fs.h 87 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568
  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 1024 /* Initial setting for nfile rlimits */
  23. #define BLOCK_SIZE_BITS 10
  24. #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
  25. #define SEEK_SET 0 /* seek relative to beginning of file */
  26. #define SEEK_CUR 1 /* seek relative to current file position */
  27. #define SEEK_END 2 /* seek relative to end of file */
  28. #define SEEK_MAX SEEK_END
  29. struct fstrim_range {
  30. __u64 start;
  31. __u64 len;
  32. __u64 minlen;
  33. };
  34. /* And dynamically-tunable limits and defaults: */
  35. struct files_stat_struct {
  36. unsigned long nr_files; /* read only */
  37. unsigned long nr_free_files; /* read only */
  38. unsigned long max_files; /* tunable */
  39. };
  40. struct inodes_stat_t {
  41. int nr_inodes;
  42. int nr_unused;
  43. int dummy[5]; /* padding for sysctl ABI compatibility */
  44. };
  45. #define NR_FILE 8192 /* this can well be larger on a larger system */
  46. #define MAY_EXEC 1
  47. #define MAY_WRITE 2
  48. #define MAY_READ 4
  49. #define MAY_APPEND 8
  50. #define MAY_ACCESS 16
  51. #define MAY_OPEN 32
  52. #define MAY_CHDIR 64
  53. /*
  54. * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
  55. * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
  56. */
  57. /* file is open for reading */
  58. #define FMODE_READ ((__force fmode_t)0x1)
  59. /* file is open for writing */
  60. #define FMODE_WRITE ((__force fmode_t)0x2)
  61. /* file is seekable */
  62. #define FMODE_LSEEK ((__force fmode_t)0x4)
  63. /* file can be accessed using pread */
  64. #define FMODE_PREAD ((__force fmode_t)0x8)
  65. /* file can be accessed using pwrite */
  66. #define FMODE_PWRITE ((__force fmode_t)0x10)
  67. /* File is opened for execution with sys_execve / sys_uselib */
  68. #define FMODE_EXEC ((__force fmode_t)0x20)
  69. /* File is opened with O_NDELAY (only set for block devices) */
  70. #define FMODE_NDELAY ((__force fmode_t)0x40)
  71. /* File is opened with O_EXCL (only set for block devices) */
  72. #define FMODE_EXCL ((__force fmode_t)0x80)
  73. /* File is opened using open(.., 3, ..) and is writeable only for ioctls
  74. (specialy hack for floppy.c) */
  75. #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
  76. /*
  77. * Don't update ctime and mtime.
  78. *
  79. * Currently a special hack for the XFS open_by_handle ioctl, but we'll
  80. * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
  81. */
  82. #define FMODE_NOCMTIME ((__force fmode_t)0x800)
  83. /* Expect random access pattern */
  84. #define FMODE_RANDOM ((__force fmode_t)0x1000)
  85. /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
  86. #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
  87. /* File was opened by fanotify and shouldn't generate fanotify events */
  88. #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
  89. /*
  90. * The below are the various read and write types that we support. Some of
  91. * them include behavioral modifiers that send information down to the
  92. * block layer and IO scheduler. Terminology:
  93. *
  94. * The block layer uses device plugging to defer IO a little bit, in
  95. * the hope that we will see more IO very shortly. This increases
  96. * coalescing of adjacent IO and thus reduces the number of IOs we
  97. * have to send to the device. It also allows for better queuing,
  98. * if the IO isn't mergeable. If the caller is going to be waiting
  99. * for the IO, then he must ensure that the device is unplugged so
  100. * that the IO is dispatched to the driver.
  101. *
  102. * All IO is handled async in Linux. This is fine for background
  103. * writes, but for reads or writes that someone waits for completion
  104. * on, we want to notify the block layer and IO scheduler so that they
  105. * know about it. That allows them to make better scheduling
  106. * decisions. So when the below references 'sync' and 'async', it
  107. * is referencing this priority hint.
  108. *
  109. * With that in mind, the available types are:
  110. *
  111. * READ A normal read operation. Device will be plugged.
  112. * READ_SYNC A synchronous read. Device is not plugged, caller can
  113. * immediately wait on this read without caring about
  114. * unplugging.
  115. * READA Used for read-ahead operations. Lower priority, and the
  116. * block layer could (in theory) choose to ignore this
  117. * request if it runs into resource problems.
  118. * WRITE A normal async write. Device will be plugged.
  119. * WRITE_SYNC_PLUG Synchronous write. Identical to WRITE, but passes down
  120. * the hint that someone will be waiting on this IO
  121. * shortly. The device must still be unplugged explicitly,
  122. * WRITE_SYNC_PLUG does not do this as we could be
  123. * submitting more writes before we actually wait on any
  124. * of them.
  125. * WRITE_SYNC Like WRITE_SYNC_PLUG, but also unplugs the device
  126. * immediately after submission. The write equivalent
  127. * of READ_SYNC.
  128. * WRITE_ODIRECT_PLUG Special case write for O_DIRECT only.
  129. * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
  130. * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
  131. * non-volatile media on completion.
  132. * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
  133. * by a cache flush and data is guaranteed to be on
  134. * non-volatile media on completion.
  135. *
  136. */
  137. #define RW_MASK REQ_WRITE
  138. #define RWA_MASK REQ_RAHEAD
  139. #define READ 0
  140. #define WRITE RW_MASK
  141. #define READA RWA_MASK
  142. #define READ_SYNC (READ | REQ_SYNC | REQ_UNPLUG)
  143. #define READ_META (READ | REQ_META)
  144. #define WRITE_SYNC_PLUG (WRITE | REQ_SYNC | REQ_NOIDLE)
  145. #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG)
  146. #define WRITE_ODIRECT_PLUG (WRITE | REQ_SYNC)
  147. #define WRITE_META (WRITE | REQ_META)
  148. #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  149. REQ_FLUSH)
  150. #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  151. REQ_FUA)
  152. #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  153. REQ_FLUSH | REQ_FUA)
  154. #define SEL_IN 1
  155. #define SEL_OUT 2
  156. #define SEL_EX 4
  157. /* public flags for file_system_type */
  158. #define FS_REQUIRES_DEV 1
  159. #define FS_BINARY_MOUNTDATA 2
  160. #define FS_HAS_SUBTYPE 4
  161. #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
  162. #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
  163. * during rename() internally.
  164. */
  165. /*
  166. * These are the fs-independent mount-flags: up to 32 flags are supported
  167. */
  168. #define MS_RDONLY 1 /* Mount read-only */
  169. #define MS_NOSUID 2 /* Ignore suid and sgid bits */
  170. #define MS_NODEV 4 /* Disallow access to device special files */
  171. #define MS_NOEXEC 8 /* Disallow program execution */
  172. #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
  173. #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
  174. #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
  175. #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
  176. #define MS_NOATIME 1024 /* Do not update access times. */
  177. #define MS_NODIRATIME 2048 /* Do not update directory access times */
  178. #define MS_BIND 4096
  179. #define MS_MOVE 8192
  180. #define MS_REC 16384
  181. #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
  182. MS_VERBOSE is deprecated. */
  183. #define MS_SILENT 32768
  184. #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
  185. #define MS_UNBINDABLE (1<<17) /* change to unbindable */
  186. #define MS_PRIVATE (1<<18) /* change to private */
  187. #define MS_SLAVE (1<<19) /* change to slave */
  188. #define MS_SHARED (1<<20) /* change to shared */
  189. #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
  190. #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
  191. #define MS_I_VERSION (1<<23) /* Update inode I_version field */
  192. #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
  193. #define MS_BORN (1<<29)
  194. #define MS_ACTIVE (1<<30)
  195. #define MS_NOUSER (1<<31)
  196. /*
  197. * Superblock flags that can be altered by MS_REMOUNT
  198. */
  199. #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
  200. /*
  201. * Old magic mount flag and mask
  202. */
  203. #define MS_MGC_VAL 0xC0ED0000
  204. #define MS_MGC_MSK 0xffff0000
  205. /* Inode flags - they have nothing to superblock flags now */
  206. #define S_SYNC 1 /* Writes are synced at once */
  207. #define S_NOATIME 2 /* Do not update access times */
  208. #define S_APPEND 4 /* Append-only file */
  209. #define S_IMMUTABLE 8 /* Immutable file */
  210. #define S_DEAD 16 /* removed, but still open directory */
  211. #define S_NOQUOTA 32 /* Inode is not counted to quota */
  212. #define S_DIRSYNC 64 /* Directory modifications are synchronous */
  213. #define S_NOCMTIME 128 /* Do not update file c/mtime */
  214. #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
  215. #define S_PRIVATE 512 /* Inode is fs-internal */
  216. #define S_IMA 1024 /* Inode has an associated IMA struct */
  217. #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
  218. /*
  219. * Note that nosuid etc flags are inode-specific: setting some file-system
  220. * flags just means all the inodes inherit those flags by default. It might be
  221. * possible to override it selectively if you really wanted to with some
  222. * ioctl() that is not currently implemented.
  223. *
  224. * Exception: MS_RDONLY is always applied to the entire file system.
