ext4.h 61 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762
  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. /*
  31. * The fourth extended filesystem constants/structures
  32. */
  33. /*
  34. * Define EXT4FS_DEBUG to produce debug messages
  35. */
  36. #undef EXT4FS_DEBUG
  37. /*
  38. * Debug code
  39. */
  40. #ifdef EXT4FS_DEBUG
  41. #define ext4_debug(f, a...) \
  42. do { \
  43. printk(KERN_DEBUG "EXT4-fs DEBUG (%s, %d): %s:", \
  44. __FILE__, __LINE__, __func__); \
  45. printk(KERN_DEBUG f, ## a); \
  46. } while (0)
  47. #else
  48. #define ext4_debug(f, a...) do {} while (0)
  49. #endif
  50. /* data type for block offset of block group */
  51. typedef int ext4_grpblk_t;
  52. /* data type for filesystem-wide blocks number */
  53. typedef unsigned long long ext4_fsblk_t;
  54. /* data type for file logical block number */
  55. typedef __u32 ext4_lblk_t;
  56. /* data type for block group number */
  57. typedef unsigned int ext4_group_t;
  58. /* prefer goal again. length */
  59. #define EXT4_MB_HINT_MERGE 0x0001
  60. /* blocks already reserved */
  61. #define EXT4_MB_HINT_RESERVED 0x0002
  62. /* metadata is being allocated */
  63. #define EXT4_MB_HINT_METADATA 0x0004
  64. /* first blocks in the file */
  65. #define EXT4_MB_HINT_FIRST 0x0008
  66. /* search for the best chunk */
  67. #define EXT4_MB_HINT_BEST 0x0010
  68. /* data is being allocated */
  69. #define EXT4_MB_HINT_DATA 0x0020
  70. /* don't preallocate (for tails) */
  71. #define EXT4_MB_HINT_NOPREALLOC 0x0040
  72. /* allocate for locality group */
  73. #define EXT4_MB_HINT_GROUP_ALLOC 0x0080
  74. /* allocate goal blocks or none */
  75. #define EXT4_MB_HINT_GOAL_ONLY 0x0100
  76. /* goal is meaningful */
  77. #define EXT4_MB_HINT_TRY_GOAL 0x0200
  78. /* blocks already pre-reserved by delayed allocation */
  79. #define EXT4_MB_DELALLOC_RESERVED 0x0400
  80. /* We are doing stream allocation */
  81. #define EXT4_MB_STREAM_ALLOC 0x0800
  82. struct ext4_allocation_request {
  83. /* target inode for block we're allocating */
  84. struct inode *inode;
  85. /* how many blocks we want to allocate */
  86. unsigned int len;
  87. /* logical block in target inode */
  88. ext4_lblk_t logical;
  89. /* the closest logical allocated block to the left */
  90. ext4_lblk_t lleft;
  91. /* the closest logical allocated block to the right */
  92. ext4_lblk_t lright;
  93. /* phys. target (a hint) */
  94. ext4_fsblk_t goal;
  95. /* phys. block for the closest logical allocated block to the left */
  96. ext4_fsblk_t pleft;
  97. /* phys. block for the closest logical allocated block to the right */
  98. ext4_fsblk_t pright;
  99. /* flags. see above EXT4_MB_HINT_* */
  100. unsigned int flags;
  101. };
  102. /*
  103. * For delayed allocation tracking
  104. */
  105. struct mpage_da_data {
  106. struct inode *inode;
  107. sector_t b_blocknr; /* start block number of extent */
  108. size_t b_size; /* size of extent */
  109. unsigned long b_state; /* state of the extent */
  110. unsigned long first_page, next_page; /* extent of pages */
  111. struct writeback_control *wbc;
  112. int io_done;
  113. int pages_written;
  114. int retval;
  115. };
  116. #define DIO_AIO_UNWRITTEN 0x1
  117. typedef struct ext4_io_end {
  118. struct list_head list; /* per-file finished AIO list */
  119. struct inode *inode; /* file being written to */
  120. unsigned int flag; /* unwritten or not */
  121. int error; /* I/O error code */
  122. ext4_lblk_t offset; /* offset in the file */
  123. size_t size; /* size of the extent */
  124. struct work_struct work; /* data work queue */
  125. } ext4_io_end_t;
  126. /*
  127. * Special inodes numbers
  128. */
  129. #define EXT4_BAD_INO 1 /* Bad blocks inode */
  130. #define EXT4_ROOT_INO 2 /* Root inode */
  131. #define EXT4_BOOT_LOADER_INO 5 /* Boot loader inode */
  132. #define EXT4_UNDEL_DIR_INO 6 /* Undelete directory inode */
  133. #define EXT4_RESIZE_INO 7 /* Reserved group descriptors inode */
  134. #define EXT4_JOURNAL_INO 8 /* Journal inode */
  135. /* First non-reserved inode for old ext4 filesystems */
  136. #define EXT4_GOOD_OLD_FIRST_INO 11
  137. /*
  138. * Maximal count of links to a file
  139. */
  140. #define EXT4_LINK_MAX 65000
  141. /*
  142. * Macro-instructions used to manage several block sizes
  143. */
  144. #define EXT4_MIN_BLOCK_SIZE 1024
  145. #define EXT4_MAX_BLOCK_SIZE 65536
  146. #define EXT4_MIN_BLOCK_LOG_SIZE 10
  147. #ifdef __KERNEL__
  148. # define EXT4_BLOCK_SIZE(s) ((s)->s_blocksize)
  149. #else
  150. # define EXT4_BLOCK_SIZE(s) (EXT4_MIN_BLOCK_SIZE << (s)->s_log_block_size)
  151. #endif
  152. #define EXT4_ADDR_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / sizeof(__u32))
  153. #ifdef __KERNEL__
  154. # define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
  155. #else
  156. # define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_log_block_size + 10)
  157. #endif
  158. #ifdef __KERNEL__
  159. #define EXT4_ADDR_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_addr_per_block_bits)
  160. #define EXT4_INODE_SIZE(s) (EXT4_SB(s)->s_inode_size)
  161. #define EXT4_FIRST_INO(s) (EXT4_SB(s)->s_first_ino)
  162. #else
  163. #define EXT4_INODE_SIZE(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
  164. EXT4_GOOD_OLD_INODE_SIZE : \
  165. (s)->s_inode_size)
  166. #define EXT4_FIRST_INO(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
  167. EXT4_GOOD_OLD_FIRST_INO : \
  168. (s)->s_first_ino)
  169. #endif
  170. #define EXT4_BLOCK_ALIGN(size, blkbits) ALIGN((size), (1 << (blkbits)))
  171. /*
  172. * Structure of a blocks group descriptor
  173. */
  174. struct ext4_group_desc
  175. {
  176. __le32 bg_block_bitmap_lo; /* Blocks bitmap block */
  177. __le32 bg_inode_bitmap_lo; /* Inodes bitmap block */
  178. __le32 bg_inode_table_lo; /* Inodes table block */
  179. __le16 bg_free_blocks_count_lo;/* Free blocks count */
  180. __le16 bg_free_inodes_count_lo;/* Free inodes count */
  181. __le16 bg_used_dirs_count_lo; /* Directories count */
  182. __le16 bg_flags; /* EXT4_BG_flags (INODE_UNINIT, etc) */
  183. __u32 bg_reserved[2]; /* Likely block/inode bitmap checksum */
  184. __le16 bg_itable_unused_lo; /* Unused inodes count */
  185. __le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
  186. __le32 bg_block_bitmap_hi; /* Blocks bitmap block MSB */
  187. __le32 bg_inode_bitmap_hi; /* Inodes bitmap block MSB */
  188. __le32 bg_inode_table_hi; /* Inodes table block MSB */
  189. __le16 bg_free_blocks_count_hi;/* Free blocks count MSB */
  190. __le16 bg_free_inodes_count_hi;/* Free inodes count MSB */
  191. __le16 bg_used_dirs_count_hi; /* Directories count MSB */
  192. __le16 bg_itable_unused_hi; /* Unused inodes count MSB */
  193. __u32 bg_reserved2[3];
  194. };
  195. /*
  196. * Structure of a flex block group info
  197. */
  198. struct flex_groups {
  199. atomic_t free_inodes;
  200. atomic_t free_blocks;
  201. atomic_t used_dirs;
  202. };
  203. #define EXT4_BG_INODE_UNINIT 0x0001 /* Inode table/bitmap not in use */
  204. #define EXT4_BG_BLOCK_UNINIT 0x0002 /* Block bitmap not in use */
  205. #define EXT4_BG_INODE_ZEROED 0x0004 /* On-disk itable initialized to zero */
  206. /*
  207. * Macro-instructions used to manage group descriptors
  208. */
  209. #define EXT4_MIN_DESC_SIZE 32
  210. #define EXT4_MIN_DESC_SIZE_64BIT 64
  211. #define EXT4_MAX_DESC_SIZE EXT4_MIN_BLOCK_SIZE
  212. #define EXT4_DESC_SIZE(s) (EXT4_SB(s)->s_desc_size)
  213. #ifdef __KERNEL__
  214. # define EXT4_BLOCKS_PER_GROUP(s) (EXT4_SB(s)->s_blocks_per_group)
  215. # define EXT4_DESC_PER_BLOCK(s) (EXT4_SB(s)->s_desc_per_block)
  216. # define EXT4_INODES_PER_GROUP(s) (EXT4_SB(s)->s_inodes_per_group)
  217. # define EXT4_DESC_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_desc_per_block_bits)
  218. #else
  219. # define EXT4_BLOCKS_PER_GROUP(s) ((s)->s_blocks_per_group)
  220. # define EXT4_DESC_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / EXT4_DESC_SIZE(s))
  221. # define EXT4_INODES_PER_GROUP(s) ((s)->s_inodes_per_group)
  222. #endif
  223. /*
  224. * Constants relative to the data blocks
  225. */
  226. #define EXT4_NDIR_BLOCKS 12
  227. #define EXT4_IND_BLOCK EXT4_NDIR_BLOCKS
  228. #define EXT4_DIND_BLOCK (EXT4_IND_BLOCK + 1)
  229. #define EXT4_TIND_BLOCK (EXT4_DIND_BLOCK + 1)
  230. #define EXT4_N_BLOCKS (EXT4_TIND_BLOCK + 1)
  231. /*
  232. * Inode flags
  233. */
  234. #define EXT4_SECRM_FL 0x00000001 /* Secure deletion */
  235. #define EXT4_UNRM_FL 0x00000002 /* Undelete */
  236. #define EXT4_COMPR_FL 0x00000004 /* Compress file */
  237. #define EXT4_SYNC_FL 0x00000008 /* Synchronous updates */
  238. #define EXT4_IMMUTABLE_FL 0x00000010 /* Immutable file */
  239. #define EXT4_APPEND_FL 0x00000020 /* writes to file may only append */
  240. #define EXT4_NODUMP_FL 0x00000040 /* do not dump file */
  241. #define EXT4_NOATIME_FL 0x00000080 /* do not update atime */
  242. /* Reserved for compression usage... */
  243. #define EXT4_DIRTY_FL 0x00000100
  244. #define EXT4_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
  245. #define EXT4_NOCOMPR_FL 0x00000400 /* Don't compress */
  246. #define EXT4_ECOMPR_FL 0x00000800 /* Compression error */
  247. /* End compression flags --- maybe not all used */
  248. #define EXT4_INDEX_FL 0x00001000 /* hash-indexed directory */
  249. #define EXT4_IMAGIC_FL 0x00002000 /* AFS directory */
  250. #define EXT4_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */
  251. #define EXT4_NOTAIL_FL 0x00008000 /* file tail should not be merged */
  252. #define EXT4_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
