
  792. struct sched_domain_attr {
  793. int relax_domain_level;
  794. };
  795. #define SD_ATTR_INIT (struct sched_domain_attr) { \
  796. .relax_domain_level = -1, \
  797. }
  798. struct sched_domain {
  799. /* These fields must be setup */
  800. struct sched_domain *parent; /* top domain must be null terminated */
  801. struct sched_domain *child; /* bottom domain must be null terminated */
  802. struct sched_group *groups; /* the balancing groups of the domain */
  803. unsigned long min_interval; /* Minimum balance interval ms */
  804. unsigned long max_interval; /* Maximum balance interval ms */
  805. unsigned int busy_factor; /* less balancing by factor if busy */
  806. unsigned int imbalance_pct; /* No balance until over watermark */
  807. unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
  808. unsigned int busy_idx;
  809. unsigned int idle_idx;
  810. unsigned int newidle_idx;
  811. unsigned int wake_idx;
  812. unsigned int forkexec_idx;
  813. unsigned int smt_gain;
  814. int flags; /* See SD_* */
  815. enum sched_domain_level level;
  816. /* Runtime fields. */
  817. unsigned long last_balance; /* init to jiffies. units in jiffies */
  818. unsigned int balance_interval; /* initialise to 1. units in ms. */
  819. unsigned int nr_balance_failed; /* initialise to 0 */
  820. u64 last_update;
  821. #ifdef CONFIG_SCHEDSTATS
  822. /* load_balance() stats */
  823. unsigned int lb_count[CPU_MAX_IDLE_TYPES];
  824. unsigned int lb_failed[CPU_MAX_IDLE_TYPES];
  825. unsigned int lb_balanced[CPU_MAX_IDLE_TYPES];
  826. unsigned int lb_imbalance[CPU_MAX_IDLE_TYPES];
  827. unsigned int lb_gained[CPU_MAX_IDLE_TYPES];
  828. unsigned int lb_hot_gained[CPU_MAX_IDLE_TYPES];
  829. unsigned int lb_nobusyg[CPU_MAX_IDLE_TYPES];
  830. unsigned int lb_nobusyq[CPU_MAX_IDLE_TYPES];
  831. /* Active load balancing */
  832. unsigned int alb_count;
  833. unsigned int alb_failed;
  834. unsigned int alb_pushed;
  835. /* SD_BALANCE_EXEC stats */
  836. unsigned int sbe_count;
  837. unsigned int sbe_balanced;
  838. unsigned int sbe_pushed;
  839. /* SD_BALANCE_FORK stats */
  840. unsigned int sbf_count;
  841. unsigned int sbf_balanced;
  842. unsigned int sbf_pushed;
  843. /* try_to_wake_up() stats */
  844. unsigned int ttwu_wake_remote;
  845. unsigned int ttwu_move_affine;
  846. unsigned int ttwu_move_balance;
  847. #endif
  848. #ifdef CONFIG_SCHED_DEBUG
  849. char *name;
  850. #endif
  851. /*
  852. * Span of all CPUs in this domain.
  853. *
  854. * NOTE: this field is variable length. (Allocated dynamically
  855. * by attaching extra space to the end of the structure,
  856. * depending on how many CPUs the kernel has booted up with)
  857. *
  858. * It is also be embedded into static data structures at build
  859. * time. (See 'struct static_sched_domain' in kernel/sched.c)
  860. */
  861. unsigned long span[0];
  862. };
  863. static inline struct cpumask *sched_domain_span(struct sched_domain *sd)
  864. {
  865. return to_cpumask(sd->span);
  866. }
  867. extern void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
  868. struct sched_domain_attr *dattr_new);
  869. /* Allocate an array of sched domains, for partition_sched_domains(). */
  870. cpumask_var_t *alloc_sched_domains(unsigned int ndoms);
  871. void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms);
  872. /* Test a flag in parent sched domain */
  873. static inline int test_sd_parent(struct sched_domain *sd, int flag)
  874. {
  875. if (sd->parent && (sd->parent->flags & flag))
  876. return 1;
  877. return 0;
  878. }
  879. unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu);
  880. unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu);
  881. #else /* CONFIG_SMP */
  882. struct sched_domain_attr;
  883. static inline void
  884. partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
  885. struct sched_domain_attr *dattr_new)
  886. {
  887. }
  888. #endif /* !CONFIG_SMP */
  889. struct io_context; /* See blkdev.h */
  890. #ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
  891. extern void prefetch_stack(struct task_struct *t);
  892. #else
  893. static inline void prefetch_stack(struct task_struct *t) { }
  894. #endif
  895. struct audit_context; /* See audit.c */
  896. struct mempolicy;
  897. struct pipe_inode_info;
  898. struct uts_namespace;
  899. struct rq;
  900. struct sched_domain;
  901. /*
  902. * wake flags
  903. */
  904. #define WF_SYNC 0x01 /* waker goes to sleep after wakup */
  905. #define WF_FORK 0x02 /* child wakeup after fork */
  906. struct sched_class {
  907. const struct sched_class *next;
  908. void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
  909. void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
  910. void (*yield_task) (struct rq *rq);
  911. void (*check_preempt_curr) (struct rq *rq, struct task_struct *p, int flags);
  912. struct task_struct * (*pick_next_task) (struct rq *rq);
  913. void (*put_prev_task) (struct rq *rq, struct task_struct *p);
  914. #ifdef CONFIG_SMP
  915. int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags);
  916. unsigned long (*load_balance) (struct rq *this_rq, int this_cpu,
  917. struct rq *busiest, unsigned long max_load_move,
  918. struct sched_domain *sd, enum cpu_idle_type idle,
  919. int *all_pinned, int *this_best_prio);
  920. int (*move_one_task) (struct rq *this_rq, int this_cpu,
  921. struct rq *busiest, struct sched_domain *sd,
  922. enum cpu_idle_type idle);
  923. void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
  924. void (*post_schedule) (struct rq *this_rq);
  925. void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
  926. void (*set_cpus_allowed)(struct task_struct *p,
  927. const struct cpumask *newmask);
  928. void (*rq_online)(struct rq *rq);
  929. void (*rq_offline)(struct rq *rq);
  930. #endif
  931. void (*set_curr_task) (struct rq *rq);
  932. void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
  933. void (*task_fork) (struct task_struct *p);
  934. void (*switched_from) (struct rq *this_rq, struct task_struct *task,
  935. int running);
  936. void (*switched_to) (struct rq *this_rq, struct task_struct *task,
