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 ** $Id: btreeInt.h,v 1.14 2007/12/07 18:55:28 drh Exp $ 13 ** 14 ** This file implements a external (disk-based) database using BTrees. 15 ** For a detailed discussion of BTrees, refer to 16 ** 17 ** Donald E. Knuth, THE ART OF COMPUTER PROGRAMMING, Volume 3: 18 ** "Sorting And Searching", pages 473-480. Addison-Wesley 19 ** Publishing Company, Reading, Massachusetts. 20 ** 21 ** The basic idea is that each page of the file contains N database 22 ** entries and N+1 pointers to subpages. 23 ** 24 ** ---------------------------------------------------------------- 25 ** | Ptr(0) | Key(0) | Ptr(1) | Key(1) | ... | Key(N-1) | Ptr(N) | 26 ** ---------------------------------------------------------------- 27 ** 28 ** All of the keys on the page that Ptr(0) points to have values less 29 ** than Key(0). All of the keys on page Ptr(1) and its subpages have 30 ** values greater than Key(0) and less than Key(1). All of the keys 31 ** on Ptr(N) and its subpages have values greater than Key(N-1). And 32 ** so forth. 33 ** 34 ** Finding a particular key requires reading O(log(M)) pages from the 35 ** disk where M is the number of entries in the tree. 36 ** 37 ** In this implementation, a single file can hold one or more separate 38 ** BTrees. Each BTree is identified by the index of its root page. The 39 ** key and data for any entry are combined to form the "payload". A 40 ** fixed amount of payload can be carried directly on the database 41 ** page. If the payload is larger than the preset amount then surplus 42 ** bytes are stored on overflow pages. The payload for an entry 43 ** and the preceding pointer are combined to form a "Cell". Each 44 ** page has a small header which contains the Ptr(N) pointer and other 45 ** information such as the size of key and data. 46 ** 47 ** FORMAT DETAILS 48 ** 49 ** The file is divided into pages. The first page is called page 1, 50 ** the second is page 2, and so forth. A page number of zero indicates 51 ** "no such page". The page size can be anything between 512 and 65536. 52 ** Each page can be either a btree page, a freelist page or an overflow 53 ** page. 54 ** 55 ** The first page is always a btree page. The first 100 bytes of the first 56 ** page contain a special header (the "file header") that describes the file. 57 ** The format of the file header is as follows: 58 ** 59 ** OFFSET SIZE DESCRIPTION 60 ** 0 16 Header string: "SQLite format 3\000" 61 ** 16 2 Page size in bytes. 62 ** 18 1 File format write version 63 ** 19 1 File format read version 64 ** 20 1 Bytes of unused space at the end of each page 65 ** 21 1 Max embedded payload fraction 66 ** 22 1 Min embedded payload fraction 67 ** 23 1 Min leaf payload fraction 68 ** 24 4 File change counter 69 ** 28 4 Reserved for future use 70 ** 32 4 First freelist page 71 ** 36 4 Number of freelist pages in the file 72 ** 40 60 15 4-byte meta values passed to higher layers 73 ** 74 ** All of the integer values are big-endian (most significant byte first). 75 ** 76 ** The file change counter is incremented when the database is changed 77 ** This counter allows other processes to know when the file has changed 78 ** and thus when they need to flush their cache. 79 ** 80 ** The max embedded payload fraction is the amount of the total usable 81 ** space in a page that can be consumed by a single cell for standard 82 ** B-tree (non-LEAFDATA) tables. A value of 255 means 100%. The default 83 ** is to limit the maximum cell size so that at least 4 cells will fit 84 ** on one page. Thus the default max embedded payload fraction is 64. 85 ** 86 ** If the payload for a cell is larger than the max payload, then extra 87 ** payload is spilled to overflow pages. Once an overflow page is allocated, 88 ** as many bytes as possible are moved into the overflow pages without letting 89 ** the cell size drop below the min embedded payload fraction. 90 ** 91 ** The min leaf payload fraction is like the min embedded payload fraction 92 ** except that it applies to leaf nodes in a LEAFDATA tree. The maximum 93 ** payload fraction for a LEAFDATA tree is always 100% (or 255) and it 94 ** not specified in the header. 