dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1 | /* |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2 | ** 2010 February 1 |
| 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 | ** |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 13 | ** This file contains the implementation of a write-ahead log (WAL) used in |
| 14 | ** "journal_mode=WAL" mode. |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 15 | ** |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 16 | ** WRITE-AHEAD LOG (WAL) FILE FORMAT |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 17 | ** |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 18 | ** A WAL file consists of a header followed by zero or more "frames". |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 19 | ** Each frame records the revised content of a single page from the |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 20 | ** database file. All changes to the database are recorded by writing |
| 21 | ** frames into the WAL. Transactions commit when a frame is written that |
| 22 | ** contains a commit marker. A single WAL can and usually does record |
| 23 | ** multiple transactions. Periodically, the content of the WAL is |
| 24 | ** transferred back into the database file in an operation called a |
| 25 | ** "checkpoint". |
| 26 | ** |
| 27 | ** A single WAL file can be used multiple times. In other words, the |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 28 | ** WAL can fill up with frames and then be checkpointed and then new |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 29 | ** frames can overwrite the old ones. A WAL always grows from beginning |
| 30 | ** toward the end. Checksums and counters attached to each frame are |
| 31 | ** used to determine which frames within the WAL are valid and which |
| 32 | ** are leftovers from prior checkpoints. |
| 33 | ** |
drh | cd28508 | 2010-06-23 22:00:35 +0000 | [diff] [blame] | 34 | ** The WAL header is 32 bytes in size and consists of the following eight |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 35 | ** big-endian 32-bit unsigned integer values: |
| 36 | ** |
drh | 1b78eaf | 2010-05-25 13:40:03 +0000 | [diff] [blame] | 37 | ** 0: Magic number. 0x377f0682 or 0x377f0683 |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 38 | ** 4: File format version. Currently 3007000 |
| 39 | ** 8: Database page size. Example: 1024 |
| 40 | ** 12: Checkpoint sequence number |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 41 | ** 16: Salt-1, random integer incremented with each checkpoint |
| 42 | ** 20: Salt-2, a different random integer changing with each ckpt |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 43 | ** 24: Checksum-1 (first part of checksum for first 24 bytes of header). |
| 44 | ** 28: Checksum-2 (second part of checksum for first 24 bytes of header). |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 45 | ** |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 46 | ** Immediately following the wal-header are zero or more frames. Each |
| 47 | ** frame consists of a 24-byte frame-header followed by a <page-size> bytes |
drh | cd28508 | 2010-06-23 22:00:35 +0000 | [diff] [blame] | 48 | ** of page data. The frame-header is six big-endian 32-bit unsigned |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 49 | ** integer values, as follows: |
| 50 | ** |
dan | 3de777f | 2010-04-17 12:31:37 +0000 | [diff] [blame] | 51 | ** 0: Page number. |
| 52 | ** 4: For commit records, the size of the database image in pages |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 53 | ** after the commit. For all other records, zero. |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 54 | ** 8: Salt-1 (copied from the header) |
| 55 | ** 12: Salt-2 (copied from the header) |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 56 | ** 16: Checksum-1. |
| 57 | ** 20: Checksum-2. |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 58 | ** |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 59 | ** A frame is considered valid if and only if the following conditions are |
| 60 | ** true: |
| 61 | ** |
| 62 | ** (1) The salt-1 and salt-2 values in the frame-header match |
| 63 | ** salt values in the wal-header |
| 64 | ** |
| 65 | ** (2) The checksum values in the final 8 bytes of the frame-header |
drh | 1b78eaf | 2010-05-25 13:40:03 +0000 | [diff] [blame] | 66 | ** exactly match the checksum computed consecutively on the |
| 67 | ** WAL header and the first 8 bytes and the content of all frames |
| 68 | ** up to and including the current frame. |
| 69 | ** |
| 70 | ** The checksum is computed using 32-bit big-endian integers if the |
| 71 | ** magic number in the first 4 bytes of the WAL is 0x377f0683 and it |
| 72 | ** is computed using little-endian if the magic number is 0x377f0682. |
drh | 51b21b1 | 2010-05-25 15:53:31 +0000 | [diff] [blame] | 73 | ** The checksum values are always stored in the frame header in a |
| 74 | ** big-endian format regardless of which byte order is used to compute |
| 75 | ** the checksum. The checksum is computed by interpreting the input as |
| 76 | ** an even number of unsigned 32-bit integers: x[0] through x[N]. The |
drh | ffca430 | 2010-06-15 11:21:54 +0000 | [diff] [blame] | 77 | ** algorithm used for the checksum is as follows: |
drh | 51b21b1 | 2010-05-25 15:53:31 +0000 | [diff] [blame] | 78 | ** |
| 79 | ** for i from 0 to n-1 step 2: |
| 80 | ** s0 += x[i] + s1; |
| 81 | ** s1 += x[i+1] + s0; |
| 82 | ** endfor |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 83 | ** |
drh | cd28508 | 2010-06-23 22:00:35 +0000 | [diff] [blame] | 84 | ** Note that s0 and s1 are both weighted checksums using fibonacci weights |
| 85 | ** in reverse order (the largest fibonacci weight occurs on the first element |
| 86 | ** of the sequence being summed.) The s1 value spans all 32-bit |
| 87 | ** terms of the sequence whereas s0 omits the final term. |
| 88 | ** |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 89 | ** On a checkpoint, the WAL is first VFS.xSync-ed, then valid content of the |
| 90 | ** WAL is transferred into the database, then the database is VFS.xSync-ed. |
drh | ffca430 | 2010-06-15 11:21:54 +0000 | [diff] [blame] | 91 | ** The VFS.xSync operations serve as write barriers - all writes launched |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 92 | ** before the xSync must complete before any write that launches after the |
| 93 | ** xSync begins. |
| 94 | ** |
| 95 | ** After each checkpoint, the salt-1 value is incremented and the salt-2 |
| 96 | ** value is randomized. This prevents old and new frames in the WAL from |
| 97 | ** being considered valid at the same time and being checkpointing together |
| 98 | ** following a crash. |
| 99 | ** |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 100 | ** READER ALGORITHM |
| 101 | ** |
| 102 | ** To read a page from the database (call it page number P), a reader |
| 103 | ** first checks the WAL to see if it contains page P. If so, then the |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 104 | ** last valid instance of page P that is a followed by a commit frame |
| 105 | ** or is a commit frame itself becomes the value read. If the WAL |
| 106 | ** contains no copies of page P that are valid and which are a commit |
| 107 | ** frame or are followed by a commit frame, then page P is read from |
| 108 | ** the database file. |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 109 | ** |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 110 | ** To start a read transaction, the reader records the index of the last |
| 111 | ** valid frame in the WAL. The reader uses this recorded "mxFrame" value |
| 112 | ** for all subsequent read operations. New transactions can be appended |
| 113 | ** to the WAL, but as long as the reader uses its original mxFrame value |
| 114 | ** and ignores the newly appended content, it will see a consistent snapshot |
| 115 | ** of the database from a single point in time. This technique allows |
| 116 | ** multiple concurrent readers to view different versions of the database |
| 117 | ** content simultaneously. |
| 118 | ** |
| 119 | ** The reader algorithm in the previous paragraphs works correctly, but |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 120 | ** because frames for page P can appear anywhere within the WAL, the |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 121 | ** reader has to scan the entire WAL looking for page P frames. If the |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 122 | ** WAL is large (multiple megabytes is typical) that scan can be slow, |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 123 | ** and read performance suffers. To overcome this problem, a separate |
| 124 | ** data structure called the wal-index is maintained to expedite the |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 125 | ** search for frames of a particular page. |
| 126 | ** |
| 127 | ** WAL-INDEX FORMAT |
| 128 | ** |
| 129 | ** Conceptually, the wal-index is shared memory, though VFS implementations |
| 130 | ** might choose to implement the wal-index using a mmapped file. Because |
| 131 | ** the wal-index is shared memory, SQLite does not support journal_mode=WAL |
| 132 | ** on a network filesystem. All users of the database must be able to |
| 133 | ** share memory. |
| 134 | ** |
| 135 | ** The wal-index is transient. After a crash, the wal-index can (and should |
| 136 | ** be) reconstructed from the original WAL file. In fact, the VFS is required |
| 137 | ** to either truncate or zero the header of the wal-index when the last |
| 138 | ** connection to it closes. Because the wal-index is transient, it can |
| 139 | ** use an architecture-specific format; it does not have to be cross-platform. |
| 140 | ** Hence, unlike the database and WAL file formats which store all values |
| 141 | ** as big endian, the wal-index can store multi-byte values in the native |
| 142 | ** byte order of the host computer. |
| 143 | ** |
| 144 | ** The purpose of the wal-index is to answer this question quickly: Given |
| 145 | ** a page number P, return the index of the last frame for page P in the WAL, |
| 146 | ** or return NULL if there are no frames for page P in the WAL. |
| 147 | ** |
| 148 | ** The wal-index consists of a header region, followed by an one or |
| 149 | ** more index blocks. |
| 150 | ** |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 151 | ** The wal-index header contains the total number of frames within the WAL |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 152 | ** in the the mxFrame field. |
| 153 | ** |
| 154 | ** Each index block except for the first contains information on |
| 155 | ** HASHTABLE_NPAGE frames. The first index block contains information on |
| 156 | ** HASHTABLE_NPAGE_ONE frames. The values of HASHTABLE_NPAGE_ONE and |
| 157 | ** HASHTABLE_NPAGE are selected so that together the wal-index header and |
| 158 | ** first index block are the same size as all other index blocks in the |
| 159 | ** wal-index. |
| 160 | ** |
| 161 | ** Each index block contains two sections, a page-mapping that contains the |
| 162 | ** database page number associated with each wal frame, and a hash-table |
drh | ffca430 | 2010-06-15 11:21:54 +0000 | [diff] [blame] | 163 | ** that allows readers to query an index block for a specific page number. |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 164 | ** The page-mapping is an array of HASHTABLE_NPAGE (or HASHTABLE_NPAGE_ONE |
| 165 | ** for the first index block) 32-bit page numbers. The first entry in the |
| 166 | ** first index-block contains the database page number corresponding to the |
| 167 | ** first frame in the WAL file. The first entry in the second index block |
| 168 | ** in the WAL file corresponds to the (HASHTABLE_NPAGE_ONE+1)th frame in |
| 169 | ** the log, and so on. |
| 170 | ** |
| 171 | ** The last index block in a wal-index usually contains less than the full |
| 172 | ** complement of HASHTABLE_NPAGE (or HASHTABLE_NPAGE_ONE) page-numbers, |
| 173 | ** depending on the contents of the WAL file. This does not change the |
| 174 | ** allocated size of the page-mapping array - the page-mapping array merely |
| 175 | ** contains unused entries. |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 176 | ** |
| 177 | ** Even without using the hash table, the last frame for page P |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 178 | ** can be found by scanning the page-mapping sections of each index block |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 179 | ** starting with the last index block and moving toward the first, and |
| 180 | ** within each index block, starting at the end and moving toward the |
| 181 | ** beginning. The first entry that equals P corresponds to the frame |
| 182 | ** holding the content for that page. |
| 183 | ** |
| 184 | ** The hash table consists of HASHTABLE_NSLOT 16-bit unsigned integers. |
| 185 | ** HASHTABLE_NSLOT = 2*HASHTABLE_NPAGE, and there is one entry in the |
| 186 | ** hash table for each page number in the mapping section, so the hash |
| 187 | ** table is never more than half full. The expected number of collisions |
| 188 | ** prior to finding a match is 1. Each entry of the hash table is an |
| 189 | ** 1-based index of an entry in the mapping section of the same |
| 190 | ** index block. Let K be the 1-based index of the largest entry in |
| 191 | ** the mapping section. (For index blocks other than the last, K will |
| 192 | ** always be exactly HASHTABLE_NPAGE (4096) and for the last index block |
| 193 | ** K will be (mxFrame%HASHTABLE_NPAGE).) Unused slots of the hash table |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 194 | ** contain a value of 0. |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 195 | ** |
| 196 | ** To look for page P in the hash table, first compute a hash iKey on |
| 197 | ** P as follows: |
| 198 | ** |
| 199 | ** iKey = (P * 383) % HASHTABLE_NSLOT |
| 200 | ** |
| 201 | ** Then start scanning entries of the hash table, starting with iKey |
| 202 | ** (wrapping around to the beginning when the end of the hash table is |
| 203 | ** reached) until an unused hash slot is found. Let the first unused slot |
| 204 | ** be at index iUnused. (iUnused might be less than iKey if there was |
| 205 | ** wrap-around.) Because the hash table is never more than half full, |
| 206 | ** the search is guaranteed to eventually hit an unused entry. Let |
| 207 | ** iMax be the value between iKey and iUnused, closest to iUnused, |
| 208 | ** where aHash[iMax]==P. If there is no iMax entry (if there exists |
| 209 | ** no hash slot such that aHash[i]==p) then page P is not in the |
| 210 | ** current index block. Otherwise the iMax-th mapping entry of the |
| 211 | ** current index block corresponds to the last entry that references |
| 212 | ** page P. |
| 213 | ** |
| 214 | ** A hash search begins with the last index block and moves toward the |
| 215 | ** first index block, looking for entries corresponding to page P. On |
| 216 | ** average, only two or three slots in each index block need to be |
| 217 | ** examined in order to either find the last entry for page P, or to |
| 218 | ** establish that no such entry exists in the block. Each index block |
| 219 | ** holds over 4000 entries. So two or three index blocks are sufficient |
| 220 | ** to cover a typical 10 megabyte WAL file, assuming 1K pages. 8 or 10 |
| 221 | ** comparisons (on average) suffice to either locate a frame in the |
| 222 | ** WAL or to establish that the frame does not exist in the WAL. This |
| 223 | ** is much faster than scanning the entire 10MB WAL. |
| 224 | ** |
| 225 | ** Note that entries are added in order of increasing K. Hence, one |
| 226 | ** reader might be using some value K0 and a second reader that started |
| 227 | ** at a later time (after additional transactions were added to the WAL |
| 228 | ** and to the wal-index) might be using a different value K1, where K1>K0. |
| 229 | ** Both readers can use the same hash table and mapping section to get |
| 230 | ** the correct result. There may be entries in the hash table with |
| 231 | ** K>K0 but to the first reader, those entries will appear to be unused |
| 232 | ** slots in the hash table and so the first reader will get an answer as |
| 233 | ** if no values greater than K0 had ever been inserted into the hash table |
| 234 | ** in the first place - which is what reader one wants. Meanwhile, the |
| 235 | ** second reader using K1 will see additional values that were inserted |
| 236 | ** later, which is exactly what reader two wants. |
| 237 | ** |
dan | 6f15014 | 2010-05-21 15:31:56 +0000 | [diff] [blame] | 238 | ** When a rollback occurs, the value of K is decreased. Hash table entries |
| 239 | ** that correspond to frames greater than the new K value are removed |
| 240 | ** from the hash table at this point. |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 241 | */ |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 242 | #ifndef SQLITE_OMIT_WAL |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 243 | |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 244 | #include "wal.h" |
| 245 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 246 | /* |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 247 | ** Trace output macros |
| 248 | */ |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 249 | #if defined(SQLITE_TEST) && defined(SQLITE_DEBUG) |
drh | 15d6809 | 2010-05-31 16:56:14 +0000 | [diff] [blame] | 250 | int sqlite3WalTrace = 0; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 251 | # define WALTRACE(X) if(sqlite3WalTrace) sqlite3DebugPrintf X |
| 252 | #else |
| 253 | # define WALTRACE(X) |
| 254 | #endif |
| 255 | |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 256 | /* |
| 257 | ** The maximum (and only) versions of the wal and wal-index formats |
| 258 | ** that may be interpreted by this version of SQLite. |
| 259 | ** |
| 260 | ** If a client begins recovering a WAL file and finds that (a) the checksum |
| 261 | ** values in the wal-header are correct and (b) the version field is not |
| 262 | ** WAL_MAX_VERSION, recovery fails and SQLite returns SQLITE_CANTOPEN. |
| 263 | ** |
| 264 | ** Similarly, if a client successfully reads a wal-index header (i.e. the |
| 265 | ** checksum test is successful) and finds that the version field is not |
| 266 | ** WALINDEX_MAX_VERSION, then no read-transaction is opened and SQLite |
| 267 | ** returns SQLITE_CANTOPEN. |
| 268 | */ |
| 269 | #define WAL_MAX_VERSION 3007000 |
| 270 | #define WALINDEX_MAX_VERSION 3007000 |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 271 | |
| 272 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 273 | ** Indices of various locking bytes. WAL_NREADER is the number |
| 274 | ** of available reader locks and should be at least 3. |
| 275 | */ |
| 276 | #define WAL_WRITE_LOCK 0 |
| 277 | #define WAL_ALL_BUT_WRITE 1 |
| 278 | #define WAL_CKPT_LOCK 1 |
| 279 | #define WAL_RECOVER_LOCK 2 |
| 280 | #define WAL_READ_LOCK(I) (3+(I)) |
| 281 | #define WAL_NREADER (SQLITE_SHM_NLOCK-3) |
| 282 | |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 283 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 284 | /* Object declarations */ |
| 285 | typedef struct WalIndexHdr WalIndexHdr; |
| 286 | typedef struct WalIterator WalIterator; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 287 | typedef struct WalCkptInfo WalCkptInfo; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 288 | |