  937. int running);
  938. void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
  939. int oldprio, int running);
  940. unsigned int (*get_rr_interval) (struct rq *rq,
  941. struct task_struct *task);
  942. #ifdef CONFIG_FAIR_GROUP_SCHED
  943. void (*moved_group) (struct task_struct *p);
  944. #endif
  945. };
  946. struct load_weight {
  947. unsigned long weight, inv_weight;
  948. };
  949. /*
  950. * CFS stats for a schedulable entity (task, task-group etc)
  951. *
  952. * Current field usage histogram:
  953. *
  954. * 4 se->block_start
  955. * 4 se->run_node
  956. * 4 se->sleep_start
  957. * 6 se->load.weight
  958. */
  959. struct sched_entity {
  960. struct load_weight load; /* for load-balancing */
  961. struct rb_node run_node;
  962. struct list_head group_node;
  963. unsigned int on_rq;
  964. u64 exec_start;
  965. u64 sum_exec_runtime;
  966. u64 vruntime;
  967. u64 prev_sum_exec_runtime;
  968. u64 last_wakeup;
  969. u64 avg_overlap;
  970. u64 nr_migrations;
  971. u64 start_runtime;
  972. u64 avg_wakeup;
  973. #ifdef CONFIG_SCHEDSTATS
  974. u64 wait_start;
  975. u64 wait_max;
  976. u64 wait_count;
  977. u64 wait_sum;
  978. u64 iowait_count;
  979. u64 iowait_sum;
  980. u64 sleep_start;
  981. u64 sleep_max;
  982. s64 sum_sleep_runtime;
  983. u64 block_start;
  984. u64 block_max;
  985. u64 exec_max;
  986. u64 slice_max;
  987. u64 nr_migrations_cold;
  988. u64 nr_failed_migrations_affine;
  989. u64 nr_failed_migrations_running;
  990. u64 nr_failed_migrations_hot;
  991. u64 nr_forced_migrations;
  992. u64 nr_wakeups;
  993. u64 nr_wakeups_sync;
  994. u64 nr_wakeups_migrate;
  995. u64 nr_wakeups_local;
  996. u64 nr_wakeups_remote;
  997. u64 nr_wakeups_affine;
  998. u64 nr_wakeups_affine_attempts;
  999. u64 nr_wakeups_passive;
  1000. u64 nr_wakeups_idle;
  1001. #endif
  1002. #ifdef CONFIG_FAIR_GROUP_SCHED
  1003. struct sched_entity *parent;
  1004. /* rq on which this entity is (to be) queued: */
  1005. struct cfs_rq *cfs_rq;
  1006. /* rq "owned" by this entity/group: */
  1007. struct cfs_rq *my_q;
  1008. #endif
  1009. };
  1010. struct sched_rt_entity {
  1011. struct list_head run_list;
  1012. unsigned long timeout;
  1013. unsigned int time_slice;
  1014. int nr_cpus_allowed;
  1015. struct sched_rt_entity *back;
  1016. #ifdef CONFIG_RT_GROUP_SCHED
  1017. struct sched_rt_entity *parent;
  1018. /* rq on which this entity is (to be) queued: */
  1019. struct rt_rq *rt_rq;
  1020. /* rq "owned" by this entity/group: */
  1021. struct rt_rq *my_q;
  1022. #endif
  1023. };
  1024. struct rcu_node;
  1025. struct task_struct {
  1026. volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
  1027. void *stack;
  1028. atomic_t usage;
  1029. unsigned int flags; /* per process flags, defined below */
  1030. unsigned int ptrace;
  1031. int lock_depth; /* BKL lock depth */
  1032. #ifdef CONFIG_SMP
  1033. #ifdef __ARCH_WANT_UNLOCKED_CTXSW
  1034. int oncpu;
  1035. #endif
  1036. #endif
  1037. int prio, static_prio, normal_prio;
  1038. unsigned int rt_priority;
  1039. const struct sched_class *sched_class;
  1040. struct sched_entity se;
  1041. struct sched_rt_entity rt;
  1042. #ifdef CONFIG_PREEMPT_NOTIFIERS
  1043. /* list of struct preempt_notifier: */
  1044. struct hlist_head preempt_notifiers;
  1045. #endif
  1046. /*
  1047. * fpu_counter contains the number of consecutive context switches
  1048. * that the FPU is used. If this is over a threshold, the lazy fpu
  1049. * saving becomes unlazy to save the trap. This is an unsigned char
  1050. * so that after 256 times the counter wraps and the behavior turns
  1051. * lazy again; this to deal with bursty apps that only use FPU for
  1052. * a short time
  1053. */
  1054. unsigned char fpu_counter;
  1055. #ifdef CONFIG_BLK_DEV_IO_TRACE
  1056. unsigned int btrace_seq;
  1057. #endif
  1058. unsigned int policy;
  1059. cpumask_t cpus_allowed;
  1060. #ifdef CONFIG_TREE_PREEMPT_RCU
  1061. int rcu_read_lock_nesting;
  1062. char rcu_read_unlock_special;
  1063. struct rcu_node *rcu_blocked_node;
  1064. struct list_head rcu_node_entry;
  1065. #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
  1066. #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
  1067. struct sched_info sched_info;
  1068. #endif
  1069. struct list_head tasks;
  1070. struct plist_node pushable_tasks;
  1071. struct mm_struct *mm, *active_mm;
  1072. /* task state */
  1073. int exit_state;
  1074. int exit_code, exit_signal;
  1075. int pdeath_signal; /* The signal sent when the parent dies */
  1076. /* ??? */
  1077. unsigned int personality;
  1078. unsigned did_exec:1;
  1079. unsigned in_execve:1; /* Tell the LSMs that the process is doing an
  1080. * execve */
  1081. unsigned in_iowait:1;
  1082. /* Revert to default priority/policy when forking */
  1083. unsigned sched_reset_on_fork:1;
  1084. pid_t pid;
  1085. pid_t tgid;
  1086. #ifdef CONFIG_CC_STACKPROTECTOR
  1087. /* Canary value for the -fstack-protector gcc feature */
  1088. unsigned long stack_canary;
  1089. #endif
  1090. /*
  1091. * pointers to (original) parent process, youngest child, younger sibling,
  1092. * older sibling, respectively. (p->father can be replaced with
  1093. * p->real_parent->pid)
  1094. */
  1095. struct task_struct *real_parent; /* real parent process */
  1096. struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
  1097. /*
  1098. * children/sibling forms the list of my natural children
  1099. */
  1100. struct list_head children; /* list of my children */
  1101. struct list_head sibling; /* linkage in my parent's children list */
  1102. struct task_struct *group_leader; /* threadgroup leader */
  1103. /*
  1104. * ptraced is the list of tasks this task is using ptrace on.
  1105. * This includes both natural children and PTRACE_ATTACH targets.
  1106. * p->ptrace_entry is p's link on the p->parent->ptraced list.
  1107. */
  1108. struct list_head ptraced;
  1109. struct list_head ptrace_entry;
  1110. /*
  1111. * This is the tracer handle for the ptrace BTS extension.
  1112. * This field actually belongs to the ptracer task.
