fs.h 87 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568256925702571257225732574257525762577257825792580258125822583258425852586258725882589259025912592259325942595259625972598
  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 is opened with O_PATH; almost nothing can be done with it */
  88. #define FMODE_PATH ((__force fmode_t)0x4000)
  89. /* File was opened by fanotify and shouldn't generate fanotify events */
  90. #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
  91. /*
  92. * The below are the various read and write types that we support. Some of
  93. * them include behavioral modifiers that send information down to the
  94. * block layer and IO scheduler. Terminology:
  95. *
  96. * The block layer uses device plugging to defer IO a little bit, in
  97. * the hope that we will see more IO very shortly. This increases
  98. * coalescing of adjacent IO and thus reduces the number of IOs we
  99. * have to send to the device. It also allows for better queuing,
  100. * if the IO isn't mergeable. If the caller is going to be waiting
  101. * for the IO, then he must ensure that the device is unplugged so
  102. * that the IO is dispatched to the driver.
  103. *
  104. * All IO is handled async in Linux. This is fine for background
  105. * writes, but for reads or writes that someone waits for completion
  106. * on, we want to notify the block layer and IO scheduler so that they
  107. * know about it. That allows them to make better scheduling
  108. * decisions. So when the below references 'sync' and 'async', it
  109. * is referencing this priority hint.
  110. *
  111. * With that in mind, the available types are:
  112. *
  113. * READ A normal read operation. Device will be plugged.
  114. * READ_SYNC A synchronous read. Device is not plugged, caller can
  115. * immediately wait on this read without caring about
  116. * unplugging.
  117. * READA Used for read-ahead operations. Lower priority, and the
  118. * block layer could (in theory) choose to ignore this
  119. * request if it runs into resource problems.
  120. * WRITE A normal async write. Device will be plugged.
  121. * WRITE_SYNC_PLUG Synchronous write. Identical to WRITE, but passes down
  122. * the hint that someone will be waiting on this IO
  123. * shortly. The device must still be unplugged explicitly,
  124. * WRITE_SYNC_PLUG does not do this as we could be
  125. * submitting more writes before we actually wait on any
  126. * of them.
  127. * WRITE_SYNC Like WRITE_SYNC_PLUG, but also unplugs the device
  128. * immediately after submission. The write equivalent
  129. * of READ_SYNC.
  130. * WRITE_ODIRECT_PLUG Special case write for O_DIRECT only.
  131. * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
  132. * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
  133. * non-volatile media on completion.
  134. * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
  135. * by a cache flush and data is guaranteed to be on
  136. * non-volatile media on completion.
  137. *
  138. */
  139. #define RW_MASK REQ_WRITE
  140. #define RWA_MASK REQ_RAHEAD
  141. #define READ 0
  142. #define WRITE RW_MASK
  143. #define READA RWA_MASK
  144. #define READ_SYNC (READ | REQ_SYNC | REQ_UNPLUG)
  145. #define READ_META (READ | REQ_META)
  146. #define WRITE_SYNC_PLUG (WRITE | REQ_SYNC | REQ_NOIDLE)
  147. #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG)
  148. #define WRITE_ODIRECT_PLUG (WRITE | REQ_SYNC)
  149. #define WRITE_META (WRITE | REQ_META)
  150. #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  151. REQ_FLUSH)
  152. #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  153. REQ_FUA)
  154. #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
  155. REQ_FLUSH | REQ_FUA)
  156. #define SEL_IN 1
  157. #define SEL_OUT 2
  158. #define SEL_EX 4
  159. /* public flags for file_system_type */
  160. #define FS_REQUIRES_DEV 1
  161. #define FS_BINARY_MOUNTDATA 2
  162. #define FS_HAS_SUBTYPE 4
  163. #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
  164. #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
  165. * during rename() internally.
  166. */
  167. /*
  168. * These are the fs-independent mount-flags: up to 32 flags are supported
  169. */
  170. #define MS_RDONLY 1 /* Mount read-only */
  171. #define MS_NOSUID 2 /* Ignore suid and sgid bits */
  172. #define MS_NODEV 4 /* Disallow access to device special files */
  173. #define MS_NOEXEC 8 /* Disallow program execution */
  174. #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
  175. #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
  176. #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
  177. #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
  178. #define MS_NOATIME 1024 /* Do not update access times. */
  179. #define MS_NODIRATIME 2048 /* Do not update directory access times */
  180. #define MS_BIND 4096
  181. #define MS_MOVE 8192
  182. #define MS_REC 16384
  183. #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
  184. MS_VERBOSE is deprecated. */
  185. #define MS_SILENT 32768
  186. #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
  187. #define MS_UNBINDABLE (1<<17) /* change to unbindable */
  188. #define MS_PRIVATE (1<<18) /* change to private */
  189. #define MS_SLAVE (1<<19) /* change to slave */
  190. #define MS_SHARED (1<<20) /* change to shared */
  191. #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
  192. #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
  193. #define MS_I_VERSION (1<<23) /* Update inode I_version field */
  194. #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
  195. #define MS_BORN (1<<29)
  196. #define MS_ACTIVE (1<<30)
  197. #define MS_NOUSER (1<<31)
  198. /*
  199. * Superblock flags that can be altered by MS_REMOUNT
  200. */
  201. #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
  202. /*
  203. * Old magic mount flag and mask
  204. */
  205. #define MS_MGC_VAL 0xC0ED0000
  206. #define MS_MGC_MSK 0xffff0000
  207. /* Inode flags - they have nothing to superblock flags now */
  208. #define S_SYNC 1 /* Writes are synced at once */
  209. #define S_NOATIME 2 /* Do not update access times */
  210. #define S_APPEND 4 /* Append-only file */
  211. #define S_IMMUTABLE 8 /* Immutable file */
  212. #define S_DEAD 16 /* removed, but still open directory */
  213. #define S_NOQUOTA 32 /* Inode is not counted to quota */
  214. #define S_DIRSYNC 64 /* Directory modifications are synchronous */
  215. #define S_NOCMTIME 128 /* Do not update file c/mtime */
  216. #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
  217. #define S_PRIVATE 512 /* Inode is fs-internal */
  218. #define S_IMA 1024 /* Inode has an associated IMA struct */
  219. #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
  220. /*
  221. * Note that nosuid etc flags are inode-specific: setting some file-system
  222. * flags just means all the inodes inherit those flags by default. It might be
  223. * possible to override it selectively if you really wanted to with some
  224. * ioctl() that is not currently implemented.
  225. *
  226. * Exception: MS_RDONLY is always applied to the entire file system.
  227. *
  228. * Unfortunately, it is possible to change a filesystems flags with it mounted
  229. * with files in use. This means that all of the inodes will not have their
  230. * i_flags updated. Hence, i_flags no longer inherit the superblock mount
  231. * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
  232. */
  233. #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
  234. #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
  235. #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
  236. ((inode)->i_flags & S_SYNC))
  237. #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
  238. ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
  239. #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
  240. #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
  241. #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
  242. #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
  243. #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
  244. #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
  245. #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
  246. #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
  247. #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
  248. #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
  249. #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
  250. #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
  251. #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
  252. /* the read-only stuff doesn't really belong here, but any other place is
  253. probably as bad and I don't want to create yet another include file. */
  254. #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
  255. #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
  256. #define BLKRRPART _IO(0x12,95) /* re-read partition table */
  257. #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
  258. #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
  259. #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
  260. #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
  261. #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
  262. #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
  263. #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
  264. #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
  265. #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
  266. #if 0
  267. #define BLKPG _IO(0x12,105)/* See blkpg.h */
  268. /* Some people are morons. Do not use sizeof! */
  269. #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
  270. #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
  271. /* This was here just to show that the number is taken -
  272. probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
  273. #endif
  274. /* A jump here: 108-111 have been used for various private purposes. */
  275. #define BLKBSZGET _IOR(0x12,112,size_t)
  276. #define BLKBSZSET _IOW(0x12,113,size_t)
  277. #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
  278. #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
  279. #define BLKTRACESTART _IO(0x12,116)
  280. #define BLKTRACESTOP _IO(0x12,117)
  281. #define BLKTRACETEARDOWN _IO(0x12,118)
  282. #define BLKDISCARD _IO(0x12,119)
  283. #define BLKIOMIN _IO(0x12,120)
  284. #define BLKIOOPT _IO(0x12,121)
  285. #define BLKALIGNOFF _IO(0x12,122)
  286. #define BLKPBSZGET _IO(0x12,123)
  287. #define BLKDISCARDZEROES _IO(0x12,124)
  288. #define BLKSECDISCARD _IO(0x12,125)
  289. #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
  290. #define FIBMAP _IO(0x00,1) /* bmap access */
  291. #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
  292. #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
  293. #define FITHAW _IOWR('X', 120, int) /* Thaw */
  294. #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
  295. #define FS_IOC_GETFLAGS _IOR('f', 1, long)
  296. #define FS_IOC_SETFLAGS _IOW('f', 2, long)
  297. #define FS_IOC_GETVERSION _IOR('v', 1, long)
  298. #define FS_IOC_SETVERSION _IOW('v', 2, long)
  299. #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
  300. #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
  301. #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
  302. #define FS_IOC32_GETVERSION _IOR('v', 1, int)
  303. #define FS_IOC32_SETVERSION _IOW('v', 2, int)
  304. /*
  305. * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
  306. */
  307. #define FS_SECRM_FL 0x00000001 /* Secure deletion */
  308. #define FS_UNRM_FL 0x00000002 /* Undelete */
  309. #define FS_COMPR_FL 0x00000004 /* Compress file */
  310. #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
  311. #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
  312. #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
  313. #define FS_NODUMP_FL 0x00000040 /* do not dump file */
  314. #define FS_NOATIME_FL 0x00000080 /* do not update atime */
  315. /* Reserved for compression usage... */
  316. #define FS_DIRTY_FL 0x00000100
  317. #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
  318. #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
  319. #define FS_ECOMPR_FL 0x00000800 /* Compression error */
  320. /* End compression flags --- maybe not all used */
  321. #define FS_BTREE_FL 0x00001000 /* btree format dir */
  322. #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
  323. #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
  324. #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
  325. #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
  326. #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
  327. #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
  328. #define FS_EXTENT_FL 0x00080000 /* Extents */
  329. #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
  330. #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
  331. #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
  332. #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
  333. #define SYNC_FILE_RANGE_WAIT_BEFORE 1
  334. #define SYNC_FILE_RANGE_WRITE 2
  335. #define SYNC_FILE_RANGE_WAIT_AFTER 4
  336. #ifdef __KERNEL__
  337. #include <linux/linkage.h>
  338. #include <linux/wait.h>
  339. #include <linux/types.h>
  340. #include <linux/kdev_t.h>
  341. #include <linux/dcache.h>
  342. #include <linux/path.h>
  343. #include <linux/stat.h>
  344. #include <linux/cache.h>
  345. #include <linux/list.h>
  346. #include <linux/radix-tree.h>
  347. #include <linux/prio_tree.h>
  348. #include <linux/init.h>
  349. #include <linux/pid.h>
  350. #include <linux/mutex.h>
  351. #include <linux/capability.h>
  352. #include <linux/semaphore.h>
  353. #include <linux/fiemap.h>
  354. #include <linux/rculist_bl.h>
  355. #include <asm/atomic.h>
  356. #include <asm/byteorder.h>
  357. struct export_operations;
  358. struct hd_geometry;
  359. struct iovec;
  360. struct nameidata;
  361. struct kiocb;
  362. struct kobject;
  363. struct pipe_inode_info;
  364. struct poll_table_struct;
  365. struct kstatfs;
  366. struct vm_area_struct;
  367. struct vfsmount;
  368. struct cred;
  369. extern void __init inode_init(void);
  370. extern void __init inode_init_early(void);
  371. extern void __init files_init(unsigned long);
  372. extern struct files_stat_struct files_stat;
  373. extern unsigned long get_max_files(void);
  374. extern int sysctl_nr_open;
  375. extern struct inodes_stat_t inodes_stat;
  376. extern int leases_enable, lease_break_time;
  377. struct buffer_head;
  378. typedef int (get_block_t)(struct inode *inode, sector_t iblock,
  379. struct buffer_head *bh_result, int create);
  380. typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
  381. ssize_t bytes, void *private, int ret,
  382. bool is_async);
  383. /*
  384. * Attribute flags. These should be or-ed together to figure out what
  385. * has been changed!