95 ** 96 ** Each btree pages is divided into three sections: The header, the 97 ** cell pointer array, and the cell content area. Page 1 also has a 100-byte 98 ** file header that occurs before the page header. 99 ** 100 ** |----------------| 101 ** | file header | 100 bytes. Page 1 only. 102 ** |----------------| 103 ** | page header | 8 bytes for leaves. 12 bytes for interior nodes 104 ** |----------------| 105 ** | cell pointer | | 2 bytes per cell. Sorted order. 106 ** | array | | Grows downward 107 ** | | v 108 ** |----------------| 109 ** | unallocated | 110 ** | space | 111 ** |----------------| ^ Grows upwards 112 ** | cell content | | Arbitrary order interspersed with freeblocks. 113 ** | area | | and free space fragments. 114 ** |----------------| 115 ** 116 ** The page headers looks like this: 117 ** 118 ** OFFSET SIZE DESCRIPTION 119 ** 0 1 Flags. 1: intkey, 2: zerodata, 4: leafdata, 8: leaf 120 ** 1 2 byte offset to the first freeblock 121 ** 3 2 number of cells on this page 122 ** 5 2 first byte of the cell content area 123 ** 7 1 number of fragmented free bytes 124 ** 8 4 Right child (the Ptr(N) value). Omitted on leaves. 125 ** 126 ** The flags define the format of this btree page. The leaf flag means that 127 ** this page has no children. The zerodata flag means that this page carries 128 ** only keys and no data. The intkey flag means that the key is a integer 129 ** which is stored in the key size entry of the cell header rather than in 130 ** the payload area. 131 ** 132 ** The cell pointer array begins on the first byte after the page header. 133 ** The cell pointer array contains zero or more 2-byte numbers which are 134 ** offsets from the beginning of the page to the cell content in the cell 135 ** content area. The cell pointers occur in sorted order. The system strives 136 ** to keep free space after the last cell pointer so that new cells can 137 ** be easily added without having to defragment the page. 138 ** 139 ** Cell content is stored at the very end of the page and grows toward the 140 ** beginning of the page. 141 ** 142 ** Unused space within the cell content area is collected into a linked list of 143 ** freeblocks. Each freeblock is at least 4 bytes in size. The byte offset 144 ** to the first freeblock is given in the header. Freeblocks occur in 145 ** increasing order. Because a freeblock must be at least 4 bytes in size, 146 ** any group of 3 or fewer unused bytes in the cell content area cannot 147 ** exist on the freeblock chain. A group of 3 or fewer free bytes is called 148 ** a fragment. The total number of bytes in all fragments is recorded. 149 ** in the page header at offset 7. 150 ** 151 ** SIZE DESCRIPTION 152 ** 2 Byte offset of the next freeblock 153 ** 2 Bytes in this freeblock 154 ** 155 ** Cells are of variable length. Cells are stored in the cell content area at 156 ** the end of the page. Pointers to the cells are in the cell pointer array 157 ** that immediately follows the page header. Cells is not necessarily 158 ** contiguous or in order, but cell pointers are contiguous and in order. 159 ** 160 ** Cell content makes use of variable length integers. A variable 161 ** length integer is 1 to 9 bytes where the lower 7 bits of each 162 ** byte are used. The integer consists of all bytes that have bit 8 set and 163 ** the first byte with bit 8 clear. The most significant byte of the integer 164 ** appears first. A variable-length integer may not be more than 9 bytes long. 165 ** As a special case, all 8 bytes of the 9th byte are used as data. This 166 ** allows a 64-bit integer to be encoded in 9 bytes. 