fs.h 87 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568256925702571257225732574257525762577257825792580258125822583258425852586258725882589259025912592259325942595259625972598259926002601260226032604
  1. #ifndef _LINUX_FS_H
  2. #define _LINUX_FS_H
  3. /*
  4. * This file has definitions for some important file table
  5. * structures etc.
  6. */
  7. #include <linux/limits.h>
  8. #include <linux/ioctl.h>
  9. #include <linux/blk_types.h>
  10. #include <linux/types.h>
  11. /*
  12. * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
  13. * the file limit at runtime and only root can increase the per-process
  14. * nr_file rlimit, so it's safe to set up a ridiculously high absolute
  15. * upper limit on files-per-process.
  16. *
  17. * Some programs (notably those using select()) may have to be
  18. * recompiled to take full advantage of the new limits..
  19. */
  20. /* Fixed constants first: */
  21. #undef NR_OPEN
  22. #define INR_OPEN_CUR 1024 /* Initial setting for nfile rlimits */
  23. #define INR_OPEN_MAX 4096 /* Hard limit for nfile rlimits */
  24. #define BLOCK_SIZE_BITS 10
  25. #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
  26. #define SEEK_SET 0 /* seek relative to beginning of file */
  27. #define SEEK_CUR 1 /* seek relative to current file position */
  28. #define SEEK_END 2 /* seek relative to end of file */
  29. #define SEEK_MAX SEEK_END
  30. struct fstrim_range {
  31. __u64 start;
  32. __u64 len;
  33. __u64 minlen;
  34. };
  35. /* And dynamically-tunable limits and defaults: */
  36. struct files_stat_struct {
  37. unsigned long nr_files; /* read only */
  38. unsigned long nr_free_files; /* read only */
  39. unsigned long max_files; /* tunable */
  40. };
  41. struct inodes_stat_t {
  42. int nr_inodes;
  43. int nr_unused;
  44. int dummy[5]; /* padding for sysctl ABI compatibility */
  45. };
  46. #define NR_FILE 8192 /* this can well be larger on a larger system */
  47. #define MAY_EXEC 1
  48. #define MAY_WRITE 2
  49. #define MAY_READ 4
  50. #define MAY_APPEND 8
  51. #define MAY_ACCESS 16
  52. #define MAY_OPEN 32
  53. #define MAY_CHDIR 64
  54. /*
  55. * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
  56. * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
  57. */
  58. /* file is open for reading */
  59. #define FMODE_READ ((__force fmode_t)0x1)
  60. /* file is open for writing */
  61. #define FMODE_WRITE ((__force fmode_t)0x2)
  62. /* file is seekable */
  63. #define FMODE_LSEEK ((__force fmode_t)0x4)
  64. /* file can be accessed using pread */
  65. #define FMODE_PREAD ((__force fmode_t)0x8)
  66. /* file can be accessed using pwrite */
  67. #define FMODE_PWRITE ((__force fmode_t)0x10)
  68. /* File is opened for execution with sys_execve / sys_uselib */
  69. #define FMODE_EXEC ((__force fmode_t)0x20)
  70. /* File is opened with O_NDELAY (only set for block devices) */
  71. #define FMODE_NDELAY ((__force fmode_t)0x40)
  72. /* File is opened with O_EXCL (only set for block devices) */
  73. #define FMODE_EXCL ((__force fmode_t)0x80)
  74. /* File is opened using open(.., 3, ..) and is writeable only for ioctls
  75. (specialy hack for floppy.c) */
  76. #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
  77. /*
  78. * Don't update ctime and mtime.
  79. *
  80. * Currently a special hack for the XFS open_by_handle ioctl, but we'll
  81. * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
  82. */
  83. #define FMODE_NOCMTIME ((__force fmode_t)0x800)
  84. /* Expect random access pattern */
  85. #define FMODE_RANDOM ((__force fmode_t)0x1000)
  86. /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
  87. #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
  88. /* File is opened with O_PATH; almost nothing can be done with it */
  89. #define FMODE_PATH ((__force fmode_t)0x4000)
  90. /* File was opened by fanotify and shouldn't generate fanotify events */
  91. #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
  92. /*
  93. * The below are the various read and write types that we support. Some of
  94. * them include behavioral modifiers that send information down to the
  95. * block layer and IO scheduler. Terminology:
  96. *
  97. * The block layer uses device plugging to defer IO a little bit, in
  98. * the hope that we will see more IO very shortly. This increases
  99. * coalescing of adjacent IO and thus reduces the number of IOs we
  100. * have to send to the device. It also allows for better queuing,
  101. * if the IO isn't mergeable. If the caller is going to be waiting
  102. * for the IO, then he must ensure that the device is unplugged so
  103. * that the IO is dispatched to the driver.
  104. *
  105. * All IO is handled async in Linux. This is fine for background
  106. * writes, but for reads or writes that someone waits for completion
  107. * on, we want to notify the block layer and IO scheduler so that they
  108. * know about it. That allows them to make better scheduling
  109. * decisions. So when the below references 'sync' and 'async', it
  110. * is referencing this priority hint.
  111. *
  112. * With that in mind, the available types are:
  113. *
  114. * READ A normal read operation. Device will be plugged.
  115. * READ_SYNC A synchronous read. Device is not plugged, caller can
  116. * immediately wait on this read without caring about
  117. * unplugging.
  118. * READA Used for read-ahead operations. Lower priority, and the
  119. * block layer could (in theory) choose to ignore this
  120. * request if it runs into resource problems.
  121. * WRITE A normal async write. Device will be plugged.
  122. * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
  123. * the hint that someone will be waiting on this IO
  124. * shortly. The write equivalent of READ_SYNC.
  125. * WRITE_ODIRECT Special case write for O_DIRECT only.
  126. * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
  127. * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
  128. * non-volatile media on completion.
  129. * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
  130. * by a cache flush and data is guaranteed to be on
  131. * non-volatile media on completion.
  132. *
  133. */
  134. #define RW_MASK REQ_WRITE
  135. #define RWA_MASK REQ_RAHEAD
  136. #define READ 0
  137. #define WRITE RW_MASK
  138. #define READA RWA_MASK
  139. #define READ_SYNC (READ | REQ_SYNC)
  140. #define READ_META (READ | REQ_META)
  141. #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
  142. #define WRITE_ODIRECT (WRITE | REQ_SYNC)
  143. #define WRITE_META (WRITE | REQ_META)
  144. #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
  145. #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
  146. #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
  147. #define SEL_IN 1
  148. #define SEL_OUT 2
  149. #define SEL_EX 4
  150. /* public flags for file_system_type */
  151. #define FS_REQUIRES_DEV 1
  152. #define FS_BINARY_MOUNTDATA 2
  153. #define FS_HAS_SUBTYPE 4
  154. #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
  155. #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
  156. * during rename() internally.
  157. */
  158. /*
  159. * These are the fs-independent mount-flags: up to 32 flags are supported
  160. */
  161. #define MS_RDONLY 1 /* Mount read-only */
  162. #define MS_NOSUID 2 /* Ignore suid and sgid bits */
  163. #define MS_NODEV 4 /* Disallow access to device special files */
  164. #define MS_NOEXEC 8 /* Disallow program execution */
  165. #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
  166. #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
  167. #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
  168. #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
  169. #define MS_NOATIME 1024 /* Do not update access times. */
  170. #define MS_NODIRATIME 2048 /* Do not update directory access times */
  171. #define MS_BIND 4096
  172. #define MS_MOVE 8192
  173. #define MS_REC 16384
  174. #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
  175. MS_VERBOSE is deprecated. */
  176. #define MS_SILENT 32768
  177. #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
  178. #define MS_UNBINDABLE (1<<17) /* change to unbindable */
  179. #define MS_PRIVATE (1<<18) /* change to private */
  180. #define MS_SLAVE (1<<19) /* change to slave */
  181. #define MS_SHARED (1<<20) /* change to shared */
  182. #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
  183. #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
  184. #define MS_I_VERSION (1<<23) /* Update inode I_version field */
  185. #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
  186. #define MS_NOSEC (1<<28)
  187. #define MS_BORN (1<<29)
  188. #define MS_ACTIVE (1<<30)
  189. #define MS_NOUSER (1<<31)
  190. /*
  191. * Superblock flags that can be altered by MS_REMOUNT
  192. */
  193. #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
  194. /*
  195. * Old magic mount flag and mask
  196. */
  197. #define MS_MGC_VAL 0xC0ED0000
  198. #define MS_MGC_MSK 0xffff0000
  199. /* Inode flags - they have nothing to superblock flags now */
  200. #define S_SYNC 1 /* Writes are synced at once */
  201. #define S_NOATIME 2 /* Do not update access times */
  202. #define S_APPEND 4 /* Append-only file */
  203. #define S_IMMUTABLE 8 /* Immutable file */
  204. #define S_DEAD 16 /* removed, but still open directory */
  205. #define S_NOQUOTA 32 /* Inode is not counted to quota */
  206. #define S_DIRSYNC 64 /* Directory modifications are synchronous */
  207. #define S_NOCMTIME 128 /* Do not update file c/mtime */
  208. #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
  209. #define S_PRIVATE 512 /* Inode is fs-internal */
  210. #define S_IMA 1024 /* Inode has an associated IMA struct */
  211. #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
  212. #define S_NOSEC 4096 /* no suid or xattr security attributes */
  213. /*
  214. * Note that nosuid etc flags are inode-specific: setting some file-system
  215. * flags just means all the inodes inherit those flags by default. It might be
  216. * possible to override it selectively if you really wanted to with some
  217. * ioctl() that is not currently implemented.
  218. *
  219. * Exception: MS_RDONLY is always applied to the entire file system.
