| OLD | NEW |
|---|
| (Empty) | |
| 1 /* |
| 2 ** 2004 April 6 |
| 3 ** |
| 4 ** The author disclaims copyright to this source code. In place of |
| 5 ** a legal notice, here is a blessing: |
| 6 ** |
| 7 ** May you do good and not evil. |
| 8 ** May you find forgiveness for yourself and forgive others. |
| 9 ** May you share freely, never taking more than you give. |
| 10 ** |
| 11 ************************************************************************* |
| 12 ** This file implements a external (disk-based) database using BTrees. |
| 13 ** For a detailed discussion of BTrees, refer to |
| 14 ** |
| 15 ** Donald E. Knuth, THE ART OF COMPUTER PROGRAMMING, Volume 3: |
| 16 ** "Sorting And Searching", pages 473-480. Addison-Wesley |
| 17 ** Publishing Company, Reading, Massachusetts. |
| 18 ** |
| 19 ** The basic idea is that each page of the file contains N database |
| 20 ** entries and N+1 pointers to subpages. |
| 21 ** |
| 22 ** ---------------------------------------------------------------- |
| 23 ** | Ptr(0) | Key(0) | Ptr(1) | Key(1) | ... | Key(N-1) | Ptr(N) | |
| 24 ** ---------------------------------------------------------------- |
| 25 ** |
| 26 ** All of the keys on the page that Ptr(0) points to have values less |
| 27 ** than Key(0). All of the keys on page Ptr(1) and its subpages have |
| 28 ** values greater than Key(0) and less than Key(1). All of the keys |
| 29 ** on Ptr(N) and its subpages have values greater than Key(N-1). And |
| 30 ** so forth. |
| 31 ** |
| 32 ** Finding a particular key requires reading O(log(M)) pages from the |
| 33 ** disk where M is the number of entries in the tree. |
| 34 ** |
| 35 ** In this implementation, a single file can hold one or more separate |
| 36 ** BTrees. Each BTree is identified by the index of its root page. The |
| 37 ** key and data for any entry are combined to form the "payload". A |
| 38 ** fixed amount of payload can be carried directly on the database |
| 39 ** page. If the payload is larger than the preset amount then surplus |
| 40 ** bytes are stored on overflow pages. The payload for an entry |
| 41 ** and the preceding pointer are combined to form a "Cell". Each |
| 42 ** page has a small header which contains the Ptr(N) pointer and other |
| 43 ** information such as the size of key and data. |
| 44 ** |
| 45 ** FORMAT DETAILS |
| 46 ** |
| 47 ** The file is divided into pages. The first page is called page 1, |
| 48 ** the second is page 2, and so forth. A page number of zero indicates |
| 49 ** "no such page". The page size can be any power of 2 between 512 and 65536. |
| 50 ** Each page can be either a btree page, a freelist page, an overflow |
| 51 ** page, or a pointer-map page. |
| 52 ** |
| 53 ** The first page is always a btree page. The first 100 bytes of the first |
| 54 ** page contain a special header (the "file header") that describes the file. |
| 55 ** The format of the file header is as follows: |
| 56 ** |
| 57 ** OFFSET SIZE DESCRIPTION |
| 58 ** 0 16 Header string: "SQLite format 3\000" |
| 59 ** 16 2 Page size in bytes. |
| 60 ** 18 1 File format write version |
| 61 ** 19 1 File format read version |
| 62 ** 20 1 Bytes of unused space at the end of each page |
| 63 ** 21 1 Max embedded payload fraction |
| 64 ** 22 1 Min embedded payload fraction |
| 65 ** 23 1 Min leaf payload fraction |
| 66 ** 24 4 File change counter |
| 67 ** 28 4 Reserved for future use |
| 68 ** 32 4 First freelist page |
| 69 ** 36 4 Number of freelist pages in the file |
| 70 ** 40 60 15 4-byte meta values passed to higher layers |
| 71 ** |
| 72 ** 40 4 Schema cookie |
| 73 ** 44 4 File format of schema layer |
| 74 ** 48 4 Size of page cache |
| 75 ** 52 4 Largest root-page (auto/incr_vacuum) |
| 76 ** 56 4 1=UTF-8 2=UTF16le 3=UTF16be |
| 77 ** 60 4 User version |
| 78 ** 64 4 Incremental vacuum mode |
| 79 ** 68 4 unused |
| 80 ** 72 4 unused |
| 81 ** 76 4 unused |
| 82 ** |
| 83 ** All of the integer values are big-endian (most significant byte first). |
