fs.h 108 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409141014111412141314141415141614171418141914201421142214231424142514261427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532153315341535153615371538153915401541154215431544154515461547154815491550155115521553155415551556155715581559156015611562156315641565156615671568156915701571157215731574157515761577157815791580158115821583158415851586158715881589159015911592159315941595159615971598159916001601160216031604160516061607160816091610161116121613161416151616161716181619162016211622162316241625162616271628162916301631163216331634163516361637163816391640164116421643164416451646164716481649165016511652165316541655165616571658165916601661166216631664166516661667166816691670167116721673167416751676167716781679168016811682168316841685168616871688168916901691169216931694169516961697169816991700170117021703170417051706170717081709171017111712171317141715171617171718171917201721172217231724172517261727172817291730173117321733173417351736173717381739174017411742174317441745174617471748174917501751175217531754175517561757175817591760176117621763176417651766176717681769177017711772177317741775177617771778177917801781178217831784178517861787178817891790179117921793179417951796179717981799180018011802180318041805180618071808180918101811181218131814181518161817181818191820182118221823182418251826182718281829183018311832183318341835183618371838183918401841184218431844184518461847184818491850185118521853185418551856185718581859186018611862186318641865186618671868186918701871187218731874187518761877187818791880188118821883188418851886188718881889189018911892189318941895189618971898189919001901190219031904190519061907190819091910191119121913191419151916191719181919192019211922192319241925192619271928192919301931193219331934193519361937193819391940194119421943194419451946194719481949195019511952195319541955195619571958195919601961196219631964196519661967196819691970197119721973197419751976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015201620172018201920202021202220232024202520262027202820292030203120322033203420352036203720382039204020412042204320442045204620472048204920502051205220532054205520562057205820592060206120622063206420652066206720682069207020712072207320742075207620772078207920802081208220832084208520862087208820892090209120922093209420952096209720982099210021012102210321042105210621072108210921102111211221132114211521162117211821192120212121222123212421252126212721282129213021312132213321342135213621372138213921402141214221432144214521462147214821492150215121522153215421552156215721582159216021612162216321642165216621672168216921702171217221732174217521762177217821792180218121822183218421852186218721882189219021912192219321942195219621972198219922002201220222032204220522062207220822092210221122122213221422152216221722182219222022212222222322242225222622272228222922302231223222332234223522362237223822392240224122422243224422452246224722482249225022512252225322542255225622572258225922602261226222632264226522662267226822692270227122722273227422752276227722782279228022812282228322842285228622872288228922902291229222932294229522962297229822992300230123022303230423052306230723082309231023112312231323142315231623172318231923202321232223232324232523262327232823292330233123322333233423352336233723382339234023412342234323442345234623472348234923502351235223532354235523562357235823592360236123622363236423652366236723682369237023712372237323742375237623772378237923802381238223832384238523862387238823892390239123922393239423952396239723982399240024012402240324042405240624072408240924102411241224132414241524162417241824192420242124222423242424252426242724282429243024312432243324342435243624372438243924402441244224432444244524462447244824492450245124522453245424552456245724582459246024612462246324642465246624672468246924702471247224732474247524762477247824792480248124822483248424852486248724882489249024912492249324942495249624972498249925002501250225032504250525062507250825092510251125122513251425152516251725182519252025212522252325242525252625272528252925302531253225332534253525362537253825392540254125422543254425452546254725482549255025512552255325542555255625572558255925602561256225632564256525662567256825692570257125722573257425752576257725782579258025812582258325842585258625872588258925902591259225932594259525962597259825992600260126022603260426052606260726082609261026112612261326142615261626172618261926202621262226232624262526262627262826292630263126322633263426352636263726382639264026412642264326442645264626472648264926502651265226532654265526562657265826592660266126622663266426652666266726682669267026712672267326742675267626772678267926802681268226832684268526862687268826892690269126922693269426952696269726982699270027012702270327042705270627072708270927102711271227132714271527162717271827192720272127222723272427252726272727282729273027312732273327342735273627372738273927402741274227432744274527462747274827492750275127522753275427552756275727582759276027612762276327642765276627672768276927702771277227732774277527762777277827792780278127822783278427852786278727882789279027912792279327942795279627972798279928002801280228032804280528062807280828092810281128122813281428152816281728182819282028212822282328242825282628272828282928302831283228332834283528362837283828392840284128422843284428452846284728482849285028512852285328542855285628572858285928602861286228632864286528662867286828692870287128722873287428752876287728782879288028812882288328842885288628872888288928902891289228932894289528962897289828992900290129022903290429052906290729082909291029112912291329142915291629172918291929202921292229232924292529262927292829292930293129322933293429352936293729382939294029412942294329442945294629472948294929502951295229532954295529562957295829592960296129622963296429652966296729682969297029712972297329742975297629772978297929802981298229832984298529862987298829892990299129922993299429952996299729982999300030013002300330043005300630073008300930103011301230133014301530163017301830193020302130223023302430253026302730283029303030313032303330343035303630373038303930403041304230433044304530463047304830493050305130523053305430553056305730583059306030613062306330643065306630673068306930703071307230733074307530763077307830793080308130823083308430853086308730883089309030913092309330943095309630973098309931003101310231033104310531063107310831093110311131123113311431153116311731183119312031213122312331243125312631273128312931303131313231333134313531363137313831393140314131423143314431453146314731483149315031513152315331543155315631573158315931603161316231633164316531663167316831693170317131723173317431753176317731783179318031813182318331843185318631873188318931903191319231933194319531963197319831993200320132023203320432053206320732083209321032113212321332143215321632173218321932203221322232233224322532263227322832293230323132323233323432353236323732383239324032413242324332443245324632473248324932503251325232533254325532563257325832593260326132623263326432653266326732683269327032713272327332743275327632773278327932803281328232833284328532863287328832893290329132923293329432953296329732983299330033013302330333043305330633073308330933103311331233133314331533163317331833193320332133223323332433253326332733283329333033313332333333343335333633373338333933403341334233433344334533463347334833493350335133523353335433553356335733583359336033613362336333643365336633673368336933703371337233733374337533763377
  1. #ifndef _LINUX_FS_H
  2. #define _LINUX_FS_H
  3. #include <linux/linkage.h>
  4. #include <linux/wait_bit.h>
  5. #include <linux/kdev_t.h>
  6. #include <linux/dcache.h>
  7. #include <linux/path.h>
  8. #include <linux/stat.h>
  9. #include <linux/cache.h>
  10. #include <linux/list.h>
  11. #include <linux/list_lru.h>
  12. #include <linux/llist.h>
  13. #include <linux/radix-tree.h>
  14. #include <linux/rbtree.h>
  15. #include <linux/init.h>
  16. #include <linux/pid.h>
  17. #include <linux/bug.h>
  18. #include <linux/mutex.h>
  19. #include <linux/rwsem.h>
  20. #include <linux/mm_types.h>
  21. #include <linux/capability.h>
  22. #include <linux/semaphore.h>
  23. #include <linux/fcntl.h>
  24. #include <linux/fiemap.h>
  25. #include <linux/rculist_bl.h>
  26. #include <linux/atomic.h>
  27. #include <linux/shrinker.h>
  28. #include <linux/migrate_mode.h>
  29. #include <linux/uidgid.h>
  30. #include <linux/lockdep.h>
  31. #include <linux/percpu-rwsem.h>
  32. #include <linux/workqueue.h>
  33. #include <linux/delayed_call.h>
  34. #include <linux/uuid.h>
  35. #include <linux/errseq.h>
  36. #include <asm/byteorder.h>
  37. #include <uapi/linux/fs.h>
  38. struct backing_dev_info;
  39. struct bdi_writeback;
  40. struct bio;
  41. struct export_operations;
  42. struct hd_geometry;
  43. struct iovec;
  44. struct kiocb;
  45. struct kobject;
  46. struct pipe_inode_info;
  47. struct poll_table_struct;
  48. struct kstatfs;
  49. struct vm_area_struct;
  50. struct vfsmount;
  51. struct cred;
  52. struct swap_info_struct;
  53. struct seq_file;
  54. struct workqueue_struct;
  55. struct iov_iter;
  56. struct fscrypt_info;
  57. struct fscrypt_operations;
  58. extern void __init inode_init(void);
  59. extern void __init inode_init_early(void);
  60. extern void __init files_init(void);
  61. extern void __init files_maxfiles_init(void);
  62. extern struct files_stat_struct files_stat;
  63. extern unsigned long get_max_files(void);
  64. extern unsigned int sysctl_nr_open;
  65. extern struct inodes_stat_t inodes_stat;
  66. extern int leases_enable, lease_break_time;
  67. extern int sysctl_protected_symlinks;
  68. extern int sysctl_protected_hardlinks;
  69. typedef __kernel_rwf_t rwf_t;
  70. struct buffer_head;
  71. typedef int (get_block_t)(struct inode *inode, sector_t iblock,
  72. struct buffer_head *bh_result, int create);
  73. typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
  74. ssize_t bytes, void *private);
  75. #define MAY_EXEC 0x00000001
  76. #define MAY_WRITE 0x00000002
  77. #define MAY_READ 0x00000004
  78. #define MAY_APPEND 0x00000008
  79. #define MAY_ACCESS 0x00000010
  80. #define MAY_OPEN 0x00000020
  81. #define MAY_CHDIR 0x00000040
  82. /* called from RCU mode, don't block */
  83. #define MAY_NOT_BLOCK 0x00000080
  84. /*
  85. * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
  86. * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
  87. */
  88. /* file is open for reading */
  89. #define FMODE_READ ((__force fmode_t)0x1)
  90. /* file is open for writing */
  91. #define FMODE_WRITE ((__force fmode_t)0x2)
  92. /* file is seekable */
  93. #define FMODE_LSEEK ((__force fmode_t)0x4)
  94. /* file can be accessed using pread */
  95. #define FMODE_PREAD ((__force fmode_t)0x8)
  96. /* file can be accessed using pwrite */
  97. #define FMODE_PWRITE ((__force fmode_t)0x10)
  98. /* File is opened for execution with sys_execve / sys_uselib */
  99. #define FMODE_EXEC ((__force fmode_t)0x20)
  100. /* File is opened with O_NDELAY (only set for block devices) */
  101. #define FMODE_NDELAY ((__force fmode_t)0x40)
  102. /* File is opened with O_EXCL (only set for block devices) */
  103. #define FMODE_EXCL ((__force fmode_t)0x80)
  104. /* File is opened using open(.., 3, ..) and is writeable only for ioctls
  105. (specialy hack for floppy.c) */
  106. #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
  107. /* 32bit hashes as llseek() offset (for directories) */
  108. #define FMODE_32BITHASH ((__force fmode_t)0x200)
  109. /* 64bit hashes as llseek() offset (for directories) */
  110. #define FMODE_64BITHASH ((__force fmode_t)0x400)
  111. /*
  112. * Don't update ctime and mtime.
  113. *
  114. * Currently a special hack for the XFS open_by_handle ioctl, but we'll
  115. * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
  116. */
  117. #define FMODE_NOCMTIME ((__force fmode_t)0x800)
  118. /* Expect random access pattern */
  119. #define FMODE_RANDOM ((__force fmode_t)0x1000)
  120. /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
  121. #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
  122. /* File is opened with O_PATH; almost nothing can be done with it */
  123. #define FMODE_PATH ((__force fmode_t)0x4000)
  124. /* File needs atomic accesses to f_pos */
  125. #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
  126. /* Write access to underlying fs */
  127. #define FMODE_WRITER ((__force fmode_t)0x10000)
  128. /* Has read method(s) */
  129. #define FMODE_CAN_READ ((__force fmode_t)0x20000)
  130. /* Has write method(s) */
  131. #define FMODE_CAN_WRITE ((__force fmode_t)0x40000)
  132. /* File was opened by fanotify and shouldn't generate fanotify events */
  133. #define FMODE_NONOTIFY ((__force fmode_t)0x4000000)
  134. /* File is capable of returning -EAGAIN if I/O will block */
  135. #define FMODE_NOWAIT ((__force fmode_t)0x8000000)
  136. /*
  137. * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
  138. * that indicates that they should check the contents of the iovec are
  139. * valid, but not check the memory that the iovec elements
  140. * points too.
  141. */
  142. #define CHECK_IOVEC_ONLY -1
  143. /*
  144. * Attribute flags. These should be or-ed together to figure out what
  145. * has been changed!
  146. */
  147. #define ATTR_MODE (1 << 0)
  148. #define ATTR_UID (1 << 1)
  149. #define ATTR_GID (1 << 2)
  150. #define ATTR_SIZE (1 << 3)
  151. #define ATTR_ATIME (1 << 4)
  152. #define ATTR_MTIME (1 << 5)
  153. #define ATTR_CTIME (1 << 6)
  154. #define ATTR_ATIME_SET (1 << 7)
  155. #define ATTR_MTIME_SET (1 << 8)
  156. #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
  157. #define ATTR_ATTR_FLAG (1 << 10)
  158. #define ATTR_KILL_SUID (1 << 11)
  159. #define ATTR_KILL_SGID (1 << 12)
  160. #define ATTR_FILE (1 << 13)
  161. #define ATTR_KILL_PRIV (1 << 14)
  162. #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
  163. #define ATTR_TIMES_SET (1 << 16)
  164. #define ATTR_TOUCH (1 << 17)
  165. /*
  166. * Whiteout is represented by a char device. The following constants define the
  167. * mode and device number to use.
  168. */
  169. #define WHITEOUT_MODE 0
  170. #define WHITEOUT_DEV 0
  171. /*
  172. * This is the Inode Attributes structure, used for notify_change(). It
  173. * uses the above definitions as flags, to know which values have changed.
  174. * Also, in this manner, a Filesystem can look at only the values it cares
  175. * about. Basically, these are the attributes that the VFS layer can
  176. * request to change from the FS layer.
  177. *
  178. * Derek Atkins <warlord@MIT.EDU> 94-10-20
  179. */
  180. struct iattr {
  181. unsigned int ia_valid;
  182. umode_t ia_mode;
  183. kuid_t ia_uid;
  184. kgid_t ia_gid;
  185. loff_t ia_size;
  186. struct timespec ia_atime;
  187. struct timespec ia_mtime;
  188. struct timespec ia_ctime;
  189. /*
  190. * Not an attribute, but an auxiliary info for filesystems wanting to
  191. * implement an ftruncate() like method. NOTE: filesystem should
  192. * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
  193. */
  194. struct file *ia_file;
  195. };
  196. /*
  197. * Includes for diskquotas.
  198. */
  199. #include <linux/quota.h>
  200. /*
  201. * Maximum number of layers of fs stack. Needs to be limited to
  202. * prevent kernel stack overflow
  203. */
  204. #define FILESYSTEM_MAX_STACK_DEPTH 2
  205. /**
  206. * enum positive_aop_returns - aop return codes with specific semantics
  207. *
  208. * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
  209. * completed, that the page is still locked, and
  210. * should be considered active. The VM uses this hint
  211. * to return the page to the active list -- it won't
  212. * be a candidate for writeback again in the near
  213. * future. Other callers must be careful to unlock
  214. * the page if they get this return. Returned by
  215. * writepage();
  216. *
  217. * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
  218. * unlocked it and the page might have been truncated.
  219. * The caller should back up to acquiring a new page and
  220. * trying again. The aop will be taking reasonable
  221. * precautions not to livelock. If the caller held a page
  222. * reference, it should drop it before retrying. Returned
  223. * by readpage().
  224. *
  225. * address_space_operation functions return these large constants to indicate
  226. * special semantics to the caller. These are much larger than the bytes in a
  227. * page to allow for functions that return the number of bytes operated on in a
  228. * given page.
  229. */
  230. enum positive_aop_returns {
  231. AOP_WRITEPAGE_ACTIVATE = 0x80000,
  232. AOP_TRUNCATED_PAGE = 0x80001,
  233. };
  234. #define AOP_FLAG_CONT_EXPAND 0x0001 /* called from cont_expand */
  235. #define AOP_FLAG_NOFS 0x0002 /* used by filesystem to direct
  236. * helper code (eg buffer layer)
  237. * to clear GFP_FS from alloc */
  238. /*
  239. * oh the beauties of C type declarations.
  240. */
  241. struct page;
  242. struct address_space;
  243. struct writeback_control;
  244. /*
  245. * Write life time hint values.