  220. *
  221. * Unfortunately, it is possible to change a filesystems flags with it mounted
  222. * with files in use. This means that all of the inodes will not have their
  223. * i_flags updated. Hence, i_flags no longer inherit the superblock mount
  224. * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
  225. */
  226. #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
  227. #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
  228. #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
  229. ((inode)->i_flags & S_SYNC))
  230. #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
  231. ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
  232. #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
  233. #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
  234. #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
  235. #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
  236. #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
  237. #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
  238. #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
  239. #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
  240. #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
  241. #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
  242. #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
  243. #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
  244. #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
  245. #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
  246. /* the read-only stuff doesn't really belong here, but any other place is
  247. probably as bad and I don't want to create yet another include file. */
  248. #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
  249. #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
  250. #define BLKRRPART _IO(0x12,95) /* re-read partition table */
  251. #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
  252. #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
  253. #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
  254. #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
  255. #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
  256. #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
  257. #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
  258. #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
  259. #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
  260. #if 0
  261. #define BLKPG _IO(0x12,105)/* See blkpg.h */
  262. /* Some people are morons. Do not use sizeof! */
  263. #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
  264. #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
  265. /* This was here just to show that the number is taken -
  266. probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
  267. #endif
  268. /* A jump here: 108-111 have been used for various private purposes. */
  269. #define BLKBSZGET _IOR(0x12,112,size_t)
  270. #define BLKBSZSET _IOW(0x12,113,size_t)
  271. #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
  272. #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
  273. #define BLKTRACESTART _IO(0x12,116)
  274. #define BLKTRACESTOP _IO(0x12,117)
  275. #define BLKTRACETEARDOWN _IO(0x12,118)
  276. #define BLKDISCARD _IO(0x12,119)
  277. #define BLKIOMIN _IO(0x12,120)
  278. #define BLKIOOPT _IO(0x12,121)
  279. #define BLKALIGNOFF _IO(0x12,122)
  280. #define BLKPBSZGET _IO(0x12,123)
  281. #define BLKDISCARDZEROES _IO(0x12,124)
  282. #define BLKSECDISCARD _IO(0x12,125)
  283. #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
  284. #define FIBMAP _IO(0x00,1) /* bmap access */
  285. #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
  286. #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
  287. #define FITHAW _IOWR('X', 120, int) /* Thaw */
  288. #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
  289. #define FS_IOC_GETFLAGS _IOR('f', 1, long)
  290. #define FS_IOC_SETFLAGS _IOW('f', 2, long)
  291. #define FS_IOC_GETVERSION _IOR('v', 1, long)
  292. #define FS_IOC_SETVERSION _IOW('v', 2, long)
  293. #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
  294. #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
  295. #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
  296. #define FS_IOC32_GETVERSION _IOR('v', 1, int)
  297. #define FS_IOC32_SETVERSION _IOW('v', 2, int)
  298. /*
  299. * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
  300. */
  301. #define FS_SECRM_FL 0x00000001 /* Secure deletion */
  302. #define FS_UNRM_FL 0x00000002 /* Undelete */
  303. #define FS_COMPR_FL 0x00000004 /* Compress file */
  304. #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
  305. #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
  306. #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
  307. #define FS_NODUMP_FL 0x00000040 /* do not dump file */
  308. #define FS_NOATIME_FL 0x00000080 /* do not update atime */
  309. /* Reserved for compression usage... */
  310. #define FS_DIRTY_FL 0x00000100
  311. #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
  312. #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
  313. #define FS_ECOMPR_FL 0x00000800 /* Compression error */
  314. /* End compression flags --- maybe not all used */
  315. #define FS_BTREE_FL 0x00001000 /* btree format dir */
  316. #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
  317. #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
  318. #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
  319. #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
  320. #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
  321. #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
  322. #define FS_EXTENT_FL 0x00080000 /* Extents */
  323. #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
  324. #define FS_NOCOW_FL 0x00800000 /* Do not cow file */
  325. #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
  326. #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
  327. #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
  328. #define SYNC_FILE_RANGE_WAIT_BEFORE 1
  329. #define SYNC_FILE_RANGE_WRITE 2
  330. #define SYNC_FILE_RANGE_WAIT_AFTER 4
  331. #ifdef __KERNEL__
  332. #include <linux/linkage.h>
  333. #include <linux/wait.h>
  334. #include <linux/types.h>
  335. #include <linux/kdev_t.h>
  336. #include <linux/dcache.h>
  337. #include <linux/path.h>
  338. #include <linux/stat.h>
  339. #include <linux/cache.h>
  340. #include <linux/list.h>
  341. #include <linux/radix-tree.h>
  342. #include <linux/prio_tree.h>
  343. #include <linux/init.h>
  344. #include <linux/pid.h>
  345. #include <linux/mutex.h>
  346. #include <linux/capability.h>
  347. #include <linux/semaphore.h>
  348. #include <linux/fiemap.h>
  349. #include <linux/rculist_bl.h>
  350. #include <asm/atomic.h>
  351. #include <asm/byteorder.h>
  352. struct export_operations;
  353. struct hd_geometry;
  354. struct iovec;
  355. struct nameidata;
  356. struct kiocb;
  357. struct kobject;
  358. struct pipe_inode_info;
  359. struct poll_table_struct;
  360. struct kstatfs;
  361. struct vm_area_struct;
  362. struct vfsmount;
  363. struct cred;
  364. extern void __init inode_init(void);
  365. extern void __init inode_init_early(void);
  366. extern void __init files_init(unsigned long);
  367. extern struct files_stat_struct files_stat;
  368. extern unsigned long get_max_files(void);
  369. extern int sysctl_nr_open;
  370. extern struct inodes_stat_t inodes_stat;
  371. extern int leases_enable, lease_break_time;
  372. struct buffer_head;
  373. typedef int (get_block_t)(struct inode *inode, sector_t iblock,
  374. struct buffer_head *bh_result, int create);
  375. typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
  376. ssize_t bytes, void *private, int ret,
  377. bool is_async);
  378. /*
  379. * Attribute flags. These should be or-ed together to figure out what
  380. * has been changed!
  381. */
  382. #define ATTR_MODE (1 << 0)
  383. #define ATTR_UID (1 << 1)
  384. #define ATTR_GID (1 << 2)
  385. #define ATTR_SIZE (1 << 3)
  386. #define ATTR_ATIME (1 << 4)
  387. #define ATTR_MTIME (1 << 5)
  388. #define ATTR_CTIME (1 << 6)
  389. #define ATTR_ATIME_SET (1 << 7)
  390. #define ATTR_MTIME_SET (1 << 8)
  391. #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
  392. #define ATTR_ATTR_FLAG (1 << 10)
  393. #define ATTR_KILL_SUID (1 << 11)
  394. #define ATTR_KILL_SGID (1 << 12)
  395. #define ATTR_FILE (1 << 13)
  396. #define ATTR_KILL_PRIV (1 << 14)
  397. #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
  398. #define ATTR_TIMES_SET (1 << 16)
  399. /*
  400. * This is the Inode Attributes structure, used for notify_change(). It
  401. * uses the above definitions as flags, to know which values have changed.
  402. * Also, in this manner, a Filesystem can look at only the values it cares
  403. * about. Basically, these are the attributes that the VFS layer can
  404. * request to change from the FS layer.
  405. *
  406. * Derek Atkins <warlord@MIT.EDU> 94-10-20
  407. */
  408. struct iattr {
  409. unsigned int ia_valid;
  410. umode_t ia_mode;
  411. uid_t ia_uid;
  412. gid_t ia_gid;
  413. loff_t ia_size;
  414. struct timespec ia_atime;
  415. struct timespec ia_mtime;
  416. struct timespec ia_ctime;
  417. /*
  418. * Not an attribute, but an auxiliary info for filesystems wanting to
  419. * implement an ftruncate() like method. NOTE: filesystem should
  420. * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
  421. */
  422. struct file *ia_file;
  423. };
  424. /*
  425. * Includes for diskquotas.
  426. */
  427. #include <linux/quota.h>
  428. /**
  429. * enum positive_aop_returns - aop return codes with specific semantics
  430. *
  431. * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
  432. * completed, that the page is still locked, and
  433. * should be considered active. The VM uses this hint
  434. * to return the page to the active list -- it won't
  435. * be a candidate for writeback again in the near
  436. * future. Other callers must be careful to unlock
  437. * the page if they get this return. Returned by
  438. * writepage();
  439. *
  440. * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
  441. * unlocked it and the page might have been truncated.
  442. * The caller should back up to acquiring a new page and
  443. * trying again. The aop will be taking reasonable
  444. * precautions not to livelock. If the caller held a page
  445. * reference, it should drop it before retrying. Returned
  446. * by readpage().
  447. *
  448. * address_space_operation functions return these large constants to indicate
  449. * special semantics to the caller. These are much larger than the bytes in a
  450. * page to allow for functions that return the number of bytes operated on in a
  451. * given page.
  452. */
  453. enum positive_aop_returns {
  454. AOP_WRITEPAGE_ACTIVATE = 0x80000,
  455. AOP_TRUNCATED_PAGE = 0x80001,
  456. };
  457. #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
  458. #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
  459. #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
  460. * helper code (eg buffer layer)
  461. * to clear GFP_FS from alloc */
  462. /*
  463. * oh the beauties of C type declarations.
  464. */
  465. struct page;
  466. struct address_space;
  467. struct writeback_control;
  468. struct iov_iter {
  469. const struct iovec *iov;
  470. unsigned long nr_segs;
  471. size_t iov_offset;
  472. size_t count;
  473. };
  474. size_t iov_iter_copy_from_user_atomic(struct page *page,
  475. struct iov_iter *i, unsigned long offset, size_t bytes);
  476. size_t iov_iter_copy_from_user(struct page *page,
  477. struct iov_iter *i, unsigned long offset, size_t bytes);
  478. void iov_iter_advance(struct iov_iter *i, size_t bytes);
  479. int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
  480. size_t iov_iter_single_seg_count(struct iov_iter *i);
  481. static inline void iov_iter_init(struct iov_iter *i,
  482. const struct iovec *iov, unsigned long nr_segs,
  483. size_t count, size_t written)
  484. {
  485. i->iov = iov;
  486. i->nr_segs = nr_segs;
  487. i->iov_offset = 0;
  488. i->count = count + written;
  489. iov_iter_advance(i, written);
  490. }
  491. static inline size_t iov_iter_count(struct iov_iter *i)
  492. {
  493. return i->count;
  494. }
  495. /*
  496. * "descriptor" for what we're up to with a read.
  497. * This allows us to use the same read code yet
  498. * have multiple different users of the data that
  499. * we read from a file.
  500. *
  501. * The simplest case just copies the data to user
  502. * mode.
  503. */
  504. typedef struct {
  505. size_t written;
  506. size_t count;
  507. union {
  508. char __user *buf;
  509. void *data;
  510. } arg;
  511. int error;
  512. } read_descriptor_t;
  513. typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
  514. unsigned long, unsigned long);
  515. struct address_space_operations {
  516. int (*writepage)(struct page *page, struct writeback_control *wbc);
  517. int (*readpage)(struct file *, struct page *);
  518. /* Write back some dirty pages from this mapping. */
  519. int (*writepages)(struct address_space *, struct writeback_control *);
  520. /* Set a page dirty. Return true if this dirtied it */
  521. int (*set_page_dirty)(struct page *page);
  522. int (*readpages)(struct file *filp, struct address_space *mapping,
  523. struct list_head *pages, unsigned nr_pages);
  524. int (*write_begin)(struct file *, struct address_space *mapping,
  525. loff_t pos, unsigned len, unsigned flags,
  526. struct page **pagep, void **fsdata);
  527. int (*write_end)(struct file *, struct address_space *mapping,
  528. loff_t pos, unsigned len, unsigned copied,
  529. struct page *page, void *fsdata);
  530. /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
  531. sector_t (*bmap)(struct address_space *, sector_t);
  532. void (*invalidatepage) (struct page *, unsigned long);
  533. int (*releasepage) (struct page *, gfp_t);
  534. void (*freepage)(struct page *);
  535. ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
  536. loff_t offset, unsigned long nr_segs);
  537. int (*get_xip_mem)(struct address_space *, pgoff_t, int,
  538. void **, unsigned long *);
  539. /* migrate the contents of a page to the specified target */
  540. int (*migratepage) (struct address_space *,
  541. struct page *, struct page *);
  542. int (*launder_page) (struct page *);
  543. int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
  544. unsigned long);
  545. int (*error_remove_page)(struct address_space *, struct page *);
  546. };
  547. extern const struct address_space_operations empty_aops;
  548. /*
  549. * pagecache_write_begin/pagecache_write_end must be used by general code
  550. * to write into the pagecache.