  1113. */
  1114. struct bts_context *bts;
  1115. /* PID/PID hash table linkage. */
  1116. struct pid_link pids[PIDTYPE_MAX];
  1117. struct list_head thread_group;
  1118. struct completion *vfork_done; /* for vfork() */
  1119. int __user *set_child_tid; /* CLONE_CHILD_SETTID */
  1120. int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
  1121. cputime_t utime, stime, utimescaled, stimescaled;
  1122. cputime_t gtime;
  1123. #ifndef CONFIG_VIRT_CPU_ACCOUNTING
  1124. cputime_t prev_utime, prev_stime;
  1125. #endif
  1126. unsigned long nvcsw, nivcsw; /* context switch counts */
  1127. struct timespec start_time; /* monotonic time */
  1128. struct timespec real_start_time; /* boot based time */
  1129. /* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
  1130. unsigned long min_flt, maj_flt;
  1131. struct task_cputime cputime_expires;
  1132. struct list_head cpu_timers[3];
  1133. /* process credentials */
  1134. const struct cred *real_cred; /* objective and real subjective task
  1135. * credentials (COW) */
  1136. const struct cred *cred; /* effective (overridable) subjective task
  1137. * credentials (COW) */
  1138. struct mutex cred_guard_mutex; /* guard against foreign influences on
  1139. * credential calculations
  1140. * (notably. ptrace) */
  1141. struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */
  1142. char comm[TASK_COMM_LEN]; /* executable name excluding path
  1143. - access with [gs]et_task_comm (which lock
  1144. it with task_lock())
  1145. - initialized normally by flush_old_exec */
  1146. /* file system info */
  1147. int link_count, total_link_count;
  1148. #ifdef CONFIG_SYSVIPC
  1149. /* ipc stuff */
  1150. struct sysv_sem sysvsem;
  1151. #endif
  1152. #ifdef CONFIG_DETECT_HUNG_TASK
  1153. /* hung task detection */
  1154. unsigned long last_switch_count;
  1155. #endif
  1156. /* CPU-specific state of this task */
  1157. struct thread_struct thread;
  1158. /* filesystem information */
  1159. struct fs_struct *fs;
  1160. /* open file information */
  1161. struct files_struct *files;
  1162. /* namespaces */
  1163. struct nsproxy *nsproxy;
  1164. /* signal handlers */
  1165. struct signal_struct *signal;
  1166. struct sighand_struct *sighand;
  1167. sigset_t blocked, real_blocked;
  1168. sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
  1169. struct sigpending pending;
  1170. unsigned long sas_ss_sp;
  1171. size_t sas_ss_size;
  1172. int (*notifier)(void *priv);
  1173. void *notifier_data;
  1174. sigset_t *notifier_mask;
  1175. struct audit_context *audit_context;
  1176. #ifdef CONFIG_AUDITSYSCALL
  1177. uid_t loginuid;
  1178. unsigned int sessionid;
  1179. #endif
  1180. seccomp_t seccomp;
  1181. /* Thread group tracking */
  1182. u32 parent_exec_id;
  1183. u32 self_exec_id;
  1184. /* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
  1185. * mempolicy */
  1186. spinlock_t alloc_lock;
  1187. #ifdef CONFIG_GENERIC_HARDIRQS
  1188. /* IRQ handler threads */
  1189. struct irqaction *irqaction;
  1190. #endif
  1191. /* Protection of the PI data structures: */
  1192. raw_spinlock_t pi_lock;
  1193. #ifdef CONFIG_RT_MUTEXES
  1194. /* PI waiters blocked on a rt_mutex held by this task */
  1195. struct plist_head pi_waiters;
  1196. /* Deadlock detection and priority inheritance handling */
  1197. struct rt_mutex_waiter *pi_blocked_on;
  1198. #endif
  1199. #ifdef CONFIG_DEBUG_MUTEXES
  1200. /* mutex deadlock detection */
  1201. struct mutex_waiter *blocked_on;
  1202. #endif
  1203. #ifdef CONFIG_TRACE_IRQFLAGS
  1204. unsigned int irq_events;
  1205. unsigned long hardirq_enable_ip;
  1206. unsigned long hardirq_disable_ip;
  1207. unsigned int hardirq_enable_event;
  1208. unsigned int hardirq_disable_event;
  1209. int hardirqs_enabled;
  1210. int hardirq_context;
  1211. unsigned long softirq_disable_ip;
  1212. unsigned long softirq_enable_ip;
  1213. unsigned int softirq_disable_event;
  1214. unsigned int softirq_enable_event;
  1215. int softirqs_enabled;
  1216. int softirq_context;
  1217. #endif
  1218. #ifdef CONFIG_LOCKDEP
  1219. # define MAX_LOCK_DEPTH 48UL
  1220. u64 curr_chain_key;
  1221. int lockdep_depth;
  1222. unsigned int lockdep_recursion;
  1223. struct held_lock held_locks[MAX_LOCK_DEPTH];
  1224. gfp_t lockdep_reclaim_gfp;
  1225. #endif
  1226. /* journalling filesystem info */
  1227. void *journal_info;
  1228. /* stacked block device info */
  1229. struct bio *bio_list, **bio_tail;
  1230. /* VM state */
  1231. struct reclaim_state *reclaim_state;
  1232. struct backing_dev_info *backing_dev_info;
  1233. struct io_context *io_context;
  1234. unsigned long ptrace_message;
  1235. siginfo_t *last_siginfo; /* For ptrace use. */
  1236. struct task_io_accounting ioac;
  1237. #if defined(CONFIG_TASK_XACCT)
  1238. u64 acct_rss_mem1; /* accumulated rss usage */
  1239. u64 acct_vm_mem1; /* accumulated virtual memory usage */
  1240. cputime_t acct_timexpd; /* stime + utime since last update */
  1241. #endif
  1242. #ifdef CONFIG_CPUSETS
  1243. nodemask_t mems_allowed; /* Protected by alloc_lock */
  1244. int cpuset_mem_spread_rotor;
  1245. #endif
  1246. #ifdef CONFIG_CGROUPS
  1247. /* Control Group info protected by css_set_lock */
  1248. struct css_set *cgroups;
  1249. /* cg_list protected by css_set_lock and tsk->alloc_lock */
  1250. struct list_head cg_list;
  1251. #endif
  1252. #ifdef CONFIG_FUTEX
  1253. struct robust_list_head __user *robust_list;
  1254. #ifdef CONFIG_COMPAT
  1255. struct compat_robust_list_head __user *compat_robust_list;
  1256. #endif
  1257. struct list_head pi_state_list;
  1258. struct futex_pi_state *pi_state_cache;
  1259. #endif
  1260. #ifdef CONFIG_PERF_EVENTS
  1261. struct perf_event_context *perf_event_ctxp;
  1262. struct mutex perf_event_mutex;
  1263. struct list_head perf_event_list;
  1264. #endif
  1265. #ifdef CONFIG_NUMA
  1266. struct mempolicy *mempolicy; /* Protected by alloc_lock */
  1267. short il_next;
  1268. #endif
  1269. atomic_t fs_excl; /* holding fs exclusive resources */
  1270. struct rcu_head rcu;
  1271. /*
  1272. * cache last used pipe for splice
  1273. */
  1274. struct pipe_inode_info *splice_pipe;
  1275. #ifdef CONFIG_TASK_DELAY_ACCT
  1276. struct task_delay_info *delays;
  1277. #endif
  1278. #ifdef CONFIG_FAULT_INJECTION
  1279. int make_it_fail;
  1280. #endif
  1281. struct prop_local_single dirties;
  1282. #ifdef CONFIG_LATENCYTOP
  1283. int latency_record_count;
  1284. struct latency_record latency_record[LT_SAVECOUNT];
  1285. #endif
  1286. /*
  1287. * time slack values; these are used to round up poll() and
  1288. * select() etc timeout values. These are in nanoseconds.
  1289. */
  1290. unsigned long timer_slack_ns;
  1291. unsigned long default_timer_slack_ns;
  1292. struct list_head *scm_work_list;
  1293. #ifdef CONFIG_FUNCTION_GRAPH_TRACER
  1294. /* Index of current stored adress in ret_stack */
  1295. int curr_ret_stack;
  1296. /* Stack of return addresses for return function tracing */
  1297. struct ftrace_ret_stack *ret_stack;
  1298. /* time stamp for last schedule */
  1299. unsigned long long ftrace_timestamp;
  1300. /*
  1301. * Number of functions that haven't been traced
  1302. * because of depth overrun.