  246. */
  247. enum rw_hint {
  248. WRITE_LIFE_NOT_SET = 0,
  249. WRITE_LIFE_NONE = RWH_WRITE_LIFE_NONE,
  250. WRITE_LIFE_SHORT = RWH_WRITE_LIFE_SHORT,
  251. WRITE_LIFE_MEDIUM = RWH_WRITE_LIFE_MEDIUM,
  252. WRITE_LIFE_LONG = RWH_WRITE_LIFE_LONG,
  253. WRITE_LIFE_EXTREME = RWH_WRITE_LIFE_EXTREME,
  254. };
  255. #define IOCB_EVENTFD (1 << 0)
  256. #define IOCB_APPEND (1 << 1)
  257. #define IOCB_DIRECT (1 << 2)
  258. #define IOCB_HIPRI (1 << 3)
  259. #define IOCB_DSYNC (1 << 4)
  260. #define IOCB_SYNC (1 << 5)
  261. #define IOCB_WRITE (1 << 6)
  262. #define IOCB_NOWAIT (1 << 7)
  263. struct kiocb {
  264. struct file *ki_filp;
  265. loff_t ki_pos;
  266. void (*ki_complete)(struct kiocb *iocb, long ret, long ret2);
  267. void *private;
  268. int ki_flags;
  269. enum rw_hint ki_hint;
  270. } __randomize_layout;
  271. static inline bool is_sync_kiocb(struct kiocb *kiocb)
  272. {
  273. return kiocb->ki_complete == NULL;
  274. }
  275. /*
  276. * "descriptor" for what we're up to with a read.
  277. * This allows us to use the same read code yet
  278. * have multiple different users of the data that
  279. * we read from a file.
  280. *
  281. * The simplest case just copies the data to user
  282. * mode.
  283. */
  284. typedef struct {
  285. size_t written;
  286. size_t count;
  287. union {
  288. char __user *buf;
  289. void *data;
  290. } arg;
  291. int error;
  292. } read_descriptor_t;
  293. typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
  294. unsigned long, unsigned long);
  295. struct address_space_operations {
  296. int (*writepage)(struct page *page, struct writeback_control *wbc);
  297. int (*readpage)(struct file *, struct page *);
  298. /* Write back some dirty pages from this mapping. */
  299. int (*writepages)(struct address_space *, struct writeback_control *);
  300. /* Set a page dirty. Return true if this dirtied it */
  301. int (*set_page_dirty)(struct page *page);
  302. int (*readpages)(struct file *filp, struct address_space *mapping,
  303. struct list_head *pages, unsigned nr_pages);
  304. int (*write_begin)(struct file *, struct address_space *mapping,
  305. loff_t pos, unsigned len, unsigned flags,
  306. struct page **pagep, void **fsdata);
  307. int (*write_end)(struct file *, struct address_space *mapping,
  308. loff_t pos, unsigned len, unsigned copied,
  309. struct page *page, void *fsdata);
  310. /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
  311. sector_t (*bmap)(struct address_space *, sector_t);
  312. void (*invalidatepage) (struct page *, unsigned int, unsigned int);
  313. int (*releasepage) (struct page *, gfp_t);
  314. void (*freepage)(struct page *);
  315. ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter);
  316. /*
  317. * migrate the contents of a page to the specified target. If
  318. * migrate_mode is MIGRATE_ASYNC, it must not block.
  319. */
  320. int (*migratepage) (struct address_space *,
  321. struct page *, struct page *, enum migrate_mode);
  322. bool (*isolate_page)(struct page *, isolate_mode_t);
  323. void (*putback_page)(struct page *);
  324. int (*launder_page) (struct page *);
  325. int (*is_partially_uptodate) (struct page *, unsigned long,
  326. unsigned long);
  327. void (*is_dirty_writeback) (struct page *, bool *, bool *);
  328. int (*error_remove_page)(struct address_space *, struct page *);
  329. /* swapfile support */
  330. int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
  331. sector_t *span);
  332. void (*swap_deactivate)(struct file *file);
  333. };
  334. extern const struct address_space_operations empty_aops;
  335. /*
  336. * pagecache_write_begin/pagecache_write_end must be used by general code
  337. * to write into the pagecache.
  338. */
  339. int pagecache_write_begin(struct file *, struct address_space *mapping,
  340. loff_t pos, unsigned len, unsigned flags,
  341. struct page **pagep, void **fsdata);
  342. int pagecache_write_end(struct file *, struct address_space *mapping,
  343. loff_t pos, unsigned len, unsigned copied,
  344. struct page *page, void *fsdata);
  345. struct address_space {
  346. struct inode *host; /* owner: inode, block_device */
  347. struct radix_tree_root page_tree; /* radix tree of all pages */
  348. spinlock_t tree_lock; /* and lock protecting it */
  349. atomic_t i_mmap_writable;/* count VM_SHARED mappings */
  350. struct rb_root_cached i_mmap; /* tree of private and shared mappings */
  351. struct rw_semaphore i_mmap_rwsem; /* protect tree, count, list */
  352. /* Protected by tree_lock together with the radix tree */
  353. unsigned long nrpages; /* number of total pages */
  354. /* number of shadow or DAX exceptional entries */
  355. unsigned long nrexceptional;
  356. pgoff_t writeback_index;/* writeback starts here */
  357. const struct address_space_operations *a_ops; /* methods */
  358. unsigned long flags; /* error bits */
  359. spinlock_t private_lock; /* for use by the address_space */
  360. gfp_t gfp_mask; /* implicit gfp mask for allocations */
  361. struct list_head private_list; /* ditto */
  362. void *private_data; /* ditto */
  363. errseq_t wb_err;
  364. } __attribute__((aligned(sizeof(long)))) __randomize_layout;
  365. /*
  366. * On most architectures that alignment is already the case; but
  367. * must be enforced here for CRIS, to let the least significant bit
  368. * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
  369. */
  370. struct request_queue;
  371. struct block_device {
  372. dev_t bd_dev; /* not a kdev_t - it's a search key */
  373. int bd_openers;
  374. struct inode * bd_inode; /* will die */
  375. struct super_block * bd_super;
  376. struct mutex bd_mutex; /* open/close mutex */
  377. void * bd_claiming;
  378. void * bd_holder;
  379. int bd_holders;
  380. bool bd_write_holder;
  381. #ifdef CONFIG_SYSFS
  382. struct list_head bd_holder_disks;
  383. #endif
  384. struct block_device * bd_contains;
  385. unsigned bd_block_size;
  386. u8 bd_partno;
  387. struct hd_struct * bd_part;
  388. /* number of times partitions within this device have been opened. */
  389. unsigned bd_part_count;
  390. int bd_invalidated;
  391. struct gendisk * bd_disk;
  392. struct request_queue * bd_queue;
  393. struct backing_dev_info *bd_bdi;
  394. struct list_head bd_list;
  395. /*
  396. * Private data. You must have bd_claim'ed the block_device
  397. * to use this. NOTE: bd_claim allows an owner to claim
  398. * the same device multiple times, the owner must take special
  399. * care to not mess up bd_private for that case.
  400. */
  401. unsigned long bd_private;
  402. /* The counter of freeze processes */
  403. int bd_fsfreeze_count;
  404. /* Mutex for freeze */
  405. struct mutex bd_fsfreeze_mutex;
  406. } __randomize_layout;
  407. /*
  408. * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
  409. * radix trees
  410. */
  411. #define PAGECACHE_TAG_DIRTY 0
  412. #define PAGECACHE_TAG_WRITEBACK 1
  413. #define PAGECACHE_TAG_TOWRITE 2
  414. int mapping_tagged(struct address_space *mapping, int tag);
  415. static inline void i_mmap_lock_write(struct address_space *mapping)
  416. {
  417. down_write(&mapping->i_mmap_rwsem);
  418. }
  419. static inline void i_mmap_unlock_write(struct address_space *mapping)
  420. {
  421. up_write(&mapping->i_mmap_rwsem);
  422. }
  423. static inline void i_mmap_lock_read(struct address_space *mapping)
  424. {
  425. down_read(&mapping->i_mmap_rwsem);
  426. }
  427. static inline void i_mmap_unlock_read(struct address_space *mapping)
  428. {
  429. up_read(&mapping->i_mmap_rwsem);
  430. }
  431. /*
  432. * Might pages of this file be mapped into userspace?
  433. */
  434. static inline int mapping_mapped(struct address_space *mapping)
  435. {
  436. return !RB_EMPTY_ROOT(&mapping->i_mmap.rb_root);
  437. }
  438. /*
  439. * Might pages of this file have been modified in userspace?
  440. * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
  441. * marks vma as VM_SHARED if it is shared, and the file was opened for
  442. * writing i.e. vma may be mprotected writable even if now readonly.
  443. *
  444. * If i_mmap_writable is negative, no new writable mappings are allowed. You
  445. * can only deny writable mappings, if none exists right now.
  446. */
  447. static inline int mapping_writably_mapped(struct address_space *mapping)
  448. {
  449. return atomic_read(&mapping->i_mmap_writable) > 0;
  450. }
  451. static inline int mapping_map_writable(struct address_space *mapping)
  452. {
  453. return atomic_inc_unless_negative(&mapping->i_mmap_writable) ?
  454. 0 : -EPERM;
  455. }
  456. static inline void mapping_unmap_writable(struct address_space *mapping)
  457. {
  458. atomic_dec(&mapping->i_mmap_writable);
  459. }
  460. static inline int mapping_deny_writable(struct address_space *mapping)
  461. {
  462. return atomic_dec_unless_positive(&mapping->i_mmap_writable) ?
  463. 0 : -EBUSY;
  464. }
  465. static inline void mapping_allow_writable(struct address_space *mapping)
  466. {
  467. atomic_inc(&mapping->i_mmap_writable);
  468. }
  469. /*
  470. * Use sequence counter to get consistent i_size on 32-bit processors.
  471. */
  472. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  473. #include <linux/seqlock.h>
  474. #define __NEED_I_SIZE_ORDERED
  475. #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
  476. #else
  477. #define i_size_ordered_init(inode) do { } while (0)
  478. #endif
  479. struct posix_acl;
  480. #define ACL_NOT_CACHED ((void *)(-1))
  481. #define ACL_DONT_CACHE ((void *)(-3))
  482. static inline struct posix_acl *
  483. uncached_acl_sentinel(struct task_struct *task)
  484. {
  485. return (void *)task + 1;
  486. }
  487. static inline bool
  488. is_uncached_acl(struct posix_acl *acl)
  489. {
  490. return (long)acl & 1;
  491. }
  492. #define IOP_FASTPERM 0x0001
  493. #define IOP_LOOKUP 0x0002
  494. #define IOP_NOFOLLOW 0x0004
  495. #define IOP_XATTR 0x0008
  496. #define IOP_DEFAULT_READLINK 0x0010
  497. struct fsnotify_mark_connector;
  498. /*
  499. * Keep mostly read-only and often accessed (especially for
  500. * the RCU path lookup and 'stat' data) fields at the beginning
  501. * of the 'struct inode'
  502. */
  503. struct inode {
  504. umode_t i_mode;
  505. unsigned short i_opflags;
  506. kuid_t i_uid;
  507. kgid_t i_gid;
  508. unsigned int i_flags;
  509. #ifdef CONFIG_FS_POSIX_ACL
  510. struct posix_acl *i_acl;
  511. struct posix_acl *i_default_acl;
  512. #endif
  513. const struct inode_operations *i_op;
  514. struct super_block *i_sb;
  515. struct address_space *i_mapping;
  516. #ifdef CONFIG_SECURITY
  517. void *i_security;
  518. #endif
  519. /* Stat data, not accessed from path walking */
  520. unsigned long i_ino;
  521. /*
  522. * Filesystems may only read i_nlink directly. They shall use the
  523. * following functions for modification:
  524. *
  525. * (set|clear|inc|drop)_nlink
  526. * inode_(inc|dec)_link_count
  527. */
  528. union {
  529. const unsigned int i_nlink;
  530. unsigned int __i_nlink;
  531. };
  532. dev_t i_rdev;
  533. loff_t i_size;
  534. struct timespec i_atime;
  535. struct timespec i_mtime;
  536. struct timespec i_ctime;
  537. spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
  538. unsigned short i_bytes;
  539. unsigned int i_blkbits;
  540. enum rw_hint i_write_hint;
  541. blkcnt_t i_blocks;
  542. #ifdef __NEED_I_SIZE_ORDERED
  543. seqcount_t i_size_seqcount;
  544. #endif
  545. /* Misc */
  546. unsigned long i_state;
  547. struct rw_semaphore i_rwsem;
  548. unsigned long dirtied_when; /* jiffies of first dirtying */
  549. unsigned long dirtied_time_when;
  550. struct hlist_node i_hash;
  551. struct list_head i_io_list; /* backing dev IO list */
  552. #ifdef CONFIG_CGROUP_WRITEBACK
  553. struct bdi_writeback *i_wb; /* the associated cgroup wb */
  554. /* foreign inode detection, see wbc_detach_inode() */
  555. int i_wb_frn_winner;
  556. u16 i_wb_frn_avg_time;
  557. u16 i_wb_frn_history;
  558. #endif
  559. struct list_head i_lru; /* inode LRU list */
  560. struct list_head i_sb_list;
  561. struct list_head i_wb_list; /* backing dev writeback list */
  562. union {
  563. struct hlist_head i_dentry;
  564. struct rcu_head i_rcu;
  565. };
  566. u64 i_version;
  567. atomic_t i_count;
  568. atomic_t i_dio_count;
  569. atomic_t i_writecount;
  570. #ifdef CONFIG_IMA
  571. atomic_t i_readcount; /* struct files open RO */
  572. #endif
  573. const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
  574. struct file_lock_context *i_flctx;
  575. struct address_space i_data;
  576. struct list_head i_devices;
  577. union {
  578. struct pipe_inode_info *i_pipe;
  579. struct block_device *i_bdev;
  580. struct cdev *i_cdev;
  581. char *i_link;
  582. unsigned i_dir_seq;
  583. };
  584. __u32 i_generation;
  585. #ifdef CONFIG_FSNOTIFY
  586. __u32 i_fsnotify_mask; /* all events this inode cares about */
  587. struct fsnotify_mark_connector __rcu *i_fsnotify_marks;
  588. #endif
  589. #if IS_ENABLED(CONFIG_FS_ENCRYPTION)
  590. struct fscrypt_info *i_crypt_info;
  591. #endif
  592. void *i_private; /* fs or device private pointer */
  593. } __randomize_layout;
  594. static inline unsigned int i_blocksize(const struct inode *node)
  595. {
  596. return (1 << node->i_blkbits);
  597. }
  598. static inline int inode_unhashed(struct inode *inode)
  599. {
  600. return hlist_unhashed(&inode->i_hash);
  601. }
  602. /*
  603. * inode->i_mutex nesting subclasses for the lock validator:
  604. *
  605. * 0: the object of the current VFS operation
  606. * 1: parent
  607. * 2: child/target
  608. * 3: xattr
  609. * 4: second non-directory
  610. * 5: second parent (when locking independent directories in rename)
  611. *
  612. * I_MUTEX_NONDIR2 is for certain operations (such as rename) which lock two
  613. * non-directories at once.
  614. *
  615. * The locking order between these classes is
  616. * parent[2] -> child -> grandchild -> normal -> xattr -> second non-directory
  617. */
  618. enum inode_i_mutex_lock_class
  619. {
  620. I_MUTEX_NORMAL,
  621. I_MUTEX_PARENT,
  622. I_MUTEX_CHILD,
  623. I_MUTEX_XATTR,
  624. I_MUTEX_NONDIR2,
  625. I_MUTEX_PARENT2,
  626. };
  627. static inline void inode_lock(struct inode *inode)
  628. {
  629. down_write(&inode->i_rwsem);
  630. }
  631. static inline void inode_unlock(struct inode *inode)
  632. {
  633. up_write(&inode->i_rwsem);
  634. }
  635. static inline void inode_lock_shared(struct inode *inode)
  636. {
  637. down_read(&inode->i_rwsem);
  638. }
  639. static inline void inode_unlock_shared(struct inode *inode)
  640. {
  641. up_read(&inode->i_rwsem);
  642. }
  643. static inline int inode_trylock(struct inode *inode)
  644. {
  645. return down_write_trylock(&inode->i_rwsem);
  646. }
  647. static inline int inode_trylock_shared(struct inode *inode)
  648. {
  649. return down_read_trylock(&inode->i_rwsem);
  650. }
  651. static inline int inode_is_locked(struct inode *inode)
  652. {
  653. return rwsem_is_locked(&inode->i_rwsem);
  654. }
  655. static inline void inode_lock_nested(struct inode *inode, unsigned subclass)
  656. {
  657. down_write_nested(&inode->i_rwsem, subclass);
  658. }
  659. void lock_two_nondirectories(struct inode *, struct inode*);
  660. void unlock_two_nondirectories(struct inode *, struct inode*);
  661. /*
  662. * NOTE: in a 32bit arch with a preemptable kernel and
  663. * an UP compile the i_size_read/write must be atomic
  664. * with respect to the local cpu (unlike with preempt disabled),
  665. * but they don't need to be atomic with respect to other cpus like in
  666. * true SMP (so they need either to either locally disable irq around
  667. * the read or for example on x86 they can be still implemented as a
  668. * cmpxchg8b without the need of the lock prefix). For SMP compiles
  669. * and 64bit archs it makes no difference if preempt is enabled or not.
