ext4.h 79 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257225822592260226122622263226422652266226722682269227022712272227322742275227622772278227922802281228222832284
  1. /*
  2. * ext4.h
  3. *
  4. * Copyright (C) 1992, 1993, 1994, 1995
  5. * Remy Card (card@masi.ibp.fr)
  6. * Laboratoire MASI - Institut Blaise Pascal
  7. * Universite Pierre et Marie Curie (Paris VI)
  8. *
  9. * from
  10. *
  11. * linux/include/linux/minix_fs.h
  12. *
  13. * Copyright (C) 1991, 1992 Linus Torvalds
  14. */
  15. #ifndef _EXT4_H
  16. #define _EXT4_H
  17. #include <linux/types.h>
  18. #include <linux/blkdev.h>
  19. #include <linux/magic.h>
  20. #include <linux/jbd2.h>
  21. #include <linux/quota.h>
  22. #include <linux/rwsem.h>
  23. #include <linux/rbtree.h>
  24. #include <linux/seqlock.h>
  25. #include <linux/mutex.h>
  26. #include <linux/timer.h>
  27. #include <linux/wait.h>
  28. #include <linux/blockgroup_lock.h>
  29. #include <linux/percpu_counter.h>
  30. #ifdef __KERNEL__
  31. #include <linux/compat.h>
  32. #endif
  33. /*
  34. * The fourth extended filesystem constants/structures
  35. */
  36. /*
  37. * Define EXT4FS_DEBUG to produce debug messages
  38. */
  39. #undef EXT4FS_DEBUG
  40. /*
  41. * Debug code
  42. */
  43. #ifdef EXT4FS_DEBUG
  44. #define ext4_debug(f, a...) \
  45. do { \
  46. printk(KERN_DEBUG "EXT4-fs DEBUG (%s, %d): %s:", \
  47. __FILE__, __LINE__, __func__); \
  48. printk(KERN_DEBUG f, ## a); \
  49. } while (0)
  50. #else
  51. #define ext4_debug(f, a...) do {} while (0)
  52. #endif
  53. #define EXT4_ERROR_INODE(inode, fmt, a...) \
  54. ext4_error_inode((inode), __func__, __LINE__, 0, (fmt), ## a)
  55. #define EXT4_ERROR_INODE_BLOCK(inode, block, fmt, a...) \
  56. ext4_error_inode((inode), __func__, __LINE__, (block), (fmt), ## a)
  57. #define EXT4_ERROR_FILE(file, block, fmt, a...) \
  58. ext4_error_file((file), __func__, __LINE__, (block), (fmt), ## a)
  59. /* data type for block offset of block group */
  60. typedef int ext4_grpblk_t;
  61. /* data type for filesystem-wide blocks number */
  62. typedef unsigned long long ext4_fsblk_t;
  63. /* data type for file logical block number */
  64. typedef __u32 ext4_lblk_t;
  65. /* data type for block group number */
  66. typedef unsigned int ext4_group_t;
  67. /*
  68. * Flags used in mballoc's allocation_context flags field.
  69. *
  70. * Also used to show what's going on for debugging purposes when the
  71. * flag field is exported via the traceport interface
  72. */
  73. /* prefer goal again. length */
  74. #define EXT4_MB_HINT_MERGE 0x0001
  75. /* blocks already reserved */
  76. #define EXT4_MB_HINT_RESERVED 0x0002
  77. /* metadata is being allocated */
  78. #define EXT4_MB_HINT_METADATA 0x0004
  79. /* first blocks in the file */
  80. #define EXT4_MB_HINT_FIRST 0x0008
  81. /* search for the best chunk */
  82. #define EXT4_MB_HINT_BEST 0x0010
  83. /* data is being allocated */
  84. #define EXT4_MB_HINT_DATA 0x0020
  85. /* don't preallocate (for tails) */
  86. #define EXT4_MB_HINT_NOPREALLOC 0x0040
  87. /* allocate for locality group */
  88. #define EXT4_MB_HINT_GROUP_ALLOC 0x0080
  89. /* allocate goal blocks or none */
  90. #define EXT4_MB_HINT_GOAL_ONLY 0x0100
  91. /* goal is meaningful */
  92. #define EXT4_MB_HINT_TRY_GOAL 0x0200
  93. /* blocks already pre-reserved by delayed allocation */
  94. #define EXT4_MB_DELALLOC_RESERVED 0x0400
  95. /* We are doing stream allocation */
  96. #define EXT4_MB_STREAM_ALLOC 0x0800
  97. /* Use reserved root blocks if needed */
  98. #define EXT4_MB_USE_ROOT_BLOCKS 0x1000
  99. struct ext4_allocation_request {
  100. /* target inode for block we're allocating */
  101. struct inode *inode;
  102. /* how many blocks we want to allocate */
  103. unsigned int len;
  104. /* logical block in target inode */
  105. ext4_lblk_t logical;
  106. /* the closest logical allocated block to the left */
  107. ext4_lblk_t lleft;
  108. /* the closest logical allocated block to the right */
  109. ext4_lblk_t lright;
  110. /* phys. target (a hint) */
  111. ext4_fsblk_t goal;
  112. /* phys. block for the closest logical allocated block to the left */
  113. ext4_fsblk_t pleft;
  114. /* phys. block for the closest logical allocated block to the right */
  115. ext4_fsblk_t pright;
  116. /* flags. see above EXT4_MB_HINT_* */
  117. unsigned int flags;
  118. };
  119. /*
  120. * Logical to physical block mapping, used by ext4_map_blocks()
  121. *
  122. * This structure is used to pass requests into ext4_map_blocks() as
  123. * well as to store the information returned by ext4_map_blocks(). It
  124. * takes less room on the stack than a struct buffer_head.
  125. */
  126. #define EXT4_MAP_NEW (1 << BH_New)
  127. #define EXT4_MAP_MAPPED (1 << BH_Mapped)
  128. #define EXT4_MAP_UNWRITTEN (1 << BH_Unwritten)
  129. #define EXT4_MAP_BOUNDARY (1 << BH_Boundary)
  130. #define EXT4_MAP_UNINIT (1 << BH_Uninit)
  131. #define EXT4_MAP_FLAGS (EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
  132. EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY |\
  133. EXT4_MAP_UNINIT)
  134. struct ext4_map_blocks {
  135. ext4_fsblk_t m_pblk;
  136. ext4_lblk_t m_lblk;
  137. unsigned int m_len;
  138. unsigned int m_flags;
  139. };
  140. /*
  141. * For delayed allocation tracking
  142. */
  143. struct mpage_da_data {
  144. struct inode *inode;
  145. sector_t b_blocknr; /* start block number of extent */
  146. size_t b_size; /* size of extent */
  147. unsigned long b_state; /* state of the extent */
  148. unsigned long first_page, next_page; /* extent of pages */
  149. struct writeback_control *wbc;
  150. int io_done;
  151. int pages_written;
  152. int retval;
  153. };
  154. /*
  155. * Flags for ext4_io_end->flags
  156. */
  157. #define EXT4_IO_END_UNWRITTEN 0x0001
  158. #define EXT4_IO_END_ERROR 0x0002
  159. struct ext4_io_page {
  160. struct page *p_page;
  161. atomic_t p_count;
  162. };
  163. #define MAX_IO_PAGES 128
  164. typedef struct ext4_io_end {
  165. struct list_head list; /* per-file finished IO list */
  166. struct inode *inode; /* file being written to */
  167. unsigned int flag; /* unwritten or not */
  168. struct page *page; /* page struct for buffer write */
  169. loff_t offset; /* offset in the file */
  170. ssize_t size; /* size of the extent */
  171. struct work_struct work; /* data work queue */
  172. struct kiocb *iocb; /* iocb struct for AIO */
  173. int result; /* error value for AIO */
  174. int num_io_pages;
  175. struct ext4_io_page *pages[MAX_IO_PAGES];
  176. } ext4_io_end_t;
  177. struct ext4_io_submit {
  178. int io_op;
  179. struct bio *io_bio;
  180. ext4_io_end_t *io_end;
  181. struct ext4_io_page *io_page;
  182. sector_t io_next_block;
  183. };
  184. /*
  185. * Special inodes numbers
  186. */
  187. #define EXT4_BAD_INO 1 /* Bad blocks inode */
  188. #define EXT4_ROOT_INO 2 /* Root inode */
  189. #define EXT4_USR_QUOTA_INO 3 /* User quota inode */
  190. #define EXT4_GRP_QUOTA_INO 4 /* Group quota inode */
  191. #define EXT4_BOOT_LOADER_INO 5 /* Boot loader inode */
  192. #define EXT4_UNDEL_DIR_INO 6 /* Undelete directory inode */
  193. #define EXT4_RESIZE_INO 7 /* Reserved group descriptors inode */
  194. #define EXT4_JOURNAL_INO 8 /* Journal inode */
  195. /* First non-reserved inode for old ext4 filesystems */
  196. #define EXT4_GOOD_OLD_FIRST_INO 11
  197. /*
  198. * Maximal count of links to a file
  199. */
  200. #define EXT4_LINK_MAX 65000
  201. /*
  202. * Macro-instructions used to manage several block sizes
  203. */
  204. #define EXT4_MIN_BLOCK_SIZE 1024
  205. #define EXT4_MAX_BLOCK_SIZE 65536
  206. #define EXT4_MIN_BLOCK_LOG_SIZE 10
  207. #define EXT4_MAX_BLOCK_LOG_SIZE 16
  208. #ifdef __KERNEL__
  209. # define EXT4_BLOCK_SIZE(s) ((s)->s_blocksize)
  210. #else
  211. # define EXT4_BLOCK_SIZE(s) (EXT4_MIN_BLOCK_SIZE << (s)->s_log_block_size)
  212. #endif
  213. #define EXT4_ADDR_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / sizeof(__u32))
  214. #ifdef __KERNEL__
  215. # define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
  216. #else
  217. # define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_log_block_size + 10)
  218. #endif
  219. #ifdef __KERNEL__
  220. #define EXT4_ADDR_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_addr_per_block_bits)
  221. #define EXT4_INODE_SIZE(s) (EXT4_SB(s)->s_inode_size)
  222. #define EXT4_FIRST_INO(s) (EXT4_SB(s)->s_first_ino)
  223. #else
  224. #define EXT4_INODE_SIZE(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
  225. EXT4_GOOD_OLD_INODE_SIZE : \
  226. (s)->s_inode_size)
  227. #define EXT4_FIRST_INO(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
  228. EXT4_GOOD_OLD_FIRST_INO : \
  229. (s)->s_first_ino)
  230. #endif
  231. #define EXT4_BLOCK_ALIGN(size, blkbits) ALIGN((size), (1 << (blkbits)))
  232. /*
  233. * Structure of a blocks group descriptor
  234. */
  235. struct ext4_group_desc
  236. {
  237. __le32 bg_block_bitmap_lo; /* Blocks bitmap block */
  238. __le32 bg_inode_bitmap_lo; /* Inodes bitmap block */
  239. __le32 bg_inode_table_lo; /* Inodes table block */
  240. __le16 bg_free_blocks_count_lo;/* Free blocks count */
  241. __le16 bg_free_inodes_count_lo;/* Free inodes count */
  242. __le16 bg_used_dirs_count_lo; /* Directories count */
  243. __le16 bg_flags; /* EXT4_BG_flags (INODE_UNINIT, etc) */
  244. __u32 bg_reserved[2]; /* Likely block/inode bitmap checksum */
  245. __le16 bg_itable_unused_lo; /* Unused inodes count */
  246. __le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
  247. __le32 bg_block_bitmap_hi; /* Blocks bitmap block MSB */
  248. __le32 bg_inode_bitmap_hi; /* Inodes bitmap block MSB */
  249. __le32 bg_inode_table_hi; /* Inodes table block MSB */
  250. __le16 bg_free_blocks_count_hi;/* Free blocks count MSB */
  251. __le16 bg_free_inodes_count_hi;/* Free inodes count MSB */
  252. __le16 bg_used_dirs_count_hi; /* Directories count MSB */
  253. __le16 bg_itable_unused_hi; /* Unused inodes count MSB */
  254. __u32 bg_reserved2[3];
  255. };
  256. /*
  257. * Structure of a flex block group info
  258. */
  259. struct flex_groups {
  260. atomic_t free_inodes;
  261. atomic_t free_blocks;
  262. atomic_t used_dirs;
  263. };
  264. #define EXT4_BG_INODE_UNINIT 0x0001 /* Inode table/bitmap not in use */
  265. #define EXT4_BG_BLOCK_UNINIT 0x0002 /* Block bitmap not in use */
  266. #define EXT4_BG_INODE_ZEROED 0x0004 /* On-disk itable initialized to zero */
  267. /*
  268. * Macro-instructions used to manage group descriptors
  269. */
  270. #define EXT4_MIN_DESC_SIZE 32
  271. #define EXT4_MIN_DESC_SIZE_64BIT 64
  272. #define EXT4_MAX_DESC_SIZE EXT4_MIN_BLOCK_SIZE
  273. #define EXT4_DESC_SIZE(s) (EXT4_SB(s)->s_desc_size)
  274. #ifdef __KERNEL__
  275. # define EXT4_BLOCKS_PER_GROUP(s) (EXT4_SB(s)->s_blocks_per_group)
  276. # define EXT4_DESC_PER_BLOCK(s) (EXT4_SB(s)->s_desc_per_block)
  277. # define EXT4_INODES_PER_GROUP(s) (EXT4_SB(s)->s_inodes_per_group)
  278. # define EXT4_DESC_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_desc_per_block_bits)
  279. #else
  280. # define EXT4_BLOCKS_PER_GROUP(s) ((s)->s_blocks_per_group)
  281. # define EXT4_DESC_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / EXT4_DESC_SIZE(s))
  282. # define EXT4_INODES_PER_GROUP(s) ((s)->s_inodes_per_group)