  253. #define EXT4_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
  254. #define EXT4_HUGE_FILE_FL 0x00040000 /* Set to each huge file */
  255. #define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
  256. #define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */
  257. #define EXT4_FL_USER_VISIBLE 0x000BDFFF /* User visible flags */
  258. #define EXT4_FL_USER_MODIFIABLE 0x000B80FF /* User modifiable flags */
  259. /* Flags that should be inherited by new inodes from their parent. */
  260. #define EXT4_FL_INHERITED (EXT4_SECRM_FL | EXT4_UNRM_FL | EXT4_COMPR_FL |\
  261. EXT4_SYNC_FL | EXT4_IMMUTABLE_FL | EXT4_APPEND_FL |\
  262. EXT4_NODUMP_FL | EXT4_NOATIME_FL |\
  263. EXT4_NOCOMPR_FL | EXT4_JOURNAL_DATA_FL |\
  264. EXT4_NOTAIL_FL | EXT4_DIRSYNC_FL)
  265. /* Flags that are appropriate for regular files (all but dir-specific ones). */
  266. #define EXT4_REG_FLMASK (~(EXT4_DIRSYNC_FL | EXT4_TOPDIR_FL))
  267. /* Flags that are appropriate for non-directories/regular files. */
  268. #define EXT4_OTHER_FLMASK (EXT4_NODUMP_FL | EXT4_NOATIME_FL)
  269. /* Mask out flags that are inappropriate for the given type of inode. */
  270. static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
  271. {
  272. if (S_ISDIR(mode))
  273. return flags;
  274. else if (S_ISREG(mode))
  275. return flags & EXT4_REG_FLMASK;
  276. else
  277. return flags & EXT4_OTHER_FLMASK;
  278. }
  279. /*
  280. * Inode dynamic state flags
  281. */
  282. #define EXT4_STATE_JDATA 0x00000001 /* journaled data exists */
  283. #define EXT4_STATE_NEW 0x00000002 /* inode is newly created */
  284. #define EXT4_STATE_XATTR 0x00000004 /* has in-inode xattrs */
  285. #define EXT4_STATE_NO_EXPAND 0x00000008 /* No space for expansion */
  286. #define EXT4_STATE_DA_ALLOC_CLOSE 0x00000010 /* Alloc DA blks on close */
  287. #define EXT4_STATE_EXT_MIGRATE 0x00000020 /* Inode is migrating */
  288. /* Used to pass group descriptor data when online resize is done */
  289. struct ext4_new_group_input {
  290. __u32 group; /* Group number for this data */
  291. __u64 block_bitmap; /* Absolute block number of block bitmap */
  292. __u64 inode_bitmap; /* Absolute block number of inode bitmap */
  293. __u64 inode_table; /* Absolute block number of inode table start */
  294. __u32 blocks_count; /* Total number of blocks in this group */
  295. __u16 reserved_blocks; /* Number of reserved blocks in this group */
  296. __u16 unused;
  297. };
  298. /* The struct ext4_new_group_input in kernel space, with free_blocks_count */
  299. struct ext4_new_group_data {
  300. __u32 group;
  301. __u64 block_bitmap;
  302. __u64 inode_bitmap;
  303. __u64 inode_table;
  304. __u32 blocks_count;
  305. __u16 reserved_blocks;
  306. __u16 unused;
  307. __u32 free_blocks_count;
  308. };
  309. /*
  310. * Flags used by ext4_get_blocks()
  311. */
  312. /* Allocate any needed blocks and/or convert an unitialized
  313. extent to be an initialized ext4 */
  314. #define EXT4_GET_BLOCKS_CREATE 0x0001
  315. /* Request the creation of an unitialized extent */
  316. #define EXT4_GET_BLOCKS_UNINIT_EXT 0x0002
  317. #define EXT4_GET_BLOCKS_CREATE_UNINIT_EXT (EXT4_GET_BLOCKS_UNINIT_EXT|\
  318. EXT4_GET_BLOCKS_CREATE)
  319. /* Caller is from the delayed allocation writeout path,
  320. so set the magic i_delalloc_reserve_flag after taking the
  321. inode allocation semaphore for */
  322. #define EXT4_GET_BLOCKS_DELALLOC_RESERVE 0x0004
  323. /* Call ext4_da_update_reserve_space() after successfully
  324. allocating the blocks */
  325. #define EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE 0x0008
  326. /* caller is from the direct IO path, request to creation of an
  327. unitialized extents if not allocated, split the uninitialized
  328. extent if blocks has been preallocated already*/
  329. #define EXT4_GET_BLOCKS_DIO 0x0010
  330. #define EXT4_GET_BLOCKS_CONVERT 0x0020
  331. #define EXT4_GET_BLOCKS_DIO_CREATE_EXT (EXT4_GET_BLOCKS_DIO|\
  332. EXT4_GET_BLOCKS_CREATE_UNINIT_EXT)
  333. /* Convert extent to initialized after direct IO complete */
  334. #define EXT4_GET_BLOCKS_DIO_CONVERT_EXT (EXT4_GET_BLOCKS_CONVERT|\
  335. EXT4_GET_BLOCKS_DIO_CREATE_EXT)
  336. /*
  337. * ioctl commands
  338. */
  339. #define EXT4_IOC_GETFLAGS FS_IOC_GETFLAGS
  340. #define EXT4_IOC_SETFLAGS FS_IOC_SETFLAGS
  341. #define EXT4_IOC_GETVERSION _IOR('f', 3, long)
  342. #define EXT4_IOC_SETVERSION _IOW('f', 4, long)
  343. #define EXT4_IOC_GETVERSION_OLD FS_IOC_GETVERSION
  344. #define EXT4_IOC_SETVERSION_OLD FS_IOC_SETVERSION
  345. #ifdef CONFIG_JBD2_DEBUG
  346. #define EXT4_IOC_WAIT_FOR_READONLY _IOR('f', 99, long)
  347. #endif
  348. #define EXT4_IOC_GETRSVSZ _IOR('f', 5, long)
  349. #define EXT4_IOC_SETRSVSZ _IOW('f', 6, long)
  350. #define EXT4_IOC_GROUP_EXTEND _IOW('f', 7, unsigned long)
  351. #define EXT4_IOC_GROUP_ADD _IOW('f', 8, struct ext4_new_group_input)
  352. #define EXT4_IOC_MIGRATE _IO('f', 9)
  353. /* note ioctl 10 reserved for an early version of the FIEMAP ioctl */
  354. /* note ioctl 11 reserved for filesystem-independent FIEMAP ioctl */
  355. #define EXT4_IOC_ALLOC_DA_BLKS _IO('f', 12)
  356. #define EXT4_IOC_MOVE_EXT _IOWR('f', 15, struct move_extent)
  357. /*
  358. * ioctl commands in 32 bit emulation
  359. */
  360. #define EXT4_IOC32_GETFLAGS FS_IOC32_GETFLAGS
  361. #define EXT4_IOC32_SETFLAGS FS_IOC32_SETFLAGS
  362. #define EXT4_IOC32_GETVERSION _IOR('f', 3, int)
  363. #define EXT4_IOC32_SETVERSION _IOW('f', 4, int)
  364. #define EXT4_IOC32_GETRSVSZ _IOR('f', 5, int)
  365. #define EXT4_IOC32_SETRSVSZ _IOW('f', 6, int)
  366. #define EXT4_IOC32_GROUP_EXTEND _IOW('f', 7, unsigned int)
  367. #ifdef CONFIG_JBD2_DEBUG
  368. #define EXT4_IOC32_WAIT_FOR_READONLY _IOR('f', 99, int)
  369. #endif
  370. #define EXT4_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
  371. #define EXT4_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
  372. /*
  373. * Mount options
  374. */
  375. struct ext4_mount_options {
  376. unsigned long s_mount_opt;
  377. uid_t s_resuid;
  378. gid_t s_resgid;
  379. unsigned long s_commit_interval;
  380. u32 s_min_batch_time, s_max_batch_time;
  381. #ifdef CONFIG_QUOTA
  382. int s_jquota_fmt;
  383. char *s_qf_names[MAXQUOTAS];
  384. #endif
  385. };
  386. /* Max physical block we can addres w/o extents */
  387. #define EXT4_MAX_BLOCK_FILE_PHYS 0xFFFFFFFF
  388. /*
  389. * Structure of an inode on the disk
  390. */
  391. struct ext4_inode {
  392. __le16 i_mode; /* File mode */
  393. __le16 i_uid; /* Low 16 bits of Owner Uid */
  394. __le32 i_size_lo; /* Size in bytes */
  395. __le32 i_atime; /* Access time */
  396. __le32 i_ctime; /* Inode Change time */
  397. __le32 i_mtime; /* Modification time */
  398. __le32 i_dtime; /* Deletion Time */
  399. __le16 i_gid; /* Low 16 bits of Group Id */
  400. __le16 i_links_count; /* Links count */
  401. __le32 i_blocks_lo; /* Blocks count */
  402. __le32 i_flags; /* File flags */
  403. union {
  404. struct {
  405. __le32 l_i_version;
  406. } linux1;
  407. struct {
  408. __u32 h_i_translator;
  409. } hurd1;
  410. struct {
  411. __u32 m_i_reserved1;
  412. } masix1;
  413. } osd1; /* OS dependent 1 */
  414. __le32 i_block[EXT4_N_BLOCKS];/* Pointers to blocks */
  415. __le32 i_generation; /* File version (for NFS) */
  416. __le32 i_file_acl_lo; /* File ACL */
  417. __le32 i_size_high;
  418. __le32 i_obso_faddr; /* Obsoleted fragment address */
  419. union {
  420. struct {
  421. __le16 l_i_blocks_high; /* were l_i_reserved1 */
  422. __le16 l_i_file_acl_high;
  423. __le16 l_i_uid_high; /* these 2 fields */
  424. __le16 l_i_gid_high; /* were reserved2[0] */
  425. __u32 l_i_reserved2;
  426. } linux2;
  427. struct {
  428. __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
  429. __u16 h_i_mode_high;
  430. __u16 h_i_uid_high;
  431. __u16 h_i_gid_high;
  432. __u32 h_i_author;
  433. } hurd2;
  434. struct {
  435. __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
  436. __le16 m_i_file_acl_high;
  437. __u32 m_i_reserved2[2];
  438. } masix2;
  439. } osd2; /* OS dependent 2 */
  440. __le16 i_extra_isize;
  441. __le16 i_pad1;
  442. __le32 i_ctime_extra; /* extra Change time (nsec << 2 | epoch) */
  443. __le32 i_mtime_extra; /* extra Modification time(nsec << 2 | epoch) */
  444. __le32 i_atime_extra; /* extra Access time (nsec << 2 | epoch) */
  445. __le32 i_crtime; /* File Creation time */
  446. __le32 i_crtime_extra; /* extra FileCreationtime (nsec << 2 | epoch) */
  447. __le32 i_version_hi; /* high 32 bits for 64-bit version */
  448. };
  449. struct move_extent {
  450. __u32 reserved; /* should be zero */
  451. __u32 donor_fd; /* donor file descriptor */
  452. __u64 orig_start; /* logical start offset in block for orig */
  453. __u64 donor_start; /* logical start offset in block for donor */
  454. __u64 len; /* block length to be moved */
  455. __u64 moved_len; /* moved block length */
  456. };
  457. #define EXT4_EPOCH_BITS 2
  458. #define EXT4_EPOCH_MASK ((1 << EXT4_EPOCH_BITS) - 1)
  459. #define EXT4_NSEC_MASK (~0UL << EXT4_EPOCH_BITS)
  460. /*
  461. * Extended fields will fit into an inode if the filesystem was formatted
  462. * with large inodes (-I 256 or larger) and there are not currently any EAs
  463. * consuming all of the available space. For new inodes we always reserve
  464. * enough space for the kernel's known extended fields, but for inodes
  465. * created with an old kernel this might not have been the case. None of
  466. * the extended inode fields is critical for correct filesystem operation.