  1303. */
  1304. atomic_t trace_overrun;
  1305. /* Pause for the tracing */
  1306. atomic_t tracing_graph_pause;
  1307. #endif
  1308. #ifdef CONFIG_TRACING
  1309. /* state flags for use by tracers */
  1310. unsigned long trace;
  1311. /* bitmask of trace recursion */
  1312. unsigned long trace_recursion;
  1313. #endif /* CONFIG_TRACING */
  1314. unsigned long stack_start;
  1315. #ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
  1316. struct memcg_batch_info {
  1317. int do_batch; /* incremented when batch uncharge started */
  1318. struct mem_cgroup *memcg; /* target memcg of uncharge */
  1319. unsigned long bytes; /* uncharged usage */
  1320. unsigned long memsw_bytes; /* uncharged mem+swap usage */
  1321. } memcg_batch;
  1322. #endif
  1323. };
  1324. /* Future-safe accessor for struct task_struct's cpus_allowed. */
  1325. #define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
  1326. /*
  1327. * Priority of a process goes from 0..MAX_PRIO-1, valid RT
  1328. * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
  1329. * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
  1330. * values are inverted: lower p->prio value means higher priority.
  1331. *
  1332. * The MAX_USER_RT_PRIO value allows the actual maximum
  1333. * RT priority to be separate from the value exported to
  1334. * user-space. This allows kernel threads to set their
  1335. * priority to a value higher than any user task. Note:
  1336. * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
  1337. */
  1338. #define MAX_USER_RT_PRIO 100
  1339. #define MAX_RT_PRIO MAX_USER_RT_PRIO
  1340. #define MAX_PRIO (MAX_RT_PRIO + 40)
  1341. #define DEFAULT_PRIO (MAX_RT_PRIO + 20)
  1342. static inline int rt_prio(int prio)
  1343. {
  1344. if (unlikely(prio < MAX_RT_PRIO))
  1345. return 1;
  1346. return 0;
  1347. }
  1348. static inline int rt_task(struct task_struct *p)
  1349. {
  1350. return rt_prio(p->prio);
  1351. }
  1352. static inline struct pid *task_pid(struct task_struct *task)
  1353. {
  1354. return task->pids[PIDTYPE_PID].pid;
  1355. }
  1356. static inline struct pid *task_tgid(struct task_struct *task)
  1357. {
  1358. return task->group_leader->pids[PIDTYPE_PID].pid;
  1359. }
  1360. /*
  1361. * Without tasklist or rcu lock it is not safe to dereference
  1362. * the result of task_pgrp/task_session even if task == current,
  1363. * we can race with another thread doing sys_setsid/sys_setpgid.
  1364. */
  1365. static inline struct pid *task_pgrp(struct task_struct *task)
  1366. {
  1367. return task->group_leader->pids[PIDTYPE_PGID].pid;
  1368. }
  1369. static inline struct pid *task_session(struct task_struct *task)
  1370. {
  1371. return task->group_leader->pids[PIDTYPE_SID].pid;
  1372. }
  1373. struct pid_namespace;
  1374. /*
  1375. * the helpers to get the task's different pids as they are seen
  1376. * from various namespaces
  1377. *
  1378. * task_xid_nr() : global id, i.e. the id seen from the init namespace;
  1379. * task_xid_vnr() : virtual id, i.e. the id seen from the pid namespace of
  1380. * current.
  1381. * task_xid_nr_ns() : id seen from the ns specified;
  1382. *
  1383. * set_task_vxid() : assigns a virtual id to a task;
  1384. *
  1385. * see also pid_nr() etc in include/linux/pid.h
  1386. */
  1387. pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
  1388. struct pid_namespace *ns);
  1389. static inline pid_t task_pid_nr(struct task_struct *tsk)
  1390. {
  1391. return tsk->pid;
  1392. }
  1393. static inline pid_t task_pid_nr_ns(struct task_struct *tsk,
  1394. struct pid_namespace *ns)
  1395. {
  1396. return __task_pid_nr_ns(tsk, PIDTYPE_PID, ns);
  1397. }
  1398. static inline pid_t task_pid_vnr(struct task_struct *tsk)
  1399. {
  1400. return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
  1401. }
  1402. static inline pid_t task_tgid_nr(struct task_struct *tsk)
  1403. {
  1404. return tsk->tgid;
  1405. }
  1406. pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
  1407. static inline pid_t task_tgid_vnr(struct task_struct *tsk)
  1408. {
  1409. return pid_vnr(task_tgid(tsk));
  1410. }
  1411. static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk,
  1412. struct pid_namespace *ns)
  1413. {
  1414. return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
  1415. }
  1416. static inline pid_t task_pgrp_vnr(struct task_struct *tsk)
  1417. {
  1418. return __task_pid_nr_ns(tsk, PIDTYPE_PGID, NULL);
  1419. }
  1420. static inline pid_t task_session_nr_ns(struct task_struct *tsk,
  1421. struct pid_namespace *ns)
  1422. {
  1423. return __task_pid_nr_ns(tsk, PIDTYPE_SID, ns);
  1424. }
  1425. static inline pid_t task_session_vnr(struct task_struct *tsk)
  1426. {
  1427. return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
  1428. }
  1429. /* obsolete, do not use */
  1430. static inline pid_t task_pgrp_nr(struct task_struct *tsk)
  1431. {
  1432. return task_pgrp_nr_ns(tsk, &init_pid_ns);
  1433. }
  1434. /**
  1435. * pid_alive - check that a task structure is not stale
  1436. * @p: Task structure to be checked.
  1437. *
  1438. * Test if a process is not yet dead (at most zombie state)
  1439. * If pid_alive fails, then pointers within the task structure
  1440. * can be stale and must not be dereferenced.
  1441. */
  1442. static inline int pid_alive(struct task_struct *p)
  1443. {
  1444. return p->pids[PIDTYPE_PID].pid != NULL;
  1445. }
  1446. /**
  1447. * is_global_init - check if a task structure is init
  1448. * @tsk: Task structure to be checked.
  1449. *
  1450. * Check if a task structure is the first user space task the kernel created.
  1451. */
  1452. static inline int is_global_init(struct task_struct *tsk)
  1453. {
  1454. return tsk->pid == 1;
  1455. }
  1456. /*
  1457. * is_container_init:
  1458. * check whether in the task is init in its own pid namespace.
  1459. */
  1460. extern int is_container_init(struct task_struct *tsk);
  1461. extern struct pid *cad_pid;
  1462. extern void free_task(struct task_struct *tsk);
  1463. #define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
  1464. extern void __put_task_struct(struct task_struct *t);
  1465. static inline void put_task_struct(struct task_struct *t)
  1466. {
  1467. if (atomic_dec_and_test(&t->usage))
  1468. __put_task_struct(t);
  1469. }
  1470. extern void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
  1471. extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st);
  1472. /*
  1473. * Per process flags
  1474. */
  1475. #define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
  1476. /* Not implemented yet, only for 486*/
  1477. #define PF_STARTING 0x00000002 /* being created */
  1478. #define PF_EXITING 0x00000004 /* getting shut down */
  1479. #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */
  1480. #define PF_VCPU 0x00000010 /* I'm a virtual CPU */
  1481. #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
  1482. #define PF_MCE_PROCESS 0x00000080 /* process policy on mce errors */
  1483. #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
  1484. #define PF_DUMPCORE 0x00000200 /* dumped core */
  1485. #define PF_SIGNALED 0x00000400 /* killed by a signal */
  1486. #define PF_MEMALLOC 0x00000800 /* Allocating memory */
  1487. #define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
  1488. #define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
  1489. #define PF_FREEZING 0x00004000 /* freeze in progress. do not account to load */
  1490. #define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
  1491. #define PF_FROZEN 0x00010000 /* frozen for system suspend */
  1492. #define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
  1493. #define PF_KSWAPD 0x00040000 /* I am kswapd */
  1494. #define PF_OOM_ORIGIN 0x00080000 /* Allocating much memory to others */
  1495. #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
  1496. #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
  1497. #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
  1498. #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
  1499. #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
  1500. #define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
  1501. #define PF_THREAD_BOUND 0x04000000 /* Thread bound to specific cpu */
  1502. #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
  1503. #define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
  1504. #define PF_MUTEX_TESTER 0x20000000 /* Thread belongs to the rt mutex tester */
  1505. #define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezeable */
  1506. #define PF_FREEZER_NOSIG 0x80000000 /* Freezer won't send signals to it */
  1507. /*
  1508. * Only the _current_ task can read/write to tsk->flags, but other
  1509. * tasks can access tsk->flags in readonly mode for example
  1510. * with tsk_used_math (like during threaded core dumping).