  283. #endif
  284. /*
  285. * Constants relative to the data blocks
  286. */
  287. #define EXT4_NDIR_BLOCKS 12
  288. #define EXT4_IND_BLOCK EXT4_NDIR_BLOCKS
  289. #define EXT4_DIND_BLOCK (EXT4_IND_BLOCK + 1)
  290. #define EXT4_TIND_BLOCK (EXT4_DIND_BLOCK + 1)
  291. #define EXT4_N_BLOCKS (EXT4_TIND_BLOCK + 1)
  292. /*
  293. * Inode flags
  294. */
  295. #define EXT4_SECRM_FL 0x00000001 /* Secure deletion */
  296. #define EXT4_UNRM_FL 0x00000002 /* Undelete */
  297. #define EXT4_COMPR_FL 0x00000004 /* Compress file */
  298. #define EXT4_SYNC_FL 0x00000008 /* Synchronous updates */
  299. #define EXT4_IMMUTABLE_FL 0x00000010 /* Immutable file */
  300. #define EXT4_APPEND_FL 0x00000020 /* writes to file may only append */
  301. #define EXT4_NODUMP_FL 0x00000040 /* do not dump file */
  302. #define EXT4_NOATIME_FL 0x00000080 /* do not update atime */
  303. /* Reserved for compression usage... */
  304. #define EXT4_DIRTY_FL 0x00000100
  305. #define EXT4_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
  306. #define EXT4_NOCOMPR_FL 0x00000400 /* Don't compress */
  307. #define EXT4_ECOMPR_FL 0x00000800 /* Compression error */
  308. /* End compression flags --- maybe not all used */
  309. #define EXT4_INDEX_FL 0x00001000 /* hash-indexed directory */
  310. #define EXT4_IMAGIC_FL 0x00002000 /* AFS directory */
  311. #define EXT4_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */
  312. #define EXT4_NOTAIL_FL 0x00008000 /* file tail should not be merged */
  313. #define EXT4_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
  314. #define EXT4_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
  315. #define EXT4_HUGE_FILE_FL 0x00040000 /* Set to each huge file */
  316. #define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
  317. #define EXT4_EA_INODE_FL 0x00200000 /* Inode used for large EA */
  318. #define EXT4_EOFBLOCKS_FL 0x00400000 /* Blocks allocated beyond EOF */
  319. #define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */
  320. #define EXT4_FL_USER_VISIBLE 0x004BDFFF /* User visible flags */
  321. #define EXT4_FL_USER_MODIFIABLE 0x004B80FF /* User modifiable flags */
  322. /* Flags that should be inherited by new inodes from their parent. */
  323. #define EXT4_FL_INHERITED (EXT4_SECRM_FL | EXT4_UNRM_FL | EXT4_COMPR_FL |\
  324. EXT4_SYNC_FL | EXT4_IMMUTABLE_FL | EXT4_APPEND_FL |\
  325. EXT4_NODUMP_FL | EXT4_NOATIME_FL |\
  326. EXT4_NOCOMPR_FL | EXT4_JOURNAL_DATA_FL |\
  327. EXT4_NOTAIL_FL | EXT4_DIRSYNC_FL)
  328. /* Flags that are appropriate for regular files (all but dir-specific ones). */
  329. #define EXT4_REG_FLMASK (~(EXT4_DIRSYNC_FL | EXT4_TOPDIR_FL))
  330. /* Flags that are appropriate for non-directories/regular files. */
  331. #define EXT4_OTHER_FLMASK (EXT4_NODUMP_FL | EXT4_NOATIME_FL)
  332. /* Mask out flags that are inappropriate for the given type of inode. */
  333. static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
  334. {
  335. if (S_ISDIR(mode))
  336. return flags;
  337. else if (S_ISREG(mode))
  338. return flags & EXT4_REG_FLMASK;
  339. else
  340. return flags & EXT4_OTHER_FLMASK;
  341. }
  342. /*
  343. * Inode flags used for atomic set/get
  344. */
  345. enum {
  346. EXT4_INODE_SECRM = 0, /* Secure deletion */
  347. EXT4_INODE_UNRM = 1, /* Undelete */
  348. EXT4_INODE_COMPR = 2, /* Compress file */
  349. EXT4_INODE_SYNC = 3, /* Synchronous updates */
  350. EXT4_INODE_IMMUTABLE = 4, /* Immutable file */
  351. EXT4_INODE_APPEND = 5, /* writes to file may only append */
  352. EXT4_INODE_NODUMP = 6, /* do not dump file */
  353. EXT4_INODE_NOATIME = 7, /* do not update atime */
  354. /* Reserved for compression usage... */
  355. EXT4_INODE_DIRTY = 8,
  356. EXT4_INODE_COMPRBLK = 9, /* One or more compressed clusters */
  357. EXT4_INODE_NOCOMPR = 10, /* Don't compress */
  358. EXT4_INODE_ECOMPR = 11, /* Compression error */
  359. /* End compression flags --- maybe not all used */
  360. EXT4_INODE_INDEX = 12, /* hash-indexed directory */
  361. EXT4_INODE_IMAGIC = 13, /* AFS directory */
  362. EXT4_INODE_JOURNAL_DATA = 14, /* file data should be journaled */
  363. EXT4_INODE_NOTAIL = 15, /* file tail should not be merged */
  364. EXT4_INODE_DIRSYNC = 16, /* dirsync behaviour (directories only) */
  365. EXT4_INODE_TOPDIR = 17, /* Top of directory hierarchies*/
  366. EXT4_INODE_HUGE_FILE = 18, /* Set to each huge file */
  367. EXT4_INODE_EXTENTS = 19, /* Inode uses extents */
  368. EXT4_INODE_EA_INODE = 21, /* Inode used for large EA */
  369. EXT4_INODE_EOFBLOCKS = 22, /* Blocks allocated beyond EOF */
  370. EXT4_INODE_RESERVED = 31, /* reserved for ext4 lib */
  371. };
  372. #define TEST_FLAG_VALUE(FLAG) (EXT4_##FLAG##_FL == (1 << EXT4_INODE_##FLAG))
  373. #define CHECK_FLAG_VALUE(FLAG) if (!TEST_FLAG_VALUE(FLAG)) { \
  374. printk(KERN_EMERG "EXT4 flag fail: " #FLAG ": %d %d\n", \
  375. EXT4_##FLAG##_FL, EXT4_INODE_##FLAG); BUG_ON(1); }
  376. /*
  377. * Since it's pretty easy to mix up bit numbers and hex values, and we
  378. * can't do a compile-time test for ENUM values, we use a run-time
  379. * test to make sure that EXT4_XXX_FL is consistent with respect to
  380. * EXT4_INODE_XXX. If all is well the printk and BUG_ON will all drop
  381. * out so it won't cost any extra space in the compiled kernel image.
  382. * But it's important that these values are the same, since we are
  383. * using EXT4_INODE_XXX to test for the flag values, but EXT4_XX_FL
  384. * must be consistent with the values of FS_XXX_FL defined in
  385. * include/linux/fs.h and the on-disk values found in ext2, ext3, and
  386. * ext4 filesystems, and of course the values defined in e2fsprogs.
  387. *
  388. * It's not paranoia if the Murphy's Law really *is* out to get you. :-)
  389. */
  390. static inline void ext4_check_flag_values(void)
  391. {
  392. CHECK_FLAG_VALUE(SECRM);
  393. CHECK_FLAG_VALUE(UNRM);
  394. CHECK_FLAG_VALUE(COMPR);
  395. CHECK_FLAG_VALUE(SYNC);
  396. CHECK_FLAG_VALUE(IMMUTABLE);
  397. CHECK_FLAG_VALUE(APPEND);
  398. CHECK_FLAG_VALUE(NODUMP);
  399. CHECK_FLAG_VALUE(NOATIME);
  400. CHECK_FLAG_VALUE(DIRTY);
  401. CHECK_FLAG_VALUE(COMPRBLK);
  402. CHECK_FLAG_VALUE(NOCOMPR);
  403. CHECK_FLAG_VALUE(ECOMPR);
  404. CHECK_FLAG_VALUE(INDEX);
  405. CHECK_FLAG_VALUE(IMAGIC);
  406. CHECK_FLAG_VALUE(JOURNAL_DATA);
  407. CHECK_FLAG_VALUE(NOTAIL);
  408. CHECK_FLAG_VALUE(DIRSYNC);
  409. CHECK_FLAG_VALUE(TOPDIR);
  410. CHECK_FLAG_VALUE(HUGE_FILE);
  411. CHECK_FLAG_VALUE(EXTENTS);
  412. CHECK_FLAG_VALUE(EA_INODE);
  413. CHECK_FLAG_VALUE(EOFBLOCKS);
  414. CHECK_FLAG_VALUE(RESERVED);
  415. }
  416. /* Used to pass group descriptor data when online resize is done */
  417. struct ext4_new_group_input {
  418. __u32 group; /* Group number for this data */
  419. __u64 block_bitmap; /* Absolute block number of block bitmap */
  420. __u64 inode_bitmap; /* Absolute block number of inode bitmap */
  421. __u64 inode_table; /* Absolute block number of inode table start */
  422. __u32 blocks_count; /* Total number of blocks in this group */
  423. __u16 reserved_blocks; /* Number of reserved blocks in this group */
  424. __u16 unused;
  425. };
  426. #if defined(__KERNEL__) && defined(CONFIG_COMPAT)
  427. struct compat_ext4_new_group_input {
  428. u32 group;
  429. compat_u64 block_bitmap;
  430. compat_u64 inode_bitmap;
  431. compat_u64 inode_table;
  432. u32 blocks_count;
  433. u16 reserved_blocks;
  434. u16 unused;
  435. };
  436. #endif
  437. /* The struct ext4_new_group_input in kernel space, with free_blocks_count */
  438. struct ext4_new_group_data {
  439. __u32 group;
  440. __u64 block_bitmap;
  441. __u64 inode_bitmap;
  442. __u64 inode_table;
  443. __u32 blocks_count;
  444. __u16 reserved_blocks;
  445. __u16 unused;
  446. __u32 free_blocks_count;
  447. };
  448. /*
  449. * Flags used by ext4_map_blocks()
  450. */
  451. /* Allocate any needed blocks and/or convert an unitialized
  452. extent to be an initialized ext4 */
  453. #define EXT4_GET_BLOCKS_CREATE 0x0001
  454. /* Request the creation of an unitialized extent */
  455. #define EXT4_GET_BLOCKS_UNINIT_EXT 0x0002
  456. #define EXT4_GET_BLOCKS_CREATE_UNINIT_EXT (EXT4_GET_BLOCKS_UNINIT_EXT|\
  457. EXT4_GET_BLOCKS_CREATE)
  458. /* Caller is from the delayed allocation writeout path,
  459. so set the magic i_delalloc_reserve_flag after taking the
  460. inode allocation semaphore for */
  461. #define EXT4_GET_BLOCKS_DELALLOC_RESERVE 0x0004
  462. /* caller is from the direct IO path, request to creation of an
  463. unitialized extents if not allocated, split the uninitialized
  464. extent if blocks has been preallocated already*/
  465. #define EXT4_GET_BLOCKS_PRE_IO 0x0008
  466. #define EXT4_GET_BLOCKS_CONVERT 0x0010
  467. #define EXT4_GET_BLOCKS_IO_CREATE_EXT (EXT4_GET_BLOCKS_PRE_IO|\
  468. EXT4_GET_BLOCKS_CREATE_UNINIT_EXT)
  469. /* Convert extent to initialized after IO complete */
  470. #define EXT4_GET_BLOCKS_IO_CONVERT_EXT (EXT4_GET_BLOCKS_CONVERT|\
  471. EXT4_GET_BLOCKS_CREATE_UNINIT_EXT)
  472. /* Punch out blocks of an extent */
  473. #define EXT4_GET_BLOCKS_PUNCH_OUT_EXT 0x0020
  474. /* Don't normalize allocation size (used for fallocate) */
  475. #define EXT4_GET_BLOCKS_NO_NORMALIZE 0x0040
  476. /*
  477. * Flags used by ext4_free_blocks
  478. */
  479. #define EXT4_FREE_BLOCKS_METADATA 0x0001
  480. #define EXT4_FREE_BLOCKS_FORGET 0x0002
  481. #define EXT4_FREE_BLOCKS_VALIDATED 0x0004
  482. #define EXT4_FREE_BLOCKS_NO_QUOT_UPDATE 0x0008
  483. /*
  484. * ioctl commands
  485. */
  486. #define EXT4_IOC_GETFLAGS FS_IOC_GETFLAGS
  487. #define EXT4_IOC_SETFLAGS FS_IOC_SETFLAGS
  488. #define EXT4_IOC_GETVERSION _IOR('f', 3, long)
  489. #define EXT4_IOC_SETVERSION _IOW('f', 4, long)
  490. #define EXT4_IOC_GETVERSION_OLD FS_IOC_GETVERSION
  491. #define EXT4_IOC_SETVERSION_OLD FS_IOC_SETVERSION
  492. #ifdef CONFIG_JBD2_DEBUG
  493. #define EXT4_IOC_WAIT_FOR_READONLY _IOR('f', 99, long)
  494. #endif
  495. #define EXT4_IOC_GETRSVSZ _IOR('f', 5, long)
  496. #define EXT4_IOC_SETRSVSZ _IOW('f', 6, long)
  497. #define EXT4_IOC_GROUP_EXTEND _IOW('f', 7, unsigned long)
  498. #define EXT4_IOC_GROUP_ADD _IOW('f', 8, struct ext4_new_group_input)
  499. #define EXT4_IOC_MIGRATE _IO('f', 9)
  500. /* note ioctl 10 reserved for an early version of the FIEMAP ioctl */
  501. /* note ioctl 11 reserved for filesystem-independent FIEMAP ioctl */
  502. #define EXT4_IOC_ALLOC_DA_BLKS _IO('f', 12)
  503. #define EXT4_IOC_MOVE_EXT _IOWR('f', 15, struct move_extent)
  504. #if defined(__KERNEL__) && defined(CONFIG_COMPAT)
  505. /*
  506. * ioctl commands in 32 bit emulation
  507. */
  508. #define EXT4_IOC32_GETFLAGS FS_IOC32_GETFLAGS
  509. #define EXT4_IOC32_SETFLAGS FS_IOC32_SETFLAGS
  510. #define EXT4_IOC32_GETVERSION _IOR('f', 3, int)
  511. #define EXT4_IOC32_SETVERSION _IOW('f', 4, int)
  512. #define EXT4_IOC32_GETRSVSZ _IOR('f', 5, int)
  513. #define EXT4_IOC32_SETRSVSZ _IOW('f', 6, int)
  514. #define EXT4_IOC32_GROUP_EXTEND _IOW('f', 7, unsigned int)
  515. #define EXT4_IOC32_GROUP_ADD _IOW('f', 8, struct compat_ext4_new_group_input)
  516. #ifdef CONFIG_JBD2_DEBUG
  517. #define EXT4_IOC32_WAIT_FOR_READONLY _IOR('f', 99, int)
  518. #endif
  519. #define EXT4_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