167 ** 168 ** 0x00 becomes 0x00000000 169 ** 0x7f becomes 0x0000007f 170 ** 0x81 0x00 becomes 0x00000080 171 ** 0x82 0x00 becomes 0x00000100 172 ** 0x80 0x7f becomes 0x0000007f 173 ** 0x8a 0x91 0xd1 0xac 0x78 becomes 0x12345678 174 ** 0x81 0x81 0x81 0x81 0x01 becomes 0x10204081 175 ** 176 ** Variable length integers are used for rowids and to hold the number of 177 ** bytes of key and data in a btree cell. 178 ** 179 ** The content of a cell looks like this: 180 ** 181 ** SIZE DESCRIPTION 182 ** 4 Page number of the left child. Omitted if leaf flag is set. 183 ** var Number of bytes of data. Omitted if the zerodata flag is set. 184 ** var Number of bytes of key. Or the key itself if intkey flag is set. 185 ** * Payload 186 ** 4 First page of the overflow chain. Omitted if no overflow 187 ** 188 ** Overflow pages form a linked list. Each page except the last is completely 189 ** filled with data (pagesize - 4 bytes). The last page can have as little 190 ** as 1 byte of data. 191 ** 192 ** SIZE DESCRIPTION 193 ** 4 Page number of next overflow page 194 ** * Data 195 ** 196 ** Freelist pages come in two subtypes: trunk pages and leaf pages. The 197 ** file header points to the first in a linked list of trunk page. Each trunk 198 ** page points to multiple leaf pages. The content of a leaf page is 199 ** unspecified. A trunk page looks like this: 200 ** 201 ** SIZE DESCRIPTION 202 ** 4 Page number of next trunk page 203 ** 4 Number of leaf pointers on this page 204 ** * zero or more pages numbers of leaves 205 */ 206 #include "sqliteInt.h" 207 #include "pager.h" 208 #include "btree.h" 209 #include "os.h" 210 #include <assert.h> 211 212 /* Round up a number to the next larger multiple of 8. This is used 213 ** to force 8-byte alignment on 64-bit architectures. 214 */ 215 #define ROUND8(x) ((x+7)&~7) 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) (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 3 bytes. Such small cells will be 225 ** exceedingly rare, but they are possible. 226 */ 227 #define MX_CELL(pBt) ((pBt->pageSize-8)/3) 228 229 /* Forward declarations */ 230 typedef struct MemPage MemPage; 231 typedef struct BtLock BtLock; 232 233 /* 234 ** This is a magic string that appears at the beginning of every 235 ** SQLite database in order to identify the file as a real database. 236 ** 237 ** You can change this value at compile-time by specifying a 238 ** -DSQLITE_FILE_HEADER="..." on the compiler command-line. The 239 ** header must be exactly 16 bytes including the zero-terminator so 240 ** the string itself should be 15 characters long. If you change 241 ** the header, then your custom library will not be able to read 242 ** databases generated by the standard tools and the standard tools 243 ** will not be able to read databases created by your custom library. 244 */ 245 #ifndef SQLITE_FILE_HEADER /* 123456789 123456 */ 246 # define SQLITE_FILE_HEADER "SQLite format 3" 247 #endif 248 249 /* 250 ** Page type flags. An ORed combination of these flags appear as the 251 ** first byte of on-disk image of every BTree page. 252 */ 253 #define PTF_INTKEY 0x01 254 #define PTF_ZERODATA 0x02 255 #define PTF_LEAFDATA 0x04 256 #define PTF_LEAF 0x08 257 258 /* 259 ** As each page of the file is loaded into memory, an instance of the following 260 ** structure is appended and initialized to zero. This structure stores 261 ** information about the page that is decoded from the raw file page. 262 ** 263 ** The pParent field points back to the parent page. This allows us to 264 ** walk up the BTree from any leaf to the root. Care must be taken to 265 ** unref() the parent page pointer when this page is no longer referenced. 266 ** The pageDestructor() routine handles that chore. 267 ** 268 ** Access to all fields of this structure is controlled by the mutex 269 ** stored in MemPage.pBt->mutex. 270 */ 271 struct MemPage { 272 u8 isInit; /* True if previously initialized. MUST BE FIRST! */ 273 u8 idxShift; /* True if Cell indices have changed */ 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 zeroData; /* True if table stores keys only */ 278 u8 leafData; /* True if tables stores data on leaves only */ 279 u8 hasData; /* True if this page stores data */ 280 u8 hdrOffset; /* 100 for page 1. 0 otherwise */ 281 u8 childPtrSize; /* 0 if leaf==1. 