  225. *
  226. * Unfortunately, it is possible to change a filesystems flags with it mounted
  227. * with files in use. This means that all of the inodes will not have their
  228. * i_flags updated. Hence, i_flags no longer inherit the superblock mount
  229. * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
  230. */
  231. #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
  232. #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
  233. #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
  234. ((inode)->i_flags & S_SYNC))
  235. #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
  236. ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
  237. #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
  238. #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
  239. #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
  240. #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
  241. #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
  242. #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
  243. #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
  244. #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
  245. #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
  246. #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
  247. #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
  248. #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
  249. #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
  250. /* the read-only stuff doesn't really belong here, but any other place is
  251. probably as bad and I don't want to create yet another include file. */
  252. #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
  253. #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
  254. #define BLKRRPART _IO(0x12,95) /* re-read partition table */
  255. #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
  256. #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
  257. #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
  258. #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
  259. #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
  260. #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
  261. #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
  262. #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
  263. #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
  264. #if 0
  265. #define BLKPG _IO(0x12,105)/* See blkpg.h */
  266. /* Some people are morons. Do not use sizeof! */
  267. #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
  268. #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
  269. /* This was here just to show that the number is taken -
  270. probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
  271. #endif
  272. /* A jump here: 108-111 have been used for various private purposes. */
  273. #define BLKBSZGET _IOR(0x12,112,size_t)
  274. #define BLKBSZSET _IOW(0x12,113,size_t)
  275. #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
  276. #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
  277. #define BLKTRACESTART _IO(0x12,116)
  278. #define BLKTRACESTOP _IO(0x12,117)
  279. #define BLKTRACETEARDOWN _IO(0x12,118)
  280. #define BLKDISCARD _IO(0x12,119)
  281. #define BLKIOMIN _IO(0x12,120)
  282. #define BLKIOOPT _IO(0x12,121)
  283. #define BLKALIGNOFF _IO(0x12,122)
  284. #define BLKPBSZGET _IO(0x12,123)
  285. #define BLKDISCARDZEROES _IO(0x12,124)
  286. #define BLKSECDISCARD _IO(0x12,125)
  287. #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
  288. #define FIBMAP _IO(0x00,1) /* bmap access */
  289. #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
  290. #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
  291. #define FITHAW _IOWR('X', 120, int) /* Thaw */
  292. #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
  293. #define FS_IOC_GETFLAGS _IOR('f', 1, long)
  294. #define FS_IOC_SETFLAGS _IOW('f', 2, long)
  295. #define FS_IOC_GETVERSION _IOR('v', 1, long)
  296. #define FS_IOC_SETVERSION _IOW('v', 2, long)
  297. #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
  298. #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
  299. #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
  300. #define FS_IOC32_GETVERSION _IOR('v', 1, int)
  301. #define FS_IOC32_SETVERSION _IOW('v', 2, int)
  302. /*
  303. * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
  304. */
  305. #define FS_SECRM_FL 0x00000001 /* Secure deletion */
  306. #define FS_UNRM_FL 0x00000002 /* Undelete */
  307. #define FS_COMPR_FL 0x00000004 /* Compress file */
  308. #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
  309. #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
  310. #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
  311. #define FS_NODUMP_FL 0x00000040 /* do not dump file */
  312. #define FS_NOATIME_FL 0x00000080 /* do not update atime */
  313. /* Reserved for compression usage... */
  314. #define FS_DIRTY_FL 0x00000100
  315. #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
  316. #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
  317. #define FS_ECOMPR_FL 0x00000800 /* Compression error */
  318. /* End compression flags --- maybe not all used */
  319. #define FS_BTREE_FL 0x00001000 /* btree format dir */
  320. #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
  321. #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
  322. #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
  323. #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
  324. #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
  325. #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
  326. #define FS_EXTENT_FL 0x00080000 /* Extents */
  327. #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
  328. #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
  329. #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
  330. #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
  331. #define SYNC_FILE_RANGE_WAIT_BEFORE 1
  332. #define SYNC_FILE_RANGE_WRITE 2
  333. #define SYNC_FILE_RANGE_WAIT_AFTER 4
  334. #ifdef __KERNEL__
  335. #include <linux/linkage.h>
  336. #include <linux/wait.h>
  337. #include <linux/types.h>
  338. #include <linux/kdev_t.h>
  339. #include <linux/dcache.h>
  340. #include <linux/path.h>
  341. #include <linux/stat.h>
  342. #include <linux/cache.h>
  343. #include <linux/list.h>
  344. #include <linux/radix-tree.h>
  345. #include <linux/prio_tree.h>
  346. #include <linux/init.h>
  347. #include <linux/pid.h>
  348. #include <linux/mutex.h>
  349. #include <linux/capability.h>
  350. #include <linux/semaphore.h>
  351. #include <linux/fiemap.h>
  352. #include <linux/rculist_bl.h>
  353. #include <asm/atomic.h>
  354. #include <asm/byteorder.h>
  355. struct export_operations;
  356. struct hd_geometry;
  357. struct iovec;
  358. struct nameidata;
  359. struct kiocb;
  360. struct kobject;
  361. struct pipe_inode_info;
  362. struct poll_table_struct;
  363. struct kstatfs;
  364. struct vm_area_struct;
  365. struct vfsmount;
  366. struct cred;
  367. extern void __init inode_init(void);
  368. extern void __init inode_init_early(void);
  369. extern void __init files_init(unsigned long);
  370. extern struct files_stat_struct files_stat;
  371. extern unsigned long get_max_files(void);
  372. extern int sysctl_nr_open;
  373. extern struct inodes_stat_t inodes_stat;
  374. extern int leases_enable, lease_break_time;
  375. struct buffer_head;
  376. typedef int (get_block_t)(struct inode *inode, sector_t iblock,
  377. struct buffer_head *bh_result, int create);
  378. typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
  379. ssize_t bytes, void *private, int ret,
  380. bool is_async);
  381. /*
  382. * Attribute flags. These should be or-ed together to figure out what
  383. * has been changed!
  384. */
  385. #define ATTR_MODE (1 << 0)
  386. #define ATTR_UID (1 << 1)
  387. #define ATTR_GID (1 << 2)
  388. #define ATTR_SIZE (1 << 3)
  389. #define ATTR_ATIME (1 << 4)
  390. #define ATTR_MTIME (1 << 5)
  391. #define ATTR_CTIME (1 << 6)
  392. #define ATTR_ATIME_SET (1 << 7)
  393. #define ATTR_MTIME_SET (1 << 8)
  394. #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
  395. #define ATTR_ATTR_FLAG (1 << 10)
  396. #define ATTR_KILL_SUID (1 << 11)
  397. #define ATTR_KILL_SGID (1 << 12)
  398. #define ATTR_FILE (1 << 13)
  399. #define ATTR_KILL_PRIV (1 << 14)
  400. #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
  401. #define ATTR_TIMES_SET (1 << 16)
  402. /*
  403. * This is the Inode Attributes structure, used for notify_change(). It
  404. * uses the above definitions as flags, to know which values have changed.
  405. * Also, in this manner, a Filesystem can look at only the values it cares
  406. * about. Basically, these are the attributes that the VFS layer can
  407. * request to change from the FS layer.
  408. *
  409. * Derek Atkins <warlord@MIT.EDU> 94-10-20
  410. */
  411. struct iattr {
  412. unsigned int ia_valid;
  413. umode_t ia_mode;
  414. uid_t ia_uid;
  415. gid_t ia_gid;
  416. loff_t ia_size;
  417. struct timespec ia_atime;
  418. struct timespec ia_mtime;
  419. struct timespec ia_ctime;
  420. /*
  421. * Not an attribute, but an auxilary info for filesystems wanting to
  422. * implement an ftruncate() like method. NOTE: filesystem should
  423. * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
  424. */
  425. struct file *ia_file;
  426. };
  427. /*
  428. * Includes for diskquotas.
  429. */
  430. #include <linux/quota.h>
  431. /**
  432. * enum positive_aop_returns - aop return codes with specific semantics
  433. *
  434. * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
  435. * completed, that the page is still locked, and
  436. * should be considered active. The VM uses this hint
  437. * to return the page to the active list -- it won't
  438. * be a candidate for writeback again in the near
  439. * future. Other callers must be careful to unlock
  440. * the page if they get this return. Returned by
  441. * writepage();
  442. *
  443. * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
  444. * unlocked it and the page might have been truncated.
  445. * The caller should back up to acquiring a new page and
  446. * trying again. The aop will be taking reasonable
  447. * precautions not to livelock. If the caller held a page
  448. * reference, it should drop it before retrying. Returned
  449. * by readpage().
  450. *
  451. * address_space_operation functions return these large constants to indicate
  452. * special semantics to the caller. These are much larger than the bytes in a
  453. * page to allow for functions that return the number of bytes operated on in a
  454. * given page.
  455. */
  456. enum positive_aop_returns {
  457. AOP_WRITEPAGE_ACTIVATE = 0x80000,
  458. AOP_TRUNCATED_PAGE = 0x80001,
  459. };
  460. #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
  461. #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
  462. #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
  463. * helper code (eg buffer layer)
  464. * to clear GFP_FS from alloc */
  465. /*
  466. * oh the beauties of C type declarations.
  467. */
  468. struct page;
  469. struct address_space;
  470. struct writeback_control;
  471. struct iov_iter {
  472. const struct iovec *iov;
  473. unsigned long nr_segs;
  474. size_t iov_offset;
  475. size_t count;
  476. };
  477. size_t iov_iter_copy_from_user_atomic(struct page *page,
  478. struct iov_iter *i, unsigned long offset, size_t bytes);
  479. size_t iov_iter_copy_from_user(struct page *page,
  480. struct iov_iter *i, unsigned long offset, size_t bytes);
  481. void iov_iter_advance(struct iov_iter *i, size_t bytes);
  482. int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
  483. size_t iov_iter_single_seg_count(struct iov_iter *i);
  484. static inline void iov_iter_init(struct iov_iter *i,
  485. const struct iovec *iov, unsigned long nr_segs,
  486. size_t count, size_t written)
  487. {
  488. i->iov = iov;
  489. i->nr_segs = nr_segs;
  490. i->iov_offset = 0;
  491. i->count = count + written;
  492. iov_iter_advance(i, written);
  493. }
  494. static inline size_t iov_iter_count(struct iov_iter *i)
  495. {
  496. return i->count;
  497. }
  498. /*
  499. * "descriptor" for what we're up to with a read.
  500. * This allows us to use the same read code yet
  501. * have multiple different users of the data that
  502. * we read from a file.
  503. *
  504. * The simplest case just copies the data to user
  505. * mode.
  506. */
  507. typedef struct {
  508. size_t written;
  509. size_t count;
  510. union {
  511. char __user *buf;
  512. void *data;
  513. } arg;
  514. int error;
  515. } read_descriptor_t;
  516. typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
  517. unsigned long, unsigned long);
  518. struct address_space_operations {
  519. int (*writepage)(struct page *page, struct writeback_control *wbc);
  520. int (*readpage)(struct file *, struct page *);
  521. void (*sync_page)(struct page *);
  522. /* Write back some dirty pages from this mapping. */
  523. int (*writepages)(struct address_space *, struct writeback_control *);
  524. /* Set a page dirty. Return true if this dirtied it */
  525. int (*set_page_dirty)(struct page *page);
  526. int (*readpages)(struct file *filp, struct address_space *mapping,
  527. struct list_head *pages, unsigned nr_pages);
  528. int (*write_begin)(struct file *, struct address_space *mapping,
  529. loff_t pos, unsigned len, unsigned flags,
  530. struct page **pagep, void **fsdata);
  531. int (*write_end)(struct file *, struct address_space *mapping,
  532. loff_t pos, unsigned len, unsigned copied,
  533. struct page *page, void *fsdata);
  534. /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
  535. sector_t (*bmap)(struct address_space *, sector_t);
  536. void (*invalidatepage) (struct page *, unsigned long);
  537. int (*releasepage) (struct page *, gfp_t);
  538. void (*freepage)(struct page *);
  539. ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
  540. loff_t offset, unsigned long nr_segs);
  541. int (*get_xip_mem)(struct address_space *, pgoff_t, int,
  542. void **, unsigned long *);
  543. /* migrate the contents of a page to the specified target */
  544. int (*migratepage) (struct address_space *,
  545. struct page *, struct page *);
  546. int (*launder_page) (struct page *);
  547. int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
  548. unsigned long);
  549. int (*error_remove_page)(struct address_space *, struct page *);
  550. };
  551. /*
  552. * pagecache_write_begin/pagecache_write_end must be used by general code
  553. * to write into the pagecache.