  520. #define EXT4_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
  521. #endif
  522. /* Max physical block we can address w/o extents */
  523. #define EXT4_MAX_BLOCK_FILE_PHYS 0xFFFFFFFF
  524. /*
  525. * Structure of an inode on the disk
  526. */
  527. struct ext4_inode {
  528. __le16 i_mode; /* File mode */
  529. __le16 i_uid; /* Low 16 bits of Owner Uid */
  530. __le32 i_size_lo; /* Size in bytes */
  531. __le32 i_atime; /* Access time */
  532. __le32 i_ctime; /* Inode Change time */
  533. __le32 i_mtime; /* Modification time */
  534. __le32 i_dtime; /* Deletion Time */
  535. __le16 i_gid; /* Low 16 bits of Group Id */
  536. __le16 i_links_count; /* Links count */
  537. __le32 i_blocks_lo; /* Blocks count */
  538. __le32 i_flags; /* File flags */
  539. union {
  540. struct {
  541. __le32 l_i_version;
  542. } linux1;
  543. struct {
  544. __u32 h_i_translator;
  545. } hurd1;
  546. struct {
  547. __u32 m_i_reserved1;
  548. } masix1;
  549. } osd1; /* OS dependent 1 */
  550. __le32 i_block[EXT4_N_BLOCKS];/* Pointers to blocks */
  551. __le32 i_generation; /* File version (for NFS) */
  552. __le32 i_file_acl_lo; /* File ACL */
  553. __le32 i_size_high;
  554. __le32 i_obso_faddr; /* Obsoleted fragment address */
  555. union {
  556. struct {
  557. __le16 l_i_blocks_high; /* were l_i_reserved1 */
  558. __le16 l_i_file_acl_high;
  559. __le16 l_i_uid_high; /* these 2 fields */
  560. __le16 l_i_gid_high; /* were reserved2[0] */
  561. __u32 l_i_reserved2;
  562. } linux2;
  563. struct {
  564. __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
  565. __u16 h_i_mode_high;
  566. __u16 h_i_uid_high;
  567. __u16 h_i_gid_high;
  568. __u32 h_i_author;
  569. } hurd2;
  570. struct {
  571. __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
  572. __le16 m_i_file_acl_high;
  573. __u32 m_i_reserved2[2];
  574. } masix2;
  575. } osd2; /* OS dependent 2 */
  576. __le16 i_extra_isize;
  577. __le16 i_pad1;
  578. __le32 i_ctime_extra; /* extra Change time (nsec << 2 | epoch) */
  579. __le32 i_mtime_extra; /* extra Modification time(nsec << 2 | epoch) */
  580. __le32 i_atime_extra; /* extra Access time (nsec << 2 | epoch) */
  581. __le32 i_crtime; /* File Creation time */
  582. __le32 i_crtime_extra; /* extra FileCreationtime (nsec << 2 | epoch) */
  583. __le32 i_version_hi; /* high 32 bits for 64-bit version */
  584. };
  585. struct move_extent {
  586. __u32 reserved; /* should be zero */
  587. __u32 donor_fd; /* donor file descriptor */
  588. __u64 orig_start; /* logical start offset in block for orig */
  589. __u64 donor_start; /* logical start offset in block for donor */
  590. __u64 len; /* block length to be moved */
  591. __u64 moved_len; /* moved block length */
  592. };
  593. #define EXT4_EPOCH_BITS 2
  594. #define EXT4_EPOCH_MASK ((1 << EXT4_EPOCH_BITS) - 1)
  595. #define EXT4_NSEC_MASK (~0UL << EXT4_EPOCH_BITS)
  596. /*
  597. * Extended fields will fit into an inode if the filesystem was formatted
  598. * with large inodes (-I 256 or larger) and there are not currently any EAs
  599. * consuming all of the available space. For new inodes we always reserve
  600. * enough space for the kernel's known extended fields, but for inodes
  601. * created with an old kernel this might not have been the case. None of
  602. * the extended inode fields is critical for correct filesystem operation.
  603. * This macro checks if a certain field fits in the inode. Note that
  604. * inode-size = GOOD_OLD_INODE_SIZE + i_extra_isize
  605. */
  606. #define EXT4_FITS_IN_INODE(ext4_inode, einode, field) \
  607. ((offsetof(typeof(*ext4_inode), field) + \
  608. sizeof((ext4_inode)->field)) \
  609. <= (EXT4_GOOD_OLD_INODE_SIZE + \
  610. (einode)->i_extra_isize)) \
  611. static inline __le32 ext4_encode_extra_time(struct timespec *time)
  612. {
  613. return cpu_to_le32((sizeof(time->tv_sec) > 4 ?
  614. (time->tv_sec >> 32) & EXT4_EPOCH_MASK : 0) |
  615. ((time->tv_nsec << EXT4_EPOCH_BITS) & EXT4_NSEC_MASK));
  616. }
  617. static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra)
  618. {
  619. if (sizeof(time->tv_sec) > 4)
  620. time->tv_sec |= (__u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK)
  621. << 32;
  622. time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> EXT4_EPOCH_BITS;
  623. }
  624. #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \
  625. do { \
  626. (raw_inode)->xtime = cpu_to_le32((inode)->xtime.tv_sec); \
  627. if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) \
  628. (raw_inode)->xtime ## _extra = \
  629. ext4_encode_extra_time(&(inode)->xtime); \
  630. } while (0)
  631. #define EXT4_EINODE_SET_XTIME(xtime, einode, raw_inode) \
  632. do { \
  633. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
  634. (raw_inode)->xtime = cpu_to_le32((einode)->xtime.tv_sec); \
  635. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
  636. (raw_inode)->xtime ## _extra = \
  637. ext4_encode_extra_time(&(einode)->xtime); \
  638. } while (0)
  639. #define EXT4_INODE_GET_XTIME(xtime, inode, raw_inode) \
  640. do { \
  641. (inode)->xtime.tv_sec = (signed)le32_to_cpu((raw_inode)->xtime); \
  642. if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) \
  643. ext4_decode_extra_time(&(inode)->xtime, \
  644. raw_inode->xtime ## _extra); \
  645. else \
  646. (inode)->xtime.tv_nsec = 0; \
  647. } while (0)
  648. #define EXT4_EINODE_GET_XTIME(xtime, einode, raw_inode) \
  649. do { \
  650. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
  651. (einode)->xtime.tv_sec = \
  652. (signed)le32_to_cpu((raw_inode)->xtime); \
  653. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
  654. ext4_decode_extra_time(&(einode)->xtime, \
  655. raw_inode->xtime ## _extra); \
  656. else \
  657. (einode)->xtime.tv_nsec = 0; \
  658. } while (0)
  659. #define i_disk_version osd1.linux1.l_i_version
  660. #if defined(__KERNEL__) || defined(__linux__)
  661. #define i_reserved1 osd1.linux1.l_i_reserved1
  662. #define i_file_acl_high osd2.linux2.l_i_file_acl_high
  663. #define i_blocks_high osd2.linux2.l_i_blocks_high
  664. #define i_uid_low i_uid
  665. #define i_gid_low i_gid
  666. #define i_uid_high osd2.linux2.l_i_uid_high
  667. #define i_gid_high osd2.linux2.l_i_gid_high
  668. #define i_reserved2 osd2.linux2.l_i_reserved2
  669. #elif defined(__GNU__)
  670. #define i_translator osd1.hurd1.h_i_translator
  671. #define i_uid_high osd2.hurd2.h_i_uid_high
  672. #define i_gid_high osd2.hurd2.h_i_gid_high
  673. #define i_author osd2.hurd2.h_i_author
  674. #elif defined(__masix__)
  675. #define i_reserved1 osd1.masix1.m_i_reserved1
  676. #define i_file_acl_high osd2.masix2.m_i_file_acl_high
  677. #define i_reserved2 osd2.masix2.m_i_reserved2
  678. #endif /* defined(__KERNEL__) || defined(__linux__) */
  679. /*
  680. * storage for cached extent
  681. * If ec_len == 0, then the cache is invalid.
  682. * If ec_start == 0, then the cache represents a gap (null mapping)
  683. */
  684. struct ext4_ext_cache {
  685. ext4_fsblk_t ec_start;
  686. ext4_lblk_t ec_block;
  687. __u32 ec_len; /* must be 32bit to return holes */
  688. };
  689. /*
  690. * fourth extended file system inode data in memory
  691. */
  692. struct ext4_inode_info {
  693. __le32 i_data[15]; /* unconverted */
  694. __u32 i_dtime;
  695. ext4_fsblk_t i_file_acl;
  696. /*
  697. * i_block_group is the number of the block group which contains
  698. * this file's inode. Constant across the lifetime of the inode,
  699. * it is ued for making block allocation decisions - we try to
  700. * place a file's data blocks near its inode block, and new inodes
  701. * near to their parent directory's inode.
  702. */
  703. ext4_group_t i_block_group;
  704. ext4_lblk_t i_dir_start_lookup;
  705. #if (BITS_PER_LONG < 64)
  706. unsigned long i_state_flags; /* Dynamic state flags */
  707. #endif
  708. unsigned long i_flags;
  709. #ifdef CONFIG_EXT4_FS_XATTR
  710. /*
  711. * Extended attributes can be read independently of the main file
  712. * data. Taking i_mutex even when reading would cause contention
  713. * between readers of EAs and writers of regular file data, so
  714. * instead we synchronize on xattr_sem when reading or changing
  715. * EAs.
  716. */
  717. struct rw_semaphore xattr_sem;
  718. #endif
  719. struct list_head i_orphan; /* unlinked but open inodes */
  720. /*
  721. * i_disksize keeps track of what the inode size is ON DISK, not
  722. * in memory. During truncate, i_size is set to the new size by
  723. * the VFS prior to calling ext4_truncate(), but the filesystem won't
  724. * set i_disksize to 0 until the truncate is actually under way.
  725. *
  726. * The intent is that i_disksize always represents the blocks which
  727. * are used by this file. This allows recovery to restart truncate
  728. * on orphans if we crash during truncate. We actually write i_disksize
  729. * into the on-disk inode when writing inodes out, instead of i_size.
  730. *
  731. * The only time when i_disksize and i_size may be different is when
  732. * a truncate is in progress. The only things which change i_disksize
  733. * are ext4_get_block (growth) and ext4_truncate (shrinkth).
  734. */
  735. loff_t i_disksize;
  736. /*
  737. * i_data_sem is for serialising ext4_truncate() against
  738. * ext4_getblock(). In the 2.4 ext2 design, great chunks of inode's
  739. * data tree are chopped off during truncate. We can't do that in
  740. * ext4 because whenever we perform intermediate commits during
  741. * truncate, the inode and all the metadata blocks *must* be in a
  742. * consistent state which allows truncation of the orphans to restart
  743. * during recovery. Hence we must fix the get_block-vs-truncate race
  744. * by other means, so we have i_data_sem.
  745. */
  746. struct rw_semaphore i_data_sem;
  747. struct inode vfs_inode;
  748. struct jbd2_inode *jinode;
  749. struct ext4_ext_cache i_cached_extent;
  750. /*
  751. * File creation time. Its function is same as that of
  752. * struct timespec i_{a,c,m}time in the generic inode.
  753. */
  754. struct timespec i_crtime;
  755. /* mballoc */
  756. struct list_head i_prealloc_list;
  757. spinlock_t i_prealloc_lock;
  758. /* ialloc */
  759. ext4_group_t i_last_alloc_group;
  760. /* allocation reservation info for delalloc */
  761. unsigned int i_reserved_data_blocks;
  762. unsigned int i_reserved_meta_blocks;
  763. unsigned int i_allocated_meta_blocks;
  764. ext4_lblk_t i_da_metadata_calc_last_lblock;
  765. int i_da_metadata_calc_len;
  766. /* on-disk additional length */
  767. __u16 i_extra_isize;
  768. #ifdef CONFIG_QUOTA
  769. /* quota space reservation, managed internally by quota code */
  770. qsize_t i_reserved_quota;
  771. #endif
  772. /* completed IOs that might need unwritten extents handling */
  773. struct list_head i_completed_io_list;
  774. spinlock_t i_completed_io_lock;
  775. atomic_t i_ioend_count; /* Number of outstanding io_end structs */
  776. /* current io_end structure for async DIO write*/
  777. ext4_io_end_t *cur_aio_dio;
  778. atomic_t i_aiodio_unwritten; /* Nr. of inflight conversions pending */
  779. spinlock_t i_block_reservation_lock;
  780. /*
  781. * Transactions that contain inode's metadata needed to complete
  782. * fsync and fdatasync, respectively.