| 84 ** |
| 85 ** The file change counter is incremented when the database is changed |
| 86 ** This counter allows other processes to know when the file has changed |
| 87 ** and thus when they need to flush their cache. |
| 88 ** |
| 89 ** The max embedded payload fraction is the amount of the total usable |
| 90 ** space in a page that can be consumed by a single cell for standard |
| 91 ** B-tree (non-LEAFDATA) tables. A value of 255 means 100%. The default |
| 92 ** is to limit the maximum cell size so that at least 4 cells will fit |
| 93 ** on one page. Thus the default max embedded payload fraction is 64. |
| 94 ** |
| 95 ** If the payload for a cell is larger than the max payload, then extra |
| 96 ** payload is spilled to overflow pages. Once an overflow page is allocated, |
| 97 ** as many bytes as possible are moved into the overflow pages without letting |
| 98 ** the cell size drop below the min embedded payload fraction. |
| 99 ** |
| 100 ** The min leaf payload fraction is like the min embedded payload fraction |
| 101 ** except that it applies to leaf nodes in a LEAFDATA tree. The maximum |
| 102 ** payload fraction for a LEAFDATA tree is always 100% (or 255) and it |
| 103 ** not specified in the header. |
| 104 ** |
| 105 ** Each btree pages is divided into three sections: The header, the |
| 106 ** cell pointer array, and the cell content area. Page 1 also has a 100-byte |
| 107 ** file header that occurs before the page header. |
| 108 ** |
| 109 ** |----------------| |
| 110 ** | file header | 100 bytes. Page 1 only. |
| 111 ** |----------------| |
| 112 ** | page header | 8 bytes for leaves. 12 bytes for interior nodes |
| 113 ** |----------------| |
| 114 ** | cell pointer | | 2 bytes per cell. Sorted order. |
| 115 ** | array | | Grows downward |
| 116 ** | | v |
| 117 ** |----------------| |
| 118 ** | unallocated | |
| 119 ** | space | |
| 120 ** |----------------| ^ Grows upwards |
| 121 ** | cell content | | Arbitrary order interspersed with freeblocks. |
| 122 ** | area | | and free space fragments. |
| 123 ** |----------------| |
| 124 ** |
| 125 ** The page headers looks like this: |
| 126 ** |
| 127 ** OFFSET SIZE DESCRIPTION |
| 128 ** 0 1 Flags. 1: intkey, 2: zerodata, 4: leafdata, 8: leaf |
| 129 ** 1 2 byte offset to the first freeblock |
| 130 ** 3 2 number of cells on this page |
| 131 ** 5 2 first byte of the cell content area |
| 132 ** 7 1 number of fragmented free bytes |
| 133 ** 8 4 Right child (the Ptr(N) value). Omitted on leaves. |
| 134 ** |
| 135 ** The flags define the format of this btree page. The leaf flag means that |
| 136 ** this page has no children. The zerodata flag means that this page carries |
| 137 ** only keys and no data. The intkey flag means that the key is a integer |
| 138 ** which is stored in the key size entry of the cell header rather than in |
| 139 ** the payload area. |
| 140 ** |
| 141 ** The cell pointer array begins on the first byte after the page header. |
| 142 ** The cell pointer array contains zero or more 2-byte numbers which are |
| 143 ** offsets from the beginning of the page to the cell content in the cell |
| 144 ** content area. The cell pointers occur in sorted order. The system strives |
| 145 ** to keep free space after the last cell pointer so that new cells can |
| 146 ** be easily added without having to defragment the page. |
| 147 ** |
| 148 ** Cell content is stored at the very end of the page and grows toward the |
| 149 ** beginning of the page. |
| 150 ** |
| 151 ** Unused space within the cell content area is collected into a linked list of |
| 152 ** freeblocks. Each freeblock is at least 4 bytes in size. The byte offset |
| 153 ** to the first freeblock is given in the header. Freeblocks occur in |
| 154 ** increasing order. Because a freeblock must be at least 4 bytes in size, |
| 155 ** any group of 3 or fewer unused bytes in the cell content area cannot |
| 156 ** exist on the freeblock chain. A group of 3 or fewer free bytes is called |