  551. */
  552. int pagecache_write_begin(struct file *, struct address_space *mapping,
  553. loff_t pos, unsigned len, unsigned flags,
  554. struct page **pagep, void **fsdata);
  555. int pagecache_write_end(struct file *, struct address_space *mapping,
  556. loff_t pos, unsigned len, unsigned copied,
  557. struct page *page, void *fsdata);
  558. struct backing_dev_info;
  559. struct address_space {
  560. struct inode *host; /* owner: inode, block_device */
  561. struct radix_tree_root page_tree; /* radix tree of all pages */
  562. spinlock_t tree_lock; /* and lock protecting it */
  563. unsigned int i_mmap_writable;/* count VM_SHARED mappings */
  564. struct prio_tree_root i_mmap; /* tree of private and shared mappings */
  565. struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
  566. struct mutex i_mmap_mutex; /* protect tree, count, list */
  567. /* Protected by tree_lock together with the radix tree */
  568. unsigned long nrpages; /* number of total pages */
  569. pgoff_t writeback_index;/* writeback starts here */
  570. const struct address_space_operations *a_ops; /* methods */
  571. unsigned long flags; /* error bits/gfp mask */
  572. struct backing_dev_info *backing_dev_info; /* device readahead, etc */
  573. spinlock_t private_lock; /* for use by the address_space */
  574. struct list_head private_list; /* ditto */
  575. struct address_space *assoc_mapping; /* ditto */
  576. } __attribute__((aligned(sizeof(long))));
  577. /*
  578. * On most architectures that alignment is already the case; but
  579. * must be enforced here for CRIS, to let the least significant bit
  580. * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
  581. */
  582. struct block_device {
  583. dev_t bd_dev; /* not a kdev_t - it's a search key */
  584. int bd_openers;
  585. struct inode * bd_inode; /* will die */
  586. struct super_block * bd_super;
  587. struct mutex bd_mutex; /* open/close mutex */
  588. struct list_head bd_inodes;
  589. void * bd_claiming;
  590. void * bd_holder;
  591. int bd_holders;
  592. bool bd_write_holder;
  593. #ifdef CONFIG_SYSFS
  594. struct list_head bd_holder_disks;
  595. #endif
  596. struct block_device * bd_contains;
  597. unsigned bd_block_size;
  598. struct hd_struct * bd_part;
  599. /* number of times partitions within this device have been opened. */
  600. unsigned bd_part_count;
  601. int bd_invalidated;
  602. struct gendisk * bd_disk;
  603. struct list_head bd_list;
  604. /*
  605. * Private data. You must have bd_claim'ed the block_device
  606. * to use this. NOTE: bd_claim allows an owner to claim
  607. * the same device multiple times, the owner must take special
  608. * care to not mess up bd_private for that case.
  609. */
  610. unsigned long bd_private;
  611. /* The counter of freeze processes */
  612. int bd_fsfreeze_count;
  613. /* Mutex for freeze */
  614. struct mutex bd_fsfreeze_mutex;
  615. };
  616. /*
  617. * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
  618. * radix trees
  619. */
  620. #define PAGECACHE_TAG_DIRTY 0
  621. #define PAGECACHE_TAG_WRITEBACK 1
  622. #define PAGECACHE_TAG_TOWRITE 2
  623. int mapping_tagged(struct address_space *mapping, int tag);
  624. /*
  625. * Might pages of this file be mapped into userspace?
  626. */
  627. static inline int mapping_mapped(struct address_space *mapping)
  628. {
  629. return !prio_tree_empty(&mapping->i_mmap) ||
  630. !list_empty(&mapping->i_mmap_nonlinear);
  631. }
  632. /*
  633. * Might pages of this file have been modified in userspace?
  634. * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
  635. * marks vma as VM_SHARED if it is shared, and the file was opened for
  636. * writing i.e. vma may be mprotected writable even if now readonly.
  637. */
  638. static inline int mapping_writably_mapped(struct address_space *mapping)
  639. {
  640. return mapping->i_mmap_writable != 0;
  641. }
  642. /*
  643. * Use sequence counter to get consistent i_size on 32-bit processors.
  644. */
  645. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  646. #include <linux/seqlock.h>
  647. #define __NEED_I_SIZE_ORDERED
  648. #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
  649. #else
  650. #define i_size_ordered_init(inode) do { } while (0)
  651. #endif
  652. struct posix_acl;
  653. #define ACL_NOT_CACHED ((void *)(-1))
  654. struct inode {
  655. /* RCU path lookup touches following: */
  656. umode_t i_mode;
  657. uid_t i_uid;
  658. gid_t i_gid;
  659. const struct inode_operations *i_op;
  660. struct super_block *i_sb;
  661. spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
  662. unsigned int i_flags;
  663. unsigned long i_state;
  664. #ifdef CONFIG_SECURITY
  665. void *i_security;
  666. #endif
  667. struct mutex i_mutex;
  668. unsigned long dirtied_when; /* jiffies of first dirtying */
  669. struct hlist_node i_hash;
  670. struct list_head i_wb_list; /* backing dev IO list */
  671. struct list_head i_lru; /* inode LRU list */
  672. struct list_head i_sb_list;
  673. union {
  674. struct list_head i_dentry;
  675. struct rcu_head i_rcu;
  676. };
  677. unsigned long i_ino;
  678. atomic_t i_count;
  679. unsigned int i_nlink;
  680. dev_t i_rdev;
  681. unsigned int i_blkbits;
  682. u64 i_version;
  683. loff_t i_size;
  684. #ifdef __NEED_I_SIZE_ORDERED
  685. seqcount_t i_size_seqcount;
  686. #endif
  687. struct timespec i_atime;
  688. struct timespec i_mtime;
  689. struct timespec i_ctime;
  690. blkcnt_t i_blocks;
  691. unsigned short i_bytes;
  692. struct rw_semaphore i_alloc_sem;
  693. const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
  694. struct file_lock *i_flock;
  695. struct address_space *i_mapping;
  696. struct address_space i_data;
  697. #ifdef CONFIG_QUOTA
  698. struct dquot *i_dquot[MAXQUOTAS];
  699. #endif
  700. struct list_head i_devices;
  701. union {
  702. struct pipe_inode_info *i_pipe;
  703. struct block_device *i_bdev;
  704. struct cdev *i_cdev;
  705. };
  706. __u32 i_generation;
  707. #ifdef CONFIG_FSNOTIFY
  708. __u32 i_fsnotify_mask; /* all events this inode cares about */
  709. struct hlist_head i_fsnotify_marks;
  710. #endif
  711. #ifdef CONFIG_IMA
  712. atomic_t i_readcount; /* struct files open RO */
  713. #endif
  714. atomic_t i_writecount;
  715. #ifdef CONFIG_FS_POSIX_ACL
  716. struct posix_acl *i_acl;
  717. struct posix_acl *i_default_acl;
  718. #endif
  719. void *i_private; /* fs or device private pointer */
  720. };
  721. static inline int inode_unhashed(struct inode *inode)
  722. {
  723. return hlist_unhashed(&inode->i_hash);
  724. }
  725. /*
  726. * inode->i_mutex nesting subclasses for the lock validator:
  727. *
  728. * 0: the object of the current VFS operation
  729. * 1: parent
  730. * 2: child/target
  731. * 3: quota file
  732. *
  733. * The locking order between these classes is
  734. * parent -> child -> normal -> xattr -> quota
  735. */
  736. enum inode_i_mutex_lock_class
  737. {
  738. I_MUTEX_NORMAL,
  739. I_MUTEX_PARENT,
  740. I_MUTEX_CHILD,
  741. I_MUTEX_XATTR,
  742. I_MUTEX_QUOTA
  743. };
  744. /*
  745. * NOTE: in a 32bit arch with a preemptable kernel and
  746. * an UP compile the i_size_read/write must be atomic
  747. * with respect to the local cpu (unlike with preempt disabled),
  748. * but they don't need to be atomic with respect to other cpus like in
  749. * true SMP (so they need either to either locally disable irq around
  750. * the read or for example on x86 they can be still implemented as a
  751. * cmpxchg8b without the need of the lock prefix). For SMP compiles
  752. * and 64bit archs it makes no difference if preempt is enabled or not.
  753. */
  754. static inline loff_t i_size_read(const struct inode *inode)
  755. {
  756. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  757. loff_t i_size;
  758. unsigned int seq;
  759. do {
  760. seq = read_seqcount_begin(&inode->i_size_seqcount);
  761. i_size = inode->i_size;
  762. } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
  763. return i_size;
  764. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  765. loff_t i_size;
  766. preempt_disable();
  767. i_size = inode->i_size;
  768. preempt_enable();
  769. return i_size;
  770. #else
  771. return inode->i_size;
  772. #endif
  773. }
  774. /*
  775. * NOTE: unlike i_size_read(), i_size_write() does need locking around it
  776. * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
  777. * can be lost, resulting in subsequent i_size_read() calls spinning forever.
  778. */
  779. static inline void i_size_write(struct inode *inode, loff_t i_size)
  780. {
  781. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  782. write_seqcount_begin(&inode->i_size_seqcount);
  783. inode->i_size = i_size;
  784. write_seqcount_end(&inode->i_size_seqcount);
  785. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  786. preempt_disable();
  787. inode->i_size = i_size;
  788. preempt_enable();
  789. #else
  790. inode->i_size = i_size;
  791. #endif
  792. }
  793. static inline unsigned iminor(const struct inode *inode)
  794. {
  795. return MINOR(inode->i_rdev);
  796. }
  797. static inline unsigned imajor(const struct inode *inode)
  798. {
  799. return MAJOR(inode->i_rdev);
  800. }
  801. extern struct block_device *I_BDEV(struct inode *inode);
  802. struct fown_struct {
  803. rwlock_t lock; /* protects pid, uid, euid fields */
  804. struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
  805. enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
  806. uid_t uid, euid; /* uid/euid of process setting the owner */
  807. int signum; /* posix.1b rt signal to be delivered on IO */
  808. };
  809. /*
  810. * Track a single file's readahead state
  811. */
  812. struct file_ra_state {
  813. pgoff_t start; /* where readahead started */
  814. unsigned int size; /* # of readahead pages */
  815. unsigned int async_size; /* do asynchronous readahead when
  816. there are only # of pages ahead */
  817. unsigned int ra_pages; /* Maximum readahead window */
  818. unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
  819. loff_t prev_pos; /* Cache last read() position */
  820. };
  821. /*
  822. * Check if @index falls in the readahead windows.
  823. */
  824. static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
  825. {
  826. return (index >= ra->start &&
  827. index < ra->start + ra->size);
  828. }
  829. #define FILE_MNT_WRITE_TAKEN 1
  830. #define FILE_MNT_WRITE_RELEASED 2
  831. struct file {
  832. /*
  833. * fu_list becomes invalid after file_free is called and queued via
  834. * fu_rcuhead for RCU freeing
  835. */
  836. union {
  837. struct list_head fu_list;
  838. struct rcu_head fu_rcuhead;
  839. } f_u;
  840. struct path f_path;
  841. #define f_dentry f_path.dentry
  842. #define f_vfsmnt f_path.mnt
  843. const struct file_operations *f_op;
  844. spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
  845. #ifdef CONFIG_SMP
  846. int f_sb_list_cpu;
  847. #endif
  848. atomic_long_t f_count;
  849. unsigned int f_flags;
  850. fmode_t f_mode;
  851. loff_t f_pos;
  852. struct fown_struct f_owner;
  853. const struct cred *f_cred;
  854. struct file_ra_state f_ra;
  855. u64 f_version;
  856. #ifdef CONFIG_SECURITY
  857. void *f_security;
  858. #endif
  859. /* needed for tty driver, and maybe others */
  860. void *private_data;
  861. #ifdef CONFIG_EPOLL
  862. /* Used by fs/eventpoll.c to link all the hooks to this file */
  863. struct list_head f_ep_links;
  864. #endif /* #ifdef CONFIG_EPOLL */
  865. struct address_space *f_mapping;
  866. #ifdef CONFIG_DEBUG_WRITECOUNT
  867. unsigned long f_mnt_write_state;
  868. #endif
  869. };
  870. struct file_handle {
  871. __u32 handle_bytes;
  872. int handle_type;
  873. /* file identifier */
  874. unsigned char f_handle[0];
  875. };
  876. #define get_file(x) atomic_long_inc(&(x)->f_count)
  877. #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
  878. #define file_count(x) atomic_long_read(&(x)->f_count)
  879. #ifdef CONFIG_DEBUG_WRITECOUNT
  880. static inline void file_take_write(struct file *f)
  881. {
  882. WARN_ON(f->f_mnt_write_state != 0);
  883. f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
  884. }
  885. static inline void file_release_write(struct file *f)
  886. {
  887. f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
  888. }
  889. static inline void file_reset_write(struct file *f)
  890. {
  891. f->f_mnt_write_state = 0;
  892. }
  893. static inline void file_check_state(struct file *f)
  894. {
  895. /*
  896. * At this point, either both or neither of these bits
  897. * should be set.