  554. */
  555. int pagecache_write_begin(struct file *, struct address_space *mapping,
  556. loff_t pos, unsigned len, unsigned flags,
  557. struct page **pagep, void **fsdata);
  558. int pagecache_write_end(struct file *, struct address_space *mapping,
  559. loff_t pos, unsigned len, unsigned copied,
  560. struct page *page, void *fsdata);
  561. struct backing_dev_info;
  562. struct address_space {
  563. struct inode *host; /* owner: inode, block_device */
  564. struct radix_tree_root page_tree; /* radix tree of all pages */
  565. spinlock_t tree_lock; /* and lock protecting it */
  566. unsigned int i_mmap_writable;/* count VM_SHARED mappings */
  567. struct prio_tree_root i_mmap; /* tree of private and shared mappings */
  568. struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
  569. spinlock_t i_mmap_lock; /* protect tree, count, list */
  570. unsigned int truncate_count; /* Cover race condition with truncate */
  571. unsigned long nrpages; /* number of total pages */
  572. pgoff_t writeback_index;/* writeback starts here */
  573. const struct address_space_operations *a_ops; /* methods */
  574. unsigned long flags; /* error bits/gfp mask */
  575. struct backing_dev_info *backing_dev_info; /* device readahead, etc */
  576. spinlock_t private_lock; /* for use by the address_space */
  577. struct list_head private_list; /* ditto */
  578. struct address_space *assoc_mapping; /* ditto */
  579. struct mutex unmap_mutex; /* to protect unmapping */
  580. } __attribute__((aligned(sizeof(long))));
  581. /*
  582. * On most architectures that alignment is already the case; but
  583. * must be enforced here for CRIS, to let the least signficant bit
  584. * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
  585. */
  586. struct block_device {
  587. dev_t bd_dev; /* not a kdev_t - it's a search key */
  588. struct inode * bd_inode; /* will die */
  589. struct super_block * bd_super;
  590. int bd_openers;
  591. struct mutex bd_mutex; /* open/close mutex */
  592. struct list_head bd_inodes;
  593. void * bd_claiming;
  594. void * bd_holder;
  595. int bd_holders;
  596. bool bd_write_holder;
  597. #ifdef CONFIG_SYSFS
  598. struct list_head bd_holder_disks;
  599. #endif
  600. struct block_device * bd_contains;
  601. unsigned bd_block_size;
  602. struct hd_struct * bd_part;
  603. /* number of times partitions within this device have been opened. */
  604. unsigned bd_part_count;
  605. int bd_invalidated;
  606. struct gendisk * bd_disk;
  607. struct list_head bd_list;
  608. /*
  609. * Private data. You must have bd_claim'ed the block_device
  610. * to use this. NOTE: bd_claim allows an owner to claim
  611. * the same device multiple times, the owner must take special
  612. * care to not mess up bd_private for that case.
  613. */
  614. unsigned long bd_private;
  615. /* The counter of freeze processes */
  616. int bd_fsfreeze_count;
  617. /* Mutex for freeze */
  618. struct mutex bd_fsfreeze_mutex;
  619. };
  620. /*
  621. * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
  622. * radix trees
  623. */
  624. #define PAGECACHE_TAG_DIRTY 0
  625. #define PAGECACHE_TAG_WRITEBACK 1
  626. #define PAGECACHE_TAG_TOWRITE 2
  627. int mapping_tagged(struct address_space *mapping, int tag);
  628. /*
  629. * Might pages of this file be mapped into userspace?
  630. */
  631. static inline int mapping_mapped(struct address_space *mapping)
  632. {
  633. return !prio_tree_empty(&mapping->i_mmap) ||
  634. !list_empty(&mapping->i_mmap_nonlinear);
  635. }
  636. /*
  637. * Might pages of this file have been modified in userspace?
  638. * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
  639. * marks vma as VM_SHARED if it is shared, and the file was opened for
  640. * writing i.e. vma may be mprotected writable even if now readonly.
  641. */
  642. static inline int mapping_writably_mapped(struct address_space *mapping)
  643. {
  644. return mapping->i_mmap_writable != 0;
  645. }
  646. /*
  647. * Use sequence counter to get consistent i_size on 32-bit processors.
  648. */
  649. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  650. #include <linux/seqlock.h>
  651. #define __NEED_I_SIZE_ORDERED
  652. #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
  653. #else
  654. #define i_size_ordered_init(inode) do { } while (0)
  655. #endif
  656. struct posix_acl;
  657. #define ACL_NOT_CACHED ((void *)(-1))
  658. struct inode {
  659. /* RCU path lookup touches following: */
  660. umode_t i_mode;
  661. uid_t i_uid;
  662. gid_t i_gid;
  663. const struct inode_operations *i_op;
  664. struct super_block *i_sb;
  665. spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
  666. unsigned int i_flags;
  667. struct mutex i_mutex;
  668. unsigned long i_state;
  669. unsigned long dirtied_when; /* jiffies of first dirtying */
  670. struct hlist_node i_hash;
  671. struct list_head i_wb_list; /* backing dev IO list */
  672. struct list_head i_lru; /* inode LRU list */
  673. struct list_head i_sb_list;
  674. union {
  675. struct list_head i_dentry;
  676. struct rcu_head i_rcu;
  677. };
  678. unsigned long i_ino;
  679. atomic_t i_count;
  680. unsigned int i_nlink;
  681. dev_t i_rdev;
  682. unsigned int i_blkbits;
  683. u64 i_version;
  684. loff_t i_size;
  685. #ifdef __NEED_I_SIZE_ORDERED
  686. seqcount_t i_size_seqcount;
  687. #endif
  688. struct timespec i_atime;
  689. struct timespec i_mtime;
  690. struct timespec i_ctime;
  691. blkcnt_t i_blocks;
  692. unsigned short i_bytes;
  693. struct rw_semaphore i_alloc_sem;
  694. const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
  695. struct file_lock *i_flock;
  696. struct address_space *i_mapping;
  697. struct address_space i_data;
  698. #ifdef CONFIG_QUOTA
  699. struct dquot *i_dquot[MAXQUOTAS];
  700. #endif
  701. struct list_head i_devices;
  702. union {
  703. struct pipe_inode_info *i_pipe;
  704. struct block_device *i_bdev;
  705. struct cdev *i_cdev;
  706. };
  707. __u32 i_generation;
  708. #ifdef CONFIG_FSNOTIFY
  709. __u32 i_fsnotify_mask; /* all events this inode cares about */
  710. struct hlist_head i_fsnotify_marks;
  711. #endif
  712. #ifdef CONFIG_IMA
  713. /* protected by i_lock */
  714. unsigned int i_readcount; /* struct files open RO */
  715. #endif
  716. atomic_t i_writecount;
  717. #ifdef CONFIG_SECURITY
  718. void *i_security;
  719. #endif
  720. #ifdef CONFIG_FS_POSIX_ACL
  721. struct posix_acl *i_acl;
  722. struct posix_acl *i_default_acl;
  723. #endif
  724. void *i_private; /* fs or device private pointer */
  725. };
  726. static inline int inode_unhashed(struct inode *inode)
  727. {
  728. return hlist_unhashed(&inode->i_hash);
  729. }
  730. /*
  731. * inode->i_mutex nesting subclasses for the lock validator:
  732. *
  733. * 0: the object of the current VFS operation
  734. * 1: parent
  735. * 2: child/target
  736. * 3: quota file
  737. *
  738. * The locking order between these classes is
  739. * parent -> child -> normal -> xattr -> quota
  740. */
  741. enum inode_i_mutex_lock_class
  742. {
  743. I_MUTEX_NORMAL,
  744. I_MUTEX_PARENT,
  745. I_MUTEX_CHILD,
  746. I_MUTEX_XATTR,
  747. I_MUTEX_QUOTA
  748. };
  749. /*
  750. * NOTE: in a 32bit arch with a preemptable kernel and
  751. * an UP compile the i_size_read/write must be atomic
  752. * with respect to the local cpu (unlike with preempt disabled),
  753. * but they don't need to be atomic with respect to other cpus like in
  754. * true SMP (so they need either to either locally disable irq around
  755. * the read or for example on x86 they can be still implemented as a
  756. * cmpxchg8b without the need of the lock prefix). For SMP compiles
  757. * and 64bit archs it makes no difference if preempt is enabled or not.
  758. */
  759. static inline loff_t i_size_read(const struct inode *inode)
  760. {
  761. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  762. loff_t i_size;
  763. unsigned int seq;
  764. do {
  765. seq = read_seqcount_begin(&inode->i_size_seqcount);
  766. i_size = inode->i_size;
  767. } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
  768. return i_size;
  769. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  770. loff_t i_size;
  771. preempt_disable();
  772. i_size = inode->i_size;
  773. preempt_enable();
  774. return i_size;
  775. #else
  776. return inode->i_size;
  777. #endif
  778. }
  779. /*
  780. * NOTE: unlike i_size_read(), i_size_write() does need locking around it
  781. * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
  782. * can be lost, resulting in subsequent i_size_read() calls spinning forever.
  783. */
  784. static inline void i_size_write(struct inode *inode, loff_t i_size)
  785. {
  786. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  787. write_seqcount_begin(&inode->i_size_seqcount);
  788. inode->i_size = i_size;
  789. write_seqcount_end(&inode->i_size_seqcount);
  790. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  791. preempt_disable();
  792. inode->i_size = i_size;
  793. preempt_enable();
  794. #else
  795. inode->i_size = i_size;
  796. #endif
  797. }
  798. static inline unsigned iminor(const struct inode *inode)
  799. {
  800. return MINOR(inode->i_rdev);
  801. }
  802. static inline unsigned imajor(const struct inode *inode)
  803. {
  804. return MAJOR(inode->i_rdev);
  805. }
  806. extern struct block_device *I_BDEV(struct inode *inode);
  807. struct fown_struct {
  808. rwlock_t lock; /* protects pid, uid, euid fields */
  809. struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
  810. enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
  811. uid_t uid, euid; /* uid/euid of process setting the owner */
  812. int signum; /* posix.1b rt signal to be delivered on IO */
  813. };
  814. /*
  815. * Track a single file's readahead state
  816. */
  817. struct file_ra_state {
  818. pgoff_t start; /* where readahead started */
  819. unsigned int size; /* # of readahead pages */
  820. unsigned int async_size; /* do asynchronous readahead when
  821. there are only # of pages ahead */
  822. unsigned int ra_pages; /* Maximum readahead window */
  823. unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
  824. loff_t prev_pos; /* Cache last read() position */
  825. };
  826. /*
  827. * Check if @index falls in the readahead windows.