  670. */
  671. static inline loff_t i_size_read(const struct inode *inode)
  672. {
  673. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  674. loff_t i_size;
  675. unsigned int seq;
  676. do {
  677. seq = read_seqcount_begin(&inode->i_size_seqcount);
  678. i_size = inode->i_size;
  679. } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
  680. return i_size;
  681. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  682. loff_t i_size;
  683. preempt_disable();
  684. i_size = inode->i_size;
  685. preempt_enable();
  686. return i_size;
  687. #else
  688. return inode->i_size;
  689. #endif
  690. }
  691. /*
  692. * NOTE: unlike i_size_read(), i_size_write() does need locking around it
  693. * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
  694. * can be lost, resulting in subsequent i_size_read() calls spinning forever.
  695. */
  696. static inline void i_size_write(struct inode *inode, loff_t i_size)
  697. {
  698. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  699. preempt_disable();
  700. write_seqcount_begin(&inode->i_size_seqcount);
  701. inode->i_size = i_size;
  702. write_seqcount_end(&inode->i_size_seqcount);
  703. preempt_enable();
  704. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  705. preempt_disable();
  706. inode->i_size = i_size;
  707. preempt_enable();
  708. #else
  709. inode->i_size = i_size;
  710. #endif
  711. }
  712. static inline unsigned iminor(const struct inode *inode)
  713. {
  714. return MINOR(inode->i_rdev);
  715. }
  716. static inline unsigned imajor(const struct inode *inode)
  717. {
  718. return MAJOR(inode->i_rdev);
  719. }
  720. extern struct block_device *I_BDEV(struct inode *inode);
  721. struct fown_struct {
  722. rwlock_t lock; /* protects pid, uid, euid fields */
  723. struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
  724. enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
  725. kuid_t uid, euid; /* uid/euid of process setting the owner */
  726. int signum; /* posix.1b rt signal to be delivered on IO */
  727. };
  728. /*
  729. * Track a single file's readahead state
  730. */
  731. struct file_ra_state {
  732. pgoff_t start; /* where readahead started */
  733. unsigned int size; /* # of readahead pages */
  734. unsigned int async_size; /* do asynchronous readahead when
  735. there are only # of pages ahead */
  736. unsigned int ra_pages; /* Maximum readahead window */
  737. unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
  738. loff_t prev_pos; /* Cache last read() position */
  739. };
  740. /*
  741. * Check if @index falls in the readahead windows.
  742. */
  743. static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
  744. {
  745. return (index >= ra->start &&
  746. index < ra->start + ra->size);
  747. }
  748. struct file {
  749. union {
  750. struct llist_node fu_llist;
  751. struct rcu_head fu_rcuhead;
  752. } f_u;
  753. struct path f_path;
  754. struct inode *f_inode; /* cached value */
  755. const struct file_operations *f_op;
  756. /*
  757. * Protects f_ep_links, f_flags.
  758. * Must not be taken from IRQ context.
  759. */
  760. spinlock_t f_lock;
  761. enum rw_hint f_write_hint;
  762. atomic_long_t f_count;
  763. unsigned int f_flags;
  764. fmode_t f_mode;
  765. struct mutex f_pos_lock;
  766. loff_t f_pos;
  767. struct fown_struct f_owner;
  768. const struct cred *f_cred;
  769. struct file_ra_state f_ra;
  770. u64 f_version;
  771. #ifdef CONFIG_SECURITY
  772. void *f_security;
  773. #endif
  774. /* needed for tty driver, and maybe others */
  775. void *private_data;
  776. #ifdef CONFIG_EPOLL
  777. /* Used by fs/eventpoll.c to link all the hooks to this file */
  778. struct list_head f_ep_links;
  779. struct list_head f_tfile_llink;
  780. #endif /* #ifdef CONFIG_EPOLL */
  781. struct address_space *f_mapping;
  782. errseq_t f_wb_err;
  783. } __randomize_layout
  784. __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
  785. struct file_handle {
  786. __u32 handle_bytes;
  787. int handle_type;
  788. /* file identifier */
  789. unsigned char f_handle[0];
  790. };
  791. static inline struct file *get_file(struct file *f)
  792. {
  793. atomic_long_inc(&f->f_count);
  794. return f;
  795. }
  796. #define get_file_rcu(x) atomic_long_inc_not_zero(&(x)->f_count)
  797. #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
  798. #define file_count(x) atomic_long_read(&(x)->f_count)
  799. #define MAX_NON_LFS ((1UL<<31) - 1)
  800. /* Page cache limit. The filesystems should put that into their s_maxbytes
  801. limits, otherwise bad things can happen in VM. */
  802. #if BITS_PER_LONG==32
  803. #define MAX_LFS_FILESIZE ((loff_t)ULONG_MAX << PAGE_SHIFT)
  804. #elif BITS_PER_LONG==64
  805. #define MAX_LFS_FILESIZE ((loff_t)LLONG_MAX)
  806. #endif
  807. #define FL_POSIX 1
  808. #define FL_FLOCK 2
  809. #define FL_DELEG 4 /* NFSv4 delegation */
  810. #define FL_ACCESS 8 /* not trying to lock, just looking */
  811. #define FL_EXISTS 16 /* when unlocking, test for existence */
  812. #define FL_LEASE 32 /* lease held on this file */
  813. #define FL_CLOSE 64 /* unlock on close */
  814. #define FL_SLEEP 128 /* A blocking lock */
  815. #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
  816. #define FL_UNLOCK_PENDING 512 /* Lease is being broken */
  817. #define FL_OFDLCK 1024 /* lock is "owned" by struct file */
  818. #define FL_LAYOUT 2048 /* outstanding pNFS layout */
  819. #define FL_CLOSE_POSIX (FL_POSIX | FL_CLOSE)
  820. /*
  821. * Special return value from posix_lock_file() and vfs_lock_file() for
  822. * asynchronous locking.
  823. */
  824. #define FILE_LOCK_DEFERRED 1
  825. /* legacy typedef, should eventually be removed */
  826. typedef void *fl_owner_t;
  827. struct file_lock;
  828. struct file_lock_operations {
  829. void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
  830. void (*fl_release_private)(struct file_lock *);
  831. };
  832. struct lock_manager_operations {
  833. int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
  834. unsigned long (*lm_owner_key)(struct file_lock *);
  835. fl_owner_t (*lm_get_owner)(fl_owner_t);
  836. void (*lm_put_owner)(fl_owner_t);
  837. void (*lm_notify)(struct file_lock *); /* unblock callback */
  838. int (*lm_grant)(struct file_lock *, int);
  839. bool (*lm_break)(struct file_lock *);
  840. int (*lm_change)(struct file_lock *, int, struct list_head *);
  841. void (*lm_setup)(struct file_lock *, void **);
  842. };
  843. struct lock_manager {
  844. struct list_head list;
  845. /*
  846. * NFSv4 and up also want opens blocked during the grace period;
  847. * NLM doesn't care:
  848. */
  849. bool block_opens;
  850. };
  851. struct net;
  852. void locks_start_grace(struct net *, struct lock_manager *);
  853. void locks_end_grace(struct lock_manager *);
  854. int locks_in_grace(struct net *);
  855. int opens_in_grace(struct net *);
  856. /* that will die - we need it for nfs_lock_info */
  857. #include <linux/nfs_fs_i.h>
  858. /*
  859. * struct file_lock represents a generic "file lock". It's used to represent
  860. * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
  861. * note that the same struct is used to represent both a request for a lock and
  862. * the lock itself, but the same object is never used for both.
  863. *
  864. * FIXME: should we create a separate "struct lock_request" to help distinguish
  865. * these two uses?
  866. *
  867. * The varous i_flctx lists are ordered by:
  868. *
  869. * 1) lock owner
  870. * 2) lock range start
  871. * 3) lock range end
  872. *
  873. * Obviously, the last two criteria only matter for POSIX locks.
  874. */
  875. struct file_lock {
  876. struct file_lock *fl_next; /* singly linked list for this inode */
  877. struct list_head fl_list; /* link into file_lock_context */
  878. struct hlist_node fl_link; /* node in global lists */
  879. struct list_head fl_block; /* circular list of blocked processes */
  880. fl_owner_t fl_owner;
  881. unsigned int fl_flags;
  882. unsigned char fl_type;
  883. unsigned int fl_pid;
  884. int fl_link_cpu; /* what cpu's list is this on? */
  885. wait_queue_head_t fl_wait;
  886. struct file *fl_file;
  887. loff_t fl_start;
  888. loff_t fl_end;
  889. struct fasync_struct * fl_fasync; /* for lease break notifications */
  890. /* for lease breaks: */
  891. unsigned long fl_break_time;
  892. unsigned long fl_downgrade_time;
  893. const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
  894. const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
  895. union {
  896. struct nfs_lock_info nfs_fl;
  897. struct nfs4_lock_info nfs4_fl;
  898. struct {
  899. struct list_head link; /* link in AFS vnode's pending_locks list */
  900. int state; /* state of grant or error if -ve */
  901. } afs;
  902. } fl_u;
  903. } __randomize_layout;
  904. struct file_lock_context {
  905. spinlock_t flc_lock;
  906. struct list_head flc_flock;
  907. struct list_head flc_posix;
  908. struct list_head flc_lease;
  909. };
  910. /* The following constant reflects the upper bound of the file/locking space */
  911. #ifndef OFFSET_MAX
  912. #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
  913. #define OFFSET_MAX INT_LIMIT(loff_t)
  914. #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
  915. #endif
  916. extern void send_sigio(struct fown_struct *fown, int fd, int band);
  917. /*
  918. * Return the inode to use for locking
  919. *
  920. * For overlayfs this should be the overlay inode, not the real inode returned
  921. * by file_inode(). For any other fs file_inode(filp) and locks_inode(filp) are
  922. * equal.
  923. */
  924. static inline struct inode *locks_inode(const struct file *f)
  925. {
  926. return f->f_path.dentry->d_inode;
  927. }
  928. #ifdef CONFIG_FILE_LOCKING
  929. extern int fcntl_getlk(struct file *, unsigned int, struct flock *);
  930. extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
  931. struct flock *);
  932. #if BITS_PER_LONG == 32
  933. extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 *);
  934. extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
  935. struct flock64 *);
  936. #endif
  937. extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
  938. extern int fcntl_getlease(struct file *filp);
  939. /* fs/locks.c */
  940. void locks_free_lock_context(struct inode *inode);
  941. void locks_free_lock(struct file_lock *fl);
  942. extern void locks_init_lock(struct file_lock *);
  943. extern struct file_lock * locks_alloc_lock(void);
  944. extern void locks_copy_lock(struct file_lock *, struct file_lock *);
  945. extern void locks_copy_conflock(struct file_lock *, struct file_lock *);
  946. extern void locks_remove_posix(struct file *, fl_owner_t);
  947. extern void locks_remove_file(struct file *);
  948. extern void locks_release_private(struct file_lock *);
  949. extern void posix_test_lock(struct file *, struct file_lock *);
  950. extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
  951. extern int posix_unblock_lock(struct file_lock *);
  952. extern int vfs_test_lock(struct file *, struct file_lock *);
  953. extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
  954. extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
  955. extern int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl);
  956. extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
  957. extern void lease_get_mtime(struct inode *, struct timespec *time);
  958. extern int generic_setlease(struct file *, long, struct file_lock **, void **priv);
  959. extern int vfs_setlease(struct file *, long, struct file_lock **, void **);
  960. extern int lease_modify(struct file_lock *, int, struct list_head *);
  961. struct files_struct;
  962. extern void show_fd_locks(struct seq_file *f,
  963. struct file *filp, struct files_struct *files);
  964. #else /* !CONFIG_FILE_LOCKING */
  965. static inline int fcntl_getlk(struct file *file, unsigned int cmd,
  966. struct flock __user *user)
  967. {
  968. return -EINVAL;
  969. }
  970. static inline int fcntl_setlk(unsigned int fd, struct file *file,
  971. unsigned int cmd, struct flock __user *user)
  972. {
  973. return -EACCES;
  974. }
  975. #if BITS_PER_LONG == 32
  976. static inline int fcntl_getlk64(struct file *file, unsigned int cmd,
  977. struct flock64 __user *user)
  978. {
  979. return -EINVAL;
  980. }
  981. static inline int fcntl_setlk64(unsigned int fd, struct file *file,
  982. unsigned int cmd, struct flock64 __user *user)
  983. {
  984. return -EACCES;
  985. }
  986. #endif
  987. static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
  988. {
  989. return -EINVAL;
  990. }
  991. static inline int fcntl_getlease(struct file *filp)
  992. {
  993. return F_UNLCK;
  994. }
  995. static inline void
  996. locks_free_lock_context(struct inode *inode)
  997. {
  998. }
  999. static inline void locks_init_lock(struct file_lock *fl)
  1000. {
  1001. return;
  1002. }
  1003. static inline void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
  1004. {
  1005. return;
  1006. }
  1007. static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1008. {
  1009. return;
  1010. }
  1011. static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
  1012. {
  1013. return;
  1014. }
  1015. static inline void locks_remove_file(struct file *filp)
  1016. {
  1017. return;
  1018. }
  1019. static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
  1020. {
  1021. return;
  1022. }
  1023. static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
  1024. struct file_lock *conflock)
  1025. {
  1026. return -ENOLCK;
  1027. }
  1028. static inline int posix_unblock_lock(struct file_lock *waiter)
  1029. {
  1030. return -ENOENT;
  1031. }
  1032. static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
  1033. {
  1034. return 0;
  1035. }
  1036. static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
  1037. struct file_lock *fl, struct file_lock *conf)
  1038. {
  1039. return -ENOLCK;
  1040. }
  1041. static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
  1042. {
  1043. return 0;
  1044. }
  1045. static inline int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl)
  1046. {
  1047. return -ENOLCK;
  1048. }
  1049. static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
  1050. {
  1051. return 0;
  1052. }
  1053. static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
  1054. {
  1055. return;
  1056. }
  1057. static inline int generic_setlease(struct file *filp, long arg,
  1058. struct file_lock **flp, void **priv)
  1059. {
  1060. return -EINVAL;
  1061. }
  1062. static inline int vfs_setlease(struct file *filp, long arg,
  1063. struct file_lock **lease, void **priv)
  1064. {
  1065. return -EINVAL;
  1066. }
  1067. static inline int lease_modify(struct file_lock *fl, int arg,
  1068. struct list_head *dispose)
  1069. {
  1070. return -EINVAL;
  1071. }
  1072. struct files_struct;
  1073. static inline void show_fd_locks(struct seq_file *f,
  1074. struct file *filp, struct files_struct *files) {}
  1075. #endif /* !CONFIG_FILE_LOCKING */
  1076. static inline struct inode *file_inode(const struct file *f)
  1077. {
  1078. return f->f_inode;
  1079. }
  1080. static inline struct dentry *file_dentry(const struct file *file)
  1081. {
  1082. return d_real(file->f_path.dentry, file_inode(file), 0, 0);
  1083. }
  1084. static inline int locks_lock_file_wait(struct file *filp, struct file_lock *fl)
  1085. {
  1086. return locks_lock_inode_wait(locks_inode(filp), fl);
  1087. }
  1088. struct fasync_struct {
  1089. spinlock_t fa_lock;
  1090. int magic;
  1091. int fa_fd;
  1092. struct fasync_struct *fa_next; /* singly linked list */
  1093. struct file *fa_file;
  1094. struct rcu_head fa_rcu;
  1095. };
  1096. #define FASYNC_MAGIC 0x4601
  1097. /* SMP safe fasync helpers: */
  1098. extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
  1099. extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
  1100. extern int fasync_remove_entry(struct file *, struct fasync_struct **);
  1101. extern struct fasync_struct *fasync_alloc(void);
  1102. extern void fasync_free(struct fasync_struct *);
  1103. /* can be called from interrupts */
  1104. extern void kill_fasync(struct fasync_struct **, int, int);
  1105. extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
  1106. extern int f_setown(struct file *filp, unsigned long arg, int force);
  1107. extern void f_delown(struct file *filp);
  1108. extern pid_t f_getown(struct file *filp);
  1109. extern int send_sigurg(struct fown_struct *fown);
  1110. /*
  1111. * sb->s_flags. Note that these mirror the equivalent MS_* flags where
  1112. * represented in both.