| 157 ** a fragment. The total number of bytes in all fragments is recorded. |
| 158 ** in the page header at offset 7. |
| 159 ** |
| 160 ** SIZE DESCRIPTION |
| 161 ** 2 Byte offset of the next freeblock |
| 162 ** 2 Bytes in this freeblock |
| 163 ** |
| 164 ** Cells are of variable length. Cells are stored in the cell content area at |
| 165 ** the end of the page. Pointers to the cells are in the cell pointer array |
| 166 ** that immediately follows the page header. Cells is not necessarily |
| 167 ** contiguous or in order, but cell pointers are contiguous and in order. |
| 168 ** |
| 169 ** Cell content makes use of variable length integers. A variable |
| 170 ** length integer is 1 to 9 bytes where the lower 7 bits of each |
| 171 ** byte are used. The integer consists of all bytes that have bit 8 set and |
| 172 ** the first byte with bit 8 clear. The most significant byte of the integer |
| 173 ** appears first. A variable-length integer may not be more than 9 bytes long. |
| 174 ** As a special case, all 8 bytes of the 9th byte are used as data. This |
| 175 ** allows a 64-bit integer to be encoded in 9 bytes. |
| 176 ** |
| 177 ** 0x00 becomes 0x00000000 |
| 178 ** 0x7f becomes 0x0000007f |
| 179 ** 0x81 0x00 becomes 0x00000080 |
| 180 ** 0x82 0x00 becomes 0x00000100 |
| 181 ** 0x80 0x7f becomes 0x0000007f |
| 182 ** 0x8a 0x91 0xd1 0xac 0x78 becomes 0x12345678 |
| 183 ** 0x81 0x81 0x81 0x81 0x01 becomes 0x10204081 |
| 184 ** |
| 185 ** Variable length integers are used for rowids and to hold the number of |
| 186 ** bytes of key and data in a btree cell. |
| 187 ** |
| 188 ** The content of a cell looks like this: |
| 189 ** |
| 190 ** SIZE DESCRIPTION |
| 191 ** 4 Page number of the left child. Omitted if leaf flag is set. |
| 192 ** var Number of bytes of data. Omitted if the zerodata flag is set. |
| 193 ** var Number of bytes of key. Or the key itself if intkey flag is set. |
| 194 ** * Payload |
| 195 ** 4 First page of the overflow chain. Omitted if no overflow |
| 196 ** |
| 197 ** Overflow pages form a linked list. Each page except the last is completely |
| 198 ** filled with data (pagesize - 4 bytes). The last page can have as little |
| 199 ** as 1 byte of data. |
| 200 ** |
| 201 ** SIZE DESCRIPTION |
| 202 ** 4 Page number of next overflow page |
| 203 ** * Data |
| 204 ** |
| 205 ** Freelist pages come in two subtypes: trunk pages and leaf pages. The |
| 206 ** file header points to the first in a linked list of trunk page. Each trunk |
| 207 ** page points to multiple leaf pages. The content of a leaf page is |
| 208 ** unspecified. A trunk page looks like this: |
| 209 ** |
| 210 ** SIZE DESCRIPTION |
| 211 ** 4 Page number of next trunk page |
| 212 ** 4 Number of leaf pointers on this page |
| 213 ** * zero or more pages numbers of leaves |
| 214 */ |
| 215 #include "sqliteInt.h" |
| 216 |
| 217 |
| 218 /* The following value is the maximum cell size assuming a maximum page |
| 219 ** size give above. |
| 220 */ |
| 221 #define MX_CELL_SIZE(pBt) ((int)(pBt->pageSize-8)) |
| 222 |
| 223 /* The maximum number of cells on a single page of the database. This |
| 224 ** assumes a minimum cell size of 6 bytes (4 bytes for the cell itself |
| 225 ** plus 2 bytes for the index to the cell in the page header). Such |
| 226 ** small cells will be rare, but they are possible. |
| 227 */ |
| 228 #define MX_CELL(pBt) ((pBt->pageSize-8)/6) |
| 229 |
| 230 /* Forward declarations */ |
| 231 typedef struct MemPage MemPage; |
| 232 typedef struct BtLock BtLock; |
| 233 |
| 234 /* |
| 235 ** This is a magic string that appears at the beginning of every |
| 236 ** SQLite database in order to identify the file as a real database. |
| 237 ** |
| 238 ** You can change this value at compile-time by specifying a |
| 239 ** -DSQLITE_FILE_HEADER="..." on the compiler command-line. The |
| 240 ** header must be exactly 16 bytes including the zero-terminator so |
| 241 ** the string itself should be 15 characters long. If you change |