  386. */
  387. #define ATTR_MODE (1 << 0)
  388. #define ATTR_UID (1 << 1)
  389. #define ATTR_GID (1 << 2)
  390. #define ATTR_SIZE (1 << 3)
  391. #define ATTR_ATIME (1 << 4)
  392. #define ATTR_MTIME (1 << 5)
  393. #define ATTR_CTIME (1 << 6)
  394. #define ATTR_ATIME_SET (1 << 7)
  395. #define ATTR_MTIME_SET (1 << 8)
  396. #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
  397. #define ATTR_ATTR_FLAG (1 << 10)
  398. #define ATTR_KILL_SUID (1 << 11)
  399. #define ATTR_KILL_SGID (1 << 12)
  400. #define ATTR_FILE (1 << 13)
  401. #define ATTR_KILL_PRIV (1 << 14)
  402. #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
  403. #define ATTR_TIMES_SET (1 << 16)
  404. /*
  405. * This is the Inode Attributes structure, used for notify_change(). It
  406. * uses the above definitions as flags, to know which values have changed.
  407. * Also, in this manner, a Filesystem can look at only the values it cares
  408. * about. Basically, these are the attributes that the VFS layer can
  409. * request to change from the FS layer.
  410. *
  411. * Derek Atkins <warlord@MIT.EDU> 94-10-20
  412. */
  413. struct iattr {
  414. unsigned int ia_valid;
  415. umode_t ia_mode;
  416. uid_t ia_uid;
  417. gid_t ia_gid;
  418. loff_t ia_size;
  419. struct timespec ia_atime;
  420. struct timespec ia_mtime;
  421. struct timespec ia_ctime;
  422. /*
  423. * Not an attribute, but an auxilary info for filesystems wanting to
  424. * implement an ftruncate() like method. NOTE: filesystem should
  425. * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
  426. */
  427. struct file *ia_file;
  428. };
  429. /*
  430. * Includes for diskquotas.
  431. */
  432. #include <linux/quota.h>
  433. /**
  434. * enum positive_aop_returns - aop return codes with specific semantics
  435. *
  436. * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
  437. * completed, that the page is still locked, and
  438. * should be considered active. The VM uses this hint
  439. * to return the page to the active list -- it won't
  440. * be a candidate for writeback again in the near
  441. * future. Other callers must be careful to unlock
  442. * the page if they get this return. Returned by
  443. * writepage();
  444. *
  445. * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
  446. * unlocked it and the page might have been truncated.
  447. * The caller should back up to acquiring a new page and
  448. * trying again. The aop will be taking reasonable
  449. * precautions not to livelock. If the caller held a page
  450. * reference, it should drop it before retrying. Returned
  451. * by readpage().
  452. *
  453. * address_space_operation functions return these large constants to indicate
  454. * special semantics to the caller. These are much larger than the bytes in a
  455. * page to allow for functions that return the number of bytes operated on in a
  456. * given page.
  457. */
  458. enum positive_aop_returns {
  459. AOP_WRITEPAGE_ACTIVATE = 0x80000,
  460. AOP_TRUNCATED_PAGE = 0x80001,
  461. };
  462. #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
  463. #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
  464. #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
  465. * helper code (eg buffer layer)
  466. * to clear GFP_FS from alloc */
  467. /*
  468. * oh the beauties of C type declarations.
  469. */
  470. struct page;
  471. struct address_space;
  472. struct writeback_control;
  473. struct iov_iter {
  474. const struct iovec *iov;
  475. unsigned long nr_segs;
  476. size_t iov_offset;
  477. size_t count;
  478. };
  479. size_t iov_iter_copy_from_user_atomic(struct page *page,
  480. struct iov_iter *i, unsigned long offset, size_t bytes);
  481. size_t iov_iter_copy_from_user(struct page *page,
  482. struct iov_iter *i, unsigned long offset, size_t bytes);
  483. void iov_iter_advance(struct iov_iter *i, size_t bytes);
  484. int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
  485. size_t iov_iter_single_seg_count(struct iov_iter *i);
  486. static inline void iov_iter_init(struct iov_iter *i,
  487. const struct iovec *iov, unsigned long nr_segs,
  488. size_t count, size_t written)
  489. {
  490. i->iov = iov;
  491. i->nr_segs = nr_segs;
  492. i->iov_offset = 0;
  493. i->count = count + written;
  494. iov_iter_advance(i, written);
  495. }
  496. static inline size_t iov_iter_count(struct iov_iter *i)
  497. {
  498. return i->count;
  499. }
  500. /*
  501. * "descriptor" for what we're up to with a read.
  502. * This allows us to use the same read code yet
  503. * have multiple different users of the data that
  504. * we read from a file.
  505. *
  506. * The simplest case just copies the data to user
  507. * mode.
  508. */
  509. typedef struct {
  510. size_t written;
  511. size_t count;
  512. union {
  513. char __user *buf;
  514. void *data;
  515. } arg;
  516. int error;
  517. } read_descriptor_t;
  518. typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
  519. unsigned long, unsigned long);
  520. struct address_space_operations {
  521. int (*writepage)(struct page *page, struct writeback_control *wbc);
  522. int (*readpage)(struct file *, struct page *);
  523. void (*sync_page)(struct page *);
  524. /* Write back some dirty pages from this mapping. */
  525. int (*writepages)(struct address_space *, struct writeback_control *);
  526. /* Set a page dirty. Return true if this dirtied it */
  527. int (*set_page_dirty)(struct page *page);
  528. int (*readpages)(struct file *filp, struct address_space *mapping,
  529. struct list_head *pages, unsigned nr_pages);
  530. int (*write_begin)(struct file *, struct address_space *mapping,
  531. loff_t pos, unsigned len, unsigned flags,
  532. struct page **pagep, void **fsdata);
  533. int (*write_end)(struct file *, struct address_space *mapping,
  534. loff_t pos, unsigned len, unsigned copied,
  535. struct page *page, void *fsdata);
  536. /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
  537. sector_t (*bmap)(struct address_space *, sector_t);
  538. void (*invalidatepage) (struct page *, unsigned long);
  539. int (*releasepage) (struct page *, gfp_t);
  540. void (*freepage)(struct page *);
  541. ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
  542. loff_t offset, unsigned long nr_segs);
  543. int (*get_xip_mem)(struct address_space *, pgoff_t, int,
  544. void **, unsigned long *);
  545. /* migrate the contents of a page to the specified target */
  546. int (*migratepage) (struct address_space *,
  547. struct page *, struct page *);
  548. int (*launder_page) (struct page *);
  549. int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
  550. unsigned long);
  551. int (*error_remove_page)(struct address_space *, struct page *);
  552. };
  553. /*
  554. * pagecache_write_begin/pagecache_write_end must be used by general code
  555. * to write into the pagecache.
  556. */
  557. int pagecache_write_begin(struct file *, struct address_space *mapping,
  558. loff_t pos, unsigned len, unsigned flags,
  559. struct page **pagep, void **fsdata);
  560. int pagecache_write_end(struct file *, struct address_space *mapping,
  561. loff_t pos, unsigned len, unsigned copied,
  562. struct page *page, void *fsdata);
  563. struct backing_dev_info;
  564. struct address_space {
  565. struct inode *host; /* owner: inode, block_device */
  566. struct radix_tree_root page_tree; /* radix tree of all pages */
  567. spinlock_t tree_lock; /* and lock protecting it */
  568. unsigned int i_mmap_writable;/* count VM_SHARED mappings */
  569. struct prio_tree_root i_mmap; /* tree of private and shared mappings */
  570. struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
  571. spinlock_t i_mmap_lock; /* protect tree, count, list */
  572. unsigned int truncate_count; /* Cover race condition with truncate */
  573. unsigned long nrpages; /* number of total pages */
  574. pgoff_t writeback_index;/* writeback starts here */
  575. const struct address_space_operations *a_ops; /* methods */
  576. unsigned long flags; /* error bits/gfp mask */
  577. struct backing_dev_info *backing_dev_info; /* device readahead, etc */
  578. spinlock_t private_lock; /* for use by the address_space */
  579. struct list_head private_list; /* ditto */
  580. struct address_space *assoc_mapping; /* ditto */
  581. struct mutex unmap_mutex; /* to protect unmapping */
  582. } __attribute__((aligned(sizeof(long))));
  583. /*
  584. * On most architectures that alignment is already the case; but
  585. * must be enforced here for CRIS, to let the least signficant bit
  586. * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
  587. */
  588. struct block_device {
  589. dev_t bd_dev; /* not a kdev_t - it's a search key */
  590. struct inode * bd_inode; /* will die */
  591. struct super_block * bd_super;
  592. int bd_openers;
  593. struct mutex bd_mutex; /* open/close mutex */
  594. struct list_head bd_inodes;
  595. void * bd_claiming;
  596. void * bd_holder;
  597. int bd_holders;
  598. bool bd_write_holder;
  599. #ifdef CONFIG_SYSFS
  600. struct list_head bd_holder_disks;
  601. #endif
  602. struct block_device * bd_contains;
  603. unsigned bd_block_size;
  604. struct hd_struct * bd_part;
  605. /* number of times partitions within this device have been opened. */
  606. unsigned bd_part_count;
  607. int bd_invalidated;
  608. struct gendisk * bd_disk;
  609. struct list_head bd_list;
  610. /*
  611. * Private data. You must have bd_claim'ed the block_device
  612. * to use this. NOTE: bd_claim allows an owner to claim
  613. * the same device multiple times, the owner must take special
  614. * care to not mess up bd_private for that case.
  615. */
  616. unsigned long bd_private;
  617. /* The counter of freeze processes */
  618. int bd_fsfreeze_count;
  619. /* Mutex for freeze */
  620. struct mutex bd_fsfreeze_mutex;
  621. };
  622. /*
  623. * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
  624. * radix trees
  625. */
  626. #define PAGECACHE_TAG_DIRTY 0
  627. #define PAGECACHE_TAG_WRITEBACK 1
  628. #define PAGECACHE_TAG_TOWRITE 2
  629. int mapping_tagged(struct address_space *mapping, int tag);
  630. /*
  631. * Might pages of this file be mapped into userspace?
  632. */
  633. static inline int mapping_mapped(struct address_space *mapping)
  634. {
  635. return !prio_tree_empty(&mapping->i_mmap) ||
  636. !list_empty(&mapping->i_mmap_nonlinear);
  637. }
  638. /*
  639. * Might pages of this file have been modified in userspace?