| 289 | |
| 290 | /* |
drh | 286a288 | 2010-05-20 23:51:06 +0000 | [diff] [blame] | 291 | ** The following object holds a copy of the wal-index header content. |
| 292 | ** |
| 293 | ** The actual header in the wal-index consists of two copies of this |
| 294 | ** object. |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 295 | ** |
| 296 | ** The szPage value can be any power of 2 between 512 and 32768, inclusive. |
| 297 | ** Or it can be 1 to represent a 65536-byte page. The latter case was |
| 298 | ** added in 3.7.1 when support for 64K pages was added. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 299 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 300 | struct WalIndexHdr { |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 301 | u32 iVersion; /* Wal-index version */ |
| 302 | u32 unused; /* Unused (padding) field */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 303 | u32 iChange; /* Counter incremented each transaction */ |
drh | 4b82c38 | 2010-05-31 18:24:19 +0000 | [diff] [blame] | 304 | u8 isInit; /* 1 when initialized */ |
| 305 | u8 bigEndCksum; /* True if checksums in WAL are big-endian */ |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 306 | u16 szPage; /* Database page size in bytes. 1==64K */ |
dan | d0aa342 | 2010-05-31 16:41:53 +0000 | [diff] [blame] | 307 | u32 mxFrame; /* Index of last valid frame in the WAL */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 308 | u32 nPage; /* Size of database in pages */ |
| 309 | u32 aFrameCksum[2]; /* Checksum of last frame in log */ |
| 310 | u32 aSalt[2]; /* Two salt values copied from WAL header */ |
| 311 | u32 aCksum[2]; /* Checksum over all prior fields */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 312 | }; |
| 313 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 314 | /* |
| 315 | ** A copy of the following object occurs in the wal-index immediately |
| 316 | ** following the second copy of the WalIndexHdr. This object stores |
| 317 | ** information used by checkpoint. |
| 318 | ** |
| 319 | ** nBackfill is the number of frames in the WAL that have been written |
| 320 | ** back into the database. (We call the act of moving content from WAL to |
| 321 | ** database "backfilling".) The nBackfill number is never greater than |
| 322 | ** WalIndexHdr.mxFrame. nBackfill can only be increased by threads |
| 323 | ** holding the WAL_CKPT_LOCK lock (which includes a recovery thread). |
| 324 | ** However, a WAL_WRITE_LOCK thread can move the value of nBackfill from |
| 325 | ** mxFrame back to zero when the WAL is reset. |
| 326 | ** |
| 327 | ** There is one entry in aReadMark[] for each reader lock. If a reader |
| 328 | ** holds read-lock K, then the value in aReadMark[K] is no greater than |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 329 | ** the mxFrame for that reader. The value READMARK_NOT_USED (0xffffffff) |
| 330 | ** for any aReadMark[] means that entry is unused. aReadMark[0] is |
| 331 | ** a special case; its value is never used and it exists as a place-holder |
| 332 | ** to avoid having to offset aReadMark[] indexs by one. Readers holding |
| 333 | ** WAL_READ_LOCK(0) always ignore the entire WAL and read all content |
| 334 | ** directly from the database. |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 335 | ** |
| 336 | ** The value of aReadMark[K] may only be changed by a thread that |
| 337 | ** is holding an exclusive lock on WAL_READ_LOCK(K). Thus, the value of |
| 338 | ** aReadMark[K] cannot changed while there is a reader is using that mark |
| 339 | ** since the reader will be holding a shared lock on WAL_READ_LOCK(K). |
| 340 | ** |
| 341 | ** The checkpointer may only transfer frames from WAL to database where |
| 342 | ** the frame numbers are less than or equal to every aReadMark[] that is |
| 343 | ** in use (that is, every aReadMark[j] for which there is a corresponding |
| 344 | ** WAL_READ_LOCK(j)). New readers (usually) pick the aReadMark[] with the |
| 345 | ** largest value and will increase an unused aReadMark[] to mxFrame if there |
| 346 | ** is not already an aReadMark[] equal to mxFrame. The exception to the |
| 347 | ** previous sentence is when nBackfill equals mxFrame (meaning that everything |
| 348 | ** in the WAL has been backfilled into the database) then new readers |
| 349 | ** will choose aReadMark[0] which has value 0 and hence such reader will |
| 350 | ** get all their all content directly from the database file and ignore |
| 351 | ** the WAL. |
| 352 | ** |
| 353 | ** Writers normally append new frames to the end of the WAL. However, |
| 354 | ** if nBackfill equals mxFrame (meaning that all WAL content has been |
| 355 | ** written back into the database) and if no readers are using the WAL |
| 356 | ** (in other words, if there are no WAL_READ_LOCK(i) where i>0) then |
| 357 | ** the writer will first "reset" the WAL back to the beginning and start |
| 358 | ** writing new content beginning at frame 1. |
| 359 | ** |
| 360 | ** We assume that 32-bit loads are atomic and so no locks are needed in |
| 361 | ** order to read from any aReadMark[] entries. |
| 362 | */ |
| 363 | struct WalCkptInfo { |
| 364 | u32 nBackfill; /* Number of WAL frames backfilled into DB */ |
| 365 | u32 aReadMark[WAL_NREADER]; /* Reader marks */ |
| 366 | }; |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 367 | #define READMARK_NOT_USED 0xffffffff |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 368 | |
| 369 | |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 370 | /* A block of WALINDEX_LOCK_RESERVED bytes beginning at |
| 371 | ** WALINDEX_LOCK_OFFSET is reserved for locks. Since some systems |
| 372 | ** only support mandatory file-locks, we do not read or write data |
| 373 | ** from the region of the file on which locks are applied. |
dan | ff20701 | 2010-04-24 04:49:15 +0000 | [diff] [blame] | 374 | */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 375 | #define WALINDEX_LOCK_OFFSET (sizeof(WalIndexHdr)*2 + sizeof(WalCkptInfo)) |
| 376 | #define WALINDEX_LOCK_RESERVED 16 |
drh | 026ac28 | 2010-05-26 15:06:38 +0000 | [diff] [blame] | 377 | #define WALINDEX_HDR_SIZE (WALINDEX_LOCK_OFFSET+WALINDEX_LOCK_RESERVED) |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 378 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 379 | /* Size of header before each frame in wal */ |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 380 | #define WAL_FRAME_HDRSIZE 24 |
dan | ff20701 | 2010-04-24 04:49:15 +0000 | [diff] [blame] | 381 | |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 382 | /* Size of write ahead log header, including checksum. */ |
| 383 | /* #define WAL_HDRSIZE 24 */ |
| 384 | #define WAL_HDRSIZE 32 |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 385 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 386 | /* WAL magic value. Either this value, or the same value with the least |
| 387 | ** significant bit also set (WAL_MAGIC | 0x00000001) is stored in 32-bit |
| 388 | ** big-endian format in the first 4 bytes of a WAL file. |
| 389 | ** |
| 390 | ** If the LSB is set, then the checksums for each frame within the WAL |
| 391 | ** file are calculated by treating all data as an array of 32-bit |
| 392 | ** big-endian words. Otherwise, they are calculated by interpreting |
| 393 | ** all data as 32-bit little-endian words. |
| 394 | */ |
| 395 | #define WAL_MAGIC 0x377f0682 |
| 396 | |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 397 | /* |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 398 | ** Return the offset of frame iFrame in the write-ahead log file, |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 399 | ** assuming a database page size of szPage bytes. The offset returned |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 400 | ** is to the start of the write-ahead log frame-header. |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 401 | */ |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 402 | #define walFrameOffset(iFrame, szPage) ( \ |
dan | bd0e907 | 2010-07-07 09:48:44 +0000 | [diff] [blame] | 403 | WAL_HDRSIZE + ((iFrame)-1)*(i64)((szPage)+WAL_FRAME_HDRSIZE) \ |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 404 | ) |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 405 | |
| 406 | /* |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 407 | ** An open write-ahead log file is represented by an instance of the |
| 408 | ** following object. |
dan | ce4f05f | 2010-04-22 19:14:13 +0000 | [diff] [blame] | 409 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 410 | struct Wal { |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 411 | sqlite3_vfs *pVfs; /* The VFS used to create pDbFd */ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 412 | sqlite3_file *pDbFd; /* File handle for the database file */ |
| 413 | sqlite3_file *pWalFd; /* File handle for WAL file */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 414 | u32 iCallback; /* Value to pass to log callback (or 0) */ |
drh | 85a8375 | 2011-05-16 21:00:27 +0000 | [diff] [blame] | 415 | i64 mxWalSize; /* Truncate WAL to this size upon reset */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 416 | int nWiData; /* Size of array apWiData */ |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 417 | int szFirstBlock; /* Size of first block written to WAL file */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 418 | volatile u32 **apWiData; /* Pointer to wal-index content in memory */ |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 419 | u32 szPage; /* Database page size */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 420 | i16 readLock; /* Which read lock is being held. -1 for none */ |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 421 | u8 syncFlags; /* Flags to use to sync header writes */ |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 422 | u8 exclusiveMode; /* Non-zero if connection is in exclusive mode */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 423 | u8 writeLock; /* True if in a write transaction */ |
| 424 | u8 ckptLock; /* True if holding a checkpoint lock */ |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 425 | u8 readOnly; /* WAL_RDWR, WAL_RDONLY, or WAL_SHM_RDONLY */ |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 426 | u8 truncateOnCommit; /* True to truncate WAL file on commit */ |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 427 | u8 noSyncHeader; /* Avoid WAL header fsyncs if true */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 428 | WalIndexHdr hdr; /* Wal-index header for current transaction */ |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 429 | const char *zWalName; /* Name of WAL file */ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 430 | u32 nCkpt; /* Checkpoint sequence counter in the wal-header */ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 431 | #ifdef SQLITE_DEBUG |
| 432 | u8 lockError; /* True if a locking error has occurred */ |
| 433 | #endif |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 434 | }; |
| 435 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 436 | /* |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 437 | ** Candidate values for Wal.exclusiveMode. |
| 438 | */ |
| 439 | #define WAL_NORMAL_MODE 0 |
| 440 | #define WAL_EXCLUSIVE_MODE 1 |
| 441 | #define WAL_HEAPMEMORY_MODE 2 |
| 442 | |
| 443 | /* |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 444 | ** Possible values for WAL.readOnly |
| 445 | */ |
| 446 | #define WAL_RDWR 0 /* Normal read/write connection */ |
| 447 | #define WAL_RDONLY 1 /* The WAL file is readonly */ |
| 448 | #define WAL_SHM_RDONLY 2 /* The SHM file is readonly */ |
| 449 | |
| 450 | /* |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 451 | ** Each page of the wal-index mapping contains a hash-table made up of |
| 452 | ** an array of HASHTABLE_NSLOT elements of the following type. |
| 453 | */ |
| 454 | typedef u16 ht_slot; |
| 455 | |
| 456 | /* |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 457 | ** This structure is used to implement an iterator that loops through |
| 458 | ** all frames in the WAL in database page order. Where two or more frames |
| 459 | ** correspond to the same database page, the iterator visits only the |
| 460 | ** frame most recently written to the WAL (in other words, the frame with |
| 461 | ** the largest index). |
| 462 | ** |
| 463 | ** The internals of this structure are only accessed by: |
| 464 | ** |
| 465 | ** walIteratorInit() - Create a new iterator, |
| 466 | ** walIteratorNext() - Step an iterator, |
| 467 | ** walIteratorFree() - Free an iterator. |
| 468 | ** |
| 469 | ** This functionality is used by the checkpoint code (see walCheckpoint()). |
| 470 | */ |
| 471 | struct WalIterator { |
| 472 | int iPrior; /* Last result returned from the iterator */ |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 473 | int nSegment; /* Number of entries in aSegment[] */ |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 474 | struct WalSegment { |
| 475 | int iNext; /* Next slot in aIndex[] not yet returned */ |
| 476 | ht_slot *aIndex; /* i0, i1, i2... such that aPgno[iN] ascend */ |
| 477 | u32 *aPgno; /* Array of page numbers. */ |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 478 | int nEntry; /* Nr. of entries in aPgno[] and aIndex[] */ |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 479 | int iZero; /* Frame number associated with aPgno[0] */ |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 480 | } aSegment[1]; /* One for every 32KB page in the wal-index */ |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 481 | }; |
| 482 | |
| 483 | /* |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 484 | ** Define the parameters of the hash tables in the wal-index file. There |
| 485 | ** is a hash-table following every HASHTABLE_NPAGE page numbers in the |
| 486 | ** wal-index. |
| 487 | ** |
| 488 | ** Changing any of these constants will alter the wal-index format and |
| 489 | ** create incompatibilities. |
| 490 | */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 491 | #define HASHTABLE_NPAGE 4096 /* Must be power of 2 */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 492 | #define HASHTABLE_HASH_1 383 /* Should be prime */ |
| 493 | #define HASHTABLE_NSLOT (HASHTABLE_NPAGE*2) /* Must be a power of 2 */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 494 | |
dan | ad3cadd | 2010-06-14 11:49:26 +0000 | [diff] [blame] | 495 | /* |
| 496 | ** The block of page numbers associated with the first hash-table in a |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 497 | ** wal-index is smaller than usual. This is so that there is a complete |
| 498 | ** hash-table on each aligned 32KB page of the wal-index. |
| 499 | */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 500 | #define HASHTABLE_NPAGE_ONE (HASHTABLE_NPAGE - (WALINDEX_HDR_SIZE/sizeof(u32))) |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 501 | |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 502 | /* The wal-index is divided into pages of WALINDEX_PGSZ bytes each. */ |
| 503 | #define WALINDEX_PGSZ ( \ |
| 504 | sizeof(ht_slot)*HASHTABLE_NSLOT + HASHTABLE_NPAGE*sizeof(u32) \ |
| 505 | ) |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 506 | |
| 507 | /* |
| 508 | ** Obtain a pointer to the iPage'th page of the wal-index. The wal-index |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 509 | ** is broken into pages of WALINDEX_PGSZ bytes. Wal-index pages are |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 510 | ** numbered from zero. |
| 511 | ** |
| 512 | ** If this call is successful, *ppPage is set to point to the wal-index |
| 513 | ** page and SQLITE_OK is returned. If an error (an OOM or VFS error) occurs, |
| 514 | ** then an SQLite error code is returned and *ppPage is set to 0. |
| 515 | */ |
| 516 | static int walIndexPage(Wal *pWal, int iPage, volatile u32 **ppPage){ |
| 517 | int rc = SQLITE_OK; |
| 518 | |
| 519 | /* Enlarge the pWal->apWiData[] array if required */ |
| 520 | if( pWal->nWiData<=iPage ){ |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 521 | int nByte = sizeof(u32*)*(iPage+1); |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 522 | volatile u32 **apNew; |
shaneh | 8a300f8 | 2010-07-02 18:15:31 +0000 | [diff] [blame] | 523 | apNew = (volatile u32 **)sqlite3_realloc((void *)pWal->apWiData, nByte); |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 524 | if( !apNew ){ |
| 525 | *ppPage = 0; |
| 526 | return SQLITE_NOMEM; |
| 527 | } |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 528 | memset((void*)&apNew[pWal->nWiData], 0, |
| 529 | sizeof(u32*)*(iPage+1-pWal->nWiData)); |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 530 | pWal->apWiData = apNew; |
| 531 | pWal->nWiData = iPage+1; |
| 532 | } |
| 533 | |
| 534 | /* Request a pointer to the required page from the VFS */ |
| 535 | if( pWal->apWiData[iPage]==0 ){ |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 536 | if( pWal->exclusiveMode==WAL_HEAPMEMORY_MODE ){ |
| 537 | pWal->apWiData[iPage] = (u32 volatile *)sqlite3MallocZero(WALINDEX_PGSZ); |
| 538 | if( !pWal->apWiData[iPage] ) rc = SQLITE_NOMEM; |
| 539 | }else{ |
| 540 | rc = sqlite3OsShmMap(pWal->pDbFd, iPage, WALINDEX_PGSZ, |
| 541 | pWal->writeLock, (void volatile **)&pWal->apWiData[iPage] |
| 542 | ); |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 543 | if( rc==SQLITE_READONLY ){ |
| 544 | pWal->readOnly |= WAL_SHM_RDONLY; |
| 545 | rc = SQLITE_OK; |
dan | 4edc6bf | 2011-05-10 17:31:29 +0000 | [diff] [blame] | 546 | } |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 547 | } |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 548 | } |
dan | b6d2f9c | 2011-05-11 14:57:33 +0000 | [diff] [blame] | 549 | |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 550 | *ppPage = pWal->apWiData[iPage]; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 551 | assert( iPage==0 || *ppPage || rc!=SQLITE_OK ); |
| 552 | return rc; |
| 553 | } |
| 554 | |
| 555 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 556 | ** Return a pointer to the WalCkptInfo structure in the wal-index. |
| 557 | */ |
| 558 | static volatile WalCkptInfo *walCkptInfo(Wal *pWal){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 559 | assert( pWal->nWiData>0 && pWal->apWiData[0] ); |
| 560 | return (volatile WalCkptInfo*)&(pWal->apWiData[0][sizeof(WalIndexHdr)/2]); |
| 561 | } |
| 562 | |
| 563 | /* |
| 564 | ** Return a pointer to the WalIndexHdr structure in the wal-index. |
| 565 | */ |
| 566 | static volatile WalIndexHdr *walIndexHdr(Wal *pWal){ |
| 567 | assert( pWal->nWiData>0 && pWal->apWiData[0] ); |
| 568 | return (volatile WalIndexHdr*)pWal->apWiData[0]; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 569 | } |
| 570 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 571 | /* |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 572 | ** The argument to this macro must be of type u32. On a little-endian |
| 573 | ** architecture, it returns the u32 value that results from interpreting |
| 574 | ** the 4 bytes as a big-endian value. On a big-endian architecture, it |
| 575 | ** returns the value that would be produced by intepreting the 4 bytes |
| 576 | ** of the input value as a little-endian integer. |
| 577 | */ |
| 578 | #define BYTESWAP32(x) ( \ |
| 579 | (((x)&0x000000FF)<<24) + (((x)&0x0000FF00)<<8) \ |