| 242 ** the header, then your custom library will not be able to read |
| 243 ** databases generated by the standard tools and the standard tools |
| 244 ** will not be able to read databases created by your custom library. |
| 245 */ |
| 246 #ifndef SQLITE_FILE_HEADER /* 123456789 123456 */ |
| 247 # define SQLITE_FILE_HEADER "SQLite format 3" |
| 248 #endif |
| 249 |
| 250 /* |
| 251 ** Page type flags. An ORed combination of these flags appear as the |
| 252 ** first byte of on-disk image of every BTree page. |
| 253 */ |
| 254 #define PTF_INTKEY 0x01 |
| 255 #define PTF_ZERODATA 0x02 |
| 256 #define PTF_LEAFDATA 0x04 |
| 257 #define PTF_LEAF 0x08 |
| 258 |
| 259 /* |
| 260 ** As each page of the file is loaded into memory, an instance of the following |
| 261 ** structure is appended and initialized to zero. This structure stores |
| 262 ** information about the page that is decoded from the raw file page. |
| 263 ** |
| 264 ** The pParent field points back to the parent page. This allows us to |
| 265 ** walk up the BTree from any leaf to the root. Care must be taken to |
| 266 ** unref() the parent page pointer when this page is no longer referenced. |
| 267 ** The pageDestructor() routine handles that chore. |
| 268 ** |
| 269 ** Access to all fields of this structure is controlled by the mutex |
| 270 ** stored in MemPage.pBt->mutex. |
| 271 */ |
| 272 struct MemPage { |
| 273 u8 isInit; /* True if previously initialized. MUST BE FIRST! */ |
| 274 u8 nOverflow; /* Number of overflow cell bodies in aCell[] */ |
| 275 u8 intKey; /* True if intkey flag is set */ |
| 276 u8 leaf; /* True if leaf flag is set */ |
| 277 u8 hasData; /* True if this page stores data */ |
| 278 u8 hdrOffset; /* 100 for page 1. 0 otherwise */ |
| 279 u8 childPtrSize; /* 0 if leaf==1. 4 if leaf==0 */ |
| 280 u16 maxLocal; /* Copy of BtShared.maxLocal or BtShared.maxLeaf */ |
| 281 u16 minLocal; /* Copy of BtShared.minLocal or BtShared.minLeaf */ |
| 282 u16 cellOffset; /* Index in aData of first cell pointer */ |
| 283 u16 nFree; /* Number of free bytes on the page */ |
| 284 u16 nCell; /* Number of cells on this page, local and ovfl */ |
| 285 u16 maskPage; /* Mask for page offset */ |
| 286 struct _OvflCell { /* Cells that will not fit on aData[] */ |
| 287 u8 *pCell; /* Pointers to the body of the overflow cell */ |
| 288 u16 idx; /* Insert this cell before idx-th non-overflow cell */ |
| 289 } aOvfl[5]; |
| 290 BtShared *pBt; /* Pointer to BtShared that this page is part of */ |
| 291 u8 *aData; /* Pointer to disk image of the page data */ |
| 292 DbPage *pDbPage; /* Pager page handle */ |
| 293 Pgno pgno; /* Page number for this page */ |
| 294 }; |
| 295 |
| 296 /* |
| 297 ** The in-memory image of a disk page has the auxiliary information appended |
| 298 ** to the end. EXTRA_SIZE is the number of bytes of space needed to hold |
| 299 ** that extra information. |
| 300 */ |
| 301 #define EXTRA_SIZE sizeof(MemPage) |
| 302 |
| 303 /* |
| 304 ** A linked list of the following structures is stored at BtShared.pLock. |
| 305 ** Locks are added (or upgraded from READ_LOCK to WRITE_LOCK) when a cursor |
| 306 ** is opened on the table with root page BtShared.iTable. Locks are removed |
| 307 ** from this list when a transaction is committed or rolled back, or when |
| 308 ** a btree handle is closed. |
| 309 */ |
| 310 struct BtLock { |
| 311 Btree *pBtree; /* Btree handle holding this lock */ |
| 312 Pgno iTable; /* Root page of table */ |
| 313 u8 eLock; /* READ_LOCK or WRITE_LOCK */ |
| 314 BtLock *pNext; /* Next in BtShared.pLock list */ |
| 315 }; |
| 316 |
| 317 /* Candidate values for BtLock.eLock */ |
| 318 #define READ_LOCK 1 |
| 319 #define WRITE_LOCK 2 |
| 320 |
| 321 /* A Btree handle |
| 322 ** |
| 323 ** A database connection contains a pointer to an instance of |
| 324 ** this object for every database file that it has open. This structure |
| 325 ** is opaque to the database connection. The database connection cannot |