  640. * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
  641. * marks vma as VM_SHARED if it is shared, and the file was opened for
  642. * writing i.e. vma may be mprotected writable even if now readonly.
  643. */
  644. static inline int mapping_writably_mapped(struct address_space *mapping)
  645. {
  646. return mapping->i_mmap_writable != 0;
  647. }
  648. /*
  649. * Use sequence counter to get consistent i_size on 32-bit processors.
  650. */
  651. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  652. #include <linux/seqlock.h>
  653. #define __NEED_I_SIZE_ORDERED
  654. #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
  655. #else
  656. #define i_size_ordered_init(inode) do { } while (0)
  657. #endif
  658. struct posix_acl;
  659. #define ACL_NOT_CACHED ((void *)(-1))
  660. struct inode {
  661. /* RCU path lookup touches following: */
  662. umode_t i_mode;
  663. uid_t i_uid;
  664. gid_t i_gid;
  665. const struct inode_operations *i_op;
  666. struct super_block *i_sb;
  667. spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
  668. unsigned int i_flags;
  669. struct mutex i_mutex;
  670. unsigned long i_state;
  671. unsigned long dirtied_when; /* jiffies of first dirtying */
  672. struct hlist_node i_hash;
  673. struct list_head i_wb_list; /* backing dev IO list */
  674. struct list_head i_lru; /* inode LRU list */
  675. struct list_head i_sb_list;
  676. union {
  677. struct list_head i_dentry;
  678. struct rcu_head i_rcu;
  679. };
  680. unsigned long i_ino;
  681. atomic_t i_count;
  682. unsigned int i_nlink;
  683. dev_t i_rdev;
  684. unsigned int i_blkbits;
  685. u64 i_version;
  686. loff_t i_size;
  687. #ifdef __NEED_I_SIZE_ORDERED
  688. seqcount_t i_size_seqcount;
  689. #endif
  690. struct timespec i_atime;
  691. struct timespec i_mtime;
  692. struct timespec i_ctime;
  693. blkcnt_t i_blocks;
  694. unsigned short i_bytes;
  695. struct rw_semaphore i_alloc_sem;
  696. const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
  697. struct file_lock *i_flock;
  698. struct address_space *i_mapping;
  699. struct address_space i_data;
  700. #ifdef CONFIG_QUOTA
  701. struct dquot *i_dquot[MAXQUOTAS];
  702. #endif
  703. struct list_head i_devices;
  704. union {
  705. struct pipe_inode_info *i_pipe;
  706. struct block_device *i_bdev;
  707. struct cdev *i_cdev;
  708. };
  709. __u32 i_generation;
  710. #ifdef CONFIG_FSNOTIFY
  711. __u32 i_fsnotify_mask; /* all events this inode cares about */
  712. struct hlist_head i_fsnotify_marks;
  713. #endif
  714. #ifdef CONFIG_IMA
  715. atomic_t i_readcount; /* struct files open RO */
  716. #endif
  717. atomic_t i_writecount;
  718. #ifdef CONFIG_SECURITY
  719. void *i_security;
  720. #endif
  721. #ifdef CONFIG_FS_POSIX_ACL
  722. struct posix_acl *i_acl;
  723. struct posix_acl *i_default_acl;
  724. #endif
  725. void *i_private; /* fs or device private pointer */
  726. };
  727. static inline int inode_unhashed(struct inode *inode)
  728. {
  729. return hlist_unhashed(&inode->i_hash);
  730. }
  731. /*
  732. * inode->i_mutex nesting subclasses for the lock validator:
  733. *
  734. * 0: the object of the current VFS operation
  735. * 1: parent
  736. * 2: child/target
  737. * 3: quota file
  738. *
  739. * The locking order between these classes is
  740. * parent -> child -> normal -> xattr -> quota
  741. */
  742. enum inode_i_mutex_lock_class
  743. {
  744. I_MUTEX_NORMAL,
  745. I_MUTEX_PARENT,
  746. I_MUTEX_CHILD,
  747. I_MUTEX_XATTR,
  748. I_MUTEX_QUOTA
  749. };
  750. /*
  751. * NOTE: in a 32bit arch with a preemptable kernel and
  752. * an UP compile the i_size_read/write must be atomic
  753. * with respect to the local cpu (unlike with preempt disabled),
  754. * but they don't need to be atomic with respect to other cpus like in
  755. * true SMP (so they need either to either locally disable irq around
  756. * the read or for example on x86 they can be still implemented as a
  757. * cmpxchg8b without the need of the lock prefix). For SMP compiles
  758. * and 64bit archs it makes no difference if preempt is enabled or not.
  759. */
  760. static inline loff_t i_size_read(const struct inode *inode)
  761. {
  762. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  763. loff_t i_size;
  764. unsigned int seq;
  765. do {
  766. seq = read_seqcount_begin(&inode->i_size_seqcount);
  767. i_size = inode->i_size;
  768. } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
  769. return i_size;
  770. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  771. loff_t i_size;
  772. preempt_disable();
  773. i_size = inode->i_size;
  774. preempt_enable();
  775. return i_size;
  776. #else
  777. return inode->i_size;
  778. #endif
  779. }
  780. /*
  781. * NOTE: unlike i_size_read(), i_size_write() does need locking around it
  782. * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
  783. * can be lost, resulting in subsequent i_size_read() calls spinning forever.
  784. */
  785. static inline void i_size_write(struct inode *inode, loff_t i_size)
  786. {
  787. #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
  788. write_seqcount_begin(&inode->i_size_seqcount);
  789. inode->i_size = i_size;
  790. write_seqcount_end(&inode->i_size_seqcount);
  791. #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
  792. preempt_disable();
  793. inode->i_size = i_size;
  794. preempt_enable();
  795. #else
  796. inode->i_size = i_size;
  797. #endif
  798. }
  799. static inline unsigned iminor(const struct inode *inode)
  800. {
  801. return MINOR(inode->i_rdev);
  802. }
  803. static inline unsigned imajor(const struct inode *inode)
  804. {
  805. return MAJOR(inode->i_rdev);
  806. }
  807. extern struct block_device *I_BDEV(struct inode *inode);
  808. struct fown_struct {
  809. rwlock_t lock; /* protects pid, uid, euid fields */
  810. struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
  811. enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
  812. uid_t uid, euid; /* uid/euid of process setting the owner */
  813. int signum; /* posix.1b rt signal to be delivered on IO */
  814. };
  815. /*
  816. * Track a single file's readahead state
  817. */
  818. struct file_ra_state {
  819. pgoff_t start; /* where readahead started */
  820. unsigned int size; /* # of readahead pages */
  821. unsigned int async_size; /* do asynchronous readahead when
  822. there are only # of pages ahead */
  823. unsigned int ra_pages; /* Maximum readahead window */
  824. unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
  825. loff_t prev_pos; /* Cache last read() position */
  826. };
  827. /*
  828. * Check if @index falls in the readahead windows.
  829. */
  830. static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
  831. {
  832. return (index >= ra->start &&
  833. index < ra->start + ra->size);
  834. }
  835. #define FILE_MNT_WRITE_TAKEN 1
  836. #define FILE_MNT_WRITE_RELEASED 2
  837. struct file {
  838. /*
  839. * fu_list becomes invalid after file_free is called and queued via
  840. * fu_rcuhead for RCU freeing
  841. */
  842. union {
  843. struct list_head fu_list;
  844. struct rcu_head fu_rcuhead;
  845. } f_u;
  846. struct path f_path;
  847. #define f_dentry f_path.dentry
  848. #define f_vfsmnt f_path.mnt
  849. const struct file_operations *f_op;
  850. spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
  851. #ifdef CONFIG_SMP
  852. int f_sb_list_cpu;
  853. #endif
  854. atomic_long_t f_count;
  855. unsigned int f_flags;
  856. fmode_t f_mode;
  857. loff_t f_pos;
  858. struct fown_struct f_owner;
  859. const struct cred *f_cred;
  860. struct file_ra_state f_ra;
  861. u64 f_version;
  862. #ifdef CONFIG_SECURITY
  863. void *f_security;
  864. #endif
  865. /* needed for tty driver, and maybe others */
  866. void *private_data;
  867. #ifdef CONFIG_EPOLL
  868. /* Used by fs/eventpoll.c to link all the hooks to this file */
  869. struct list_head f_ep_links;
  870. #endif /* #ifdef CONFIG_EPOLL */
  871. struct address_space *f_mapping;
  872. #ifdef CONFIG_DEBUG_WRITECOUNT
  873. unsigned long f_mnt_write_state;
  874. #endif
  875. };
  876. struct file_handle {
  877. __u32 handle_bytes;
  878. int handle_type;
  879. /* file identifier */
  880. unsigned char f_handle[0];
  881. };
  882. #define get_file(x) atomic_long_inc(&(x)->f_count)
  883. #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
  884. #define file_count(x) atomic_long_read(&(x)->f_count)
  885. #ifdef CONFIG_DEBUG_WRITECOUNT
  886. static inline void file_take_write(struct file *f)
  887. {
  888. WARN_ON(f->f_mnt_write_state != 0);
  889. f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
  890. }
  891. static inline void file_release_write(struct file *f)
  892. {
  893. f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
  894. }
  895. static inline void file_reset_write(struct file *f)
  896. {
  897. f->f_mnt_write_state = 0;
  898. }
  899. static inline void file_check_state(struct file *f)
  900. {
  901. /*
  902. * At this point, either both or neither of these bits
  903. * should be set.
  904. */
  905. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
  906. WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
  907. }
  908. static inline int file_check_writeable(struct file *f)
  909. {
  910. if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
  911. return 0;
  912. printk(KERN_WARNING "writeable file with no "
  913. "mnt_want_write()\n");
  914. WARN_ON(1);
  915. return -EINVAL;
  916. }
  917. #else /* !CONFIG_DEBUG_WRITECOUNT */
  918. static inline void file_take_write(struct file *filp) {}
  919. static inline void file_release_write(struct file *filp) {}
  920. static inline void file_reset_write(struct file *filp) {}
  921. static inline void file_check_state(struct file *filp) {}
  922. static inline int file_check_writeable(struct file *filp)
  923. {
  924. return 0;
  925. }
  926. #endif /* CONFIG_DEBUG_WRITECOUNT */
  927. #define MAX_NON_LFS ((1UL<<31) - 1)
  928. /* Page cache limit. The filesystems should put that into their s_maxbytes
  929. limits, otherwise bad things can happen in VM. */
  930. #if BITS_PER_LONG==32
  931. #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
  932. #elif BITS_PER_LONG==64
  933. #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
  934. #endif
  935. #define FL_POSIX 1
  936. #define FL_FLOCK 2
  937. #define FL_ACCESS 8 /* not trying to lock, just looking */
  938. #define FL_EXISTS 16 /* when unlocking, test for existence */
  939. #define FL_LEASE 32 /* lease held on this file */
  940. #define FL_CLOSE 64 /* unlock on close */
  941. #define FL_SLEEP 128 /* A blocking lock */
  942. /*
  943. * Special return value from posix_lock_file() and vfs_lock_file() for
  944. * asynchronous locking.
  945. */
  946. #define FILE_LOCK_DEFERRED 1
  947. /*
  948. * The POSIX file lock owner is determined by
  949. * the "struct files_struct" in the thread group
  950. * (or NULL for no owner - BSD locks).