  1113. */
  1114. #define SB_RDONLY 1 /* Mount read-only */
  1115. #define SB_NOSUID 2 /* Ignore suid and sgid bits */
  1116. #define SB_NODEV 4 /* Disallow access to device special files */
  1117. #define SB_NOEXEC 8 /* Disallow program execution */
  1118. #define SB_SYNCHRONOUS 16 /* Writes are synced at once */
  1119. #define SB_MANDLOCK 64 /* Allow mandatory locks on an FS */
  1120. #define SB_DIRSYNC 128 /* Directory modifications are synchronous */
  1121. #define SB_NOATIME 1024 /* Do not update access times. */
  1122. #define SB_NODIRATIME 2048 /* Do not update directory access times */
  1123. #define SB_SILENT 32768
  1124. #define SB_POSIXACL (1<<16) /* VFS does not apply the umask */
  1125. #define SB_KERNMOUNT (1<<22) /* this is a kern_mount call */
  1126. #define SB_I_VERSION (1<<23) /* Update inode I_version field */
  1127. #define SB_LAZYTIME (1<<25) /* Update the on-disk [acm]times lazily */
  1128. /* These sb flags are internal to the kernel */
  1129. #define SB_SUBMOUNT (1<<26)
  1130. #define SB_NOREMOTELOCK (1<<27)
  1131. #define SB_NOSEC (1<<28)
  1132. #define SB_BORN (1<<29)
  1133. #define SB_ACTIVE (1<<30)
  1134. #define SB_NOUSER (1<<31)
  1135. /*
  1136. * Umount options
  1137. */
  1138. #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
  1139. #define MNT_DETACH 0x00000002 /* Just detach from the tree */
  1140. #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
  1141. #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
  1142. #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
  1143. /* sb->s_iflags */
  1144. #define SB_I_CGROUPWB 0x00000001 /* cgroup-aware writeback enabled */
  1145. #define SB_I_NOEXEC 0x00000002 /* Ignore executables on this fs */
  1146. #define SB_I_NODEV 0x00000004 /* Ignore devices on this fs */
  1147. /* sb->s_iflags to limit user namespace mounts */
  1148. #define SB_I_USERNS_VISIBLE 0x00000010 /* fstype already mounted */
  1149. /* Possible states of 'frozen' field */
  1150. enum {
  1151. SB_UNFROZEN = 0, /* FS is unfrozen */
  1152. SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
  1153. SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
  1154. SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
  1155. * internal threads if needed) */
  1156. SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
  1157. };
  1158. #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
  1159. struct sb_writers {
  1160. int frozen; /* Is sb frozen? */
  1161. wait_queue_head_t wait_unfrozen; /* for get_super_thawed() */
  1162. struct percpu_rw_semaphore rw_sem[SB_FREEZE_LEVELS];
  1163. };
  1164. struct super_block {
  1165. struct list_head s_list; /* Keep this first */
  1166. dev_t s_dev; /* search index; _not_ kdev_t */
  1167. unsigned char s_blocksize_bits;
  1168. unsigned long s_blocksize;
  1169. loff_t s_maxbytes; /* Max file size */
  1170. struct file_system_type *s_type;
  1171. const struct super_operations *s_op;
  1172. const struct dquot_operations *dq_op;
  1173. const struct quotactl_ops *s_qcop;
  1174. const struct export_operations *s_export_op;
  1175. unsigned long s_flags;
  1176. unsigned long s_iflags; /* internal SB_I_* flags */
  1177. unsigned long s_magic;
  1178. struct dentry *s_root;
  1179. struct rw_semaphore s_umount;
  1180. int s_count;
  1181. atomic_t s_active;
  1182. #ifdef CONFIG_SECURITY
  1183. void *s_security;
  1184. #endif
  1185. const struct xattr_handler **s_xattr;
  1186. const struct fscrypt_operations *s_cop;
  1187. struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
  1188. struct list_head s_mounts; /* list of mounts; _not_ for fs use */
  1189. struct block_device *s_bdev;
  1190. struct backing_dev_info *s_bdi;
  1191. struct mtd_info *s_mtd;
  1192. struct hlist_node s_instances;
  1193. unsigned int s_quota_types; /* Bitmask of supported quota types */
  1194. struct quota_info s_dquot; /* Diskquota specific options */
  1195. struct sb_writers s_writers;
  1196. char s_id[32]; /* Informational name */
  1197. uuid_t s_uuid; /* UUID */
  1198. void *s_fs_info; /* Filesystem private info */
  1199. unsigned int s_max_links;
  1200. fmode_t s_mode;
  1201. /* Granularity of c/m/atime in ns.
  1202. Cannot be worse than a second */
  1203. u32 s_time_gran;
  1204. /*
  1205. * The next field is for VFS *only*. No filesystems have any business
  1206. * even looking at it. You had been warned.
  1207. */
  1208. struct mutex s_vfs_rename_mutex; /* Kludge */
  1209. /*
  1210. * Filesystem subtype. If non-empty the filesystem type field
  1211. * in /proc/mounts will be "type.subtype"
  1212. */
  1213. char *s_subtype;
  1214. const struct dentry_operations *s_d_op; /* default d_op for dentries */
  1215. /*
  1216. * Saved pool identifier for cleancache (-1 means none)
  1217. */
  1218. int cleancache_poolid;
  1219. struct shrinker s_shrink; /* per-sb shrinker handle */
  1220. /* Number of inodes with nlink == 0 but still referenced */
  1221. atomic_long_t s_remove_count;
  1222. /* Being remounted read-only */
  1223. int s_readonly_remount;
  1224. /* AIO completions deferred from interrupt context */
  1225. struct workqueue_struct *s_dio_done_wq;
  1226. struct hlist_head s_pins;
  1227. /*
  1228. * Owning user namespace and default context in which to
  1229. * interpret filesystem uids, gids, quotas, device nodes,
  1230. * xattrs and security labels.
  1231. */
  1232. struct user_namespace *s_user_ns;
  1233. /*
  1234. * Keep the lru lists last in the structure so they always sit on their
  1235. * own individual cachelines.
  1236. */
  1237. struct list_lru s_dentry_lru ____cacheline_aligned_in_smp;
  1238. struct list_lru s_inode_lru ____cacheline_aligned_in_smp;
  1239. struct rcu_head rcu;
  1240. struct work_struct destroy_work;
  1241. struct mutex s_sync_lock; /* sync serialisation lock */
  1242. /*
  1243. * Indicates how deep in a filesystem stack this SB is
  1244. */
  1245. int s_stack_depth;
  1246. /* s_inode_list_lock protects s_inodes */
  1247. spinlock_t s_inode_list_lock ____cacheline_aligned_in_smp;
  1248. struct list_head s_inodes; /* all inodes */
  1249. spinlock_t s_inode_wblist_lock;
  1250. struct list_head s_inodes_wb; /* writeback inodes */
  1251. } __randomize_layout;
  1252. /* Helper functions so that in most cases filesystems will
  1253. * not need to deal directly with kuid_t and kgid_t and can
  1254. * instead deal with the raw numeric values that are stored
  1255. * in the filesystem.
  1256. */
  1257. static inline uid_t i_uid_read(const struct inode *inode)
  1258. {
  1259. return from_kuid(inode->i_sb->s_user_ns, inode->i_uid);
  1260. }
  1261. static inline gid_t i_gid_read(const struct inode *inode)
  1262. {
  1263. return from_kgid(inode->i_sb->s_user_ns, inode->i_gid);
  1264. }
  1265. static inline void i_uid_write(struct inode *inode, uid_t uid)
  1266. {
  1267. inode->i_uid = make_kuid(inode->i_sb->s_user_ns, uid);
  1268. }
  1269. static inline void i_gid_write(struct inode *inode, gid_t gid)
  1270. {
  1271. inode->i_gid = make_kgid(inode->i_sb->s_user_ns, gid);
  1272. }
  1273. extern struct timespec current_time(struct inode *inode);
  1274. /*
  1275. * Snapshotting support.
  1276. */
  1277. void __sb_end_write(struct super_block *sb, int level);
  1278. int __sb_start_write(struct super_block *sb, int level, bool wait);
  1279. #define __sb_writers_acquired(sb, lev) \
  1280. percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
  1281. #define __sb_writers_release(sb, lev) \
  1282. percpu_rwsem_release(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
  1283. /**
  1284. * sb_end_write - drop write access to a superblock
  1285. * @sb: the super we wrote to
  1286. *
  1287. * Decrement number of writers to the filesystem. Wake up possible waiters
  1288. * wanting to freeze the filesystem.
  1289. */
  1290. static inline void sb_end_write(struct super_block *sb)
  1291. {
  1292. __sb_end_write(sb, SB_FREEZE_WRITE);
  1293. }
  1294. /**
  1295. * sb_end_pagefault - drop write access to a superblock from a page fault
  1296. * @sb: the super we wrote to
  1297. *
  1298. * Decrement number of processes handling write page fault to the filesystem.
  1299. * Wake up possible waiters wanting to freeze the filesystem.
  1300. */
  1301. static inline void sb_end_pagefault(struct super_block *sb)
  1302. {
  1303. __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
  1304. }
  1305. /**
  1306. * sb_end_intwrite - drop write access to a superblock for internal fs purposes
  1307. * @sb: the super we wrote to
  1308. *
  1309. * Decrement fs-internal number of writers to the filesystem. Wake up possible
  1310. * waiters wanting to freeze the filesystem.
  1311. */
  1312. static inline void sb_end_intwrite(struct super_block *sb)
  1313. {
  1314. __sb_end_write(sb, SB_FREEZE_FS);
  1315. }
  1316. /**
  1317. * sb_start_write - get write access to a superblock
  1318. * @sb: the super we write to
  1319. *
  1320. * When a process wants to write data or metadata to a file system (i.e. dirty
  1321. * a page or an inode), it should embed the operation in a sb_start_write() -
  1322. * sb_end_write() pair to get exclusion against file system freezing. This
  1323. * function increments number of writers preventing freezing. If the file
  1324. * system is already frozen, the function waits until the file system is
  1325. * thawed.
  1326. *
  1327. * Since freeze protection behaves as a lock, users have to preserve
  1328. * ordering of freeze protection and other filesystem locks. Generally,
  1329. * freeze protection should be the outermost lock. In particular, we have:
  1330. *
  1331. * sb_start_write
  1332. * -> i_mutex (write path, truncate, directory ops, ...)
  1333. * -> s_umount (freeze_super, thaw_super)
  1334. */
  1335. static inline void sb_start_write(struct super_block *sb)
  1336. {
  1337. __sb_start_write(sb, SB_FREEZE_WRITE, true);
  1338. }
  1339. static inline int sb_start_write_trylock(struct super_block *sb)
  1340. {
  1341. return __sb_start_write(sb, SB_FREEZE_WRITE, false);
  1342. }
  1343. /**
  1344. * sb_start_pagefault - get write access to a superblock from a page fault
  1345. * @sb: the super we write to
  1346. *
  1347. * When a process starts handling write page fault, it should embed the
  1348. * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
  1349. * exclusion against file system freezing. This is needed since the page fault
  1350. * is going to dirty a page. This function increments number of running page
  1351. * faults preventing freezing. If the file system is already frozen, the
  1352. * function waits until the file system is thawed.
  1353. *
  1354. * Since page fault freeze protection behaves as a lock, users have to preserve
  1355. * ordering of freeze protection and other filesystem locks. It is advised to
  1356. * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
  1357. * handling code implies lock dependency:
  1358. *
  1359. * mmap_sem
  1360. * -> sb_start_pagefault
  1361. */
  1362. static inline void sb_start_pagefault(struct super_block *sb)
  1363. {
  1364. __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
  1365. }
  1366. /*
  1367. * sb_start_intwrite - get write access to a superblock for internal fs purposes
  1368. * @sb: the super we write to
  1369. *
  1370. * This is the third level of protection against filesystem freezing. It is
  1371. * free for use by a filesystem. The only requirement is that it must rank
  1372. * below sb_start_pagefault.
  1373. *
  1374. * For example filesystem can call sb_start_intwrite() when starting a
  1375. * transaction which somewhat eases handling of freezing for internal sources
  1376. * of filesystem changes (internal fs threads, discarding preallocation on file
  1377. * close, etc.).
  1378. */
  1379. static inline void sb_start_intwrite(struct super_block *sb)
  1380. {
  1381. __sb_start_write(sb, SB_FREEZE_FS, true);
  1382. }
  1383. extern bool inode_owner_or_capable(const struct inode *inode);
  1384. /*
  1385. * VFS helper functions..
  1386. */
  1387. extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
  1388. extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
  1389. extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
  1390. extern int vfs_symlink(struct inode *, struct dentry *, const char *);
  1391. extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
  1392. extern int vfs_rmdir(struct inode *, struct dentry *);
  1393. extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
  1394. extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int);
  1395. extern int vfs_whiteout(struct inode *, struct dentry *);
  1396. extern struct dentry *vfs_tmpfile(struct dentry *dentry, umode_t mode,
  1397. int open_flag);
  1398. /*
  1399. * VFS file helper functions.
  1400. */
  1401. extern void inode_init_owner(struct inode *inode, const struct inode *dir,
  1402. umode_t mode);
  1403. extern bool may_open_dev(const struct path *path);
  1404. /*
  1405. * VFS FS_IOC_FIEMAP helper definitions.
  1406. */
  1407. struct fiemap_extent_info {
  1408. unsigned int fi_flags; /* Flags as passed from user */
  1409. unsigned int fi_extents_mapped; /* Number of mapped extents */
  1410. unsigned int fi_extents_max; /* Size of fiemap_extent array */
  1411. struct fiemap_extent __user *fi_extents_start; /* Start of
  1412. fiemap_extent array */
  1413. };
  1414. int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
  1415. u64 phys, u64 len, u32 flags);
  1416. int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
  1417. /*
  1418. * File types
  1419. *
  1420. * NOTE! These match bits 12..15 of stat.st_mode
  1421. * (ie "(i_mode >> 12) & 15").
  1422. */
  1423. #define DT_UNKNOWN 0
  1424. #define DT_FIFO 1
  1425. #define DT_CHR 2
  1426. #define DT_DIR 4
  1427. #define DT_BLK 6
  1428. #define DT_REG 8
  1429. #define DT_LNK 10
  1430. #define DT_SOCK 12
  1431. #define DT_WHT 14
  1432. /*
  1433. * This is the "filldir" function type, used by readdir() to let
  1434. * the kernel specify what kind of dirent layout it wants to have.
  1435. * This allows the kernel to read directories into kernel space or
  1436. * to have different dirent layouts depending on the binary type.
  1437. */
  1438. struct dir_context;
  1439. typedef int (*filldir_t)(struct dir_context *, const char *, int, loff_t, u64,
  1440. unsigned);
  1441. struct dir_context {
  1442. const filldir_t actor;
  1443. loff_t pos;
  1444. };
  1445. struct block_device_operations;
  1446. /* These macros are for out of kernel modules to test that
  1447. * the kernel supports the unlocked_ioctl and compat_ioctl
  1448. * fields in struct file_operations. */
  1449. #define HAVE_COMPAT_IOCTL 1
  1450. #define HAVE_UNLOCKED_IOCTL 1
  1451. /*
  1452. * These flags let !MMU mmap() govern direct device mapping vs immediate
  1453. * copying more easily for MAP_PRIVATE, especially for ROM filesystems.