4 if leaf==0 */ 282 u16 maxLocal; /* Copy of BtShared.maxLocal or BtShared.maxLeaf */ 283 u16 minLocal; /* Copy of BtShared.minLocal or BtShared.minLeaf */ 284 u16 cellOffset; /* Index in aData of first cell pointer */ 285 u16 idxParent; /* Index in parent of this node */ 286 u16 nFree; /* Number of free bytes on the page */ 287 u16 nCell; /* Number of cells on this page, local and ovfl */ 288 struct _OvflCell { /* Cells that will not fit on aData[] */ 289 u8 *pCell; /* Pointers to the body of the overflow cell */ 290 u16 idx; /* Insert this cell before idx-th non-overflow cell */ 291 } aOvfl[5]; 292 BtShared *pBt; /* Pointer to BtShared that this page is part of */ 293 u8 *aData; /* Pointer to disk image of the page data */ 294 DbPage *pDbPage; /* Pager page handle */ 295 Pgno pgno; /* Page number for this page */ 296 MemPage *pParent; /* The parent of this page. NULL for root */ 297 }; 298 299 /* 300 ** The in-memory image of a disk page has the auxiliary information appended 301 ** to the end. EXTRA_SIZE is the number of bytes of space needed to hold 302 ** that extra information. 303 */ 304 #define EXTRA_SIZE sizeof(MemPage) 305 306 /* A Btree handle 307 ** 308 ** A database connection contains a pointer to an instance of 309 ** this object for every database file that it has open. This structure 310 ** is opaque to the database connection. The database connection cannot 311 ** see the internals of this structure and only deals with pointers to 312 ** this structure. 313 ** 314 ** For some database files, the same underlying database cache might be 315 ** shared between multiple connections. In that case, each contection 316 ** has it own pointer to this object. But each instance of this object 317 ** points to the same BtShared object. The database cache and the 318 ** schema associated with the database file are all contained within 319 ** the BtShared object. 320 ** 321 ** All fields in this structure are accessed under sqlite3.mutex. 322 ** The pBt pointer itself may not be changed while there exists cursors 323 ** in the referenced BtShared that point back to this Btree since those 324 ** cursors have to do go through this Btree to find their BtShared and 325 ** they often do so without holding sqlite3.mutex. 326 */ 327 struct Btree { 328 sqlite3 *db; /* The database connection holding this btree */ 329 BtShared *pBt; /* Sharable content of this btree */ 330 u8 inTrans; /* TRANS_NONE, TRANS_READ or TRANS_WRITE */ 331 u8 sharable; /* True if we can share pBt with another db */ 332 u8 locked; /* True if db currently has pBt locked */ 333 int wantToLock; /* Number of nested calls to sqlite3BtreeEnter() */ 334 Btree *pNext; /* List of other sharable Btrees from the same db */ 335 Btree *pPrev; /* Back pointer of the same list */ 336 }; 337 338 /* 339 ** Btree.inTrans may take one of the following values. 340 ** 341 ** If the shared-data extension is enabled, there may be multiple users 342 ** of the Btree structure. At most one of these may open a write transaction, 343 ** but any number may have active read transactions. 344 */ 345 #define TRANS_NONE 0 346 #define TRANS_READ 1 347 #define TRANS_WRITE 2 348 349 /* 350 ** An instance of this object represents a single database file. 351 ** 352 ** A single database file can be in use as the same time by two 353 ** or more database connections. When two or more connections are 354 ** sharing the same database file, each connection has it own 355 ** private Btree object for the file and each of those Btrees points 356 ** to this one BtShared object. BtShared.nRef is the number of 357 ** connections currently sharing this database file. 358 ** 359 ** Fields in this structure are accessed under the BtShared.mutex 360 ** mutex, except for nRef and pNext which are accessed under the 361 ** global SQLITE_MUTEX_STATIC_MASTER mutex. The pPager field 362 ** may not be modified once it is initially set as long as nRef>0. 363 ** The pSchema field may be set once under BtShared.mutex and 364 ** thereafter is unchanged as long as nRef>0. 