  951. *
  952. * Lockd stuffs a "host" pointer into this.
  953. */
  954. typedef struct files_struct *fl_owner_t;
  955. struct file_lock_operations {
  956. void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
  957. void (*fl_release_private)(struct file_lock *);
  958. };
  959. struct lock_manager_operations {
  960. int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
  961. void (*fl_notify)(struct file_lock *); /* unblock callback */
  962. int (*fl_grant)(struct file_lock *, struct file_lock *, int);
  963. void (*fl_release_private)(struct file_lock *);
  964. void (*fl_break)(struct file_lock *);
  965. int (*fl_change)(struct file_lock **, int);
  966. };
  967. struct lock_manager {
  968. struct list_head list;
  969. };
  970. void locks_start_grace(struct lock_manager *);
  971. void locks_end_grace(struct lock_manager *);
  972. int locks_in_grace(void);
  973. /* that will die - we need it for nfs_lock_info */
  974. #include <linux/nfs_fs_i.h>
  975. struct file_lock {
  976. struct file_lock *fl_next; /* singly linked list for this inode */
  977. struct list_head fl_link; /* doubly linked list of all locks */
  978. struct list_head fl_block; /* circular list of blocked processes */
  979. fl_owner_t fl_owner;
  980. unsigned char fl_flags;
  981. unsigned char fl_type;
  982. unsigned int fl_pid;
  983. struct pid *fl_nspid;
  984. wait_queue_head_t fl_wait;
  985. struct file *fl_file;
  986. loff_t fl_start;
  987. loff_t fl_end;
  988. struct fasync_struct * fl_fasync; /* for lease break notifications */
  989. unsigned long fl_break_time; /* for nonblocking lease breaks */
  990. const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
  991. const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
  992. union {
  993. struct nfs_lock_info nfs_fl;
  994. struct nfs4_lock_info nfs4_fl;
  995. struct {
  996. struct list_head link; /* link in AFS vnode's pending_locks list */
  997. int state; /* state of grant or error if -ve */
  998. } afs;
  999. } fl_u;
  1000. };
  1001. /* The following constant reflects the upper bound of the file/locking space */
  1002. #ifndef OFFSET_MAX
  1003. #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
  1004. #define OFFSET_MAX INT_LIMIT(loff_t)
  1005. #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
  1006. #endif
  1007. #include <linux/fcntl.h>
  1008. extern void send_sigio(struct fown_struct *fown, int fd, int band);
  1009. #ifdef CONFIG_FILE_LOCKING
  1010. extern int fcntl_getlk(struct file *, struct flock __user *);
  1011. extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
  1012. struct flock __user *);
  1013. #if BITS_PER_LONG == 32
  1014. extern int fcntl_getlk64(struct file *, struct flock64 __user *);
  1015. extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
  1016. struct flock64 __user *);
  1017. #endif
  1018. extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
  1019. extern int fcntl_getlease(struct file *filp);
  1020. /* fs/locks.c */
  1021. void locks_free_lock(struct file_lock *fl);
  1022. extern void locks_init_lock(struct file_lock *);
  1023. extern struct file_lock * locks_alloc_lock(void);
  1024. extern void locks_copy_lock(struct file_lock *, struct file_lock *);
  1025. extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
  1026. extern void locks_remove_posix(struct file *, fl_owner_t);
  1027. extern void locks_remove_flock(struct file *);
  1028. extern void locks_release_private(struct file_lock *);
  1029. extern void posix_test_lock(struct file *, struct file_lock *);
  1030. extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
  1031. extern int posix_lock_file_wait(struct file *, struct file_lock *);
  1032. extern int posix_unblock_lock(struct file *, struct file_lock *);
  1033. extern int vfs_test_lock(struct file *, struct file_lock *);
  1034. extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
  1035. extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
  1036. extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
  1037. extern int __break_lease(struct inode *inode, unsigned int flags);
  1038. extern void lease_get_mtime(struct inode *, struct timespec *time);
  1039. extern int generic_setlease(struct file *, long, struct file_lock **);
  1040. extern int vfs_setlease(struct file *, long, struct file_lock **);
  1041. extern int lease_modify(struct file_lock **, int);
  1042. extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
  1043. extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
  1044. extern void lock_flocks(void);
  1045. extern void unlock_flocks(void);
  1046. #else /* !CONFIG_FILE_LOCKING */
  1047. static inline int fcntl_getlk(struct file *file, struct flock __user *user)
  1048. {
  1049. return -EINVAL;
  1050. }
  1051. static inline int fcntl_setlk(unsigned int fd, struct file *file,
  1052. unsigned int cmd, struct flock __user *user)
  1053. {
  1054. return -EACCES;
  1055. }
  1056. #if BITS_PER_LONG == 32
  1057. static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
  1058. {
  1059. return -EINVAL;
  1060. }
  1061. static inline int fcntl_setlk64(unsigned int fd, struct file *file,
  1062. unsigned int cmd, struct flock64 __user *user)
  1063. {
  1064. return -EACCES;
  1065. }
  1066. #endif
  1067. static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
  1068. {
  1069. return 0;
  1070. }
  1071. static inline int fcntl_getlease(struct file *filp)
  1072. {
  1073. return 0;
  1074. }
  1075. static inline void locks_init_lock(struct file_lock *fl)
  1076. {
  1077. return;
  1078. }
  1079. static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1080. {
  1081. return;
  1082. }
  1083. static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
  1084. {
  1085. return;
  1086. }
  1087. static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
  1088. {
  1089. return;
  1090. }
  1091. static inline void locks_remove_flock(struct file *filp)
  1092. {
  1093. return;
  1094. }
  1095. static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
  1096. {
  1097. return;
  1098. }
  1099. static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
  1100. struct file_lock *conflock)
  1101. {
  1102. return -ENOLCK;
  1103. }
  1104. static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
  1105. {
  1106. return -ENOLCK;
  1107. }
  1108. static inline int posix_unblock_lock(struct file *filp,
  1109. struct file_lock *waiter)
  1110. {
  1111. return -ENOENT;
  1112. }
  1113. static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
  1114. {
  1115. return 0;
  1116. }
  1117. static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
  1118. struct file_lock *fl, struct file_lock *conf)
  1119. {
  1120. return -ENOLCK;
  1121. }
  1122. static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
  1123. {
  1124. return 0;
  1125. }
  1126. static inline int flock_lock_file_wait(struct file *filp,
  1127. struct file_lock *request)
  1128. {
  1129. return -ENOLCK;
  1130. }
  1131. static inline int __break_lease(struct inode *inode, unsigned int mode)
  1132. {
  1133. return 0;
  1134. }
  1135. static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
  1136. {
  1137. return;
  1138. }
  1139. static inline int generic_setlease(struct file *filp, long arg,
  1140. struct file_lock **flp)
  1141. {
  1142. return -EINVAL;
  1143. }
  1144. static inline int vfs_setlease(struct file *filp, long arg,
  1145. struct file_lock **lease)
  1146. {
  1147. return -EINVAL;
  1148. }
  1149. static inline int lease_modify(struct file_lock **before, int arg)
  1150. {
  1151. return -EINVAL;
  1152. }
  1153. static inline int lock_may_read(struct inode *inode, loff_t start,
  1154. unsigned long len)
  1155. {
  1156. return 1;
  1157. }
  1158. static inline int lock_may_write(struct inode *inode, loff_t start,
  1159. unsigned long len)
  1160. {
  1161. return 1;
  1162. }
  1163. static inline void lock_flocks(void)
  1164. {
  1165. }
  1166. static inline void unlock_flocks(void)
  1167. {
  1168. }
  1169. #endif /* !CONFIG_FILE_LOCKING */
  1170. struct fasync_struct {
  1171. spinlock_t fa_lock;
  1172. int magic;
  1173. int fa_fd;
  1174. struct fasync_struct *fa_next; /* singly linked list */
  1175. struct file *fa_file;
  1176. struct rcu_head fa_rcu;
  1177. };
  1178. #define FASYNC_MAGIC 0x4601
  1179. /* SMP safe fasync helpers: */
  1180. extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
  1181. extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
  1182. extern int fasync_remove_entry(struct file *, struct fasync_struct **);
  1183. extern struct fasync_struct *fasync_alloc(void);
  1184. extern void fasync_free(struct fasync_struct *);
  1185. /* can be called from interrupts */
  1186. extern void kill_fasync(struct fasync_struct **, int, int);
  1187. extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
  1188. extern int f_setown(struct file *filp, unsigned long arg, int force);
  1189. extern void f_delown(struct file *filp);
  1190. extern pid_t f_getown(struct file *filp);
  1191. extern int send_sigurg(struct fown_struct *fown);
  1192. /*
  1193. * Umount options
  1194. */
  1195. #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
  1196. #define MNT_DETACH 0x00000002 /* Just detach from the tree */
  1197. #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
  1198. #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
  1199. #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
  1200. extern struct list_head super_blocks;
  1201. extern spinlock_t sb_lock;
  1202. struct super_block {
  1203. struct list_head s_list; /* Keep this first */
  1204. dev_t s_dev; /* search index; _not_ kdev_t */
  1205. unsigned char s_dirt;
  1206. unsigned char s_blocksize_bits;
  1207. unsigned long s_blocksize;
  1208. loff_t s_maxbytes; /* Max file size */
  1209. struct file_system_type *s_type;
  1210. const struct super_operations *s_op;
  1211. const struct dquot_operations *dq_op;
  1212. const struct quotactl_ops *s_qcop;
  1213. const struct export_operations *s_export_op;
  1214. unsigned long s_flags;
  1215. unsigned long s_magic;
  1216. struct dentry *s_root;
  1217. struct rw_semaphore s_umount;
  1218. struct mutex s_lock;
  1219. int s_count;
  1220. atomic_t s_active;
  1221. #ifdef CONFIG_SECURITY
  1222. void *s_security;
  1223. #endif
  1224. const struct xattr_handler **s_xattr;
  1225. struct list_head s_inodes; /* all inodes */
  1226. struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
  1227. #ifdef CONFIG_SMP
  1228. struct list_head __percpu *s_files;
  1229. #else
  1230. struct list_head s_files;
  1231. #endif
  1232. /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
  1233. struct list_head s_dentry_lru; /* unused dentry lru */
  1234. int s_nr_dentry_unused; /* # of dentry on lru */
  1235. struct block_device *s_bdev;
  1236. struct backing_dev_info *s_bdi;
  1237. struct mtd_info *s_mtd;
  1238. struct list_head s_instances;
  1239. struct quota_info s_dquot; /* Diskquota specific options */
  1240. int s_frozen;
  1241. wait_queue_head_t s_wait_unfrozen;
  1242. char s_id[32]; /* Informational name */
  1243. u8 s_uuid[16]; /* UUID */
  1244. void *s_fs_info; /* Filesystem private info */
  1245. fmode_t s_mode;
  1246. /* Granularity of c/m/atime in ns.
  1247. Cannot be worse than a second */
  1248. u32 s_time_gran;
  1249. /*
  1250. * The next field is for VFS *only*. No filesystems have any business
  1251. * even looking at it. You had been warned.