| 580 | + (((x)&0x00FF0000)>>8) + (((x)&0xFF000000)>>24) \ |
| 581 | ) |
dan | 64d039e | 2010-04-13 19:27:31 +0000 | [diff] [blame] | 582 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 583 | /* |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 584 | ** Generate or extend an 8 byte checksum based on the data in |
| 585 | ** array aByte[] and the initial values of aIn[0] and aIn[1] (or |
| 586 | ** initial values of 0 and 0 if aIn==NULL). |
| 587 | ** |
| 588 | ** The checksum is written back into aOut[] before returning. |
| 589 | ** |
| 590 | ** nByte must be a positive multiple of 8. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 591 | */ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 592 | static void walChecksumBytes( |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 593 | int nativeCksum, /* True for native byte-order, false for non-native */ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 594 | u8 *a, /* Content to be checksummed */ |
| 595 | int nByte, /* Bytes of content in a[]. Must be a multiple of 8. */ |
| 596 | const u32 *aIn, /* Initial checksum value input */ |
| 597 | u32 *aOut /* OUT: Final checksum value output */ |
| 598 | ){ |
| 599 | u32 s1, s2; |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 600 | u32 *aData = (u32 *)a; |
| 601 | u32 *aEnd = (u32 *)&a[nByte]; |
| 602 | |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 603 | if( aIn ){ |
| 604 | s1 = aIn[0]; |
| 605 | s2 = aIn[1]; |
| 606 | }else{ |
| 607 | s1 = s2 = 0; |
| 608 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 609 | |
drh | 584c754 | 2010-05-19 18:08:10 +0000 | [diff] [blame] | 610 | assert( nByte>=8 ); |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 611 | assert( (nByte&0x00000007)==0 ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 612 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 613 | if( nativeCksum ){ |
| 614 | do { |
| 615 | s1 += *aData++ + s2; |
| 616 | s2 += *aData++ + s1; |
| 617 | }while( aData<aEnd ); |
| 618 | }else{ |
| 619 | do { |
| 620 | s1 += BYTESWAP32(aData[0]) + s2; |
| 621 | s2 += BYTESWAP32(aData[1]) + s1; |
| 622 | aData += 2; |
| 623 | }while( aData<aEnd ); |
| 624 | } |
| 625 | |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 626 | aOut[0] = s1; |
| 627 | aOut[1] = s2; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 628 | } |
| 629 | |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 630 | static void walShmBarrier(Wal *pWal){ |
| 631 | if( pWal->exclusiveMode!=WAL_HEAPMEMORY_MODE ){ |
| 632 | sqlite3OsShmBarrier(pWal->pDbFd); |
| 633 | } |
| 634 | } |
| 635 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 636 | /* |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 637 | ** Write the header information in pWal->hdr into the wal-index. |
| 638 | ** |
| 639 | ** The checksum on pWal->hdr is updated before it is written. |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 640 | */ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 641 | static void walIndexWriteHdr(Wal *pWal){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 642 | volatile WalIndexHdr *aHdr = walIndexHdr(pWal); |
| 643 | const int nCksum = offsetof(WalIndexHdr, aCksum); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 644 | |
| 645 | assert( pWal->writeLock ); |
drh | 4b82c38 | 2010-05-31 18:24:19 +0000 | [diff] [blame] | 646 | pWal->hdr.isInit = 1; |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 647 | pWal->hdr.iVersion = WALINDEX_MAX_VERSION; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 648 | walChecksumBytes(1, (u8*)&pWal->hdr, nCksum, 0, pWal->hdr.aCksum); |
| 649 | memcpy((void *)&aHdr[1], (void *)&pWal->hdr, sizeof(WalIndexHdr)); |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 650 | walShmBarrier(pWal); |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 651 | memcpy((void *)&aHdr[0], (void *)&pWal->hdr, sizeof(WalIndexHdr)); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 652 | } |
| 653 | |
| 654 | /* |
| 655 | ** This function encodes a single frame header and writes it to a buffer |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 656 | ** supplied by the caller. A frame-header is made up of a series of |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 657 | ** 4-byte big-endian integers, as follows: |
| 658 | ** |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 659 | ** 0: Page number. |
| 660 | ** 4: For commit records, the size of the database image in pages |
| 661 | ** after the commit. For all other records, zero. |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 662 | ** 8: Salt-1 (copied from the wal-header) |
| 663 | ** 12: Salt-2 (copied from the wal-header) |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 664 | ** 16: Checksum-1. |
| 665 | ** 20: Checksum-2. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 666 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 667 | static void walEncodeFrame( |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 668 | Wal *pWal, /* The write-ahead log */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 669 | u32 iPage, /* Database page number for frame */ |
| 670 | u32 nTruncate, /* New db size (or 0 for non-commit frames) */ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 671 | u8 *aData, /* Pointer to page data */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 672 | u8 *aFrame /* OUT: Write encoded frame here */ |
| 673 | ){ |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 674 | int nativeCksum; /* True for native byte-order checksums */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 675 | u32 *aCksum = pWal->hdr.aFrameCksum; |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 676 | assert( WAL_FRAME_HDRSIZE==24 ); |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 677 | sqlite3Put4byte(&aFrame[0], iPage); |
| 678 | sqlite3Put4byte(&aFrame[4], nTruncate); |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 679 | memcpy(&aFrame[8], pWal->hdr.aSalt, 8); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 680 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 681 | nativeCksum = (pWal->hdr.bigEndCksum==SQLITE_BIGENDIAN); |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 682 | walChecksumBytes(nativeCksum, aFrame, 8, aCksum, aCksum); |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 683 | walChecksumBytes(nativeCksum, aData, pWal->szPage, aCksum, aCksum); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 684 | |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 685 | sqlite3Put4byte(&aFrame[16], aCksum[0]); |
| 686 | sqlite3Put4byte(&aFrame[20], aCksum[1]); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 687 | } |
| 688 | |
| 689 | /* |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 690 | ** Check to see if the frame with header in aFrame[] and content |
| 691 | ** in aData[] is valid. If it is a valid frame, fill *piPage and |
| 692 | ** *pnTruncate and return true. Return if the frame is not valid. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 693 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 694 | static int walDecodeFrame( |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 695 | Wal *pWal, /* The write-ahead log */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 696 | u32 *piPage, /* OUT: Database page number for frame */ |
| 697 | u32 *pnTruncate, /* OUT: New db size (or 0 if not commit) */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 698 | u8 *aData, /* Pointer to page data (for checksum) */ |
| 699 | u8 *aFrame /* Frame data */ |
| 700 | ){ |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 701 | int nativeCksum; /* True for native byte-order checksums */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 702 | u32 *aCksum = pWal->hdr.aFrameCksum; |
drh | c817915 | 2010-05-24 13:28:36 +0000 | [diff] [blame] | 703 | u32 pgno; /* Page number of the frame */ |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 704 | assert( WAL_FRAME_HDRSIZE==24 ); |
| 705 | |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 706 | /* A frame is only valid if the salt values in the frame-header |
| 707 | ** match the salt values in the wal-header. |
| 708 | */ |
| 709 | if( memcmp(&pWal->hdr.aSalt, &aFrame[8], 8)!=0 ){ |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 710 | return 0; |
| 711 | } |
dan | 4a4b01d | 2010-04-16 11:30:18 +0000 | [diff] [blame] | 712 | |
drh | c817915 | 2010-05-24 13:28:36 +0000 | [diff] [blame] | 713 | /* A frame is only valid if the page number is creater than zero. |
| 714 | */ |
| 715 | pgno = sqlite3Get4byte(&aFrame[0]); |
| 716 | if( pgno==0 ){ |
| 717 | return 0; |
| 718 | } |
| 719 | |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 720 | /* A frame is only valid if a checksum of the WAL header, |
| 721 | ** all prior frams, the first 16 bytes of this frame-header, |
| 722 | ** and the frame-data matches the checksum in the last 8 |
| 723 | ** bytes of this frame-header. |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 724 | */ |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 725 | nativeCksum = (pWal->hdr.bigEndCksum==SQLITE_BIGENDIAN); |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 726 | walChecksumBytes(nativeCksum, aFrame, 8, aCksum, aCksum); |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 727 | walChecksumBytes(nativeCksum, aData, pWal->szPage, aCksum, aCksum); |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 728 | if( aCksum[0]!=sqlite3Get4byte(&aFrame[16]) |
| 729 | || aCksum[1]!=sqlite3Get4byte(&aFrame[20]) |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 730 | ){ |
| 731 | /* Checksum failed. */ |
| 732 | return 0; |
| 733 | } |
| 734 | |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 735 | /* If we reach this point, the frame is valid. Return the page number |
| 736 | ** and the new database size. |
| 737 | */ |
drh | c817915 | 2010-05-24 13:28:36 +0000 | [diff] [blame] | 738 | *piPage = pgno; |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 739 | *pnTruncate = sqlite3Get4byte(&aFrame[4]); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 740 | return 1; |
| 741 | } |
| 742 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 743 | |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 744 | #if defined(SQLITE_TEST) && defined(SQLITE_DEBUG) |
| 745 | /* |
drh | 181e091 | 2010-06-01 01:08:08 +0000 | [diff] [blame] | 746 | ** Names of locks. This routine is used to provide debugging output and is not |
| 747 | ** a part of an ordinary build. |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 748 | */ |
| 749 | static const char *walLockName(int lockIdx){ |
| 750 | if( lockIdx==WAL_WRITE_LOCK ){ |
| 751 | return "WRITE-LOCK"; |
| 752 | }else if( lockIdx==WAL_CKPT_LOCK ){ |
| 753 | return "CKPT-LOCK"; |
| 754 | }else if( lockIdx==WAL_RECOVER_LOCK ){ |
| 755 | return "RECOVER-LOCK"; |
| 756 | }else{ |
| 757 | static char zName[15]; |
| 758 | sqlite3_snprintf(sizeof(zName), zName, "READ-LOCK[%d]", |
| 759 | lockIdx-WAL_READ_LOCK(0)); |
| 760 | return zName; |
| 761 | } |
| 762 | } |
| 763 | #endif /*defined(SQLITE_TEST) || defined(SQLITE_DEBUG) */ |
| 764 | |
| 765 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 766 | /* |
drh | 181e091 | 2010-06-01 01:08:08 +0000 | [diff] [blame] | 767 | ** Set or release locks on the WAL. Locks are either shared or exclusive. |
| 768 | ** A lock cannot be moved directly between shared and exclusive - it must go |
| 769 | ** through the unlocked state first. |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 770 | ** |
| 771 | ** In locking_mode=EXCLUSIVE, all of these routines become no-ops. |
| 772 | */ |
| 773 | static int walLockShared(Wal *pWal, int lockIdx){ |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 774 | int rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 775 | if( pWal->exclusiveMode ) return SQLITE_OK; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 776 | rc = sqlite3OsShmLock(pWal->pDbFd, lockIdx, 1, |
| 777 | SQLITE_SHM_LOCK | SQLITE_SHM_SHARED); |
| 778 | WALTRACE(("WAL%p: acquire SHARED-%s %s\n", pWal, |
| 779 | walLockName(lockIdx), rc ? "failed" : "ok")); |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 780 | VVA_ONLY( pWal->lockError = (u8)(rc!=SQLITE_OK && rc!=SQLITE_BUSY); ) |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 781 | return rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 782 | } |
| 783 | static void walUnlockShared(Wal *pWal, int lockIdx){ |
| 784 | if( pWal->exclusiveMode ) return; |
| 785 | (void)sqlite3OsShmLock(pWal->pDbFd, lockIdx, 1, |
| 786 | SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED); |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 787 | WALTRACE(("WAL%p: release SHARED-%s\n", pWal, walLockName(lockIdx))); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 788 | } |
| 789 | static int walLockExclusive(Wal *pWal, int lockIdx, int n){ |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 790 | int rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 791 | if( pWal->exclusiveMode ) return SQLITE_OK; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 792 | rc = sqlite3OsShmLock(pWal->pDbFd, lockIdx, n, |
| 793 | SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE); |
| 794 | WALTRACE(("WAL%p: acquire EXCLUSIVE-%s cnt=%d %s\n", pWal, |
| 795 | walLockName(lockIdx), n, rc ? "failed" : "ok")); |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 796 | VVA_ONLY( pWal->lockError = (u8)(rc!=SQLITE_OK && rc!=SQLITE_BUSY); ) |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 797 | return rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 798 | } |
| 799 | static void walUnlockExclusive(Wal *pWal, int lockIdx, int n){ |
| 800 | if( pWal->exclusiveMode ) return; |
| 801 | (void)sqlite3OsShmLock(pWal->pDbFd, lockIdx, n, |
| 802 | SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE); |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 803 | WALTRACE(("WAL%p: release EXCLUSIVE-%s cnt=%d\n", pWal, |
| 804 | walLockName(lockIdx), n)); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 805 | } |
| 806 | |
| 807 | /* |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 808 | ** Compute a hash on a page number. The resulting hash value must land |
drh | 181e091 | 2010-06-01 01:08:08 +0000 | [diff] [blame] | 809 | ** between 0 and (HASHTABLE_NSLOT-1). The walHashNext() function advances |
| 810 | ** the hash to the next value in the event of a collision. |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 811 | */ |
| 812 | static int walHash(u32 iPage){ |
| 813 | assert( iPage>0 ); |
| 814 | assert( (HASHTABLE_NSLOT & (HASHTABLE_NSLOT-1))==0 ); |
| 815 | return (iPage*HASHTABLE_HASH_1) & (HASHTABLE_NSLOT-1); |
| 816 | } |
| 817 | static int walNextHash(int iPriorHash){ |
| 818 | return (iPriorHash+1)&(HASHTABLE_NSLOT-1); |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 819 | } |
| 820 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 821 | /* |
| 822 | ** Return pointers to the hash table and page number array stored on |
| 823 | ** page iHash of the wal-index. The wal-index is broken into 32KB pages |
| 824 | ** numbered starting from 0. |
| 825 | ** |
| 826 | ** Set output variable *paHash to point to the start of the hash table |
| 827 | ** in the wal-index file. Set *piZero to one less than the frame |
| 828 | ** number of the first frame indexed by this hash table. If a |
| 829 | ** slot in the hash table is set to N, it refers to frame number |
| 830 | ** (*piZero+N) in the log. |
| 831 | ** |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 832 | ** Finally, set *paPgno so that *paPgno[1] is the page number of the |
| 833 | ** first frame indexed by the hash table, frame (*piZero+1). |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 834 | */ |
| 835 | static int walHashGet( |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 836 | Wal *pWal, /* WAL handle */ |
| 837 | int iHash, /* Find the iHash'th table */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 838 | volatile ht_slot **paHash, /* OUT: Pointer to hash index */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 839 | volatile u32 **paPgno, /* OUT: Pointer to page number array */ |
| 840 | u32 *piZero /* OUT: Frame associated with *paPgno[0] */ |
| 841 | ){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 842 | int rc; /* Return code */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 843 | volatile u32 *aPgno; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 844 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 845 | rc = walIndexPage(pWal, iHash, &aPgno); |
| 846 | assert( rc==SQLITE_OK || iHash>0 ); |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 847 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 848 | if( rc==SQLITE_OK ){ |
| 849 | u32 iZero; |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 850 | volatile ht_slot *aHash; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 851 | |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 852 | aHash = (volatile ht_slot *)&aPgno[HASHTABLE_NPAGE]; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 853 | if( iHash==0 ){ |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 854 | aPgno = &aPgno[WALINDEX_HDR_SIZE/sizeof(u32)]; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 855 | iZero = 0; |
| 856 | }else{ |
| 857 | iZero = HASHTABLE_NPAGE_ONE + (iHash-1)*HASHTABLE_NPAGE; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 858 | } |
| 859 | |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 860 | *paPgno = &aPgno[-1]; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 861 | *paHash = aHash; |
| 862 | *piZero = iZero; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 863 | } |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 864 | return rc; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 865 | } |
| 866 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 867 | /* |
| 868 | ** Return the number of the wal-index page that contains the hash-table |
| 869 | ** and page-number array that contain entries corresponding to WAL frame |
| 870 | ** iFrame. The wal-index is broken up into 32KB pages. Wal-index pages |
| 871 | ** are numbered starting from 0. |
| 872 | */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 873 | static int walFramePage(u32 iFrame){ |
| 874 | int iHash = (iFrame+HASHTABLE_NPAGE-HASHTABLE_NPAGE_ONE-1) / HASHTABLE_NPAGE; |
| 875 | assert( (iHash==0 || iFrame>HASHTABLE_NPAGE_ONE) |
| 876 | && (iHash>=1 || iFrame<=HASHTABLE_NPAGE_ONE) |
| 877 | && (iHash<=1 || iFrame>(HASHTABLE_NPAGE_ONE+HASHTABLE_NPAGE)) |
| 878 | && (iHash>=2 || iFrame<=HASHTABLE_NPAGE_ONE+HASHTABLE_NPAGE) |
| 879 | && (iHash<=2 || iFrame>(HASHTABLE_NPAGE_ONE+2*HASHTABLE_NPAGE)) |
| 880 | ); |
| 881 | return iHash; |
| 882 | } |
| 883 | |
| 884 | /* |
| 885 | ** Return the page number associated with frame iFrame in this WAL. |
| 886 | */ |
| 887 | static u32 walFramePgno(Wal *pWal, u32 iFrame){ |
| 888 | int iHash = walFramePage(iFrame); |
| 889 | if( iHash==0 ){ |
| 890 | return pWal->apWiData[0][WALINDEX_HDR_SIZE/sizeof(u32) + iFrame - 1]; |
| 891 | } |
| 892 | return pWal->apWiData[iHash][(iFrame-1-HASHTABLE_NPAGE_ONE)%HASHTABLE_NPAGE]; |
| 893 | } |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 894 | |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 895 | /* |
| 896 | ** Remove entries from the hash table that point to WAL slots greater |
| 897 | ** than pWal->hdr.mxFrame. |
| 898 | ** |
| 899 | ** This function is called whenever pWal->hdr.mxFrame is decreased due |