| 326 ** see the internals of this structure and only deals with pointers to |
| 327 ** this structure. |
| 328 ** |
| 329 ** For some database files, the same underlying database cache might be |
| 330 ** shared between multiple connections. In that case, each connection |
| 331 ** has it own instance of this object. But each instance of this object |
| 332 ** points to the same BtShared object. The database cache and the |
| 333 ** schema associated with the database file are all contained within |
| 334 ** the BtShared object. |
| 335 ** |
| 336 ** All fields in this structure are accessed under sqlite3.mutex. |
| 337 ** The pBt pointer itself may not be changed while there exists cursors |
| 338 ** in the referenced BtShared that point back to this Btree since those |
| 339 ** cursors have to go through this Btree to find their BtShared and |
| 340 ** they often do so without holding sqlite3.mutex. |
| 341 */ |
| 342 struct Btree { |
| 343 sqlite3 *db; /* The database connection holding this btree */ |
| 344 BtShared *pBt; /* Sharable content of this btree */ |
| 345 u8 inTrans; /* TRANS_NONE, TRANS_READ or TRANS_WRITE */ |
| 346 u8 sharable; /* True if we can share pBt with another db */ |
| 347 u8 locked; /* True if db currently has pBt locked */ |
| 348 int wantToLock; /* Number of nested calls to sqlite3BtreeEnter() */ |
| 349 int nBackup; /* Number of backup operations reading this btree */ |
| 350 Btree *pNext; /* List of other sharable Btrees from the same db */ |
| 351 Btree *pPrev; /* Back pointer of the same list */ |
| 352 #ifndef SQLITE_OMIT_SHARED_CACHE |
| 353 BtLock lock; /* Object used to lock page 1 */ |
| 354 #endif |
| 355 }; |
| 356 |
| 357 /* |
| 358 ** Btree.inTrans may take one of the following values. |
| 359 ** |
| 360 ** If the shared-data extension is enabled, there may be multiple users |
| 361 ** of the Btree structure. At most one of these may open a write transaction, |
| 362 ** but any number may have active read transactions. |
| 363 */ |
| 364 #define TRANS_NONE 0 |
| 365 #define TRANS_READ 1 |
| 366 #define TRANS_WRITE 2 |
| 367 |
| 368 /* |
| 369 ** An instance of this object represents a single database file. |
| 370 ** |
| 371 ** A single database file can be in use as the same time by two |
| 372 ** or more database connections. When two or more connections are |
| 373 ** sharing the same database file, each connection has it own |
| 374 ** private Btree object for the file and each of those Btrees points |
| 375 ** to this one BtShared object. BtShared.nRef is the number of |
| 376 ** connections currently sharing this database file. |
| 377 ** |
| 378 ** Fields in this structure are accessed under the BtShared.mutex |
| 379 ** mutex, except for nRef and pNext which are accessed under the |
| 380 ** global SQLITE_MUTEX_STATIC_MASTER mutex. The pPager field |
| 381 ** may not be modified once it is initially set as long as nRef>0. |
| 382 ** The pSchema field may be set once under BtShared.mutex and |
| 383 ** thereafter is unchanged as long as nRef>0. |
| 384 ** |
| 385 ** isPending: |
| 386 ** |
| 387 ** If a BtShared client fails to obtain a write-lock on a database |
| 388 ** table (because there exists one or more read-locks on the table), |
| 389 ** the shared-cache enters 'pending-lock' state and isPending is |
| 390 ** set to true. |
| 391 ** |
| 392 ** The shared-cache leaves the 'pending lock' state when either of |
| 393 ** the following occur: |
| 394 ** |
| 395 ** 1) The current writer (BtShared.pWriter) concludes its transaction, OR |
| 396 ** 2) The number of locks held by other connections drops to zero. |
| 397 ** |
| 398 ** while in the 'pending-lock' state, no connection may start a new |
| 399 ** transaction. |
| 400 ** |
| 401 ** This feature is included to help prevent writer-starvation. |
| 402 */ |
| 403 struct BtShared { |
| 404 Pager *pPager; /* The page cache */ |
| 405 sqlite3 *db; /* Database connection currently using this Btree */ |