  898. */
  899. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
  900. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
  901. }
  902. static inline int file_check_writeable(struct file *f)
  903. {
  904. if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
  905. return 0;
  906. printk(KERN_WARNING "writeable file with no "
  907. "mnt_want_write()\n");
  908. WARN_ON(1);
  909. return -EINVAL;
  910. }
  911. #else /* !CONFIG_DEBUG_WRITECOUNT */
  912. static inline void file_take_write(struct file *filp) {}
  913. static inline void file_release_write(struct file *filp) {}
  914. static inline void file_reset_write(struct file *filp) {}
  915. static inline void file_check_state(struct file *filp) {}
  916. static inline int file_check_writeable(struct file *filp)
  917. {
  918. return 0;
  919. }
  920. #endif /* CONFIG_DEBUG_WRITECOUNT */
  921. #define MAX_NON_LFS ((1UL<<31) - 1)
  922. /* Page cache limit. The filesystems should put that into their s_maxbytes
  923. limits, otherwise bad things can happen in VM. */
  924. #if BITS_PER_LONG==32
  925. #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
  926. #elif BITS_PER_LONG==64
  927. #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
  928. #endif
  929. #define FL_POSIX 1
  930. #define FL_FLOCK 2
  931. #define FL_ACCESS 8 /* not trying to lock, just looking */
  932. #define FL_EXISTS 16 /* when unlocking, test for existence */
  933. #define FL_LEASE 32 /* lease held on this file */
  934. #define FL_CLOSE 64 /* unlock on close */
  935. #define FL_SLEEP 128 /* A blocking lock */
  936. /*
  937. * Special return value from posix_lock_file() and vfs_lock_file() for
  938. * asynchronous locking.
  939. */
  940. #define FILE_LOCK_DEFERRED 1
  941. /*
  942. * The POSIX file lock owner is determined by
  943. * the "struct files_struct" in the thread group
  944. * (or NULL for no owner - BSD locks).
  945. *
  946. * Lockd stuffs a "host" pointer into this.
  947. */
  948. typedef struct files_struct *fl_owner_t;
  949. struct file_lock_operations {
  950. void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
  951. void (*fl_release_private)(struct file_lock *);
  952. };
  953. struct lock_manager_operations {
  954. int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
  955. void (*fl_notify)(struct file_lock *); /* unblock callback */
  956. int (*fl_grant)(struct file_lock *, struct file_lock *, int);
  957. void (*fl_release_private)(struct file_lock *);
  958. void (*fl_break)(struct file_lock *);
  959. int (*fl_change)(struct file_lock **, int);
  960. };
  961. struct lock_manager {
  962. struct list_head list;
  963. };
  964. void locks_start_grace(struct lock_manager *);
  965. void locks_end_grace(struct lock_manager *);
  966. int locks_in_grace(void);
  967. /* that will die - we need it for nfs_lock_info */
  968. #include <linux/nfs_fs_i.h>
  969. struct file_lock {
  970. struct file_lock *fl_next; /* singly linked list for this inode */
  971. struct list_head fl_link; /* doubly linked list of all locks */
  972. struct list_head fl_block; /* circular list of blocked processes */
  973. fl_owner_t fl_owner;
  974. unsigned char fl_flags;
  975. unsigned char fl_type;
  976. unsigned int fl_pid;
  977. struct pid *fl_nspid;
  978. wait_queue_head_t fl_wait;
  979. struct file *fl_file;
  980. loff_t fl_start;
  981. loff_t fl_end;
  982. struct fasync_struct * fl_fasync; /* for lease break notifications */
  983. unsigned long fl_break_time; /* for nonblocking lease breaks */
  984. const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
  985. const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
  986. union {
  987. struct nfs_lock_info nfs_fl;
  988. struct nfs4_lock_info nfs4_fl;
  989. struct {
  990. struct list_head link; /* link in AFS vnode's pending_locks list */
  991. int state; /* state of grant or error if -ve */
  992. } afs;
  993. } fl_u;
  994. };
  995. /* The following constant reflects the upper bound of the file/locking space */
  996. #ifndef OFFSET_MAX
  997. #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
  998. #define OFFSET_MAX INT_LIMIT(loff_t)
  999. #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
  1000. #endif
  1001. #include <linux/fcntl.h>
  1002. extern void send_sigio(struct fown_struct *fown, int fd, int band);
  1003. #ifdef CONFIG_FILE_LOCKING
  1004. extern int fcntl_getlk(struct file *, struct flock __user *);
  1005. extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
  1006. struct flock __user *);
  1007. #if BITS_PER_LONG == 32
  1008. extern int fcntl_getlk64(struct file *, struct flock64 __user *);
  1009. extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
  1010. struct flock64 __user *);
  1011. #endif
  1012. extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
  1013. extern int fcntl_getlease(struct file *filp);
  1014. /* fs/locks.c */
  1015. void locks_free_lock(struct file_lock *fl);
  1016. extern void locks_init_lock(struct file_lock *);
  1017. extern struct file_lock * locks_alloc_lock(void);
  1018. extern void locks_copy_lock(struct file_lock *, struct file_lock *);
  1019. extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
  1020. extern void locks_remove_posix(struct file *, fl_owner_t);
  1021. extern void locks_remove_flock(struct file *);
  1022. extern void locks_release_private(struct file_lock *);
  1023. extern void posix_test_lock(struct file *, struct file_lock *);
  1024. extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
  1025. extern int posix_lock_file_wait(struct file *, struct file_lock *);
  1026. extern int posix_unblock_lock(struct file *, struct file_lock *);
  1027. extern int vfs_test_lock(struct file *, struct file_lock *);
  1028. extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
  1029. extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
  1030. extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
  1031. extern int __break_lease(struct inode *inode, unsigned int flags);
  1032. extern void lease_get_mtime(struct inode *, struct timespec *time);
  1033. extern int generic_setlease(struct file *, long, struct file_lock **);
  1034. extern int vfs_setlease(struct file *, long, struct file_lock **);
  1035. extern int lease_modify(struct file_lock **, int);
  1036. extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
  1037. extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
  1038. extern void lock_flocks(void);
  1039. extern void unlock_flocks(void);
  1040. #else /* !CONFIG_FILE_LOCKING */
  1041. static inline int fcntl_getlk(struct file *file, struct flock __user *user)
  1042. {
  1043. return -EINVAL;
  1044. }
  1045. static inline int fcntl_setlk(unsigned int fd, struct file *file,
  1046. unsigned int cmd, struct flock __user *user)
  1047. {
  1048. return -EACCES;
  1049. }
  1050. #if BITS_PER_LONG == 32
  1051. static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
  1052. {
  1053. return -EINVAL;
  1054. }
  1055. static inline int fcntl_setlk64(unsigned int fd, struct file *file,
  1056. unsigned int cmd, struct flock64 __user *user)
  1057. {
  1058. return -EACCES;
  1059. }
  1060. #endif
  1061. static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
  1062. {
  1063. return 0;
  1064. }
  1065. static inline int fcntl_getlease(struct file *filp)
  1066. {
  1067. return 0;
  1068. }
  1069. static inline void locks_init_lock(struct file_lock *fl)
  1070. {
  1071. return;
  1072. }
  1073. static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1074. {
  1075. return;
  1076. }
  1077. static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1078. {
  1079. return;
  1080. }
  1081. static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
  1082. {
  1083. return;
  1084. }
  1085. static inline void locks_remove_flock(struct file *filp)
  1086. {
  1087. return;
  1088. }
  1089. static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
  1090. {
  1091. return;
  1092. }
  1093. static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
  1094. struct file_lock *conflock)
  1095. {
  1096. return -ENOLCK;
  1097. }
  1098. static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
  1099. {
  1100. return -ENOLCK;
  1101. }
  1102. static inline int posix_unblock_lock(struct file *filp,
  1103. struct file_lock *waiter)
  1104. {
  1105. return -ENOENT;
  1106. }
  1107. static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
  1108. {
  1109. return 0;
  1110. }
  1111. static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
  1112. struct file_lock *fl, struct file_lock *conf)
  1113. {
  1114. return -ENOLCK;
  1115. }
  1116. static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
  1117. {
  1118. return 0;
  1119. }
  1120. static inline int flock_lock_file_wait(struct file *filp,
  1121. struct file_lock *request)
  1122. {
  1123. return -ENOLCK;
  1124. }
  1125. static inline int __break_lease(struct inode *inode, unsigned int mode)
  1126. {
  1127. return 0;
  1128. }
  1129. static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
  1130. {
  1131. return;
  1132. }
  1133. static inline int generic_setlease(struct file *filp, long arg,
  1134. struct file_lock **flp)
  1135. {
  1136. return -EINVAL;
  1137. }
  1138. static inline int vfs_setlease(struct file *filp, long arg,
  1139. struct file_lock **lease)
  1140. {
  1141. return -EINVAL;
  1142. }
  1143. static inline int lease_modify(struct file_lock **before, int arg)
  1144. {
  1145. return -EINVAL;
  1146. }
  1147. static inline int lock_may_read(struct inode *inode, loff_t start,
  1148. unsigned long len)
  1149. {
  1150. return 1;
  1151. }
  1152. static inline int lock_may_write(struct inode *inode, loff_t start,
  1153. unsigned long len)
  1154. {
  1155. return 1;
  1156. }
  1157. static inline void lock_flocks(void)
  1158. {
  1159. }
  1160. static inline void unlock_flocks(void)
  1161. {
  1162. }
  1163. #endif /* !CONFIG_FILE_LOCKING */
  1164. struct fasync_struct {
  1165. spinlock_t fa_lock;
  1166. int magic;
  1167. int fa_fd;
  1168. struct fasync_struct *fa_next; /* singly linked list */
  1169. struct file *fa_file;
  1170. struct rcu_head fa_rcu;
  1171. };
  1172. #define FASYNC_MAGIC 0x4601
  1173. /* SMP safe fasync helpers: */
  1174. extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
  1175. extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
  1176. extern int fasync_remove_entry(struct file *, struct fasync_struct **);
  1177. extern struct fasync_struct *fasync_alloc(void);
  1178. extern void fasync_free(struct fasync_struct *);
  1179. /* can be called from interrupts */
  1180. extern void kill_fasync(struct fasync_struct **, int, int);
  1181. extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
  1182. extern int f_setown(struct file *filp, unsigned long arg, int force);
  1183. extern void f_delown(struct file *filp);
  1184. extern pid_t f_getown(struct file *filp);
  1185. extern int send_sigurg(struct fown_struct *fown);
  1186. /*
  1187. * Umount options
  1188. */
  1189. #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
  1190. #define MNT_DETACH 0x00000002 /* Just detach from the tree */
  1191. #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
  1192. #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
  1193. #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
  1194. extern struct list_head super_blocks;
  1195. extern spinlock_t sb_lock;
  1196. struct super_block {
  1197. struct list_head s_list; /* Keep this first */
  1198. dev_t s_dev; /* search index; _not_ kdev_t */
  1199. unsigned char s_dirt;
  1200. unsigned char s_blocksize_bits;
  1201. unsigned long s_blocksize;
  1202. loff_t s_maxbytes; /* Max file size */
  1203. struct file_system_type *s_type;
  1204. const struct super_operations *s_op;
  1205. const struct dquot_operations *dq_op;
  1206. const struct quotactl_ops *s_qcop;
  1207. const struct export_operations *s_export_op;
  1208. unsigned long s_flags;
  1209. unsigned long s_magic;
  1210. struct dentry *s_root;
  1211. struct rw_semaphore s_umount;
  1212. struct mutex s_lock;
  1213. int s_count;
  1214. atomic_t s_active;
  1215. #ifdef CONFIG_SECURITY
  1216. void *s_security;
  1217. #endif
  1218. const struct xattr_handler **s_xattr;
  1219. struct list_head s_inodes; /* all inodes */
  1220. struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
  1221. #ifdef CONFIG_SMP
  1222. struct list_head __percpu *s_files;
  1223. #else
  1224. struct list_head s_files;
  1225. #endif
  1226. /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
  1227. struct list_head s_dentry_lru; /* unused dentry lru */
  1228. int s_nr_dentry_unused; /* # of dentry on lru */
  1229. struct block_device *s_bdev;
  1230. struct backing_dev_info *s_bdi;
  1231. struct mtd_info *s_mtd;
  1232. struct list_head s_instances;
  1233. struct quota_info s_dquot; /* Diskquota specific options */
  1234. int s_frozen;
  1235. wait_queue_head_t s_wait_unfrozen;
  1236. char s_id[32]; /* Informational name */
  1237. u8 s_uuid[16]; /* UUID */
  1238. void *s_fs_info; /* Filesystem private info */
  1239. fmode_t s_mode;
  1240. /* Granularity of c/m/atime in ns.