  1511. * There is however an exception to this rule during ptrace
  1512. * or during fork: the ptracer task is allowed to write to the
  1513. * child->flags of its traced child (same goes for fork, the parent
  1514. * can write to the child->flags), because we're guaranteed the
  1515. * child is not running and in turn not changing child->flags
  1516. * at the same time the parent does it.
  1517. */
  1518. #define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
  1519. #define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
  1520. #define clear_used_math() clear_stopped_child_used_math(current)
  1521. #define set_used_math() set_stopped_child_used_math(current)
  1522. #define conditional_stopped_child_used_math(condition, child) \
  1523. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
  1524. #define conditional_used_math(condition) \
  1525. conditional_stopped_child_used_math(condition, current)
  1526. #define copy_to_stopped_child_used_math(child) \
  1527. do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
  1528. /* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
  1529. #define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
  1530. #define used_math() tsk_used_math(current)
  1531. #ifdef CONFIG_TREE_PREEMPT_RCU
  1532. #define RCU_READ_UNLOCK_BLOCKED (1 << 0) /* blocked while in RCU read-side. */
  1533. #define RCU_READ_UNLOCK_NEED_QS (1 << 1) /* RCU core needs CPU response. */
  1534. static inline void rcu_copy_process(struct task_struct *p)
  1535. {
  1536. p->rcu_read_lock_nesting = 0;
  1537. p->rcu_read_unlock_special = 0;
  1538. p->rcu_blocked_node = NULL;
  1539. INIT_LIST_HEAD(&p->rcu_node_entry);
  1540. }
  1541. #else
  1542. static inline void rcu_copy_process(struct task_struct *p)
  1543. {
  1544. }
  1545. #endif
  1546. #ifdef CONFIG_SMP
  1547. extern int set_cpus_allowed_ptr(struct task_struct *p,
  1548. const struct cpumask *new_mask);
  1549. #else
  1550. static inline int set_cpus_allowed_ptr(struct task_struct *p,
  1551. const struct cpumask *new_mask)
  1552. {
  1553. if (!cpumask_test_cpu(0, new_mask))
  1554. return -EINVAL;
  1555. return 0;
  1556. }
  1557. #endif
  1558. #ifndef CONFIG_CPUMASK_OFFSTACK
  1559. static inline int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
  1560. {
  1561. return set_cpus_allowed_ptr(p, &new_mask);
  1562. }
  1563. #endif
  1564. /*
  1565. * Architectures can set this to 1 if they have specified
  1566. * CONFIG_HAVE_UNSTABLE_SCHED_CLOCK in their arch Kconfig,
  1567. * but then during bootup it turns out that sched_clock()
  1568. * is reliable after all:
  1569. */
  1570. #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
  1571. extern int sched_clock_stable;
  1572. #endif
  1573. /* ftrace calls sched_clock() directly */
  1574. extern unsigned long long notrace sched_clock(void);
  1575. extern void sched_clock_init(void);
  1576. extern u64 sched_clock_cpu(int cpu);
  1577. #ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
  1578. static inline void sched_clock_tick(void)
  1579. {
  1580. }
  1581. static inline void sched_clock_idle_sleep_event(void)
  1582. {
  1583. }
  1584. static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
  1585. {
  1586. }
  1587. #else
  1588. extern void sched_clock_tick(void);
  1589. extern void sched_clock_idle_sleep_event(void);
  1590. extern void sched_clock_idle_wakeup_event(u64 delta_ns);
  1591. #endif
  1592. /*
  1593. * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
  1594. * clock constructed from sched_clock():
  1595. */
  1596. extern unsigned long long cpu_clock(int cpu);
  1597. extern unsigned long long
  1598. task_sched_runtime(struct task_struct *task);
  1599. extern unsigned long long thread_group_sched_runtime(struct task_struct *task);
  1600. /* sched_exec is called by processes performing an exec */
  1601. #ifdef CONFIG_SMP
  1602. extern void sched_exec(void);
  1603. #else
  1604. #define sched_exec() {}
  1605. #endif
  1606. extern void sched_clock_idle_sleep_event(void);
  1607. extern void sched_clock_idle_wakeup_event(u64 delta_ns);
  1608. #ifdef CONFIG_HOTPLUG_CPU
  1609. extern void idle_task_exit(void);
  1610. #else
  1611. static inline void idle_task_exit(void) {}
  1612. #endif
  1613. extern void sched_idle_next(void);
  1614. #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP)
  1615. extern void wake_up_idle_cpu(int cpu);
  1616. #else
  1617. static inline void wake_up_idle_cpu(int cpu) { }
  1618. #endif
  1619. extern unsigned int sysctl_sched_latency;
  1620. extern unsigned int sysctl_sched_min_granularity;
  1621. extern unsigned int sysctl_sched_wakeup_granularity;
  1622. extern unsigned int sysctl_sched_shares_ratelimit;
  1623. extern unsigned int sysctl_sched_shares_thresh;
  1624. extern unsigned int sysctl_sched_child_runs_first;
  1625. enum sched_tunable_scaling {
  1626. SCHED_TUNABLESCALING_NONE,
  1627. SCHED_TUNABLESCALING_LOG,
  1628. SCHED_TUNABLESCALING_LINEAR,
  1629. SCHED_TUNABLESCALING_END,
  1630. };
  1631. extern enum sched_tunable_scaling sysctl_sched_tunable_scaling;
  1632. #ifdef CONFIG_SCHED_DEBUG
  1633. extern unsigned int sysctl_sched_migration_cost;
  1634. extern unsigned int sysctl_sched_nr_migrate;
  1635. extern unsigned int sysctl_sched_time_avg;
  1636. extern unsigned int sysctl_timer_migration;
  1637. int sched_proc_update_handler(struct ctl_table *table, int write,
  1638. void __user *buffer, size_t *length,
  1639. loff_t *ppos);
  1640. #endif
  1641. #ifdef CONFIG_SCHED_DEBUG
  1642. static inline unsigned int get_sysctl_timer_migration(void)
  1643. {
  1644. return sysctl_timer_migration;
  1645. }
  1646. #else
  1647. static inline unsigned int get_sysctl_timer_migration(void)
  1648. {
  1649. return 1;
  1650. }
  1651. #endif
  1652. extern unsigned int sysctl_sched_rt_period;
  1653. extern int sysctl_sched_rt_runtime;
  1654. int sched_rt_handler(struct ctl_table *table, int write,
  1655. void __user *buffer, size_t *lenp,
  1656. loff_t *ppos);
  1657. extern unsigned int sysctl_sched_compat_yield;
  1658. #ifdef CONFIG_RT_MUTEXES
  1659. extern int rt_mutex_getprio(struct task_struct *p);
  1660. extern void rt_mutex_setprio(struct task_struct *p, int prio);
  1661. extern void rt_mutex_adjust_pi(struct task_struct *p);
  1662. #else
  1663. static inline int rt_mutex_getprio(struct task_struct *p)
  1664. {
  1665. return p->normal_prio;
  1666. }
  1667. # define rt_mutex_adjust_pi(p) do { } while (0)
  1668. #endif
  1669. extern void set_user_nice(struct task_struct *p, long nice);
  1670. extern int task_prio(const struct task_struct *p);
  1671. extern int task_nice(const struct task_struct *p);
  1672. extern int can_nice(const struct task_struct *p, const int nice);
  1673. extern int task_curr(const struct task_struct *p);
  1674. extern int idle_cpu(int cpu);
  1675. extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
  1676. extern int sched_setscheduler_nocheck(struct task_struct *, int,
  1677. struct sched_param *);
  1678. extern struct task_struct *idle_task(int cpu);
  1679. extern struct task_struct *curr_task(int cpu);
  1680. extern void set_curr_task(int cpu, struct task_struct *p);
  1681. void yield(void);
  1682. /*
  1683. * The default (Linux) execution domain.