| 406 BtCursor *pCursor; /* A list of all open cursors */ |
| 407 MemPage *pPage1; /* First page of the database */ |
| 408 u8 readOnly; /* True if the underlying file is readonly */ |
| 409 u8 pageSizeFixed; /* True if the page size can no longer be changed */ |
| 410 u8 secureDelete; /* True if secure_delete is enabled */ |
| 411 u8 initiallyEmpty; /* Database is empty at start of transaction */ |
| 412 u8 openFlags; /* Flags to sqlite3BtreeOpen() */ |
| 413 #ifndef SQLITE_OMIT_AUTOVACUUM |
| 414 u8 autoVacuum; /* True if auto-vacuum is enabled */ |
| 415 u8 incrVacuum; /* True if incr-vacuum is enabled */ |
| 416 #endif |
| 417 u8 inTransaction; /* Transaction state */ |
| 418 u8 doNotUseWAL; /* If true, do not open write-ahead-log file */ |
| 419 u16 maxLocal; /* Maximum local payload in non-LEAFDATA tables */ |
| 420 u16 minLocal; /* Minimum local payload in non-LEAFDATA tables */ |
| 421 u16 maxLeaf; /* Maximum local payload in a LEAFDATA table */ |
| 422 u16 minLeaf; /* Minimum local payload in a LEAFDATA table */ |
| 423 u32 pageSize; /* Total number of bytes on a page */ |
| 424 u32 usableSize; /* Number of usable bytes on each page */ |
| 425 int nTransaction; /* Number of open transactions (read + write) */ |
| 426 u32 nPage; /* Number of pages in the database */ |
| 427 void *pSchema; /* Pointer to space allocated by sqlite3BtreeSchema() */ |
| 428 void (*xFreeSchema)(void*); /* Destructor for BtShared.pSchema */ |
| 429 sqlite3_mutex *mutex; /* Non-recursive mutex required to access this object */ |
| 430 Bitvec *pHasContent; /* Set of pages moved to free-list this transaction */ |
| 431 #ifndef SQLITE_OMIT_SHARED_CACHE |
| 432 int nRef; /* Number of references to this structure */ |
| 433 BtShared *pNext; /* Next on a list of sharable BtShared structs */ |
| 434 BtLock *pLock; /* List of locks held on this shared-btree struct */ |
| 435 Btree *pWriter; /* Btree with currently open write transaction */ |
| 436 u8 isExclusive; /* True if pWriter has an EXCLUSIVE lock on the db */ |
| 437 u8 isPending; /* If waiting for read-locks to clear */ |
| 438 #endif |
| 439 u8 *pTmpSpace; /* BtShared.pageSize bytes of space for tmp use */ |
| 440 }; |
| 441 |
| 442 /* |
| 443 ** An instance of the following structure is used to hold information |
| 444 ** about a cell. The parseCellPtr() function fills in this structure |
| 445 ** based on information extract from the raw disk page. |
| 446 */ |
| 447 typedef struct CellInfo CellInfo; |
| 448 struct CellInfo { |
| 449 i64 nKey; /* The key for INTKEY tables, or number of bytes in key */ |
| 450 u8 *pCell; /* Pointer to the start of cell content */ |
| 451 u32 nData; /* Number of bytes of data */ |
| 452 u32 nPayload; /* Total amount of payload */ |
| 453 u16 nHeader; /* Size of the cell content header in bytes */ |
| 454 u16 nLocal; /* Amount of payload held locally */ |
| 455 u16 iOverflow; /* Offset to overflow page number. Zero if no overflow */ |
| 456 u16 nSize; /* Size of the cell content on the main b-tree page */ |
| 457 }; |
| 458 |
| 459 /* |
| 460 ** Maximum depth of an SQLite B-Tree structure. Any B-Tree deeper than |
| 461 ** this will be declared corrupt. This value is calculated based on a |
| 462 ** maximum database size of 2^31 pages a minimum fanout of 2 for a |
| 463 ** root-node and 3 for all other internal nodes. |
| 464 ** |
| 465 ** If a tree that appears to be taller than this is encountered, it is |
| 466 ** assumed that the database is corrupt. |
| 467 */ |
| 468 #define BTCURSOR_MAX_DEPTH 20 |
| 469 |
| 470 /* |
| 471 ** A cursor is a pointer to a particular entry within a particular |
| 472 ** b-tree within a database file. |
| 473 ** |
| 474 ** The entry is identified by its MemPage and the index in |
| 475 ** MemPage.aCell[] of the entry. |
| 476 ** |
| 477 ** A single database file can shared by two more database connections, |
| 478 ** but cursors cannot be shared. Each cursor is associated with a |
| 479 ** particular database connection identified BtCursor.pBtree.db. |