  467. * This macro checks if a certain field fits in the inode. Note that
  468. * inode-size = GOOD_OLD_INODE_SIZE + i_extra_isize
  469. */
  470. #define EXT4_FITS_IN_INODE(ext4_inode, einode, field) \
  471. ((offsetof(typeof(*ext4_inode), field) + \
  472. sizeof((ext4_inode)->field)) \
  473. <= (EXT4_GOOD_OLD_INODE_SIZE + \
  474. (einode)->i_extra_isize)) \
  475. static inline __le32 ext4_encode_extra_time(struct timespec *time)
  476. {
  477. return cpu_to_le32((sizeof(time->tv_sec) > 4 ?
  478. time->tv_sec >> 32 : 0) |
  479. ((time->tv_nsec << 2) & EXT4_NSEC_MASK));
  480. }
  481. static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra)
  482. {
  483. if (sizeof(time->tv_sec) > 4)
  484. time->tv_sec |= (__u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK)
  485. << 32;
  486. time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> 2;
  487. }
  488. #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \
  489. do { \
  490. (raw_inode)->xtime = cpu_to_le32((inode)->xtime.tv_sec); \
  491. if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) \
  492. (raw_inode)->xtime ## _extra = \
  493. ext4_encode_extra_time(&(inode)->xtime); \
  494. } while (0)
  495. #define EXT4_EINODE_SET_XTIME(xtime, einode, raw_inode) \
  496. do { \
  497. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
  498. (raw_inode)->xtime = cpu_to_le32((einode)->xtime.tv_sec); \
  499. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
  500. (raw_inode)->xtime ## _extra = \
  501. ext4_encode_extra_time(&(einode)->xtime); \
  502. } while (0)
  503. #define EXT4_INODE_GET_XTIME(xtime, inode, raw_inode) \
  504. do { \
  505. (inode)->xtime.tv_sec = (signed)le32_to_cpu((raw_inode)->xtime); \
  506. if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) \
  507. ext4_decode_extra_time(&(inode)->xtime, \
  508. raw_inode->xtime ## _extra); \
  509. } while (0)
  510. #define EXT4_EINODE_GET_XTIME(xtime, einode, raw_inode) \
  511. do { \
  512. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
  513. (einode)->xtime.tv_sec = \
  514. (signed)le32_to_cpu((raw_inode)->xtime); \
  515. if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
  516. ext4_decode_extra_time(&(einode)->xtime, \
  517. raw_inode->xtime ## _extra); \
  518. } while (0)
  519. #define i_disk_version osd1.linux1.l_i_version
  520. #if defined(__KERNEL__) || defined(__linux__)
  521. #define i_reserved1 osd1.linux1.l_i_reserved1
  522. #define i_file_acl_high osd2.linux2.l_i_file_acl_high
  523. #define i_blocks_high osd2.linux2.l_i_blocks_high
  524. #define i_uid_low i_uid
  525. #define i_gid_low i_gid
  526. #define i_uid_high osd2.linux2.l_i_uid_high
  527. #define i_gid_high osd2.linux2.l_i_gid_high
  528. #define i_reserved2 osd2.linux2.l_i_reserved2
  529. #elif defined(__GNU__)
  530. #define i_translator osd1.hurd1.h_i_translator
  531. #define i_uid_high osd2.hurd2.h_i_uid_high
  532. #define i_gid_high osd2.hurd2.h_i_gid_high
  533. #define i_author osd2.hurd2.h_i_author
  534. #elif defined(__masix__)
  535. #define i_reserved1 osd1.masix1.m_i_reserved1
  536. #define i_file_acl_high osd2.masix2.m_i_file_acl_high
  537. #define i_reserved2 osd2.masix2.m_i_reserved2
  538. #endif /* defined(__KERNEL__) || defined(__linux__) */
  539. /*
  540. * storage for cached extent
  541. */
  542. struct ext4_ext_cache {
  543. ext4_fsblk_t ec_start;
  544. ext4_lblk_t ec_block;
  545. __u32 ec_len; /* must be 32bit to return holes */
  546. __u32 ec_type;
  547. };
  548. /*
  549. * fourth extended file system inode data in memory
  550. */
  551. struct ext4_inode_info {
  552. __le32 i_data[15]; /* unconverted */
  553. __u32 i_flags;
  554. ext4_fsblk_t i_file_acl;
  555. __u32 i_dtime;
  556. /*
  557. * i_block_group is the number of the block group which contains
  558. * this file's inode. Constant across the lifetime of the inode,
  559. * it is ued for making block allocation decisions - we try to
  560. * place a file's data blocks near its inode block, and new inodes
  561. * near to their parent directory's inode.
  562. */
  563. ext4_group_t i_block_group;
  564. __u32 i_state; /* Dynamic state flags for ext4 */
  565. ext4_lblk_t i_dir_start_lookup;
  566. #ifdef CONFIG_EXT4_FS_XATTR
  567. /*
  568. * Extended attributes can be read independently of the main file
  569. * data. Taking i_mutex even when reading would cause contention
  570. * between readers of EAs and writers of regular file data, so
  571. * instead we synchronize on xattr_sem when reading or changing
  572. * EAs.
  573. */
  574. struct rw_semaphore xattr_sem;
  575. #endif
  576. struct list_head i_orphan; /* unlinked but open inodes */
  577. /*
  578. * i_disksize keeps track of what the inode size is ON DISK, not
  579. * in memory. During truncate, i_size is set to the new size by
  580. * the VFS prior to calling ext4_truncate(), but the filesystem won't
  581. * set i_disksize to 0 until the truncate is actually under way.
  582. *
  583. * The intent is that i_disksize always represents the blocks which
  584. * are used by this file. This allows recovery to restart truncate
  585. * on orphans if we crash during truncate. We actually write i_disksize
  586. * into the on-disk inode when writing inodes out, instead of i_size.
  587. *
  588. * The only time when i_disksize and i_size may be different is when
  589. * a truncate is in progress. The only things which change i_disksize
  590. * are ext4_get_block (growth) and ext4_truncate (shrinkth).
  591. */
  592. loff_t i_disksize;
  593. /*
  594. * i_data_sem is for serialising ext4_truncate() against
  595. * ext4_getblock(). In the 2.4 ext2 design, great chunks of inode's
  596. * data tree are chopped off during truncate. We can't do that in
  597. * ext4 because whenever we perform intermediate commits during
  598. * truncate, the inode and all the metadata blocks *must* be in a
  599. * consistent state which allows truncation of the orphans to restart
  600. * during recovery. Hence we must fix the get_block-vs-truncate race
  601. * by other means, so we have i_data_sem.
  602. */
  603. struct rw_semaphore i_data_sem;
  604. struct inode vfs_inode;
  605. struct jbd2_inode jinode;
  606. struct ext4_ext_cache i_cached_extent;
  607. /*
  608. * File creation time. Its function is same as that of
  609. * struct timespec i_{a,c,m}time in the generic inode.