  1684. */
  1685. extern struct exec_domain default_exec_domain;
  1686. union thread_union {
  1687. struct thread_info thread_info;
  1688. unsigned long stack[THREAD_SIZE/sizeof(long)];
  1689. };
  1690. #ifndef __HAVE_ARCH_KSTACK_END
  1691. static inline int kstack_end(void *addr)
  1692. {
  1693. /* Reliable end of stack detection:
  1694. * Some APM bios versions misalign the stack
  1695. */
  1696. return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
  1697. }
  1698. #endif
  1699. extern union thread_union init_thread_union;
  1700. extern struct task_struct init_task;
  1701. extern struct mm_struct init_mm;
  1702. extern struct pid_namespace init_pid_ns;
  1703. /*
  1704. * find a task by one of its numerical ids
  1705. *
  1706. * find_task_by_pid_ns():
  1707. * finds a task by its pid in the specified namespace
  1708. * find_task_by_vpid():
  1709. * finds a task by its virtual pid
  1710. *
  1711. * see also find_vpid() etc in include/linux/pid.h
  1712. */
  1713. extern struct task_struct *find_task_by_vpid(pid_t nr);
  1714. extern struct task_struct *find_task_by_pid_ns(pid_t nr,
  1715. struct pid_namespace *ns);
  1716. extern void __set_special_pids(struct pid *pid);
  1717. /* per-UID process charging. */
  1718. extern struct user_struct * alloc_uid(struct user_namespace *, uid_t);
  1719. static inline struct user_struct *get_uid(struct user_struct *u)
  1720. {
  1721. atomic_inc(&u->__count);
  1722. return u;
  1723. }
  1724. extern void free_uid(struct user_struct *);
  1725. extern void release_uids(struct user_namespace *ns);
  1726. #include <asm/current.h>
  1727. extern void do_timer(unsigned long ticks);
  1728. extern int wake_up_state(struct task_struct *tsk, unsigned int state);
  1729. extern int wake_up_process(struct task_struct *tsk);
  1730. extern void wake_up_new_task(struct task_struct *tsk,
  1731. unsigned long clone_flags);
  1732. #ifdef CONFIG_SMP
  1733. extern void kick_process(struct task_struct *tsk);
  1734. #else
  1735. static inline void kick_process(struct task_struct *tsk) { }
  1736. #endif
  1737. extern void sched_fork(struct task_struct *p, int clone_flags);
  1738. extern void sched_dead(struct task_struct *p);
  1739. extern void proc_caches_init(void);
  1740. extern void flush_signals(struct task_struct *);
  1741. extern void __flush_signals(struct task_struct *);
  1742. extern void ignore_signals(struct task_struct *);
  1743. extern void flush_signal_handlers(struct task_struct *, int force_default);
  1744. extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
  1745. static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
  1746. {
  1747. unsigned long flags;
  1748. int ret;
  1749. spin_lock_irqsave(&tsk->sighand->siglock, flags);
  1750. ret = dequeue_signal(tsk, mask, info);
  1751. spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
  1752. return ret;
  1753. }
  1754. extern void block_all_signals(int (*notifier)(void *priv), void *priv,
  1755. sigset_t *mask);
  1756. extern void unblock_all_signals(void);
  1757. extern void release_task(struct task_struct * p);
  1758. extern int send_sig_info(int, struct siginfo *, struct task_struct *);
  1759. extern int force_sigsegv(int, struct task_struct *);
  1760. extern int force_sig_info(int, struct siginfo *, struct task_struct *);
  1761. extern int __kill_pgrp_info(int sig, struct siginfo *info, struct pid *pgrp);
  1762. extern int kill_pid_info(int sig, struct siginfo *info, struct pid *pid);
  1763. extern int kill_pid_info_as_uid(int, struct siginfo *, struct pid *, uid_t, uid_t, u32);
  1764. extern int kill_pgrp(struct pid *pid, int sig, int priv);
  1765. extern int kill_pid(struct pid *pid, int sig, int priv);
  1766. extern int kill_proc_info(int, struct siginfo *, pid_t);
  1767. extern int do_notify_parent(struct task_struct *, int);
  1768. extern void __wake_up_parent(struct task_struct *p, struct task_struct *parent);
  1769. extern void force_sig(int, struct task_struct *);
  1770. extern int send_sig(int, struct task_struct *, int);
  1771. extern void zap_other_threads(struct task_struct *p);
  1772. extern struct sigqueue *sigqueue_alloc(void);
  1773. extern void sigqueue_free(struct sigqueue *);
  1774. extern int send_sigqueue(struct sigqueue *, struct task_struct *, int group);
  1775. extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
  1776. extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
  1777. static inline int kill_cad_pid(int sig, int priv)
  1778. {
  1779. return kill_pid(cad_pid, sig, priv);
  1780. }
  1781. /* These can be the second arg to send_sig_info/send_group_sig_info. */
  1782. #define SEND_SIG_NOINFO ((struct siginfo *) 0)
  1783. #define SEND_SIG_PRIV ((struct siginfo *) 1)
  1784. #define SEND_SIG_FORCED ((struct siginfo *) 2)
  1785. /*
  1786. * True if we are on the alternate signal stack.