365 */ 366 struct BtShared { 367 Pager *pPager; /* The page cache */ 368 sqlite3 *db; /* Database connection currently using this Btree */ 369 BtCursor *pCursor; /* A list of all open cursors */ 370 MemPage *pPage1; /* First page of the database */ 371 u8 inStmt; /* True if we are in a statement subtransaction */ 372 u8 readOnly; /* True if the underlying file is readonly */ 373 u8 maxEmbedFrac; /* Maximum payload as % of total page size */ 374 u8 minEmbedFrac; /* Minimum payload as % of total page size */ 375 u8 minLeafFrac; /* Minimum leaf payload as % of total page size */ 376 u8 pageSizeFixed; /* True if the page size can no longer be changed */ 377 #ifndef SQLITE_OMIT_AUTOVACUUM 378 u8 autoVacuum; /* True if auto-vacuum is enabled */ 379 u8 incrVacuum; /* True if incr-vacuum is enabled */ 380 Pgno nTrunc; /* Non-zero if the db will be truncated (incr vacuum) */ 381 #endif 382 u16 pageSize; /* Total number of bytes on a page */ 383 u16 usableSize; /* Number of usable bytes on each page */ 384 int maxLocal; /* Maximum local payload in non-LEAFDATA tables */ 385 int minLocal; /* Minimum local payload in non-LEAFDATA tables */ 386 int maxLeaf; /* Maximum local payload in a LEAFDATA table */ 387 int minLeaf; /* Minimum local payload in a LEAFDATA table */ 388 u8 inTransaction; /* Transaction state */ 389 int nTransaction; /* Number of open transactions (read + write) */ 390 void *pSchema; /* Pointer to space allocated by sqlite3BtreeSchema() */ 391 void (*xFreeSchema)(void*); /* Destructor for BtShared.pSchema */ 392 sqlite3_mutex *mutex; /* Non-recursive mutex required to access this struct */ 393 BusyHandler busyHdr; /* The busy handler for this btree */ 394 #ifndef SQLITE_OMIT_SHARED_CACHE 395 int nRef; /* Number of references to this structure */ 396 BtShared *pNext; /* Next on a list of sharable BtShared structs */ 397 BtLock *pLock; /* List of locks held on this shared-btree struct */ 398 #endif 399 }; 400 401 /* 402 ** An instance of the following structure is used to hold information 403 ** about a cell. The parseCellPtr() function fills in this structure 404 ** based on information extract from the raw disk page. 405 */ 406 typedef struct CellInfo CellInfo; 407 struct CellInfo { 408 u8 *pCell; /* Pointer to the start of cell content */ 409 i64 nKey; /* The key for INTKEY tables, or number of bytes in key */ 410 u32 nData; /* Number of bytes of data */ 411 u32 nPayload; /* Total amount of payload */ 412 u16 nHeader; /* Size of the cell content header in bytes */ 413 u16 nLocal; /* Amount of payload held locally */ 414 u16 iOverflow; /* Offset to overflow page number. Zero if no overflow */ 415 u16 nSize; /* Size of the cell content on the main b-tree page */ 416 }; 417 418 /* 419 ** A cursor is a pointer to a particular entry within a particular 420 ** b-tree within a database file. 421 ** 422 ** The entry is identified by its MemPage and the index in 423 ** MemPage.aCell[] of the entry. 424 ** 425 ** When a single database file can shared by two more database connections, 426 ** but cursors cannot be shared. Each cursor is associated with a 427 ** particular database connection identified BtCursor.pBtree.db. 428 ** 429 ** Fields in this structure are accessed under the BtShared.mutex 430 ** found at self->pBt->mutex. 431 */ 432 struct BtCursor { 433 Btree *pBtree; /* The Btree to which this cursor belongs */ 434 BtShared *pBt; /* The BtShared this cursor points to */ 435 BtCursor *pNext, *pPrev; /* Forms a linked list of all cursors */ 436 int (*xCompare)(void*,int,const void*,int,const void*); /* Key comp func */ 437 void *pArg; /* First arg to xCompare() */ 438 Pgno pgnoRoot; /* The root page of this tree */ 439 MemPage *pPage; /* Page that contains the entry */ 440 int idx; /* Index of the entry in pPage->aCell[] */ 441 CellInfo info; /* A parse of the cell we are pointing at */ 442 u8 wrFlag; /* True if writable */ 443 u8 eState; /* One of the CURSOR_XXX constants (see below) */ 444 void *pKey; /* Saved key that was cursor's last known position */ 445 i64 nKey; /* Size of pKey, or last integer key */ 446 int skip; /* (skip<0) -> Prev() is a no-op. (skip>0) -> Next() is */ 447 #ifndef SQLITE_OMIT_INCRBLOB 448 u8 isIncrblobHandle; /* True if this cursor is an incr. io handle */ 449 Pgno *aOverflow; /* Cache of overflow page locations */ 450 #endif 451 }; 452 453 /* 454 ** Potential values for BtCursor.eState. 