  610. */
  611. struct timespec i_crtime;
  612. /* mballoc */
  613. struct list_head i_prealloc_list;
  614. spinlock_t i_prealloc_lock;
  615. /* ialloc */
  616. ext4_group_t i_last_alloc_group;
  617. /* allocation reservation info for delalloc */
  618. unsigned int i_reserved_data_blocks;
  619. unsigned int i_reserved_meta_blocks;
  620. unsigned int i_allocated_meta_blocks;
  621. unsigned short i_delalloc_reserved_flag;
  622. /* on-disk additional length */
  623. __u16 i_extra_isize;
  624. spinlock_t i_block_reservation_lock;
  625. /* completed async DIOs that might need unwritten extents handling */
  626. struct list_head i_aio_dio_complete_list;
  627. /* current io_end structure for async DIO write*/
  628. ext4_io_end_t *cur_aio_dio;
  629. };
  630. /*
  631. * File system states
  632. */
  633. #define EXT4_VALID_FS 0x0001 /* Unmounted cleanly */
  634. #define EXT4_ERROR_FS 0x0002 /* Errors detected */
  635. #define EXT4_ORPHAN_FS 0x0004 /* Orphans being recovered */
  636. /*
  637. * Misc. filesystem flags
  638. */
  639. #define EXT2_FLAGS_SIGNED_HASH 0x0001 /* Signed dirhash in use */
  640. #define EXT2_FLAGS_UNSIGNED_HASH 0x0002 /* Unsigned dirhash in use */
  641. #define EXT2_FLAGS_TEST_FILESYS 0x0004 /* to test development code */
  642. /*
  643. * Mount flags
  644. */
  645. #define EXT4_MOUNT_OLDALLOC 0x00002 /* Don't use the new Orlov allocator */
  646. #define EXT4_MOUNT_GRPID 0x00004 /* Create files with directory's group */
  647. #define EXT4_MOUNT_DEBUG 0x00008 /* Some debugging messages */
  648. #define EXT4_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
  649. #define EXT4_MOUNT_ERRORS_RO 0x00020 /* Remount fs ro on errors */
  650. #define EXT4_MOUNT_ERRORS_PANIC 0x00040 /* Panic on errors */
  651. #define EXT4_MOUNT_MINIX_DF 0x00080 /* Mimics the Minix statfs */
  652. #define EXT4_MOUNT_NOLOAD 0x00100 /* Don't use existing journal*/
  653. #define EXT4_MOUNT_DATA_FLAGS 0x00C00 /* Mode for data writes: */
  654. #define EXT4_MOUNT_JOURNAL_DATA 0x00400 /* Write data to journal */
  655. #define EXT4_MOUNT_ORDERED_DATA 0x00800 /* Flush data before commit */
  656. #define EXT4_MOUNT_WRITEBACK_DATA 0x00C00 /* No data ordering */
  657. #define EXT4_MOUNT_UPDATE_JOURNAL 0x01000 /* Update the journal format */
  658. #define EXT4_MOUNT_NO_UID32 0x02000 /* Disable 32-bit UIDs */
  659. #define EXT4_MOUNT_XATTR_USER 0x04000 /* Extended user attributes */
  660. #define EXT4_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */
  661. #define EXT4_MOUNT_NO_AUTO_DA_ALLOC 0x10000 /* No auto delalloc mapping */
  662. #define EXT4_MOUNT_BARRIER 0x20000 /* Use block barriers */
  663. #define EXT4_MOUNT_NOBH 0x40000 /* No bufferheads */
  664. #define EXT4_MOUNT_QUOTA 0x80000 /* Some quota option set */
  665. #define EXT4_MOUNT_USRQUOTA 0x100000 /* "old" user quota */
  666. #define EXT4_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */
  667. #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT 0x1000000 /* Journal Async Commit */
  668. #define EXT4_MOUNT_I_VERSION 0x2000000 /* i_version support */
  669. #define EXT4_MOUNT_DELALLOC 0x8000000 /* Delalloc support */
  670. #define EXT4_MOUNT_DATA_ERR_ABORT 0x10000000 /* Abort on file data write */
  671. #define EXT4_MOUNT_BLOCK_VALIDITY 0x20000000 /* Block validity checking */
  672. #define clear_opt(o, opt) o &= ~EXT4_MOUNT_##opt
  673. #define set_opt(o, opt) o |= EXT4_MOUNT_##opt
  674. #define test_opt(sb, opt) (EXT4_SB(sb)->s_mount_opt & \
  675. EXT4_MOUNT_##opt)
  676. #define ext4_set_bit ext2_set_bit
  677. #define ext4_set_bit_atomic ext2_set_bit_atomic
  678. #define ext4_clear_bit ext2_clear_bit
  679. #define ext4_clear_bit_atomic ext2_clear_bit_atomic
  680. #define ext4_test_bit ext2_test_bit
  681. #define ext4_find_first_zero_bit ext2_find_first_zero_bit
  682. #define ext4_find_next_zero_bit ext2_find_next_zero_bit
  683. #define ext4_find_next_bit ext2_find_next_bit
  684. /*
  685. * Maximal mount counts between two filesystem checks
  686. */
  687. #define EXT4_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
  688. #define EXT4_DFL_CHECKINTERVAL 0 /* Don't use interval check */
  689. /*
  690. * Behaviour when detecting errors
  691. */
  692. #define EXT4_ERRORS_CONTINUE 1 /* Continue execution */
  693. #define EXT4_ERRORS_RO 2 /* Remount fs read-only */
  694. #define EXT4_ERRORS_PANIC 3 /* Panic */
  695. #define EXT4_ERRORS_DEFAULT EXT4_ERRORS_CONTINUE
  696. /*
  697. * Structure of the super block
  698. */
  699. struct ext4_super_block {
  700. /*00*/ __le32 s_inodes_count; /* Inodes count */
  701. __le32 s_blocks_count_lo; /* Blocks count */
  702. __le32 s_r_blocks_count_lo; /* Reserved blocks count */
  703. __le32 s_free_blocks_count_lo; /* Free blocks count */
  704. /*10*/ __le32 s_free_inodes_count; /* Free inodes count */
  705. __le32 s_first_data_block; /* First Data Block */
  706. __le32 s_log_block_size; /* Block size */
  707. __le32 s_obso_log_frag_size; /* Obsoleted fragment size */
  708. /*20*/ __le32 s_blocks_per_group; /* # Blocks per group */
  709. __le32 s_obso_frags_per_group; /* Obsoleted fragments per group */
  710. __le32 s_inodes_per_group; /* # Inodes per group */
  711. __le32 s_mtime; /* Mount time */
  712. /*30*/ __le32 s_wtime; /* Write time */
  713. __le16 s_mnt_count; /* Mount count */
  714. __le16 s_max_mnt_count; /* Maximal mount count */
  715. __le16 s_magic; /* Magic signature */
  716. __le16 s_state; /* File system state */
  717. __le16 s_errors; /* Behaviour when detecting errors */
  718. __le16 s_minor_rev_level; /* minor revision level */
  719. /*40*/ __le32 s_lastcheck; /* time of last check */
  720. __le32 s_checkinterval; /* max. time between checks */
  721. __le32 s_creator_os; /* OS */
  722. __le32 s_rev_level; /* Revision level */
  723. /*50*/ __le16 s_def_resuid; /* Default uid for reserved blocks */
  724. __le16 s_def_resgid; /* Default gid for reserved blocks */
  725. /*
  726. * These fields are for EXT4_DYNAMIC_REV superblocks only.
  727. *
  728. * Note: the difference between the compatible feature set and
  729. * the incompatible feature set is that if there is a bit set
  730. * in the incompatible feature set that the kernel doesn't
  731. * know about, it should refuse to mount the filesystem.
  732. *
  733. * e2fsck's requirements are more strict; if it doesn't know
  734. * about a feature in either the compatible or incompatible
  735. * feature set, it must abort and not try to meddle with
  736. * things it doesn't understand...
  737. */
  738. __le32 s_first_ino; /* First non-reserved inode */
  739. __le16 s_inode_size; /* size of inode structure */
  740. __le16 s_block_group_nr; /* block group # of this superblock */
  741. __le32 s_feature_compat; /* compatible feature set */
  742. /*60*/ __le32 s_feature_incompat; /* incompatible feature set */
  743. __le32 s_feature_ro_compat; /* readonly-compatible feature set */
  744. /*68*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */
  745. /*78*/ char s_volume_name[16]; /* volume name */
  746. /*88*/ char s_last_mounted[64]; /* directory where last mounted */
  747. /*C8*/ __le32 s_algorithm_usage_bitmap; /* For compression */
  748. /*
  749. * Performance hints. Directory preallocation should only
  750. * happen if the EXT4_FEATURE_COMPAT_DIR_PREALLOC flag is on.
  751. */
  752. __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
  753. __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
  754. __le16 s_reserved_gdt_blocks; /* Per group desc for online growth */
  755. /*
  756. * Journaling support valid if EXT4_FEATURE_COMPAT_HAS_JOURNAL set.