  828. */
  829. static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
  830. {
  831. return (index >= ra->start &&
  832. index < ra->start + ra->size);
  833. }
  834. #define FILE_MNT_WRITE_TAKEN 1
  835. #define FILE_MNT_WRITE_RELEASED 2
  836. struct file {
  837. /*
  838. * fu_list becomes invalid after file_free is called and queued via
  839. * fu_rcuhead for RCU freeing
  840. */
  841. union {
  842. struct list_head fu_list;
  843. struct rcu_head fu_rcuhead;
  844. } f_u;
  845. struct path f_path;
  846. #define f_dentry f_path.dentry
  847. #define f_vfsmnt f_path.mnt
  848. const struct file_operations *f_op;
  849. spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
  850. #ifdef CONFIG_SMP
  851. int f_sb_list_cpu;
  852. #endif
  853. atomic_long_t f_count;
  854. unsigned int f_flags;
  855. fmode_t f_mode;
  856. loff_t f_pos;
  857. struct fown_struct f_owner;
  858. const struct cred *f_cred;
  859. struct file_ra_state f_ra;
  860. u64 f_version;
  861. #ifdef CONFIG_SECURITY
  862. void *f_security;
  863. #endif
  864. /* needed for tty driver, and maybe others */
  865. void *private_data;
  866. #ifdef CONFIG_EPOLL
  867. /* Used by fs/eventpoll.c to link all the hooks to this file */
  868. struct list_head f_ep_links;
  869. #endif /* #ifdef CONFIG_EPOLL */
  870. struct address_space *f_mapping;
  871. #ifdef CONFIG_DEBUG_WRITECOUNT
  872. unsigned long f_mnt_write_state;
  873. #endif
  874. };
  875. #define get_file(x) atomic_long_inc(&(x)->f_count)
  876. #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
  877. #define file_count(x) atomic_long_read(&(x)->f_count)
  878. #ifdef CONFIG_DEBUG_WRITECOUNT
  879. static inline void file_take_write(struct file *f)
  880. {
  881. WARN_ON(f->f_mnt_write_state != 0);
  882. f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
  883. }
  884. static inline void file_release_write(struct file *f)
  885. {
  886. f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
  887. }
  888. static inline void file_reset_write(struct file *f)
  889. {
  890. f->f_mnt_write_state = 0;
  891. }
  892. static inline void file_check_state(struct file *f)
  893. {
  894. /*
  895. * At this point, either both or neither of these bits
  896. * should be set.
  897. */
  898. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
  899. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
  900. }
  901. static inline int file_check_writeable(struct file *f)
  902. {
  903. if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
  904. return 0;
  905. printk(KERN_WARNING "writeable file with no "
  906. "mnt_want_write()\n");
  907. WARN_ON(1);
  908. return -EINVAL;
  909. }
  910. #else /* !CONFIG_DEBUG_WRITECOUNT */
  911. static inline void file_take_write(struct file *filp) {}
  912. static inline void file_release_write(struct file *filp) {}
  913. static inline void file_reset_write(struct file *filp) {}
  914. static inline void file_check_state(struct file *filp) {}
  915. static inline int file_check_writeable(struct file *filp)
  916. {
  917. return 0;
  918. }
  919. #endif /* CONFIG_DEBUG_WRITECOUNT */
  920. #define MAX_NON_LFS ((1UL<<31) - 1)
  921. /* Page cache limit. The filesystems should put that into their s_maxbytes
  922. limits, otherwise bad things can happen in VM. */
  923. #if BITS_PER_LONG==32
  924. #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
  925. #elif BITS_PER_LONG==64
  926. #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
  927. #endif
  928. #define FL_POSIX 1
  929. #define FL_FLOCK 2
  930. #define FL_ACCESS 8 /* not trying to lock, just looking */
  931. #define FL_EXISTS 16 /* when unlocking, test for existence */
  932. #define FL_LEASE 32 /* lease held on this file */
  933. #define FL_CLOSE 64 /* unlock on close */
  934. #define FL_SLEEP 128 /* A blocking lock */
  935. /*
  936. * Special return value from posix_lock_file() and vfs_lock_file() for
  937. * asynchronous locking.
  938. */
  939. #define FILE_LOCK_DEFERRED 1
  940. /*
  941. * The POSIX file lock owner is determined by
  942. * the "struct files_struct" in the thread group
  943. * (or NULL for no owner - BSD locks).
  944. *
  945. * Lockd stuffs a "host" pointer into this.
  946. */
  947. typedef struct files_struct *fl_owner_t;
  948. struct file_lock_operations {
  949. void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
  950. void (*fl_release_private)(struct file_lock *);
  951. };
  952. struct lock_manager_operations {
  953. int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
  954. void (*fl_notify)(struct file_lock *); /* unblock callback */
  955. int (*fl_grant)(struct file_lock *, struct file_lock *, int);
  956. void (*fl_release_private)(struct file_lock *);
  957. void (*fl_break)(struct file_lock *);
  958. int (*fl_change)(struct file_lock **, int);
  959. };
  960. struct lock_manager {
  961. struct list_head list;
  962. };
  963. void locks_start_grace(struct lock_manager *);
  964. void locks_end_grace(struct lock_manager *);
  965. int locks_in_grace(void);
  966. /* that will die - we need it for nfs_lock_info */
  967. #include <linux/nfs_fs_i.h>
  968. struct file_lock {
  969. struct file_lock *fl_next; /* singly linked list for this inode */
  970. struct list_head fl_link; /* doubly linked list of all locks */
  971. struct list_head fl_block; /* circular list of blocked processes */
  972. fl_owner_t fl_owner;
  973. unsigned char fl_flags;
  974. unsigned char fl_type;
  975. unsigned int fl_pid;
  976. struct pid *fl_nspid;
  977. wait_queue_head_t fl_wait;
  978. struct file *fl_file;
  979. loff_t fl_start;
  980. loff_t fl_end;
  981. struct fasync_struct * fl_fasync; /* for lease break notifications */
  982. unsigned long fl_break_time; /* for nonblocking lease breaks */
  983. const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
  984. const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
  985. union {
  986. struct nfs_lock_info nfs_fl;
  987. struct nfs4_lock_info nfs4_fl;
  988. struct {
  989. struct list_head link; /* link in AFS vnode's pending_locks list */
  990. int state; /* state of grant or error if -ve */
  991. } afs;
  992. } fl_u;
  993. };
  994. /* The following constant reflects the upper bound of the file/locking space */
  995. #ifndef OFFSET_MAX
  996. #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
  997. #define OFFSET_MAX INT_LIMIT(loff_t)
  998. #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
  999. #endif
  1000. #include <linux/fcntl.h>
  1001. extern void send_sigio(struct fown_struct *fown, int fd, int band);
  1002. #ifdef CONFIG_FILE_LOCKING
  1003. extern int fcntl_getlk(struct file *, struct flock __user *);
  1004. extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
  1005. struct flock __user *);
  1006. #if BITS_PER_LONG == 32
  1007. extern int fcntl_getlk64(struct file *, struct flock64 __user *);
  1008. extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
  1009. struct flock64 __user *);
  1010. #endif
  1011. extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
  1012. extern int fcntl_getlease(struct file *filp);
  1013. /* fs/locks.c */
  1014. void locks_free_lock(struct file_lock *fl);
  1015. extern void locks_init_lock(struct file_lock *);
  1016. extern struct file_lock * locks_alloc_lock(void);
  1017. extern void locks_copy_lock(struct file_lock *, struct file_lock *);
  1018. extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
  1019. extern void locks_remove_posix(struct file *, fl_owner_t);
  1020. extern void locks_remove_flock(struct file *);
  1021. extern void locks_release_private(struct file_lock *);
  1022. extern void posix_test_lock(struct file *, struct file_lock *);
  1023. extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
  1024. extern int posix_lock_file_wait(struct file *, struct file_lock *);
  1025. extern int posix_unblock_lock(struct file *, struct file_lock *);
  1026. extern int vfs_test_lock(struct file *, struct file_lock *);
  1027. extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
  1028. extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
  1029. extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
  1030. extern int __break_lease(struct inode *inode, unsigned int flags);
  1031. extern void lease_get_mtime(struct inode *, struct timespec *time);
  1032. extern int generic_setlease(struct file *, long, struct file_lock **);
  1033. extern int vfs_setlease(struct file *, long, struct file_lock **);
  1034. extern int lease_modify(struct file_lock **, int);
  1035. extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
  1036. extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
  1037. extern void lock_flocks(void);
  1038. extern void unlock_flocks(void);
  1039. #else /* !CONFIG_FILE_LOCKING */
  1040. static inline int fcntl_getlk(struct file *file, struct flock __user *user)
  1041. {
  1042. return -EINVAL;
  1043. }
  1044. static inline int fcntl_setlk(unsigned int fd, struct file *file,
  1045. unsigned int cmd, struct flock __user *user)
  1046. {
  1047. return -EACCES;
  1048. }
  1049. #if BITS_PER_LONG == 32
  1050. static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
  1051. {
  1052. return -EINVAL;
  1053. }
  1054. static inline int fcntl_setlk64(unsigned int fd, struct file *file,
  1055. unsigned int cmd, struct flock64 __user *user)
  1056. {
  1057. return -EACCES;
  1058. }
  1059. #endif
  1060. static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
  1061. {
  1062. return 0;
  1063. }
  1064. static inline int fcntl_getlease(struct file *filp)
  1065. {
  1066. return 0;
  1067. }
  1068. static inline void locks_init_lock(struct file_lock *fl)
  1069. {
  1070. return;
  1071. }
  1072. static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1073. {
  1074. return;
  1075. }
  1076. static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1077. {
  1078. return;
  1079. }
  1080. static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
  1081. {
  1082. return;
  1083. }
  1084. static inline void locks_remove_flock(struct file *filp)
  1085. {
  1086. return;
  1087. }
  1088. static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
  1089. {
  1090. return;
  1091. }
  1092. static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
  1093. struct file_lock *conflock)
  1094. {
  1095. return -ENOLCK;
  1096. }
  1097. static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
  1098. {
  1099. return -ENOLCK;
  1100. }
  1101. static inline int posix_unblock_lock(struct file *filp,
  1102. struct file_lock *waiter)
  1103. {
  1104. return -ENOENT;
  1105. }
  1106. static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
  1107. {
  1108. return 0;
  1109. }
  1110. static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
  1111. struct file_lock *fl, struct file_lock *conf)
  1112. {
  1113. return -ENOLCK;
  1114. }
  1115. static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
  1116. {
  1117. return 0;
  1118. }
  1119. static inline int flock_lock_file_wait(struct file *filp,
  1120. struct file_lock *request)
  1121. {
  1122. return -ENOLCK;
  1123. }
  1124. static inline int __break_lease(struct inode *inode, unsigned int mode)
  1125. {
  1126. return 0;
  1127. }
  1128. static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
  1129. {
  1130. return;
  1131. }
  1132. static inline int generic_setlease(struct file *filp, long arg,
  1133. struct file_lock **flp)
  1134. {
  1135. return -EINVAL;
  1136. }
  1137. static inline int vfs_setlease(struct file *filp, long arg,
  1138. struct file_lock **lease)
  1139. {
  1140. return -EINVAL;
  1141. }
  1142. static inline int lease_modify(struct file_lock **before, int arg)
  1143. {
  1144. return -EINVAL;
  1145. }
  1146. static inline int lock_may_read(struct inode *inode, loff_t start,
  1147. unsigned long len)
  1148. {
  1149. return 1;
  1150. }
  1151. static inline int lock_may_write(struct inode *inode, loff_t start,
  1152. unsigned long len)
  1153. {
  1154. return 1;
  1155. }
  1156. static inline void lock_flocks(void)
  1157. {
  1158. }
  1159. static inline void unlock_flocks(void)
  1160. {
  1161. }
  1162. #endif /* !CONFIG_FILE_LOCKING */
  1163. struct fasync_struct {
  1164. spinlock_t fa_lock;
  1165. int magic;
  1166. int fa_fd;
  1167. struct fasync_struct *fa_next; /* singly linked list */
  1168. struct file *fa_file;
  1169. struct rcu_head fa_rcu;
  1170. };
  1171. #define FASYNC_MAGIC 0x4601
  1172. /* SMP safe fasync helpers: */
  1173. extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
  1174. extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
  1175. extern int fasync_remove_entry(struct file *, struct fasync_struct **);
  1176. extern struct fasync_struct *fasync_alloc(void);
  1177. extern void fasync_free(struct fasync_struct *);
  1178. /* can be called from interrupts */
  1179. extern void kill_fasync(struct fasync_struct **, int, int);
  1180. extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
  1181. extern int f_setown(struct file *filp, unsigned long arg, int force);
  1182. extern void f_delown(struct file *filp);
  1183. extern pid_t f_getown(struct file *filp);
  1184. extern int send_sigurg(struct fown_struct *fown);
  1185. /*
  1186. * Umount options
  1187. */
  1188. #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
  1189. #define MNT_DETACH 0x00000002 /* Just detach from the tree */
  1190. #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
  1191. #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
  1192. #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
  1193. extern struct list_head super_blocks;
  1194. extern spinlock_t sb_lock;
  1195. struct super_block {
  1196. struct list_head s_list; /* Keep this first */
  1197. dev_t s_dev; /* search index; _not_ kdev_t */
  1198. unsigned char s_dirt;
  1199. unsigned char s_blocksize_bits;
  1200. unsigned long s_blocksize;
  1201. loff_t s_maxbytes; /* Max file size */
  1202. struct file_system_type *s_type;
  1203. const struct super_operations *s_op;
  1204. const struct dquot_operations *dq_op;
  1205. const struct quotactl_ops *s_qcop;
  1206. const struct export_operations *s_export_op;
  1207. unsigned long s_flags;
  1208. unsigned long s_magic;
  1209. struct dentry *s_root;
  1210. struct rw_semaphore s_umount;
  1211. struct mutex s_lock;
  1212. int s_count;
  1213. atomic_t s_active;
  1214. #ifdef CONFIG_SECURITY
  1215. void *s_security;
  1216. #endif
  1217. const struct xattr_handler **s_xattr;
  1218. struct list_head s_inodes; /* all inodes */
  1219. struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
  1220. #ifdef CONFIG_SMP
  1221. struct list_head __percpu *s_files;
  1222. #else
  1223. struct list_head s_files;
  1224. #endif
  1225. /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
  1226. struct list_head s_dentry_lru; /* unused dentry lru */
  1227. int s_nr_dentry_unused; /* # of dentry on lru */
  1228. struct block_device *s_bdev;
  1229. struct backing_dev_info *s_bdi;
  1230. struct mtd_info *s_mtd;
  1231. struct list_head s_instances;
  1232. struct quota_info s_dquot; /* Diskquota specific options */
  1233. int s_frozen;
  1234. wait_queue_head_t s_wait_unfrozen;
  1235. char s_id[32]; /* Informational name */
  1236. void *s_fs_info; /* Filesystem private info */
  1237. fmode_t s_mode;
  1238. /* Granularity of c/m/atime in ns.