  1241. Cannot be worse than a second */
  1242. u32 s_time_gran;
  1243. /*
  1244. * The next field is for VFS *only*. No filesystems have any business
  1245. * even looking at it. You had been warned.
  1246. */
  1247. struct mutex s_vfs_rename_mutex; /* Kludge */
  1248. /*
  1249. * Filesystem subtype. If non-empty the filesystem type field
  1250. * in /proc/mounts will be "type.subtype"
  1251. */
  1252. char *s_subtype;
  1253. /*
  1254. * Saved mount options for lazy filesystems using
  1255. * generic_show_options()
  1256. */
  1257. char __rcu *s_options;
  1258. const struct dentry_operations *s_d_op; /* default d_op for dentries */
  1259. /*
  1260. * Saved pool identifier for cleancache (-1 means none)
  1261. */
  1262. int cleancache_poolid;
  1263. };
  1264. extern struct timespec current_fs_time(struct super_block *sb);
  1265. /*
  1266. * Snapshotting support.
  1267. */
  1268. enum {
  1269. SB_UNFROZEN = 0,
  1270. SB_FREEZE_WRITE = 1,
  1271. SB_FREEZE_TRANS = 2,
  1272. };
  1273. #define vfs_check_frozen(sb, level) \
  1274. wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
  1275. #define get_fs_excl() atomic_inc(&current->fs_excl)
  1276. #define put_fs_excl() atomic_dec(&current->fs_excl)
  1277. #define has_fs_excl() atomic_read(&current->fs_excl)
  1278. /*
  1279. * until VFS tracks user namespaces for inodes, just make all files
  1280. * belong to init_user_ns
  1281. */
  1282. extern struct user_namespace init_user_ns;
  1283. #define inode_userns(inode) (&init_user_ns)
  1284. extern bool inode_owner_or_capable(const struct inode *inode);
  1285. /* not quite ready to be deprecated, but... */
  1286. extern void lock_super(struct super_block *);
  1287. extern void unlock_super(struct super_block *);
  1288. /*
  1289. * VFS helper functions..
  1290. */
  1291. extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
  1292. extern int vfs_mkdir(struct inode *, struct dentry *, int);
  1293. extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
  1294. extern int vfs_symlink(struct inode *, struct dentry *, const char *);
  1295. extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
  1296. extern int vfs_rmdir(struct inode *, struct dentry *);
  1297. extern int vfs_unlink(struct inode *, struct dentry *);
  1298. extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  1299. /*
  1300. * VFS dentry helper functions.
  1301. */
  1302. extern void dentry_unhash(struct dentry *dentry);
  1303. /*
  1304. * VFS file helper functions.
  1305. */
  1306. extern int file_permission(struct file *, int);
  1307. extern void inode_init_owner(struct inode *inode, const struct inode *dir,
  1308. mode_t mode);
  1309. /*
  1310. * VFS FS_IOC_FIEMAP helper definitions.
  1311. */
  1312. struct fiemap_extent_info {
  1313. unsigned int fi_flags; /* Flags as passed from user */
  1314. unsigned int fi_extents_mapped; /* Number of mapped extents */
  1315. unsigned int fi_extents_max; /* Size of fiemap_extent array */
  1316. struct fiemap_extent __user *fi_extents_start; /* Start of
  1317. fiemap_extent array */
  1318. };
  1319. int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
  1320. u64 phys, u64 len, u32 flags);
  1321. int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
  1322. /*
  1323. * File types
  1324. *
  1325. * NOTE! These match bits 12..15 of stat.st_mode
  1326. * (ie "(i_mode >> 12) & 15").
  1327. */
  1328. #define DT_UNKNOWN 0
  1329. #define DT_FIFO 1
  1330. #define DT_CHR 2
  1331. #define DT_DIR 4
  1332. #define DT_BLK 6
  1333. #define DT_REG 8
  1334. #define DT_LNK 10
  1335. #define DT_SOCK 12
  1336. #define DT_WHT 14
  1337. /*
  1338. * This is the "filldir" function type, used by readdir() to let
  1339. * the kernel specify what kind of dirent layout it wants to have.
  1340. * This allows the kernel to read directories into kernel space or
  1341. * to have different dirent layouts depending on the binary type.
  1342. */
  1343. typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
  1344. struct block_device_operations;
  1345. /* These macros are for out of kernel modules to test that
  1346. * the kernel supports the unlocked_ioctl and compat_ioctl
  1347. * fields in struct file_operations. */
  1348. #define HAVE_COMPAT_IOCTL 1
  1349. #define HAVE_UNLOCKED_IOCTL 1
  1350. /*
  1351. * NOTE:
  1352. * all file operations except setlease can be called without
  1353. * the big kernel lock held in all filesystems.
  1354. */
  1355. struct file_operations {
  1356. struct module *owner;
  1357. loff_t (*llseek) (struct file *, loff_t, int);
  1358. ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
  1359. ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
  1360. ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1361. ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1362. int (*readdir) (struct file *, void *, filldir_t);
  1363. unsigned int (*poll) (struct file *, struct poll_table_struct *);
  1364. long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
  1365. long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
  1366. int (*mmap) (struct file *, struct vm_area_struct *);
  1367. int (*open) (struct inode *, struct file *);
  1368. int (*flush) (struct file *, fl_owner_t id);
  1369. int (*release) (struct inode *, struct file *);
  1370. int (*fsync) (struct file *, int datasync);
  1371. int (*aio_fsync) (struct kiocb *, int datasync);
  1372. int (*fasync) (int, struct file *, int);
  1373. int (*lock) (struct file *, int, struct file_lock *);
  1374. ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
  1375. unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
  1376. int (*check_flags)(int);
  1377. int (*flock) (struct file *, int, struct file_lock *);
  1378. ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
  1379. ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
  1380. int (*setlease)(struct file *, long, struct file_lock **);
  1381. long (*fallocate)(struct file *file, int mode, loff_t offset,
  1382. loff_t len);
  1383. };
  1384. #define IPERM_FLAG_RCU 0x0001
  1385. struct inode_operations {
  1386. struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
  1387. void * (*follow_link) (struct dentry *, struct nameidata *);
  1388. int (*permission) (struct inode *, int, unsigned int);
  1389. int (*check_acl)(struct inode *, int, unsigned int);
  1390. int (*readlink) (struct dentry *, char __user *,int);
  1391. void (*put_link) (struct dentry *, struct nameidata *, void *);
  1392. int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
  1393. int (*link) (struct dentry *,struct inode *,struct dentry *);
  1394. int (*unlink) (struct inode *,struct dentry *);
  1395. int (*symlink) (struct inode *,struct dentry *,const char *);
  1396. int (*mkdir) (struct inode *,struct dentry *,int);
  1397. int (*rmdir) (struct inode *,struct dentry *);
  1398. int (*mknod) (struct inode *,struct dentry *,int,dev_t);
  1399. int (*rename) (struct inode *, struct dentry *,
  1400. struct inode *, struct dentry *);
  1401. void (*truncate) (struct inode *);
  1402. int (*setattr) (struct dentry *, struct iattr *);
  1403. int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
  1404. int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
  1405. ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
  1406. ssize_t (*listxattr) (struct dentry *, char *, size_t);
  1407. int (*removexattr) (struct dentry *, const char *);
  1408. void (*truncate_range)(struct inode *, loff_t, loff_t);
  1409. int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
  1410. u64 len);
  1411. } ____cacheline_aligned;
  1412. struct seq_file;
  1413. ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
  1414. unsigned long nr_segs, unsigned long fast_segs,
  1415. struct iovec *fast_pointer,
  1416. struct iovec **ret_pointer);
  1417. extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
  1418. extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
  1419. extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
  1420. unsigned long, loff_t *);
  1421. extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
  1422. unsigned long, loff_t *);
  1423. struct super_operations {
  1424. struct inode *(*alloc_inode)(struct super_block *sb);
  1425. void (*destroy_inode)(struct inode *);
  1426. void (*dirty_inode) (struct inode *, int flags);
  1427. int (*write_inode) (struct inode *, struct writeback_control *wbc);
  1428. int (*drop_inode) (struct inode *);
  1429. void (*evict_inode) (struct inode *);
  1430. void (*put_super) (struct super_block *);
  1431. void (*write_super) (struct super_block *);
  1432. int (*sync_fs)(struct super_block *sb, int wait);
  1433. int (*freeze_fs) (struct super_block *);
  1434. int (*unfreeze_fs) (struct super_block *);
  1435. int (*statfs) (struct dentry *, struct kstatfs *);
  1436. int (*remount_fs) (struct super_block *, int *, char *);
  1437. void (*umount_begin) (struct super_block *);
  1438. int (*show_options)(struct seq_file *, struct vfsmount *);
  1439. int (*show_devname)(struct seq_file *, struct vfsmount *);
  1440. int (*show_path)(struct seq_file *, struct vfsmount *);
  1441. int (*show_stats)(struct seq_file *, struct vfsmount *);
  1442. #ifdef CONFIG_QUOTA
  1443. ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
  1444. ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
  1445. #endif
  1446. int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
  1447. };
  1448. /*
  1449. * Inode state bits. Protected by inode->i_lock
  1450. *
  1451. * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
  1452. * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
  1453. *
  1454. * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
  1455. * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
  1456. * various stages of removing an inode.
  1457. *
  1458. * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
  1459. *
  1460. * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
  1461. * fdatasync(). i_atime is the usual cause.
  1462. * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
  1463. * these changes separately from I_DIRTY_SYNC so that we
  1464. * don't have to write inode on fdatasync() when only
  1465. * mtime has changed in it.
  1466. * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
  1467. * I_NEW Serves as both a mutex and completion notification.
  1468. * New inodes set I_NEW. If two processes both create
  1469. * the same inode, one of them will release its inode and
  1470. * wait for I_NEW to be released before returning.
  1471. * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
  1472. * also cause waiting on I_NEW, without I_NEW actually
  1473. * being set. find_inode() uses this to prevent returning
  1474. * nearly-dead inodes.
  1475. * I_WILL_FREE Must be set when calling write_inode_now() if i_count
  1476. * is zero. I_FREEING must be set when I_WILL_FREE is
  1477. * cleared.
  1478. * I_FREEING Set when inode is about to be freed but still has dirty
  1479. * pages or buffers attached or the inode itself is still
  1480. * dirty.
  1481. * I_CLEAR Added by end_writeback(). In this state the inode is clean
  1482. * and can be destroyed. Inode keeps I_FREEING.
  1483. *
  1484. * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
  1485. * prohibited for many purposes. iget() must wait for
  1486. * the inode to be completely released, then create it
  1487. * anew. Other functions will just ignore such inodes,
  1488. * if appropriate. I_NEW is used for waiting.