  1454. *
  1455. * NOMMU_MAP_COPY: Copy can be mapped (MAP_PRIVATE)
  1456. * NOMMU_MAP_DIRECT: Can be mapped directly (MAP_SHARED)
  1457. * NOMMU_MAP_READ: Can be mapped for reading
  1458. * NOMMU_MAP_WRITE: Can be mapped for writing
  1459. * NOMMU_MAP_EXEC: Can be mapped for execution
  1460. */
  1461. #define NOMMU_MAP_COPY 0x00000001
  1462. #define NOMMU_MAP_DIRECT 0x00000008
  1463. #define NOMMU_MAP_READ VM_MAYREAD
  1464. #define NOMMU_MAP_WRITE VM_MAYWRITE
  1465. #define NOMMU_MAP_EXEC VM_MAYEXEC
  1466. #define NOMMU_VMFLAGS \
  1467. (NOMMU_MAP_READ | NOMMU_MAP_WRITE | NOMMU_MAP_EXEC)
  1468. struct iov_iter;
  1469. struct file_operations {
  1470. struct module *owner;
  1471. loff_t (*llseek) (struct file *, loff_t, int);
  1472. ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
  1473. ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
  1474. ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
  1475. ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
  1476. int (*iterate) (struct file *, struct dir_context *);
  1477. int (*iterate_shared) (struct file *, struct dir_context *);
  1478. unsigned int (*poll) (struct file *, struct poll_table_struct *);
  1479. long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
  1480. long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
  1481. int (*mmap) (struct file *, struct vm_area_struct *);
  1482. int (*open) (struct inode *, struct file *);
  1483. int (*flush) (struct file *, fl_owner_t id);
  1484. int (*release) (struct inode *, struct file *);
  1485. int (*fsync) (struct file *, loff_t, loff_t, int datasync);
  1486. int (*fasync) (int, struct file *, int);
  1487. int (*lock) (struct file *, int, struct file_lock *);
  1488. ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
  1489. unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
  1490. int (*check_flags)(int);
  1491. int (*flock) (struct file *, int, struct file_lock *);
  1492. ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
  1493. ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
  1494. int (*setlease)(struct file *, long, struct file_lock **, void **);
  1495. long (*fallocate)(struct file *file, int mode, loff_t offset,
  1496. loff_t len);
  1497. void (*show_fdinfo)(struct seq_file *m, struct file *f);
  1498. #ifndef CONFIG_MMU
  1499. unsigned (*mmap_capabilities)(struct file *);
  1500. #endif
  1501. ssize_t (*copy_file_range)(struct file *, loff_t, struct file *,
  1502. loff_t, size_t, unsigned int);
  1503. int (*clone_file_range)(struct file *, loff_t, struct file *, loff_t,
  1504. u64);
  1505. ssize_t (*dedupe_file_range)(struct file *, u64, u64, struct file *,
  1506. u64);
  1507. } __randomize_layout;
  1508. struct inode_operations {
  1509. struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
  1510. const char * (*get_link) (struct dentry *, struct inode *, struct delayed_call *);
  1511. int (*permission) (struct inode *, int);
  1512. struct posix_acl * (*get_acl)(struct inode *, int);
  1513. int (*readlink) (struct dentry *, char __user *,int);
  1514. int (*create) (struct inode *,struct dentry *, umode_t, bool);
  1515. int (*link) (struct dentry *,struct inode *,struct dentry *);
  1516. int (*unlink) (struct inode *,struct dentry *);
  1517. int (*symlink) (struct inode *,struct dentry *,const char *);
  1518. int (*mkdir) (struct inode *,struct dentry *,umode_t);
  1519. int (*rmdir) (struct inode *,struct dentry *);
  1520. int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
  1521. int (*rename) (struct inode *, struct dentry *,
  1522. struct inode *, struct dentry *, unsigned int);
  1523. int (*setattr) (struct dentry *, struct iattr *);
  1524. int (*getattr) (const struct path *, struct kstat *, u32, unsigned int);
  1525. ssize_t (*listxattr) (struct dentry *, char *, size_t);
  1526. int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
  1527. u64 len);
  1528. int (*update_time)(struct inode *, struct timespec *, int);
  1529. int (*atomic_open)(struct inode *, struct dentry *,
  1530. struct file *, unsigned open_flag,
  1531. umode_t create_mode, int *opened);
  1532. int (*tmpfile) (struct inode *, struct dentry *, umode_t);
  1533. int (*set_acl)(struct inode *, struct posix_acl *, int);
  1534. } ____cacheline_aligned;
  1535. static inline ssize_t call_read_iter(struct file *file, struct kiocb *kio,
  1536. struct iov_iter *iter)
  1537. {
  1538. return file->f_op->read_iter(kio, iter);
  1539. }
  1540. static inline ssize_t call_write_iter(struct file *file, struct kiocb *kio,
  1541. struct iov_iter *iter)
  1542. {
  1543. return file->f_op->write_iter(kio, iter);
  1544. }
  1545. static inline int call_mmap(struct file *file, struct vm_area_struct *vma)
  1546. {
  1547. return file->f_op->mmap(file, vma);
  1548. }
  1549. ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
  1550. unsigned long nr_segs, unsigned long fast_segs,
  1551. struct iovec *fast_pointer,
  1552. struct iovec **ret_pointer);
  1553. extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *);
  1554. extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
  1555. extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
  1556. extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
  1557. unsigned long, loff_t *, rwf_t);
  1558. extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *,
  1559. loff_t, size_t, unsigned int);
  1560. extern int vfs_clone_file_prep_inodes(struct inode *inode_in, loff_t pos_in,
  1561. struct inode *inode_out, loff_t pos_out,
  1562. u64 *len, bool is_dedupe);
  1563. extern int vfs_clone_file_range(struct file *file_in, loff_t pos_in,
  1564. struct file *file_out, loff_t pos_out, u64 len);
  1565. extern int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff,
  1566. struct inode *dest, loff_t destoff,
  1567. loff_t len, bool *is_same);
  1568. extern int vfs_dedupe_file_range(struct file *file,
  1569. struct file_dedupe_range *same);
  1570. struct super_operations {
  1571. struct inode *(*alloc_inode)(struct super_block *sb);
  1572. void (*destroy_inode)(struct inode *);
  1573. void (*dirty_inode) (struct inode *, int flags);
  1574. int (*write_inode) (struct inode *, struct writeback_control *wbc);
  1575. int (*drop_inode) (struct inode *);
  1576. void (*evict_inode) (struct inode *);
  1577. void (*put_super) (struct super_block *);
  1578. int (*sync_fs)(struct super_block *sb, int wait);
  1579. int (*freeze_super) (struct super_block *);
  1580. int (*freeze_fs) (struct super_block *);
  1581. int (*thaw_super) (struct super_block *);
  1582. int (*unfreeze_fs) (struct super_block *);
  1583. int (*statfs) (struct dentry *, struct kstatfs *);
  1584. int (*remount_fs) (struct super_block *, int *, char *);
  1585. void (*umount_begin) (struct super_block *);
  1586. int (*show_options)(struct seq_file *, struct dentry *);
  1587. int (*show_devname)(struct seq_file *, struct dentry *);
  1588. int (*show_path)(struct seq_file *, struct dentry *);
  1589. int (*show_stats)(struct seq_file *, struct dentry *);
  1590. #ifdef CONFIG_QUOTA
  1591. ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
  1592. ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
  1593. struct dquot **(*get_dquots)(struct inode *);
  1594. #endif
  1595. int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
  1596. long (*nr_cached_objects)(struct super_block *,
  1597. struct shrink_control *);
  1598. long (*free_cached_objects)(struct super_block *,
  1599. struct shrink_control *);
  1600. };
  1601. /*
  1602. * Inode flags - they have no relation to superblock flags now
  1603. */
  1604. #define S_SYNC 1 /* Writes are synced at once */
  1605. #define S_NOATIME 2 /* Do not update access times */
  1606. #define S_APPEND 4 /* Append-only file */
  1607. #define S_IMMUTABLE 8 /* Immutable file */
  1608. #define S_DEAD 16 /* removed, but still open directory */
  1609. #define S_NOQUOTA 32 /* Inode is not counted to quota */
  1610. #define S_DIRSYNC 64 /* Directory modifications are synchronous */
  1611. #define S_NOCMTIME 128 /* Do not update file c/mtime */
  1612. #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
  1613. #define S_PRIVATE 512 /* Inode is fs-internal */
  1614. #define S_IMA 1024 /* Inode has an associated IMA struct */
  1615. #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
  1616. #define S_NOSEC 4096 /* no suid or xattr security attributes */
  1617. #ifdef CONFIG_FS_DAX
  1618. #define S_DAX 8192 /* Direct Access, avoiding the page cache */
  1619. #else
  1620. #define S_DAX 0 /* Make all the DAX code disappear */
  1621. #endif
  1622. /*
  1623. * Note that nosuid etc flags are inode-specific: setting some file-system
  1624. * flags just means all the inodes inherit those flags by default. It might be
  1625. * possible to override it selectively if you really wanted to with some
  1626. * ioctl() that is not currently implemented.
  1627. *
  1628. * Exception: SB_RDONLY is always applied to the entire file system.
  1629. *
  1630. * Unfortunately, it is possible to change a filesystems flags with it mounted
  1631. * with files in use. This means that all of the inodes will not have their
  1632. * i_flags updated. Hence, i_flags no longer inherit the superblock mount
  1633. * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
  1634. */
  1635. #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
  1636. static inline bool sb_rdonly(const struct super_block *sb) { return sb->s_flags & MS_RDONLY; }
  1637. #define IS_RDONLY(inode) sb_rdonly((inode)->i_sb)
  1638. #define IS_SYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS) || \
  1639. ((inode)->i_flags & S_SYNC))
  1640. #define IS_DIRSYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS|SB_DIRSYNC) || \
  1641. ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
  1642. #define IS_MANDLOCK(inode) __IS_FLG(inode, SB_MANDLOCK)
  1643. #define IS_NOATIME(inode) __IS_FLG(inode, SB_RDONLY|SB_NOATIME)
  1644. #define IS_I_VERSION(inode) __IS_FLG(inode, SB_I_VERSION)
  1645. #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
  1646. #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
  1647. #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
  1648. #define IS_POSIXACL(inode) __IS_FLG(inode, SB_POSIXACL)
  1649. #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
  1650. #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
  1651. #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
  1652. #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
  1653. #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
  1654. #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
  1655. #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
  1656. #define IS_DAX(inode) ((inode)->i_flags & S_DAX)
  1657. #define IS_WHITEOUT(inode) (S_ISCHR(inode->i_mode) && \
  1658. (inode)->i_rdev == WHITEOUT_DEV)
  1659. static inline bool HAS_UNMAPPED_ID(struct inode *inode)
  1660. {
  1661. return !uid_valid(inode->i_uid) || !gid_valid(inode->i_gid);
  1662. }
  1663. static inline enum rw_hint file_write_hint(struct file *file)
  1664. {
  1665. if (file->f_write_hint != WRITE_LIFE_NOT_SET)
  1666. return file->f_write_hint;
  1667. return file_inode(file)->i_write_hint;
  1668. }
  1669. static inline int iocb_flags(struct file *file);
  1670. static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp)
  1671. {
  1672. *kiocb = (struct kiocb) {
  1673. .ki_filp = filp,
  1674. .ki_flags = iocb_flags(filp),
  1675. .ki_hint = file_write_hint(filp),
  1676. };
  1677. }
  1678. /*
  1679. * Inode state bits. Protected by inode->i_lock
  1680. *
  1681. * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
  1682. * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
  1683. *
  1684. * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
  1685. * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
  1686. * various stages of removing an inode.
  1687. *
  1688. * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
  1689. *
  1690. * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
  1691. * fdatasync(). i_atime is the usual cause.
  1692. * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
  1693. * these changes separately from I_DIRTY_SYNC so that we
  1694. * don't have to write inode on fdatasync() when only
  1695. * mtime has changed in it.
  1696. * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
  1697. * I_NEW Serves as both a mutex and completion notification.
  1698. * New inodes set I_NEW. If two processes both create
  1699. * the same inode, one of them will release its inode and
  1700. * wait for I_NEW to be released before returning.
  1701. * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
  1702. * also cause waiting on I_NEW, without I_NEW actually
  1703. * being set. find_inode() uses this to prevent returning
  1704. * nearly-dead inodes.
  1705. * I_WILL_FREE Must be set when calling write_inode_now() if i_count
  1706. * is zero. I_FREEING must be set when I_WILL_FREE is
  1707. * cleared.
  1708. * I_FREEING Set when inode is about to be freed but still has dirty
  1709. * pages or buffers attached or the inode itself is still
  1710. * dirty.
  1711. * I_CLEAR Added by clear_inode(). In this state the inode is
  1712. * clean and can be destroyed. Inode keeps I_FREEING.
  1713. *
  1714. * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
  1715. * prohibited for many purposes. iget() must wait for
  1716. * the inode to be completely released, then create it
  1717. * anew. Other functions will just ignore such inodes,
  1718. * if appropriate. I_NEW is used for waiting.
  1719. *
  1720. * I_SYNC Writeback of inode is running. The bit is set during
  1721. * data writeback, and cleared with a wakeup on the bit
  1722. * address once it is done. The bit is also used to pin
  1723. * the inode in memory for flusher thread.
  1724. *
  1725. * I_REFERENCED Marks the inode as recently references on the LRU list.
  1726. *
  1727. * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
  1728. *
  1729. * I_WB_SWITCH Cgroup bdi_writeback switching in progress. Used to
  1730. * synchronize competing switching instances and to tell
  1731. * wb stat updates to grab mapping->tree_lock. See
  1732. * inode_switch_wb_work_fn() for details.
  1733. *
  1734. * I_OVL_INUSE Used by overlayfs to get exclusive ownership on upper
  1735. * and work dirs among overlayfs mounts.
  1736. *
  1737. * Q: What is the difference between I_WILL_FREE and I_FREEING?
  1738. */
  1739. #define I_DIRTY_SYNC (1 << 0)
  1740. #define I_DIRTY_DATASYNC (1 << 1)
  1741. #define I_DIRTY_PAGES (1 << 2)
  1742. #define __I_NEW 3
  1743. #define I_NEW (1 << __I_NEW)
  1744. #define I_WILL_FREE (1 << 4)
  1745. #define I_FREEING (1 << 5)
  1746. #define I_CLEAR (1 << 6)
  1747. #define __I_SYNC 7
  1748. #define I_SYNC (1 << __I_SYNC)
  1749. #define I_REFERENCED (1 << 8)
  1750. #define __I_DIO_WAKEUP 9
  1751. #define I_DIO_WAKEUP (1 << __I_DIO_WAKEUP)
  1752. #define I_LINKABLE (1 << 10)
  1753. #define I_DIRTY_TIME (1 << 11)
  1754. #define __I_DIRTY_TIME_EXPIRED 12
  1755. #define I_DIRTY_TIME_EXPIRED (1 << __I_DIRTY_TIME_EXPIRED)
  1756. #define I_WB_SWITCH (1 << 13)
  1757. #define I_OVL_INUSE (1 << 14)
  1758. #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
  1759. #define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME)
  1760. extern void __mark_inode_dirty(struct inode *, int);
  1761. static inline void mark_inode_dirty(struct inode *inode)
  1762. {
  1763. __mark_inode_dirty(inode, I_DIRTY);
  1764. }
  1765. static inline void mark_inode_dirty_sync(struct inode *inode)
  1766. {
  1767. __mark_inode_dirty(inode, I_DIRTY_SYNC);
  1768. }
  1769. extern void inc_nlink(struct inode *inode);
  1770. extern void drop_nlink(struct inode *inode);
  1771. extern void clear_nlink(struct inode *inode);
  1772. extern void set_nlink(struct inode *inode, unsigned int nlink);
  1773. static inline void inode_inc_link_count(struct inode *inode)
  1774. {
  1775. inc_nlink(inode);
  1776. mark_inode_dirty(inode);
  1777. }
  1778. static inline void inode_dec_link_count(struct inode *inode)
  1779. {
  1780. drop_nlink(inode);
  1781. mark_inode_dirty(inode);
  1782. }
  1783. /**
  1784. * inode_inc_iversion - increments i_version
  1785. * @inode: inode that need to be updated
  1786. *
  1787. * Every time the inode is modified, the i_version field will be incremented.