| 900 | ** to a rollback or savepoint. |
| 901 | ** |
drh | 181e091 | 2010-06-01 01:08:08 +0000 | [diff] [blame] | 902 | ** At most only the hash table containing pWal->hdr.mxFrame needs to be |
| 903 | ** updated. Any later hash tables will be automatically cleared when |
| 904 | ** pWal->hdr.mxFrame advances to the point where those hash tables are |
| 905 | ** actually needed. |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 906 | */ |
| 907 | static void walCleanupHash(Wal *pWal){ |
drh | ff82894 | 2010-06-26 21:34:06 +0000 | [diff] [blame] | 908 | volatile ht_slot *aHash = 0; /* Pointer to hash table to clear */ |
| 909 | volatile u32 *aPgno = 0; /* Page number array for hash table */ |
| 910 | u32 iZero = 0; /* frame == (aHash[x]+iZero) */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 911 | int iLimit = 0; /* Zero values greater than this */ |
| 912 | int nByte; /* Number of bytes to zero in aPgno[] */ |
| 913 | int i; /* Used to iterate through aHash[] */ |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 914 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 915 | assert( pWal->writeLock ); |
drh | ffca430 | 2010-06-15 11:21:54 +0000 | [diff] [blame] | 916 | testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE-1 ); |
| 917 | testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE ); |
| 918 | testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE+1 ); |
drh | 9c15647 | 2010-06-01 12:58:41 +0000 | [diff] [blame] | 919 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 920 | if( pWal->hdr.mxFrame==0 ) return; |
| 921 | |
| 922 | /* Obtain pointers to the hash-table and page-number array containing |
| 923 | ** the entry that corresponds to frame pWal->hdr.mxFrame. It is guaranteed |
| 924 | ** that the page said hash-table and array reside on is already mapped. |
| 925 | */ |
| 926 | assert( pWal->nWiData>walFramePage(pWal->hdr.mxFrame) ); |
| 927 | assert( pWal->apWiData[walFramePage(pWal->hdr.mxFrame)] ); |
| 928 | walHashGet(pWal, walFramePage(pWal->hdr.mxFrame), &aHash, &aPgno, &iZero); |
| 929 | |
| 930 | /* Zero all hash-table entries that correspond to frame numbers greater |
| 931 | ** than pWal->hdr.mxFrame. |
| 932 | */ |
| 933 | iLimit = pWal->hdr.mxFrame - iZero; |
| 934 | assert( iLimit>0 ); |
| 935 | for(i=0; i<HASHTABLE_NSLOT; i++){ |
| 936 | if( aHash[i]>iLimit ){ |
| 937 | aHash[i] = 0; |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 938 | } |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 939 | } |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 940 | |
| 941 | /* Zero the entries in the aPgno array that correspond to frames with |
| 942 | ** frame numbers greater than pWal->hdr.mxFrame. |
| 943 | */ |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 944 | nByte = (int)((char *)aHash - (char *)&aPgno[iLimit+1]); |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 945 | memset((void *)&aPgno[iLimit+1], 0, nByte); |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 946 | |
| 947 | #ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT |
| 948 | /* Verify that the every entry in the mapping region is still reachable |
| 949 | ** via the hash table even after the cleanup. |
| 950 | */ |
drh | f77bbd9 | 2010-06-01 13:17:44 +0000 | [diff] [blame] | 951 | if( iLimit ){ |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 952 | int i; /* Loop counter */ |
| 953 | int iKey; /* Hash key */ |
| 954 | for(i=1; i<=iLimit; i++){ |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 955 | for(iKey=walHash(aPgno[i]); aHash[iKey]; iKey=walNextHash(iKey)){ |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 956 | if( aHash[iKey]==i ) break; |
| 957 | } |
| 958 | assert( aHash[iKey]==i ); |
| 959 | } |
| 960 | } |
| 961 | #endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */ |
| 962 | } |
| 963 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 964 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 965 | /* |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 966 | ** Set an entry in the wal-index that will map database page number |
| 967 | ** pPage into WAL frame iFrame. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 968 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 969 | static int walIndexAppend(Wal *pWal, u32 iFrame, u32 iPage){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 970 | int rc; /* Return code */ |
drh | ff82894 | 2010-06-26 21:34:06 +0000 | [diff] [blame] | 971 | u32 iZero = 0; /* One less than frame number of aPgno[1] */ |
| 972 | volatile u32 *aPgno = 0; /* Page number array */ |
| 973 | volatile ht_slot *aHash = 0; /* Hash table */ |
dan | ce4f05f | 2010-04-22 19:14:13 +0000 | [diff] [blame] | 974 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 975 | rc = walHashGet(pWal, walFramePage(iFrame), &aHash, &aPgno, &iZero); |
| 976 | |
| 977 | /* Assuming the wal-index file was successfully mapped, populate the |
| 978 | ** page number array and hash table entry. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 979 | */ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 980 | if( rc==SQLITE_OK ){ |
| 981 | int iKey; /* Hash table key */ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 982 | int idx; /* Value to write to hash-table slot */ |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 983 | int nCollide; /* Number of hash collisions */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 984 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 985 | idx = iFrame - iZero; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 986 | assert( idx <= HASHTABLE_NSLOT/2 + 1 ); |
| 987 | |
| 988 | /* If this is the first entry to be added to this hash-table, zero the |
| 989 | ** entire hash table and aPgno[] array before proceding. |
| 990 | */ |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 991 | if( idx==1 ){ |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 992 | int nByte = (int)((u8 *)&aHash[HASHTABLE_NSLOT] - (u8 *)&aPgno[1]); |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 993 | memset((void*)&aPgno[1], 0, nByte); |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 994 | } |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 995 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 996 | /* If the entry in aPgno[] is already set, then the previous writer |
| 997 | ** must have exited unexpectedly in the middle of a transaction (after |
| 998 | ** writing one or more dirty pages to the WAL to free up memory). |
| 999 | ** Remove the remnants of that writers uncommitted transaction from |
| 1000 | ** the hash-table before writing any new entries. |
| 1001 | */ |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 1002 | if( aPgno[idx] ){ |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 1003 | walCleanupHash(pWal); |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 1004 | assert( !aPgno[idx] ); |
dan | ca6b5ba | 2010-05-25 10:50:56 +0000 | [diff] [blame] | 1005 | } |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1006 | |
| 1007 | /* Write the aPgno[] array entry and the hash-table slot. */ |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 1008 | nCollide = idx; |
dan | 6f15014 | 2010-05-21 15:31:56 +0000 | [diff] [blame] | 1009 | for(iKey=walHash(iPage); aHash[iKey]; iKey=walNextHash(iKey)){ |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 1010 | if( (nCollide--)==0 ) return SQLITE_CORRUPT_BKPT; |
drh | 29d4dbe | 2010-05-18 23:29:52 +0000 | [diff] [blame] | 1011 | } |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 1012 | aPgno[idx] = iPage; |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 1013 | aHash[iKey] = (ht_slot)idx; |
drh | 4fa95bf | 2010-05-22 00:55:39 +0000 | [diff] [blame] | 1014 | |
| 1015 | #ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT |
| 1016 | /* Verify that the number of entries in the hash table exactly equals |
| 1017 | ** the number of entries in the mapping region. |
| 1018 | */ |
| 1019 | { |
| 1020 | int i; /* Loop counter */ |
| 1021 | int nEntry = 0; /* Number of entries in the hash table */ |
| 1022 | for(i=0; i<HASHTABLE_NSLOT; i++){ if( aHash[i] ) nEntry++; } |
| 1023 | assert( nEntry==idx ); |
| 1024 | } |
| 1025 | |
| 1026 | /* Verify that the every entry in the mapping region is reachable |
| 1027 | ** via the hash table. This turns out to be a really, really expensive |
| 1028 | ** thing to check, so only do this occasionally - not on every |
| 1029 | ** iteration. |
| 1030 | */ |
| 1031 | if( (idx&0x3ff)==0 ){ |
| 1032 | int i; /* Loop counter */ |
| 1033 | for(i=1; i<=idx; i++){ |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 1034 | for(iKey=walHash(aPgno[i]); aHash[iKey]; iKey=walNextHash(iKey)){ |
drh | 4fa95bf | 2010-05-22 00:55:39 +0000 | [diff] [blame] | 1035 | if( aHash[iKey]==i ) break; |
| 1036 | } |
| 1037 | assert( aHash[iKey]==i ); |
| 1038 | } |
| 1039 | } |
| 1040 | #endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1041 | } |
dan | 31f98fc | 2010-04-27 05:42:32 +0000 | [diff] [blame] | 1042 | |
drh | 4fa95bf | 2010-05-22 00:55:39 +0000 | [diff] [blame] | 1043 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 1044 | return rc; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1045 | } |
| 1046 | |
| 1047 | |
| 1048 | /* |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1049 | ** Recover the wal-index by reading the write-ahead log file. |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1050 | ** |
| 1051 | ** This routine first tries to establish an exclusive lock on the |
| 1052 | ** wal-index to prevent other threads/processes from doing anything |
| 1053 | ** with the WAL or wal-index while recovery is running. The |
| 1054 | ** WAL_RECOVER_LOCK is also held so that other threads will know |
| 1055 | ** that this thread is running recovery. If unable to establish |
| 1056 | ** the necessary locks, this routine returns SQLITE_BUSY. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1057 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1058 | static int walIndexRecover(Wal *pWal){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1059 | int rc; /* Return Code */ |
| 1060 | i64 nSize; /* Size of log file */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1061 | u32 aFrameCksum[2] = {0, 0}; |
dan | d0aa342 | 2010-05-31 16:41:53 +0000 | [diff] [blame] | 1062 | int iLock; /* Lock offset to lock for checkpoint */ |
| 1063 | int nLock; /* Number of locks to hold */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1064 | |
dan | d0aa342 | 2010-05-31 16:41:53 +0000 | [diff] [blame] | 1065 | /* Obtain an exclusive lock on all byte in the locking range not already |
| 1066 | ** locked by the caller. The caller is guaranteed to have locked the |
| 1067 | ** WAL_WRITE_LOCK byte, and may have also locked the WAL_CKPT_LOCK byte. |
| 1068 | ** If successful, the same bytes that are locked here are unlocked before |
| 1069 | ** this function returns. |
| 1070 | */ |
| 1071 | assert( pWal->ckptLock==1 || pWal->ckptLock==0 ); |
| 1072 | assert( WAL_ALL_BUT_WRITE==WAL_WRITE_LOCK+1 ); |
| 1073 | assert( WAL_CKPT_LOCK==WAL_ALL_BUT_WRITE ); |
| 1074 | assert( pWal->writeLock ); |
| 1075 | iLock = WAL_ALL_BUT_WRITE + pWal->ckptLock; |
| 1076 | nLock = SQLITE_SHM_NLOCK - iLock; |
| 1077 | rc = walLockExclusive(pWal, iLock, nLock); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1078 | if( rc ){ |
| 1079 | return rc; |
| 1080 | } |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 1081 | WALTRACE(("WAL%p: recovery begin...\n", pWal)); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1082 | |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1083 | memset(&pWal->hdr, 0, sizeof(WalIndexHdr)); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1084 | |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1085 | rc = sqlite3OsFileSize(pWal->pWalFd, &nSize); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1086 | if( rc!=SQLITE_OK ){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1087 | goto recovery_error; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1088 | } |
| 1089 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1090 | if( nSize>WAL_HDRSIZE ){ |
| 1091 | u8 aBuf[WAL_HDRSIZE]; /* Buffer to load WAL header into */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1092 | u8 *aFrame = 0; /* Malloc'd buffer to load entire frame */ |
drh | 584c754 | 2010-05-19 18:08:10 +0000 | [diff] [blame] | 1093 | int szFrame; /* Number of bytes in buffer aFrame[] */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1094 | u8 *aData; /* Pointer to data part of aFrame buffer */ |
| 1095 | int iFrame; /* Index of last frame read */ |
| 1096 | i64 iOffset; /* Next offset to read from log file */ |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 1097 | int szPage; /* Page size according to the log */ |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1098 | u32 magic; /* Magic value read from WAL header */ |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 1099 | u32 version; /* Magic value read from WAL header */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1100 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1101 | /* Read in the WAL header. */ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1102 | rc = sqlite3OsRead(pWal->pWalFd, aBuf, WAL_HDRSIZE, 0); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1103 | if( rc!=SQLITE_OK ){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1104 | goto recovery_error; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1105 | } |
| 1106 | |
| 1107 | /* If the database page size is not a power of two, or is greater than |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1108 | ** SQLITE_MAX_PAGE_SIZE, conclude that the WAL file contains no valid |
| 1109 | ** data. Similarly, if the 'magic' value is invalid, ignore the whole |
| 1110 | ** WAL file. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1111 | */ |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1112 | magic = sqlite3Get4byte(&aBuf[0]); |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 1113 | szPage = sqlite3Get4byte(&aBuf[8]); |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1114 | if( (magic&0xFFFFFFFE)!=WAL_MAGIC |
| 1115 | || szPage&(szPage-1) |
| 1116 | || szPage>SQLITE_MAX_PAGE_SIZE |
| 1117 | || szPage<512 |
| 1118 | ){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1119 | goto finished; |
| 1120 | } |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 1121 | pWal->hdr.bigEndCksum = (u8)(magic&0x00000001); |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 1122 | pWal->szPage = szPage; |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 1123 | pWal->nCkpt = sqlite3Get4byte(&aBuf[12]); |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 1124 | memcpy(&pWal->hdr.aSalt, &aBuf[16], 8); |
drh | cd28508 | 2010-06-23 22:00:35 +0000 | [diff] [blame] | 1125 | |
| 1126 | /* Verify that the WAL header checksum is correct */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1127 | walChecksumBytes(pWal->hdr.bigEndCksum==SQLITE_BIGENDIAN, |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 1128 | aBuf, WAL_HDRSIZE-2*4, 0, pWal->hdr.aFrameCksum |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1129 | ); |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 1130 | if( pWal->hdr.aFrameCksum[0]!=sqlite3Get4byte(&aBuf[24]) |
| 1131 | || pWal->hdr.aFrameCksum[1]!=sqlite3Get4byte(&aBuf[28]) |
| 1132 | ){ |
| 1133 | goto finished; |
| 1134 | } |
| 1135 | |
drh | cd28508 | 2010-06-23 22:00:35 +0000 | [diff] [blame] | 1136 | /* Verify that the version number on the WAL format is one that |
| 1137 | ** are able to understand */ |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 1138 | version = sqlite3Get4byte(&aBuf[4]); |
| 1139 | if( version!=WAL_MAX_VERSION ){ |
| 1140 | rc = SQLITE_CANTOPEN_BKPT; |
| 1141 | goto finished; |
| 1142 | } |
| 1143 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1144 | /* Malloc a buffer to read frames into. */ |
drh | 584c754 | 2010-05-19 18:08:10 +0000 | [diff] [blame] | 1145 | szFrame = szPage + WAL_FRAME_HDRSIZE; |
| 1146 | aFrame = (u8 *)sqlite3_malloc(szFrame); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1147 | if( !aFrame ){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1148 | rc = SQLITE_NOMEM; |
| 1149 | goto recovery_error; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1150 | } |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1151 | aData = &aFrame[WAL_FRAME_HDRSIZE]; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1152 | |
| 1153 | /* Read all frames from the log file. */ |
| 1154 | iFrame = 0; |
drh | 584c754 | 2010-05-19 18:08:10 +0000 | [diff] [blame] | 1155 | for(iOffset=WAL_HDRSIZE; (iOffset+szFrame)<=nSize; iOffset+=szFrame){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1156 | u32 pgno; /* Database page number for frame */ |
| 1157 | u32 nTruncate; /* dbsize field from frame header */ |
| 1158 | int isValid; /* True if this frame is valid */ |
| 1159 | |
| 1160 | /* Read and decode the next log frame. */ |
drh | 584c754 | 2010-05-19 18:08:10 +0000 | [diff] [blame] | 1161 | rc = sqlite3OsRead(pWal->pWalFd, aFrame, szFrame, iOffset); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1162 | if( rc!=SQLITE_OK ) break; |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 1163 | isValid = walDecodeFrame(pWal, &pgno, &nTruncate, aData, aFrame); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1164 | if( !isValid ) break; |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 1165 | rc = walIndexAppend(pWal, ++iFrame, pgno); |
| 1166 | if( rc!=SQLITE_OK ) break; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1167 | |
| 1168 | /* If nTruncate is non-zero, this is a commit record. */ |
| 1169 | if( nTruncate ){ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1170 | pWal->hdr.mxFrame = iFrame; |
| 1171 | pWal->hdr.nPage = nTruncate; |
shaneh | 1df2db7 | 2010-08-18 02:28:48 +0000 | [diff] [blame] | 1172 | pWal->hdr.szPage = (u16)((szPage&0xff00) | (szPage>>16)); |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 1173 | testcase( szPage<=32768 ); |
| 1174 | testcase( szPage>=65536 ); |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1175 | aFrameCksum[0] = pWal->hdr.aFrameCksum[0]; |
| 1176 | aFrameCksum[1] = pWal->hdr.aFrameCksum[1]; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1177 | } |
| 1178 | } |
| 1179 | |
| 1180 | sqlite3_free(aFrame); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1181 | } |
| 1182 | |
| 1183 | finished: |
dan | 576bc32 | 2010-05-06 18:04:50 +0000 | [diff] [blame] | 1184 | if( rc==SQLITE_OK ){ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 1185 | volatile WalCkptInfo *pInfo; |
| 1186 | int i; |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 1187 | pWal->hdr.aFrameCksum[0] = aFrameCksum[0]; |
| 1188 | pWal->hdr.aFrameCksum[1] = aFrameCksum[1]; |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 1189 | walIndexWriteHdr(pWal); |
dan | 3dee6da | 2010-05-31 16:17:54 +0000 | [diff] [blame] | 1190 | |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 1191 | /* Reset the checkpoint-header. This is safe because this thread is |
dan | 3dee6da | 2010-05-31 16:17:54 +0000 | [diff] [blame] | 1192 | ** currently holding locks that exclude all other readers, writers and |
| 1193 | ** checkpointers. |
| 1194 | */ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 1195 | pInfo = walCkptInfo(pWal); |
| 1196 | pInfo->nBackfill = 0; |
| 1197 | pInfo->aReadMark[0] = 0; |