  1489. *
  1490. * I_SYNC Synchonized write of dirty inode data. The bits is
  1491. * set during data writeback, and cleared with a wakeup
  1492. * on the bit address once it is done.
  1493. *
  1494. * Q: What is the difference between I_WILL_FREE and I_FREEING?
  1495. */
  1496. #define I_DIRTY_SYNC (1 << 0)
  1497. #define I_DIRTY_DATASYNC (1 << 1)
  1498. #define I_DIRTY_PAGES (1 << 2)
  1499. #define __I_NEW 3
  1500. #define I_NEW (1 << __I_NEW)
  1501. #define I_WILL_FREE (1 << 4)
  1502. #define I_FREEING (1 << 5)
  1503. #define I_CLEAR (1 << 6)
  1504. #define __I_SYNC 7
  1505. #define I_SYNC (1 << __I_SYNC)
  1506. #define I_REFERENCED (1 << 8)
  1507. #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
  1508. extern void __mark_inode_dirty(struct inode *, int);
  1509. static inline void mark_inode_dirty(struct inode *inode)
  1510. {
  1511. __mark_inode_dirty(inode, I_DIRTY);
  1512. }
  1513. static inline void mark_inode_dirty_sync(struct inode *inode)
  1514. {
  1515. __mark_inode_dirty(inode, I_DIRTY_SYNC);
  1516. }
  1517. /**
  1518. * inc_nlink - directly increment an inode's link count
  1519. * @inode: inode
  1520. *
  1521. * This is a low-level filesystem helper to replace any
  1522. * direct filesystem manipulation of i_nlink. Currently,
  1523. * it is only here for parity with dec_nlink().
  1524. */
  1525. static inline void inc_nlink(struct inode *inode)
  1526. {
  1527. inode->i_nlink++;
  1528. }
  1529. static inline void inode_inc_link_count(struct inode *inode)
  1530. {
  1531. inc_nlink(inode);
  1532. mark_inode_dirty(inode);
  1533. }
  1534. /**
  1535. * drop_nlink - directly drop an inode's link count
  1536. * @inode: inode
  1537. *
  1538. * This is a low-level filesystem helper to replace any
  1539. * direct filesystem manipulation of i_nlink. In cases
  1540. * where we are attempting to track writes to the
  1541. * filesystem, a decrement to zero means an imminent
  1542. * write when the file is truncated and actually unlinked
  1543. * on the filesystem.
  1544. */
  1545. static inline void drop_nlink(struct inode *inode)
  1546. {
  1547. inode->i_nlink--;
  1548. }
  1549. /**
  1550. * clear_nlink - directly zero an inode's link count
  1551. * @inode: inode
  1552. *
  1553. * This is a low-level filesystem helper to replace any
  1554. * direct filesystem manipulation of i_nlink. See
  1555. * drop_nlink() for why we care about i_nlink hitting zero.
  1556. */
  1557. static inline void clear_nlink(struct inode *inode)
  1558. {
  1559. inode->i_nlink = 0;
  1560. }
  1561. static inline void inode_dec_link_count(struct inode *inode)
  1562. {
  1563. drop_nlink(inode);
  1564. mark_inode_dirty(inode);
  1565. }
  1566. /**
  1567. * inode_inc_iversion - increments i_version
  1568. * @inode: inode that need to be updated
  1569. *
  1570. * Every time the inode is modified, the i_version field will be incremented.
  1571. * The filesystem has to be mounted with i_version flag
  1572. */
  1573. static inline void inode_inc_iversion(struct inode *inode)
  1574. {
  1575. spin_lock(&inode->i_lock);
  1576. inode->i_version++;
  1577. spin_unlock(&inode->i_lock);
  1578. }
  1579. extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
  1580. static inline void file_accessed(struct file *file)
  1581. {
  1582. if (!(file->f_flags & O_NOATIME))
  1583. touch_atime(file->f_path.mnt, file->f_path.dentry);
  1584. }
  1585. int sync_inode(struct inode *inode, struct writeback_control *wbc);
  1586. int sync_inode_metadata(struct inode *inode, int wait);
  1587. struct file_system_type {
  1588. const char *name;
  1589. int fs_flags;
  1590. struct dentry *(*mount) (struct file_system_type *, int,
  1591. const char *, void *);
  1592. void (*kill_sb) (struct super_block *);
  1593. struct module *owner;
  1594. struct file_system_type * next;
  1595. struct list_head fs_supers;
  1596. struct lock_class_key s_lock_key;
  1597. struct lock_class_key s_umount_key;
  1598. struct lock_class_key s_vfs_rename_key;
  1599. struct lock_class_key i_lock_key;
  1600. struct lock_class_key i_mutex_key;
  1601. struct lock_class_key i_mutex_dir_key;
  1602. struct lock_class_key i_alloc_sem_key;
  1603. };
  1604. extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
  1605. void *data, int (*fill_super)(struct super_block *, void *, int));
  1606. extern struct dentry *mount_bdev(struct file_system_type *fs_type,
  1607. int flags, const char *dev_name, void *data,
  1608. int (*fill_super)(struct super_block *, void *, int));
  1609. extern struct dentry *mount_single(struct file_system_type *fs_type,
  1610. int flags, void *data,
  1611. int (*fill_super)(struct super_block *, void *, int));
  1612. extern struct dentry *mount_nodev(struct file_system_type *fs_type,
  1613. int flags, void *data,
  1614. int (*fill_super)(struct super_block *, void *, int));
  1615. void generic_shutdown_super(struct super_block *sb);
  1616. void kill_block_super(struct super_block *sb);
  1617. void kill_anon_super(struct super_block *sb);
  1618. void kill_litter_super(struct super_block *sb);
  1619. void deactivate_super(struct super_block *sb);
  1620. void deactivate_locked_super(struct super_block *sb);
  1621. int set_anon_super(struct super_block *s, void *data);
  1622. struct super_block *sget(struct file_system_type *type,
  1623. int (*test)(struct super_block *,void *),
  1624. int (*set)(struct super_block *,void *),
  1625. void *data);
  1626. extern struct dentry *mount_pseudo(struct file_system_type *, char *,
  1627. const struct super_operations *ops,
  1628. const struct dentry_operations *dops,
  1629. unsigned long);
  1630. static inline void sb_mark_dirty(struct super_block *sb)
  1631. {
  1632. sb->s_dirt = 1;
  1633. }
  1634. static inline void sb_mark_clean(struct super_block *sb)
  1635. {
  1636. sb->s_dirt = 0;
  1637. }
  1638. static inline int sb_is_dirty(struct super_block *sb)
  1639. {
  1640. return sb->s_dirt;
  1641. }
  1642. /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
  1643. #define fops_get(fops) \
  1644. (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
  1645. #define fops_put(fops) \
  1646. do { if (fops) module_put((fops)->owner); } while(0)
  1647. extern int register_filesystem(struct file_system_type *);
  1648. extern int unregister_filesystem(struct file_system_type *);
  1649. extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
  1650. #define kern_mount(type) kern_mount_data(type, NULL)
  1651. extern int may_umount_tree(struct vfsmount *);
  1652. extern int may_umount(struct vfsmount *);
  1653. extern long do_mount(char *, char *, char *, unsigned long, void *);
  1654. extern struct vfsmount *collect_mounts(struct path *);
  1655. extern void drop_collected_mounts(struct vfsmount *);
  1656. extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
  1657. struct vfsmount *);
  1658. extern int vfs_statfs(struct path *, struct kstatfs *);
  1659. extern int user_statfs(const char __user *, struct kstatfs *);
  1660. extern int fd_statfs(int, struct kstatfs *);
  1661. extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
  1662. extern int freeze_super(struct super_block *super);
  1663. extern int thaw_super(struct super_block *super);
  1664. extern int current_umask(void);
  1665. /* /sys/fs */
  1666. extern struct kobject *fs_kobj;
  1667. #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
  1668. extern int rw_verify_area(int, struct file *, loff_t *, size_t);
  1669. #define FLOCK_VERIFY_READ 1
  1670. #define FLOCK_VERIFY_WRITE 2
  1671. #ifdef CONFIG_FILE_LOCKING
  1672. extern int locks_mandatory_locked(struct inode *);
  1673. extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
  1674. /*
  1675. * Candidates for mandatory locking have the setgid bit set
  1676. * but no group execute bit - an otherwise meaningless combination.