  783. */
  784. tid_t i_sync_tid;
  785. tid_t i_datasync_tid;
  786. };
  787. /*
  788. * File system states
  789. */
  790. #define EXT4_VALID_FS 0x0001 /* Unmounted cleanly */
  791. #define EXT4_ERROR_FS 0x0002 /* Errors detected */
  792. #define EXT4_ORPHAN_FS 0x0004 /* Orphans being recovered */
  793. /*
  794. * Misc. filesystem flags
  795. */
  796. #define EXT2_FLAGS_SIGNED_HASH 0x0001 /* Signed dirhash in use */
  797. #define EXT2_FLAGS_UNSIGNED_HASH 0x0002 /* Unsigned dirhash in use */
  798. #define EXT2_FLAGS_TEST_FILESYS 0x0004 /* to test development code */
  799. /*
  800. * Mount flags
  801. */
  802. #define EXT4_MOUNT_OLDALLOC 0x00002 /* Don't use the new Orlov allocator */
  803. #define EXT4_MOUNT_GRPID 0x00004 /* Create files with directory's group */
  804. #define EXT4_MOUNT_DEBUG 0x00008 /* Some debugging messages */
  805. #define EXT4_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
  806. #define EXT4_MOUNT_ERRORS_RO 0x00020 /* Remount fs ro on errors */
  807. #define EXT4_MOUNT_ERRORS_PANIC 0x00040 /* Panic on errors */
  808. #define EXT4_MOUNT_MINIX_DF 0x00080 /* Mimics the Minix statfs */
  809. #define EXT4_MOUNT_NOLOAD 0x00100 /* Don't use existing journal*/
  810. #define EXT4_MOUNT_DATA_FLAGS 0x00C00 /* Mode for data writes: */
  811. #define EXT4_MOUNT_JOURNAL_DATA 0x00400 /* Write data to journal */
  812. #define EXT4_MOUNT_ORDERED_DATA 0x00800 /* Flush data before commit */
  813. #define EXT4_MOUNT_WRITEBACK_DATA 0x00C00 /* No data ordering */
  814. #define EXT4_MOUNT_UPDATE_JOURNAL 0x01000 /* Update the journal format */
  815. #define EXT4_MOUNT_NO_UID32 0x02000 /* Disable 32-bit UIDs */
  816. #define EXT4_MOUNT_XATTR_USER 0x04000 /* Extended user attributes */
  817. #define EXT4_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */
  818. #define EXT4_MOUNT_NO_AUTO_DA_ALLOC 0x10000 /* No auto delalloc mapping */
  819. #define EXT4_MOUNT_BARRIER 0x20000 /* Use block barriers */
  820. #define EXT4_MOUNT_QUOTA 0x80000 /* Some quota option set */
  821. #define EXT4_MOUNT_USRQUOTA 0x100000 /* "old" user quota */
  822. #define EXT4_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */
  823. #define EXT4_MOUNT_DIOREAD_NOLOCK 0x400000 /* Enable support for dio read nolocking */
  824. #define EXT4_MOUNT_JOURNAL_CHECKSUM 0x800000 /* Journal checksums */
  825. #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT 0x1000000 /* Journal Async Commit */
  826. #define EXT4_MOUNT_I_VERSION 0x2000000 /* i_version support */
  827. #define EXT4_MOUNT_MBLK_IO_SUBMIT 0x4000000 /* multi-block io submits */
  828. #define EXT4_MOUNT_DELALLOC 0x8000000 /* Delalloc support */
  829. #define EXT4_MOUNT_DATA_ERR_ABORT 0x10000000 /* Abort on file data write */
  830. #define EXT4_MOUNT_BLOCK_VALIDITY 0x20000000 /* Block validity checking */
  831. #define EXT4_MOUNT_DISCARD 0x40000000 /* Issue DISCARD requests */
  832. #define EXT4_MOUNT_INIT_INODE_TABLE 0x80000000 /* Initialize uninitialized itables */
  833. #define clear_opt(sb, opt) EXT4_SB(sb)->s_mount_opt &= \
  834. ~EXT4_MOUNT_##opt
  835. #define set_opt(sb, opt) EXT4_SB(sb)->s_mount_opt |= \
  836. EXT4_MOUNT_##opt
  837. #define test_opt(sb, opt) (EXT4_SB(sb)->s_mount_opt & \
  838. EXT4_MOUNT_##opt)
  839. #define clear_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 &= \
  840. ~EXT4_MOUNT2_##opt
  841. #define set_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 |= \
  842. EXT4_MOUNT2_##opt
  843. #define test_opt2(sb, opt) (EXT4_SB(sb)->s_mount_opt2 & \
  844. EXT4_MOUNT2_##opt)
  845. #define ext4_set_bit __test_and_set_bit_le
  846. #define ext4_set_bit_atomic ext2_set_bit_atomic
  847. #define ext4_clear_bit __test_and_clear_bit_le
  848. #define ext4_clear_bit_atomic ext2_clear_bit_atomic
  849. #define ext4_test_bit test_bit_le
  850. #define ext4_find_first_zero_bit find_first_zero_bit_le
  851. #define ext4_find_next_zero_bit find_next_zero_bit_le
  852. #define ext4_find_next_bit find_next_bit_le
  853. extern void ext4_set_bits(void *bm, int cur, int len);
  854. /*
  855. * Maximal mount counts between two filesystem checks
  856. */
  857. #define EXT4_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
  858. #define EXT4_DFL_CHECKINTERVAL 0 /* Don't use interval check */
  859. /*
  860. * Behaviour when detecting errors
  861. */
  862. #define EXT4_ERRORS_CONTINUE 1 /* Continue execution */
  863. #define EXT4_ERRORS_RO 2 /* Remount fs read-only */
  864. #define EXT4_ERRORS_PANIC 3 /* Panic */
  865. #define EXT4_ERRORS_DEFAULT EXT4_ERRORS_CONTINUE
  866. /*
  867. * Structure of the super block
  868. */
  869. struct ext4_super_block {
  870. /*00*/ __le32 s_inodes_count; /* Inodes count */
  871. __le32 s_blocks_count_lo; /* Blocks count */
  872. __le32 s_r_blocks_count_lo; /* Reserved blocks count */
  873. __le32 s_free_blocks_count_lo; /* Free blocks count */
  874. /*10*/ __le32 s_free_inodes_count; /* Free inodes count */
  875. __le32 s_first_data_block; /* First Data Block */
  876. __le32 s_log_block_size; /* Block size */
  877. __le32 s_obso_log_frag_size; /* Obsoleted fragment size */
  878. /*20*/ __le32 s_blocks_per_group; /* # Blocks per group */
  879. __le32 s_obso_frags_per_group; /* Obsoleted fragments per group */
  880. __le32 s_inodes_per_group; /* # Inodes per group */
  881. __le32 s_mtime; /* Mount time */
  882. /*30*/ __le32 s_wtime; /* Write time */
  883. __le16 s_mnt_count; /* Mount count */
  884. __le16 s_max_mnt_count; /* Maximal mount count */
  885. __le16 s_magic; /* Magic signature */
  886. __le16 s_state; /* File system state */
  887. __le16 s_errors; /* Behaviour when detecting errors */
  888. __le16 s_minor_rev_level; /* minor revision level */
  889. /*40*/ __le32 s_lastcheck; /* time of last check */
  890. __le32 s_checkinterval; /* max. time between checks */
  891. __le32 s_creator_os; /* OS */
  892. __le32 s_rev_level; /* Revision level */
  893. /*50*/ __le16 s_def_resuid; /* Default uid for reserved blocks */
  894. __le16 s_def_resgid; /* Default gid for reserved blocks */
  895. /*
  896. * These fields are for EXT4_DYNAMIC_REV superblocks only.
  897. *
  898. * Note: the difference between the compatible feature set and
  899. * the incompatible feature set is that if there is a bit set
  900. * in the incompatible feature set that the kernel doesn't
  901. * know about, it should refuse to mount the filesystem.
  902. *
  903. * e2fsck's requirements are more strict; if it doesn't know
  904. * about a feature in either the compatible or incompatible
  905. * feature set, it must abort and not try to meddle with
  906. * things it doesn't understand...
  907. */
  908. __le32 s_first_ino; /* First non-reserved inode */
  909. __le16 s_inode_size; /* size of inode structure */
  910. __le16 s_block_group_nr; /* block group # of this superblock */
  911. __le32 s_feature_compat; /* compatible feature set */
  912. /*60*/ __le32 s_feature_incompat; /* incompatible feature set */
  913. __le32 s_feature_ro_compat; /* readonly-compatible feature set */
  914. /*68*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */
  915. /*78*/ char s_volume_name[16]; /* volume name */
  916. /*88*/ char s_last_mounted[64]; /* directory where last mounted */
  917. /*C8*/ __le32 s_algorithm_usage_bitmap; /* For compression */
  918. /*
  919. * Performance hints. Directory preallocation should only
  920. * happen if the EXT4_FEATURE_COMPAT_DIR_PREALLOC flag is on.
  921. */
  922. __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
  923. __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
  924. __le16 s_reserved_gdt_blocks; /* Per group desc for online growth */
  925. /*
  926. * Journaling support valid if EXT4_FEATURE_COMPAT_HAS_JOURNAL set.
  927. */
  928. /*D0*/ __u8 s_journal_uuid[16]; /* uuid of journal superblock */
  929. /*E0*/ __le32 s_journal_inum; /* inode number of journal file */
  930. __le32 s_journal_dev; /* device number of journal file */
  931. __le32 s_last_orphan; /* start of list of inodes to delete */
  932. __le32 s_hash_seed[4]; /* HTREE hash seed */
  933. __u8 s_def_hash_version; /* Default hash version to use */
  934. __u8 s_jnl_backup_type;
  935. __le16 s_desc_size; /* size of group descriptor */
  936. /*100*/ __le32 s_default_mount_opts;
  937. __le32 s_first_meta_bg; /* First metablock block group */
  938. __le32 s_mkfs_time; /* When the filesystem was created */
  939. __le32 s_jnl_blocks[17]; /* Backup of the journal inode */
  940. /* 64bit support valid if EXT4_FEATURE_COMPAT_64BIT */
  941. /*150*/ __le32 s_blocks_count_hi; /* Blocks count */
  942. __le32 s_r_blocks_count_hi; /* Reserved blocks count */
  943. __le32 s_free_blocks_count_hi; /* Free blocks count */
  944. __le16 s_min_extra_isize; /* All inodes have at least # bytes */
  945. __le16 s_want_extra_isize; /* New inodes should reserve # bytes */
  946. __le32 s_flags; /* Miscellaneous flags */
  947. __le16 s_raid_stride; /* RAID stride */
  948. __le16 s_mmp_update_interval; /* # seconds to wait in MMP checking */
  949. __le64 s_mmp_block; /* Block for multi-mount protection */
  950. __le32 s_raid_stripe_width; /* blocks on all data disks (N*stride)*/
  951. __u8 s_log_groups_per_flex; /* FLEX_BG group size */
  952. __u8 s_reserved_char_pad;
  953. __le16 s_reserved_pad;
  954. __le64 s_kbytes_written; /* nr of lifetime kilobytes written */
  955. __le32 s_snapshot_inum; /* Inode number of active snapshot */
  956. __le32 s_snapshot_id; /* sequential ID of active snapshot */
  957. __le64 s_snapshot_r_blocks_count; /* reserved blocks for active
  958. snapshot's future use */
  959. __le32 s_snapshot_list; /* inode number of the head of the
  960. on-disk snapshot list */
  961. #define EXT4_S_ERR_START offsetof(struct ext4_super_block, s_error_count)
  962. __le32 s_error_count; /* number of fs errors */
  963. __le32 s_first_error_time; /* first time an error happened */
  964. __le32 s_first_error_ino; /* inode involved in first error */
  965. __le64 s_first_error_block; /* block involved of first error */
  966. __u8 s_first_error_func[32]; /* function where the error happened */
  967. __le32 s_first_error_line; /* line number where error happened */
  968. __le32 s_last_error_time; /* most recent time of an error */
  969. __le32 s_last_error_ino; /* inode involved in last error */
  970. __le32 s_last_error_line; /* line number where error happened */
  971. __le64 s_last_error_block; /* block involved of last error */
  972. __u8 s_last_error_func[32]; /* function where the error happened */
  973. #define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_mount_opts)
  974. __u8 s_mount_opts[64];
  975. __le32 s_reserved[112]; /* Padding to the end of the block */
  976. };
  977. #define EXT4_S_ERR_LEN (EXT4_S_ERR_END - EXT4_S_ERR_START)
  978. #ifdef __KERNEL__
  979. /*
  980. * run-time mount flags
  981. */
  982. #define EXT4_MF_MNTDIR_SAMPLED 0x0001
  983. #define EXT4_MF_FS_ABORTED 0x0002 /* Fatal error detected */
  984. /*
  985. * fourth extended-fs super-block data in memory
  986. */
  987. struct ext4_sb_info {
  988. unsigned long s_desc_size; /* Size of a group descriptor in bytes */
  989. unsigned long s_inodes_per_block;/* Number of inodes per block */
  990. unsigned long s_blocks_per_group;/* Number of blocks in a group */
  991. unsigned long s_inodes_per_group;/* Number of inodes in a group */
  992. unsigned long s_itb_per_group; /* Number of inode table blocks per group */
  993. unsigned long s_gdb_count; /* Number of group descriptor blocks */
  994. unsigned long s_desc_per_block; /* Number of group descriptors per block */
  995. ext4_group_t s_groups_count; /* Number of groups in the fs */
  996. ext4_group_t s_blockfile_groups;/* Groups acceptable for non-extent files */
  997. unsigned long s_overhead_last; /* Last calculated overhead */
  998. unsigned long s_blocks_last; /* Last seen block count */
  999. loff_t s_bitmap_maxbytes; /* max bytes for bitmap files */
  1000. struct buffer_head * s_sbh; /* Buffer containing the super block */
  1001. struct ext4_super_block *s_es; /* Pointer to the super block in the buffer */
  1002. struct buffer_head **s_group_desc;
  1003. unsigned int s_mount_opt;
  1004. unsigned int s_mount_opt2;
  1005. unsigned int s_mount_flags;
  1006. ext4_fsblk_t s_sb_block;
  1007. uid_t s_resuid;
  1008. gid_t s_resgid;
  1009. unsigned short s_mount_state;
  1010. unsigned short s_pad;
  1011. int s_addr_per_block_bits;
  1012. int s_desc_per_block_bits;
  1013. int s_inode_size;
  1014. int s_first_ino;
  1015. unsigned int s_inode_readahead_blks;
  1016. unsigned int s_inode_goal;
  1017. spinlock_t s_next_gen_lock;
  1018. u32 s_next_generation;
  1019. u32 s_hash_seed[4];
  1020. int s_def_hash_version;
  1021. int s_hash_unsigned; /* 3 if hash should be signed, 0 if not */
  1022. struct percpu_counter s_freeblocks_counter;
  1023. struct percpu_counter s_freeinodes_counter;
  1024. struct percpu_counter s_dirs_counter;
  1025. struct percpu_counter s_dirtyblocks_counter;
  1026. struct blockgroup_lock *s_blockgroup_lock;