  757. */
  758. /*D0*/ __u8 s_journal_uuid[16]; /* uuid of journal superblock */
  759. /*E0*/ __le32 s_journal_inum; /* inode number of journal file */
  760. __le32 s_journal_dev; /* device number of journal file */
  761. __le32 s_last_orphan; /* start of list of inodes to delete */
  762. __le32 s_hash_seed[4]; /* HTREE hash seed */
  763. __u8 s_def_hash_version; /* Default hash version to use */
  764. __u8 s_reserved_char_pad;
  765. __le16 s_desc_size; /* size of group descriptor */
  766. /*100*/ __le32 s_default_mount_opts;
  767. __le32 s_first_meta_bg; /* First metablock block group */
  768. __le32 s_mkfs_time; /* When the filesystem was created */
  769. __le32 s_jnl_blocks[17]; /* Backup of the journal inode */
  770. /* 64bit support valid if EXT4_FEATURE_COMPAT_64BIT */
  771. /*150*/ __le32 s_blocks_count_hi; /* Blocks count */
  772. __le32 s_r_blocks_count_hi; /* Reserved blocks count */
  773. __le32 s_free_blocks_count_hi; /* Free blocks count */
  774. __le16 s_min_extra_isize; /* All inodes have at least # bytes */
  775. __le16 s_want_extra_isize; /* New inodes should reserve # bytes */
  776. __le32 s_flags; /* Miscellaneous flags */
  777. __le16 s_raid_stride; /* RAID stride */
  778. __le16 s_mmp_interval; /* # seconds to wait in MMP checking */
  779. __le64 s_mmp_block; /* Block for multi-mount protection */
  780. __le32 s_raid_stripe_width; /* blocks on all data disks (N*stride)*/
  781. __u8 s_log_groups_per_flex; /* FLEX_BG group size */
  782. __u8 s_reserved_char_pad2;
  783. __le16 s_reserved_pad;
  784. __le64 s_kbytes_written; /* nr of lifetime kilobytes written */
  785. __u32 s_reserved[160]; /* Padding to the end of the block */
  786. };
  787. #ifdef __KERNEL__
  788. /*
  789. * run-time mount flags
  790. */
  791. #define EXT4_MF_MNTDIR_SAMPLED 0x0001
  792. #define EXT4_MF_FS_ABORTED 0x0002 /* Fatal error detected */
  793. /*
  794. * fourth extended-fs super-block data in memory
  795. */
  796. struct ext4_sb_info {
  797. unsigned long s_desc_size; /* Size of a group descriptor in bytes */
  798. unsigned long s_inodes_per_block;/* Number of inodes per block */
  799. unsigned long s_blocks_per_group;/* Number of blocks in a group */
  800. unsigned long s_inodes_per_group;/* Number of inodes in a group */
  801. unsigned long s_itb_per_group; /* Number of inode table blocks per group */
  802. unsigned long s_gdb_count; /* Number of group descriptor blocks */
  803. unsigned long s_desc_per_block; /* Number of group descriptors per block */
  804. ext4_group_t s_groups_count; /* Number of groups in the fs */
  805. ext4_group_t s_blockfile_groups;/* Groups acceptable for non-extent files */
  806. unsigned long s_overhead_last; /* Last calculated overhead */
  807. unsigned long s_blocks_last; /* Last seen block count */
  808. loff_t s_bitmap_maxbytes; /* max bytes for bitmap files */
  809. struct buffer_head * s_sbh; /* Buffer containing the super block */
  810. struct ext4_super_block *s_es; /* Pointer to the super block in the buffer */
  811. struct buffer_head **s_group_desc;
  812. unsigned int s_mount_opt;
  813. unsigned int s_mount_flags;
  814. ext4_fsblk_t s_sb_block;
  815. uid_t s_resuid;
  816. gid_t s_resgid;
  817. unsigned short s_mount_state;
  818. unsigned short s_pad;
  819. int s_addr_per_block_bits;
  820. int s_desc_per_block_bits;
  821. int s_inode_size;
  822. int s_first_ino;
  823. unsigned int s_inode_readahead_blks;
  824. unsigned int s_inode_goal;
  825. spinlock_t s_next_gen_lock;
  826. u32 s_next_generation;
  827. u32 s_hash_seed[4];
  828. int s_def_hash_version;
  829. int s_hash_unsigned; /* 3 if hash should be signed, 0 if not */
  830. struct percpu_counter s_freeblocks_counter;
  831. struct percpu_counter s_freeinodes_counter;
  832. struct percpu_counter s_dirs_counter;
  833. struct percpu_counter s_dirtyblocks_counter;
  834. struct blockgroup_lock *s_blockgroup_lock;
  835. struct proc_dir_entry *s_proc;
  836. struct kobject s_kobj;
  837. struct completion s_kobj_unregister;
  838. /* Journaling */
  839. struct inode *s_journal_inode;
  840. struct journal_s *s_journal;
  841. struct list_head s_orphan;
  842. struct mutex s_orphan_lock;
  843. struct mutex s_resize_lock;
  844. unsigned long s_commit_interval;
  845. u32 s_max_batch_time;
  846. u32 s_min_batch_time;
  847. struct block_device *journal_bdev;
  848. #ifdef CONFIG_JBD2_DEBUG
  849. struct timer_list turn_ro_timer; /* For turning read-only (crash simulation) */
  850. wait_queue_head_t ro_wait_queue; /* For people waiting for the fs to go read-only */
  851. #endif
  852. #ifdef CONFIG_QUOTA
  853. char *s_qf_names[MAXQUOTAS]; /* Names of quota files with journalled quota */
  854. int s_jquota_fmt; /* Format of quota to use */
  855. #endif
  856. unsigned int s_want_extra_isize; /* New inodes should reserve # bytes */
  857. struct rb_root system_blks;
  858. #ifdef EXTENTS_STATS
  859. /* ext4 extents stats */
  860. unsigned long s_ext_min;
  861. unsigned long s_ext_max;
  862. unsigned long s_depth_max;
  863. spinlock_t s_ext_stats_lock;
  864. unsigned long s_ext_blocks;
  865. unsigned long s_ext_extents;
  866. #endif
  867. /* for buddy allocator */
  868. struct ext4_group_info ***s_group_info;
  869. struct inode *s_buddy_cache;
  870. long s_blocks_reserved;
  871. spinlock_t s_reserve_lock;
  872. spinlock_t s_md_lock;
  873. tid_t s_last_transaction;
  874. unsigned short *s_mb_offsets;
  875. unsigned int *s_mb_maxs;
  876. /* tunables */
  877. unsigned long s_stripe;
  878. unsigned int s_mb_stream_request;
  879. unsigned int s_mb_max_to_scan;
  880. unsigned int s_mb_min_to_scan;
  881. unsigned int s_mb_stats;
  882. unsigned int s_mb_order2_reqs;
  883. unsigned int s_mb_group_prealloc;
  884. unsigned int s_max_writeback_mb_bump;
  885. /* where last allocation was done - for stream allocation */
  886. unsigned long s_mb_last_group;
  887. unsigned long s_mb_last_start;
  888. /* history to debug policy */
  889. struct ext4_mb_history *s_mb_history;
  890. int s_mb_history_cur;
  891. int s_mb_history_max;
  892. int s_mb_history_num;
  893. spinlock_t s_mb_history_lock;
  894. int s_mb_history_filter;
  895. /* stats for buddy allocator */
  896. spinlock_t s_mb_pa_lock;
  897. atomic_t s_bal_reqs; /* number of reqs with len > 1 */
  898. atomic_t s_bal_success; /* we found long enough chunks */
  899. atomic_t s_bal_allocated; /* in blocks */
  900. atomic_t s_bal_ex_scanned; /* total extents scanned */
  901. atomic_t s_bal_goals; /* goal hits */
  902. atomic_t s_bal_breaks; /* too long searches */
  903. atomic_t s_bal_2orders; /* 2^order hits */
  904. spinlock_t s_bal_lock;
  905. unsigned long s_mb_buddies_generated;
  906. unsigned long long s_mb_generation_time;
  907. atomic_t s_mb_lost_chunks;
  908. atomic_t s_mb_preallocated;
  909. atomic_t s_mb_discarded;
  910. atomic_t s_lock_busy;
  911. /* locality groups */
  912. struct ext4_locality_group *s_locality_groups;
  913. /* for write statistics */
  914. unsigned long s_sectors_written_start;
  915. u64 s_kbytes_written;
  916. unsigned int s_log_groups_per_flex;
  917. struct flex_groups *s_flex_groups;
  918. /* workqueue for dio unwritten */
  919. struct workqueue_struct *dio_unwritten_wq;
  920. };
  921. static inline struct ext4_sb_info *EXT4_SB(struct super_block *sb)
  922. {
  923. return sb->s_fs_info;
  924. }
  925. static inline struct ext4_inode_info *EXT4_I(struct inode *inode)
  926. {
  927. return container_of(inode, struct ext4_inode_info, vfs_inode);
  928. }
  929. static inline struct timespec ext4_current_time(struct inode *inode)
  930. {
  931. return (inode->i_sb->s_time_gran < NSEC_PER_SEC) ?
  932. current_fs_time(inode->i_sb) : CURRENT_TIME_SEC;
  933. }
  934. static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino)
  935. {
  936. return ino == EXT4_ROOT_INO ||
  937. ino == EXT4_JOURNAL_INO ||
  938. ino == EXT4_RESIZE_INO ||
  939. (ino >= EXT4_FIRST_INO(sb) &&
  940. ino <= le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count));
  941. }
  942. #else
  943. /* Assume that user mode programs are passing in an ext4fs superblock, not
  944. * a kernel struct super_block. This will allow us to call the feature-test
  945. * macros from user land. */
  946. #define EXT4_SB(sb) (sb)
  947. #endif
  948. #define NEXT_ORPHAN(inode) EXT4_I(inode)->i_dtime
  949. /*
  950. * Codes for operating systems
  951. */
  952. #define EXT4_OS_LINUX 0
  953. #define EXT4_OS_HURD 1
  954. #define EXT4_OS_MASIX 2
  955. #define EXT4_OS_FREEBSD 3
  956. #define EXT4_OS_LITES 4
  957. /*
  958. * Revision levels
  959. */
  960. #define EXT4_GOOD_OLD_REV 0 /* The good old (original) format */
  961. #define EXT4_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
  962. #define EXT4_CURRENT_REV EXT4_GOOD_OLD_REV