  1677. */
  1678. static inline int __mandatory_lock(struct inode *ino)
  1679. {
  1680. return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
  1681. }
  1682. /*
  1683. * ... and these candidates should be on MS_MANDLOCK mounted fs,
  1684. * otherwise these will be advisory locks
  1685. */
  1686. static inline int mandatory_lock(struct inode *ino)
  1687. {
  1688. return IS_MANDLOCK(ino) && __mandatory_lock(ino);
  1689. }
  1690. static inline int locks_verify_locked(struct inode *inode)
  1691. {
  1692. if (mandatory_lock(inode))
  1693. return locks_mandatory_locked(inode);
  1694. return 0;
  1695. }
  1696. static inline int locks_verify_truncate(struct inode *inode,
  1697. struct file *filp,
  1698. loff_t size)
  1699. {
  1700. if (inode->i_flock && mandatory_lock(inode))
  1701. return locks_mandatory_area(
  1702. FLOCK_VERIFY_WRITE, inode, filp,
  1703. size < inode->i_size ? size : inode->i_size,
  1704. (size < inode->i_size ? inode->i_size - size
  1705. : size - inode->i_size)
  1706. );
  1707. return 0;
  1708. }
  1709. static inline int break_lease(struct inode *inode, unsigned int mode)
  1710. {
  1711. if (inode->i_flock)
  1712. return __break_lease(inode, mode);
  1713. return 0;
  1714. }
  1715. #else /* !CONFIG_FILE_LOCKING */
  1716. static inline int locks_mandatory_locked(struct inode *inode)
  1717. {
  1718. return 0;
  1719. }
  1720. static inline int locks_mandatory_area(int rw, struct inode *inode,
  1721. struct file *filp, loff_t offset,
  1722. size_t count)
  1723. {
  1724. return 0;
  1725. }
  1726. static inline int __mandatory_lock(struct inode *inode)
  1727. {
  1728. return 0;
  1729. }
  1730. static inline int mandatory_lock(struct inode *inode)
  1731. {
  1732. return 0;
  1733. }
  1734. static inline int locks_verify_locked(struct inode *inode)
  1735. {
  1736. return 0;
  1737. }
  1738. static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
  1739. size_t size)
  1740. {
  1741. return 0;
  1742. }
  1743. static inline int break_lease(struct inode *inode, unsigned int mode)
  1744. {
  1745. return 0;
  1746. }
  1747. #endif /* CONFIG_FILE_LOCKING */
  1748. /* fs/open.c */
  1749. extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
  1750. struct file *filp);
  1751. extern int do_fallocate(struct file *file, int mode, loff_t offset,
  1752. loff_t len);
  1753. extern long do_sys_open(int dfd, const char __user *filename, int flags,
  1754. int mode);
  1755. extern struct file *filp_open(const char *, int, int);
  1756. extern struct file *file_open_root(struct dentry *, struct vfsmount *,
  1757. const char *, int);
  1758. extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
  1759. const struct cred *);
  1760. extern int filp_close(struct file *, fl_owner_t id);
  1761. extern char * getname(const char __user *);
  1762. /* fs/ioctl.c */
  1763. extern int ioctl_preallocate(struct file *filp, void __user *argp);
  1764. /* fs/dcache.c */
  1765. extern void __init vfs_caches_init_early(void);
  1766. extern void __init vfs_caches_init(unsigned long);
  1767. extern struct kmem_cache *names_cachep;
  1768. #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
  1769. #define __getname() __getname_gfp(GFP_KERNEL)
  1770. #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
  1771. #ifndef CONFIG_AUDITSYSCALL
  1772. #define putname(name) __putname(name)
  1773. #else
  1774. extern void putname(const char *name);
  1775. #endif
  1776. #ifdef CONFIG_BLOCK
  1777. extern int register_blkdev(unsigned int, const char *);
  1778. extern void unregister_blkdev(unsigned int, const char *);
  1779. extern struct block_device *bdget(dev_t);
  1780. extern struct block_device *bdgrab(struct block_device *bdev);
  1781. extern void bd_set_size(struct block_device *, loff_t size);
  1782. extern void bd_forget(struct inode *inode);
  1783. extern void bdput(struct block_device *);
  1784. extern void invalidate_bdev(struct block_device *);
  1785. extern int sync_blockdev(struct block_device *bdev);
  1786. extern struct super_block *freeze_bdev(struct block_device *);
  1787. extern void emergency_thaw_all(void);
  1788. extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
  1789. extern int fsync_bdev(struct block_device *);
  1790. #else
  1791. static inline void bd_forget(struct inode *inode) {}
  1792. static inline int sync_blockdev(struct block_device *bdev) { return 0; }
  1793. static inline void invalidate_bdev(struct block_device *bdev) {}
  1794. static inline struct super_block *freeze_bdev(struct block_device *sb)
  1795. {
  1796. return NULL;
  1797. }
  1798. static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
  1799. {
  1800. return 0;
  1801. }
  1802. #endif
  1803. extern int sync_filesystem(struct super_block *);
  1804. extern const struct file_operations def_blk_fops;
  1805. extern const struct file_operations def_chr_fops;
  1806. extern const struct file_operations bad_sock_fops;
  1807. extern const struct file_operations def_fifo_fops;
  1808. #ifdef CONFIG_BLOCK
  1809. extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
  1810. extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
  1811. extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
  1812. extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
  1813. extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
  1814. void *holder);
  1815. extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
  1816. void *holder);
  1817. extern int blkdev_put(struct block_device *bdev, fmode_t mode);
  1818. #ifdef CONFIG_SYSFS
  1819. extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
  1820. extern void bd_unlink_disk_holder(struct block_device *bdev,
  1821. struct gendisk *disk);
  1822. #else
  1823. static inline int bd_link_disk_holder(struct block_device *bdev,
  1824. struct gendisk *disk)
  1825. {
  1826. return 0;
  1827. }
  1828. static inline void bd_unlink_disk_holder(struct block_device *bdev,
  1829. struct gendisk *disk)
  1830. {
  1831. }
  1832. #endif
  1833. #endif
  1834. /* fs/char_dev.c */
  1835. #define CHRDEV_MAJOR_HASH_SIZE 255
  1836. extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
  1837. extern int register_chrdev_region(dev_t, unsigned, const char *);
  1838. extern int __register_chrdev(unsigned int major, unsigned int baseminor,
  1839. unsigned int count, const char *name,
  1840. const struct file_operations *fops);
  1841. extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
  1842. unsigned int count, const char *name);
  1843. extern void unregister_chrdev_region(dev_t, unsigned);
  1844. extern void chrdev_show(struct seq_file *,off_t);
  1845. static inline int register_chrdev(unsigned int major, const char *name,
  1846. const struct file_operations *fops)
  1847. {
  1848. return __register_chrdev(major, 0, 256, name, fops);
  1849. }
  1850. static inline void unregister_chrdev(unsigned int major, const char *name)
  1851. {
  1852. __unregister_chrdev(major, 0, 256, name);
  1853. }
  1854. /* fs/block_dev.c */
  1855. #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
  1856. #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
  1857. #ifdef CONFIG_BLOCK
  1858. #define BLKDEV_MAJOR_HASH_SIZE 255
  1859. extern const char *__bdevname(dev_t, char *buffer);
  1860. extern const char *bdevname(struct block_device *bdev, char *buffer);
  1861. extern struct block_device *lookup_bdev(const char *);
  1862. extern void blkdev_show(struct seq_file *,off_t);
  1863. #else
  1864. #define BLKDEV_MAJOR_HASH_SIZE 0
  1865. #endif
  1866. extern void init_special_inode(struct inode *, umode_t, dev_t);
  1867. /* Invalid inode operations -- fs/bad_inode.c */
  1868. extern void make_bad_inode(struct inode *);
  1869. extern int is_bad_inode(struct inode *);
  1870. extern const struct file_operations read_pipefifo_fops;
  1871. extern const struct file_operations write_pipefifo_fops;
  1872. extern const struct file_operations rdwr_pipefifo_fops;
  1873. extern int fs_may_remount_ro(struct super_block *);
  1874. #ifdef CONFIG_BLOCK
  1875. /*
  1876. * return READ, READA, or WRITE
  1877. */
  1878. #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
  1879. /*
  1880. * return data direction, READ or WRITE
  1881. */
  1882. #define bio_data_dir(bio) ((bio)->bi_rw & 1)
  1883. extern void check_disk_size_change(struct gendisk *disk,
  1884. struct block_device *bdev);
  1885. extern int revalidate_disk(struct gendisk *);
  1886. extern int check_disk_change(struct block_device *);
  1887. extern int __invalidate_device(struct block_device *, bool);
  1888. extern int invalidate_partition(struct gendisk *, int);
  1889. #endif
  1890. unsigned long invalidate_mapping_pages(struct address_space *mapping,
  1891. pgoff_t start, pgoff_t end);
  1892. static inline void invalidate_remote_inode(struct inode *inode)
  1893. {
  1894. if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
  1895. S_ISLNK(inode->i_mode))
  1896. invalidate_mapping_pages(inode->i_mapping, 0, -1);
  1897. }
  1898. extern int invalidate_inode_pages2(struct address_space *mapping);
  1899. extern int invalidate_inode_pages2_range(struct address_space *mapping,
  1900. pgoff_t start, pgoff_t end);
  1901. extern int write_inode_now(struct inode *, int);
  1902. extern int filemap_fdatawrite(struct address_space *);
  1903. extern int filemap_flush(struct address_space *);
  1904. extern int filemap_fdatawait(struct address_space *);
  1905. extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
  1906. loff_t lend);
  1907. extern int filemap_write_and_wait(struct address_space *mapping);
  1908. extern int filemap_write_and_wait_range(struct address_space *mapping,
  1909. loff_t lstart, loff_t lend);
  1910. extern int __filemap_fdatawrite_range(struct address_space *mapping,
  1911. loff_t start, loff_t end, int sync_mode);
  1912. extern int filemap_fdatawrite_range(struct address_space *mapping,
  1913. loff_t start, loff_t end);
  1914. extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
  1915. int datasync);
  1916. extern int vfs_fsync(struct file *file, int datasync);
  1917. extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
  1918. extern void sync_supers(void);
  1919. extern void emergency_sync(void);
  1920. extern void emergency_remount(void);
  1921. #ifdef CONFIG_BLOCK
  1922. extern sector_t bmap(struct inode *, sector_t);
  1923. #endif
  1924. extern int notify_change(struct dentry *, struct iattr *);
  1925. extern int inode_permission(struct inode *, int);
  1926. extern int generic_permission(struct inode *, int, unsigned int,
  1927. int (*check_acl)(struct inode *, int, unsigned int));
  1928. static inline bool execute_ok(struct inode *inode)
  1929. {
  1930. return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
  1931. }
  1932. extern int get_write_access(struct inode *);
  1933. extern int deny_write_access(struct file *);
  1934. static inline void put_write_access(struct inode * inode)
  1935. {
  1936. atomic_dec(&inode->i_writecount);
  1937. }
  1938. static inline void allow_write_access(struct file *file)
  1939. {
  1940. if (file)
  1941. atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
  1942. }
  1943. #ifdef CONFIG_IMA
  1944. static inline void i_readcount_dec(struct inode *inode)
  1945. {
  1946. BUG_ON(!atomic_read(&inode->i_readcount));
  1947. atomic_dec(&inode->i_readcount);
  1948. }
  1949. static inline void i_readcount_inc(struct inode *inode)
  1950. {
  1951. atomic_inc(&inode->i_readcount);
  1952. }
  1953. #else
  1954. static inline void i_readcount_dec(struct inode *inode)
  1955. {
  1956. return;
  1957. }
  1958. static inline void i_readcount_inc(struct inode *inode)
  1959. {
  1960. return;
  1961. }
  1962. #endif
  1963. extern int do_pipe_flags(int *, int);
  1964. extern struct file *create_read_pipe(struct file *f, int flags);
  1965. extern struct file *create_write_pipe(int flags);
  1966. extern void free_write_pipe(struct file *);
  1967. extern int kernel_read(struct file *, loff_t, char *, unsigned long);
  1968. extern struct file * open_exec(const char *);
  1969. /* fs/dcache.c -- generic fs support functions */
  1970. extern int is_subdir(struct dentry *, struct dentry *);
  1971. extern int path_is_under(struct path *, struct path *);
  1972. extern ino_t find_inode_number(struct dentry *, struct qstr *);
  1973. #include <linux/err.h>
  1974. /* needed for stackable file system support */
  1975. extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
  1976. extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
  1977. extern int inode_init_always(struct super_block *, struct inode *);
  1978. extern void inode_init_once(struct inode *);
  1979. extern void address_space_init_once(struct address_space *mapping);
  1980. extern void ihold(struct inode * inode);
  1981. extern void iput(struct inode *);
  1982. extern struct inode * igrab(struct inode *);
  1983. extern ino_t iunique(struct super_block *, ino_t);
  1984. extern int inode_needs_sync(struct inode *inode);
  1985. extern int generic_delete_inode(struct inode *inode);
  1986. extern int generic_drop_inode(struct inode *inode);
  1987. extern struct inode *ilookup5_nowait(struct super_block *sb,
  1988. unsigned long hashval, int (*test)(struct inode *, void *),
  1989. void *data);
  1990. extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
  1991. int (*test)(struct inode *, void *), void *data);
  1992. extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
  1993. extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
  1994. extern struct inode * iget_locked(struct super_block *, unsigned long);
  1995. extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
  1996. extern int insert_inode_locked(struct inode *);
  1997. extern void unlock_new_inode(struct inode *);
  1998. extern unsigned int get_next_ino(void);
  1999. extern void __iget(struct inode * inode);
  2000. extern void iget_failed(struct inode *);
  2001. extern void end_writeback(struct inode *);
  2002. extern void __destroy_inode(struct inode *);
  2003. extern struct inode *new_inode(struct super_block *);
  2004. extern void free_inode_nonrcu(struct inode *inode);
  2005. extern int should_remove_suid(struct dentry *);
  2006. extern int file_remove_suid(struct file *);