  1027. struct proc_dir_entry *s_proc;
  1028. struct kobject s_kobj;
  1029. struct completion s_kobj_unregister;
  1030. /* Journaling */
  1031. struct journal_s *s_journal;
  1032. struct list_head s_orphan;
  1033. struct mutex s_orphan_lock;
  1034. unsigned long s_resize_flags; /* Flags indicating if there
  1035. is a resizer */
  1036. unsigned long s_commit_interval;
  1037. u32 s_max_batch_time;
  1038. u32 s_min_batch_time;
  1039. struct block_device *journal_bdev;
  1040. #ifdef CONFIG_JBD2_DEBUG
  1041. struct timer_list turn_ro_timer; /* For turning read-only (crash simulation) */
  1042. wait_queue_head_t ro_wait_queue; /* For people waiting for the fs to go read-only */
  1043. #endif
  1044. #ifdef CONFIG_QUOTA
  1045. char *s_qf_names[MAXQUOTAS]; /* Names of quota files with journalled quota */
  1046. int s_jquota_fmt; /* Format of quota to use */
  1047. #endif
  1048. unsigned int s_want_extra_isize; /* New inodes should reserve # bytes */
  1049. struct rb_root system_blks;
  1050. #ifdef EXTENTS_STATS
  1051. /* ext4 extents stats */
  1052. unsigned long s_ext_min;
  1053. unsigned long s_ext_max;
  1054. unsigned long s_depth_max;
  1055. spinlock_t s_ext_stats_lock;
  1056. unsigned long s_ext_blocks;
  1057. unsigned long s_ext_extents;
  1058. #endif
  1059. /* ext4 extent cache stats */
  1060. unsigned long extent_cache_hits;
  1061. unsigned long extent_cache_misses;
  1062. /* for buddy allocator */
  1063. struct ext4_group_info ***s_group_info;
  1064. struct inode *s_buddy_cache;
  1065. spinlock_t s_md_lock;
  1066. unsigned short *s_mb_offsets;
  1067. unsigned int *s_mb_maxs;
  1068. /* tunables */
  1069. unsigned long s_stripe;
  1070. unsigned int s_mb_stream_request;
  1071. unsigned int s_mb_max_to_scan;
  1072. unsigned int s_mb_min_to_scan;
  1073. unsigned int s_mb_stats;
  1074. unsigned int s_mb_order2_reqs;
  1075. unsigned int s_mb_group_prealloc;
  1076. unsigned int s_max_writeback_mb_bump;
  1077. /* where last allocation was done - for stream allocation */
  1078. unsigned long s_mb_last_group;
  1079. unsigned long s_mb_last_start;
  1080. /* stats for buddy allocator */
  1081. atomic_t s_bal_reqs; /* number of reqs with len > 1 */
  1082. atomic_t s_bal_success; /* we found long enough chunks */
  1083. atomic_t s_bal_allocated; /* in blocks */
  1084. atomic_t s_bal_ex_scanned; /* total extents scanned */
  1085. atomic_t s_bal_goals; /* goal hits */
  1086. atomic_t s_bal_breaks; /* too long searches */
  1087. atomic_t s_bal_2orders; /* 2^order hits */
  1088. spinlock_t s_bal_lock;
  1089. unsigned long s_mb_buddies_generated;
  1090. unsigned long long s_mb_generation_time;
  1091. atomic_t s_mb_lost_chunks;
  1092. atomic_t s_mb_preallocated;
  1093. atomic_t s_mb_discarded;
  1094. atomic_t s_lock_busy;
  1095. /* locality groups */
  1096. struct ext4_locality_group __percpu *s_locality_groups;
  1097. /* for write statistics */
  1098. unsigned long s_sectors_written_start;
  1099. u64 s_kbytes_written;
  1100. unsigned int s_log_groups_per_flex;
  1101. struct flex_groups *s_flex_groups;
  1102. /* workqueue for dio unwritten */
  1103. struct workqueue_struct *dio_unwritten_wq;
  1104. /* timer for periodic error stats printing */
  1105. struct timer_list s_err_report;
  1106. /* Lazy inode table initialization info */
  1107. struct ext4_li_request *s_li_request;
  1108. /* Wait multiplier for lazy initialization thread */
  1109. unsigned int s_li_wait_mult;
  1110. /* Kernel thread for multiple mount protection */
  1111. struct task_struct *s_mmp_tsk;
  1112. /* record the last minlen when FITRIM is called. */
  1113. atomic_t s_last_trim_minblks;
  1114. };
  1115. static inline struct ext4_sb_info *EXT4_SB(struct super_block *sb)
  1116. {
  1117. return sb->s_fs_info;
  1118. }
  1119. static inline struct ext4_inode_info *EXT4_I(struct inode *inode)
  1120. {
  1121. return container_of(inode, struct ext4_inode_info, vfs_inode);
  1122. }
  1123. static inline struct timespec ext4_current_time(struct inode *inode)
  1124. {
  1125. return (inode->i_sb->s_time_gran < NSEC_PER_SEC) ?
  1126. current_fs_time(inode->i_sb) : CURRENT_TIME_SEC;
  1127. }
  1128. static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino)
  1129. {
  1130. return ino == EXT4_ROOT_INO ||
  1131. ino == EXT4_JOURNAL_INO ||
  1132. ino == EXT4_RESIZE_INO ||
  1133. (ino >= EXT4_FIRST_INO(sb) &&
  1134. ino <= le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count));
  1135. }
  1136. /*
  1137. * Inode dynamic state flags
  1138. */
  1139. enum {
  1140. EXT4_STATE_JDATA, /* journaled data exists */
  1141. EXT4_STATE_NEW, /* inode is newly created */
  1142. EXT4_STATE_XATTR, /* has in-inode xattrs */
  1143. EXT4_STATE_NO_EXPAND, /* No space for expansion */
  1144. EXT4_STATE_DA_ALLOC_CLOSE, /* Alloc DA blks on close */
  1145. EXT4_STATE_EXT_MIGRATE, /* Inode is migrating */
  1146. EXT4_STATE_DIO_UNWRITTEN, /* need convert on dio done*/
  1147. EXT4_STATE_NEWENTRY, /* File just added to dir */
  1148. EXT4_STATE_DELALLOC_RESERVED, /* blks already reserved for delalloc */
  1149. };
  1150. #define EXT4_INODE_BIT_FNS(name, field, offset) \
  1151. static inline int ext4_test_inode_##name(struct inode *inode, int bit) \
  1152. { \
  1153. return test_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
  1154. } \
  1155. static inline void ext4_set_inode_##name(struct inode *inode, int bit) \
  1156. { \
  1157. set_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
  1158. } \
  1159. static inline void ext4_clear_inode_##name(struct inode *inode, int bit) \
  1160. { \
  1161. clear_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
  1162. }
  1163. EXT4_INODE_BIT_FNS(flag, flags, 0)
  1164. #if (BITS_PER_LONG < 64)
  1165. EXT4_INODE_BIT_FNS(state, state_flags, 0)
  1166. static inline void ext4_clear_state_flags(struct ext4_inode_info *ei)
  1167. {
  1168. (ei)->i_state_flags = 0;
  1169. }
  1170. #else
  1171. EXT4_INODE_BIT_FNS(state, flags, 32)
  1172. static inline void ext4_clear_state_flags(struct ext4_inode_info *ei)
  1173. {
  1174. /* We depend on the fact that callers will set i_flags */
  1175. }
  1176. #endif
  1177. #else
  1178. /* Assume that user mode programs are passing in an ext4fs superblock, not
  1179. * a kernel struct super_block. This will allow us to call the feature-test
  1180. * macros from user land. */
  1181. #define EXT4_SB(sb) (sb)
  1182. #endif
  1183. #define NEXT_ORPHAN(inode) EXT4_I(inode)->i_dtime
  1184. /*
  1185. * Codes for operating systems
  1186. */
  1187. #define EXT4_OS_LINUX 0
  1188. #define EXT4_OS_HURD 1
  1189. #define EXT4_OS_MASIX 2
  1190. #define EXT4_OS_FREEBSD 3
  1191. #define EXT4_OS_LITES 4
  1192. /*
  1193. * Revision levels
  1194. */
  1195. #define EXT4_GOOD_OLD_REV 0 /* The good old (original) format */
  1196. #define EXT4_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
  1197. #define EXT4_CURRENT_REV EXT4_GOOD_OLD_REV
  1198. #define EXT4_MAX_SUPP_REV EXT4_DYNAMIC_REV
  1199. #define EXT4_GOOD_OLD_INODE_SIZE 128
  1200. /*
  1201. * Feature set definitions
  1202. */
  1203. #define EXT4_HAS_COMPAT_FEATURE(sb,mask) \
  1204. ((EXT4_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask)) != 0)
  1205. #define EXT4_HAS_RO_COMPAT_FEATURE(sb,mask) \
  1206. ((EXT4_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask)) != 0)
  1207. #define EXT4_HAS_INCOMPAT_FEATURE(sb,mask) \
  1208. ((EXT4_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask)) != 0)
  1209. #define EXT4_SET_COMPAT_FEATURE(sb,mask) \
  1210. EXT4_SB(sb)->s_es->s_feature_compat |= cpu_to_le32(mask)
  1211. #define EXT4_SET_RO_COMPAT_FEATURE(sb,mask) \
  1212. EXT4_SB(sb)->s_es->s_feature_ro_compat |= cpu_to_le32(mask)
  1213. #define EXT4_SET_INCOMPAT_FEATURE(sb,mask) \
  1214. EXT4_SB(sb)->s_es->s_feature_incompat |= cpu_to_le32(mask)
  1215. #define EXT4_CLEAR_COMPAT_FEATURE(sb,mask) \
  1216. EXT4_SB(sb)->s_es->s_feature_compat &= ~cpu_to_le32(mask)
  1217. #define EXT4_CLEAR_RO_COMPAT_FEATURE(sb,mask) \
  1218. EXT4_SB(sb)->s_es->s_feature_ro_compat &= ~cpu_to_le32(mask)
  1219. #define EXT4_CLEAR_INCOMPAT_FEATURE(sb,mask) \
  1220. EXT4_SB(sb)->s_es->s_feature_incompat &= ~cpu_to_le32(mask)
  1221. #define EXT4_FEATURE_COMPAT_DIR_PREALLOC 0x0001
  1222. #define EXT4_FEATURE_COMPAT_IMAGIC_INODES 0x0002
  1223. #define EXT4_FEATURE_COMPAT_HAS_JOURNAL 0x0004
  1224. #define EXT4_FEATURE_COMPAT_EXT_ATTR 0x0008
  1225. #define EXT4_FEATURE_COMPAT_RESIZE_INODE 0x0010
  1226. #define EXT4_FEATURE_COMPAT_DIR_INDEX 0x0020
  1227. #define EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
  1228. #define EXT4_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
  1229. #define EXT4_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
  1230. #define EXT4_FEATURE_RO_COMPAT_HUGE_FILE 0x0008
  1231. #define EXT4_FEATURE_RO_COMPAT_GDT_CSUM 0x0010
  1232. #define EXT4_FEATURE_RO_COMPAT_DIR_NLINK 0x0020
  1233. #define EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 0x0040
  1234. #define EXT4_FEATURE_RO_COMPAT_QUOTA 0x0100
  1235. #define EXT4_FEATURE_INCOMPAT_COMPRESSION 0x0001
  1236. #define EXT4_FEATURE_INCOMPAT_FILETYPE 0x0002
  1237. #define EXT4_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */
  1238. #define EXT4_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */
  1239. #define EXT4_FEATURE_INCOMPAT_META_BG 0x0010
  1240. #define EXT4_FEATURE_INCOMPAT_EXTENTS 0x0040 /* extents support */
  1241. #define EXT4_FEATURE_INCOMPAT_64BIT 0x0080
  1242. #define EXT4_FEATURE_INCOMPAT_MMP 0x0100
  1243. #define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
  1244. #define EXT4_FEATURE_INCOMPAT_EA_INODE 0x0400 /* EA in inode */
  1245. #define EXT4_FEATURE_INCOMPAT_DIRDATA 0x1000 /* data in dirent */
  1246. #define EXT2_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
  1247. #define EXT2_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
  1248. EXT4_FEATURE_INCOMPAT_META_BG)
  1249. #define EXT2_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
  1250. EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
  1251. EXT4_FEATURE_RO_COMPAT_BTREE_DIR)
  1252. #define EXT3_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
  1253. #define EXT3_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
  1254. EXT4_FEATURE_INCOMPAT_RECOVER| \
  1255. EXT4_FEATURE_INCOMPAT_META_BG)
  1256. #define EXT3_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
  1257. EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
  1258. EXT4_FEATURE_RO_COMPAT_BTREE_DIR)
  1259. #define EXT4_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR
  1260. #define EXT4_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
  1261. EXT4_FEATURE_INCOMPAT_RECOVER| \
  1262. EXT4_FEATURE_INCOMPAT_META_BG| \
  1263. EXT4_FEATURE_INCOMPAT_EXTENTS| \
  1264. EXT4_FEATURE_INCOMPAT_64BIT| \
  1265. EXT4_FEATURE_INCOMPAT_FLEX_BG| \
  1266. EXT4_FEATURE_INCOMPAT_MMP)
  1267. #define EXT4_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
  1268. EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
  1269. EXT4_FEATURE_RO_COMPAT_GDT_CSUM| \
  1270. EXT4_FEATURE_RO_COMPAT_DIR_NLINK | \
  1271. EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE | \
  1272. EXT4_FEATURE_RO_COMPAT_BTREE_DIR |\
  1273. EXT4_FEATURE_RO_COMPAT_HUGE_FILE)
  1274. /*
  1275. * Default values for user and/or group using reserved blocks
  1276. */
  1277. #define EXT4_DEF_RESUID 0
  1278. #define EXT4_DEF_RESGID 0
  1279. #define EXT4_DEF_INODE_READAHEAD_BLKS 32
  1280. /*
  1281. * Default mount options
  1282. */
  1283. #define EXT4_DEFM_DEBUG 0x0001
  1284. #define EXT4_DEFM_BSDGROUPS 0x0002
  1285. #define EXT4_DEFM_XATTR_USER 0x0004
  1286. #define EXT4_DEFM_ACL 0x0008
  1287. #define EXT4_DEFM_UID16 0x0010
  1288. #define EXT4_DEFM_JMODE 0x0060
  1289. #define EXT4_DEFM_JMODE_DATA 0x0020
  1290. #define EXT4_DEFM_JMODE_ORDERED 0x0040
  1291. #define EXT4_DEFM_JMODE_WBACK 0x0060
  1292. #define EXT4_DEFM_NOBARRIER 0x0100
  1293. #define EXT4_DEFM_BLOCK_VALIDITY 0x0200
  1294. #define EXT4_DEFM_DISCARD 0x0400
  1295. #define EXT4_DEFM_NODELALLOC 0x0800
  1296. /*
  1297. * Default journal batch times
  1298. */
  1299. #define EXT4_DEF_MIN_BATCH_TIME 0
  1300. #define EXT4_DEF_MAX_BATCH_TIME 15000 /* 15ms */
  1301. /*
  1302. * Minimum number of groups in a flexgroup before we separate out
  1303. * directories into the first block group of a flexgroup
  1304. */
  1305. #define EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME 4
  1306. /*
  1307. * Structure of a directory entry
  1308. */
  1309. #define EXT4_NAME_LEN 255
  1310. struct ext4_dir_entry {
  1311. __le32 inode; /* Inode number */
  1312. __le16 rec_len; /* Directory entry length */
  1313. __le16 name_len; /* Name length */
  1314. char name[EXT4_NAME_LEN]; /* File name */
  1315. };
  1316. /*
  1317. * The new version of the directory entry. Since EXT4 structures are
  1318. * stored in intel byte order, and the name_len field could never be
  1319. * bigger than 255 chars, it's safe to reclaim the extra byte for the
  1320. * file_type field.