  1788. * The filesystem has to be mounted with i_version flag
  1789. */
  1790. static inline void inode_inc_iversion(struct inode *inode)
  1791. {
  1792. spin_lock(&inode->i_lock);
  1793. inode->i_version++;
  1794. spin_unlock(&inode->i_lock);
  1795. }
  1796. enum file_time_flags {
  1797. S_ATIME = 1,
  1798. S_MTIME = 2,
  1799. S_CTIME = 4,
  1800. S_VERSION = 8,
  1801. };
  1802. extern void touch_atime(const struct path *);
  1803. static inline void file_accessed(struct file *file)
  1804. {
  1805. if (!(file->f_flags & O_NOATIME))
  1806. touch_atime(&file->f_path);
  1807. }
  1808. int sync_inode(struct inode *inode, struct writeback_control *wbc);
  1809. int sync_inode_metadata(struct inode *inode, int wait);
  1810. struct file_system_type {
  1811. const char *name;
  1812. int fs_flags;
  1813. #define FS_REQUIRES_DEV 1
  1814. #define FS_BINARY_MOUNTDATA 2
  1815. #define FS_HAS_SUBTYPE 4
  1816. #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
  1817. #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
  1818. struct dentry *(*mount) (struct file_system_type *, int,
  1819. const char *, void *);
  1820. void (*kill_sb) (struct super_block *);
  1821. struct module *owner;
  1822. struct file_system_type * next;
  1823. struct hlist_head fs_supers;
  1824. struct lock_class_key s_lock_key;
  1825. struct lock_class_key s_umount_key;
  1826. struct lock_class_key s_vfs_rename_key;
  1827. struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
  1828. struct lock_class_key i_lock_key;
  1829. struct lock_class_key i_mutex_key;
  1830. struct lock_class_key i_mutex_dir_key;
  1831. };
  1832. #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
  1833. extern struct dentry *mount_ns(struct file_system_type *fs_type,
  1834. int flags, void *data, void *ns, struct user_namespace *user_ns,
  1835. int (*fill_super)(struct super_block *, void *, int));
  1836. extern struct dentry *mount_bdev(struct file_system_type *fs_type,
  1837. int flags, const char *dev_name, void *data,
  1838. int (*fill_super)(struct super_block *, void *, int));
  1839. extern struct dentry *mount_single(struct file_system_type *fs_type,
  1840. int flags, void *data,
  1841. int (*fill_super)(struct super_block *, void *, int));
  1842. extern struct dentry *mount_nodev(struct file_system_type *fs_type,
  1843. int flags, void *data,
  1844. int (*fill_super)(struct super_block *, void *, int));
  1845. extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
  1846. void generic_shutdown_super(struct super_block *sb);
  1847. void kill_block_super(struct super_block *sb);
  1848. void kill_anon_super(struct super_block *sb);
  1849. void kill_litter_super(struct super_block *sb);
  1850. void deactivate_super(struct super_block *sb);
  1851. void deactivate_locked_super(struct super_block *sb);
  1852. int set_anon_super(struct super_block *s, void *data);
  1853. int get_anon_bdev(dev_t *);
  1854. void free_anon_bdev(dev_t);
  1855. struct super_block *sget_userns(struct file_system_type *type,
  1856. int (*test)(struct super_block *,void *),
  1857. int (*set)(struct super_block *,void *),
  1858. int flags, struct user_namespace *user_ns,
  1859. void *data);
  1860. struct super_block *sget(struct file_system_type *type,
  1861. int (*test)(struct super_block *,void *),
  1862. int (*set)(struct super_block *,void *),
  1863. int flags, void *data);
  1864. extern struct dentry *mount_pseudo_xattr(struct file_system_type *, char *,
  1865. const struct super_operations *ops,
  1866. const struct xattr_handler **xattr,
  1867. const struct dentry_operations *dops,
  1868. unsigned long);
  1869. static inline struct dentry *
  1870. mount_pseudo(struct file_system_type *fs_type, char *name,
  1871. const struct super_operations *ops,
  1872. const struct dentry_operations *dops, unsigned long magic)
  1873. {
  1874. return mount_pseudo_xattr(fs_type, name, ops, NULL, dops, magic);
  1875. }
  1876. /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
  1877. #define fops_get(fops) \
  1878. (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
  1879. #define fops_put(fops) \
  1880. do { if (fops) module_put((fops)->owner); } while(0)
  1881. /*
  1882. * This one is to be used *ONLY* from ->open() instances.
  1883. * fops must be non-NULL, pinned down *and* module dependencies
  1884. * should be sufficient to pin the caller down as well.
  1885. */
  1886. #define replace_fops(f, fops) \
  1887. do { \
  1888. struct file *__file = (f); \
  1889. fops_put(__file->f_op); \
  1890. BUG_ON(!(__file->f_op = (fops))); \
  1891. } while(0)
  1892. extern int register_filesystem(struct file_system_type *);
  1893. extern int unregister_filesystem(struct file_system_type *);
  1894. extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
  1895. #define kern_mount(type) kern_mount_data(type, NULL)
  1896. extern void kern_unmount(struct vfsmount *mnt);
  1897. extern int may_umount_tree(struct vfsmount *);
  1898. extern int may_umount(struct vfsmount *);
  1899. extern long do_mount(const char *, const char __user *,
  1900. const char *, unsigned long, void *);
  1901. extern struct vfsmount *collect_mounts(const struct path *);
  1902. extern void drop_collected_mounts(struct vfsmount *);
  1903. extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
  1904. struct vfsmount *);
  1905. extern int vfs_statfs(const struct path *, struct kstatfs *);
  1906. extern int user_statfs(const char __user *, struct kstatfs *);
  1907. extern int fd_statfs(int, struct kstatfs *);
  1908. extern int freeze_super(struct super_block *super);
  1909. extern int thaw_super(struct super_block *super);
  1910. extern bool our_mnt(struct vfsmount *mnt);
  1911. extern __printf(2, 3)
  1912. int super_setup_bdi_name(struct super_block *sb, char *fmt, ...);
  1913. extern int super_setup_bdi(struct super_block *sb);
  1914. extern int current_umask(void);
  1915. extern void ihold(struct inode * inode);
  1916. extern void iput(struct inode *);
  1917. extern int generic_update_time(struct inode *, struct timespec *, int);
  1918. /* /sys/fs */
  1919. extern struct kobject *fs_kobj;
  1920. #define MAX_RW_COUNT (INT_MAX & PAGE_MASK)
  1921. #ifdef CONFIG_MANDATORY_FILE_LOCKING
  1922. extern int locks_mandatory_locked(struct file *);
  1923. extern int locks_mandatory_area(struct inode *, struct file *, loff_t, loff_t, unsigned char);
  1924. /*
  1925. * Candidates for mandatory locking have the setgid bit set
  1926. * but no group execute bit - an otherwise meaningless combination.
  1927. */
  1928. static inline int __mandatory_lock(struct inode *ino)
  1929. {
  1930. return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
  1931. }
  1932. /*
  1933. * ... and these candidates should be on SB_MANDLOCK mounted fs,
  1934. * otherwise these will be advisory locks
  1935. */
  1936. static inline int mandatory_lock(struct inode *ino)
  1937. {
  1938. return IS_MANDLOCK(ino) && __mandatory_lock(ino);
  1939. }
  1940. static inline int locks_verify_locked(struct file *file)
  1941. {
  1942. if (mandatory_lock(locks_inode(file)))
  1943. return locks_mandatory_locked(file);
  1944. return 0;
  1945. }
  1946. static inline int locks_verify_truncate(struct inode *inode,
  1947. struct file *f,
  1948. loff_t size)
  1949. {
  1950. if (!inode->i_flctx || !mandatory_lock(inode))
  1951. return 0;
  1952. if (size < inode->i_size) {
  1953. return locks_mandatory_area(inode, f, size, inode->i_size - 1,
  1954. F_WRLCK);
  1955. } else {
  1956. return locks_mandatory_area(inode, f, inode->i_size, size - 1,
  1957. F_WRLCK);
  1958. }
  1959. }
  1960. #else /* !CONFIG_MANDATORY_FILE_LOCKING */
  1961. static inline int locks_mandatory_locked(struct file *file)
  1962. {
  1963. return 0;
  1964. }
  1965. static inline int locks_mandatory_area(struct inode *inode, struct file *filp,
  1966. loff_t start, loff_t end, unsigned char type)
  1967. {
  1968. return 0;
  1969. }
  1970. static inline int __mandatory_lock(struct inode *inode)
  1971. {
  1972. return 0;
  1973. }
  1974. static inline int mandatory_lock(struct inode *inode)
  1975. {
  1976. return 0;
  1977. }
  1978. static inline int locks_verify_locked(struct file *file)
  1979. {
  1980. return 0;
  1981. }
  1982. static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
  1983. size_t size)
  1984. {
  1985. return 0;
  1986. }
  1987. #endif /* CONFIG_MANDATORY_FILE_LOCKING */
  1988. #ifdef CONFIG_FILE_LOCKING
  1989. static inline int break_lease(struct inode *inode, unsigned int mode)
  1990. {
  1991. /*
  1992. * Since this check is lockless, we must ensure that any refcounts
  1993. * taken are done before checking i_flctx->flc_lease. Otherwise, we
  1994. * could end up racing with tasks trying to set a new lease on this
  1995. * file.
  1996. */
  1997. smp_mb();
  1998. if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
  1999. return __break_lease(inode, mode, FL_LEASE);
  2000. return 0;
  2001. }
  2002. static inline int break_deleg(struct inode *inode, unsigned int mode)
  2003. {
  2004. /*
  2005. * Since this check is lockless, we must ensure that any refcounts
  2006. * taken are done before checking i_flctx->flc_lease. Otherwise, we
  2007. * could end up racing with tasks trying to set a new lease on this
  2008. * file.
  2009. */
  2010. smp_mb();
  2011. if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
  2012. return __break_lease(inode, mode, FL_DELEG);
  2013. return 0;
  2014. }
  2015. static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
  2016. {
  2017. int ret;
  2018. ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
  2019. if (ret == -EWOULDBLOCK && delegated_inode) {
  2020. *delegated_inode = inode;
  2021. ihold(inode);
  2022. }
  2023. return ret;
  2024. }
  2025. static inline int break_deleg_wait(struct inode **delegated_inode)
  2026. {
  2027. int ret;
  2028. ret = break_deleg(*delegated_inode, O_WRONLY);
  2029. iput(*delegated_inode);
  2030. *delegated_inode = NULL;
  2031. return ret;
  2032. }
  2033. static inline int break_layout(struct inode *inode, bool wait)
  2034. {
  2035. smp_mb();
  2036. if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
  2037. return __break_lease(inode,
  2038. wait ? O_WRONLY : O_WRONLY | O_NONBLOCK,
  2039. FL_LAYOUT);
  2040. return 0;
  2041. }
  2042. #else /* !CONFIG_FILE_LOCKING */
  2043. static inline int break_lease(struct inode *inode, unsigned int mode)
  2044. {
  2045. return 0;
  2046. }
  2047. static inline int break_deleg(struct inode *inode, unsigned int mode)
  2048. {
  2049. return 0;
  2050. }
  2051. static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
  2052. {
  2053. return 0;
  2054. }
  2055. static inline int break_deleg_wait(struct inode **delegated_inode)
  2056. {
  2057. BUG();
  2058. return 0;
  2059. }
  2060. static inline int break_layout(struct inode *inode, bool wait)
  2061. {
  2062. return 0;
  2063. }
  2064. #endif /* CONFIG_FILE_LOCKING */
  2065. /* fs/open.c */
  2066. struct audit_names;
  2067. struct filename {
  2068. const char *name; /* pointer to actual string */
  2069. const __user char *uptr; /* original userland pointer */
  2070. struct audit_names *aname;
  2071. int refcnt;
  2072. const char iname[];
  2073. };
  2074. extern long vfs_truncate(const struct path *, loff_t);
  2075. extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
  2076. struct file *filp);
  2077. extern int vfs_fallocate(struct file *file, int mode, loff_t offset,
  2078. loff_t len);
  2079. extern long do_sys_open(int dfd, const char __user *filename, int flags,
  2080. umode_t mode);
  2081. extern struct file *file_open_name(struct filename *, int, umode_t);
  2082. extern struct file *filp_open(const char *, int, umode_t);
  2083. extern struct file *file_open_root(struct dentry *, struct vfsmount *,
  2084. const char *, int, umode_t);
  2085. extern struct file * dentry_open(const struct path *, int, const struct cred *);
  2086. extern int filp_close(struct file *, fl_owner_t id);
  2087. extern struct filename *getname_flags(const char __user *, int, int *);
  2088. extern struct filename *getname(const char __user *);
  2089. extern struct filename *getname_kernel(const char *);
  2090. extern void putname(struct filename *name);
  2091. enum {
  2092. FILE_CREATED = 1,
  2093. FILE_OPENED = 2
  2094. };
  2095. extern int finish_open(struct file *file, struct dentry *dentry,
  2096. int (*open)(struct inode *, struct file *),
  2097. int *opened);
  2098. extern int finish_no_open(struct file *file, struct dentry *dentry);
  2099. /* fs/ioctl.c */
  2100. extern int ioctl_preallocate(struct file *filp, void __user *argp);
  2101. /* fs/dcache.c */
  2102. extern void __init vfs_caches_init_early(void);
  2103. extern void __init vfs_caches_init(void);
  2104. extern struct kmem_cache *names_cachep;
  2105. #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
  2106. #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
  2107. #ifdef CONFIG_BLOCK
  2108. extern int register_blkdev(unsigned int, const char *);
  2109. extern void unregister_blkdev(unsigned int, const char *);
  2110. extern void bdev_unhash_inode(dev_t dev);
  2111. extern struct block_device *bdget(dev_t);
  2112. extern struct block_device *bdgrab(struct block_device *bdev);
  2113. extern void bd_set_size(struct block_device *, loff_t size);
  2114. extern void bd_forget(struct inode *inode);
  2115. extern void bdput(struct block_device *);
  2116. extern void invalidate_bdev(struct block_device *);
  2117. extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
  2118. extern int sync_blockdev(struct block_device *bdev);
  2119. extern void kill_bdev(struct block_device *);
  2120. extern struct super_block *freeze_bdev(struct block_device *);
  2121. extern void emergency_thaw_all(void);
  2122. extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
  2123. extern int fsync_bdev(struct block_device *);
  2124. extern struct super_block *blockdev_superblock;
  2125. static inline bool sb_is_blkdev_sb(struct super_block *sb)
  2126. {
  2127. return sb == blockdev_superblock;
  2128. }
  2129. #else
  2130. static inline void bd_forget(struct inode *inode) {}
  2131. static inline int sync_blockdev(struct block_device *bdev) { return 0; }
  2132. static inline void kill_bdev(struct block_device *bdev) {}
  2133. static inline void invalidate_bdev(struct block_device *bdev) {}
  2134. static inline struct super_block *freeze_bdev(struct block_device *sb)
  2135. {
  2136. return NULL;
  2137. }
  2138. static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
  2139. {
  2140. return 0;
  2141. }
  2142. static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
  2143. {
  2144. }
  2145. static inline bool sb_is_blkdev_sb(struct super_block *sb)
  2146. {
  2147. return false;
  2148. }
  2149. #endif
  2150. extern int sync_filesystem(struct super_block *);
  2151. extern const struct file_operations def_blk_fops;
  2152. extern const struct file_operations def_chr_fops;
  2153. #ifdef CONFIG_BLOCK
  2154. extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
  2155. extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
  2156. extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
  2157. extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
  2158. extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
  2159. void *holder);
  2160. extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
  2161. void *holder);
  2162. extern void blkdev_put(struct block_device *bdev, fmode_t mode);
  2163. extern int __blkdev_reread_part(struct block_device *bdev);
  2164. extern int blkdev_reread_part(struct block_device *bdev);
  2165. #ifdef CONFIG_SYSFS
  2166. extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
  2167. extern void bd_unlink_disk_holder(struct block_device *bdev,
  2168. struct gendisk *disk);
  2169. #else
  2170. static inline int bd_link_disk_holder(struct block_device *bdev,
  2171. struct gendisk *disk)
  2172. {
  2173. return 0;
  2174. }
  2175. static inline void bd_unlink_disk_holder(struct block_device *bdev,
  2176. struct gendisk *disk)
  2177. {
  2178. }
  2179. #endif
  2180. #endif
  2181. /* fs/char_dev.c */
  2182. #define CHRDEV_MAJOR_MAX 512
  2183. /* Marks the bottom of the first segment of free char majors */
  2184. #define CHRDEV_MAJOR_DYN_END 234
  2185. /* Marks the top and bottom of the second segment of free char majors */
  2186. #define CHRDEV_MAJOR_DYN_EXT_START 511
  2187. #define CHRDEV_MAJOR_DYN_EXT_END 384
  2188. extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
  2189. extern int register_chrdev_region(dev_t, unsigned, const char *);
  2190. extern int __register_chrdev(unsigned int major, unsigned int baseminor,
  2191. unsigned int count, const char *name,
  2192. const struct file_operations *fops);
  2193. extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
  2194. unsigned int count, const char *name);
  2195. extern void unregister_chrdev_region(dev_t, unsigned);
  2196. extern void chrdev_show(struct seq_file *,off_t);
  2197. static inline int register_chrdev(unsigned int major, const char *name,
  2198. const struct file_operations *fops)
  2199. {
  2200. return __register_chrdev(major, 0, 256, name, fops);
  2201. }
  2202. static inline void unregister_chrdev(unsigned int major, const char *name)
  2203. {
  2204. __unregister_chrdev(major, 0, 256, name);
  2205. }
  2206. /* fs/block_dev.c */
  2207. #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
  2208. #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
  2209. #ifdef CONFIG_BLOCK
  2210. #define BLKDEV_MAJOR_MAX 512
  2211. extern const char *__bdevname(dev_t, char *buffer);
  2212. extern const char *bdevname(struct block_device *bdev, char *buffer);
  2213. extern struct block_device *lookup_bdev(const char *);
  2214. extern void blkdev_show(struct seq_file *,off_t);
  2215. #else
  2216. #define BLKDEV_MAJOR_MAX 0
  2217. #endif
  2218. extern void init_special_inode(struct inode *, umode_t, dev_t);
  2219. /* Invalid inode operations -- fs/bad_inode.c */
  2220. extern void make_bad_inode(struct inode *);
  2221. extern bool is_bad_inode(struct inode *);
  2222. #ifdef CONFIG_BLOCK
  2223. extern void check_disk_size_change(struct gendisk *disk,
  2224. struct block_device *bdev);
  2225. extern int revalidate_disk(struct gendisk *);
  2226. extern int check_disk_change(struct block_device *);
  2227. extern int __invalidate_device(struct block_device *, bool);
  2228. extern int invalidate_partition(struct gendisk *, int);
  2229. #endif
  2230. unsigned long invalidate_mapping_pages(struct address_space *mapping,
  2231. pgoff_t start, pgoff_t end);
  2232. static inline void invalidate_remote_inode(struct inode *inode)
  2233. {
  2234. if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
  2235. S_ISLNK(inode->i_mode))
  2236. invalidate_mapping_pages(inode->i_mapping, 0, -1);
  2237. }
  2238. extern int invalidate_inode_pages2(struct address_space *mapping);
  2239. extern int invalidate_inode_pages2_range(struct address_space *mapping,
  2240. pgoff_t start, pgoff_t end);
  2241. extern int write_inode_now(struct inode *, int);
  2242. extern int filemap_fdatawrite(struct address_space *);
  2243. extern int filemap_flush(struct address_space *);
  2244. extern int filemap_fdatawait_keep_errors(struct address_space *mapping);
  2245. extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
  2246. loff_t lend);
  2247. static inline int filemap_fdatawait(struct address_space *mapping)
  2248. {
  2249. return filemap_fdatawait_range(mapping, 0, LLONG_MAX);
  2250. }
  2251. extern bool filemap_range_has_page(struct address_space *, loff_t lstart,
  2252. loff_t lend);
  2253. extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart,
  2254. loff_t lend);
  2255. extern int filemap_write_and_wait(struct address_space *mapping);
  2256. extern int filemap_write_and_wait_range(struct address_space *mapping,
  2257. loff_t lstart, loff_t lend);
  2258. extern int __filemap_fdatawrite_range(struct address_space *mapping,
  2259. loff_t start, loff_t end, int sync_mode);
  2260. extern int filemap_fdatawrite_range(struct address_space *mapping,
  2261. loff_t start, loff_t end);
  2262. extern int filemap_check_errors(struct address_space *mapping);
  2263. extern void __filemap_set_wb_err(struct address_space *mapping, int err);
  2264. extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart,
  2265. loff_t lend);
  2266. extern int __must_check file_check_and_advance_wb_err(struct file *file);
  2267. extern int __must_check file_write_and_wait_range(struct file *file,
  2268. loff_t start, loff_t end);
  2269. static inline int file_write_and_wait(struct file *file)
  2270. {
  2271. return file_write_and_wait_range(file, 0, LLONG_MAX);
  2272. }
  2273. /**
  2274. * filemap_set_wb_err - set a writeback error on an address_space
  2275. * @mapping: mapping in which to set writeback error
  2276. * @err: error to be set in mapping
  2277. *
  2278. * When writeback fails in some way, we must record that error so that
  2279. * userspace can be informed when fsync and the like are called. We endeavor
  2280. * to report errors on any file that was open at the time of the error. Some
  2281. * internal callers also need to know when writeback errors have occurred.