455 ** 456 ** CURSOR_VALID: 457 ** Cursor points to a valid entry. getPayload() etc. may be called. 458 ** 459 ** CURSOR_INVALID: 460 ** Cursor does not point to a valid entry. This can happen (for example) 461 ** because the table is empty or because BtreeCursorFirst() has not been 462 ** called. 463 ** 464 ** CURSOR_REQUIRESEEK: 465 ** The table that this cursor was opened on still exists, but has been 466 ** modified since the cursor was last used. The cursor position is saved 467 ** in variables BtCursor.pKey and BtCursor.nKey. When a cursor is in 468 ** this state, restoreOrClearCursorPosition() can be called to attempt to 469 ** seek the cursor to the saved position. 470 ** 471 ** CURSOR_FAULT: 472 ** A unrecoverable error (an I/O error or a malloc failure) has occurred 473 ** on a different connection that shares the BtShared cache with this 474 ** cursor. The error has left the cache in an inconsistent state. 475 ** Do nothing else with this cursor. Any attempt to use the cursor 476 ** should return the error code stored in BtCursor.skip 477 */ 478 #define CURSOR_INVALID 0 479 #define CURSOR_VALID 1 480 #define CURSOR_REQUIRESEEK 2 481 #define CURSOR_FAULT 3 482 483 /* 484 ** The TRACE macro will print high-level status information about the 485 ** btree operation when the global variable sqlite3_btree_trace is 486 ** enabled. 487 */ 488 #if SQLITE_TEST 489 # define TRACE(X) if( sqlite3_btree_trace ){ printf X; fflush(stdout); } 490 #else 491 # define TRACE(X) 492 #endif 493 494 /* 495 ** Routines to read and write variable-length integers. These used to 496 ** be defined locally, but now we use the varint routines in the util.c 497 ** file. 498 */ 499 #define getVarint sqlite3GetVarint 500 #define getVarint32(A,B) ((*B=*(A))<=0x7f?1:sqlite3GetVarint32(A,B)) 501 #define putVarint sqlite3PutVarint 502 503 /* The database page the PENDING_BYTE occupies. This page is never used. 504 ** TODO: This macro is very similary to PAGER_MJ_PGNO() in pager.c. They 505 ** should possibly be consolidated (presumably in pager.h). 506 ** 507 ** If disk I/O is omitted (meaning that the database is stored purely 508 ** in memory) then there is no pending byte. 509 */ 510 #ifdef SQLITE_OMIT_DISKIO 511 # define PENDING_BYTE_PAGE(pBt) 0x7fffffff 512 #else 513 # define PENDING_BYTE_PAGE(pBt) ((PENDING_BYTE/(pBt)->pageSize)+1) 514 #endif 515 516 /* 517 ** A linked list of the following structures is stored at BtShared.pLock. 518 ** Locks are added (or upgraded from READ_LOCK to WRITE_LOCK) when a cursor 519 ** is opened on the table with root page BtShared.iTable. Locks are removed 520 ** from this list when a transaction is committed or rolled back, or when 521 ** a btree handle is closed. 522 */ 523 struct BtLock { 524 Btree *pBtree; /* Btree handle holding this lock */ 525 Pgno iTable; /* Root page of table */ 526 u8 eLock; /* READ_LOCK or WRITE_LOCK */ 527 BtLock *pNext; /* Next in BtShared.pLock list */ 528 }; 529 530 /* Candidate values for BtLock.eLock */ 531 #define READ_LOCK 1 532 #define WRITE_LOCK 2 533 534 /* 535 ** These macros define the location of the pointer-map entry for a 536 ** database page. The first argument to each is the number of usable 537 ** bytes on each page of the database (often 1024). The second is the 538 ** page number to look up in the pointer map. 539 ** 540 ** PTRMAP_PAGENO returns the database page number of the pointer-map 541 ** page that stores the required pointer. PTRMAP_PTROFFSET returns 542 ** the offset of the requested map entry. 543 ** 544 ** If the pgno argument passed to PTRMAP_PAGENO is a pointer-map page, 545 ** then pgno is returned. So (pgno==PTRMAP_PAGENO(pgsz, pgno)) can be 546 ** used to test if pgno is a pointer-map page. PTRMAP_ISPAGE implements 547 ** this test. 548 */ 549 #define PTRMAP_PAGENO(pBt, pgno) ptrmapPageno(pBt, pgno) 550 #define PTRMAP_PTROFFSET(pBt, pgno) (5*(pgno-ptrmapPageno(pBt, pgno)-1)) 551 #define PTRMAP_ISPAGE(pBt, pgno) (PTRMAP_PAGENO((pBt),(pgno))==(pgno)) 552 553 /* 554 ** The pointer map is a lookup table that identifies the parent page for 555 ** each child page in the database file. The parent page is the page that 556 ** contains a pointer to the child. Every page in the database contains 557 ** 0 or 1 parent pages. (In this context 'database page' refers 558 ** to any page that is not part of the pointer map itself.) Each pointer map 559 ** entry consists of a single byte 'type' and a 4 byte parent page number. 