| 480 ** |
| 481 ** Fields in this structure are accessed under the BtShared.mutex |
| 482 ** found at self->pBt->mutex. |
| 483 */ |
| 484 struct BtCursor { |
| 485 Btree *pBtree; /* The Btree to which this cursor belongs */ |
| 486 BtShared *pBt; /* The BtShared this cursor points to */ |
| 487 BtCursor *pNext, *pPrev; /* Forms a linked list of all cursors */ |
| 488 struct KeyInfo *pKeyInfo; /* Argument passed to comparison function */ |
| 489 Pgno pgnoRoot; /* The root page of this tree */ |
| 490 sqlite3_int64 cachedRowid; /* Next rowid cache. 0 means not valid */ |
| 491 CellInfo info; /* A parse of the cell we are pointing at */ |
| 492 i64 nKey; /* Size of pKey, or last integer key */ |
| 493 void *pKey; /* Saved key that was cursor's last known position */ |
| 494 int skipNext; /* Prev() is noop if negative. Next() is noop if positive */ |
| 495 u8 wrFlag; /* True if writable */ |
| 496 u8 atLast; /* Cursor pointing to the last entry */ |
| 497 u8 validNKey; /* True if info.nKey is valid */ |
| 498 u8 eState; /* One of the CURSOR_XXX constants (see below) */ |
| 499 #ifndef SQLITE_OMIT_INCRBLOB |
| 500 Pgno *aOverflow; /* Cache of overflow page locations */ |
| 501 u8 isIncrblobHandle; /* True if this cursor is an incr. io handle */ |
| 502 #endif |
| 503 i16 iPage; /* Index of current page in apPage */ |
| 504 u16 aiIdx[BTCURSOR_MAX_DEPTH]; /* Current index in apPage[i] */ |
| 505 MemPage *apPage[BTCURSOR_MAX_DEPTH]; /* Pages from root to current page */ |
| 506 }; |
| 507 |
| 508 /* |
| 509 ** Potential values for BtCursor.eState. |
| 510 ** |
| 511 ** CURSOR_VALID: |
| 512 ** Cursor points to a valid entry. getPayload() etc. may be called. |
| 513 ** |
| 514 ** CURSOR_INVALID: |
| 515 ** Cursor does not point to a valid entry. This can happen (for example) |
| 516 ** because the table is empty or because BtreeCursorFirst() has not been |
| 517 ** called. |
| 518 ** |
| 519 ** CURSOR_REQUIRESEEK: |
| 520 ** The table that this cursor was opened on still exists, but has been |
| 521 ** modified since the cursor was last used. The cursor position is saved |
| 522 ** in variables BtCursor.pKey and BtCursor.nKey. When a cursor is in |
| 523 ** this state, restoreCursorPosition() can be called to attempt to |
| 524 ** seek the cursor to the saved position. |
| 525 ** |
| 526 ** CURSOR_FAULT: |
| 527 ** A unrecoverable error (an I/O error or a malloc failure) has occurred |
| 528 ** on a different connection that shares the BtShared cache with this |
| 529 ** cursor. The error has left the cache in an inconsistent state. |
| 530 ** Do nothing else with this cursor. Any attempt to use the cursor |
| 531 ** should return the error code stored in BtCursor.skip |
| 532 */ |
| 533 #define CURSOR_INVALID 0 |
| 534 #define CURSOR_VALID 1 |
| 535 #define CURSOR_REQUIRESEEK 2 |
| 536 #define CURSOR_FAULT 3 |
| 537 |
| 538 /* |
| 539 ** The database page the PENDING_BYTE occupies. This page is never used. |
| 540 */ |
| 541 # define PENDING_BYTE_PAGE(pBt) PAGER_MJ_PGNO(pBt) |
| 542 |
| 543 /* |
| 544 ** These macros define the location of the pointer-map entry for a |
| 545 ** database page. The first argument to each is the number of usable |
| 546 ** bytes on each page of the database (often 1024). The second is the |
| 547 ** page number to look up in the pointer map. |
| 548 ** |
| 549 ** PTRMAP_PAGENO returns the database page number of the pointer-map |
| 550 ** page that stores the required pointer. PTRMAP_PTROFFSET returns |
| 551 ** the offset of the requested map entry. |
| 552 ** |
| 553 ** If the pgno argument passed to PTRMAP_PAGENO is a pointer-map page, |
| 554 ** then pgno is returned. So (pgno==PTRMAP_PAGENO(pgsz, pgno)) can be |
| 555 ** used to test if pgno is a pointer-map page. PTRMAP_ISPAGE implements |
| 556 ** this test. |
| 557 */ |
| 558 #define PTRMAP_PAGENO(pBt, pgno) ptrmapPageno(pBt, pgno) |
| 559 #define PTRMAP_PTROFFSET(pgptrmap, pgno) (5*(pgno-pgptrmap-1)) |