| 1198 | for(i=1; i<WAL_NREADER; i++) pInfo->aReadMark[i] = READMARK_NOT_USED; |
dan | eb8763d | 2010-08-17 14:52:22 +0000 | [diff] [blame] | 1199 | |
| 1200 | /* If more than one frame was recovered from the log file, report an |
| 1201 | ** event via sqlite3_log(). This is to help with identifying performance |
| 1202 | ** problems caused by applications routinely shutting down without |
| 1203 | ** checkpointing the log file. |
| 1204 | */ |
| 1205 | if( pWal->hdr.nPage ){ |
| 1206 | sqlite3_log(SQLITE_OK, "Recovered %d frames from WAL file %s", |
| 1207 | pWal->hdr.nPage, pWal->zWalName |
| 1208 | ); |
| 1209 | } |
dan | 576bc32 | 2010-05-06 18:04:50 +0000 | [diff] [blame] | 1210 | } |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1211 | |
| 1212 | recovery_error: |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 1213 | WALTRACE(("WAL%p: recovery %s\n", pWal, rc ? "failed" : "ok")); |
dan | d0aa342 | 2010-05-31 16:41:53 +0000 | [diff] [blame] | 1214 | walUnlockExclusive(pWal, iLock, nLock); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1215 | return rc; |
| 1216 | } |
| 1217 | |
drh | a8e654e | 2010-05-04 17:38:42 +0000 | [diff] [blame] | 1218 | /* |
dan | 1018e90 | 2010-05-05 15:33:05 +0000 | [diff] [blame] | 1219 | ** Close an open wal-index. |
drh | a8e654e | 2010-05-04 17:38:42 +0000 | [diff] [blame] | 1220 | */ |
dan | 1018e90 | 2010-05-05 15:33:05 +0000 | [diff] [blame] | 1221 | static void walIndexClose(Wal *pWal, int isDelete){ |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 1222 | if( pWal->exclusiveMode==WAL_HEAPMEMORY_MODE ){ |
| 1223 | int i; |
| 1224 | for(i=0; i<pWal->nWiData; i++){ |
| 1225 | sqlite3_free((void *)pWal->apWiData[i]); |
| 1226 | pWal->apWiData[i] = 0; |
| 1227 | } |
| 1228 | }else{ |
| 1229 | sqlite3OsShmUnmap(pWal->pDbFd, isDelete); |
| 1230 | } |
drh | a8e654e | 2010-05-04 17:38:42 +0000 | [diff] [blame] | 1231 | } |
| 1232 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1233 | /* |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 1234 | ** Open a connection to the WAL file zWalName. The database file must |
| 1235 | ** already be opened on connection pDbFd. The buffer that zWalName points |
| 1236 | ** to must remain valid for the lifetime of the returned Wal* handle. |
dan | 3de777f | 2010-04-17 12:31:37 +0000 | [diff] [blame] | 1237 | ** |
| 1238 | ** A SHARED lock should be held on the database file when this function |
| 1239 | ** is called. The purpose of this SHARED lock is to prevent any other |
drh | 181e091 | 2010-06-01 01:08:08 +0000 | [diff] [blame] | 1240 | ** client from unlinking the WAL or wal-index file. If another process |
dan | 3de777f | 2010-04-17 12:31:37 +0000 | [diff] [blame] | 1241 | ** were to do this just after this client opened one of these files, the |
| 1242 | ** system would be badly broken. |
dan | ef37802 | 2010-05-04 11:06:03 +0000 | [diff] [blame] | 1243 | ** |
| 1244 | ** If the log file is successfully opened, SQLITE_OK is returned and |
| 1245 | ** *ppWal is set to point to a new WAL handle. If an error occurs, |
| 1246 | ** an SQLite error code is returned and *ppWal is left unmodified. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1247 | */ |
drh | c438efd | 2010-04-26 00:19:45 +0000 | [diff] [blame] | 1248 | int sqlite3WalOpen( |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1249 | sqlite3_vfs *pVfs, /* vfs module to open wal and wal-index */ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1250 | sqlite3_file *pDbFd, /* The open database file */ |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 1251 | const char *zWalName, /* Name of the WAL file */ |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 1252 | int bNoShm, /* True to run in heap-memory mode */ |
drh | 85a8375 | 2011-05-16 21:00:27 +0000 | [diff] [blame] | 1253 | i64 mxWalSize, /* Truncate WAL to this size on reset */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1254 | Wal **ppWal /* OUT: Allocated Wal handle */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1255 | ){ |
dan | ef37802 | 2010-05-04 11:06:03 +0000 | [diff] [blame] | 1256 | int rc; /* Return Code */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1257 | Wal *pRet; /* Object to allocate and return */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1258 | int flags; /* Flags passed to OsOpen() */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1259 | |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 1260 | assert( zWalName && zWalName[0] ); |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1261 | assert( pDbFd ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1262 | |
drh | 1b78eaf | 2010-05-25 13:40:03 +0000 | [diff] [blame] | 1263 | /* In the amalgamation, the os_unix.c and os_win.c source files come before |
| 1264 | ** this source file. Verify that the #defines of the locking byte offsets |
| 1265 | ** in os_unix.c and os_win.c agree with the WALINDEX_LOCK_OFFSET value. |
| 1266 | */ |
| 1267 | #ifdef WIN_SHM_BASE |
| 1268 | assert( WIN_SHM_BASE==WALINDEX_LOCK_OFFSET ); |
| 1269 | #endif |
| 1270 | #ifdef UNIX_SHM_BASE |
| 1271 | assert( UNIX_SHM_BASE==WALINDEX_LOCK_OFFSET ); |
| 1272 | #endif |
| 1273 | |
| 1274 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1275 | /* Allocate an instance of struct Wal to return. */ |
| 1276 | *ppWal = 0; |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 1277 | pRet = (Wal*)sqlite3MallocZero(sizeof(Wal) + pVfs->szOsFile); |
dan | 76ed3bc | 2010-05-03 17:18:24 +0000 | [diff] [blame] | 1278 | if( !pRet ){ |
| 1279 | return SQLITE_NOMEM; |
| 1280 | } |
| 1281 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1282 | pRet->pVfs = pVfs; |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1283 | pRet->pWalFd = (sqlite3_file *)&pRet[1]; |
| 1284 | pRet->pDbFd = pDbFd; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1285 | pRet->readLock = -1; |
drh | 85a8375 | 2011-05-16 21:00:27 +0000 | [diff] [blame] | 1286 | pRet->mxWalSize = mxWalSize; |
dan | 3e875ef | 2010-07-05 19:03:35 +0000 | [diff] [blame] | 1287 | pRet->zWalName = zWalName; |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 1288 | pRet->exclusiveMode = (bNoShm ? WAL_HEAPMEMORY_MODE: WAL_NORMAL_MODE); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1289 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1290 | /* Open file handle on the write-ahead log file. */ |
dan | ddb0ac4 | 2010-07-14 14:48:58 +0000 | [diff] [blame] | 1291 | flags = (SQLITE_OPEN_READWRITE|SQLITE_OPEN_CREATE|SQLITE_OPEN_WAL); |
dan | da9fe0c | 2010-07-13 18:44:03 +0000 | [diff] [blame] | 1292 | rc = sqlite3OsOpen(pVfs, zWalName, pRet->pWalFd, flags, &flags); |
dan | 50833e3 | 2010-07-14 16:37:17 +0000 | [diff] [blame] | 1293 | if( rc==SQLITE_OK && flags&SQLITE_OPEN_READONLY ){ |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 1294 | pRet->readOnly = WAL_RDONLY; |
dan | 50833e3 | 2010-07-14 16:37:17 +0000 | [diff] [blame] | 1295 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1296 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1297 | if( rc!=SQLITE_OK ){ |
dan | 1018e90 | 2010-05-05 15:33:05 +0000 | [diff] [blame] | 1298 | walIndexClose(pRet, 0); |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1299 | sqlite3OsClose(pRet->pWalFd); |
dan | ef37802 | 2010-05-04 11:06:03 +0000 | [diff] [blame] | 1300 | sqlite3_free(pRet); |
| 1301 | }else{ |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 1302 | int iDC = sqlite3OsDeviceCharacteristics(pRet->pWalFd); |
| 1303 | if( iDC & SQLITE_IOCAP_SEQUENTIAL ){ pRet->noSyncHeader = 1; } |
dan | ef37802 | 2010-05-04 11:06:03 +0000 | [diff] [blame] | 1304 | *ppWal = pRet; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 1305 | WALTRACE(("WAL%d: opened\n", pRet)); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1306 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1307 | return rc; |
| 1308 | } |
| 1309 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1310 | /* |
drh | 85a8375 | 2011-05-16 21:00:27 +0000 | [diff] [blame] | 1311 | ** Change the size to which the WAL file is trucated on each reset. |
| 1312 | */ |
| 1313 | void sqlite3WalLimit(Wal *pWal, i64 iLimit){ |
| 1314 | if( pWal ) pWal->mxWalSize = iLimit; |
| 1315 | } |
| 1316 | |
| 1317 | /* |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1318 | ** Find the smallest page number out of all pages held in the WAL that |
| 1319 | ** has not been returned by any prior invocation of this method on the |
| 1320 | ** same WalIterator object. Write into *piFrame the frame index where |
| 1321 | ** that page was last written into the WAL. Write into *piPage the page |
| 1322 | ** number. |
| 1323 | ** |
| 1324 | ** Return 0 on success. If there are no pages in the WAL with a page |
| 1325 | ** number larger than *piPage, then return 1. |
| 1326 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1327 | static int walIteratorNext( |
| 1328 | WalIterator *p, /* Iterator */ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1329 | u32 *piPage, /* OUT: The page number of the next page */ |
| 1330 | u32 *piFrame /* OUT: Wal frame index of next page */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1331 | ){ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1332 | u32 iMin; /* Result pgno must be greater than iMin */ |
| 1333 | u32 iRet = 0xFFFFFFFF; /* 0xffffffff is never a valid page number */ |
| 1334 | int i; /* For looping through segments */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1335 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1336 | iMin = p->iPrior; |
| 1337 | assert( iMin<0xffffffff ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1338 | for(i=p->nSegment-1; i>=0; i--){ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1339 | struct WalSegment *pSegment = &p->aSegment[i]; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1340 | while( pSegment->iNext<pSegment->nEntry ){ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1341 | u32 iPg = pSegment->aPgno[pSegment->aIndex[pSegment->iNext]]; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1342 | if( iPg>iMin ){ |
| 1343 | if( iPg<iRet ){ |
| 1344 | iRet = iPg; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1345 | *piFrame = pSegment->iZero + pSegment->aIndex[pSegment->iNext]; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1346 | } |
| 1347 | break; |
| 1348 | } |
| 1349 | pSegment->iNext++; |
| 1350 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1351 | } |
| 1352 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1353 | *piPage = p->iPrior = iRet; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1354 | return (iRet==0xFFFFFFFF); |
| 1355 | } |
| 1356 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1357 | /* |
| 1358 | ** This function merges two sorted lists into a single sorted list. |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 1359 | ** |
| 1360 | ** aLeft[] and aRight[] are arrays of indices. The sort key is |
| 1361 | ** aContent[aLeft[]] and aContent[aRight[]]. Upon entry, the following |
| 1362 | ** is guaranteed for all J<K: |
| 1363 | ** |
| 1364 | ** aContent[aLeft[J]] < aContent[aLeft[K]] |
| 1365 | ** aContent[aRight[J]] < aContent[aRight[K]] |
| 1366 | ** |
| 1367 | ** This routine overwrites aRight[] with a new (probably longer) sequence |
| 1368 | ** of indices such that the aRight[] contains every index that appears in |
| 1369 | ** either aLeft[] or the old aRight[] and such that the second condition |
| 1370 | ** above is still met. |
| 1371 | ** |
| 1372 | ** The aContent[aLeft[X]] values will be unique for all X. And the |
| 1373 | ** aContent[aRight[X]] values will be unique too. But there might be |
| 1374 | ** one or more combinations of X and Y such that |
| 1375 | ** |
| 1376 | ** aLeft[X]!=aRight[Y] && aContent[aLeft[X]] == aContent[aRight[Y]] |
| 1377 | ** |
| 1378 | ** When that happens, omit the aLeft[X] and use the aRight[Y] index. |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1379 | */ |
| 1380 | static void walMerge( |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 1381 | const u32 *aContent, /* Pages in wal - keys for the sort */ |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1382 | ht_slot *aLeft, /* IN: Left hand input list */ |
| 1383 | int nLeft, /* IN: Elements in array *paLeft */ |
| 1384 | ht_slot **paRight, /* IN/OUT: Right hand input list */ |
| 1385 | int *pnRight, /* IN/OUT: Elements in *paRight */ |
| 1386 | ht_slot *aTmp /* Temporary buffer */ |
| 1387 | ){ |
| 1388 | int iLeft = 0; /* Current index in aLeft */ |
| 1389 | int iRight = 0; /* Current index in aRight */ |
| 1390 | int iOut = 0; /* Current index in output buffer */ |
| 1391 | int nRight = *pnRight; |
| 1392 | ht_slot *aRight = *paRight; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1393 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1394 | assert( nLeft>0 && nRight>0 ); |
| 1395 | while( iRight<nRight || iLeft<nLeft ){ |
| 1396 | ht_slot logpage; |
| 1397 | Pgno dbpage; |
| 1398 | |
| 1399 | if( (iLeft<nLeft) |
| 1400 | && (iRight>=nRight || aContent[aLeft[iLeft]]<aContent[aRight[iRight]]) |
| 1401 | ){ |
| 1402 | logpage = aLeft[iLeft++]; |
| 1403 | }else{ |
| 1404 | logpage = aRight[iRight++]; |
| 1405 | } |
| 1406 | dbpage = aContent[logpage]; |
| 1407 | |
| 1408 | aTmp[iOut++] = logpage; |
| 1409 | if( iLeft<nLeft && aContent[aLeft[iLeft]]==dbpage ) iLeft++; |
| 1410 | |
| 1411 | assert( iLeft>=nLeft || aContent[aLeft[iLeft]]>dbpage ); |
| 1412 | assert( iRight>=nRight || aContent[aRight[iRight]]>dbpage ); |
| 1413 | } |
| 1414 | |
| 1415 | *paRight = aLeft; |
| 1416 | *pnRight = iOut; |
| 1417 | memcpy(aLeft, aTmp, sizeof(aTmp[0])*iOut); |
| 1418 | } |
| 1419 | |
| 1420 | /* |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 1421 | ** Sort the elements in list aList using aContent[] as the sort key. |
| 1422 | ** Remove elements with duplicate keys, preferring to keep the |
| 1423 | ** larger aList[] values. |
| 1424 | ** |
| 1425 | ** The aList[] entries are indices into aContent[]. The values in |
| 1426 | ** aList[] are to be sorted so that for all J<K: |
| 1427 | ** |
| 1428 | ** aContent[aList[J]] < aContent[aList[K]] |
| 1429 | ** |
| 1430 | ** For any X and Y such that |
| 1431 | ** |
| 1432 | ** aContent[aList[X]] == aContent[aList[Y]] |
| 1433 | ** |
| 1434 | ** Keep the larger of the two values aList[X] and aList[Y] and discard |
| 1435 | ** the smaller. |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1436 | */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1437 | static void walMergesort( |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 1438 | const u32 *aContent, /* Pages in wal */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 1439 | ht_slot *aBuffer, /* Buffer of at least *pnList items to use */ |
| 1440 | ht_slot *aList, /* IN/OUT: List to sort */ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1441 | int *pnList /* IN/OUT: Number of elements in aList[] */ |
| 1442 | ){ |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1443 | struct Sublist { |
| 1444 | int nList; /* Number of elements in aList */ |
| 1445 | ht_slot *aList; /* Pointer to sub-list content */ |
| 1446 | }; |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1447 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1448 | const int nList = *pnList; /* Size of input list */ |
drh | ff82894 | 2010-06-26 21:34:06 +0000 | [diff] [blame] | 1449 | int nMerge = 0; /* Number of elements in list aMerge */ |
| 1450 | ht_slot *aMerge = 0; /* List to be merged */ |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1451 | int iList; /* Index into input list */ |
drh | 7d113eb | 2010-06-26 20:00:54 +0000 | [diff] [blame] | 1452 | int iSub = 0; /* Index into aSub array */ |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1453 | struct Sublist aSub[13]; /* Array of sub-lists */ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1454 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1455 | memset(aSub, 0, sizeof(aSub)); |
| 1456 | assert( nList<=HASHTABLE_NPAGE && nList>0 ); |
| 1457 | assert( HASHTABLE_NPAGE==(1<<(ArraySize(aSub)-1)) ); |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1458 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1459 | for(iList=0; iList<nList; iList++){ |
| 1460 | nMerge = 1; |
| 1461 | aMerge = &aList[iList]; |
| 1462 | for(iSub=0; iList & (1<<iSub); iSub++){ |
| 1463 | struct Sublist *p = &aSub[iSub]; |
| 1464 | assert( p->aList && p->nList<=(1<<iSub) ); |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1465 | assert( p->aList==&aList[iList&~((2<<iSub)-1)] ); |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1466 | walMerge(aContent, p->aList, p->nList, &aMerge, &nMerge, aBuffer); |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1467 | } |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1468 | aSub[iSub].aList = aMerge; |
| 1469 | aSub[iSub].nList = nMerge; |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1470 | } |
| 1471 | |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1472 | for(iSub++; iSub<ArraySize(aSub); iSub++){ |
| 1473 | if( nList & (1<<iSub) ){ |
| 1474 | struct Sublist *p = &aSub[iSub]; |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1475 | assert( p->nList<=(1<<iSub) ); |
| 1476 | assert( p->aList==&aList[nList&~((2<<iSub)-1)] ); |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1477 | walMerge(aContent, p->aList, p->nList, &aMerge, &nMerge, aBuffer); |
| 1478 | } |
| 1479 | } |
| 1480 | assert( aMerge==aList ); |
| 1481 | *pnList = nMerge; |
| 1482 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1483 | #ifdef SQLITE_DEBUG |
| 1484 | { |
| 1485 | int i; |
| 1486 | for(i=1; i<*pnList; i++){ |
| 1487 | assert( aContent[aList[i]] > aContent[aList[i-1]] ); |
| 1488 | } |
| 1489 | } |
| 1490 | #endif |
| 1491 | } |
| 1492 | |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 1493 | /* |
| 1494 | ** Free an iterator allocated by walIteratorInit(). |
| 1495 | */ |
| 1496 | static void walIteratorFree(WalIterator *p){ |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1497 | sqlite3ScratchFree(p); |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 1498 | } |
| 1499 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1500 | /* |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1501 | ** Construct a WalInterator object that can be used to loop over all |
| 1502 | ** pages in the WAL in ascending order. The caller must hold the checkpoint |
drh | d9c9b78 | 2010-12-15 21:02:06 +0000 | [diff] [blame] | 1503 | ** lock. |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1504 | ** |
| 1505 | ** On success, make *pp point to the newly allocated WalInterator object |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1506 | ** return SQLITE_OK. Otherwise, return an error code. If this routine |
| 1507 | ** returns an error, the value of *pp is undefined. |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1508 | ** |
| 1509 | ** The calling routine should invoke walIteratorFree() to destroy the |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1510 | ** WalIterator object when it has finished with it. |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1511 | */ |
| 1512 | static int walIteratorInit(Wal *pWal, WalIterator **pp){ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 1513 | WalIterator *p; /* Return value */ |
| 1514 | int nSegment; /* Number of segments to merge */ |
| 1515 | u32 iLast; /* Last frame in log */ |
| 1516 | int nByte; /* Number of bytes to allocate */ |
| 1517 | int i; /* Iterator variable */ |