  1252. */
  1253. struct mutex s_vfs_rename_mutex; /* Kludge */
  1254. /*
  1255. * Filesystem subtype. If non-empty the filesystem type field
  1256. * in /proc/mounts will be "type.subtype"
  1257. */
  1258. char *s_subtype;
  1259. /*
  1260. * Saved mount options for lazy filesystems using
  1261. * generic_show_options()
  1262. */
  1263. char __rcu *s_options;
  1264. const struct dentry_operations *s_d_op; /* default d_op for dentries */
  1265. };
  1266. extern struct timespec current_fs_time(struct super_block *sb);
  1267. /*
  1268. * Snapshotting support.
  1269. */
  1270. enum {
  1271. SB_UNFROZEN = 0,
  1272. SB_FREEZE_WRITE = 1,
  1273. SB_FREEZE_TRANS = 2,
  1274. };
  1275. #define vfs_check_frozen(sb, level) \
  1276. wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
  1277. #define get_fs_excl() atomic_inc(&current->fs_excl)
  1278. #define put_fs_excl() atomic_dec(&current->fs_excl)
  1279. #define has_fs_excl() atomic_read(&current->fs_excl)
  1280. #define is_owner_or_cap(inode) \
  1281. ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
  1282. /* not quite ready to be deprecated, but... */
  1283. extern void lock_super(struct super_block *);
  1284. extern void unlock_super(struct super_block *);
  1285. /*
  1286. * VFS helper functions..
  1287. */
  1288. extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
  1289. extern int vfs_mkdir(struct inode *, struct dentry *, int);
  1290. extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
  1291. extern int vfs_symlink(struct inode *, struct dentry *, const char *);
  1292. extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
  1293. extern int vfs_rmdir(struct inode *, struct dentry *);
  1294. extern int vfs_unlink(struct inode *, struct dentry *);
  1295. extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  1296. /*
  1297. * VFS dentry helper functions.
  1298. */
  1299. extern void dentry_unhash(struct dentry *dentry);
  1300. /*
  1301. * VFS file helper functions.
  1302. */
  1303. extern int file_permission(struct file *, int);
  1304. extern void inode_init_owner(struct inode *inode, const struct inode *dir,
  1305. mode_t mode);
  1306. /*
  1307. * VFS FS_IOC_FIEMAP helper definitions.
  1308. */
  1309. struct fiemap_extent_info {
  1310. unsigned int fi_flags; /* Flags as passed from user */
  1311. unsigned int fi_extents_mapped; /* Number of mapped extents */
  1312. unsigned int fi_extents_max; /* Size of fiemap_extent array */
  1313. struct fiemap_extent __user *fi_extents_start; /* Start of
  1314. fiemap_extent array */
  1315. };
  1316. int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
  1317. u64 phys, u64 len, u32 flags);
  1318. int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
  1319. /*
  1320. * File types
  1321. *
  1322. * NOTE! These match bits 12..15 of stat.st_mode
  1323. * (ie "(i_mode >> 12) & 15").
  1324. */
  1325. #define DT_UNKNOWN 0
  1326. #define DT_FIFO 1
  1327. #define DT_CHR 2
  1328. #define DT_DIR 4
  1329. #define DT_BLK 6
  1330. #define DT_REG 8
  1331. #define DT_LNK 10
  1332. #define DT_SOCK 12
  1333. #define DT_WHT 14
  1334. /*
  1335. * This is the "filldir" function type, used by readdir() to let
  1336. * the kernel specify what kind of dirent layout it wants to have.
  1337. * This allows the kernel to read directories into kernel space or
  1338. * to have different dirent layouts depending on the binary type.
  1339. */
  1340. typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
  1341. struct block_device_operations;
  1342. /* These macros are for out of kernel modules to test that
  1343. * the kernel supports the unlocked_ioctl and compat_ioctl
  1344. * fields in struct file_operations. */
  1345. #define HAVE_COMPAT_IOCTL 1
  1346. #define HAVE_UNLOCKED_IOCTL 1
  1347. /*
  1348. * NOTE:
  1349. * all file operations except setlease can be called without
  1350. * the big kernel lock held in all filesystems.
  1351. */
  1352. struct file_operations {
  1353. struct module *owner;
  1354. loff_t (*llseek) (struct file *, loff_t, int);
  1355. ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
  1356. ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
  1357. ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1358. ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
  1359. int (*readdir) (struct file *, void *, filldir_t);
  1360. unsigned int (*poll) (struct file *, struct poll_table_struct *);
  1361. long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
  1362. long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
  1363. int (*mmap) (struct file *, struct vm_area_struct *);
  1364. int (*open) (struct inode *, struct file *);
  1365. int (*flush) (struct file *, fl_owner_t id);
  1366. int (*release) (struct inode *, struct file *);
  1367. int (*fsync) (struct file *, int datasync);
  1368. int (*aio_fsync) (struct kiocb *, int datasync);
  1369. int (*fasync) (int, struct file *, int);
  1370. int (*lock) (struct file *, int, struct file_lock *);
  1371. ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
  1372. unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
  1373. int (*check_flags)(int);
  1374. int (*flock) (struct file *, int, struct file_lock *);
  1375. ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
  1376. ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
  1377. int (*setlease)(struct file *, long, struct file_lock **);
  1378. long (*fallocate)(struct file *file, int mode, loff_t offset,
  1379. loff_t len);
  1380. };
  1381. #define IPERM_FLAG_RCU 0x0001
  1382. struct inode_operations {
  1383. struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
  1384. void * (*follow_link) (struct dentry *, struct nameidata *);
  1385. int (*permission) (struct inode *, int, unsigned int);
  1386. int (*check_acl)(struct inode *, int, unsigned int);
  1387. int (*readlink) (struct dentry *, char __user *,int);
  1388. void (*put_link) (struct dentry *, struct nameidata *, void *);
  1389. int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
  1390. int (*link) (struct dentry *,struct inode *,struct dentry *);
  1391. int (*unlink) (struct inode *,struct dentry *);
  1392. int (*symlink) (struct inode *,struct dentry *,const char *);
  1393. int (*mkdir) (struct inode *,struct dentry *,int);
  1394. int (*rmdir) (struct inode *,struct dentry *);
  1395. int (*mknod) (struct inode *,struct dentry *,int,dev_t);
  1396. int (*rename) (struct inode *, struct dentry *,
  1397. struct inode *, struct dentry *);
  1398. void (*truncate) (struct inode *);
  1399. int (*setattr) (struct dentry *, struct iattr *);
  1400. int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
  1401. int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
  1402. ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
  1403. ssize_t (*listxattr) (struct dentry *, char *, size_t);
  1404. int (*removexattr) (struct dentry *, const char *);
  1405. void (*truncate_range)(struct inode *, loff_t, loff_t);
  1406. int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
  1407. u64 len);
  1408. } ____cacheline_aligned;
  1409. struct seq_file;
  1410. ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
  1411. unsigned long nr_segs, unsigned long fast_segs,
  1412. struct iovec *fast_pointer,
  1413. struct iovec **ret_pointer);
  1414. extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
  1415. extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
  1416. extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
  1417. unsigned long, loff_t *);
  1418. extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
  1419. unsigned long, loff_t *);
  1420. struct super_operations {
  1421. struct inode *(*alloc_inode)(struct super_block *sb);
  1422. void (*destroy_inode)(struct inode *);
  1423. void (*dirty_inode) (struct inode *);
  1424. int (*write_inode) (struct inode *, struct writeback_control *wbc);
  1425. int (*drop_inode) (struct inode *);
  1426. void (*evict_inode) (struct inode *);
  1427. void (*put_super) (struct super_block *);
  1428. void (*write_super) (struct super_block *);
  1429. int (*sync_fs)(struct super_block *sb, int wait);
  1430. int (*freeze_fs) (struct super_block *);
  1431. int (*unfreeze_fs) (struct super_block *);
  1432. int (*statfs) (struct dentry *, struct kstatfs *);
  1433. int (*remount_fs) (struct super_block *, int *, char *);
  1434. void (*umount_begin) (struct super_block *);
  1435. int (*show_options)(struct seq_file *, struct vfsmount *);
  1436. int (*show_stats)(struct seq_file *, struct vfsmount *);
  1437. #ifdef CONFIG_QUOTA
  1438. ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
  1439. ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
  1440. #endif
  1441. int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
  1442. };
  1443. /*
  1444. * Inode state bits. Protected by inode_lock.
  1445. *
  1446. * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
  1447. * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
  1448. *
  1449. * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
  1450. * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
  1451. * various stages of removing an inode.
  1452. *
  1453. * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
  1454. *
  1455. * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
  1456. * fdatasync(). i_atime is the usual cause.
  1457. * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
  1458. * these changes separately from I_DIRTY_SYNC so that we
  1459. * don't have to write inode on fdatasync() when only
  1460. * mtime has changed in it.
  1461. * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
  1462. * I_NEW Serves as both a mutex and completion notification.
  1463. * New inodes set I_NEW. If two processes both create
  1464. * the same inode, one of them will release its inode and
  1465. * wait for I_NEW to be released before returning.
  1466. * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
  1467. * also cause waiting on I_NEW, without I_NEW actually
  1468. * being set. find_inode() uses this to prevent returning
  1469. * nearly-dead inodes.
  1470. * I_WILL_FREE Must be set when calling write_inode_now() if i_count
  1471. * is zero. I_FREEING must be set when I_WILL_FREE is
  1472. * cleared.
  1473. * I_FREEING Set when inode is about to be freed but still has dirty
  1474. * pages or buffers attached or the inode itself is still
  1475. * dirty.
  1476. * I_CLEAR Added by end_writeback(). In this state the inode is clean
  1477. * and can be destroyed. Inode keeps I_FREEING.
  1478. *
  1479. * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
  1480. * prohibited for many purposes. iget() must wait for
  1481. * the inode to be completely released, then create it
  1482. * anew. Other functions will just ignore such inodes,
  1483. * if appropriate. I_NEW is used for waiting.
  1484. *
  1485. * I_SYNC Synchonized write of dirty inode data. The bits is
  1486. * set during data writeback, and cleared with a wakeup
  1487. * on the bit address once it is done.
  1488. *
  1489. * Q: What is the difference between I_WILL_FREE and I_FREEING?
  1490. */
  1491. #define I_DIRTY_SYNC (1 << 0)
  1492. #define I_DIRTY_DATASYNC (1 << 1)
  1493. #define I_DIRTY_PAGES (1 << 2)
  1494. #define __I_NEW 3
  1495. #define I_NEW (1 << __I_NEW)
  1496. #define I_WILL_FREE (1 << 4)
  1497. #define I_FREEING (1 << 5)
  1498. #define I_CLEAR (1 << 6)
  1499. #define __I_SYNC 7
  1500. #define I_SYNC (1 << __I_SYNC)
  1501. #define I_REFERENCED (1 << 8)
  1502. #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
  1503. extern void __mark_inode_dirty(struct inode *, int);
  1504. static inline void mark_inode_dirty(struct inode *inode)
  1505. {
  1506. __mark_inode_dirty(inode, I_DIRTY);
  1507. }
  1508. static inline void mark_inode_dirty_sync(struct inode *inode)
  1509. {
  1510. __mark_inode_dirty(inode, I_DIRTY_SYNC);
  1511. }
  1512. /**
  1513. * inc_nlink - directly increment an inode's link count
  1514. * @inode: inode
  1515. *
  1516. * This is a low-level filesystem helper to replace any
  1517. * direct filesystem manipulation of i_nlink. Currently,
  1518. * it is only here for parity with dec_nlink().