  1239. Cannot be worse than a second */
  1240. u32 s_time_gran;
  1241. /*
  1242. * The next field is for VFS *only*. No filesystems have any business
  1243. * even looking at it. You had been warned.
  1244. */
  1245. struct mutex s_vfs_rename_mutex; /* Kludge */
  1246. /*
  1247. * Filesystem subtype. If non-empty the filesystem type field
  1248. * in /proc/mounts will be "type.subtype"
  1249. */
  1250. char *s_subtype;
  1251. /*
  1252. * Saved mount options for lazy filesystems using
  1253. * generic_show_options()
  1254. */
  1255. char __rcu *s_options;
  1256. const struct dentry_operations *s_d_op; /* default d_op for dentries */
  1257. };
  1258. extern struct timespec current_fs_time(struct super_block *sb);
  1259. /*
  1260. * Snapshotting support.
  1261. */
  1262. enum {
  1263. SB_UNFROZEN = 0,
  1264. SB_FREEZE_WRITE = 1,
  1265. SB_FREEZE_TRANS = 2,
  1266. };
  1267. #define vfs_check_frozen(sb, level) \
  1268. wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
  1269. #define get_fs_excl() atomic_inc(&current->fs_excl)
  1270. #define put_fs_excl() atomic_dec(&current->fs_excl)
  1271. #define has_fs_excl() atomic_read(&current->fs_excl)
  1272. #define is_owner_or_cap(inode) \
  1273. ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
  1274. /* not quite ready to be deprecated, but... */
  1275. extern void lock_super(struct super_block *);
  1276. extern void unlock_super(struct super_block *);
  1277. /*
  1278. * VFS helper functions..
  1279. */
  1280. extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
  1281. extern int vfs_mkdir(struct inode *, struct dentry *, int);
  1282. extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
  1283. extern int vfs_symlink(struct inode *, struct dentry *, const char *);
  1284. extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
  1285. extern int vfs_rmdir(struct inode *, struct dentry *);
  1286. extern int vfs_unlink(struct inode *, struct dentry *);
  1287. extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  1288. /*
  1289. * VFS dentry helper functions.
  1290. */
  1291. extern void dentry_unhash(struct dentry *dentry);
  1292. /*
  1293. * VFS file helper functions.
  1294. */
  1295. extern int file_permission(struct file *, int);
  1296. extern void inode_init_owner(struct inode *inode, const struct inode *dir,
  1297. mode_t mode);
  1298. /*
  1299. * VFS FS_IOC_FIEMAP helper definitions.
  1300. */
  1301. struct fiemap_extent_info {
  1302. unsigned int fi_flags; /* Flags as passed from user */
  1303. unsigned int fi_extents_mapped; /* Number of mapped extents */
  1304. unsigned int fi_extents_max; /* Size of fiemap_extent array */
  1305. struct fiemap_extent __user *fi_extents_start; /* Start of
  1306. fiemap_extent array */
  1307. };
  1308. int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
  1309. u64 phys, u64 len, u32 flags);
  1310. int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
  1311. /*
  1312. * File types
  1313. *
  1314. * NOTE! These match bits 12..15 of stat.st_mode
  1315. * (ie "(i_mode >> 12) & 15").
  1316. */
  1317. #define DT_UNKNOWN 0
  1318. #define DT_FIFO 1
  1319. #define DT_CHR 2
  1320. #define DT_DIR 4
  1321. #define DT_BLK 6
  1322. #define DT_REG 8
  1323. #define DT_LNK 10
  1324. #define DT_SOCK 12
  1325. #define DT_WHT 14
  1326. /*
  1327. * This is the "filldir" function type, used by readdir() to let
  1328. * the kernel specify what kind of dirent layout it wants to have.
  1329. * This allows the kernel to read directories into kernel space or
  1330. * to have different dirent layouts depending on the binary type.
  1331. */
  1332. typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
  1333. struct block_device_operations;
  1334. /* These macros are for out of kernel modules to test that
  1335. * the kernel supports the unlocked_ioctl and compat_ioctl
  1336. * fields in struct file_operations. */
  1337. #define HAVE_COMPAT_IOCTL 1
  1338. #define HAVE_UNLOCKED_IOCTL 1
  1339. /*
  1340. * NOTE:
  1341. * all file operations except setlease can be called without
  1342. * the big kernel lock held in all filesystems.
  1343. */
  1344. struct file_operations {
  1345. struct module *owner;
  1346. loff_t (*llseek) (struct file *, loff_t, int);
  1347. ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
  1348. ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
  1349. ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1350. ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1351. int (*readdir) (struct file *, void *, filldir_t);
  1352. unsigned int (*poll) (struct file *, struct poll_table_struct *);
  1353. long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
  1354. long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
  1355. int (*mmap) (struct file *, struct vm_area_struct *);
  1356. int (*open) (struct inode *, struct file *);
  1357. int (*flush) (struct file *, fl_owner_t id);
  1358. int (*release) (struct inode *, struct file *);
  1359. int (*fsync) (struct file *, int datasync);
  1360. int (*aio_fsync) (struct kiocb *, int datasync);
  1361. int (*fasync) (int, struct file *, int);
  1362. int (*lock) (struct file *, int, struct file_lock *);
  1363. ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
  1364. unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
  1365. int (*check_flags)(int);
  1366. int (*flock) (struct file *, int, struct file_lock *);
  1367. ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
  1368. ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
  1369. int (*setlease)(struct file *, long, struct file_lock **);
  1370. long (*fallocate)(struct file *file, int mode, loff_t offset,
  1371. loff_t len);
  1372. };
  1373. #define IPERM_FLAG_RCU 0x0001
  1374. struct inode_operations {
  1375. struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
  1376. void * (*follow_link) (struct dentry *, struct nameidata *);
  1377. int (*permission) (struct inode *, int, unsigned int);
  1378. int (*check_acl)(struct inode *, int, unsigned int);
  1379. int (*readlink) (struct dentry *, char __user *,int);
  1380. void (*put_link) (struct dentry *, struct nameidata *, void *);
  1381. int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
  1382. int (*link) (struct dentry *,struct inode *,struct dentry *);
  1383. int (*unlink) (struct inode *,struct dentry *);
  1384. int (*symlink) (struct inode *,struct dentry *,const char *);
  1385. int (*mkdir) (struct inode *,struct dentry *,int);
  1386. int (*rmdir) (struct inode *,struct dentry *);
  1387. int (*mknod) (struct inode *,struct dentry *,int,dev_t);
  1388. int (*rename) (struct inode *, struct dentry *,
  1389. struct inode *, struct dentry *);
  1390. void (*truncate) (struct inode *);
  1391. int (*setattr) (struct dentry *, struct iattr *);
  1392. int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
  1393. int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
  1394. ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
  1395. ssize_t (*listxattr) (struct dentry *, char *, size_t);
  1396. int (*removexattr) (struct dentry *, const char *);
  1397. void (*truncate_range)(struct inode *, loff_t, loff_t);
  1398. int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
  1399. u64 len);
  1400. } ____cacheline_aligned;
  1401. struct seq_file;
  1402. ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
  1403. unsigned long nr_segs, unsigned long fast_segs,
  1404. struct iovec *fast_pointer,
  1405. struct iovec **ret_pointer);
  1406. extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
  1407. extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
  1408. extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
  1409. unsigned long, loff_t *);
  1410. extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
  1411. unsigned long, loff_t *);
  1412. struct super_operations {
  1413. struct inode *(*alloc_inode)(struct super_block *sb);
  1414. void (*destroy_inode)(struct inode *);
  1415. void (*dirty_inode) (struct inode *);
  1416. int (*write_inode) (struct inode *, struct writeback_control *wbc);
  1417. int (*drop_inode) (struct inode *);
  1418. void (*evict_inode) (struct inode *);
  1419. void (*put_super) (struct super_block *);
  1420. void (*write_super) (struct super_block *);
  1421. int (*sync_fs)(struct super_block *sb, int wait);
  1422. int (*freeze_fs) (struct super_block *);
  1423. int (*unfreeze_fs) (struct super_block *);
  1424. int (*statfs) (struct dentry *, struct kstatfs *);
  1425. int (*remount_fs) (struct super_block *, int *, char *);
  1426. void (*umount_begin) (struct super_block *);
  1427. int (*show_options)(struct seq_file *, struct vfsmount *);
  1428. int (*show_stats)(struct seq_file *, struct vfsmount *);
  1429. #ifdef CONFIG_QUOTA
  1430. ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
  1431. ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
  1432. #endif
  1433. int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
  1434. };
  1435. /*
  1436. * Inode state bits. Protected by inode_lock.