560 ** The PTRMAP_XXX identifiers below are the valid types. 561 ** 562 ** The purpose of the pointer map is to facility moving pages from one 563 ** position in the file to another as part of autovacuum. When a page 564 ** is moved, the pointer in its parent must be updated to point to the 565 ** new location. The pointer map is used to locate the parent page quickly. 566 ** 567 ** PTRMAP_ROOTPAGE: The database page is a root-page. The page-number is not 568 ** used in this case. 569 ** 570 ** PTRMAP_FREEPAGE: The database page is an unused (free) page. The page-number 571 ** is not used in this case. 572 ** 573 ** PTRMAP_OVERFLOW1: The database page is the first page in a list of 574 ** overflow pages. The page number identifies the page that 575 ** contains the cell with a pointer to this overflow page. 576 ** 577 ** PTRMAP_OVERFLOW2: The database page is the second or later page in a list of 578 ** overflow pages. The page-number identifies the previous 579 ** page in the overflow page list. 580 ** 581 ** PTRMAP_BTREE: The database page is a non-root btree page. The page number 582 ** identifies the parent page in the btree. 583 */ 584 #define PTRMAP_ROOTPAGE 1 585 #define PTRMAP_FREEPAGE 2 586 #define PTRMAP_OVERFLOW1 3 587 #define PTRMAP_OVERFLOW2 4 588 #define PTRMAP_BTREE 5 589 590 /* A bunch of assert() statements to check the transaction state variables 591 ** of handle p (type Btree*) are internally consistent. 592 */ 593 #define btreeIntegrity(p) \ 594 assert( p->pBt->inTransaction!=TRANS_NONE || p->pBt->nTransaction==0 ); \ 595 assert( p->pBt->inTransaction>=p->inTrans ); 596 597 598 /* 599 ** The ISAUTOVACUUM macro is used within balance_nonroot() to determine 600 ** if the database supports auto-vacuum or not. Because it is used 601 ** within an expression that is an argument to another macro 602 ** (sqliteMallocRaw), it is not possible to use conditional compilation. 603 ** So, this macro is defined instead. 604 */ 605 #ifndef SQLITE_OMIT_AUTOVACUUM 606 #define ISAUTOVACUUM (pBt->autoVacuum) 607 #else 608 #define ISAUTOVACUUM 0 609 #endif 610 611 612 /* 613 ** This structure is passed around through all the sanity checking routines 614 ** in order to keep track of some global state information. 615 */ 616 typedef struct IntegrityCk IntegrityCk; 617 struct IntegrityCk { 618 BtShared *pBt; /* The tree being checked out */ 619 Pager *pPager; /* The associated pager. Also accessible by pBt->pPager */ 620 int nPage; /* Number of pages in the database */ 621 int *anRef; /* Number of times each page is referenced */ 622 int mxErr; /* Stop accumulating errors when this reaches zero */ 623 char *zErrMsg; /* An error message. NULL if no errors seen. */ 624 int nErr; /* Number of messages written to zErrMsg so far */ 625 }; 626 627 /* 628 ** Read or write a two- and four-byte big-endian integer values. 629 */ 630 #define get2byte(x) ((x)[0]<<8 | (x)[1]) 631 #define put2byte(p,v) ((p)[0] = (v)>>8, (p)[1] = (v)) 632 #define get4byte sqlite3Get4byte 633 #define put4byte sqlite3Put4byte 634 635 /* 636 ** Internal routines that should be accessed by the btree layer only. 637 */ 638 int sqlite3BtreeGetPage(BtShared*, Pgno, MemPage**, int); 639 int sqlite3BtreeInitPage(MemPage *pPage, MemPage *pParent); 640 void sqlite3BtreeParseCellPtr(MemPage*, u8*, CellInfo*); 641 void sqlite3BtreeParseCell(MemPage*, int, CellInfo*); 642 #ifdef SQLITE_TEST 643 u8 *sqlite3BtreeFindCell(MemPage *pPage, int iCell); 644 #endif 645 int sqlite3BtreeRestoreOrClearCursorPosition(BtCursor *pCur); 646 void sqlite3BtreeGetTempCursor(BtCursor *pCur, BtCursor *pTempCur); 647 void sqlite3BtreeReleaseTempCursor(BtCursor *pCur); 648 int sqlite3BtreeIsRootPage(MemPage *pPage); 649 void sqlite3BtreeMoveToParent(BtCursor *pCur); 650