  1519. */
  1520. static inline void inc_nlink(struct inode *inode)
  1521. {
  1522. inode->i_nlink++;
  1523. }
  1524. static inline void inode_inc_link_count(struct inode *inode)
  1525. {
  1526. inc_nlink(inode);
  1527. mark_inode_dirty(inode);
  1528. }
  1529. /**
  1530. * drop_nlink - directly drop an inode's link count
  1531. * @inode: inode
  1532. *
  1533. * This is a low-level filesystem helper to replace any
  1534. * direct filesystem manipulation of i_nlink. In cases
  1535. * where we are attempting to track writes to the
  1536. * filesystem, a decrement to zero means an imminent
  1537. * write when the file is truncated and actually unlinked
  1538. * on the filesystem.
  1539. */
  1540. static inline void drop_nlink(struct inode *inode)
  1541. {
  1542. inode->i_nlink--;
  1543. }
  1544. /**
  1545. * clear_nlink - directly zero an inode's link count
  1546. * @inode: inode
  1547. *
  1548. * This is a low-level filesystem helper to replace any
  1549. * direct filesystem manipulation of i_nlink. See
  1550. * drop_nlink() for why we care about i_nlink hitting zero.
  1551. */
  1552. static inline void clear_nlink(struct inode *inode)
  1553. {
  1554. inode->i_nlink = 0;
  1555. }
  1556. static inline void inode_dec_link_count(struct inode *inode)
  1557. {
  1558. drop_nlink(inode);
  1559. mark_inode_dirty(inode);
  1560. }
  1561. /**
  1562. * inode_inc_iversion - increments i_version
  1563. * @inode: inode that need to be updated
  1564. *
  1565. * Every time the inode is modified, the i_version field will be incremented.
  1566. * The filesystem has to be mounted with i_version flag
  1567. */
  1568. static inline void inode_inc_iversion(struct inode *inode)
  1569. {
  1570. spin_lock(&inode->i_lock);
  1571. inode->i_version++;
  1572. spin_unlock(&inode->i_lock);
  1573. }
  1574. extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
  1575. static inline void file_accessed(struct file *file)
  1576. {
  1577. if (!(file->f_flags & O_NOATIME))
  1578. touch_atime(file->f_path.mnt, file->f_path.dentry);
  1579. }
  1580. int sync_inode(struct inode *inode, struct writeback_control *wbc);
  1581. int sync_inode_metadata(struct inode *inode, int wait);
  1582. struct file_system_type {
  1583. const char *name;
  1584. int fs_flags;
  1585. int (*get_sb) (struct file_system_type *, int,
  1586. const char *, void *, struct vfsmount *);
  1587. struct dentry *(*mount) (struct file_system_type *, int,
  1588. const char *, void *);
  1589. void (*kill_sb) (struct super_block *);
  1590. struct module *owner;
  1591. struct file_system_type * next;
  1592. struct list_head fs_supers;
  1593. struct lock_class_key s_lock_key;
  1594. struct lock_class_key s_umount_key;
  1595. struct lock_class_key s_vfs_rename_key;
  1596. struct lock_class_key i_lock_key;
  1597. struct lock_class_key i_mutex_key;
  1598. struct lock_class_key i_mutex_dir_key;
  1599. struct lock_class_key i_alloc_sem_key;
  1600. };
  1601. extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
  1602. void *data, int (*fill_super)(struct super_block *, void *, int));
  1603. extern struct dentry *mount_bdev(struct file_system_type *fs_type,
  1604. int flags, const char *dev_name, void *data,
  1605. int (*fill_super)(struct super_block *, void *, int));
  1606. extern int get_sb_bdev(struct file_system_type *fs_type,
  1607. int flags, const char *dev_name, void *data,
  1608. int (*fill_super)(struct super_block *, void *, int),
  1609. struct vfsmount *mnt);
  1610. extern struct dentry *mount_single(struct file_system_type *fs_type,
  1611. int flags, void *data,
  1612. int (*fill_super)(struct super_block *, void *, int));
  1613. extern int get_sb_single(struct file_system_type *fs_type,
  1614. int flags, void *data,
  1615. int (*fill_super)(struct super_block *, void *, int),
  1616. struct vfsmount *mnt);
  1617. extern struct dentry *mount_nodev(struct file_system_type *fs_type,
  1618. int flags, void *data,
  1619. int (*fill_super)(struct super_block *, void *, int));
  1620. extern int get_sb_nodev(struct file_system_type *fs_type,
  1621. int flags, void *data,
  1622. int (*fill_super)(struct super_block *, void *, int),
  1623. struct vfsmount *mnt);
  1624. void generic_shutdown_super(struct super_block *sb);
  1625. void kill_block_super(struct super_block *sb);
  1626. void kill_anon_super(struct super_block *sb);
  1627. void kill_litter_super(struct super_block *sb);
  1628. void deactivate_super(struct super_block *sb);
  1629. void deactivate_locked_super(struct super_block *sb);
  1630. int set_anon_super(struct super_block *s, void *data);
  1631. struct super_block *sget(struct file_system_type *type,
  1632. int (*test)(struct super_block *,void *),
  1633. int (*set)(struct super_block *,void *),
  1634. void *data);
  1635. extern struct dentry *mount_pseudo(struct file_system_type *, char *,
  1636. const struct super_operations *ops,
  1637. const struct dentry_operations *dops,
  1638. unsigned long);
  1639. extern void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb);
  1640. static inline void sb_mark_dirty(struct super_block *sb)
  1641. {
  1642. sb->s_dirt = 1;
  1643. }
  1644. static inline void sb_mark_clean(struct super_block *sb)
  1645. {
  1646. sb->s_dirt = 0;
  1647. }
  1648. static inline int sb_is_dirty(struct super_block *sb)
  1649. {
  1650. return sb->s_dirt;
  1651. }
  1652. /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
  1653. #define fops_get(fops) \
  1654. (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
  1655. #define fops_put(fops) \
  1656. do { if (fops) module_put((fops)->owner); } while(0)
  1657. extern int register_filesystem(struct file_system_type *);
  1658. extern int unregister_filesystem(struct file_system_type *);
  1659. extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
  1660. #define kern_mount(type) kern_mount_data(type, NULL)
  1661. extern int may_umount_tree(struct vfsmount *);
  1662. extern int may_umount(struct vfsmount *);
  1663. extern long do_mount(char *, char *, char *, unsigned long, void *);
  1664. extern struct vfsmount *collect_mounts(struct path *);
  1665. extern void drop_collected_mounts(struct vfsmount *);
  1666. extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
  1667. struct vfsmount *);
  1668. extern int vfs_statfs(struct path *, struct kstatfs *);
  1669. extern int user_statfs(const char __user *, struct kstatfs *);
  1670. extern int fd_statfs(int, struct kstatfs *);
  1671. extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
  1672. extern int freeze_super(struct super_block *super);
  1673. extern int thaw_super(struct super_block *super);
  1674. extern int current_umask(void);
  1675. /* /sys/fs */
  1676. extern struct kobject *fs_kobj;
  1677. #define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
  1678. extern int rw_verify_area(int, struct file *, loff_t *, size_t);
  1679. #define FLOCK_VERIFY_READ 1
  1680. #define FLOCK_VERIFY_WRITE 2
  1681. #ifdef CONFIG_FILE_LOCKING
  1682. extern int locks_mandatory_locked(struct inode *);
  1683. extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
  1684. /*
  1685. * Candidates for mandatory locking have the setgid bit set
  1686. * but no group execute bit - an otherwise meaningless combination.
  1687. */
  1688. static inline int __mandatory_lock(struct inode *ino)
  1689. {
  1690. return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
  1691. }
  1692. /*
  1693. * ... and these candidates should be on MS_MANDLOCK mounted fs,
  1694. * otherwise these will be advisory locks
  1695. */
  1696. static inline int mandatory_lock(struct inode *ino)
  1697. {
  1698. return IS_MANDLOCK(ino) && __mandatory_lock(ino);
  1699. }
  1700. static inline int locks_verify_locked(struct inode *inode)
  1701. {
  1702. if (mandatory_lock(inode))
  1703. return locks_mandatory_locked(inode);
  1704. return 0;
  1705. }
  1706. static inline int locks_verify_truncate(struct inode *inode,
  1707. struct file *filp,
  1708. loff_t size)
  1709. {
  1710. if (inode->i_flock && mandatory_lock(inode))
  1711. return locks_mandatory_area(
  1712. FLOCK_VERIFY_WRITE, inode, filp,
  1713. size < inode->i_size ? size : inode->i_size,
  1714. (size < inode->i_size ? inode->i_size - size
  1715. : size - inode->i_size)
  1716. );
  1717. return 0;
  1718. }
  1719. static inline int break_lease(struct inode *inode, unsigned int mode)
  1720. {
  1721. if (inode->i_flock)
  1722. return __break_lease(inode, mode);
  1723. return 0;
  1724. }
  1725. #else /* !CONFIG_FILE_LOCKING */
  1726. static inline int locks_mandatory_locked(struct inode *inode)
  1727. {
  1728. return 0;
  1729. }
  1730. static inline int locks_mandatory_area(int rw, struct inode *inode,
  1731. struct file *filp, loff_t offset,
  1732. size_t count)
  1733. {
  1734. return 0;
  1735. }
  1736. static inline int __mandatory_lock(struct inode *inode)
  1737. {
  1738. return 0;
  1739. }
  1740. static inline int mandatory_lock(struct inode *inode)
  1741. {
  1742. return 0;
  1743. }
  1744. static inline int locks_verify_locked(struct inode *inode)
  1745. {
  1746. return 0;
  1747. }
  1748. static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
  1749. size_t size)
  1750. {
  1751. return 0;
  1752. }
  1753. static inline int break_lease(struct inode *inode, unsigned int mode)
  1754. {
  1755. return 0;
  1756. }
  1757. #endif /* CONFIG_FILE_LOCKING */
  1758. /* fs/open.c */
  1759. extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
  1760. struct file *filp);
  1761. extern int do_fallocate(struct file *file, int mode, loff_t offset,
  1762. loff_t len);
  1763. extern long do_sys_open(int dfd, const char __user *filename, int flags,
  1764. int mode);
  1765. extern struct file *filp_open(const char *, int, int);
  1766. extern struct file *file_open_root(struct dentry *, struct vfsmount *,
  1767. const char *, int);
  1768. extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
  1769. const struct cred *);
  1770. extern int filp_close(struct file *, fl_owner_t id);
  1771. extern char * getname(const char __user *);
  1772. /* fs/ioctl.c */
  1773. extern int ioctl_preallocate(struct file *filp, void __user *argp);
  1774. /* fs/dcache.c */
  1775. extern void __init vfs_caches_init_early(void);
  1776. extern void __init vfs_caches_init(unsigned long);
  1777. extern struct kmem_cache *names_cachep;
  1778. #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
  1779. #define __getname() __getname_gfp(GFP_KERNEL)
  1780. #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
  1781. #ifndef CONFIG_AUDITSYSCALL
  1782. #define putname(name) __putname(name)
  1783. #else
  1784. extern void putname(const char *name);
  1785. #endif
  1786. #ifdef CONFIG_BLOCK
  1787. extern int register_blkdev(unsigned int, const char *);
  1788. extern void unregister_blkdev(unsigned int, const char *);
  1789. extern struct block_device *bdget(dev_t);
  1790. extern struct block_device *bdgrab(struct block_device *bdev);
  1791. extern void bd_set_size(struct block_device *, loff_t size);
  1792. extern void bd_forget(struct inode *inode);
  1793. extern void bdput(struct block_device *);
  1794. extern void invalidate_bdev(struct block_device *);
  1795. extern int sync_blockdev(struct block_device *bdev);
  1796. extern struct super_block *freeze_bdev(struct block_device *);
  1797. extern void emergency_thaw_all(void);
  1798. extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
  1799. extern int fsync_bdev(struct block_device *);
  1800. #else
  1801. static inline void bd_forget(struct inode *inode) {}
  1802. static inline int sync_blockdev(struct block_device *bdev) { return 0; }
  1803. static inline void invalidate_bdev(struct block_device *bdev) {}