  1437. *
  1438. * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
  1439. * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
  1440. *
  1441. * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
  1442. * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
  1443. * various stages of removing an inode.
  1444. *
  1445. * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
  1446. *
  1447. * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
  1448. * fdatasync(). i_atime is the usual cause.
  1449. * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
  1450. * these changes separately from I_DIRTY_SYNC so that we
  1451. * don't have to write inode on fdatasync() when only
  1452. * mtime has changed in it.
  1453. * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
  1454. * I_NEW Serves as both a mutex and completion notification.
  1455. * New inodes set I_NEW. If two processes both create
  1456. * the same inode, one of them will release its inode and
  1457. * wait for I_NEW to be released before returning.
  1458. * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
  1459. * also cause waiting on I_NEW, without I_NEW actually
  1460. * being set. find_inode() uses this to prevent returning
  1461. * nearly-dead inodes.
  1462. * I_WILL_FREE Must be set when calling write_inode_now() if i_count
  1463. * is zero. I_FREEING must be set when I_WILL_FREE is
  1464. * cleared.
  1465. * I_FREEING Set when inode is about to be freed but still has dirty
  1466. * pages or buffers attached or the inode itself is still
  1467. * dirty.
  1468. * I_CLEAR Added by end_writeback(). In this state the inode is clean
  1469. * and can be destroyed. Inode keeps I_FREEING.
  1470. *
  1471. * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
  1472. * prohibited for many purposes. iget() must wait for
  1473. * the inode to be completely released, then create it
  1474. * anew. Other functions will just ignore such inodes,
  1475. * if appropriate. I_NEW is used for waiting.
  1476. *
  1477. * I_SYNC Synchonized write of dirty inode data. The bits is
  1478. * set during data writeback, and cleared with a wakeup
  1479. * on the bit address once it is done.
  1480. *
  1481. * Q: What is the difference between I_WILL_FREE and I_FREEING?
  1482. */
  1483. #define I_DIRTY_SYNC (1 << 0)
  1484. #define I_DIRTY_DATASYNC (1 << 1)
  1485. #define I_DIRTY_PAGES (1 << 2)
  1486. #define __I_NEW 3
  1487. #define I_NEW (1 << __I_NEW)
  1488. #define I_WILL_FREE (1 << 4)
  1489. #define I_FREEING (1 << 5)
  1490. #define I_CLEAR (1 << 6)
  1491. #define __I_SYNC 7
  1492. #define I_SYNC (1 << __I_SYNC)
  1493. #define I_REFERENCED (1 << 8)
  1494. #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
  1495. extern void __mark_inode_dirty(struct inode *, int);
  1496. static inline void mark_inode_dirty(struct inode *inode)
  1497. {
  1498. __mark_inode_dirty(inode, I_DIRTY);
  1499. }
  1500. static inline void mark_inode_dirty_sync(struct inode *inode)
  1501. {
  1502. __mark_inode_dirty(inode, I_DIRTY_SYNC);
  1503. }
  1504. /**
  1505. * inc_nlink - directly increment an inode's link count
  1506. * @inode: inode
  1507. *
  1508. * This is a low-level filesystem helper to replace any
  1509. * direct filesystem manipulation of i_nlink. Currently,
  1510. * it is only here for parity with dec_nlink().
  1511. */
  1512. static inline void inc_nlink(struct inode *inode)
  1513. {
  1514. inode->i_nlink++;
  1515. }
  1516. static inline void inode_inc_link_count(struct inode *inode)
  1517. {
  1518. inc_nlink(inode);
  1519. mark_inode_dirty(inode);
  1520. }
  1521. /**
  1522. * drop_nlink - directly drop an inode's link count
  1523. * @inode: inode
  1524. *
  1525. * This is a low-level filesystem helper to replace any
  1526. * direct filesystem manipulation of i_nlink. In cases
  1527. * where we are attempting to track writes to the
  1528. * filesystem, a decrement to zero means an imminent
  1529. * write when the file is truncated and actually unlinked
  1530. * on the filesystem.
  1531. */
  1532. static inline void drop_nlink(struct inode *inode)
  1533. {
  1534. inode->i_nlink--;
  1535. }
  1536. /**
  1537. * clear_nlink - directly zero an inode's link count
  1538. * @inode: inode
  1539. *
  1540. * This is a low-level filesystem helper to replace any
  1541. * direct filesystem manipulation of i_nlink. See
  1542. * drop_nlink() for why we care about i_nlink hitting zero.
  1543. */
  1544. static inline void clear_nlink(struct inode *inode)
  1545. {
  1546. inode->i_nlink = 0;
  1547. }
  1548. static inline void inode_dec_link_count(struct inode *inode)
  1549. {
  1550. drop_nlink(inode);
  1551. mark_inode_dirty(inode);
  1552. }
  1553. /**
  1554. * inode_inc_iversion - increments i_version
  1555. * @inode: inode that need to be updated
  1556. *
  1557. * Every time the inode is modified, the i_version field will be incremented.
  1558. * The filesystem has to be mounted with i_version flag
  1559. */
  1560. static inline void inode_inc_iversion(struct inode *inode)
  1561. {
  1562. spin_lock(&inode->i_lock);
  1563. inode->i_version++;
  1564. spin_unlock(&inode->i_lock);
  1565. }
  1566. extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
  1567. static inline void file_accessed(struct file *file)
  1568. {
  1569. if (!(file->f_flags & O_NOATIME))
  1570. touch_atime(file->f_path.mnt, file->f_path.dentry);
  1571. }
  1572. int sync_inode(struct inode *inode, struct writeback_control *wbc);
  1573. int sync_inode_metadata(struct inode *inode, int wait);
  1574. struct file_system_type {
  1575. const char *name;
  1576. int fs_flags;
  1577. int (*get_sb) (struct file_system_type *, int,
  1578. const char *, void *, struct vfsmount *);
  1579. struct dentry *(*mount) (struct file_system_type *, int,
  1580. const char *, void *);
  1581. void (*kill_sb) (struct super_block *);
  1582. struct module *owner;
  1583. struct file_system_type * next;
  1584. struct list_head fs_supers;
  1585. struct lock_class_key s_lock_key;
  1586. struct lock_class_key s_umount_key;
  1587. struct lock_class_key s_vfs_rename_key;
  1588. struct lock_class_key i_lock_key;
  1589. struct lock_class_key i_mutex_key;
  1590. struct lock_class_key i_mutex_dir_key;
  1591. struct lock_class_key i_alloc_sem_key;
  1592. };
  1593. extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
  1594. void *data, int (*fill_super)(struct super_block *, void *, int));
  1595. extern struct dentry *mount_bdev(struct file_system_type *fs_type,
  1596. int flags, const char *dev_name, void *data,
  1597. int (*fill_super)(struct super_block *, void *, int));
  1598. extern int get_sb_bdev(struct file_system_type *fs_type,
  1599. int flags, const char *dev_name, void *data,
  1600. int (*fill_super)(struct super_block *, void *, int),
  1601. struct vfsmount *mnt);
  1602. extern struct dentry *mount_single(struct file_system_type *fs_type,
  1603. int flags, void *data,
  1604. int (*fill_super)(struct super_block *, void *, int));
  1605. extern int get_sb_single(struct file_system_type *fs_type,
  1606. int flags, void *data,
  1607. int (*fill_super)(struct super_block *, void *, int),
  1608. struct vfsmount *mnt);
  1609. extern struct dentry *mount_nodev(struct file_system_type *fs_type,
  1610. int flags, void *data,
  1611. int (*fill_super)(struct super_block *, void *, int));
  1612. extern int get_sb_nodev(struct file_system_type *fs_type,
  1613. int flags, void *data,
  1614. int (*fill_super)(struct super_block *, void *, int),
  1615. struct vfsmount *mnt);
  1616. void generic_shutdown_super(struct super_block *sb);
  1617. void kill_block_super(struct super_block *sb);
  1618. void kill_anon_super(struct super_block *sb);
  1619. void kill_litter_super(struct super_block *sb);
  1620. void deactivate_super(struct super_block *sb);
  1621. void deactivate_locked_super(struct super_block *sb);
  1622. int set_anon_super(struct super_block *s, void *data);
  1623. struct super_block *sget(struct file_system_type *type,
  1624. int (*test)(struct super_block *,void *),
  1625. int (*set)(struct super_block *,void *),
  1626. void *data);
  1627. extern struct dentry *mount_pseudo(struct file_system_type *, char *,
  1628. const struct super_operations *ops,
  1629. const struct dentry_operations *dops,
  1630. unsigned long);
  1631. extern void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb);
  1632. static inline void sb_mark_dirty(struct super_block *sb)
  1633. {
  1634. sb->s_dirt = 1;
  1635. }
  1636. static inline void sb_mark_clean(struct super_block *sb)
  1637. {
  1638. sb->s_dirt = 0;
  1639. }
  1640. static inline int sb_is_dirty(struct super_block *sb)
  1641. {
  1642. return sb->s_dirt;
  1643. }
  1644. /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
  1645. #define fops_get(fops) \
  1646. (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
  1647. #define fops_put(fops) \
  1648. do { if (fops) module_put((fops)->owner); } while(0)
  1649. extern int register_filesystem(struct file_system_type *);
  1650. extern int unregister_filesystem(struct file_system_type *);
  1651. extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
  1652. #define kern_mount(type) kern_mount_data(type, NULL)
  1653. extern int may_umount_tree(struct vfsmount *);
  1654. extern int may_umount(struct vfsmount *);
  1655. extern long do_mount(char *, char *, char *, unsigned long, void *);
  1656. extern struct vfsmount *collect_mounts(struct path *);
  1657. extern void drop_collected_mounts(struct vfsmount *);
  1658. extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
  1659. struct vfsmount *);
  1660. extern int vfs_statfs(struct path *, 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 * dentry_open(struct dentry *, struct vfsmount *, int,
  1757. const struct cred *);
  1758. extern int filp_close(struct file *, fl_owner_t id);
  1759. extern char * getname(const char __user *);
  1760. /* fs/ioctl.c */
  1761. extern int ioctl_preallocate(struct file *filp, void __user *argp);
  1762. /* fs/dcache.c */
  1763. extern void __init vfs_caches_init_early(void);
  1764. extern void __init vfs_caches_init(unsigned long);
  1765. extern struct kmem_cache *names_cachep;
  1766. #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
  1767. #define __getname() __getname_gfp(GFP_KERNEL)
  1768. #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
  1769. #ifndef CONFIG_AUDITSYSCALL
  1770. #define putname(name) __putname(name)
  1771. #else
  1772. extern void putname(const char *name);
  1773. #endif
  1774. #ifdef CONFIG_BLOCK
  1775. extern int register_blkdev(unsigned int, const char *);
  1776. extern void unregister_blkdev(unsigned int, const char *);
  1777. extern struct block_device *bdget(dev_t);
  1778. extern struct block_device *bdgrab(struct block_device *bdev);
  1779. extern void bd_set_size(struct block_device *, loff_t size);
  1780. extern void bd_forget(struct inode *inode);
  1781. extern void bdput(struct block_device *);
  1782. extern void invalidate_bdev(struct block_device *);
  1783. extern int sync_blockdev(struct block_device *bdev);
  1784. extern struct super_block *freeze_bdev(struct block_device *);
  1785. extern void emergency_thaw_all(void);
  1786. extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
  1787. extern int fsync_bdev(struct block_device *);
  1788. #else
  1789. static inline void bd_forget(struct inode *inode) {}
  1790. static inline int sync_blockdev(struct block_device *bdev) { return 0; }
  1791. static inline void invalidate_bdev(struct block_device *bdev) {}