| 1518 | ht_slot *aTmp; /* Temp space used by merge-sort */ |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1519 | int rc = SQLITE_OK; /* Return Code */ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1520 | |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1521 | /* This routine only runs while holding the checkpoint lock. And |
| 1522 | ** it only runs if there is actually content in the log (mxFrame>0). |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1523 | */ |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1524 | assert( pWal->ckptLock && pWal->hdr.mxFrame>0 ); |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1525 | iLast = pWal->hdr.mxFrame; |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1526 | |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1527 | /* Allocate space for the WalIterator object. */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1528 | nSegment = walFramePage(iLast) + 1; |
| 1529 | nByte = sizeof(WalIterator) |
dan | 52d6fc0 | 2010-06-25 16:34:32 +0000 | [diff] [blame] | 1530 | + (nSegment-1)*sizeof(struct WalSegment) |
| 1531 | + iLast*sizeof(ht_slot); |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1532 | p = (WalIterator *)sqlite3ScratchMalloc(nByte); |
dan | 8f6097c | 2010-05-06 07:43:58 +0000 | [diff] [blame] | 1533 | if( !p ){ |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1534 | return SQLITE_NOMEM; |
| 1535 | } |
| 1536 | memset(p, 0, nByte); |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1537 | p->nSegment = nSegment; |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1538 | |
| 1539 | /* Allocate temporary space used by the merge-sort routine. This block |
| 1540 | ** of memory will be freed before this function returns. |
| 1541 | */ |
dan | 52d6fc0 | 2010-06-25 16:34:32 +0000 | [diff] [blame] | 1542 | aTmp = (ht_slot *)sqlite3ScratchMalloc( |
| 1543 | sizeof(ht_slot) * (iLast>HASHTABLE_NPAGE?HASHTABLE_NPAGE:iLast) |
| 1544 | ); |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1545 | if( !aTmp ){ |
| 1546 | rc = SQLITE_NOMEM; |
| 1547 | } |
| 1548 | |
| 1549 | for(i=0; rc==SQLITE_OK && i<nSegment; i++){ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 1550 | volatile ht_slot *aHash; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1551 | u32 iZero; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1552 | volatile u32 *aPgno; |
| 1553 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1554 | rc = walHashGet(pWal, i, &aHash, &aPgno, &iZero); |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1555 | if( rc==SQLITE_OK ){ |
dan | 52d6fc0 | 2010-06-25 16:34:32 +0000 | [diff] [blame] | 1556 | int j; /* Counter variable */ |
| 1557 | int nEntry; /* Number of entries in this segment */ |
| 1558 | ht_slot *aIndex; /* Sorted index for this segment */ |
| 1559 | |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1560 | aPgno++; |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 1561 | if( (i+1)==nSegment ){ |
| 1562 | nEntry = (int)(iLast - iZero); |
| 1563 | }else{ |
shaneh | 5589796 | 2010-07-09 12:57:53 +0000 | [diff] [blame] | 1564 | nEntry = (int)((u32*)aHash - (u32*)aPgno); |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 1565 | } |
dan | 52d6fc0 | 2010-06-25 16:34:32 +0000 | [diff] [blame] | 1566 | aIndex = &((ht_slot *)&p->aSegment[p->nSegment])[iZero]; |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1567 | iZero++; |
| 1568 | |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1569 | for(j=0; j<nEntry; j++){ |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 1570 | aIndex[j] = (ht_slot)j; |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1571 | } |
| 1572 | walMergesort((u32 *)aPgno, aTmp, aIndex, &nEntry); |
| 1573 | p->aSegment[i].iZero = iZero; |
| 1574 | p->aSegment[i].nEntry = nEntry; |
| 1575 | p->aSegment[i].aIndex = aIndex; |
| 1576 | p->aSegment[i].aPgno = (u32 *)aPgno; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1577 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1578 | } |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1579 | sqlite3ScratchFree(aTmp); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1580 | |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1581 | if( rc!=SQLITE_OK ){ |
| 1582 | walIteratorFree(p); |
| 1583 | } |
dan | 8f6097c | 2010-05-06 07:43:58 +0000 | [diff] [blame] | 1584 | *pp = p; |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1585 | return rc; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1586 | } |
| 1587 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1588 | /* |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 1589 | ** Attempt to obtain the exclusive WAL lock defined by parameters lockIdx and |
| 1590 | ** n. If the attempt fails and parameter xBusy is not NULL, then it is a |
| 1591 | ** busy-handler function. Invoke it and retry the lock until either the |
| 1592 | ** lock is successfully obtained or the busy-handler returns 0. |
| 1593 | */ |
| 1594 | static int walBusyLock( |
| 1595 | Wal *pWal, /* WAL connection */ |
| 1596 | int (*xBusy)(void*), /* Function to call when busy */ |
| 1597 | void *pBusyArg, /* Context argument for xBusyHandler */ |
| 1598 | int lockIdx, /* Offset of first byte to lock */ |
| 1599 | int n /* Number of bytes to lock */ |
| 1600 | ){ |
| 1601 | int rc; |
| 1602 | do { |
| 1603 | rc = walLockExclusive(pWal, lockIdx, n); |
| 1604 | }while( xBusy && rc==SQLITE_BUSY && xBusy(pBusyArg) ); |
| 1605 | return rc; |
| 1606 | } |
| 1607 | |
| 1608 | /* |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1609 | ** The cache of the wal-index header must be valid to call this function. |
| 1610 | ** Return the page-size in bytes used by the database. |
| 1611 | */ |
| 1612 | static int walPagesize(Wal *pWal){ |
| 1613 | return (pWal->hdr.szPage&0xfe00) + ((pWal->hdr.szPage&0x0001)<<16); |
| 1614 | } |
| 1615 | |
| 1616 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1617 | ** Copy as much content as we can from the WAL back into the database file |
| 1618 | ** in response to an sqlite3_wal_checkpoint() request or the equivalent. |
| 1619 | ** |
| 1620 | ** The amount of information copies from WAL to database might be limited |
| 1621 | ** by active readers. This routine will never overwrite a database page |
| 1622 | ** that a concurrent reader might be using. |
| 1623 | ** |
| 1624 | ** All I/O barrier operations (a.k.a fsyncs) occur in this routine when |
| 1625 | ** SQLite is in WAL-mode in synchronous=NORMAL. That means that if |
| 1626 | ** checkpoints are always run by a background thread or background |
| 1627 | ** process, foreground threads will never block on a lengthy fsync call. |
| 1628 | ** |
| 1629 | ** Fsync is called on the WAL before writing content out of the WAL and |
| 1630 | ** into the database. This ensures that if the new content is persistent |
| 1631 | ** in the WAL and can be recovered following a power-loss or hard reset. |
| 1632 | ** |
| 1633 | ** Fsync is also called on the database file if (and only if) the entire |
| 1634 | ** WAL content is copied into the database file. This second fsync makes |
| 1635 | ** it safe to delete the WAL since the new content will persist in the |
| 1636 | ** database file. |
| 1637 | ** |
| 1638 | ** This routine uses and updates the nBackfill field of the wal-index header. |
| 1639 | ** This is the only routine tha will increase the value of nBackfill. |
| 1640 | ** (A WAL reset or recovery will revert nBackfill to zero, but not increase |
| 1641 | ** its value.) |
| 1642 | ** |
| 1643 | ** The caller must be holding sufficient locks to ensure that no other |
| 1644 | ** checkpoint is running (in any other thread or process) at the same |
| 1645 | ** time. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1646 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1647 | static int walCheckpoint( |
| 1648 | Wal *pWal, /* Wal connection */ |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 1649 | int eMode, /* One of PASSIVE, FULL or RESTART */ |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1650 | int (*xBusyCall)(void*), /* Function to call when busy */ |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 1651 | void *pBusyArg, /* Context argument for xBusyHandler */ |
dan | c511878 | 2010-04-17 17:34:41 +0000 | [diff] [blame] | 1652 | int sync_flags, /* Flags for OsSync() (or 0) */ |
dan | 9c5e368 | 2011-02-07 15:12:12 +0000 | [diff] [blame] | 1653 | u8 *zBuf /* Temporary buffer to use */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1654 | ){ |
| 1655 | int rc; /* Return code */ |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 1656 | int szPage; /* Database page-size */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1657 | WalIterator *pIter = 0; /* Wal iterator context */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1658 | u32 iDbpage = 0; /* Next database page to write */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1659 | u32 iFrame = 0; /* Wal frame containing data for iDbpage */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1660 | u32 mxSafeFrame; /* Max frame that can be backfilled */ |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1661 | u32 mxPage; /* Max database page to write */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1662 | int i; /* Loop counter */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1663 | volatile WalCkptInfo *pInfo; /* The checkpoint status information */ |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1664 | int (*xBusy)(void*) = 0; /* Function to call when waiting for locks */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1665 | |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1666 | szPage = walPagesize(pWal); |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 1667 | testcase( szPage<=32768 ); |
| 1668 | testcase( szPage>=65536 ); |
drh | 7d20844 | 2010-12-16 02:06:29 +0000 | [diff] [blame] | 1669 | pInfo = walCkptInfo(pWal); |
| 1670 | if( pInfo->nBackfill>=pWal->hdr.mxFrame ) return SQLITE_OK; |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1671 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1672 | /* Allocate the iterator */ |
dan | 8f6097c | 2010-05-06 07:43:58 +0000 | [diff] [blame] | 1673 | rc = walIteratorInit(pWal, &pIter); |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1674 | if( rc!=SQLITE_OK ){ |
dan | bdf1e24 | 2010-06-25 15:16:25 +0000 | [diff] [blame] | 1675 | return rc; |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 1676 | } |
dan | f544b4c | 2010-06-25 11:35:52 +0000 | [diff] [blame] | 1677 | assert( pIter ); |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 1678 | |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1679 | if( eMode!=SQLITE_CHECKPOINT_PASSIVE ) xBusy = xBusyCall; |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 1680 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1681 | /* Compute in mxSafeFrame the index of the last frame of the WAL that is |
| 1682 | ** safe to write into the database. Frames beyond mxSafeFrame might |
| 1683 | ** overwrite database pages that are in use by active readers and thus |
| 1684 | ** cannot be backfilled from the WAL. |
| 1685 | */ |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 1686 | mxSafeFrame = pWal->hdr.mxFrame; |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1687 | mxPage = pWal->hdr.nPage; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1688 | for(i=1; i<WAL_NREADER; i++){ |
| 1689 | u32 y = pInfo->aReadMark[i]; |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1690 | if( mxSafeFrame>y ){ |
dan | 83f42d1 | 2010-06-04 10:37:05 +0000 | [diff] [blame] | 1691 | assert( y<=pWal->hdr.mxFrame ); |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1692 | rc = walBusyLock(pWal, xBusy, pBusyArg, WAL_READ_LOCK(i), 1); |
dan | 83f42d1 | 2010-06-04 10:37:05 +0000 | [diff] [blame] | 1693 | if( rc==SQLITE_OK ){ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 1694 | pInfo->aReadMark[i] = READMARK_NOT_USED; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1695 | walUnlockExclusive(pWal, WAL_READ_LOCK(i), 1); |
drh | 2d37e1c | 2010-06-02 20:38:20 +0000 | [diff] [blame] | 1696 | }else if( rc==SQLITE_BUSY ){ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 1697 | mxSafeFrame = y; |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1698 | xBusy = 0; |
drh | 2d37e1c | 2010-06-02 20:38:20 +0000 | [diff] [blame] | 1699 | }else{ |
dan | 83f42d1 | 2010-06-04 10:37:05 +0000 | [diff] [blame] | 1700 | goto walcheckpoint_out; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1701 | } |
| 1702 | } |
dan | c511878 | 2010-04-17 17:34:41 +0000 | [diff] [blame] | 1703 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1704 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1705 | if( pInfo->nBackfill<mxSafeFrame |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 1706 | && (rc = walBusyLock(pWal, xBusy, pBusyArg, WAL_READ_LOCK(0), 1))==SQLITE_OK |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1707 | ){ |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1708 | i64 nSize; /* Current size of database file */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1709 | u32 nBackfill = pInfo->nBackfill; |
| 1710 | |
| 1711 | /* Sync the WAL to disk */ |
| 1712 | if( sync_flags ){ |
| 1713 | rc = sqlite3OsSync(pWal->pWalFd, sync_flags); |
| 1714 | } |
| 1715 | |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1716 | /* If the database file may grow as a result of this checkpoint, hint |
| 1717 | ** about the eventual size of the db file to the VFS layer. |
| 1718 | */ |
dan | 007820d | 2010-08-09 07:51:40 +0000 | [diff] [blame] | 1719 | if( rc==SQLITE_OK ){ |
| 1720 | i64 nReq = ((i64)mxPage * szPage); |
| 1721 | rc = sqlite3OsFileSize(pWal->pDbFd, &nSize); |
| 1722 | if( rc==SQLITE_OK && nSize<nReq ){ |
| 1723 | sqlite3OsFileControl(pWal->pDbFd, SQLITE_FCNTL_SIZE_HINT, &nReq); |
| 1724 | } |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1725 | } |
| 1726 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1727 | /* Iterate through the contents of the WAL, copying data to the db file. */ |
| 1728 | while( rc==SQLITE_OK && 0==walIteratorNext(pIter, &iDbpage, &iFrame) ){ |
drh | 3e8e7ec | 2010-07-07 13:43:19 +0000 | [diff] [blame] | 1729 | i64 iOffset; |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 1730 | assert( walFramePgno(pWal, iFrame)==iDbpage ); |
dan | 502019c | 2010-07-28 14:26:17 +0000 | [diff] [blame] | 1731 | if( iFrame<=nBackfill || iFrame>mxSafeFrame || iDbpage>mxPage ) continue; |
drh | 3e8e7ec | 2010-07-07 13:43:19 +0000 | [diff] [blame] | 1732 | iOffset = walFrameOffset(iFrame, szPage) + WAL_FRAME_HDRSIZE; |
drh | 09b5dbc | 2010-07-07 14:35:58 +0000 | [diff] [blame] | 1733 | /* testcase( IS_BIG_INT(iOffset) ); // requires a 4GiB WAL file */ |
drh | 3e8e7ec | 2010-07-07 13:43:19 +0000 | [diff] [blame] | 1734 | rc = sqlite3OsRead(pWal->pWalFd, zBuf, szPage, iOffset); |
| 1735 | if( rc!=SQLITE_OK ) break; |
| 1736 | iOffset = (iDbpage-1)*(i64)szPage; |
| 1737 | testcase( IS_BIG_INT(iOffset) ); |
| 1738 | rc = sqlite3OsWrite(pWal->pDbFd, zBuf, szPage, iOffset); |
| 1739 | if( rc!=SQLITE_OK ) break; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1740 | } |
| 1741 | |
| 1742 | /* If work was actually accomplished... */ |
dan | d764c7d | 2010-06-04 11:56:22 +0000 | [diff] [blame] | 1743 | if( rc==SQLITE_OK ){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1744 | if( mxSafeFrame==walIndexHdr(pWal)->mxFrame ){ |
drh | 3e8e7ec | 2010-07-07 13:43:19 +0000 | [diff] [blame] | 1745 | i64 szDb = pWal->hdr.nPage*(i64)szPage; |
| 1746 | testcase( IS_BIG_INT(szDb) ); |
| 1747 | rc = sqlite3OsTruncate(pWal->pDbFd, szDb); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1748 | if( rc==SQLITE_OK && sync_flags ){ |
| 1749 | rc = sqlite3OsSync(pWal->pDbFd, sync_flags); |
| 1750 | } |
| 1751 | } |
dan | d764c7d | 2010-06-04 11:56:22 +0000 | [diff] [blame] | 1752 | if( rc==SQLITE_OK ){ |
| 1753 | pInfo->nBackfill = mxSafeFrame; |
| 1754 | } |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1755 | } |
| 1756 | |
| 1757 | /* Release the reader lock held while backfilling */ |
| 1758 | walUnlockExclusive(pWal, WAL_READ_LOCK(0), 1); |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 1759 | } |
| 1760 | |
| 1761 | if( rc==SQLITE_BUSY ){ |
drh | 34116ea | 2010-05-31 12:30:52 +0000 | [diff] [blame] | 1762 | /* Reset the return code so as not to report a checkpoint failure |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 1763 | ** just because there are active readers. */ |
drh | 34116ea | 2010-05-31 12:30:52 +0000 | [diff] [blame] | 1764 | rc = SQLITE_OK; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1765 | } |
| 1766 | |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1767 | /* If this is an SQLITE_CHECKPOINT_RESTART operation, and the entire wal |
| 1768 | ** file has been copied into the database file, then block until all |
| 1769 | ** readers have finished using the wal file. This ensures that the next |
| 1770 | ** process to write to the database restarts the wal file. |
| 1771 | */ |
| 1772 | if( rc==SQLITE_OK && eMode!=SQLITE_CHECKPOINT_PASSIVE ){ |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 1773 | assert( pWal->writeLock ); |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 1774 | if( pInfo->nBackfill<pWal->hdr.mxFrame ){ |
| 1775 | rc = SQLITE_BUSY; |
| 1776 | }else if( eMode==SQLITE_CHECKPOINT_RESTART ){ |
| 1777 | assert( mxSafeFrame==pWal->hdr.mxFrame ); |
| 1778 | rc = walBusyLock(pWal, xBusy, pBusyArg, WAL_READ_LOCK(1), WAL_NREADER-1); |
| 1779 | if( rc==SQLITE_OK ){ |
| 1780 | walUnlockExclusive(pWal, WAL_READ_LOCK(1), WAL_NREADER-1); |
| 1781 | } |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 1782 | } |
| 1783 | } |
| 1784 | |
dan | 83f42d1 | 2010-06-04 10:37:05 +0000 | [diff] [blame] | 1785 | walcheckpoint_out: |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1786 | walIteratorFree(pIter); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1787 | return rc; |
| 1788 | } |
| 1789 | |
| 1790 | /* |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 1791 | ** If the WAL file is currently larger than nMax bytes in size, truncate |
| 1792 | ** it to exactly nMax bytes. If an error occurs while doing so, ignore it. |
drh | 8dd4afa | 2011-12-08 19:50:32 +0000 | [diff] [blame] | 1793 | */ |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 1794 | static void walLimitSize(Wal *pWal, i64 nMax){ |
| 1795 | i64 sz; |
| 1796 | int rx; |
| 1797 | sqlite3BeginBenignMalloc(); |
| 1798 | rx = sqlite3OsFileSize(pWal->pWalFd, &sz); |
| 1799 | if( rx==SQLITE_OK && (sz > nMax ) ){ |
| 1800 | rx = sqlite3OsTruncate(pWal->pWalFd, nMax); |
| 1801 | } |
| 1802 | sqlite3EndBenignMalloc(); |
| 1803 | if( rx ){ |
| 1804 | sqlite3_log(rx, "cannot limit WAL size: %s", pWal->zWalName); |
drh | 8dd4afa | 2011-12-08 19:50:32 +0000 | [diff] [blame] | 1805 | } |
| 1806 | } |
| 1807 | |
| 1808 | /* |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1809 | ** Close a connection to a log file. |
| 1810 | */ |
drh | c438efd | 2010-04-26 00:19:45 +0000 | [diff] [blame] | 1811 | int sqlite3WalClose( |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1812 | Wal *pWal, /* Wal to close */ |
dan | c511878 | 2010-04-17 17:34:41 +0000 | [diff] [blame] | 1813 | int sync_flags, /* Flags to pass to OsSync() (or 0) */ |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 1814 | int nBuf, |
| 1815 | u8 *zBuf /* Buffer of at least nBuf bytes */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1816 | ){ |