  1321. */
  1322. struct ext4_dir_entry_2 {
  1323. __le32 inode; /* Inode number */
  1324. __le16 rec_len; /* Directory entry length */
  1325. __u8 name_len; /* Name length */
  1326. __u8 file_type;
  1327. char name[EXT4_NAME_LEN]; /* File name */
  1328. };
  1329. /*
  1330. * Ext4 directory file types. Only the low 3 bits are used. The
  1331. * other bits are reserved for now.
  1332. */
  1333. #define EXT4_FT_UNKNOWN 0
  1334. #define EXT4_FT_REG_FILE 1
  1335. #define EXT4_FT_DIR 2
  1336. #define EXT4_FT_CHRDEV 3
  1337. #define EXT4_FT_BLKDEV 4
  1338. #define EXT4_FT_FIFO 5
  1339. #define EXT4_FT_SOCK 6
  1340. #define EXT4_FT_SYMLINK 7
  1341. #define EXT4_FT_MAX 8
  1342. /*
  1343. * EXT4_DIR_PAD defines the directory entries boundaries
  1344. *
  1345. * NOTE: It must be a multiple of 4
  1346. */
  1347. #define EXT4_DIR_PAD 4
  1348. #define EXT4_DIR_ROUND (EXT4_DIR_PAD - 1)
  1349. #define EXT4_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT4_DIR_ROUND) & \
  1350. ~EXT4_DIR_ROUND)
  1351. #define EXT4_MAX_REC_LEN ((1<<16)-1)
  1352. /*
  1353. * If we ever get support for fs block sizes > page_size, we'll need
  1354. * to remove the #if statements in the next two functions...
  1355. */
  1356. static inline unsigned int
  1357. ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize)
  1358. {
  1359. unsigned len = le16_to_cpu(dlen);
  1360. #if (PAGE_CACHE_SIZE >= 65536)
  1361. if (len == EXT4_MAX_REC_LEN || len == 0)
  1362. return blocksize;
  1363. return (len & 65532) | ((len & 3) << 16);
  1364. #else
  1365. return len;
  1366. #endif
  1367. }
  1368. static inline __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize)
  1369. {
  1370. if ((len > blocksize) || (blocksize > (1 << 18)) || (len & 3))
  1371. BUG();
  1372. #if (PAGE_CACHE_SIZE >= 65536)
  1373. if (len < 65536)
  1374. return cpu_to_le16(len);
  1375. if (len == blocksize) {
  1376. if (blocksize == 65536)
  1377. return cpu_to_le16(EXT4_MAX_REC_LEN);
  1378. else
  1379. return cpu_to_le16(0);
  1380. }
  1381. return cpu_to_le16((len & 65532) | ((len >> 16) & 3));
  1382. #else
  1383. return cpu_to_le16(len);
  1384. #endif
  1385. }
  1386. /*
  1387. * Hash Tree Directory indexing
  1388. * (c) Daniel Phillips, 2001
  1389. */
  1390. #define is_dx(dir) (EXT4_HAS_COMPAT_FEATURE(dir->i_sb, \
  1391. EXT4_FEATURE_COMPAT_DIR_INDEX) && \
  1392. ext4_test_inode_flag((dir), EXT4_INODE_INDEX))
  1393. #define EXT4_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT4_LINK_MAX)
  1394. #define EXT4_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
  1395. /* Legal values for the dx_root hash_version field: */
  1396. #define DX_HASH_LEGACY 0
  1397. #define DX_HASH_HALF_MD4 1
  1398. #define DX_HASH_TEA 2
  1399. #define DX_HASH_LEGACY_UNSIGNED 3
  1400. #define DX_HASH_HALF_MD4_UNSIGNED 4
  1401. #define DX_HASH_TEA_UNSIGNED 5
  1402. #ifdef __KERNEL__
  1403. /* hash info structure used by the directory hash */
  1404. struct dx_hash_info
  1405. {
  1406. u32 hash;
  1407. u32 minor_hash;
  1408. int hash_version;
  1409. u32 *seed;
  1410. };
  1411. #define EXT4_HTREE_EOF 0x7fffffff
  1412. /*
  1413. * Control parameters used by ext4_htree_next_block
  1414. */
  1415. #define HASH_NB_ALWAYS 1
  1416. /*
  1417. * Describe an inode's exact location on disk and in memory
  1418. */
  1419. struct ext4_iloc
  1420. {
  1421. struct buffer_head *bh;
  1422. unsigned long offset;
  1423. ext4_group_t block_group;
  1424. };
  1425. static inline struct ext4_inode *ext4_raw_inode(struct ext4_iloc *iloc)
  1426. {
  1427. return (struct ext4_inode *) (iloc->bh->b_data + iloc->offset);
  1428. }
  1429. /*
  1430. * This structure is stuffed into the struct file's private_data field
  1431. * for directories. It is where we put information so that we can do
  1432. * readdir operations in hash tree order.
  1433. */
  1434. struct dir_private_info {
  1435. struct rb_root root;
  1436. struct rb_node *curr_node;
  1437. struct fname *extra_fname;
  1438. loff_t last_pos;
  1439. __u32 curr_hash;
  1440. __u32 curr_minor_hash;
  1441. __u32 next_hash;
  1442. };
  1443. /* calculate the first block number of the group */
  1444. static inline ext4_fsblk_t
  1445. ext4_group_first_block_no(struct super_block *sb, ext4_group_t group_no)
  1446. {
  1447. return group_no * (ext4_fsblk_t)EXT4_BLOCKS_PER_GROUP(sb) +
  1448. le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
  1449. }
  1450. /*
  1451. * Special error return code only used by dx_probe() and its callers.
  1452. */
  1453. #define ERR_BAD_DX_DIR -75000
  1454. void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr,
  1455. ext4_group_t *blockgrpp, ext4_grpblk_t *offsetp);
  1456. /*
  1457. * Timeout and state flag for lazy initialization inode thread.
  1458. */
  1459. #define EXT4_DEF_LI_WAIT_MULT 10
  1460. #define EXT4_DEF_LI_MAX_START_DELAY 5
  1461. #define EXT4_LAZYINIT_QUIT 0x0001
  1462. #define EXT4_LAZYINIT_RUNNING 0x0002
  1463. /*
  1464. * Lazy inode table initialization info
  1465. */
  1466. struct ext4_lazy_init {
  1467. unsigned long li_state;
  1468. struct list_head li_request_list;
  1469. struct mutex li_list_mtx;
  1470. };
  1471. struct ext4_li_request {
  1472. struct super_block *lr_super;
  1473. struct ext4_sb_info *lr_sbi;
  1474. ext4_group_t lr_next_group;
  1475. struct list_head lr_request;
  1476. unsigned long lr_next_sched;
  1477. unsigned long lr_timeout;
  1478. };
  1479. struct ext4_features {
  1480. struct kobject f_kobj;
  1481. struct completion f_kobj_unregister;
  1482. };
  1483. /*
  1484. * This structure will be used for multiple mount protection. It will be
  1485. * written into the block number saved in the s_mmp_block field in the
  1486. * superblock. Programs that check MMP should assume that if
  1487. * SEQ_FSCK (or any unknown code above SEQ_MAX) is present then it is NOT safe
  1488. * to use the filesystem, regardless of how old the timestamp is.
  1489. */
  1490. #define EXT4_MMP_MAGIC 0x004D4D50U /* ASCII for MMP */
  1491. #define EXT4_MMP_SEQ_CLEAN 0xFF4D4D50U /* mmp_seq value for clean unmount */
  1492. #define EXT4_MMP_SEQ_FSCK 0xE24D4D50U /* mmp_seq value when being fscked */
  1493. #define EXT4_MMP_SEQ_MAX 0xE24D4D4FU /* maximum valid mmp_seq value */
  1494. struct mmp_struct {
  1495. __le32 mmp_magic; /* Magic number for MMP */
  1496. __le32 mmp_seq; /* Sequence no. updated periodically */
  1497. /*
  1498. * mmp_time, mmp_nodename & mmp_bdevname are only used for information
  1499. * purposes and do not affect the correctness of the algorithm
  1500. */
  1501. __le64 mmp_time; /* Time last updated */
  1502. char mmp_nodename[64]; /* Node which last updated MMP block */
  1503. char mmp_bdevname[32]; /* Bdev which last updated MMP block */
  1504. /*
  1505. * mmp_check_interval is used to verify if the MMP block has been
  1506. * updated on the block device. The value is updated based on the
  1507. * maximum time to write the MMP block during an update cycle.
  1508. */
  1509. __le16 mmp_check_interval;
  1510. __le16 mmp_pad1;
  1511. __le32 mmp_pad2[227];
  1512. };
  1513. /* arguments passed to the mmp thread */
  1514. struct mmpd_data {
  1515. struct buffer_head *bh; /* bh from initial read_mmp_block() */
  1516. struct super_block *sb; /* super block of the fs */
  1517. };
  1518. /*
  1519. * Check interval multiplier
  1520. * The MMP block is written every update interval and initially checked every
  1521. * update interval x the multiplier (the value is then adapted based on the
  1522. * write latency). The reason is that writes can be delayed under load and we
  1523. * don't want readers to incorrectly assume that the filesystem is no longer
  1524. * in use.
  1525. */
  1526. #define EXT4_MMP_CHECK_MULT 2UL
  1527. /*
  1528. * Minimum interval for MMP checking in seconds.
  1529. */
  1530. #define EXT4_MMP_MIN_CHECK_INTERVAL 5UL
  1531. /*
  1532. * Maximum interval for MMP checking in seconds.
  1533. */
  1534. #define EXT4_MMP_MAX_CHECK_INTERVAL 300UL
  1535. /*
  1536. * Function prototypes
  1537. */
  1538. /*
  1539. * Ok, these declarations are also in <linux/kernel.h> but none of the
  1540. * ext4 source programs needs to include it so they are duplicated here.