  1804. static inline struct super_block *freeze_bdev(struct block_device *sb)
  1805. {
  1806. return NULL;
  1807. }
  1808. static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
  1809. {
  1810. return 0;
  1811. }
  1812. #endif
  1813. extern int sync_filesystem(struct super_block *);
  1814. extern const struct file_operations def_blk_fops;
  1815. extern const struct file_operations def_chr_fops;
  1816. extern const struct file_operations bad_sock_fops;
  1817. extern const struct file_operations def_fifo_fops;
  1818. #ifdef CONFIG_BLOCK
  1819. extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
  1820. extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
  1821. extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
  1822. extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
  1823. extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
  1824. void *holder);
  1825. extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
  1826. void *holder);
  1827. extern int blkdev_put(struct block_device *bdev, fmode_t mode);
  1828. #ifdef CONFIG_SYSFS
  1829. extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
  1830. extern void bd_unlink_disk_holder(struct block_device *bdev,
  1831. struct gendisk *disk);
  1832. #else
  1833. static inline int bd_link_disk_holder(struct block_device *bdev,
  1834. struct gendisk *disk)
  1835. {
  1836. return 0;
  1837. }
  1838. static inline void bd_unlink_disk_holder(struct block_device *bdev,
  1839. struct gendisk *disk)
  1840. {
  1841. }
  1842. #endif
  1843. #endif
  1844. /* fs/char_dev.c */
  1845. #define CHRDEV_MAJOR_HASH_SIZE 255
  1846. extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
  1847. extern int register_chrdev_region(dev_t, unsigned, const char *);
  1848. extern int __register_chrdev(unsigned int major, unsigned int baseminor,
  1849. unsigned int count, const char *name,
  1850. const struct file_operations *fops);
  1851. extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
  1852. unsigned int count, const char *name);
  1853. extern void unregister_chrdev_region(dev_t, unsigned);
  1854. extern void chrdev_show(struct seq_file *,off_t);
  1855. static inline int register_chrdev(unsigned int major, const char *name,
  1856. const struct file_operations *fops)
  1857. {
  1858. return __register_chrdev(major, 0, 256, name, fops);
  1859. }
  1860. static inline void unregister_chrdev(unsigned int major, const char *name)
  1861. {
  1862. __unregister_chrdev(major, 0, 256, name);
  1863. }
  1864. /* fs/block_dev.c */
  1865. #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
  1866. #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
  1867. #ifdef CONFIG_BLOCK
  1868. #define BLKDEV_MAJOR_HASH_SIZE 255
  1869. extern const char *__bdevname(dev_t, char *buffer);
  1870. extern const char *bdevname(struct block_device *bdev, char *buffer);
  1871. extern struct block_device *lookup_bdev(const char *);
  1872. extern void blkdev_show(struct seq_file *,off_t);
  1873. #else
  1874. #define BLKDEV_MAJOR_HASH_SIZE 0
  1875. #endif
  1876. extern void init_special_inode(struct inode *, umode_t, dev_t);
  1877. /* Invalid inode operations -- fs/bad_inode.c */
  1878. extern void make_bad_inode(struct inode *);
  1879. extern int is_bad_inode(struct inode *);
  1880. extern const struct file_operations read_pipefifo_fops;
  1881. extern const struct file_operations write_pipefifo_fops;
  1882. extern const struct file_operations rdwr_pipefifo_fops;
  1883. extern int fs_may_remount_ro(struct super_block *);
  1884. #ifdef CONFIG_BLOCK
  1885. /*
  1886. * return READ, READA, or WRITE
  1887. */
  1888. #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
  1889. /*
  1890. * return data direction, READ or WRITE
  1891. */
  1892. #define bio_data_dir(bio) ((bio)->bi_rw & 1)
  1893. extern void check_disk_size_change(struct gendisk *disk,
  1894. struct block_device *bdev);
  1895. extern int revalidate_disk(struct gendisk *);
  1896. extern int check_disk_change(struct block_device *);
  1897. extern int __invalidate_device(struct block_device *, bool);
  1898. extern int invalidate_partition(struct gendisk *, int);
  1899. #endif
  1900. unsigned long invalidate_mapping_pages(struct address_space *mapping,
  1901. pgoff_t start, pgoff_t end);
  1902. static inline void invalidate_remote_inode(struct inode *inode)
  1903. {
  1904. if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
  1905. S_ISLNK(inode->i_mode))
  1906. invalidate_mapping_pages(inode->i_mapping, 0, -1);
  1907. }
  1908. extern int invalidate_inode_pages2(struct address_space *mapping);
  1909. extern int invalidate_inode_pages2_range(struct address_space *mapping,
  1910. pgoff_t start, pgoff_t end);
  1911. extern int write_inode_now(struct inode *, int);
  1912. extern int filemap_fdatawrite(struct address_space *);
  1913. extern int filemap_flush(struct address_space *);
  1914. extern int filemap_fdatawait(struct address_space *);
  1915. extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
  1916. loff_t lend);
  1917. extern int filemap_write_and_wait(struct address_space *mapping);
  1918. extern int filemap_write_and_wait_range(struct address_space *mapping,
  1919. loff_t lstart, loff_t lend);
  1920. extern int __filemap_fdatawrite_range(struct address_space *mapping,
  1921. loff_t start, loff_t end, int sync_mode);
  1922. extern int filemap_fdatawrite_range(struct address_space *mapping,
  1923. loff_t start, loff_t end);
  1924. extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
  1925. int datasync);
  1926. extern int vfs_fsync(struct file *file, int datasync);
  1927. extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
  1928. extern void sync_supers(void);
  1929. extern void emergency_sync(void);
  1930. extern void emergency_remount(void);
  1931. #ifdef CONFIG_BLOCK
  1932. extern sector_t bmap(struct inode *, sector_t);
  1933. #endif
  1934. extern int notify_change(struct dentry *, struct iattr *);
  1935. extern int inode_permission(struct inode *, int);
  1936. extern int generic_permission(struct inode *, int, unsigned int,
  1937. int (*check_acl)(struct inode *, int, unsigned int));
  1938. static inline bool execute_ok(struct inode *inode)
  1939. {
  1940. return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
  1941. }
  1942. extern int get_write_access(struct inode *);
  1943. extern int deny_write_access(struct file *);
  1944. static inline void put_write_access(struct inode * inode)
  1945. {
  1946. atomic_dec(&inode->i_writecount);
  1947. }
  1948. static inline void allow_write_access(struct file *file)
  1949. {
  1950. if (file)
  1951. atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
  1952. }
  1953. #ifdef CONFIG_IMA
  1954. static inline void i_readcount_dec(struct inode *inode)
  1955. {
  1956. BUG_ON(!atomic_read(&inode->i_readcount));
  1957. atomic_dec(&inode->i_readcount);
  1958. }
  1959. static inline void i_readcount_inc(struct inode *inode)
  1960. {
  1961. atomic_inc(&inode->i_readcount);
  1962. }
  1963. #else
  1964. static inline void i_readcount_dec(struct inode *inode)
  1965. {
  1966. return;
  1967. }
  1968. static inline void i_readcount_inc(struct inode *inode)
  1969. {
  1970. return;
  1971. }
  1972. #endif
  1973. extern int do_pipe_flags(int *, int);
  1974. extern struct file *create_read_pipe(struct file *f, int flags);
  1975. extern struct file *create_write_pipe(int flags);
  1976. extern void free_write_pipe(struct file *);
  1977. extern int kernel_read(struct file *, loff_t, char *, unsigned long);
  1978. extern struct file * open_exec(const char *);
  1979. /* fs/dcache.c -- generic fs support functions */
  1980. extern int is_subdir(struct dentry *, struct dentry *);
  1981. extern int path_is_under(struct path *, struct path *);
  1982. extern ino_t find_inode_number(struct dentry *, struct qstr *);
  1983. #include <linux/err.h>
  1984. /* needed for stackable file system support */
  1985. extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
  1986. extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
  1987. extern int inode_init_always(struct super_block *, struct inode *);
  1988. extern void inode_init_once(struct inode *);
  1989. extern void address_space_init_once(struct address_space *mapping);
  1990. extern void ihold(struct inode * inode);
  1991. extern void iput(struct inode *);
  1992. extern struct inode * igrab(struct inode *);
  1993. extern ino_t iunique(struct super_block *, ino_t);
  1994. extern int inode_needs_sync(struct inode *inode);
  1995. extern int generic_delete_inode(struct inode *inode);
  1996. extern int generic_drop_inode(struct inode *inode);
  1997. extern struct inode *ilookup5_nowait(struct super_block *sb,
  1998. unsigned long hashval, int (*test)(struct inode *, void *),
  1999. void *data);
  2000. extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
  2001. int (*test)(struct inode *, void *), void *data);
  2002. extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
  2003. extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
  2004. extern struct inode * iget_locked(struct super_block *, unsigned long);
  2005. extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
  2006. extern int insert_inode_locked(struct inode *);
  2007. extern void unlock_new_inode(struct inode *);
  2008. extern unsigned int get_next_ino(void);
  2009. extern void __iget(struct inode * inode);
  2010. extern void iget_failed(struct inode *);
  2011. extern void end_writeback(struct inode *);
  2012. extern void __destroy_inode(struct inode *);
  2013. extern struct inode *new_inode(struct super_block *);
  2014. extern void free_inode_nonrcu(struct inode *inode);
  2015. extern int should_remove_suid(struct dentry *);
  2016. extern int file_remove_suid(struct file *);
  2017. extern void __insert_inode_hash(struct inode *, unsigned long hashval);
  2018. extern void remove_inode_hash(struct inode *);
  2019. static inline void insert_inode_hash(struct inode *inode)
  2020. {
  2021. __insert_inode_hash(inode, inode->i_ino);
  2022. }
  2023. extern void inode_sb_list_add(struct inode *inode);
  2024. #ifdef CONFIG_BLOCK
  2025. extern void submit_bio(int, struct bio *);
  2026. extern int bdev_read_only(struct block_device *);
  2027. #endif
  2028. extern int set_blocksize(struct block_device *, int);
  2029. extern int sb_set_blocksize(struct super_block *, int);
  2030. extern int sb_min_blocksize(struct super_block *, int);
  2031. extern int generic_file_mmap(struct file *, struct vm_area_struct *);
  2032. extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
  2033. extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
  2034. int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
  2035. extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2036. extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
  2037. loff_t *);
  2038. extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
  2039. extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
  2040. unsigned long *, loff_t, loff_t *, size_t, size_t);
  2041. extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
  2042. unsigned long, loff_t, loff_t *, size_t, ssize_t);
  2043. extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
  2044. extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
  2045. extern int generic_segment_checks(const struct iovec *iov,
  2046. unsigned long *nr_segs, size_t *count, int access_flags);
  2047. /* fs/block_dev.c */
  2048. extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
  2049. unsigned long nr_segs, loff_t pos);
  2050. extern int blkdev_fsync(struct file *filp, int datasync);
  2051. /* fs/splice.c */
  2052. extern ssize_t generic_file_splice_read(struct file *, loff_t *,
  2053. struct pipe_inode_info *, size_t, unsigned int);