  2282. *
  2283. * When a writeback error occurs, most filesystems will want to call
  2284. * filemap_set_wb_err to record the error in the mapping so that it will be
  2285. * automatically reported whenever fsync is called on the file.
  2286. */
  2287. static inline void filemap_set_wb_err(struct address_space *mapping, int err)
  2288. {
  2289. /* Fastpath for common case of no error */
  2290. if (unlikely(err))
  2291. __filemap_set_wb_err(mapping, err);
  2292. }
  2293. /**
  2294. * filemap_check_wb_error - has an error occurred since the mark was sampled?
  2295. * @mapping: mapping to check for writeback errors
  2296. * @since: previously-sampled errseq_t
  2297. *
  2298. * Grab the errseq_t value from the mapping, and see if it has changed "since"
  2299. * the given value was sampled.
  2300. *
  2301. * If it has then report the latest error set, otherwise return 0.
  2302. */
  2303. static inline int filemap_check_wb_err(struct address_space *mapping,
  2304. errseq_t since)
  2305. {
  2306. return errseq_check(&mapping->wb_err, since);
  2307. }
  2308. /**
  2309. * filemap_sample_wb_err - sample the current errseq_t to test for later errors
  2310. * @mapping: mapping to be sampled
  2311. *
  2312. * Writeback errors are always reported relative to a particular sample point
  2313. * in the past. This function provides those sample points.
  2314. */
  2315. static inline errseq_t filemap_sample_wb_err(struct address_space *mapping)
  2316. {
  2317. return errseq_sample(&mapping->wb_err);
  2318. }
  2319. extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
  2320. int datasync);
  2321. extern int vfs_fsync(struct file *file, int datasync);
  2322. /*
  2323. * Sync the bytes written if this was a synchronous write. Expect ki_pos
  2324. * to already be updated for the write, and will return either the amount
  2325. * of bytes passed in, or an error if syncing the file failed.
  2326. */
  2327. static inline ssize_t generic_write_sync(struct kiocb *iocb, ssize_t count)
  2328. {
  2329. if (iocb->ki_flags & IOCB_DSYNC) {
  2330. int ret = vfs_fsync_range(iocb->ki_filp,
  2331. iocb->ki_pos - count, iocb->ki_pos - 1,
  2332. (iocb->ki_flags & IOCB_SYNC) ? 0 : 1);
  2333. if (ret)
  2334. return ret;
  2335. }
  2336. return count;
  2337. }
  2338. extern void emergency_sync(void);
  2339. extern void emergency_remount(void);
  2340. #ifdef CONFIG_BLOCK
  2341. extern sector_t bmap(struct inode *, sector_t);
  2342. #endif
  2343. extern int notify_change(struct dentry *, struct iattr *, struct inode **);
  2344. extern int inode_permission(struct inode *, int);
  2345. extern int __inode_permission(struct inode *, int);
  2346. extern int generic_permission(struct inode *, int);
  2347. extern int __check_sticky(struct inode *dir, struct inode *inode);
  2348. static inline bool execute_ok(struct inode *inode)
  2349. {
  2350. return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
  2351. }
  2352. static inline void file_start_write(struct file *file)
  2353. {
  2354. if (!S_ISREG(file_inode(file)->i_mode))
  2355. return;
  2356. __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
  2357. }
  2358. static inline bool file_start_write_trylock(struct file *file)
  2359. {
  2360. if (!S_ISREG(file_inode(file)->i_mode))
  2361. return true;
  2362. return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
  2363. }
  2364. static inline void file_end_write(struct file *file)
  2365. {
  2366. if (!S_ISREG(file_inode(file)->i_mode))
  2367. return;
  2368. __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
  2369. }
  2370. static inline int do_clone_file_range(struct file *file_in, loff_t pos_in,
  2371. struct file *file_out, loff_t pos_out,
  2372. u64 len)
  2373. {
  2374. int ret;
  2375. file_start_write(file_out);
  2376. ret = vfs_clone_file_range(file_in, pos_in, file_out, pos_out, len);
  2377. file_end_write(file_out);
  2378. return ret;
  2379. }
  2380. /*
  2381. * get_write_access() gets write permission for a file.
  2382. * put_write_access() releases this write permission.
  2383. * This is used for regular files.
  2384. * We cannot support write (and maybe mmap read-write shared) accesses and
  2385. * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
  2386. * can have the following values:
  2387. * 0: no writers, no VM_DENYWRITE mappings
  2388. * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
  2389. * > 0: (i_writecount) users are writing to the file.
  2390. *
  2391. * Normally we operate on that counter with atomic_{inc,dec} and it's safe
  2392. * except for the cases where we don't hold i_writecount yet. Then we need to
  2393. * use {get,deny}_write_access() - these functions check the sign and refuse
  2394. * to do the change if sign is wrong.
  2395. */
  2396. static inline int get_write_access(struct inode *inode)
  2397. {
  2398. return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
  2399. }
  2400. static inline int deny_write_access(struct file *file)
  2401. {
  2402. struct inode *inode = file_inode(file);
  2403. return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
  2404. }
  2405. static inline void put_write_access(struct inode * inode)
  2406. {
  2407. atomic_dec(&inode->i_writecount);
  2408. }
  2409. static inline void allow_write_access(struct file *file)
  2410. {
  2411. if (file)
  2412. atomic_inc(&file_inode(file)->i_writecount);
  2413. }
  2414. static inline bool inode_is_open_for_write(const struct inode *inode)
  2415. {
  2416. return atomic_read(&inode->i_writecount) > 0;
  2417. }
  2418. #ifdef CONFIG_IMA
  2419. static inline void i_readcount_dec(struct inode *inode)
  2420. {
  2421. BUG_ON(!atomic_read(&inode->i_readcount));
  2422. atomic_dec(&inode->i_readcount);
  2423. }
  2424. static inline void i_readcount_inc(struct inode *inode)
  2425. {
  2426. atomic_inc(&inode->i_readcount);
  2427. }
  2428. #else
  2429. static inline void i_readcount_dec(struct inode *inode)
  2430. {
  2431. return;
  2432. }
  2433. static inline void i_readcount_inc(struct inode *inode)
  2434. {
  2435. return;
  2436. }
  2437. #endif
  2438. extern int do_pipe_flags(int *, int);
  2439. #define __kernel_read_file_id(id) \
  2440. id(UNKNOWN, unknown) \
  2441. id(FIRMWARE, firmware) \
  2442. id(FIRMWARE_PREALLOC_BUFFER, firmware) \
  2443. id(MODULE, kernel-module) \
  2444. id(KEXEC_IMAGE, kexec-image) \
  2445. id(KEXEC_INITRAMFS, kexec-initramfs) \
  2446. id(POLICY, security-policy) \
  2447. id(MAX_ID, )
  2448. #define __fid_enumify(ENUM, dummy) READING_ ## ENUM,
  2449. #define __fid_stringify(dummy, str) #str,
  2450. enum kernel_read_file_id {
  2451. __kernel_read_file_id(__fid_enumify)
  2452. };
  2453. static const char * const kernel_read_file_str[] = {
  2454. __kernel_read_file_id(__fid_stringify)
  2455. };
  2456. static inline const char *kernel_read_file_id_str(enum kernel_read_file_id id)
  2457. {
  2458. if ((unsigned)id >= READING_MAX_ID)
  2459. return kernel_read_file_str[READING_UNKNOWN];
  2460. return kernel_read_file_str[id];
  2461. }
  2462. extern int kernel_read_file(struct file *, void **, loff_t *, loff_t,
  2463. enum kernel_read_file_id);
  2464. extern int kernel_read_file_from_path(const char *, void **, loff_t *, loff_t,
  2465. enum kernel_read_file_id);
  2466. extern int kernel_read_file_from_fd(int, void **, loff_t *, loff_t,
  2467. enum kernel_read_file_id);
  2468. extern ssize_t kernel_read(struct file *, void *, size_t, loff_t *);
  2469. extern ssize_t kernel_write(struct file *, const void *, size_t, loff_t *);
  2470. extern ssize_t __kernel_write(struct file *, const void *, size_t, loff_t *);
  2471. extern struct file * open_exec(const char *);
  2472. /* fs/dcache.c -- generic fs support functions */
  2473. extern bool is_subdir(struct dentry *, struct dentry *);
  2474. extern bool path_is_under(const struct path *, const struct path *);
  2475. extern char *file_path(struct file *, char *, int);
  2476. #include <linux/err.h>
  2477. /* needed for stackable file system support */
  2478. extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
  2479. extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
  2480. extern int inode_init_always(struct super_block *, struct inode *);
  2481. extern void inode_init_once(struct inode *);
  2482. extern void address_space_init_once(struct address_space *mapping);
  2483. extern struct inode * igrab(struct inode *);
  2484. extern ino_t iunique(struct super_block *, ino_t);
  2485. extern int inode_needs_sync(struct inode *inode);
  2486. extern int generic_delete_inode(struct inode *inode);
  2487. static inline int generic_drop_inode(struct inode *inode)
  2488. {
  2489. return !inode->i_nlink || inode_unhashed(inode);
  2490. }
  2491. extern struct inode *ilookup5_nowait(struct super_block *sb,
  2492. unsigned long hashval, int (*test)(struct inode *, void *),
  2493. void *data);
  2494. extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
  2495. int (*test)(struct inode *, void *), void *data);
  2496. extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
  2497. extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
  2498. extern struct inode * iget_locked(struct super_block *, unsigned long);
  2499. extern struct inode *find_inode_nowait(struct super_block *,
  2500. unsigned long,
  2501. int (*match)(struct inode *,
  2502. unsigned long, void *),
  2503. void *data);
  2504. extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
  2505. extern int insert_inode_locked(struct inode *);
  2506. #ifdef CONFIG_DEBUG_LOCK_ALLOC
  2507. extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
  2508. #else
  2509. static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
  2510. #endif
  2511. extern void unlock_new_inode(struct inode *);
  2512. extern unsigned int get_next_ino(void);
  2513. extern void evict_inodes(struct super_block *sb);
  2514. extern void __iget(struct inode * inode);
  2515. extern void iget_failed(struct inode *);
  2516. extern void clear_inode(struct inode *);
  2517. extern void __destroy_inode(struct inode *);
  2518. extern struct inode *new_inode_pseudo(struct super_block *sb);
  2519. extern struct inode *new_inode(struct super_block *sb);
  2520. extern void free_inode_nonrcu(struct inode *inode);
  2521. extern int should_remove_suid(struct dentry *);
  2522. extern int file_remove_privs(struct file *);
  2523. extern void __insert_inode_hash(struct inode *, unsigned long hashval);
  2524. static inline void insert_inode_hash(struct inode *inode)
  2525. {
  2526. __insert_inode_hash(inode, inode->i_ino);
  2527. }
  2528. extern void __remove_inode_hash(struct inode *);
  2529. static inline void remove_inode_hash(struct inode *inode)
  2530. {
  2531. if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash))
  2532. __remove_inode_hash(inode);
  2533. }
  2534. extern void inode_sb_list_add(struct inode *inode);
  2535. #ifdef CONFIG_BLOCK
  2536. extern int bdev_read_only(struct block_device *);
  2537. #endif
  2538. extern int set_blocksize(struct block_device *, int);
  2539. extern int sb_set_blocksize(struct super_block *, int);
  2540. extern int sb_min_blocksize(struct super_block *, int);
  2541. extern int generic_file_mmap(struct file *, struct vm_area_struct *);
  2542. extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
  2543. extern ssize_t generic_write_checks(struct kiocb *, struct iov_iter *);
  2544. extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
  2545. extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *);
  2546. extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *);
  2547. extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *);
  2548. extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t);
  2549. ssize_t vfs_iter_read(struct file *file, struct iov_iter *iter, loff_t *ppos,
  2550. rwf_t flags);
  2551. ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos,
  2552. rwf_t flags);
  2553. /* fs/block_dev.c */
  2554. extern ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to);
  2555. extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from);
  2556. extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
  2557. int datasync);
  2558. extern void block_sync_page(struct page *page);
  2559. /* fs/splice.c */
  2560. extern ssize_t generic_file_splice_read(struct file *, loff_t *,
  2561. struct pipe_inode_info *, size_t, unsigned int);
  2562. extern ssize_t iter_file_splice_write(struct pipe_inode_info *,
  2563. struct file *, loff_t *, size_t, unsigned int);
  2564. extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
  2565. struct file *out, loff_t *, size_t len, unsigned int flags);
  2566. extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
  2567. loff_t *opos, size_t len, unsigned int flags);
  2568. extern void
  2569. file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
  2570. extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
  2571. extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
  2572. extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
  2573. extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
  2574. extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
  2575. int whence, loff_t maxsize, loff_t eof);
  2576. extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
  2577. int whence, loff_t size);
  2578. extern loff_t no_seek_end_llseek_size(struct file *, loff_t, int, loff_t);
  2579. extern loff_t no_seek_end_llseek(struct file *, loff_t, int);
  2580. extern int generic_file_open(struct inode * inode, struct file * filp);
  2581. extern int nonseekable_open(struct inode * inode, struct file * filp);
  2582. #ifdef CONFIG_BLOCK
  2583. typedef void (dio_submit_t)(struct bio *bio, struct inode *inode,
  2584. loff_t file_offset);
  2585. enum {
  2586. /* need locking between buffered and direct access */
  2587. DIO_LOCKING = 0x01,
  2588. /* filesystem does not support filling holes */
  2589. DIO_SKIP_HOLES = 0x02,
  2590. /* filesystem can handle aio writes beyond i_size */
  2591. DIO_ASYNC_EXTEND = 0x04,
  2592. /* inode/fs/bdev does not need truncate protection */
  2593. DIO_SKIP_DIO_COUNT = 0x08,
  2594. };
  2595. void dio_end_io(struct bio *bio);
  2596. ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
  2597. struct block_device *bdev, struct iov_iter *iter,
  2598. get_block_t get_block,
  2599. dio_iodone_t end_io, dio_submit_t submit_io,
  2600. int flags);
  2601. static inline ssize_t blockdev_direct_IO(struct kiocb *iocb,
  2602. struct inode *inode,
  2603. struct iov_iter *iter,
  2604. get_block_t get_block)
  2605. {
  2606. return __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter,
  2607. get_block, NULL, NULL, DIO_LOCKING | DIO_SKIP_HOLES);
  2608. }
  2609. #endif
  2610. void inode_dio_wait(struct inode *inode);
  2611. /*
  2612. * inode_dio_begin - signal start of a direct I/O requests
  2613. * @inode: inode the direct I/O happens on
  2614. *
  2615. * This is called once we've finished processing a direct I/O request,
  2616. * and is used to wake up callers waiting for direct I/O to be quiesced.