| 1817 | int rc = SQLITE_OK; |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1818 | if( pWal ){ |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1819 | int isDelete = 0; /* True to unlink wal and wal-index files */ |
| 1820 | |
| 1821 | /* If an EXCLUSIVE lock can be obtained on the database file (using the |
| 1822 | ** ordinary, rollback-mode locking methods, this guarantees that the |
| 1823 | ** connection associated with this log file is the only connection to |
| 1824 | ** the database. In this case checkpoint the database and unlink both |
| 1825 | ** the wal and wal-index files. |
| 1826 | ** |
| 1827 | ** The EXCLUSIVE lock is not released before returning. |
| 1828 | */ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1829 | rc = sqlite3OsLock(pWal->pDbFd, SQLITE_LOCK_EXCLUSIVE); |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1830 | if( rc==SQLITE_OK ){ |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 1831 | if( pWal->exclusiveMode==WAL_NORMAL_MODE ){ |
| 1832 | pWal->exclusiveMode = WAL_EXCLUSIVE_MODE; |
| 1833 | } |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 1834 | rc = sqlite3WalCheckpoint( |
| 1835 | pWal, SQLITE_CHECKPOINT_PASSIVE, 0, 0, sync_flags, nBuf, zBuf, 0, 0 |
| 1836 | ); |
drh | eed4250 | 2011-12-16 15:38:52 +0000 | [diff] [blame] | 1837 | if( rc==SQLITE_OK ){ |
| 1838 | int bPersist = -1; |
| 1839 | sqlite3OsFileControl(pWal->pDbFd, SQLITE_FCNTL_PERSIST_WAL, &bPersist); |
| 1840 | if( bPersist!=1 ){ |
| 1841 | /* Try to delete the WAL file if the checkpoint completed and |
| 1842 | ** fsyned (rc==SQLITE_OK) and if we are not in persistent-wal |
| 1843 | ** mode (!bPersist) */ |
| 1844 | isDelete = 1; |
| 1845 | }else if( pWal->mxWalSize>=0 ){ |
| 1846 | /* Try to truncate the WAL file to zero bytes if the checkpoint |
| 1847 | ** completed and fsynced (rc==SQLITE_OK) and we are in persistent |
| 1848 | ** WAL mode (bPersist) and if the PRAGMA journal_size_limit is a |
| 1849 | ** non-negative value (pWal->mxWalSize>=0). Note that we truncate |
| 1850 | ** to zero bytes as truncating to the journal_size_limit might |
| 1851 | ** leave a corrupt WAL file on disk. */ |
| 1852 | walLimitSize(pWal, 0); |
| 1853 | } |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1854 | } |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1855 | } |
| 1856 | |
dan | 1018e90 | 2010-05-05 15:33:05 +0000 | [diff] [blame] | 1857 | walIndexClose(pWal, isDelete); |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1858 | sqlite3OsClose(pWal->pWalFd); |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1859 | if( isDelete ){ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 1860 | sqlite3OsDelete(pWal->pVfs, pWal->zWalName, 0); |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 1861 | } |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 1862 | WALTRACE(("WAL%p: closed\n", pWal)); |
shaneh | 8a300f8 | 2010-07-02 18:15:31 +0000 | [diff] [blame] | 1863 | sqlite3_free((void *)pWal->apWiData); |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1864 | sqlite3_free(pWal); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 1865 | } |
| 1866 | return rc; |
| 1867 | } |
| 1868 | |
| 1869 | /* |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1870 | ** Try to read the wal-index header. Return 0 on success and 1 if |
| 1871 | ** there is a problem. |
| 1872 | ** |
| 1873 | ** The wal-index is in shared memory. Another thread or process might |
| 1874 | ** be writing the header at the same time this procedure is trying to |
| 1875 | ** read it, which might result in inconsistency. A dirty read is detected |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1876 | ** by verifying that both copies of the header are the same and also by |
| 1877 | ** a checksum on the header. |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1878 | ** |
| 1879 | ** If and only if the read is consistent and the header is different from |
| 1880 | ** pWal->hdr, then pWal->hdr is updated to the content of the new header |
| 1881 | ** and *pChanged is set to 1. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1882 | ** |
dan | 8467050 | 2010-05-07 05:46:23 +0000 | [diff] [blame] | 1883 | ** If the checksum cannot be verified return non-zero. If the header |
| 1884 | ** is read successfully and the checksum verified, return zero. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1885 | */ |
drh | 7750ab4 | 2010-06-26 22:16:02 +0000 | [diff] [blame] | 1886 | static int walIndexTryHdr(Wal *pWal, int *pChanged){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1887 | u32 aCksum[2]; /* Checksum on the header content */ |
| 1888 | WalIndexHdr h1, h2; /* Two copies of the header content */ |
| 1889 | WalIndexHdr volatile *aHdr; /* Header in shared memory */ |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1890 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1891 | /* The first page of the wal-index must be mapped at this point. */ |
| 1892 | assert( pWal->nWiData>0 && pWal->apWiData[0] ); |
drh | 79e6c78 | 2010-04-30 02:13:26 +0000 | [diff] [blame] | 1893 | |
drh | 6cef0cf | 2010-08-16 16:31:43 +0000 | [diff] [blame] | 1894 | /* Read the header. This might happen concurrently with a write to the |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1895 | ** same area of shared memory on a different CPU in a SMP, |
| 1896 | ** meaning it is possible that an inconsistent snapshot is read |
dan | 8467050 | 2010-05-07 05:46:23 +0000 | [diff] [blame] | 1897 | ** from the file. If this happens, return non-zero. |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1898 | ** |
| 1899 | ** There are two copies of the header at the beginning of the wal-index. |
| 1900 | ** When reading, read [0] first then [1]. Writes are in the reverse order. |
| 1901 | ** Memory barriers are used to prevent the compiler or the hardware from |
| 1902 | ** reordering the reads and writes. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1903 | */ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1904 | aHdr = walIndexHdr(pWal); |
| 1905 | memcpy(&h1, (void *)&aHdr[0], sizeof(h1)); |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 1906 | walShmBarrier(pWal); |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1907 | memcpy(&h2, (void *)&aHdr[1], sizeof(h2)); |
drh | 286a288 | 2010-05-20 23:51:06 +0000 | [diff] [blame] | 1908 | |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1909 | if( memcmp(&h1, &h2, sizeof(h1))!=0 ){ |
| 1910 | return 1; /* Dirty read */ |
drh | 286a288 | 2010-05-20 23:51:06 +0000 | [diff] [blame] | 1911 | } |
drh | 4b82c38 | 2010-05-31 18:24:19 +0000 | [diff] [blame] | 1912 | if( h1.isInit==0 ){ |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1913 | return 1; /* Malformed header - probably all zeros */ |
| 1914 | } |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 1915 | walChecksumBytes(1, (u8*)&h1, sizeof(h1)-sizeof(h1.aCksum), 0, aCksum); |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1916 | if( aCksum[0]!=h1.aCksum[0] || aCksum[1]!=h1.aCksum[1] ){ |
| 1917 | return 1; /* Checksum does not match */ |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1918 | } |
| 1919 | |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1920 | if( memcmp(&pWal->hdr, &h1, sizeof(WalIndexHdr)) ){ |
dan | a861469 | 2010-05-06 14:42:34 +0000 | [diff] [blame] | 1921 | *pChanged = 1; |
drh | f0b20f8 | 2010-05-21 13:16:18 +0000 | [diff] [blame] | 1922 | memcpy(&pWal->hdr, &h1, sizeof(WalIndexHdr)); |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 1923 | pWal->szPage = (pWal->hdr.szPage&0xfe00) + ((pWal->hdr.szPage&0x0001)<<16); |
| 1924 | testcase( pWal->szPage<=32768 ); |
| 1925 | testcase( pWal->szPage>=65536 ); |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1926 | } |
dan | 8467050 | 2010-05-07 05:46:23 +0000 | [diff] [blame] | 1927 | |
| 1928 | /* The header was successfully read. Return zero. */ |
| 1929 | return 0; |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1930 | } |
| 1931 | |
| 1932 | /* |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1933 | ** Read the wal-index header from the wal-index and into pWal->hdr. |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 1934 | ** If the wal-header appears to be corrupt, try to reconstruct the |
| 1935 | ** wal-index from the WAL before returning. |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 1936 | ** |
| 1937 | ** Set *pChanged to 1 if the wal-index header value in pWal->hdr is |
| 1938 | ** changed by this opertion. If pWal->hdr is unchanged, set *pChanged |
| 1939 | ** to 0. |
| 1940 | ** |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1941 | ** If the wal-index header is successfully read, return SQLITE_OK. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1942 | ** Otherwise an SQLite error code. |
| 1943 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1944 | static int walIndexReadHdr(Wal *pWal, int *pChanged){ |
dan | 8467050 | 2010-05-07 05:46:23 +0000 | [diff] [blame] | 1945 | int rc; /* Return code */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1946 | int badHdr; /* True if a header read failed */ |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 1947 | volatile u32 *page0; /* Chunk of wal-index containing header */ |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1948 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1949 | /* Ensure that page 0 of the wal-index (the page that contains the |
| 1950 | ** wal-index header) is mapped. Return early if an error occurs here. |
| 1951 | */ |
dan | a861469 | 2010-05-06 14:42:34 +0000 | [diff] [blame] | 1952 | assert( pChanged ); |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1953 | rc = walIndexPage(pWal, 0, &page0); |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 1954 | if( rc!=SQLITE_OK ){ |
| 1955 | return rc; |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1956 | }; |
| 1957 | assert( page0 || pWal->writeLock==0 ); |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 1958 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1959 | /* If the first page of the wal-index has been mapped, try to read the |
| 1960 | ** wal-index header immediately, without holding any lock. This usually |
| 1961 | ** works, but may fail if the wal-index header is corrupt or currently |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 1962 | ** being modified by another thread or process. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1963 | */ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 1964 | badHdr = (page0 ? walIndexTryHdr(pWal, pChanged) : 1); |
drh | bab7b91 | 2010-05-26 17:31:58 +0000 | [diff] [blame] | 1965 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1966 | /* If the first attempt failed, it might have been due to a race |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 1967 | ** with a writer. So get a WRITE lock and try again. |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 1968 | */ |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 1969 | assert( badHdr==0 || pWal->writeLock==0 ); |
dan | 4edc6bf | 2011-05-10 17:31:29 +0000 | [diff] [blame] | 1970 | if( badHdr ){ |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 1971 | if( pWal->readOnly & WAL_SHM_RDONLY ){ |
dan | 4edc6bf | 2011-05-10 17:31:29 +0000 | [diff] [blame] | 1972 | if( SQLITE_OK==(rc = walLockShared(pWal, WAL_WRITE_LOCK)) ){ |
| 1973 | walUnlockShared(pWal, WAL_WRITE_LOCK); |
| 1974 | rc = SQLITE_READONLY_RECOVERY; |
drh | bab7b91 | 2010-05-26 17:31:58 +0000 | [diff] [blame] | 1975 | } |
dan | 4edc6bf | 2011-05-10 17:31:29 +0000 | [diff] [blame] | 1976 | }else if( SQLITE_OK==(rc = walLockExclusive(pWal, WAL_WRITE_LOCK, 1)) ){ |
| 1977 | pWal->writeLock = 1; |
| 1978 | if( SQLITE_OK==(rc = walIndexPage(pWal, 0, &page0)) ){ |
| 1979 | badHdr = walIndexTryHdr(pWal, pChanged); |
| 1980 | if( badHdr ){ |
| 1981 | /* If the wal-index header is still malformed even while holding |
| 1982 | ** a WRITE lock, it can only mean that the header is corrupted and |
| 1983 | ** needs to be reconstructed. So run recovery to do exactly that. |
| 1984 | */ |
| 1985 | rc = walIndexRecover(pWal); |
| 1986 | *pChanged = 1; |
| 1987 | } |
| 1988 | } |
| 1989 | pWal->writeLock = 0; |
| 1990 | walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1); |
drh | bab7b91 | 2010-05-26 17:31:58 +0000 | [diff] [blame] | 1991 | } |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 1992 | } |
| 1993 | |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 1994 | /* If the header is read successfully, check the version number to make |
| 1995 | ** sure the wal-index was not constructed with some future format that |
| 1996 | ** this version of SQLite cannot understand. |
| 1997 | */ |
| 1998 | if( badHdr==0 && pWal->hdr.iVersion!=WALINDEX_MAX_VERSION ){ |
| 1999 | rc = SQLITE_CANTOPEN_BKPT; |
| 2000 | } |
| 2001 | |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2002 | return rc; |
| 2003 | } |
| 2004 | |
| 2005 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2006 | ** This is the value that walTryBeginRead returns when it needs to |
| 2007 | ** be retried. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2008 | */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2009 | #define WAL_RETRY (-1) |
dan | 64d039e | 2010-04-13 19:27:31 +0000 | [diff] [blame] | 2010 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2011 | /* |
| 2012 | ** Attempt to start a read transaction. This might fail due to a race or |
| 2013 | ** other transient condition. When that happens, it returns WAL_RETRY to |
| 2014 | ** indicate to the caller that it is safe to retry immediately. |
| 2015 | ** |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2016 | ** On success return SQLITE_OK. On a permanent failure (such an |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2017 | ** I/O error or an SQLITE_BUSY because another process is running |
| 2018 | ** recovery) return a positive error code. |
| 2019 | ** |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2020 | ** The useWal parameter is true to force the use of the WAL and disable |
| 2021 | ** the case where the WAL is bypassed because it has been completely |
| 2022 | ** checkpointed. If useWal==0 then this routine calls walIndexReadHdr() |
| 2023 | ** to make a copy of the wal-index header into pWal->hdr. If the |
| 2024 | ** wal-index header has changed, *pChanged is set to 1 (as an indication |
| 2025 | ** to the caller that the local paget cache is obsolete and needs to be |
| 2026 | ** flushed.) When useWal==1, the wal-index header is assumed to already |
| 2027 | ** be loaded and the pChanged parameter is unused. |
| 2028 | ** |
| 2029 | ** The caller must set the cnt parameter to the number of prior calls to |
| 2030 | ** this routine during the current read attempt that returned WAL_RETRY. |
| 2031 | ** This routine will start taking more aggressive measures to clear the |
| 2032 | ** race conditions after multiple WAL_RETRY returns, and after an excessive |
| 2033 | ** number of errors will ultimately return SQLITE_PROTOCOL. The |
| 2034 | ** SQLITE_PROTOCOL return indicates that some other process has gone rogue |
| 2035 | ** and is not honoring the locking protocol. There is a vanishingly small |
| 2036 | ** chance that SQLITE_PROTOCOL could be returned because of a run of really |
| 2037 | ** bad luck when there is lots of contention for the wal-index, but that |
| 2038 | ** possibility is so small that it can be safely neglected, we believe. |
| 2039 | ** |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2040 | ** On success, this routine obtains a read lock on |
| 2041 | ** WAL_READ_LOCK(pWal->readLock). The pWal->readLock integer is |
| 2042 | ** in the range 0 <= pWal->readLock < WAL_NREADER. If pWal->readLock==(-1) |
| 2043 | ** that means the Wal does not hold any read lock. The reader must not |
| 2044 | ** access any database page that is modified by a WAL frame up to and |
| 2045 | ** including frame number aReadMark[pWal->readLock]. The reader will |
| 2046 | ** use WAL frames up to and including pWal->hdr.mxFrame if pWal->readLock>0 |
| 2047 | ** Or if pWal->readLock==0, then the reader will ignore the WAL |
| 2048 | ** completely and get all content directly from the database file. |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2049 | ** If the useWal parameter is 1 then the WAL will never be ignored and |
| 2050 | ** this routine will always set pWal->readLock>0 on success. |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2051 | ** When the read transaction is completed, the caller must release the |
| 2052 | ** lock on WAL_READ_LOCK(pWal->readLock) and set pWal->readLock to -1. |
| 2053 | ** |
| 2054 | ** This routine uses the nBackfill and aReadMark[] fields of the header |
| 2055 | ** to select a particular WAL_READ_LOCK() that strives to let the |
| 2056 | ** checkpoint process do as much work as possible. This routine might |
| 2057 | ** update values of the aReadMark[] array in the header, but if it does |
| 2058 | ** so it takes care to hold an exclusive lock on the corresponding |
| 2059 | ** WAL_READ_LOCK() while changing values. |
| 2060 | */ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2061 | static int walTryBeginRead(Wal *pWal, int *pChanged, int useWal, int cnt){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2062 | volatile WalCkptInfo *pInfo; /* Checkpoint information in wal-index */ |
| 2063 | u32 mxReadMark; /* Largest aReadMark[] value */ |
| 2064 | int mxI; /* Index of largest aReadMark[] value */ |
| 2065 | int i; /* Loop counter */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 2066 | int rc = SQLITE_OK; /* Return code */ |
dan | 64d039e | 2010-04-13 19:27:31 +0000 | [diff] [blame] | 2067 | |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2068 | assert( pWal->readLock<0 ); /* Not currently locked */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2069 | |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2070 | /* Take steps to avoid spinning forever if there is a protocol error. |
| 2071 | ** |
| 2072 | ** Circumstances that cause a RETRY should only last for the briefest |
| 2073 | ** instances of time. No I/O or other system calls are done while the |
| 2074 | ** locks are held, so the locks should not be held for very long. But |
| 2075 | ** if we are unlucky, another process that is holding a lock might get |
| 2076 | ** paged out or take a page-fault that is time-consuming to resolve, |
| 2077 | ** during the few nanoseconds that it is holding the lock. In that case, |
| 2078 | ** it might take longer than normal for the lock to free. |
| 2079 | ** |
| 2080 | ** After 5 RETRYs, we begin calling sqlite3OsSleep(). The first few |
| 2081 | ** calls to sqlite3OsSleep() have a delay of 1 microsecond. Really this |
| 2082 | ** is more of a scheduler yield than an actual delay. But on the 10th |
| 2083 | ** an subsequent retries, the delays start becoming longer and longer, |
| 2084 | ** so that on the 100th (and last) RETRY we delay for 21 milliseconds. |
| 2085 | ** The total delay time before giving up is less than 1 second. |
| 2086 | */ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2087 | if( cnt>5 ){ |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2088 | int nDelay = 1; /* Pause time in microseconds */ |
drh | 03c6967 | 2011-02-19 23:18:12 +0000 | [diff] [blame] | 2089 | if( cnt>100 ){ |
| 2090 | VVA_ONLY( pWal->lockError = 1; ) |
| 2091 | return SQLITE_PROTOCOL; |
| 2092 | } |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2093 | if( cnt>=10 ) nDelay = (cnt-9)*238; /* Max delay 21ms. Total delay 996ms */ |