  1792. static inline struct super_block *freeze_bdev(struct block_device *sb)
  1793. {
  1794. return NULL;
  1795. }
  1796. static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
  1797. {
  1798. return 0;
  1799. }
  1800. #endif
  1801. extern int sync_filesystem(struct super_block *);
  1802. extern const struct file_operations def_blk_fops;
  1803. extern const struct file_operations def_chr_fops;
  1804. extern const struct file_operations bad_sock_fops;
  1805. extern const struct file_operations def_fifo_fops;
  1806. #ifdef CONFIG_BLOCK
  1807. extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
  1808. extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
  1809. extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
  1810. extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
  1811. extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
  1812. void *holder);
  1813. extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
  1814. void *holder);
  1815. extern int blkdev_put(struct block_device *bdev, fmode_t mode);
  1816. #ifdef CONFIG_SYSFS
  1817. extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
  1818. extern void bd_unlink_disk_holder(struct block_device *bdev,
  1819. struct gendisk *disk);
  1820. #else
  1821. static inline int bd_link_disk_holder(struct block_device *bdev,
  1822. struct gendisk *disk)
  1823. {
  1824. return 0;
  1825. }
  1826. static inline void bd_unlink_disk_holder(struct block_device *bdev,
  1827. struct gendisk *disk)
  1828. {
  1829. }
  1830. #endif
  1831. #endif
  1832. /* fs/char_dev.c */
  1833. #define CHRDEV_MAJOR_HASH_SIZE 255
  1834. extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
  1835. extern int register_chrdev_region(dev_t, unsigned, const char *);
  1836. extern int __register_chrdev(unsigned int major, unsigned int baseminor,
  1837. unsigned int count, const char *name,
  1838. const struct file_operations *fops);
  1839. extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
  1840. unsigned int count, const char *name);
  1841. extern void unregister_chrdev_region(dev_t, unsigned);
  1842. extern void chrdev_show(struct seq_file *,off_t);
  1843. static inline int register_chrdev(unsigned int major, const char *name,
  1844. const struct file_operations *fops)
  1845. {
  1846. return __register_chrdev(major, 0, 256, name, fops);
  1847. }
  1848. static inline void unregister_chrdev(unsigned int major, const char *name)
  1849. {
  1850. __unregister_chrdev(major, 0, 256, name);
  1851. }
  1852. /* fs/block_dev.c */
  1853. #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
  1854. #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
  1855. #ifdef CONFIG_BLOCK
  1856. #define BLKDEV_MAJOR_HASH_SIZE 255
  1857. extern const char *__bdevname(dev_t, char *buffer);
  1858. extern const char *bdevname(struct block_device *bdev, char *buffer);
  1859. extern struct block_device *lookup_bdev(const char *);
  1860. extern void blkdev_show(struct seq_file *,off_t);
  1861. #else
  1862. #define BLKDEV_MAJOR_HASH_SIZE 0
  1863. #endif
  1864. extern void init_special_inode(struct inode *, umode_t, dev_t);
  1865. /* Invalid inode operations -- fs/bad_inode.c */
  1866. extern void make_bad_inode(struct inode *);
  1867. extern int is_bad_inode(struct inode *);
  1868. extern const struct file_operations read_pipefifo_fops;
  1869. extern const struct file_operations write_pipefifo_fops;
  1870. extern const struct file_operations rdwr_pipefifo_fops;
  1871. extern int fs_may_remount_ro(struct super_block *);
  1872. #ifdef CONFIG_BLOCK
  1873. /*
  1874. * return READ, READA, or WRITE
  1875. */
  1876. #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
  1877. /*
  1878. * return data direction, READ or WRITE
  1879. */
  1880. #define bio_data_dir(bio) ((bio)->bi_rw & 1)
  1881. extern void check_disk_size_change(struct gendisk *disk,
  1882. struct block_device *bdev);
  1883. extern int revalidate_disk(struct gendisk *);
  1884. extern int check_disk_change(struct block_device *);
  1885. extern int __invalidate_device(struct block_device *, bool);
  1886. extern int invalidate_partition(struct gendisk *, int);
  1887. #endif
  1888. unsigned long invalidate_mapping_pages(struct address_space *mapping,
  1889. pgoff_t start, pgoff_t end);
  1890. static inline void invalidate_remote_inode(struct inode *inode)
  1891. {
  1892. if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
  1893. S_ISLNK(inode->i_mode))
  1894. invalidate_mapping_pages(inode->i_mapping, 0, -1);
  1895. }
  1896. extern int invalidate_inode_pages2(struct address_space *mapping);
  1897. extern int invalidate_inode_pages2_range(struct address_space *mapping,
  1898. pgoff_t start, pgoff_t end);
  1899. extern int write_inode_now(struct inode *, int);
  1900. extern int filemap_fdatawrite(struct address_space *);
  1901. extern int filemap_flush(struct address_space *);
  1902. extern int filemap_fdatawait(struct address_space *);
  1903. extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
  1904. loff_t lend);
  1905. extern int filemap_write_and_wait(struct address_space *mapping);
  1906. extern int filemap_write_and_wait_range(struct address_space *mapping,
  1907. loff_t lstart, loff_t lend);
  1908. extern int __filemap_fdatawrite_range(struct address_space *mapping,
  1909. loff_t start, loff_t end, int sync_mode);
  1910. extern int filemap_fdatawrite_range(struct address_space *mapping,
  1911. loff_t start, loff_t end);
  1912. extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
  1913. int datasync);
  1914. extern int vfs_fsync(struct file *file, int datasync);
  1915. extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
  1916. extern void sync_supers(void);
  1917. extern void emergency_sync(void);
  1918. extern void emergency_remount(void);
  1919. #ifdef CONFIG_BLOCK
  1920. extern sector_t bmap(struct inode *, sector_t);
  1921. #endif
  1922. extern int notify_change(struct dentry *, struct iattr *);
  1923. extern int inode_permission(struct inode *, int);
  1924. extern int generic_permission(struct inode *, int, unsigned int,
  1925. int (*check_acl)(struct inode *, int, unsigned int));
  1926. static inline bool execute_ok(struct inode *inode)
  1927. {
  1928. return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
  1929. }
  1930. extern int get_write_access(struct inode *);
  1931. extern int deny_write_access(struct file *);
  1932. static inline void put_write_access(struct inode * inode)
  1933. {
  1934. atomic_dec(&inode->i_writecount);
  1935. }
  1936. static inline void allow_write_access(struct file *file)
  1937. {
  1938. if (file)
  1939. atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
  1940. }
  1941. extern int do_pipe_flags(int *, int);
  1942. extern struct file *create_read_pipe(struct file *f, int flags);
  1943. extern struct file *create_write_pipe(int flags);
  1944. extern void free_write_pipe(struct file *);
  1945. extern struct file *do_filp_open(int dfd, const char *pathname,
  1946. int open_flag, int mode, int acc_mode);
  1947. extern int may_open(struct path *, int, int);
  1948. extern int kernel_read(struct file *, loff_t, char *, unsigned long);
  1949. extern struct file * open_exec(const char *);
  1950. /* fs/dcache.c -- generic fs support functions */
  1951. extern int is_subdir(struct dentry *, struct dentry *);
  1952. extern int path_is_under(struct path *, struct path *);
  1953. extern ino_t find_inode_number(struct dentry *, struct qstr *);
  1954. #include <linux/err.h>
  1955. /* needed for stackable file system support */
  1956. extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
  1957. extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
  1958. extern int inode_init_always(struct super_block *, struct inode *);
  1959. extern void inode_init_once(struct inode *);
  1960. extern void address_space_init_once(struct address_space *mapping);
  1961. extern void ihold(struct inode * inode);
  1962. extern void iput(struct inode *);
  1963. extern struct inode * igrab(struct inode *);
  1964. extern ino_t iunique(struct super_block *, ino_t);
  1965. extern int inode_needs_sync(struct inode *inode);
  1966. extern int generic_delete_inode(struct inode *inode);
  1967. extern int generic_drop_inode(struct inode *inode);
  1968. extern struct inode *ilookup5_nowait(struct super_block *sb,
  1969. unsigned long hashval, int (*test)(struct inode *, void *),
  1970. void *data);
  1971. extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
  1972. int (*test)(struct inode *, void *), void *data);
  1973. extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
  1974. extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
  1975. extern struct inode * iget_locked(struct super_block *, unsigned long);
  1976. extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
  1977. extern int insert_inode_locked(struct inode *);
  1978. extern void unlock_new_inode(struct inode *);
  1979. extern unsigned int get_next_ino(void);
  1980. extern void __iget(struct inode * inode);
  1981. extern void iget_failed(struct inode *);
  1982. extern void end_writeback(struct inode *);
  1983. extern void __destroy_inode(struct inode *);
  1984. extern struct inode *new_inode(struct super_block *);
  1985. extern void free_inode_nonrcu(struct inode *inode);
  1986. extern int should_remove_suid(struct dentry *);
  1987. extern int file_remove_suid(struct file *);
  1988. extern void __insert_inode_hash(struct inode *, unsigned long hashval);
  1989. extern void remove_inode_hash(struct inode *);
  1990. static inline void insert_inode_hash(struct inode *inode)
  1991. {
  1992. __insert_inode_hash(inode, inode->i_ino);
  1993. }
  1994. extern void inode_sb_list_add(struct inode *inode);
  1995. #ifdef CONFIG_BLOCK
  1996. extern void submit_bio(int, struct bio *);
  1997. extern int bdev_read_only(struct block_device *);
  1998. #endif
  1999. extern int set_blocksize(struct block_device *, int);
  2000. extern int sb_set_blocksize(struct super_block *, int);
  2001. extern int sb_min_blocksize(struct super_block *, int);
  2002. extern int generic_file_mmap(struct file *, struct vm_area_struct *);
  2003. extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
  2004. extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
  2005. int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
  2006. extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2007. extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
  2008. loff_t *);
  2009. extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2010. extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
  2011. unsigned long *, loff_t, loff_t *, size_t, size_t);
  2012. extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
  2013. unsigned long, loff_t, loff_t *, size_t, ssize_t);
  2014. extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
  2015. extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
  2016. extern int generic_segment_checks(const struct iovec *iov,
  2017. unsigned long *nr_segs, size_t *count, int access_flags);
  2018. /* fs/block_dev.c */
  2019. extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
  2020. unsigned long nr_segs, loff_t pos);
  2021. extern int blkdev_fsync(struct file *filp, int datasync);
  2022. /* fs/splice.c */
  2023. extern ssize_t generic_file_splice_read(struct file *, loff_t *,
  2024. struct pipe_inode_info *, size_t, unsigned int);
  2025. extern ssize_t default_file_splice_read(struct file *, loff_t *,
  2026. struct pipe_inode_info *, size_t, unsigned int);
  2027. extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