  1787. */
  1788. static inline int on_sig_stack(unsigned long sp)
  1789. {
  1790. #ifdef CONFIG_STACK_GROWSUP
  1791. return sp >= current->sas_ss_sp &&
  1792. sp - current->sas_ss_sp < current->sas_ss_size;
  1793. #else
  1794. return sp > current->sas_ss_sp &&
  1795. sp - current->sas_ss_sp <= current->sas_ss_size;
  1796. #endif
  1797. }
  1798. static inline int sas_ss_flags(unsigned long sp)
  1799. {
  1800. return (current->sas_ss_size == 0 ? SS_DISABLE
  1801. : on_sig_stack(sp) ? SS_ONSTACK : 0);
  1802. }
  1803. /*
  1804. * Routines for handling mm_structs
  1805. */
  1806. extern struct mm_struct * mm_alloc(void);
  1807. /* mmdrop drops the mm and the page tables */
  1808. extern void __mmdrop(struct mm_struct *);
  1809. static inline void mmdrop(struct mm_struct * mm)
  1810. {
  1811. if (unlikely(atomic_dec_and_test(&mm->mm_count)))
  1812. __mmdrop(mm);
  1813. }
  1814. /* mmput gets rid of the mappings and all user-space */
  1815. extern void mmput(struct mm_struct *);
  1816. /* Grab a reference to a task's mm, if it is not already going away */
  1817. extern struct mm_struct *get_task_mm(struct task_struct *task);
  1818. /* Remove the current tasks stale references to the old mm_struct */
  1819. extern void mm_release(struct task_struct *, struct mm_struct *);
  1820. /* Allocate a new mm structure and copy contents from tsk->mm */
  1821. extern struct mm_struct *dup_mm(struct task_struct *tsk);
  1822. extern int copy_thread(unsigned long, unsigned long, unsigned long,
  1823. struct task_struct *, struct pt_regs *);
  1824. extern void flush_thread(void);
  1825. extern void exit_thread(void);
  1826. extern void exit_files(struct task_struct *);
  1827. extern void __cleanup_signal(struct signal_struct *);
  1828. extern void __cleanup_sighand(struct sighand_struct *);
  1829. extern void exit_itimers(struct signal_struct *);
  1830. extern void flush_itimer_signals(void);
  1831. extern NORET_TYPE void do_group_exit(int);
  1832. extern void daemonize(const char *, ...);
  1833. extern int allow_signal(int);
  1834. extern int disallow_signal(int);
  1835. extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
  1836. extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
  1837. struct task_struct *fork_idle(int);
  1838. extern void set_task_comm(struct task_struct *tsk, char *from);
  1839. extern char *get_task_comm(char *to, struct task_struct *tsk);
  1840. #ifdef CONFIG_SMP
  1841. extern void wait_task_context_switch(struct task_struct *p);
  1842. extern unsigned long wait_task_inactive(struct task_struct *, long match_state);
  1843. #else
  1844. static inline void wait_task_context_switch(struct task_struct *p) {}
  1845. static inline unsigned long wait_task_inactive(struct task_struct *p,
  1846. long match_state)
  1847. {
  1848. return 1;
  1849. }
  1850. #endif
  1851. #define next_task(p) \
  1852. list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
  1853. #define for_each_process(p) \
  1854. for (p = &init_task ; (p = next_task(p)) != &init_task ; )
  1855. extern bool current_is_single_threaded(void);
  1856. /*
  1857. * Careful: do_each_thread/while_each_thread is a double loop so
  1858. * 'break' will not work as expected - use goto instead.
  1859. */
  1860. #define do_each_thread(g, t) \
  1861. for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
  1862. #define while_each_thread(g, t) \
  1863. while ((t = next_thread(t)) != g)
  1864. /* de_thread depends on thread_group_leader not being a pid based check */
  1865. #define thread_group_leader(p) (p == p->group_leader)
  1866. /* Do to the insanities of de_thread it is possible for a process
  1867. * to have the pid of the thread group leader without actually being
  1868. * the thread group leader. For iteration through the pids in proc
  1869. * all we care about is that we have a task with the appropriate
  1870. * pid, we don't actually care if we have the right task.
  1871. */
  1872. static inline int has_group_leader_pid(struct task_struct *p)
  1873. {
  1874. return p->pid == p->tgid;
  1875. }
  1876. static inline
  1877. int same_thread_group(struct task_struct *p1, struct task_struct *p2)
  1878. {
  1879. return p1->tgid == p2->tgid;
  1880. }
  1881. static inline struct task_struct *next_thread(const struct task_struct *p)
  1882. {
  1883. return list_entry_rcu(p->thread_group.next,
  1884. struct task_struct, thread_group);
  1885. }
  1886. static inline int thread_group_empty(struct task_struct *p)
  1887. {
  1888. return list_empty(&p->thread_group);
  1889. }
  1890. #define delay_group_leader(p) \
  1891. (thread_group_leader(p) && !thread_group_empty(p))
  1892. static inline int task_detached(struct task_struct *p)
  1893. {
  1894. return p->exit_signal == -1;
  1895. }
  1896. /*
  1897. * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
  1898. * subscriptions and synchronises with wait4(). Also used in procfs. Also
  1899. * pins the final release of task.io_context. Also protects ->cpuset and
  1900. * ->cgroup.subsys[].
  1901. *
  1902. * Nests both inside and outside of read_lock(&tasklist_lock).
  1903. * It must not be nested with write_lock_irq(&tasklist_lock),
  1904. * neither inside nor outside.
  1905. */
  1906. static inline void task_lock(struct task_struct *p)
  1907. {
  1908. spin_lock(&p->alloc_lock);
  1909. }
  1910. static inline void task_unlock(struct task_struct *p)
  1911. {
  1912. spin_unlock(&p->alloc_lock);
  1913. }
  1914. extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
  1915. unsigned long *flags);
  1916. static inline void unlock_task_sighand(struct task_struct *tsk,
  1917. unsigned long *flags)
  1918. {
  1919. spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
  1920. }
  1921. #ifndef __HAVE_THREAD_FUNCTIONS
  1922. #define task_thread_info(task) ((struct thread_info *)(task)->stack)
  1923. #define task_stack_page(task) ((task)->stack)
  1924. static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
  1925. {
  1926. *task_thread_info(p) = *task_thread_info(org);
  1927. task_thread_info(p)->task = p;
  1928. }
  1929. static inline unsigned long *end_of_stack(struct task_struct *p)
  1930. {
  1931. return (unsigned long *)(task_thread_info(p) + 1);
  1932. }
  1933. #endif
  1934. static inline int object_is_on_stack(void *obj)
  1935. {
  1936. void *stack = task_stack_page(current);
  1937. return (obj >= stack) && (obj < (stack + THREAD_SIZE));
  1938. }
  1939. extern void thread_info_cache_init(void);
  1940. #ifdef CONFIG_DEBUG_STACK_USAGE
  1941. static inline unsigned long stack_not_used(struct task_struct *p)
  1942. {
  1943. unsigned long *n = end_of_stack(p);
  1944. do { /* Skip over canary */
  1945. n++;
  1946. } while (!*n);
  1947. return (unsigned long)n - (unsigned long)end_of_stack(p);
  1948. }
  1949. #endif
  1950. /* set thread flags in other task's structures
  1951. * - see asm/thread_info.h for TIF_xxxx flags available
  1952. */
  1953. static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1954. {
  1955. set_ti_thread_flag(task_thread_info(tsk), flag);
  1956. }
  1957. static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1958. {
  1959. clear_ti_thread_flag(task_thread_info(tsk), flag);
  1960. }
  1961. static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
  1962. {
  1963. return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
  1964. }
  1965. static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
  1966. {
  1967. return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
  1968. }
  1969. static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
  1970. {
  1971. return test_ti_thread_flag(task_thread_info(tsk), flag);
  1972. }
  1973. static inline void set_tsk_need_resched(struct task_struct *tsk)
  1974. {
  1975. set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1976. }
  1977. static inline void clear_tsk_need_resched(struct task_struct *tsk)
  1978. {
  1979. clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
  1980. }
  1981. static inline int test_tsk_need_resched(struct task_struct *tsk)
  1982. {
  1983. return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
  1984. }
  1985. static inline int restart_syscall(void)
  1986. {
  1987. set_tsk_thread_flag(current, TIF_SIGPENDING);
  1988. return -ERESTARTNOINTR;
  1989. }
  1990. static inline int signal_pending(struct task_struct *p)
  1991. {
  1992. return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
  1993. }
  1994. static inline int __fatal_signal_pending(struct task_struct *p)
  1995. {
  1996. return unlikely(sigismember(&p->pending.signal, SIGKILL));
  1997. }
  1998. static inline int fatal_signal_pending(struct task_struct *p)
  1999. {
  2000. return signal_pending(p) && __fatal_signal_pending(p);
  2001. }
  2002. static inline int signal_pending_state(long state, struct task_struct *p)
  2003. {
  2004. if (!(state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
  2005. return 0;
  2006. if (!signal_pending(p))
  2007. return 0;
  2008. return (state & TASK_INTERRUPTIBLE) || __fatal_signal_pending(p);
  2009. }
  2010. static inline int need_resched(void)
  2011. {
  2012. return unlikely(test_thread_flag(TIF_NEED_RESCHED));
  2013. }
  2014. /*
  2015. * cond_resched() and cond_resched_lock(): latency reduction via
  2016. * explicit rescheduling in places that are safe. The return
  2017. * value indicates whether a reschedule was done in fact.