  2054. extern ssize_t default_file_splice_read(struct file *, loff_t *,
  2055. struct pipe_inode_info *, size_t, unsigned int);
  2056. extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
  2057. struct file *, loff_t *, size_t, unsigned int);
  2058. extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
  2059. struct file *out, loff_t *, size_t len, unsigned int flags);
  2060. extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
  2061. size_t len, unsigned int flags);
  2062. extern void
  2063. file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
  2064. extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
  2065. extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
  2066. extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
  2067. extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
  2068. int origin);
  2069. extern int generic_file_open(struct inode * inode, struct file * filp);
  2070. extern int nonseekable_open(struct inode * inode, struct file * filp);
  2071. #ifdef CONFIG_FS_XIP
  2072. extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
  2073. loff_t *ppos);
  2074. extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
  2075. extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
  2076. size_t len, loff_t *ppos);
  2077. extern int xip_truncate_page(struct address_space *mapping, loff_t from);
  2078. #else
  2079. static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
  2080. {
  2081. return 0;
  2082. }
  2083. #endif
  2084. #ifdef CONFIG_BLOCK
  2085. typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
  2086. loff_t file_offset);
  2087. enum {
  2088. /* need locking between buffered and direct access */
  2089. DIO_LOCKING = 0x01,
  2090. /* filesystem does not support filling holes */
  2091. DIO_SKIP_HOLES = 0x02,
  2092. };
  2093. void dio_end_io(struct bio *bio, int error);
  2094. ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
  2095. struct block_device *bdev, const struct iovec *iov, loff_t offset,
  2096. unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
  2097. dio_submit_t submit_io, int flags);
  2098. static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
  2099. struct inode *inode, struct block_device *bdev, const struct iovec *iov,
  2100. loff_t offset, unsigned long nr_segs, get_block_t get_block,
  2101. dio_iodone_t end_io)
  2102. {
  2103. return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
  2104. nr_segs, get_block, end_io, NULL,
  2105. DIO_LOCKING | DIO_SKIP_HOLES);
  2106. }
  2107. #endif
  2108. extern const struct file_operations generic_ro_fops;
  2109. #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
  2110. extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
  2111. extern int vfs_follow_link(struct nameidata *, const char *);
  2112. extern int page_readlink(struct dentry *, char __user *, int);
  2113. extern void *page_follow_link_light(struct dentry *, struct nameidata *);
  2114. extern void page_put_link(struct dentry *, struct nameidata *, void *);
  2115. extern int __page_symlink(struct inode *inode, const char *symname, int len,
  2116. int nofs);
  2117. extern int page_symlink(struct inode *inode, const char *symname, int len);
  2118. extern const struct inode_operations page_symlink_inode_operations;
  2119. extern int generic_readlink(struct dentry *, char __user *, int);
  2120. extern void generic_fillattr(struct inode *, struct kstat *);
  2121. extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2122. void __inode_add_bytes(struct inode *inode, loff_t bytes);
  2123. void inode_add_bytes(struct inode *inode, loff_t bytes);
  2124. void inode_sub_bytes(struct inode *inode, loff_t bytes);
  2125. loff_t inode_get_bytes(struct inode *inode);
  2126. void inode_set_bytes(struct inode *inode, loff_t bytes);
  2127. extern int vfs_readdir(struct file *, filldir_t, void *);
  2128. extern int vfs_stat(const char __user *, struct kstat *);
  2129. extern int vfs_lstat(const char __user *, struct kstat *);
  2130. extern int vfs_fstat(unsigned int, struct kstat *);
  2131. extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
  2132. extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
  2133. unsigned long arg);
  2134. extern int __generic_block_fiemap(struct inode *inode,
  2135. struct fiemap_extent_info *fieinfo,
  2136. loff_t start, loff_t len,
  2137. get_block_t *get_block);
  2138. extern int generic_block_fiemap(struct inode *inode,
  2139. struct fiemap_extent_info *fieinfo, u64 start,
  2140. u64 len, get_block_t *get_block);
  2141. extern void get_filesystem(struct file_system_type *fs);
  2142. extern void put_filesystem(struct file_system_type *fs);
  2143. extern struct file_system_type *get_fs_type(const char *name);
  2144. extern struct super_block *get_super(struct block_device *);
  2145. extern struct super_block *get_active_super(struct block_device *bdev);
  2146. extern struct super_block *user_get_super(dev_t);
  2147. extern void drop_super(struct super_block *sb);
  2148. extern void iterate_supers(void (*)(struct super_block *, void *), void *);
  2149. extern int dcache_dir_open(struct inode *, struct file *);
  2150. extern int dcache_dir_close(struct inode *, struct file *);
  2151. extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
  2152. extern int dcache_readdir(struct file *, void *, filldir_t);
  2153. extern int simple_setattr(struct dentry *, struct iattr *);
  2154. extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
  2155. extern int simple_statfs(struct dentry *, struct kstatfs *);
  2156. extern int simple_link(struct dentry *, struct inode *, struct dentry *);
  2157. extern int simple_unlink(struct inode *, struct dentry *);
  2158. extern int simple_rmdir(struct inode *, struct dentry *);
  2159. extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
  2160. extern int noop_fsync(struct file *, int);
  2161. extern int simple_empty(struct dentry *);
  2162. extern int simple_readpage(struct file *file, struct page *page);
  2163. extern int simple_write_begin(struct file *file, struct address_space *mapping,
  2164. loff_t pos, unsigned len, unsigned flags,
  2165. struct page **pagep, void **fsdata);
  2166. extern int simple_write_end(struct file *file, struct address_space *mapping,
  2167. loff_t pos, unsigned len, unsigned copied,
  2168. struct page *page, void *fsdata);
  2169. extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
  2170. extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
  2171. extern const struct file_operations simple_dir_operations;
  2172. extern const struct inode_operations simple_dir_inode_operations;
  2173. struct tree_descr { char *name; const struct file_operations *ops; int mode; };
  2174. struct dentry *d_alloc_name(struct dentry *, const char *);
  2175. extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
  2176. extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
  2177. extern void simple_release_fs(struct vfsmount **mount, int *count);
  2178. extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
  2179. loff_t *ppos, const void *from, size_t available);
  2180. extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
  2181. const void __user *from, size_t count);
  2182. extern int generic_file_fsync(struct file *, int);
  2183. extern int generic_check_addressable(unsigned, u64);
  2184. #ifdef CONFIG_MIGRATION
  2185. extern int buffer_migrate_page(struct address_space *,
  2186. struct page *, struct page *);
  2187. #else
  2188. #define buffer_migrate_page NULL
  2189. #endif
  2190. extern int inode_change_ok(const struct inode *, struct iattr *);
  2191. extern int inode_newsize_ok(const struct inode *, loff_t offset);
  2192. extern void setattr_copy(struct inode *inode, const struct iattr *attr);
  2193. extern void file_update_time(struct file *file);
  2194. extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
  2195. extern void save_mount_options(struct super_block *sb, char *options);
  2196. extern void replace_mount_options(struct super_block *sb, char *options);
  2197. static inline ino_t parent_ino(struct dentry *dentry)
  2198. {
  2199. ino_t res;
  2200. /*
  2201. * Don't strictly need d_lock here? If the parent ino could change
  2202. * then surely we'd have a deeper race in the caller?
  2203. */
  2204. spin_lock(&dentry->d_lock);
  2205. res = dentry->d_parent->d_inode->i_ino;
  2206. spin_unlock(&dentry->d_lock);
  2207. return res;
  2208. }
  2209. /* Transaction based IO helpers */
  2210. /*
  2211. * An argresp is stored in an allocated page and holds the
  2212. * size of the argument or response, along with its content
  2213. */
  2214. struct simple_transaction_argresp {
  2215. ssize_t size;
  2216. char data[0];
  2217. };
  2218. #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
  2219. char *simple_transaction_get(struct file *file, const char __user *buf,
  2220. size_t size);
  2221. ssize_t simple_transaction_read(struct file *file, char __user *buf,
  2222. size_t size, loff_t *pos);
  2223. int simple_transaction_release(struct inode *inode, struct file *file);
  2224. void simple_transaction_set(struct file *file, size_t n);
  2225. /*
  2226. * simple attribute files
  2227. *
  2228. * These attributes behave similar to those in sysfs:
  2229. *
  2230. * Writing to an attribute immediately sets a value, an open file can be
  2231. * written to multiple times.
  2232. *
  2233. * Reading from an attribute creates a buffer from the value that might get
  2234. * read with multiple read calls. When the attribute has been read
  2235. * completely, no further read calls are possible until the file is opened
  2236. * again.
  2237. *
  2238. * All attributes contain a text representation of a numeric value
  2239. * that are accessed with the get() and set() functions.
  2240. */
  2241. #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
  2242. static int __fops ## _open(struct inode *inode, struct file *file) \
  2243. { \
  2244. __simple_attr_check_format(__fmt, 0ull); \
  2245. return simple_attr_open(inode, file, __get, __set, __fmt); \
  2246. } \
  2247. static const struct file_operations __fops = { \
  2248. .owner = THIS_MODULE, \
  2249. .open = __fops ## _open, \
  2250. .release = simple_attr_release, \
  2251. .read = simple_attr_read, \
  2252. .write = simple_attr_write, \
  2253. .llseek = generic_file_llseek, \
  2254. };
  2255. static inline void __attribute__((format(printf, 1, 2)))
  2256. __simple_attr_check_format(const char *fmt, ...)
  2257. {
  2258. /* don't do anything, just let the compiler check the arguments; */
  2259. }
  2260. int simple_attr_open(struct inode *inode, struct file *file,
  2261. int (*get)(void *, u64 *), int (*set)(void *, u64),
  2262. const char *fmt);
  2263. int simple_attr_release(struct inode *inode, struct file *file);
  2264. ssize_t simple_attr_read(struct file *file, char __user *buf,
  2265. size_t len, loff_t *ppos);
  2266. ssize_t simple_attr_write(struct file *file, const char __user *buf,
  2267. size_t len, loff_t *ppos);
  2268. struct ctl_table;
  2269. int proc_nr_files(struct ctl_table *table, int write,
  2270. void __user *buffer, size_t *lenp, loff_t *ppos);
  2271. int proc_nr_dentry(struct ctl_table *table, int write,
  2272. void __user *buffer, size_t *lenp, loff_t *ppos);
  2273. int proc_nr_inodes(struct ctl_table *table, int write,
  2274. void __user *buffer, size_t *lenp, loff_t *ppos);
  2275. int __init get_filesystem_list(char *buf);
  2276. #define __FMODE_EXEC ((__force int) FMODE_EXEC)
  2277. #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
  2278. #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
  2279. #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
  2280. (flag & __FMODE_NONOTIFY)))
  2281. #endif /* __KERNEL__ */
  2282. #endif /* _LINUX_FS_H */