  2617. */
  2618. static inline void inode_dio_begin(struct inode *inode)
  2619. {
  2620. atomic_inc(&inode->i_dio_count);
  2621. }
  2622. /*
  2623. * inode_dio_end - signal finish of a direct I/O requests
  2624. * @inode: inode the direct I/O happens on
  2625. *
  2626. * This is called once we've finished processing a direct I/O request,
  2627. * and is used to wake up callers waiting for direct I/O to be quiesced.
  2628. */
  2629. static inline void inode_dio_end(struct inode *inode)
  2630. {
  2631. if (atomic_dec_and_test(&inode->i_dio_count))
  2632. wake_up_bit(&inode->i_state, __I_DIO_WAKEUP);
  2633. }
  2634. extern void inode_set_flags(struct inode *inode, unsigned int flags,
  2635. unsigned int mask);
  2636. extern const struct file_operations generic_ro_fops;
  2637. #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
  2638. extern int readlink_copy(char __user *, int, const char *);
  2639. extern int page_readlink(struct dentry *, char __user *, int);
  2640. extern const char *page_get_link(struct dentry *, struct inode *,
  2641. struct delayed_call *);
  2642. extern void page_put_link(void *);
  2643. extern int __page_symlink(struct inode *inode, const char *symname, int len,
  2644. int nofs);
  2645. extern int page_symlink(struct inode *inode, const char *symname, int len);
  2646. extern const struct inode_operations page_symlink_inode_operations;
  2647. extern void kfree_link(void *);
  2648. extern void generic_fillattr(struct inode *, struct kstat *);
  2649. extern int vfs_getattr_nosec(const struct path *, struct kstat *, u32, unsigned int);
  2650. extern int vfs_getattr(const struct path *, struct kstat *, u32, unsigned int);
  2651. void __inode_add_bytes(struct inode *inode, loff_t bytes);
  2652. void inode_add_bytes(struct inode *inode, loff_t bytes);
  2653. void __inode_sub_bytes(struct inode *inode, loff_t bytes);
  2654. void inode_sub_bytes(struct inode *inode, loff_t bytes);
  2655. static inline loff_t __inode_get_bytes(struct inode *inode)
  2656. {
  2657. return (((loff_t)inode->i_blocks) << 9) + inode->i_bytes;
  2658. }
  2659. loff_t inode_get_bytes(struct inode *inode);
  2660. void inode_set_bytes(struct inode *inode, loff_t bytes);
  2661. const char *simple_get_link(struct dentry *, struct inode *,
  2662. struct delayed_call *);
  2663. extern const struct inode_operations simple_symlink_inode_operations;
  2664. extern int iterate_dir(struct file *, struct dir_context *);
  2665. extern int vfs_statx(int, const char __user *, int, struct kstat *, u32);
  2666. extern int vfs_statx_fd(unsigned int, struct kstat *, u32, unsigned int);
  2667. static inline int vfs_stat(const char __user *filename, struct kstat *stat)
  2668. {
  2669. return vfs_statx(AT_FDCWD, filename, AT_NO_AUTOMOUNT,
  2670. stat, STATX_BASIC_STATS);
  2671. }
  2672. static inline int vfs_lstat(const char __user *name, struct kstat *stat)
  2673. {
  2674. return vfs_statx(AT_FDCWD, name, AT_SYMLINK_NOFOLLOW | AT_NO_AUTOMOUNT,
  2675. stat, STATX_BASIC_STATS);
  2676. }
  2677. static inline int vfs_fstatat(int dfd, const char __user *filename,
  2678. struct kstat *stat, int flags)
  2679. {
  2680. return vfs_statx(dfd, filename, flags, stat, STATX_BASIC_STATS);
  2681. }
  2682. static inline int vfs_fstat(int fd, struct kstat *stat)
  2683. {
  2684. return vfs_statx_fd(fd, stat, STATX_BASIC_STATS, 0);
  2685. }
  2686. extern const char *vfs_get_link(struct dentry *, struct delayed_call *);
  2687. extern int vfs_readlink(struct dentry *, char __user *, int);
  2688. extern int __generic_block_fiemap(struct inode *inode,
  2689. struct fiemap_extent_info *fieinfo,
  2690. loff_t start, loff_t len,
  2691. get_block_t *get_block);
  2692. extern int generic_block_fiemap(struct inode *inode,
  2693. struct fiemap_extent_info *fieinfo, u64 start,
  2694. u64 len, get_block_t *get_block);
  2695. extern struct file_system_type *get_filesystem(struct file_system_type *fs);
  2696. extern void put_filesystem(struct file_system_type *fs);
  2697. extern struct file_system_type *get_fs_type(const char *name);
  2698. extern struct super_block *get_super(struct block_device *);
  2699. extern struct super_block *get_super_thawed(struct block_device *);
  2700. extern struct super_block *get_super_exclusive_thawed(struct block_device *bdev);
  2701. extern struct super_block *get_active_super(struct block_device *bdev);
  2702. extern void drop_super(struct super_block *sb);
  2703. extern void drop_super_exclusive(struct super_block *sb);
  2704. extern void iterate_supers(void (*)(struct super_block *, void *), void *);
  2705. extern void iterate_supers_type(struct file_system_type *,
  2706. void (*)(struct super_block *, void *), void *);
  2707. extern int dcache_dir_open(struct inode *, struct file *);
  2708. extern int dcache_dir_close(struct inode *, struct file *);
  2709. extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
  2710. extern int dcache_readdir(struct file *, struct dir_context *);
  2711. extern int simple_setattr(struct dentry *, struct iattr *);
  2712. extern int simple_getattr(const struct path *, struct kstat *, u32, unsigned int);
  2713. extern int simple_statfs(struct dentry *, struct kstatfs *);
  2714. extern int simple_open(struct inode *inode, struct file *file);
  2715. extern int simple_link(struct dentry *, struct inode *, struct dentry *);
  2716. extern int simple_unlink(struct inode *, struct dentry *);
  2717. extern int simple_rmdir(struct inode *, struct dentry *);
  2718. extern int simple_rename(struct inode *, struct dentry *,
  2719. struct inode *, struct dentry *, unsigned int);
  2720. extern int noop_fsync(struct file *, loff_t, loff_t, int);
  2721. extern int simple_empty(struct dentry *);
  2722. extern int simple_readpage(struct file *file, struct page *page);
  2723. extern int simple_write_begin(struct file *file, struct address_space *mapping,
  2724. loff_t pos, unsigned len, unsigned flags,
  2725. struct page **pagep, void **fsdata);
  2726. extern int simple_write_end(struct file *file, struct address_space *mapping,
  2727. loff_t pos, unsigned len, unsigned copied,
  2728. struct page *page, void *fsdata);
  2729. extern int always_delete_dentry(const struct dentry *);
  2730. extern struct inode *alloc_anon_inode(struct super_block *);
  2731. extern int simple_nosetlease(struct file *, long, struct file_lock **, void **);
  2732. extern const struct dentry_operations simple_dentry_operations;
  2733. extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
  2734. extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
  2735. extern const struct file_operations simple_dir_operations;
  2736. extern const struct inode_operations simple_dir_inode_operations;
  2737. extern void make_empty_dir_inode(struct inode *inode);
  2738. extern bool is_empty_dir_inode(struct inode *inode);
  2739. struct tree_descr { const char *name; const struct file_operations *ops; int mode; };
  2740. struct dentry *d_alloc_name(struct dentry *, const char *);
  2741. extern int simple_fill_super(struct super_block *, unsigned long,
  2742. const struct tree_descr *);
  2743. extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
  2744. extern void simple_release_fs(struct vfsmount **mount, int *count);
  2745. extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
  2746. loff_t *ppos, const void *from, size_t available);
  2747. extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
  2748. const void __user *from, size_t count);
  2749. extern int __generic_file_fsync(struct file *, loff_t, loff_t, int);
  2750. extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
  2751. extern int generic_check_addressable(unsigned, u64);
  2752. #ifdef CONFIG_MIGRATION
  2753. extern int buffer_migrate_page(struct address_space *,
  2754. struct page *, struct page *,
  2755. enum migrate_mode);
  2756. #else
  2757. #define buffer_migrate_page NULL
  2758. #endif
  2759. extern int setattr_prepare(struct dentry *, struct iattr *);
  2760. extern int inode_newsize_ok(const struct inode *, loff_t offset);
  2761. extern void setattr_copy(struct inode *inode, const struct iattr *attr);
  2762. extern int file_update_time(struct file *file);
  2763. static inline bool io_is_direct(struct file *filp)
  2764. {
  2765. return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host);
  2766. }
  2767. static inline bool vma_is_dax(struct vm_area_struct *vma)
  2768. {
  2769. return vma->vm_file && IS_DAX(vma->vm_file->f_mapping->host);
  2770. }
  2771. static inline int iocb_flags(struct file *file)
  2772. {
  2773. int res = 0;
  2774. if (file->f_flags & O_APPEND)
  2775. res |= IOCB_APPEND;
  2776. if (io_is_direct(file))
  2777. res |= IOCB_DIRECT;
  2778. if ((file->f_flags & O_DSYNC) || IS_SYNC(file->f_mapping->host))
  2779. res |= IOCB_DSYNC;
  2780. if (file->f_flags & __O_SYNC)
  2781. res |= IOCB_SYNC;
  2782. return res;
  2783. }
  2784. static inline int kiocb_set_rw_flags(struct kiocb *ki, rwf_t flags)
  2785. {
  2786. if (unlikely(flags & ~RWF_SUPPORTED))
  2787. return -EOPNOTSUPP;
  2788. if (flags & RWF_NOWAIT) {
  2789. if (!(ki->ki_filp->f_mode & FMODE_NOWAIT))
  2790. return -EOPNOTSUPP;
  2791. ki->ki_flags |= IOCB_NOWAIT;
  2792. }
  2793. if (flags & RWF_HIPRI)
  2794. ki->ki_flags |= IOCB_HIPRI;
  2795. if (flags & RWF_DSYNC)
  2796. ki->ki_flags |= IOCB_DSYNC;
  2797. if (flags & RWF_SYNC)
  2798. ki->ki_flags |= (IOCB_DSYNC | IOCB_SYNC);
  2799. return 0;
  2800. }
  2801. static inline ino_t parent_ino(struct dentry *dentry)
  2802. {
  2803. ino_t res;
  2804. /*
  2805. * Don't strictly need d_lock here? If the parent ino could change
  2806. * then surely we'd have a deeper race in the caller?
  2807. */
  2808. spin_lock(&dentry->d_lock);
  2809. res = dentry->d_parent->d_inode->i_ino;
  2810. spin_unlock(&dentry->d_lock);
  2811. return res;
  2812. }
  2813. /* Transaction based IO helpers */
  2814. /*
  2815. * An argresp is stored in an allocated page and holds the
  2816. * size of the argument or response, along with its content
  2817. */
  2818. struct simple_transaction_argresp {
  2819. ssize_t size;
  2820. char data[0];
  2821. };
  2822. #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
  2823. char *simple_transaction_get(struct file *file, const char __user *buf,
  2824. size_t size);
  2825. ssize_t simple_transaction_read(struct file *file, char __user *buf,
  2826. size_t size, loff_t *pos);
  2827. int simple_transaction_release(struct inode *inode, struct file *file);
  2828. void simple_transaction_set(struct file *file, size_t n);
  2829. /*
  2830. * simple attribute files
  2831. *
  2832. * These attributes behave similar to those in sysfs:
  2833. *
  2834. * Writing to an attribute immediately sets a value, an open file can be
  2835. * written to multiple times.
  2836. *
  2837. * Reading from an attribute creates a buffer from the value that might get
  2838. * read with multiple read calls. When the attribute has been read
  2839. * completely, no further read calls are possible until the file is opened
  2840. * again.
  2841. *
  2842. * All attributes contain a text representation of a numeric value
  2843. * that are accessed with the get() and set() functions.
  2844. */
  2845. #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
  2846. static int __fops ## _open(struct inode *inode, struct file *file) \
  2847. { \
  2848. __simple_attr_check_format(__fmt, 0ull); \
  2849. return simple_attr_open(inode, file, __get, __set, __fmt); \
  2850. } \
  2851. static const struct file_operations __fops = { \
  2852. .owner = THIS_MODULE, \
  2853. .open = __fops ## _open, \
  2854. .release = simple_attr_release, \
  2855. .read = simple_attr_read, \
  2856. .write = simple_attr_write, \
  2857. .llseek = generic_file_llseek, \
  2858. }
  2859. static inline __printf(1, 2)
  2860. void __simple_attr_check_format(const char *fmt, ...)
  2861. {
  2862. /* don't do anything, just let the compiler check the arguments; */
  2863. }
  2864. int simple_attr_open(struct inode *inode, struct file *file,
  2865. int (*get)(void *, u64 *), int (*set)(void *, u64),
  2866. const char *fmt);
  2867. int simple_attr_release(struct inode *inode, struct file *file);
  2868. ssize_t simple_attr_read(struct file *file, char __user *buf,
  2869. size_t len, loff_t *ppos);
  2870. ssize_t simple_attr_write(struct file *file, const char __user *buf,
  2871. size_t len, loff_t *ppos);
  2872. struct ctl_table;
  2873. int proc_nr_files(struct ctl_table *table, int write,
  2874. void __user *buffer, size_t *lenp, loff_t *ppos);
  2875. int proc_nr_dentry(struct ctl_table *table, int write,
  2876. void __user *buffer, size_t *lenp, loff_t *ppos);
  2877. int proc_nr_inodes(struct ctl_table *table, int write,
  2878. void __user *buffer, size_t *lenp, loff_t *ppos);
  2879. int __init get_filesystem_list(char *buf);
  2880. #define __FMODE_EXEC ((__force int) FMODE_EXEC)
  2881. #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
  2882. #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
  2883. #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
  2884. (flag & __FMODE_NONOTIFY)))
  2885. static inline bool is_sxid(umode_t mode)
  2886. {
  2887. return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
  2888. }
  2889. static inline int check_sticky(struct inode *dir, struct inode *inode)
  2890. {
  2891. if (!(dir->i_mode & S_ISVTX))
  2892. return 0;
  2893. return __check_sticky(dir, inode);
  2894. }
  2895. static inline void inode_has_no_xattr(struct inode *inode)
  2896. {
  2897. if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & SB_NOSEC))
  2898. inode->i_flags |= S_NOSEC;
  2899. }
  2900. static inline bool is_root_inode(struct inode *inode)
  2901. {
  2902. return inode == inode->i_sb->s_root->d_inode;
  2903. }
  2904. static inline bool dir_emit(struct dir_context *ctx,
  2905. const char *name, int namelen,
  2906. u64 ino, unsigned type)
  2907. {
  2908. return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
  2909. }
  2910. static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
  2911. {
  2912. return ctx->actor(ctx, ".", 1, ctx->pos,
  2913. file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
  2914. }
  2915. static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
  2916. {
  2917. return ctx->actor(ctx, "..", 2, ctx->pos,
  2918. parent_ino(file->f_path.dentry), DT_DIR) == 0;
  2919. }
  2920. static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
  2921. {
  2922. if (ctx->pos == 0) {
  2923. if (!dir_emit_dot(file, ctx))
  2924. return false;
  2925. ctx->pos = 1;
  2926. }
  2927. if (ctx->pos == 1) {
  2928. if (!dir_emit_dotdot(file, ctx))
  2929. return false;
  2930. ctx->pos = 2;
  2931. }
  2932. return true;
  2933. }
  2934. static inline bool dir_relax(struct inode *inode)
  2935. {
  2936. inode_unlock(inode);
  2937. inode_lock(inode);
  2938. return !IS_DEADDIR(inode);
  2939. }
  2940. static inline bool dir_relax_shared(struct inode *inode)
  2941. {
  2942. inode_unlock_shared(inode);
  2943. inode_lock_shared(inode);
  2944. return !IS_DEADDIR(inode);
  2945. }
  2946. extern bool path_noexec(const struct path *path);
  2947. extern void inode_nohighmem(struct inode *inode);
  2948. #endif /* _LINUX_FS_H */