  2018. * cond_resched_lock() will drop the spinlock before scheduling,
  2019. * cond_resched_softirq() will enable bhs before scheduling.
  2020. */
  2021. extern int _cond_resched(void);
  2022. #define cond_resched() ({ \
  2023. __might_sleep(__FILE__, __LINE__, 0); \
  2024. _cond_resched(); \
  2025. })
  2026. extern int __cond_resched_lock(spinlock_t *lock);
  2027. #ifdef CONFIG_PREEMPT
  2028. #define PREEMPT_LOCK_OFFSET PREEMPT_OFFSET
  2029. #else
  2030. #define PREEMPT_LOCK_OFFSET 0
  2031. #endif
  2032. #define cond_resched_lock(lock) ({ \
  2033. __might_sleep(__FILE__, __LINE__, PREEMPT_LOCK_OFFSET); \
  2034. __cond_resched_lock(lock); \
  2035. })
  2036. extern int __cond_resched_softirq(void);
  2037. #define cond_resched_softirq() ({ \
  2038. __might_sleep(__FILE__, __LINE__, SOFTIRQ_OFFSET); \
  2039. __cond_resched_softirq(); \
  2040. })
  2041. /*
  2042. * Does a critical section need to be broken due to another
  2043. * task waiting?: (technically does not depend on CONFIG_PREEMPT,
  2044. * but a general need for low latency)
  2045. */
  2046. static inline int spin_needbreak(spinlock_t *lock)
  2047. {
  2048. #ifdef CONFIG_PREEMPT
  2049. return spin_is_contended(lock);
  2050. #else
  2051. return 0;
  2052. #endif
  2053. }
  2054. /*
  2055. * Thread group CPU time accounting.
  2056. */
  2057. void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
  2058. void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
  2059. static inline void thread_group_cputime_init(struct signal_struct *sig)
  2060. {
  2061. sig->cputimer.cputime = INIT_CPUTIME;
  2062. spin_lock_init(&sig->cputimer.lock);
  2063. sig->cputimer.running = 0;
  2064. }
  2065. static inline void thread_group_cputime_free(struct signal_struct *sig)
  2066. {
  2067. }
  2068. /*
  2069. * Reevaluate whether the task has signals pending delivery.
  2070. * Wake the task if so.
  2071. * This is required every time the blocked sigset_t changes.
  2072. * callers must hold sighand->siglock.
  2073. */
  2074. extern void recalc_sigpending_and_wake(struct task_struct *t);
  2075. extern void recalc_sigpending(void);
  2076. extern void signal_wake_up(struct task_struct *t, int resume_stopped);
  2077. /*
  2078. * Wrappers for p->thread_info->cpu access. No-op on UP.
  2079. */
  2080. #ifdef CONFIG_SMP
  2081. static inline unsigned int task_cpu(const struct task_struct *p)
  2082. {
  2083. return task_thread_info(p)->cpu;
  2084. }
  2085. extern void set_task_cpu(struct task_struct *p, unsigned int cpu);
  2086. #else
  2087. static inline unsigned int task_cpu(const struct task_struct *p)
  2088. {
  2089. return 0;
  2090. }
  2091. static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
  2092. {
  2093. }
  2094. #endif /* CONFIG_SMP */
  2095. extern void arch_pick_mmap_layout(struct mm_struct *mm);
  2096. #ifdef CONFIG_TRACING
  2097. extern void
  2098. __trace_special(void *__tr, void *__data,
  2099. unsigned long arg1, unsigned long arg2, unsigned long arg3);
  2100. #else
  2101. static inline void
  2102. __trace_special(void *__tr, void *__data,
  2103. unsigned long arg1, unsigned long arg2, unsigned long arg3)
  2104. {
  2105. }
  2106. #endif
  2107. extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
  2108. extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
  2109. extern void normalize_rt_tasks(void);
  2110. #ifdef CONFIG_GROUP_SCHED
  2111. extern struct task_group init_task_group;
  2112. #ifdef CONFIG_USER_SCHED
  2113. extern struct task_group root_task_group;
  2114. extern void set_tg_uid(struct user_struct *user);
  2115. #endif
  2116. extern struct task_group *sched_create_group(struct task_group *parent);
  2117. extern void sched_destroy_group(struct task_group *tg);
  2118. extern void sched_move_task(struct task_struct *tsk);
  2119. #ifdef CONFIG_FAIR_GROUP_SCHED
  2120. extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
  2121. extern unsigned long sched_group_shares(struct task_group *tg);
  2122. #endif
  2123. #ifdef CONFIG_RT_GROUP_SCHED
  2124. extern int sched_group_set_rt_runtime(struct task_group *tg,
  2125. long rt_runtime_us);
  2126. extern long sched_group_rt_runtime(struct task_group *tg);
  2127. extern int sched_group_set_rt_period(struct task_group *tg,
  2128. long rt_period_us);
  2129. extern long sched_group_rt_period(struct task_group *tg);
  2130. extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
  2131. #endif
  2132. #endif
  2133. extern int task_can_switch_user(struct user_struct *up,
  2134. struct task_struct *tsk);
  2135. #ifdef CONFIG_TASK_XACCT
  2136. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  2137. {
  2138. tsk->ioac.rchar += amt;
  2139. }
  2140. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  2141. {
  2142. tsk->ioac.wchar += amt;
  2143. }
  2144. static inline void inc_syscr(struct task_struct *tsk)
  2145. {
  2146. tsk->ioac.syscr++;
  2147. }
  2148. static inline void inc_syscw(struct task_struct *tsk)
  2149. {
  2150. tsk->ioac.syscw++;
  2151. }
  2152. #else
  2153. static inline void add_rchar(struct task_struct *tsk, ssize_t amt)
  2154. {
  2155. }
  2156. static inline void add_wchar(struct task_struct *tsk, ssize_t amt)
  2157. {
  2158. }
  2159. static inline void inc_syscr(struct task_struct *tsk)
  2160. {
  2161. }
  2162. static inline void inc_syscw(struct task_struct *tsk)
  2163. {
  2164. }
  2165. #endif
  2166. #ifndef TASK_SIZE_OF
  2167. #define TASK_SIZE_OF(tsk) TASK_SIZE
  2168. #endif
  2169. /*
  2170. * Call the function if the target task is executing on a CPU right now:
  2171. */
  2172. extern void task_oncpu_function_call(struct task_struct *p,
  2173. void (*func) (void *info), void *info);
  2174. #ifdef CONFIG_MM_OWNER
  2175. extern void mm_update_next_owner(struct mm_struct *mm);
  2176. extern void mm_init_owner(struct mm_struct *mm, struct task_struct *p);
  2177. #else
  2178. static inline void mm_update_next_owner(struct mm_struct *mm)
  2179. {
  2180. }
  2181. static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
  2182. {
  2183. }
  2184. #endif /* CONFIG_MM_OWNER */
  2185. #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
  2186. #endif /* __KERNEL__ */
  2187. #endif