  2007. extern void __insert_inode_hash(struct inode *, unsigned long hashval);
  2008. extern void remove_inode_hash(struct inode *);
  2009. static inline void insert_inode_hash(struct inode *inode)
  2010. {
  2011. __insert_inode_hash(inode, inode->i_ino);
  2012. }
  2013. extern void inode_sb_list_add(struct inode *inode);
  2014. #ifdef CONFIG_BLOCK
  2015. extern void submit_bio(int, struct bio *);
  2016. extern int bdev_read_only(struct block_device *);
  2017. #endif
  2018. extern int set_blocksize(struct block_device *, int);
  2019. extern int sb_set_blocksize(struct super_block *, int);
  2020. extern int sb_min_blocksize(struct super_block *, int);
  2021. extern int generic_file_mmap(struct file *, struct vm_area_struct *);
  2022. extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
  2023. extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
  2024. int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
  2025. extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2026. extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
  2027. loff_t *);
  2028. extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2029. extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
  2030. unsigned long *, loff_t, loff_t *, size_t, size_t);
  2031. extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
  2032. unsigned long, loff_t, loff_t *, size_t, ssize_t);
  2033. extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
  2034. extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
  2035. extern int generic_segment_checks(const struct iovec *iov,
  2036. unsigned long *nr_segs, size_t *count, int access_flags);
  2037. /* fs/block_dev.c */
  2038. extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
  2039. unsigned long nr_segs, loff_t pos);
  2040. extern int blkdev_fsync(struct file *filp, int datasync);
  2041. /* fs/splice.c */
  2042. extern ssize_t generic_file_splice_read(struct file *, loff_t *,
  2043. struct pipe_inode_info *, size_t, unsigned int);
  2044. extern ssize_t default_file_splice_read(struct file *, loff_t *,
  2045. struct pipe_inode_info *, size_t, unsigned int);
  2046. extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
  2047. struct file *, loff_t *, size_t, unsigned int);
  2048. extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
  2049. struct file *out, loff_t *, size_t len, unsigned int flags);
  2050. extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
  2051. size_t len, unsigned int flags);
  2052. extern void
  2053. file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
  2054. extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
  2055. extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
  2056. extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
  2057. extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
  2058. int origin);
  2059. extern int generic_file_open(struct inode * inode, struct file * filp);
  2060. extern int nonseekable_open(struct inode * inode, struct file * filp);
  2061. #ifdef CONFIG_FS_XIP
  2062. extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
  2063. loff_t *ppos);
  2064. extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
  2065. extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
  2066. size_t len, loff_t *ppos);
  2067. extern int xip_truncate_page(struct address_space *mapping, loff_t from);
  2068. #else
  2069. static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
  2070. {
  2071. return 0;
  2072. }
  2073. #endif
  2074. #ifdef CONFIG_BLOCK
  2075. typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
  2076. loff_t file_offset);
  2077. enum {
  2078. /* need locking between buffered and direct access */
  2079. DIO_LOCKING = 0x01,
  2080. /* filesystem does not support filling holes */
  2081. DIO_SKIP_HOLES = 0x02,
  2082. };
  2083. void dio_end_io(struct bio *bio, int error);
  2084. ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
  2085. struct block_device *bdev, const struct iovec *iov, loff_t offset,
  2086. unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
  2087. dio_submit_t submit_io, int flags);
  2088. static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
  2089. struct inode *inode, struct block_device *bdev, const struct iovec *iov,
  2090. loff_t offset, unsigned long nr_segs, get_block_t get_block,
  2091. dio_iodone_t end_io)
  2092. {
  2093. return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
  2094. nr_segs, get_block, end_io, NULL,
  2095. DIO_LOCKING | DIO_SKIP_HOLES);
  2096. }
  2097. #endif
  2098. extern const struct file_operations generic_ro_fops;
  2099. #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
  2100. extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
  2101. extern int vfs_follow_link(struct nameidata *, const char *);
  2102. extern int page_readlink(struct dentry *, char __user *, int);
  2103. extern void *page_follow_link_light(struct dentry *, struct nameidata *);
  2104. extern void page_put_link(struct dentry *, struct nameidata *, void *);
  2105. extern int __page_symlink(struct inode *inode, const char *symname, int len,
  2106. int nofs);
  2107. extern int page_symlink(struct inode *inode, const char *symname, int len);
  2108. extern const struct inode_operations page_symlink_inode_operations;
  2109. extern int generic_readlink(struct dentry *, char __user *, int);
  2110. extern void generic_fillattr(struct inode *, struct kstat *);
  2111. extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2112. void __inode_add_bytes(struct inode *inode, loff_t bytes);
  2113. void inode_add_bytes(struct inode *inode, loff_t bytes);
  2114. void inode_sub_bytes(struct inode *inode, loff_t bytes);
  2115. loff_t inode_get_bytes(struct inode *inode);
  2116. void inode_set_bytes(struct inode *inode, loff_t bytes);
  2117. extern int vfs_readdir(struct file *, filldir_t, void *);
  2118. extern int vfs_stat(const char __user *, struct kstat *);
  2119. extern int vfs_lstat(const char __user *, struct kstat *);
  2120. extern int vfs_fstat(unsigned int, struct kstat *);
  2121. extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
  2122. extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
  2123. unsigned long arg);
  2124. extern int __generic_block_fiemap(struct inode *inode,
  2125. struct fiemap_extent_info *fieinfo,
  2126. loff_t start, loff_t len,
  2127. get_block_t *get_block);
  2128. extern int generic_block_fiemap(struct inode *inode,
  2129. struct fiemap_extent_info *fieinfo, u64 start,
  2130. u64 len, get_block_t *get_block);
  2131. extern void get_filesystem(struct file_system_type *fs);
  2132. extern void put_filesystem(struct file_system_type *fs);
  2133. extern struct file_system_type *get_fs_type(const char *name);
  2134. extern struct super_block *get_super(struct block_device *);
  2135. extern struct super_block *get_active_super(struct block_device *bdev);
  2136. extern struct super_block *user_get_super(dev_t);
  2137. extern void drop_super(struct super_block *sb);
  2138. extern void iterate_supers(void (*)(struct super_block *, void *), void *);
  2139. extern void iterate_supers_type(struct file_system_type *,
  2140. void (*)(struct super_block *, void *), void *);
  2141. extern int dcache_dir_open(struct inode *, struct file *);
  2142. extern int dcache_dir_close(struct inode *, struct file *);
  2143. extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
  2144. extern int dcache_readdir(struct file *, void *, filldir_t);
  2145. extern int simple_setattr(struct dentry *, struct iattr *);
  2146. extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2147. extern int simple_statfs(struct dentry *, struct kstatfs *);
  2148. extern int simple_link(struct dentry *, struct inode *, struct dentry *);
  2149. extern int simple_unlink(struct inode *, struct dentry *);
  2150. extern int simple_rmdir(struct inode *, struct dentry *);
  2151. extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  2152. extern int noop_fsync(struct file *, int);
  2153. extern int simple_empty(struct dentry *);
  2154. extern int simple_readpage(struct file *file, struct page *page);
  2155. extern int simple_write_begin(struct file *file, struct address_space *mapping,
  2156. loff_t pos, unsigned len, unsigned flags,
  2157. struct page **pagep, void **fsdata);
  2158. extern int simple_write_end(struct file *file, struct address_space *mapping,
  2159. loff_t pos, unsigned len, unsigned copied,
  2160. struct page *page, void *fsdata);
  2161. extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
  2162. extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
  2163. extern const struct file_operations simple_dir_operations;
  2164. extern const struct inode_operations simple_dir_inode_operations;
  2165. struct tree_descr { char *name; const struct file_operations *ops; int mode; };
  2166. struct dentry *d_alloc_name(struct dentry *, const char *);
  2167. extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
  2168. extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
  2169. extern void simple_release_fs(struct vfsmount **mount, int *count);
  2170. extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
  2171. loff_t *ppos, const void *from, size_t available);
  2172. extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
  2173. const void __user *from, size_t count);
  2174. extern int generic_file_fsync(struct file *, int);
  2175. extern int generic_check_addressable(unsigned, u64);
  2176. #ifdef CONFIG_MIGRATION
  2177. extern int buffer_migrate_page(struct address_space *,
  2178. struct page *, struct page *);
  2179. #else
  2180. #define buffer_migrate_page NULL
  2181. #endif
  2182. extern int inode_change_ok(const struct inode *, struct iattr *);
  2183. extern int inode_newsize_ok(const struct inode *, loff_t offset);
  2184. extern void setattr_copy(struct inode *inode, const struct iattr *attr);
  2185. extern void file_update_time(struct file *file);
  2186. extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
  2187. extern void save_mount_options(struct super_block *sb, char *options);
  2188. extern void replace_mount_options(struct super_block *sb, char *options);
  2189. static inline ino_t parent_ino(struct dentry *dentry)
  2190. {
  2191. ino_t res;
  2192. /*
  2193. * Don't strictly need d_lock here? If the parent ino could change
  2194. * then surely we'd have a deeper race in the caller?
  2195. */
  2196. spin_lock(&dentry->d_lock);
  2197. res = dentry->d_parent->d_inode->i_ino;
  2198. spin_unlock(&dentry->d_lock);
  2199. return res;
  2200. }
  2201. /* Transaction based IO helpers */
  2202. /*
  2203. * An argresp is stored in an allocated page and holds the
  2204. * size of the argument or response, along with its content
  2205. */
  2206. struct simple_transaction_argresp {
  2207. ssize_t size;
  2208. char data[0];
  2209. };
  2210. #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
  2211. char *simple_transaction_get(struct file *file, const char __user *buf,
  2212. size_t size);
  2213. ssize_t simple_transaction_read(struct file *file, char __user *buf,
  2214. size_t size, loff_t *pos);
  2215. int simple_transaction_release(struct inode *inode, struct file *file);
  2216. void simple_transaction_set(struct file *file, size_t n);
  2217. /*
  2218. * simple attribute files
  2219. *
  2220. * These attributes behave similar to those in sysfs:
  2221. *
  2222. * Writing to an attribute immediately sets a value, an open file can be
  2223. * written to multiple times.
  2224. *
  2225. * Reading from an attribute creates a buffer from the value that might get
  2226. * read with multiple read calls. When the attribute has been read
  2227. * completely, no further read calls are possible until the file is opened
  2228. * again.
  2229. *
  2230. * All attributes contain a text representation of a numeric value
  2231. * that are accessed with the get() and set() functions.
  2232. */
  2233. #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
  2234. static int __fops ## _open(struct inode *inode, struct file *file) \
  2235. { \
  2236. __simple_attr_check_format(__fmt, 0ull); \
  2237. return simple_attr_open(inode, file, __get, __set, __fmt); \
  2238. } \
  2239. static const struct file_operations __fops = { \
  2240. .owner = THIS_MODULE, \
  2241. .open = __fops ## _open, \
  2242. .release = simple_attr_release, \
  2243. .read = simple_attr_read, \
  2244. .write = simple_attr_write, \
  2245. .llseek = generic_file_llseek, \
  2246. };
  2247. static inline void __attribute__((format(printf, 1, 2)))
  2248. __simple_attr_check_format(const char *fmt, ...)
  2249. {
  2250. /* don't do anything, just let the compiler check the arguments; */
  2251. }
  2252. int simple_attr_open(struct inode *inode, struct file *file,
  2253. int (*get)(void *, u64 *), int (*set)(void *, u64),
  2254. const char *fmt);
  2255. int simple_attr_release(struct inode *inode, struct file *file);
  2256. ssize_t simple_attr_read(struct file *file, char __user *buf,
  2257. size_t len, loff_t *ppos);
  2258. ssize_t simple_attr_write(struct file *file, const char __user *buf,
  2259. size_t len, loff_t *ppos);
  2260. struct ctl_table;
  2261. int proc_nr_files(struct ctl_table *table, int write,
  2262. void __user *buffer, size_t *lenp, loff_t *ppos);
  2263. int proc_nr_dentry(struct ctl_table *table, int write,
  2264. void __user *buffer, size_t *lenp, loff_t *ppos);
  2265. int proc_nr_inodes(struct ctl_table *table, int write,
  2266. void __user *buffer, size_t *lenp, loff_t *ppos);
  2267. int __init get_filesystem_list(char *buf);
  2268. #define __FMODE_EXEC ((__force int) FMODE_EXEC)
  2269. #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
  2270. #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
  2271. #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
  2272. (flag & __FMODE_NONOTIFY)))
  2273. static inline int is_sxid(mode_t mode)
  2274. {
  2275. return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
  2276. }
  2277. static inline void inode_has_no_xattr(struct inode *inode)
  2278. {
  2279. if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
  2280. inode->i_flags |= S_NOSEC;
  2281. }
  2282. #endif /* __KERNEL__ */
  2283. #endif /* _LINUX_FS_H */