  963. #define EXT4_MAX_SUPP_REV EXT4_DYNAMIC_REV
  964. #define EXT4_GOOD_OLD_INODE_SIZE 128
  965. /*
  966. * Feature set definitions
  967. */
  968. #define EXT4_HAS_COMPAT_FEATURE(sb,mask) \
  969. ((EXT4_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask)) != 0)
  970. #define EXT4_HAS_RO_COMPAT_FEATURE(sb,mask) \
  971. ((EXT4_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask)) != 0)
  972. #define EXT4_HAS_INCOMPAT_FEATURE(sb,mask) \
  973. ((EXT4_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask)) != 0)
  974. #define EXT4_SET_COMPAT_FEATURE(sb,mask) \
  975. EXT4_SB(sb)->s_es->s_feature_compat |= cpu_to_le32(mask)
  976. #define EXT4_SET_RO_COMPAT_FEATURE(sb,mask) \
  977. EXT4_SB(sb)->s_es->s_feature_ro_compat |= cpu_to_le32(mask)
  978. #define EXT4_SET_INCOMPAT_FEATURE(sb,mask) \
  979. EXT4_SB(sb)->s_es->s_feature_incompat |= cpu_to_le32(mask)
  980. #define EXT4_CLEAR_COMPAT_FEATURE(sb,mask) \
  981. EXT4_SB(sb)->s_es->s_feature_compat &= ~cpu_to_le32(mask)
  982. #define EXT4_CLEAR_RO_COMPAT_FEATURE(sb,mask) \
  983. EXT4_SB(sb)->s_es->s_feature_ro_compat &= ~cpu_to_le32(mask)
  984. #define EXT4_CLEAR_INCOMPAT_FEATURE(sb,mask) \
  985. EXT4_SB(sb)->s_es->s_feature_incompat &= ~cpu_to_le32(mask)
  986. #define EXT4_FEATURE_COMPAT_DIR_PREALLOC 0x0001
  987. #define EXT4_FEATURE_COMPAT_IMAGIC_INODES 0x0002
  988. #define EXT4_FEATURE_COMPAT_HAS_JOURNAL 0x0004
  989. #define EXT4_FEATURE_COMPAT_EXT_ATTR 0x0008
  990. #define EXT4_FEATURE_COMPAT_RESIZE_INODE 0x0010
  991. #define EXT4_FEATURE_COMPAT_DIR_INDEX 0x0020
  992. #define EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
  993. #define EXT4_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
  994. #define EXT4_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
  995. #define EXT4_FEATURE_RO_COMPAT_HUGE_FILE 0x0008
  996. #define EXT4_FEATURE_RO_COMPAT_GDT_CSUM 0x0010
  997. #define EXT4_FEATURE_RO_COMPAT_DIR_NLINK 0x0020
  998. #define EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 0x0040
  999. #define EXT4_FEATURE_INCOMPAT_COMPRESSION 0x0001
  1000. #define EXT4_FEATURE_INCOMPAT_FILETYPE 0x0002
  1001. #define EXT4_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */
  1002. #define EXT4_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */
  1003. #define EXT4_FEATURE_INCOMPAT_META_BG 0x0010
  1004. #define EXT4_FEATURE_INCOMPAT_EXTENTS 0x0040 /* extents support */
  1005. #define EXT4_FEATURE_INCOMPAT_64BIT 0x0080
  1006. #define EXT4_FEATURE_INCOMPAT_MMP 0x0100
  1007. #define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
  1008. #define EXT4_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR
  1009. #define EXT4_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
  1010. EXT4_FEATURE_INCOMPAT_RECOVER| \
  1011. EXT4_FEATURE_INCOMPAT_META_BG| \
  1012. EXT4_FEATURE_INCOMPAT_EXTENTS| \
  1013. EXT4_FEATURE_INCOMPAT_64BIT| \
  1014. EXT4_FEATURE_INCOMPAT_FLEX_BG)
  1015. #define EXT4_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
  1016. EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
  1017. EXT4_FEATURE_RO_COMPAT_GDT_CSUM| \
  1018. EXT4_FEATURE_RO_COMPAT_DIR_NLINK | \
  1019. EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE | \
  1020. EXT4_FEATURE_RO_COMPAT_BTREE_DIR |\
  1021. EXT4_FEATURE_RO_COMPAT_HUGE_FILE)
  1022. /*
  1023. * Default values for user and/or group using reserved blocks
  1024. */
  1025. #define EXT4_DEF_RESUID 0
  1026. #define EXT4_DEF_RESGID 0
  1027. #define EXT4_DEF_INODE_READAHEAD_BLKS 32
  1028. /*
  1029. * Default mount options
  1030. */
  1031. #define EXT4_DEFM_DEBUG 0x0001
  1032. #define EXT4_DEFM_BSDGROUPS 0x0002
  1033. #define EXT4_DEFM_XATTR_USER 0x0004
  1034. #define EXT4_DEFM_ACL 0x0008
  1035. #define EXT4_DEFM_UID16 0x0010
  1036. #define EXT4_DEFM_JMODE 0x0060
  1037. #define EXT4_DEFM_JMODE_DATA 0x0020
  1038. #define EXT4_DEFM_JMODE_ORDERED 0x0040
  1039. #define EXT4_DEFM_JMODE_WBACK 0x0060
  1040. /*
  1041. * Default journal batch times
  1042. */
  1043. #define EXT4_DEF_MIN_BATCH_TIME 0
  1044. #define EXT4_DEF_MAX_BATCH_TIME 15000 /* 15ms */
  1045. /*
  1046. * Minimum number of groups in a flexgroup before we separate out
  1047. * directories into the first block group of a flexgroup
  1048. */
  1049. #define EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME 4
  1050. /*
  1051. * Structure of a directory entry
  1052. */
  1053. #define EXT4_NAME_LEN 255
  1054. struct ext4_dir_entry {
  1055. __le32 inode; /* Inode number */
  1056. __le16 rec_len; /* Directory entry length */
  1057. __le16 name_len; /* Name length */
  1058. char name[EXT4_NAME_LEN]; /* File name */
  1059. };
  1060. /*
  1061. * The new version of the directory entry. Since EXT4 structures are
  1062. * stored in intel byte order, and the name_len field could never be
  1063. * bigger than 255 chars, it's safe to reclaim the extra byte for the
  1064. * file_type field.
  1065. */
  1066. struct ext4_dir_entry_2 {
  1067. __le32 inode; /* Inode number */
  1068. __le16 rec_len; /* Directory entry length */
  1069. __u8 name_len; /* Name length */
  1070. __u8 file_type;
  1071. char name[EXT4_NAME_LEN]; /* File name */
  1072. };
  1073. /*
  1074. * Ext4 directory file types. Only the low 3 bits are used. The
  1075. * other bits are reserved for now.
  1076. */
  1077. #define EXT4_FT_UNKNOWN 0
  1078. #define EXT4_FT_REG_FILE 1
  1079. #define EXT4_FT_DIR 2
  1080. #define EXT4_FT_CHRDEV 3
  1081. #define EXT4_FT_BLKDEV 4
  1082. #define EXT4_FT_FIFO 5
  1083. #define EXT4_FT_SOCK 6
  1084. #define EXT4_FT_SYMLINK 7
  1085. #define EXT4_FT_MAX 8
  1086. /*
  1087. * EXT4_DIR_PAD defines the directory entries boundaries
  1088. *
  1089. * NOTE: It must be a multiple of 4
  1090. */
  1091. #define EXT4_DIR_PAD 4
  1092. #define EXT4_DIR_ROUND (EXT4_DIR_PAD - 1)
  1093. #define EXT4_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT4_DIR_ROUND) & \
  1094. ~EXT4_DIR_ROUND)
  1095. #define EXT4_MAX_REC_LEN ((1<<16)-1)
  1096. /*
  1097. * Hash Tree Directory indexing
  1098. * (c) Daniel Phillips, 2001
  1099. */
  1100. #define is_dx(dir) (EXT4_HAS_COMPAT_FEATURE(dir->i_sb, \
  1101. EXT4_FEATURE_COMPAT_DIR_INDEX) && \
  1102. (EXT4_I(dir)->i_flags & EXT4_INDEX_FL))
  1103. #define EXT4_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT4_LINK_MAX)
  1104. #define EXT4_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
  1105. /* Legal values for the dx_root hash_version field: */
  1106. #define DX_HASH_LEGACY 0
  1107. #define DX_HASH_HALF_MD4 1
  1108. #define DX_HASH_TEA 2
  1109. #define DX_HASH_LEGACY_UNSIGNED 3
  1110. #define DX_HASH_HALF_MD4_UNSIGNED 4
  1111. #define DX_HASH_TEA_UNSIGNED 5
  1112. #ifdef __KERNEL__
  1113. /* hash info structure used by the directory hash */
  1114. struct dx_hash_info
  1115. {
  1116. u32 hash;
  1117. u32 minor_hash;
  1118. int hash_version;
  1119. u32 *seed;
  1120. };
  1121. #define EXT4_HTREE_EOF 0x7fffffff
  1122. /*
  1123. * Control parameters used by ext4_htree_next_block
  1124. */
  1125. #define HASH_NB_ALWAYS 1
  1126. /*
  1127. * Describe an inode's exact location on disk and in memory
  1128. */
  1129. struct ext4_iloc
  1130. {
  1131. struct buffer_head *bh;
  1132. unsigned long offset;
  1133. ext4_group_t block_group;
  1134. };
  1135. static inline struct ext4_inode *ext4_raw_inode(struct ext4_iloc *iloc)
  1136. {
  1137. return (struct ext4_inode *) (iloc->bh->b_data + iloc->offset);
  1138. }
  1139. /*
  1140. * This structure is stuffed into the struct file's private_data field
  1141. * for directories. It is where we put information so that we can do
  1142. * readdir operations in hash tree order.
  1143. */
  1144. struct dir_private_info {
  1145. struct rb_root root;
  1146. struct rb_node *curr_node;
  1147. struct fname *extra_fname;
  1148. loff_t last_pos;
  1149. __u32 curr_hash;
  1150. __u32 curr_minor_hash;
  1151. __u32 next_hash;
  1152. };
  1153. /* calculate the first block number of the group */
  1154. static inline ext4_fsblk_t
  1155. ext4_group_first_block_no(struct super_block *sb, ext4_group_t group_no)
  1156. {
  1157. return group_no * (ext4_fsblk_t)EXT4_BLOCKS_PER_GROUP(sb) +
  1158. le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
  1159. }
  1160. /*
  1161. * Special error return code only used by dx_probe() and its callers.
  1162. */
  1163. #define ERR_BAD_DX_DIR -75000
  1164. void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr,
  1165. ext4_group_t *blockgrpp, ext4_grpblk_t *offsetp);
  1166. extern struct proc_dir_entry *ext4_proc_root;
  1167. /*
  1168. * Function prototypes
  1169. */
  1170. /*
  1171. * Ok, these declarations are also in <linux/kernel.h> but none of the
  1172. * ext4 source programs needs to include it so they are duplicated here.