  1541. */
  1542. # define NORET_TYPE /**/
  1543. # define ATTRIB_NORET __attribute__((noreturn))
  1544. # define NORET_AND noreturn,
  1545. /* bitmap.c */
  1546. extern unsigned int ext4_count_free(struct buffer_head *, unsigned);
  1547. /* balloc.c */
  1548. extern unsigned int ext4_block_group(struct super_block *sb,
  1549. ext4_fsblk_t blocknr);
  1550. extern ext4_grpblk_t ext4_block_group_offset(struct super_block *sb,
  1551. ext4_fsblk_t blocknr);
  1552. extern int ext4_bg_has_super(struct super_block *sb, ext4_group_t group);
  1553. extern unsigned long ext4_bg_num_gdb(struct super_block *sb,
  1554. ext4_group_t group);
  1555. extern ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
  1556. ext4_fsblk_t goal,
  1557. unsigned int flags,
  1558. unsigned long *count,
  1559. int *errp);
  1560. extern int ext4_claim_free_blocks(struct ext4_sb_info *sbi,
  1561. s64 nblocks, unsigned int flags);
  1562. extern ext4_fsblk_t ext4_count_free_blocks(struct super_block *);
  1563. extern void ext4_check_blocks_bitmap(struct super_block *);
  1564. extern struct ext4_group_desc * ext4_get_group_desc(struct super_block * sb,
  1565. ext4_group_t block_group,
  1566. struct buffer_head ** bh);
  1567. extern int ext4_should_retry_alloc(struct super_block *sb, int *retries);
  1568. struct buffer_head *ext4_read_block_bitmap(struct super_block *sb,
  1569. ext4_group_t block_group);
  1570. extern unsigned ext4_init_block_bitmap(struct super_block *sb,
  1571. struct buffer_head *bh,
  1572. ext4_group_t group,
  1573. struct ext4_group_desc *desc);
  1574. #define ext4_free_blocks_after_init(sb, group, desc) \
  1575. ext4_init_block_bitmap(sb, NULL, group, desc)
  1576. ext4_fsblk_t ext4_inode_to_goal_block(struct inode *);
  1577. /* dir.c */
  1578. extern int __ext4_check_dir_entry(const char *, unsigned int, struct inode *,
  1579. struct file *,
  1580. struct ext4_dir_entry_2 *,
  1581. struct buffer_head *, unsigned int);
  1582. #define ext4_check_dir_entry(dir, filp, de, bh, offset) \
  1583. unlikely(__ext4_check_dir_entry(__func__, __LINE__, (dir), (filp), \
  1584. (de), (bh), (offset)))
  1585. extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
  1586. __u32 minor_hash,
  1587. struct ext4_dir_entry_2 *dirent);
  1588. extern void ext4_htree_free_dir_info(struct dir_private_info *p);
  1589. /* fsync.c */
  1590. extern int ext4_sync_file(struct file *, loff_t, loff_t, int);
  1591. extern int ext4_flush_completed_IO(struct inode *);
  1592. /* hash.c */
  1593. extern int ext4fs_dirhash(const char *name, int len, struct
  1594. dx_hash_info *hinfo);
  1595. /* ialloc.c */
  1596. extern struct inode *ext4_new_inode(handle_t *, struct inode *, int,
  1597. const struct qstr *qstr, __u32 goal);
  1598. extern void ext4_free_inode(handle_t *, struct inode *);
  1599. extern struct inode * ext4_orphan_get(struct super_block *, unsigned long);
  1600. extern unsigned long ext4_count_free_inodes(struct super_block *);
  1601. extern unsigned long ext4_count_dirs(struct super_block *);
  1602. extern void ext4_check_inodes_bitmap(struct super_block *);
  1603. extern void ext4_mark_bitmap_end(int start_bit, int end_bit, char *bitmap);
  1604. extern int ext4_init_inode_table(struct super_block *sb,
  1605. ext4_group_t group, int barrier);
  1606. /* mballoc.c */
  1607. extern long ext4_mb_stats;
  1608. extern long ext4_mb_max_to_scan;
  1609. extern int ext4_mb_init(struct super_block *, int);
  1610. extern int ext4_mb_release(struct super_block *);
  1611. extern ext4_fsblk_t ext4_mb_new_blocks(handle_t *,
  1612. struct ext4_allocation_request *, int *);
  1613. extern int ext4_mb_reserve_blocks(struct super_block *, int);
  1614. extern void ext4_discard_preallocations(struct inode *);
  1615. extern int __init ext4_init_mballoc(void);
  1616. extern void ext4_exit_mballoc(void);
  1617. extern void ext4_free_blocks(handle_t *handle, struct inode *inode,
  1618. struct buffer_head *bh, ext4_fsblk_t block,
  1619. unsigned long count, int flags);
  1620. extern int ext4_mb_add_groupinfo(struct super_block *sb,
  1621. ext4_group_t i, struct ext4_group_desc *desc);
  1622. extern int ext4_group_add_blocks(handle_t *handle, struct super_block *sb,
  1623. ext4_fsblk_t block, unsigned long count);
  1624. extern int ext4_trim_fs(struct super_block *, struct fstrim_range *);
  1625. /* inode.c */
  1626. struct buffer_head *ext4_getblk(handle_t *, struct inode *,
  1627. ext4_lblk_t, int, int *);
  1628. struct buffer_head *ext4_bread(handle_t *, struct inode *,
  1629. ext4_lblk_t, int, int *);
  1630. int ext4_get_block(struct inode *inode, sector_t iblock,
  1631. struct buffer_head *bh_result, int create);
  1632. extern struct inode *ext4_iget(struct super_block *, unsigned long);
  1633. extern int ext4_write_inode(struct inode *, struct writeback_control *);
  1634. extern int ext4_setattr(struct dentry *, struct iattr *);
  1635. extern int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry,
  1636. struct kstat *stat);
  1637. extern void ext4_evict_inode(struct inode *);
  1638. extern void ext4_clear_inode(struct inode *);
  1639. extern int ext4_sync_inode(handle_t *, struct inode *);
  1640. extern void ext4_dirty_inode(struct inode *, int);
  1641. extern int ext4_change_inode_journal_flag(struct inode *, int);
  1642. extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *);
  1643. extern int ext4_can_truncate(struct inode *inode);
  1644. extern void ext4_truncate(struct inode *);
  1645. extern int ext4_punch_hole(struct file *file, loff_t offset, loff_t length);
  1646. extern int ext4_truncate_restart_trans(handle_t *, struct inode *, int nblocks);
  1647. extern void ext4_set_inode_flags(struct inode *);
  1648. extern void ext4_get_inode_flags(struct ext4_inode_info *);
  1649. extern int ext4_alloc_da_blocks(struct inode *inode);
  1650. extern void ext4_set_aops(struct inode *inode);
  1651. extern int ext4_writepage_trans_blocks(struct inode *);
  1652. extern int ext4_chunk_trans_blocks(struct inode *, int nrblocks);
  1653. extern int ext4_block_truncate_page(handle_t *handle,
  1654. struct address_space *mapping, loff_t from);
  1655. extern int ext4_block_zero_page_range(handle_t *handle,
  1656. struct address_space *mapping, loff_t from, loff_t length);
  1657. extern int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
  1658. extern qsize_t *ext4_get_reserved_space(struct inode *inode);
  1659. extern void ext4_da_update_reserve_space(struct inode *inode,
  1660. int used, int quota_claim);
  1661. /* indirect.c */
  1662. extern int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
  1663. struct ext4_map_blocks *map, int flags);
  1664. extern ssize_t ext4_ind_direct_IO(int rw, struct kiocb *iocb,
  1665. const struct iovec *iov, loff_t offset,
  1666. unsigned long nr_segs);
  1667. extern int ext4_ind_calc_metadata_amount(struct inode *inode, sector_t lblock);
  1668. extern int ext4_ind_trans_blocks(struct inode *inode, int nrblocks, int chunk);
  1669. extern void ext4_ind_truncate(struct inode *inode);
  1670. /* ioctl.c */
  1671. extern long ext4_ioctl(struct file *, unsigned int, unsigned long);
  1672. extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long);
  1673. /* migrate.c */
  1674. extern int ext4_ext_migrate(struct inode *);
  1675. /* namei.c */
  1676. extern int ext4_orphan_add(handle_t *, struct inode *);
  1677. extern int ext4_orphan_del(handle_t *, struct inode *);
  1678. extern int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
  1679. __u32 start_minor_hash, __u32 *next_hash);
  1680. /* resize.c */
  1681. extern int ext4_group_add(struct super_block *sb,
  1682. struct ext4_new_group_data *input);
  1683. extern int ext4_group_extend(struct super_block *sb,
  1684. struct ext4_super_block *es,
  1685. ext4_fsblk_t n_blocks_count);
  1686. /* super.c */
  1687. extern void *ext4_kvmalloc(size_t size, gfp_t flags);
  1688. extern void *ext4_kvzalloc(size_t size, gfp_t flags);
  1689. extern void ext4_kvfree(void *ptr);
  1690. extern void __ext4_error(struct super_block *, const char *, unsigned int,
  1691. const char *, ...)
  1692. __attribute__ ((format (printf, 4, 5)));
  1693. #define ext4_error(sb, message...) __ext4_error(sb, __func__, \
  1694. __LINE__, ## message)
  1695. extern void ext4_error_inode(struct inode *, const char *, unsigned int,
  1696. ext4_fsblk_t, const char *, ...)
  1697. __attribute__ ((format (printf, 5, 6)));
  1698. extern void ext4_error_file(struct file *, const char *, unsigned int,
  1699. ext4_fsblk_t, const char *, ...)
  1700. __attribute__ ((format (printf, 5, 6)));
  1701. extern void __ext4_std_error(struct super_block *, const char *,
  1702. unsigned int, int);
  1703. extern void __ext4_abort(struct super_block *, const char *, unsigned int,
  1704. const char *, ...)
  1705. __attribute__ ((format (printf, 4, 5)));
  1706. #define ext4_abort(sb, message...) __ext4_abort(sb, __func__, \
  1707. __LINE__, ## message)
  1708. extern void __ext4_warning(struct super_block *, const char *, unsigned int,
  1709. const char *, ...)
  1710. __attribute__ ((format (printf, 4, 5)));
  1711. #define ext4_warning(sb, message...) __ext4_warning(sb, __func__, \
  1712. __LINE__, ## message)
  1713. extern void ext4_msg(struct super_block *, const char *, const char *, ...)
  1714. __attribute__ ((format (printf, 3, 4)));
  1715. extern void __dump_mmp_msg(struct super_block *, struct mmp_struct *mmp,
  1716. const char *, unsigned int, const char *);
  1717. #define dump_mmp_msg(sb, mmp, msg) __dump_mmp_msg(sb, mmp, __func__, \
  1718. __LINE__, msg)
  1719. extern void __ext4_grp_locked_error(const char *, unsigned int, \
  1720. struct super_block *, ext4_group_t, \
  1721. unsigned long, ext4_fsblk_t, \
  1722. const char *, ...)