  2028. struct file *, loff_t *, size_t, unsigned int);
  2029. extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
  2030. struct file *out, loff_t *, size_t len, unsigned int flags);
  2031. extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
  2032. size_t len, unsigned int flags);
  2033. extern void
  2034. file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
  2035. extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
  2036. extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
  2037. extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
  2038. extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
  2039. int origin);
  2040. extern int generic_file_open(struct inode * inode, struct file * filp);
  2041. extern int nonseekable_open(struct inode * inode, struct file * filp);
  2042. #ifdef CONFIG_FS_XIP
  2043. extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
  2044. loff_t *ppos);
  2045. extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
  2046. extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
  2047. size_t len, loff_t *ppos);
  2048. extern int xip_truncate_page(struct address_space *mapping, loff_t from);
  2049. #else
  2050. static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
  2051. {
  2052. return 0;
  2053. }
  2054. #endif
  2055. #ifdef CONFIG_BLOCK
  2056. typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
  2057. loff_t file_offset);
  2058. enum {
  2059. /* need locking between buffered and direct access */
  2060. DIO_LOCKING = 0x01,
  2061. /* filesystem does not support filling holes */
  2062. DIO_SKIP_HOLES = 0x02,
  2063. };
  2064. void dio_end_io(struct bio *bio, int error);
  2065. ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
  2066. struct block_device *bdev, const struct iovec *iov, loff_t offset,
  2067. unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
  2068. dio_submit_t submit_io, int flags);
  2069. static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
  2070. struct inode *inode, struct block_device *bdev, const struct iovec *iov,
  2071. loff_t offset, unsigned long nr_segs, get_block_t get_block,
  2072. dio_iodone_t end_io)
  2073. {
  2074. return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
  2075. nr_segs, get_block, end_io, NULL,
  2076. DIO_LOCKING | DIO_SKIP_HOLES);
  2077. }
  2078. #endif
  2079. extern const struct file_operations generic_ro_fops;
  2080. #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
  2081. extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
  2082. extern int vfs_follow_link(struct nameidata *, const char *);
  2083. extern int page_readlink(struct dentry *, char __user *, int);
  2084. extern void *page_follow_link_light(struct dentry *, struct nameidata *);
  2085. extern void page_put_link(struct dentry *, struct nameidata *, void *);
  2086. extern int __page_symlink(struct inode *inode, const char *symname, int len,
  2087. int nofs);
  2088. extern int page_symlink(struct inode *inode, const char *symname, int len);
  2089. extern const struct inode_operations page_symlink_inode_operations;
  2090. extern int generic_readlink(struct dentry *, char __user *, int);
  2091. extern void generic_fillattr(struct inode *, struct kstat *);
  2092. extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2093. void __inode_add_bytes(struct inode *inode, loff_t bytes);
  2094. void inode_add_bytes(struct inode *inode, loff_t bytes);
  2095. void inode_sub_bytes(struct inode *inode, loff_t bytes);
  2096. loff_t inode_get_bytes(struct inode *inode);
  2097. void inode_set_bytes(struct inode *inode, loff_t bytes);
  2098. extern int vfs_readdir(struct file *, filldir_t, void *);
  2099. extern int vfs_stat(const char __user *, struct kstat *);
  2100. extern int vfs_lstat(const char __user *, struct kstat *);
  2101. extern int vfs_fstat(unsigned int, struct kstat *);
  2102. extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
  2103. extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
  2104. unsigned long arg);
  2105. extern int __generic_block_fiemap(struct inode *inode,
  2106. struct fiemap_extent_info *fieinfo,
  2107. loff_t start, loff_t len,
  2108. get_block_t *get_block);
  2109. extern int generic_block_fiemap(struct inode *inode,
  2110. struct fiemap_extent_info *fieinfo, u64 start,
  2111. u64 len, get_block_t *get_block);
  2112. extern void get_filesystem(struct file_system_type *fs);
  2113. extern void put_filesystem(struct file_system_type *fs);
  2114. extern struct file_system_type *get_fs_type(const char *name);
  2115. extern struct super_block *get_super(struct block_device *);
  2116. extern struct super_block *get_active_super(struct block_device *bdev);
  2117. extern struct super_block *user_get_super(dev_t);
  2118. extern void drop_super(struct super_block *sb);
  2119. extern void iterate_supers(void (*)(struct super_block *, void *), void *);
  2120. extern int dcache_dir_open(struct inode *, struct file *);
  2121. extern int dcache_dir_close(struct inode *, struct file *);
  2122. extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
  2123. extern int dcache_readdir(struct file *, void *, filldir_t);
  2124. extern int simple_setattr(struct dentry *, struct iattr *);
  2125. extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2126. extern int simple_statfs(struct dentry *, struct kstatfs *);
  2127. extern int simple_link(struct dentry *, struct inode *, struct dentry *);
  2128. extern int simple_unlink(struct inode *, struct dentry *);
  2129. extern int simple_rmdir(struct inode *, struct dentry *);
  2130. extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  2131. extern int noop_fsync(struct file *, int);
  2132. extern int simple_empty(struct dentry *);
  2133. extern int simple_readpage(struct file *file, struct page *page);
  2134. extern int simple_write_begin(struct file *file, struct address_space *mapping,
  2135. loff_t pos, unsigned len, unsigned flags,
  2136. struct page **pagep, void **fsdata);
  2137. extern int simple_write_end(struct file *file, struct address_space *mapping,
  2138. loff_t pos, unsigned len, unsigned copied,
  2139. struct page *page, void *fsdata);
  2140. extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
  2141. extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
  2142. extern const struct file_operations simple_dir_operations;
  2143. extern const struct inode_operations simple_dir_inode_operations;
  2144. struct tree_descr { char *name; const struct file_operations *ops; int mode; };
  2145. struct dentry *d_alloc_name(struct dentry *, const char *);
  2146. extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
  2147. extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
  2148. extern void simple_release_fs(struct vfsmount **mount, int *count);
  2149. extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
  2150. loff_t *ppos, const void *from, size_t available);
  2151. extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
  2152. const void __user *from, size_t count);
  2153. extern int generic_file_fsync(struct file *, int);
  2154. extern int generic_check_addressable(unsigned, u64);
  2155. #ifdef CONFIG_MIGRATION
  2156. extern int buffer_migrate_page(struct address_space *,
  2157. struct page *, struct page *);
  2158. #else
  2159. #define buffer_migrate_page NULL
  2160. #endif
  2161. extern int inode_change_ok(const struct inode *, struct iattr *);
  2162. extern int inode_newsize_ok(const struct inode *, loff_t offset);
  2163. extern void setattr_copy(struct inode *inode, const struct iattr *attr);
  2164. extern void file_update_time(struct file *file);
  2165. extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
  2166. extern void save_mount_options(struct super_block *sb, char *options);
  2167. extern void replace_mount_options(struct super_block *sb, char *options);
  2168. static inline ino_t parent_ino(struct dentry *dentry)
  2169. {
  2170. ino_t res;
  2171. /*
  2172. * Don't strictly need d_lock here? If the parent ino could change
  2173. * then surely we'd have a deeper race in the caller?
  2174. */
  2175. spin_lock(&dentry->d_lock);
  2176. res = dentry->d_parent->d_inode->i_ino;
  2177. spin_unlock(&dentry->d_lock);
  2178. return res;
  2179. }
  2180. /* Transaction based IO helpers */
  2181. /*
  2182. * An argresp is stored in an allocated page and holds the
  2183. * size of the argument or response, along with its content
  2184. */
  2185. struct simple_transaction_argresp {
  2186. ssize_t size;
  2187. char data[0];
  2188. };
  2189. #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
  2190. char *simple_transaction_get(struct file *file, const char __user *buf,
  2191. size_t size);
  2192. ssize_t simple_transaction_read(struct file *file, char __user *buf,
  2193. size_t size, loff_t *pos);
  2194. int simple_transaction_release(struct inode *inode, struct file *file);
  2195. void simple_transaction_set(struct file *file, size_t n);
  2196. /*
  2197. * simple attribute files
  2198. *
  2199. * These attributes behave similar to those in sysfs:
  2200. *
  2201. * Writing to an attribute immediately sets a value, an open file can be
  2202. * written to multiple times.
  2203. *
  2204. * Reading from an attribute creates a buffer from the value that might get
  2205. * read with multiple read calls. When the attribute has been read
  2206. * completely, no further read calls are possible until the file is opened
  2207. * again.
  2208. *
  2209. * All attributes contain a text representation of a numeric value
  2210. * that are accessed with the get() and set() functions.
  2211. */
  2212. #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
  2213. static int __fops ## _open(struct inode *inode, struct file *file) \
  2214. { \
  2215. __simple_attr_check_format(__fmt, 0ull); \
  2216. return simple_attr_open(inode, file, __get, __set, __fmt); \
  2217. } \
  2218. static const struct file_operations __fops = { \
  2219. .owner = THIS_MODULE, \
  2220. .open = __fops ## _open, \
  2221. .release = simple_attr_release, \
  2222. .read = simple_attr_read, \
  2223. .write = simple_attr_write, \
  2224. .llseek = generic_file_llseek, \
  2225. };
  2226. static inline void __attribute__((format(printf, 1, 2)))
  2227. __simple_attr_check_format(const char *fmt, ...)
  2228. {
  2229. /* don't do anything, just let the compiler check the arguments; */
  2230. }
  2231. int simple_attr_open(struct inode *inode, struct file *file,
  2232. int (*get)(void *, u64 *), int (*set)(void *, u64),
  2233. const char *fmt);
  2234. int simple_attr_release(struct inode *inode, struct file *file);
  2235. ssize_t simple_attr_read(struct file *file, char __user *buf,
  2236. size_t len, loff_t *ppos);
  2237. ssize_t simple_attr_write(struct file *file, const char __user *buf,
  2238. size_t len, loff_t *ppos);
  2239. struct ctl_table;
  2240. int proc_nr_files(struct ctl_table *table, int write,
  2241. void __user *buffer, size_t *lenp, loff_t *ppos);
  2242. int proc_nr_dentry(struct ctl_table *table, int write,
  2243. void __user *buffer, size_t *lenp, loff_t *ppos);
  2244. int proc_nr_inodes(struct ctl_table *table, int write,
  2245. void __user *buffer, size_t *lenp, loff_t *ppos);
  2246. int __init get_filesystem_list(char *buf);
  2247. #define __FMODE_EXEC ((__force int) FMODE_EXEC)
  2248. #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
  2249. #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
  2250. #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
  2251. (flag & __FMODE_NONOTIFY)))
  2252. #endif /* __KERNEL__ */
  2253. #endif /* _LINUX_FS_H */