  1173. */
  1174. # define NORET_TYPE /**/
  1175. # define ATTRIB_NORET __attribute__((noreturn))
  1176. # define NORET_AND noreturn,
  1177. /* bitmap.c */
  1178. extern unsigned int ext4_count_free(struct buffer_head *, unsigned);
  1179. /* balloc.c */
  1180. extern unsigned int ext4_block_group(struct super_block *sb,
  1181. ext4_fsblk_t blocknr);
  1182. extern ext4_grpblk_t ext4_block_group_offset(struct super_block *sb,
  1183. ext4_fsblk_t blocknr);
  1184. extern int ext4_bg_has_super(struct super_block *sb, ext4_group_t group);
  1185. extern unsigned long ext4_bg_num_gdb(struct super_block *sb,
  1186. ext4_group_t group);
  1187. extern ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
  1188. ext4_fsblk_t goal, unsigned long *count, int *errp);
  1189. extern int ext4_claim_free_blocks(struct ext4_sb_info *sbi, s64 nblocks);
  1190. extern int ext4_has_free_blocks(struct ext4_sb_info *sbi, s64 nblocks);
  1191. extern void ext4_free_blocks(handle_t *handle, struct inode *inode,
  1192. ext4_fsblk_t block, unsigned long count, int metadata);
  1193. extern void ext4_add_groupblocks(handle_t *handle, struct super_block *sb,
  1194. ext4_fsblk_t block, unsigned long count);
  1195. extern ext4_fsblk_t ext4_count_free_blocks(struct super_block *);
  1196. extern void ext4_check_blocks_bitmap(struct super_block *);
  1197. extern struct ext4_group_desc * ext4_get_group_desc(struct super_block * sb,
  1198. ext4_group_t block_group,
  1199. struct buffer_head ** bh);
  1200. extern int ext4_should_retry_alloc(struct super_block *sb, int *retries);
  1201. struct buffer_head *ext4_read_block_bitmap(struct super_block *sb,
  1202. ext4_group_t block_group);
  1203. extern unsigned ext4_init_block_bitmap(struct super_block *sb,
  1204. struct buffer_head *bh,
  1205. ext4_group_t group,
  1206. struct ext4_group_desc *desc);
  1207. #define ext4_free_blocks_after_init(sb, group, desc) \
  1208. ext4_init_block_bitmap(sb, NULL, group, desc)
  1209. /* dir.c */
  1210. extern int ext4_check_dir_entry(const char *, struct inode *,
  1211. struct ext4_dir_entry_2 *,
  1212. struct buffer_head *, unsigned int);
  1213. extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
  1214. __u32 minor_hash,
  1215. struct ext4_dir_entry_2 *dirent);
  1216. extern void ext4_htree_free_dir_info(struct dir_private_info *p);
  1217. /* fsync.c */
  1218. extern int ext4_sync_file(struct file *, struct dentry *, int);
  1219. /* hash.c */
  1220. extern int ext4fs_dirhash(const char *name, int len, struct
  1221. dx_hash_info *hinfo);
  1222. /* ialloc.c */
  1223. extern struct inode *ext4_new_inode(handle_t *, struct inode *, int,
  1224. const struct qstr *qstr, __u32 goal);
  1225. extern void ext4_free_inode(handle_t *, struct inode *);
  1226. extern struct inode * ext4_orphan_get(struct super_block *, unsigned long);
  1227. extern unsigned long ext4_count_free_inodes(struct super_block *);
  1228. extern unsigned long ext4_count_dirs(struct super_block *);
  1229. extern void ext4_check_inodes_bitmap(struct super_block *);
  1230. extern unsigned ext4_init_inode_bitmap(struct super_block *sb,
  1231. struct buffer_head *bh,
  1232. ext4_group_t group,
  1233. struct ext4_group_desc *desc);
  1234. extern void mark_bitmap_end(int start_bit, int end_bit, char *bitmap);
  1235. /* mballoc.c */
  1236. extern long ext4_mb_stats;
  1237. extern long ext4_mb_max_to_scan;
  1238. extern int ext4_mb_init(struct super_block *, int);
  1239. extern int ext4_mb_release(struct super_block *);
  1240. extern ext4_fsblk_t ext4_mb_new_blocks(handle_t *,
  1241. struct ext4_allocation_request *, int *);
  1242. extern int ext4_mb_reserve_blocks(struct super_block *, int);
  1243. extern void ext4_discard_preallocations(struct inode *);
  1244. extern int __init init_ext4_mballoc(void);
  1245. extern void exit_ext4_mballoc(void);
  1246. extern void ext4_mb_free_blocks(handle_t *, struct inode *,
  1247. ext4_fsblk_t, unsigned long, int, unsigned long *);
  1248. extern int ext4_mb_add_groupinfo(struct super_block *sb,
  1249. ext4_group_t i, struct ext4_group_desc *desc);
  1250. extern int ext4_mb_get_buddy_cache_lock(struct super_block *, ext4_group_t);
  1251. extern void ext4_mb_put_buddy_cache_lock(struct super_block *,
  1252. ext4_group_t, int);
  1253. /* inode.c */
  1254. int ext4_forget(handle_t *handle, int is_metadata, struct inode *inode,
  1255. struct buffer_head *bh, ext4_fsblk_t blocknr);
  1256. struct buffer_head *ext4_getblk(handle_t *, struct inode *,
  1257. ext4_lblk_t, int, int *);
  1258. struct buffer_head *ext4_bread(handle_t *, struct inode *,
  1259. ext4_lblk_t, int, int *);
  1260. int ext4_get_block(struct inode *inode, sector_t iblock,
  1261. struct buffer_head *bh_result, int create);
  1262. extern struct inode *ext4_iget(struct super_block *, unsigned long);
  1263. extern int ext4_write_inode(struct inode *, int);
  1264. extern int ext4_setattr(struct dentry *, struct iattr *);
  1265. extern int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry,
  1266. struct kstat *stat);
  1267. extern void ext4_delete_inode(struct inode *);
  1268. extern int ext4_sync_inode(handle_t *, struct inode *);
  1269. extern void ext4_dirty_inode(struct inode *);
  1270. extern int ext4_change_inode_journal_flag(struct inode *, int);
  1271. extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *);
  1272. extern int ext4_can_truncate(struct inode *inode);
  1273. extern void ext4_truncate(struct inode *);
  1274. extern int ext4_truncate_restart_trans(handle_t *, struct inode *, int nblocks);
  1275. extern void ext4_set_inode_flags(struct inode *);
  1276. extern void ext4_get_inode_flags(struct ext4_inode_info *);
  1277. extern int ext4_alloc_da_blocks(struct inode *inode);
  1278. extern void ext4_set_aops(struct inode *inode);
  1279. extern int ext4_writepage_trans_blocks(struct inode *);
  1280. extern int ext4_meta_trans_blocks(struct inode *, int nrblocks, int idxblocks);
  1281. extern int ext4_chunk_trans_blocks(struct inode *, int nrblocks);
  1282. extern int ext4_block_truncate_page(handle_t *handle,
  1283. struct address_space *mapping, loff_t from);
  1284. extern int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
  1285. extern qsize_t ext4_get_reserved_space(struct inode *inode);
  1286. extern int flush_aio_dio_completed_IO(struct inode *inode);
  1287. /* ioctl.c */
  1288. extern long ext4_ioctl(struct file *, unsigned int, unsigned long);
  1289. extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long);
  1290. /* migrate.c */
  1291. extern int ext4_ext_migrate(struct inode *);
  1292. /* namei.c */
  1293. extern unsigned int ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize);
  1294. extern __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize);
  1295. extern int ext4_orphan_add(handle_t *, struct inode *);
  1296. extern int ext4_orphan_del(handle_t *, struct inode *);
  1297. extern int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
  1298. __u32 start_minor_hash, __u32 *next_hash);
  1299. /* resize.c */
  1300. extern int ext4_group_add(struct super_block *sb,
  1301. struct ext4_new_group_data *input);
  1302. extern int ext4_group_extend(struct super_block *sb,
  1303. struct ext4_super_block *es,
  1304. ext4_fsblk_t n_blocks_count);
  1305. /* super.c */
  1306. extern void ext4_error(struct super_block *, const char *, const char *, ...)
  1307. __attribute__ ((format (printf, 3, 4)));
  1308. extern void __ext4_std_error(struct super_block *, const char *, int);
  1309. extern void ext4_abort(struct super_block *, const char *, const char *, ...)
  1310. __attribute__ ((format (printf, 3, 4)));
  1311. extern void ext4_warning(struct super_block *, const char *, const char *, ...)
  1312. __attribute__ ((format (printf, 3, 4)));
  1313. extern void ext4_msg(struct super_block *, const char *, const char *, ...)
  1314. __attribute__ ((format (printf, 3, 4)));
  1315. extern void ext4_grp_locked_error(struct super_block *, ext4_group_t,
  1316. const char *, const char *, ...)