| 560 #define PTRMAP_ISPAGE(pBt, pgno) (PTRMAP_PAGENO((pBt),(pgno))==(pgno)) |
| 561 |
| 562 /* |
| 563 ** The pointer map is a lookup table that identifies the parent page for |
| 564 ** each child page in the database file. The parent page is the page that |
| 565 ** contains a pointer to the child. Every page in the database contains |
| 566 ** 0 or 1 parent pages. (In this context 'database page' refers |
| 567 ** to any page that is not part of the pointer map itself.) Each pointer map |
| 568 ** entry consists of a single byte 'type' and a 4 byte parent page number. |
| 569 ** The PTRMAP_XXX identifiers below are the valid types. |
| 570 ** |
| 571 ** The purpose of the pointer map is to facility moving pages from one |
| 572 ** position in the file to another as part of autovacuum. When a page |
| 573 ** is moved, the pointer in its parent must be updated to point to the |
| 574 ** new location. The pointer map is used to locate the parent page quickly. |
| 575 ** |
| 576 ** PTRMAP_ROOTPAGE: The database page is a root-page. The page-number is not |
| 577 ** used in this case. |
| 578 ** |
| 579 ** PTRMAP_FREEPAGE: The database page is an unused (free) page. The page-number |
| 580 ** is not used in this case. |
| 581 ** |
| 582 ** PTRMAP_OVERFLOW1: The database page is the first page in a list of |
| 583 ** overflow pages. The page number identifies the page that |
| 584 ** contains the cell with a pointer to this overflow page. |
| 585 ** |
| 586 ** PTRMAP_OVERFLOW2: The database page is the second or later page in a list of |
| 587 ** overflow pages. The page-number identifies the previous |
| 588 ** page in the overflow page list. |
| 589 ** |
| 590 ** PTRMAP_BTREE: The database page is a non-root btree page. The page number |
| 591 ** identifies the parent page in the btree. |
| 592 */ |
| 593 #define PTRMAP_ROOTPAGE 1 |
| 594 #define PTRMAP_FREEPAGE 2 |
| 595 #define PTRMAP_OVERFLOW1 3 |
| 596 #define PTRMAP_OVERFLOW2 4 |
| 597 #define PTRMAP_BTREE 5 |
| 598 |
| 599 /* A bunch of assert() statements to check the transaction state variables |
| 600 ** of handle p (type Btree*) are internally consistent. |
| 601 */ |
| 602 #define btreeIntegrity(p) \ |
| 603 assert( p->pBt->inTransaction!=TRANS_NONE || p->pBt->nTransaction==0 ); \ |
| 604 assert( p->pBt->inTransaction>=p->inTrans ); |
| 605 |
| 606 |
| 607 /* |
| 608 ** The ISAUTOVACUUM macro is used within balance_nonroot() to determine |
| 609 ** if the database supports auto-vacuum or not. Because it is used |
| 610 ** within an expression that is an argument to another macro |
| 611 ** (sqliteMallocRaw), it is not possible to use conditional compilation. |
| 612 ** So, this macro is defined instead. |
| 613 */ |
| 614 #ifndef SQLITE_OMIT_AUTOVACUUM |
| 615 #define ISAUTOVACUUM (pBt->autoVacuum) |
| 616 #else |
| 617 #define ISAUTOVACUUM 0 |
| 618 #endif |
| 619 |
| 620 |
| 621 /* |
| 622 ** This structure is passed around through all the sanity checking routines |
| 623 ** in order to keep track of some global state information. |
| 624 */ |
| 625 typedef struct IntegrityCk IntegrityCk; |
| 626 struct IntegrityCk { |
| 627 BtShared *pBt; /* The tree being checked out */ |
| 628 Pager *pPager; /* The associated pager. Also accessible by pBt->pPager */ |
| 629 Pgno nPage; /* Number of pages in the database */ |
| 630 int *anRef; /* Number of times each page is referenced */ |
| 631 int mxErr; /* Stop accumulating errors when this reaches zero */ |
| 632 int nErr; /* Number of messages written to zErrMsg so far */ |
| 633 int mallocFailed; /* A memory allocation error has occurred */ |
| 634 StrAccum errMsg; /* Accumulate the error message text here */ |
| 635 }; |
| 636 |
| 637 /* |
| 638 ** Read or write a two- and four-byte big-endian integer values. |
| 639 */ |
| 640 #define get2byte(x) ((x)[0]<<8 | (x)[1]) |
| 641 #define put2byte(p,v) ((p)[0] = (u8)((v)>>8), (p)[1] = (u8)(v)) |
| 642 #define get4byte sqlite3Get4byte |
| 643 #define put4byte sqlite3Put4byte |
| OLD | NEW |
|---|
Chromium Code Reviews