  1723. __attribute__ ((format (printf, 7, 8)));
  1724. #define ext4_grp_locked_error(sb, grp, message...) \
  1725. __ext4_grp_locked_error(__func__, __LINE__, (sb), (grp), ## message)
  1726. extern void ext4_update_dynamic_rev(struct super_block *sb);
  1727. extern int ext4_update_compat_feature(handle_t *handle, struct super_block *sb,
  1728. __u32 compat);
  1729. extern int ext4_update_rocompat_feature(handle_t *handle,
  1730. struct super_block *sb, __u32 rocompat);
  1731. extern int ext4_update_incompat_feature(handle_t *handle,
  1732. struct super_block *sb, __u32 incompat);
  1733. extern ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
  1734. struct ext4_group_desc *bg);
  1735. extern ext4_fsblk_t ext4_inode_bitmap(struct super_block *sb,
  1736. struct ext4_group_desc *bg);
  1737. extern ext4_fsblk_t ext4_inode_table(struct super_block *sb,
  1738. struct ext4_group_desc *bg);
  1739. extern __u32 ext4_free_blks_count(struct super_block *sb,
  1740. struct ext4_group_desc *bg);
  1741. extern __u32 ext4_free_inodes_count(struct super_block *sb,
  1742. struct ext4_group_desc *bg);
  1743. extern __u32 ext4_used_dirs_count(struct super_block *sb,
  1744. struct ext4_group_desc *bg);
  1745. extern __u32 ext4_itable_unused_count(struct super_block *sb,
  1746. struct ext4_group_desc *bg);
  1747. extern void ext4_block_bitmap_set(struct super_block *sb,
  1748. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1749. extern void ext4_inode_bitmap_set(struct super_block *sb,
  1750. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1751. extern void ext4_inode_table_set(struct super_block *sb,
  1752. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1753. extern void ext4_free_blks_set(struct super_block *sb,
  1754. struct ext4_group_desc *bg, __u32 count);
  1755. extern void ext4_free_inodes_set(struct super_block *sb,
  1756. struct ext4_group_desc *bg, __u32 count);
  1757. extern void ext4_used_dirs_set(struct super_block *sb,
  1758. struct ext4_group_desc *bg, __u32 count);
  1759. extern void ext4_itable_unused_set(struct super_block *sb,
  1760. struct ext4_group_desc *bg, __u32 count);
  1761. extern __le16 ext4_group_desc_csum(struct ext4_sb_info *sbi, __u32 group,
  1762. struct ext4_group_desc *gdp);
  1763. extern int ext4_group_desc_csum_verify(struct ext4_sb_info *sbi, __u32 group,
  1764. struct ext4_group_desc *gdp);
  1765. static inline ext4_fsblk_t ext4_blocks_count(struct ext4_super_block *es)
  1766. {
  1767. return ((ext4_fsblk_t)le32_to_cpu(es->s_blocks_count_hi) << 32) |
  1768. le32_to_cpu(es->s_blocks_count_lo);
  1769. }
  1770. static inline ext4_fsblk_t ext4_r_blocks_count(struct ext4_super_block *es)
  1771. {
  1772. return ((ext4_fsblk_t)le32_to_cpu(es->s_r_blocks_count_hi) << 32) |
  1773. le32_to_cpu(es->s_r_blocks_count_lo);
  1774. }
  1775. static inline ext4_fsblk_t ext4_free_blocks_count(struct ext4_super_block *es)
  1776. {
  1777. return ((ext4_fsblk_t)le32_to_cpu(es->s_free_blocks_count_hi) << 32) |
  1778. le32_to_cpu(es->s_free_blocks_count_lo);
  1779. }
  1780. static inline void ext4_blocks_count_set(struct ext4_super_block *es,
  1781. ext4_fsblk_t blk)
  1782. {
  1783. es->s_blocks_count_lo = cpu_to_le32((u32)blk);
  1784. es->s_blocks_count_hi = cpu_to_le32(blk >> 32);
  1785. }
  1786. static inline void ext4_free_blocks_count_set(struct ext4_super_block *es,
  1787. ext4_fsblk_t blk)
  1788. {
  1789. es->s_free_blocks_count_lo = cpu_to_le32((u32)blk);
  1790. es->s_free_blocks_count_hi = cpu_to_le32(blk >> 32);
  1791. }
  1792. static inline void ext4_r_blocks_count_set(struct ext4_super_block *es,
  1793. ext4_fsblk_t blk)
  1794. {
  1795. es->s_r_blocks_count_lo = cpu_to_le32((u32)blk);
  1796. es->s_r_blocks_count_hi = cpu_to_le32(blk >> 32);
  1797. }
  1798. static inline loff_t ext4_isize(struct ext4_inode *raw_inode)
  1799. {
  1800. if (S_ISREG(le16_to_cpu(raw_inode->i_mode)))
  1801. return ((loff_t)le32_to_cpu(raw_inode->i_size_high) << 32) |
  1802. le32_to_cpu(raw_inode->i_size_lo);
  1803. else
  1804. return (loff_t) le32_to_cpu(raw_inode->i_size_lo);
  1805. }
  1806. static inline void ext4_isize_set(struct ext4_inode *raw_inode, loff_t i_size)
  1807. {
  1808. raw_inode->i_size_lo = cpu_to_le32(i_size);
  1809. raw_inode->i_size_high = cpu_to_le32(i_size >> 32);
  1810. }
  1811. static inline
  1812. struct ext4_group_info *ext4_get_group_info(struct super_block *sb,
  1813. ext4_group_t group)
  1814. {
  1815. struct ext4_group_info ***grp_info;
  1816. long indexv, indexh;
  1817. grp_info = EXT4_SB(sb)->s_group_info;
  1818. indexv = group >> (EXT4_DESC_PER_BLOCK_BITS(sb));
  1819. indexh = group & ((EXT4_DESC_PER_BLOCK(sb)) - 1);
  1820. return grp_info[indexv][indexh];
  1821. }
  1822. /*
  1823. * Reading s_groups_count requires using smp_rmb() afterwards. See
  1824. * the locking protocol documented in the comments of ext4_group_add()
  1825. * in resize.c
  1826. */
  1827. static inline ext4_group_t ext4_get_groups_count(struct super_block *sb)
  1828. {
  1829. ext4_group_t ngroups = EXT4_SB(sb)->s_groups_count;
  1830. smp_rmb();
  1831. return ngroups;
  1832. }
  1833. static inline ext4_group_t ext4_flex_group(struct ext4_sb_info *sbi,
  1834. ext4_group_t block_group)
  1835. {
  1836. return block_group >> sbi->s_log_groups_per_flex;
  1837. }
  1838. static inline unsigned int ext4_flex_bg_size(struct ext4_sb_info *sbi)
  1839. {
  1840. return 1 << sbi->s_log_groups_per_flex;
  1841. }
  1842. #define ext4_std_error(sb, errno) \
  1843. do { \
  1844. if ((errno)) \
  1845. __ext4_std_error((sb), __func__, __LINE__, (errno)); \
  1846. } while (0)
  1847. #ifdef CONFIG_SMP
  1848. /* Each CPU can accumulate percpu_counter_batch blocks in their local
  1849. * counters. So we need to make sure we have free blocks more
  1850. * than percpu_counter_batch * nr_cpu_ids. Also add a window of 4 times.
  1851. */
  1852. #define EXT4_FREEBLOCKS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
  1853. #else
  1854. #define EXT4_FREEBLOCKS_WATERMARK 0
  1855. #endif
  1856. static inline void ext4_update_i_disksize(struct inode *inode, loff_t newsize)
  1857. {
  1858. /*
  1859. * XXX: replace with spinlock if seen contended -bzzz
  1860. */
  1861. down_write(&EXT4_I(inode)->i_data_sem);
  1862. if (newsize > EXT4_I(inode)->i_disksize)
  1863. EXT4_I(inode)->i_disksize = newsize;
  1864. up_write(&EXT4_I(inode)->i_data_sem);
  1865. return ;
  1866. }
  1867. struct ext4_group_info {
  1868. unsigned long bb_state;
  1869. struct rb_root bb_free_root;
  1870. ext4_grpblk_t bb_first_free; /* first free block */
  1871. ext4_grpblk_t bb_free; /* total free blocks */
  1872. ext4_grpblk_t bb_fragments; /* nr of freespace fragments */
  1873. ext4_grpblk_t bb_largest_free_order;/* order of largest frag in BG */
  1874. struct list_head bb_prealloc_list;
  1875. #ifdef DOUBLE_CHECK
  1876. void *bb_bitmap;
  1877. #endif
  1878. struct rw_semaphore alloc_sem;
  1879. ext4_grpblk_t bb_counters[]; /* Nr of free power-of-two-block
  1880. * regions, index is order.
  1881. * bb_counters[3] = 5 means
  1882. * 5 free 8-block regions. */
  1883. };
  1884. #define EXT4_GROUP_INFO_NEED_INIT_BIT 0
  1885. #define EXT4_GROUP_INFO_WAS_TRIMMED_BIT 1
  1886. #define EXT4_MB_GRP_NEED_INIT(grp) \
  1887. (test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
  1888. #define EXT4_MB_GRP_WAS_TRIMMED(grp) \
  1889. (test_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
  1890. #define EXT4_MB_GRP_SET_TRIMMED(grp) \
  1891. (set_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
  1892. #define EXT4_MB_GRP_CLEAR_TRIMMED(grp) \
  1893. (clear_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
  1894. #define EXT4_MAX_CONTENTION 8
  1895. #define EXT4_CONTENTION_THRESHOLD 2
  1896. static inline spinlock_t *ext4_group_lock_ptr(struct super_block *sb,
  1897. ext4_group_t group)
  1898. {
  1899. return bgl_lock_ptr(EXT4_SB(sb)->s_blockgroup_lock, group);
  1900. }
  1901. /*
  1902. * Returns true if the filesystem is busy enough that attempts to
  1903. * access the block group locks has run into contention.
  1904. */
  1905. static inline int ext4_fs_is_busy(struct ext4_sb_info *sbi)
  1906. {
  1907. return (atomic_read(&sbi->s_lock_busy) > EXT4_CONTENTION_THRESHOLD);
  1908. }
  1909. static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
  1910. {
  1911. spinlock_t *lock = ext4_group_lock_ptr(sb, group);
  1912. if (spin_trylock(lock))
  1913. /*
  1914. * We're able to grab the lock right away, so drop the
  1915. * lock contention counter.
  1916. */
  1917. atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, -1, 0);
  1918. else {
  1919. /*
  1920. * The lock is busy, so bump the contention counter,
  1921. * and then wait on the spin lock.
  1922. */
  1923. atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, 1,
  1924. EXT4_MAX_CONTENTION);
  1925. spin_lock(lock);
  1926. }
  1927. }
  1928. static inline void ext4_unlock_group(struct super_block *sb,
  1929. ext4_group_t group)
  1930. {
  1931. spin_unlock(ext4_group_lock_ptr(sb, group));
  1932. }
  1933. static inline void ext4_mark_super_dirty(struct super_block *sb)
  1934. {
  1935. if (EXT4_SB(sb)->s_journal == NULL)
  1936. sb->s_dirt =1;
  1937. }
  1938. /*
  1939. * Block validity checking
  1940. */
  1941. #define ext4_check_indirect_blockref(inode, bh) \
  1942. ext4_check_blockref(__func__, __LINE__, inode, \
  1943. (__le32 *)(bh)->b_data, \
  1944. EXT4_ADDR_PER_BLOCK((inode)->i_sb))
  1945. #define ext4_ind_check_inode(inode) \
  1946. ext4_check_blockref(__func__, __LINE__, inode, \
  1947. EXT4_I(inode)->i_data, \
  1948. EXT4_NDIR_BLOCKS)
  1949. /*
  1950. * Inodes and files operations
  1951. */
  1952. /* dir.c */
  1953. extern const struct file_operations ext4_dir_operations;
  1954. /* file.c */
  1955. extern const struct inode_operations ext4_file_inode_operations;
  1956. extern const struct file_operations ext4_file_operations;
  1957. extern loff_t ext4_llseek(struct file *file, loff_t offset, int origin);
  1958. /* namei.c */
  1959. extern const struct inode_operations ext4_dir_inode_operations;
  1960. extern const struct inode_operations ext4_special_inode_operations;
  1961. extern struct dentry *ext4_get_parent(struct dentry *child);
  1962. /* symlink.c */
  1963. extern const struct inode_operations ext4_symlink_inode_operations;
  1964. extern const struct inode_operations ext4_fast_symlink_inode_operations;
  1965. /* block_validity */
  1966. extern void ext4_release_system_zone(struct super_block *sb);
  1967. extern int ext4_setup_system_zone(struct super_block *sb);
  1968. extern int __init ext4_init_system_zone(void);
  1969. extern void ext4_exit_system_zone(void);
  1970. extern int ext4_data_block_valid(struct ext4_sb_info *sbi,
  1971. ext4_fsblk_t start_blk,
  1972. unsigned int count);
  1973. extern int ext4_check_blockref(const char *, unsigned int,
  1974. struct inode *, __le32 *, unsigned int);
  1975. /* extents.c */
  1976. extern int ext4_ext_tree_init(handle_t *handle, struct inode *);
  1977. extern int ext4_ext_writepage_trans_blocks(struct inode *, int);
  1978. extern int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks,
  1979. int chunk);
  1980. extern int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
  1981. struct ext4_map_blocks *map, int flags);
  1982. extern void ext4_ext_truncate(struct inode *);
  1983. extern int ext4_ext_punch_hole(struct file *file, loff_t offset,
  1984. loff_t length);
  1985. extern void ext4_ext_init(struct super_block *);
  1986. extern void ext4_ext_release(struct super_block *);
  1987. extern long ext4_fallocate(struct file *file, int mode, loff_t offset,
  1988. loff_t len);
  1989. extern int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
  1990. ssize_t len);
  1991. extern int ext4_map_blocks(handle_t *handle, struct inode *inode,
  1992. struct ext4_map_blocks *map, int flags);
  1993. extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
  1994. __u64 start, __u64 len);
  1995. /* move_extent.c */
  1996. extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
  1997. __u64 start_orig, __u64 start_donor,
  1998. __u64 len, __u64 *moved_len);
  1999. /* page-io.c */
  2000. extern int __init ext4_init_pageio(void);
  2001. extern void ext4_exit_pageio(void);
  2002. extern void ext4_ioend_wait(struct inode *);
  2003. extern void ext4_free_io_end(ext4_io_end_t *io);
  2004. extern ext4_io_end_t *ext4_init_io_end(struct inode *inode, gfp_t flags);
  2005. extern int ext4_end_io_nolock(ext4_io_end_t *io);
  2006. extern void ext4_io_submit(struct ext4_io_submit *io);
  2007. extern int ext4_bio_write_page(struct ext4_io_submit *io,
  2008. struct page *page,
  2009. int len,
  2010. struct writeback_control *wbc);
  2011. /* mmp.c */
  2012. extern int ext4_multi_mount_protect(struct super_block *, ext4_fsblk_t);
  2013. /* BH_Uninit flag: blocks are allocated but uninitialized on disk */
  2014. enum ext4_state_bits {
  2015. BH_Uninit /* blocks are allocated but uninitialized on disk */
  2016. = BH_JBDPrivateStart,
  2017. };
  2018. BUFFER_FNS(Uninit, uninit)
  2019. TAS_BUFFER_FNS(Uninit, uninit)
  2020. /*
  2021. * Add new method to test wether block and inode bitmaps are properly
  2022. * initialized. With uninit_bg reading the block from disk is not enough
  2023. * to mark the bitmap uptodate. We need to also zero-out the bitmap
  2024. */
  2025. #define BH_BITMAP_UPTODATE BH_JBDPrivateStart
  2026. static inline int bitmap_uptodate(struct buffer_head *bh)
  2027. {
  2028. return (buffer_uptodate(bh) &&
  2029. test_bit(BH_BITMAP_UPTODATE, &(bh)->b_state));
  2030. }
  2031. static inline void set_bitmap_uptodate(struct buffer_head *bh)
  2032. {
  2033. set_bit(BH_BITMAP_UPTODATE, &(bh)->b_state);
  2034. }
  2035. #define in_range(b, first, len) ((b) >= (first) && (b) <= (first) + (len) - 1)
  2036. /* For ioend & aio unwritten conversion wait queues */
  2037. #define EXT4_WQ_HASH_SZ 37
  2038. #define ext4_ioend_wq(v) (&ext4__ioend_wq[((unsigned long)(v)) %\
  2039. EXT4_WQ_HASH_SZ])
  2040. #define ext4_aio_mutex(v) (&ext4__aio_mutex[((unsigned long)(v)) %\
  2041. EXT4_WQ_HASH_SZ])
  2042. extern wait_queue_head_t ext4__ioend_wq[EXT4_WQ_HASH_SZ];
  2043. extern struct mutex ext4__aio_mutex[EXT4_WQ_HASH_SZ];
  2044. #define EXT4_RESIZING 0
  2045. extern int ext4_resize_begin(struct super_block *sb);
  2046. extern void ext4_resize_end(struct super_block *sb);
  2047. #endif /* __KERNEL__ */
  2048. #endif /* _EXT4_H */