  1317. __attribute__ ((format (printf, 4, 5)));
  1318. extern void ext4_update_dynamic_rev(struct super_block *sb);
  1319. extern int ext4_update_compat_feature(handle_t *handle, struct super_block *sb,
  1320. __u32 compat);
  1321. extern int ext4_update_rocompat_feature(handle_t *handle,
  1322. struct super_block *sb, __u32 rocompat);
  1323. extern int ext4_update_incompat_feature(handle_t *handle,
  1324. struct super_block *sb, __u32 incompat);
  1325. extern ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
  1326. struct ext4_group_desc *bg);
  1327. extern ext4_fsblk_t ext4_inode_bitmap(struct super_block *sb,
  1328. struct ext4_group_desc *bg);
  1329. extern ext4_fsblk_t ext4_inode_table(struct super_block *sb,
  1330. struct ext4_group_desc *bg);
  1331. extern __u32 ext4_free_blks_count(struct super_block *sb,
  1332. struct ext4_group_desc *bg);
  1333. extern __u32 ext4_free_inodes_count(struct super_block *sb,
  1334. struct ext4_group_desc *bg);
  1335. extern __u32 ext4_used_dirs_count(struct super_block *sb,
  1336. struct ext4_group_desc *bg);
  1337. extern __u32 ext4_itable_unused_count(struct super_block *sb,
  1338. struct ext4_group_desc *bg);
  1339. extern void ext4_block_bitmap_set(struct super_block *sb,
  1340. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1341. extern void ext4_inode_bitmap_set(struct super_block *sb,
  1342. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1343. extern void ext4_inode_table_set(struct super_block *sb,
  1344. struct ext4_group_desc *bg, ext4_fsblk_t blk);
  1345. extern void ext4_free_blks_set(struct super_block *sb,
  1346. struct ext4_group_desc *bg, __u32 count);
  1347. extern void ext4_free_inodes_set(struct super_block *sb,
  1348. struct ext4_group_desc *bg, __u32 count);
  1349. extern void ext4_used_dirs_set(struct super_block *sb,
  1350. struct ext4_group_desc *bg, __u32 count);
  1351. extern void ext4_itable_unused_set(struct super_block *sb,
  1352. struct ext4_group_desc *bg, __u32 count);
  1353. extern __le16 ext4_group_desc_csum(struct ext4_sb_info *sbi, __u32 group,
  1354. struct ext4_group_desc *gdp);
  1355. extern int ext4_group_desc_csum_verify(struct ext4_sb_info *sbi, __u32 group,
  1356. struct ext4_group_desc *gdp);
  1357. static inline ext4_fsblk_t ext4_blocks_count(struct ext4_super_block *es)
  1358. {
  1359. return ((ext4_fsblk_t)le32_to_cpu(es->s_blocks_count_hi) << 32) |
  1360. le32_to_cpu(es->s_blocks_count_lo);
  1361. }
  1362. static inline ext4_fsblk_t ext4_r_blocks_count(struct ext4_super_block *es)
  1363. {
  1364. return ((ext4_fsblk_t)le32_to_cpu(es->s_r_blocks_count_hi) << 32) |
  1365. le32_to_cpu(es->s_r_blocks_count_lo);
  1366. }
  1367. static inline ext4_fsblk_t ext4_free_blocks_count(struct ext4_super_block *es)
  1368. {
  1369. return ((ext4_fsblk_t)le32_to_cpu(es->s_free_blocks_count_hi) << 32) |
  1370. le32_to_cpu(es->s_free_blocks_count_lo);
  1371. }
  1372. static inline void ext4_blocks_count_set(struct ext4_super_block *es,
  1373. ext4_fsblk_t blk)
  1374. {
  1375. es->s_blocks_count_lo = cpu_to_le32((u32)blk);
  1376. es->s_blocks_count_hi = cpu_to_le32(blk >> 32);
  1377. }
  1378. static inline void ext4_free_blocks_count_set(struct ext4_super_block *es,
  1379. ext4_fsblk_t blk)
  1380. {
  1381. es->s_free_blocks_count_lo = cpu_to_le32((u32)blk);
  1382. es->s_free_blocks_count_hi = cpu_to_le32(blk >> 32);
  1383. }
  1384. static inline void ext4_r_blocks_count_set(struct ext4_super_block *es,
  1385. ext4_fsblk_t blk)
  1386. {
  1387. es->s_r_blocks_count_lo = cpu_to_le32((u32)blk);
  1388. es->s_r_blocks_count_hi = cpu_to_le32(blk >> 32);
  1389. }
  1390. static inline loff_t ext4_isize(struct ext4_inode *raw_inode)
  1391. {
  1392. if (S_ISREG(le16_to_cpu(raw_inode->i_mode)))
  1393. return ((loff_t)le32_to_cpu(raw_inode->i_size_high) << 32) |
  1394. le32_to_cpu(raw_inode->i_size_lo);
  1395. else
  1396. return (loff_t) le32_to_cpu(raw_inode->i_size_lo);
  1397. }
  1398. static inline void ext4_isize_set(struct ext4_inode *raw_inode, loff_t i_size)
  1399. {
  1400. raw_inode->i_size_lo = cpu_to_le32(i_size);
  1401. raw_inode->i_size_high = cpu_to_le32(i_size >> 32);
  1402. }
  1403. static inline
  1404. struct ext4_group_info *ext4_get_group_info(struct super_block *sb,
  1405. ext4_group_t group)
  1406. {
  1407. struct ext4_group_info ***grp_info;
  1408. long indexv, indexh;
  1409. grp_info = EXT4_SB(sb)->s_group_info;
  1410. indexv = group >> (EXT4_DESC_PER_BLOCK_BITS(sb));
  1411. indexh = group & ((EXT4_DESC_PER_BLOCK(sb)) - 1);
  1412. return grp_info[indexv][indexh];
  1413. }
  1414. /*
  1415. * Reading s_groups_count requires using smp_rmb() afterwards. See
  1416. * the locking protocol documented in the comments of ext4_group_add()
  1417. * in resize.c
  1418. */
  1419. static inline ext4_group_t ext4_get_groups_count(struct super_block *sb)
  1420. {
  1421. ext4_group_t ngroups = EXT4_SB(sb)->s_groups_count;
  1422. smp_rmb();
  1423. return ngroups;
  1424. }
  1425. static inline ext4_group_t ext4_flex_group(struct ext4_sb_info *sbi,
  1426. ext4_group_t block_group)
  1427. {
  1428. return block_group >> sbi->s_log_groups_per_flex;
  1429. }
  1430. static inline unsigned int ext4_flex_bg_size(struct ext4_sb_info *sbi)
  1431. {
  1432. return 1 << sbi->s_log_groups_per_flex;
  1433. }
  1434. #define ext4_std_error(sb, errno) \
  1435. do { \
  1436. if ((errno)) \
  1437. __ext4_std_error((sb), __func__, (errno)); \
  1438. } while (0)
  1439. #ifdef CONFIG_SMP
  1440. /* Each CPU can accumulate percpu_counter_batch blocks in their local
  1441. * counters. So we need to make sure we have free blocks more
  1442. * than percpu_counter_batch * nr_cpu_ids. Also add a window of 4 times.
  1443. */
  1444. #define EXT4_FREEBLOCKS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
  1445. #else
  1446. #define EXT4_FREEBLOCKS_WATERMARK 0
  1447. #endif
  1448. static inline void ext4_update_i_disksize(struct inode *inode, loff_t newsize)
  1449. {
  1450. /*
  1451. * XXX: replace with spinlock if seen contended -bzzz
  1452. */
  1453. down_write(&EXT4_I(inode)->i_data_sem);
  1454. if (newsize > EXT4_I(inode)->i_disksize)
  1455. EXT4_I(inode)->i_disksize = newsize;
  1456. up_write(&EXT4_I(inode)->i_data_sem);
  1457. return ;
  1458. }
  1459. struct ext4_group_info {
  1460. unsigned long bb_state;
  1461. struct rb_root bb_free_root;
  1462. ext4_grpblk_t bb_first_free; /* first free block */
  1463. ext4_grpblk_t bb_free; /* total free blocks */
  1464. ext4_grpblk_t bb_fragments; /* nr of freespace fragments */
  1465. struct list_head bb_prealloc_list;
  1466. #ifdef DOUBLE_CHECK
  1467. void *bb_bitmap;
  1468. #endif
  1469. struct rw_semaphore alloc_sem;
  1470. ext4_grpblk_t bb_counters[]; /* Nr of free power-of-two-block
  1471. * regions, index is order.
  1472. * bb_counters[3] = 5 means
  1473. * 5 free 8-block regions. */
  1474. };
  1475. #define EXT4_GROUP_INFO_NEED_INIT_BIT 0
  1476. #define EXT4_MB_GRP_NEED_INIT(grp) \
  1477. (test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
  1478. #define EXT4_MAX_CONTENTION 8
  1479. #define EXT4_CONTENTION_THRESHOLD 2
  1480. static inline spinlock_t *ext4_group_lock_ptr(struct super_block *sb,
  1481. ext4_group_t group)
  1482. {
  1483. return bgl_lock_ptr(EXT4_SB(sb)->s_blockgroup_lock, group);
  1484. }
  1485. /*
  1486. * Returns true if the filesystem is busy enough that attempts to
  1487. * access the block group locks has run into contention.
  1488. */
  1489. static inline int ext4_fs_is_busy(struct ext4_sb_info *sbi)
  1490. {
  1491. return (atomic_read(&sbi->s_lock_busy) > EXT4_CONTENTION_THRESHOLD);
  1492. }
  1493. static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
  1494. {
  1495. spinlock_t *lock = ext4_group_lock_ptr(sb, group);
  1496. if (spin_trylock(lock))
  1497. /*
  1498. * We're able to grab the lock right away, so drop the
  1499. * lock contention counter.
  1500. */
  1501. atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, -1, 0);
  1502. else {
  1503. /*
  1504. * The lock is busy, so bump the contention counter,
  1505. * and then wait on the spin lock.
  1506. */
  1507. atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, 1,
  1508. EXT4_MAX_CONTENTION);
  1509. spin_lock(lock);
  1510. }
  1511. }
  1512. static inline void ext4_unlock_group(struct super_block *sb,
  1513. ext4_group_t group)
  1514. {
  1515. spin_unlock(ext4_group_lock_ptr(sb, group));
  1516. }
  1517. /*
  1518. * Inodes and files operations
  1519. */
  1520. /* dir.c */
  1521. extern const struct file_operations ext4_dir_operations;
  1522. /* file.c */
  1523. extern const struct inode_operations ext4_file_inode_operations;
  1524. extern const struct file_operations ext4_file_operations;
  1525. /* namei.c */
  1526. extern const struct inode_operations ext4_dir_inode_operations;
  1527. extern const struct inode_operations ext4_special_inode_operations;
  1528. extern struct dentry *ext4_get_parent(struct dentry *child);
  1529. /* symlink.c */
  1530. extern const struct inode_operations ext4_symlink_inode_operations;
  1531. extern const struct inode_operations ext4_fast_symlink_inode_operations;
  1532. /* block_validity */
  1533. extern void ext4_release_system_zone(struct super_block *sb);
  1534. extern int ext4_setup_system_zone(struct super_block *sb);
  1535. extern int __init init_ext4_system_zone(void);
  1536. extern void exit_ext4_system_zone(void);
  1537. extern int ext4_data_block_valid(struct ext4_sb_info *sbi,
  1538. ext4_fsblk_t start_blk,
  1539. unsigned int count);
  1540. /* extents.c */
  1541. extern int ext4_ext_tree_init(handle_t *handle, struct inode *);
  1542. extern int ext4_ext_writepage_trans_blocks(struct inode *, int);
  1543. extern int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks,
  1544. int chunk);
  1545. extern int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
  1546. ext4_lblk_t iblock, unsigned int max_blocks,
  1547. struct buffer_head *bh_result, int flags);
  1548. extern void ext4_ext_truncate(struct inode *);
  1549. extern void ext4_ext_init(struct super_block *);
  1550. extern void ext4_ext_release(struct super_block *);
  1551. extern long ext4_fallocate(struct inode *inode, int mode, loff_t offset,
  1552. loff_t len);
  1553. extern int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
  1554. loff_t len);
  1555. extern int ext4_get_blocks(handle_t *handle, struct inode *inode,
  1556. sector_t block, unsigned int max_blocks,
  1557. struct buffer_head *bh, int flags);
  1558. extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
  1559. __u64 start, __u64 len);
  1560. /* move_extent.c */
  1561. extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
  1562. __u64 start_orig, __u64 start_donor,
  1563. __u64 len, __u64 *moved_len);
  1564. /*
  1565. * Add new method to test wether block and inode bitmaps are properly
  1566. * initialized. With uninit_bg reading the block from disk is not enough
  1567. * to mark the bitmap uptodate. We need to also zero-out the bitmap
  1568. */
  1569. #define BH_BITMAP_UPTODATE BH_JBDPrivateStart
  1570. static inline int bitmap_uptodate(struct buffer_head *bh)
  1571. {
  1572. return (buffer_uptodate(bh) &&
  1573. test_bit(BH_BITMAP_UPTODATE, &(bh)->b_state));
  1574. }
  1575. static inline void set_bitmap_uptodate(struct buffer_head *bh)
  1576. {
  1577. set_bit(BH_BITMAP_UPTODATE, &(bh)->b_state);
  1578. }
  1579. #endif /* __KERNEL__ */
  1580. #endif /* _EXT4_H */