| 2094 | sqlite3OsSleep(pWal->pVfs, nDelay); |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2095 | } |
| 2096 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2097 | if( !useWal ){ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2098 | rc = walIndexReadHdr(pWal, pChanged); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2099 | if( rc==SQLITE_BUSY ){ |
| 2100 | /* If there is not a recovery running in another thread or process |
| 2101 | ** then convert BUSY errors to WAL_RETRY. If recovery is known to |
| 2102 | ** be running, convert BUSY to BUSY_RECOVERY. There is a race here |
| 2103 | ** which might cause WAL_RETRY to be returned even if BUSY_RECOVERY |
| 2104 | ** would be technically correct. But the race is benign since with |
| 2105 | ** WAL_RETRY this routine will be called again and will probably be |
| 2106 | ** right on the second iteration. |
| 2107 | */ |
dan | 7d4514a | 2010-07-15 17:54:14 +0000 | [diff] [blame] | 2108 | if( pWal->apWiData[0]==0 ){ |
| 2109 | /* This branch is taken when the xShmMap() method returns SQLITE_BUSY. |
| 2110 | ** We assume this is a transient condition, so return WAL_RETRY. The |
| 2111 | ** xShmMap() implementation used by the default unix and win32 VFS |
| 2112 | ** modules may return SQLITE_BUSY due to a race condition in the |
| 2113 | ** code that determines whether or not the shared-memory region |
| 2114 | ** must be zeroed before the requested page is returned. |
| 2115 | */ |
| 2116 | rc = WAL_RETRY; |
| 2117 | }else if( SQLITE_OK==(rc = walLockShared(pWal, WAL_RECOVER_LOCK)) ){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2118 | walUnlockShared(pWal, WAL_RECOVER_LOCK); |
| 2119 | rc = WAL_RETRY; |
| 2120 | }else if( rc==SQLITE_BUSY ){ |
| 2121 | rc = SQLITE_BUSY_RECOVERY; |
| 2122 | } |
| 2123 | } |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2124 | if( rc!=SQLITE_OK ){ |
| 2125 | return rc; |
| 2126 | } |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2127 | } |
| 2128 | |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 2129 | pInfo = walCkptInfo(pWal); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2130 | if( !useWal && pInfo->nBackfill==pWal->hdr.mxFrame ){ |
| 2131 | /* The WAL has been completely backfilled (or it is empty). |
| 2132 | ** and can be safely ignored. |
| 2133 | */ |
| 2134 | rc = walLockShared(pWal, WAL_READ_LOCK(0)); |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 2135 | walShmBarrier(pWal); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2136 | if( rc==SQLITE_OK ){ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 2137 | if( memcmp((void *)walIndexHdr(pWal), &pWal->hdr, sizeof(WalIndexHdr)) ){ |
dan | 493cc59 | 2010-06-05 18:12:23 +0000 | [diff] [blame] | 2138 | /* It is not safe to allow the reader to continue here if frames |
| 2139 | ** may have been appended to the log before READ_LOCK(0) was obtained. |
| 2140 | ** When holding READ_LOCK(0), the reader ignores the entire log file, |
| 2141 | ** which implies that the database file contains a trustworthy |
| 2142 | ** snapshoT. Since holding READ_LOCK(0) prevents a checkpoint from |
| 2143 | ** happening, this is usually correct. |
| 2144 | ** |
| 2145 | ** However, if frames have been appended to the log (or if the log |
| 2146 | ** is wrapped and written for that matter) before the READ_LOCK(0) |
| 2147 | ** is obtained, that is not necessarily true. A checkpointer may |
| 2148 | ** have started to backfill the appended frames but crashed before |
| 2149 | ** it finished. Leaving a corrupt image in the database file. |
| 2150 | */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2151 | walUnlockShared(pWal, WAL_READ_LOCK(0)); |
| 2152 | return WAL_RETRY; |
| 2153 | } |
| 2154 | pWal->readLock = 0; |
| 2155 | return SQLITE_OK; |
| 2156 | }else if( rc!=SQLITE_BUSY ){ |
| 2157 | return rc; |
dan | 64d039e | 2010-04-13 19:27:31 +0000 | [diff] [blame] | 2158 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2159 | } |
dan | ba51590 | 2010-04-30 09:32:06 +0000 | [diff] [blame] | 2160 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2161 | /* If we get this far, it means that the reader will want to use |
| 2162 | ** the WAL to get at content from recent commits. The job now is |
| 2163 | ** to select one of the aReadMark[] entries that is closest to |
| 2164 | ** but not exceeding pWal->hdr.mxFrame and lock that entry. |
| 2165 | */ |
| 2166 | mxReadMark = 0; |
| 2167 | mxI = 0; |
| 2168 | for(i=1; i<WAL_NREADER; i++){ |
| 2169 | u32 thisMark = pInfo->aReadMark[i]; |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 2170 | if( mxReadMark<=thisMark && thisMark<=pWal->hdr.mxFrame ){ |
| 2171 | assert( thisMark!=READMARK_NOT_USED ); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2172 | mxReadMark = thisMark; |
| 2173 | mxI = i; |
| 2174 | } |
| 2175 | } |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2176 | /* There was once an "if" here. The extra "{" is to preserve indentation. */ |
| 2177 | { |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 2178 | if( (pWal->readOnly & WAL_SHM_RDONLY)==0 |
| 2179 | && (mxReadMark<pWal->hdr.mxFrame || mxI==0) |
| 2180 | ){ |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 2181 | for(i=1; i<WAL_NREADER; i++){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2182 | rc = walLockExclusive(pWal, WAL_READ_LOCK(i), 1); |
| 2183 | if( rc==SQLITE_OK ){ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 2184 | mxReadMark = pInfo->aReadMark[i] = pWal->hdr.mxFrame; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2185 | mxI = i; |
| 2186 | walUnlockExclusive(pWal, WAL_READ_LOCK(i), 1); |
| 2187 | break; |
drh | 38933f2 | 2010-06-02 15:43:18 +0000 | [diff] [blame] | 2188 | }else if( rc!=SQLITE_BUSY ){ |
| 2189 | return rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2190 | } |
| 2191 | } |
| 2192 | } |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2193 | if( mxI==0 ){ |
drh | 5bf3934 | 2011-06-02 17:24:49 +0000 | [diff] [blame] | 2194 | assert( rc==SQLITE_BUSY || (pWal->readOnly & WAL_SHM_RDONLY)!=0 ); |
dan | 4edc6bf | 2011-05-10 17:31:29 +0000 | [diff] [blame] | 2195 | return rc==SQLITE_BUSY ? WAL_RETRY : SQLITE_READONLY_CANTLOCK; |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2196 | } |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2197 | |
| 2198 | rc = walLockShared(pWal, WAL_READ_LOCK(mxI)); |
| 2199 | if( rc ){ |
| 2200 | return rc==SQLITE_BUSY ? WAL_RETRY : rc; |
| 2201 | } |
dan | eb8cb3a | 2010-06-05 18:34:26 +0000 | [diff] [blame] | 2202 | /* Now that the read-lock has been obtained, check that neither the |
| 2203 | ** value in the aReadMark[] array or the contents of the wal-index |
| 2204 | ** header have changed. |
| 2205 | ** |
| 2206 | ** It is necessary to check that the wal-index header did not change |
| 2207 | ** between the time it was read and when the shared-lock was obtained |
| 2208 | ** on WAL_READ_LOCK(mxI) was obtained to account for the possibility |
| 2209 | ** that the log file may have been wrapped by a writer, or that frames |
| 2210 | ** that occur later in the log than pWal->hdr.mxFrame may have been |
| 2211 | ** copied into the database by a checkpointer. If either of these things |
| 2212 | ** happened, then reading the database with the current value of |
| 2213 | ** pWal->hdr.mxFrame risks reading a corrupted snapshot. So, retry |
| 2214 | ** instead. |
| 2215 | ** |
dan | 640aac4 | 2010-06-05 19:18:59 +0000 | [diff] [blame] | 2216 | ** This does not guarantee that the copy of the wal-index header is up to |
| 2217 | ** date before proceeding. That would not be possible without somehow |
| 2218 | ** blocking writers. It only guarantees that a dangerous checkpoint or |
dan | eb8cb3a | 2010-06-05 18:34:26 +0000 | [diff] [blame] | 2219 | ** log-wrap (either of which would require an exclusive lock on |
| 2220 | ** WAL_READ_LOCK(mxI)) has not occurred since the snapshot was valid. |
| 2221 | */ |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 2222 | walShmBarrier(pWal); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2223 | if( pInfo->aReadMark[mxI]!=mxReadMark |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 2224 | || memcmp((void *)walIndexHdr(pWal), &pWal->hdr, sizeof(WalIndexHdr)) |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2225 | ){ |
| 2226 | walUnlockShared(pWal, WAL_READ_LOCK(mxI)); |
| 2227 | return WAL_RETRY; |
| 2228 | }else{ |
drh | db7f647 | 2010-06-09 14:45:12 +0000 | [diff] [blame] | 2229 | assert( mxReadMark<=pWal->hdr.mxFrame ); |
shaneh | 5eba1f6 | 2010-07-02 17:05:03 +0000 | [diff] [blame] | 2230 | pWal->readLock = (i16)mxI; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2231 | } |
| 2232 | } |
| 2233 | return rc; |
| 2234 | } |
| 2235 | |
| 2236 | /* |
| 2237 | ** Begin a read transaction on the database. |
| 2238 | ** |
| 2239 | ** This routine used to be called sqlite3OpenSnapshot() and with good reason: |
| 2240 | ** it takes a snapshot of the state of the WAL and wal-index for the current |
| 2241 | ** instant in time. The current thread will continue to use this snapshot. |
| 2242 | ** Other threads might append new content to the WAL and wal-index but |
| 2243 | ** that extra content is ignored by the current thread. |
| 2244 | ** |
| 2245 | ** If the database contents have changes since the previous read |
| 2246 | ** transaction, then *pChanged is set to 1 before returning. The |
| 2247 | ** Pager layer will use this to know that is cache is stale and |
| 2248 | ** needs to be flushed. |
| 2249 | */ |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 2250 | int sqlite3WalBeginReadTransaction(Wal *pWal, int *pChanged){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2251 | int rc; /* Return code */ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2252 | int cnt = 0; /* Number of TryBeginRead attempts */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2253 | |
| 2254 | do{ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2255 | rc = walTryBeginRead(pWal, pChanged, 0, ++cnt); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2256 | }while( rc==WAL_RETRY ); |
drh | ab1cc74 | 2011-02-19 16:51:45 +0000 | [diff] [blame] | 2257 | testcase( (rc&0xff)==SQLITE_BUSY ); |
| 2258 | testcase( (rc&0xff)==SQLITE_IOERR ); |
| 2259 | testcase( rc==SQLITE_PROTOCOL ); |
| 2260 | testcase( rc==SQLITE_OK ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2261 | return rc; |
| 2262 | } |
| 2263 | |
| 2264 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2265 | ** Finish with a read transaction. All this does is release the |
| 2266 | ** read-lock. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2267 | */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2268 | void sqlite3WalEndReadTransaction(Wal *pWal){ |
dan | 73d66fd | 2010-08-07 16:17:48 +0000 | [diff] [blame] | 2269 | sqlite3WalEndWriteTransaction(pWal); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2270 | if( pWal->readLock>=0 ){ |
| 2271 | walUnlockShared(pWal, WAL_READ_LOCK(pWal->readLock)); |
| 2272 | pWal->readLock = -1; |
| 2273 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2274 | } |
| 2275 | |
dan | 5e0ce87 | 2010-04-28 17:48:44 +0000 | [diff] [blame] | 2276 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2277 | ** Read a page from the WAL, if it is present in the WAL and if the |
| 2278 | ** current read transaction is configured to use the WAL. |
| 2279 | ** |
| 2280 | ** The *pInWal is set to 1 if the requested page is in the WAL and |
| 2281 | ** has been loaded. Or *pInWal is set to 0 if the page was not in |
| 2282 | ** the WAL and needs to be read out of the database. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2283 | */ |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 2284 | int sqlite3WalRead( |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2285 | Wal *pWal, /* WAL handle */ |
| 2286 | Pgno pgno, /* Database page number to read data for */ |
| 2287 | int *pInWal, /* OUT: True if data is read from WAL */ |
| 2288 | int nOut, /* Size of buffer pOut in bytes */ |
| 2289 | u8 *pOut /* Buffer to write page data to */ |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 2290 | ){ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2291 | u32 iRead = 0; /* If !=0, WAL frame to return data from */ |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 2292 | u32 iLast = pWal->hdr.mxFrame; /* Last page in WAL for this reader */ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2293 | int iHash; /* Used to loop through N hash tables */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2294 | |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2295 | /* This routine is only be called from within a read transaction. */ |
| 2296 | assert( pWal->readLock>=0 || pWal->lockError ); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2297 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2298 | /* If the "last page" field of the wal-index header snapshot is 0, then |
| 2299 | ** no data will be read from the wal under any circumstances. Return early |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2300 | ** in this case as an optimization. Likewise, if pWal->readLock==0, |
| 2301 | ** then the WAL is ignored by the reader so return early, as if the |
| 2302 | ** WAL were empty. |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2303 | */ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2304 | if( iLast==0 || pWal->readLock==0 ){ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2305 | *pInWal = 0; |
| 2306 | return SQLITE_OK; |
| 2307 | } |
| 2308 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2309 | /* Search the hash table or tables for an entry matching page number |
| 2310 | ** pgno. Each iteration of the following for() loop searches one |
| 2311 | ** hash table (each hash table indexes up to HASHTABLE_NPAGE frames). |
| 2312 | ** |
drh | a927e94 | 2010-06-24 02:46:48 +0000 | [diff] [blame] | 2313 | ** This code might run concurrently to the code in walIndexAppend() |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2314 | ** that adds entries to the wal-index (and possibly to this hash |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 2315 | ** table). This means the value just read from the hash |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2316 | ** slot (aHash[iKey]) may have been added before or after the |
| 2317 | ** current read transaction was opened. Values added after the |
| 2318 | ** read transaction was opened may have been written incorrectly - |
| 2319 | ** i.e. these slots may contain garbage data. However, we assume |
| 2320 | ** that any slots written before the current read transaction was |
| 2321 | ** opened remain unmodified. |
| 2322 | ** |
| 2323 | ** For the reasons above, the if(...) condition featured in the inner |
| 2324 | ** loop of the following block is more stringent that would be required |
| 2325 | ** if we had exclusive access to the hash-table: |
| 2326 | ** |
| 2327 | ** (aPgno[iFrame]==pgno): |
| 2328 | ** This condition filters out normal hash-table collisions. |
| 2329 | ** |
| 2330 | ** (iFrame<=iLast): |
| 2331 | ** This condition filters out entries that were added to the hash |
| 2332 | ** table after the current read-transaction had started. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2333 | */ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 2334 | for(iHash=walFramePage(iLast); iHash>=0 && iRead==0; iHash--){ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 2335 | volatile ht_slot *aHash; /* Pointer to hash table */ |
| 2336 | volatile u32 *aPgno; /* Pointer to array of page numbers */ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2337 | u32 iZero; /* Frame number corresponding to aPgno[0] */ |
| 2338 | int iKey; /* Hash slot index */ |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 2339 | int nCollide; /* Number of hash collisions remaining */ |
| 2340 | int rc; /* Error code */ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2341 | |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 2342 | rc = walHashGet(pWal, iHash, &aHash, &aPgno, &iZero); |
| 2343 | if( rc!=SQLITE_OK ){ |
| 2344 | return rc; |
| 2345 | } |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 2346 | nCollide = HASHTABLE_NSLOT; |
dan | 6f15014 | 2010-05-21 15:31:56 +0000 | [diff] [blame] | 2347 | for(iKey=walHash(pgno); aHash[iKey]; iKey=walNextHash(iKey)){ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2348 | u32 iFrame = aHash[iKey] + iZero; |
dan | d60bf11 | 2010-06-14 11:18:50 +0000 | [diff] [blame] | 2349 | if( iFrame<=iLast && aPgno[aHash[iKey]]==pgno ){ |
drh | d515660 | 2011-11-12 16:46:55 +0000 | [diff] [blame] | 2350 | /* assert( iFrame>iRead ); -- not true if there is corruption */ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2351 | iRead = iFrame; |
| 2352 | } |
drh | 519426a | 2010-07-09 03:19:07 +0000 | [diff] [blame] | 2353 | if( (nCollide--)==0 ){ |
| 2354 | return SQLITE_CORRUPT_BKPT; |
| 2355 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2356 | } |
| 2357 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2358 | |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2359 | #ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT |
| 2360 | /* If expensive assert() statements are available, do a linear search |
| 2361 | ** of the wal-index file content. Make sure the results agree with the |
| 2362 | ** result obtained using the hash indexes above. */ |
| 2363 | { |
| 2364 | u32 iRead2 = 0; |
| 2365 | u32 iTest; |
| 2366 | for(iTest=iLast; iTest>0; iTest--){ |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 2367 | if( walFramePgno(pWal, iTest)==pgno ){ |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2368 | iRead2 = iTest; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2369 | break; |
| 2370 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2371 | } |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2372 | assert( iRead==iRead2 ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2373 | } |
dan | bb23aff | 2010-05-10 14:46:09 +0000 | [diff] [blame] | 2374 | #endif |
dan | cd11fb2 | 2010-04-26 10:40:52 +0000 | [diff] [blame] | 2375 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2376 | /* If iRead is non-zero, then it is the log frame number that contains the |
| 2377 | ** required page. Read and return data from the log file. |
| 2378 | */ |
| 2379 | if( iRead ){ |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 2380 | int sz; |
| 2381 | i64 iOffset; |
| 2382 | sz = pWal->hdr.szPage; |
drh | b07028f | 2011-10-14 21:49:18 +0000 | [diff] [blame] | 2383 | sz = (sz&0xfe00) + ((sz&0x0001)<<16); |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 2384 | testcase( sz<=32768 ); |
| 2385 | testcase( sz>=65536 ); |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 2386 | iOffset = walFrameOffset(iRead, sz) + WAL_FRAME_HDRSIZE; |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2387 | *pInWal = 1; |
drh | 09b5dbc | 2010-07-07 14:35:58 +0000 | [diff] [blame] | 2388 | /* testcase( IS_BIG_INT(iOffset) ); // requires a 4GiB WAL */ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 2389 | return sqlite3OsRead(pWal->pWalFd, pOut, nOut, iOffset); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2390 | } |
| 2391 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2392 | *pInWal = 0; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2393 | return SQLITE_OK; |
| 2394 | } |
| 2395 | |
| 2396 | |
| 2397 | /* |
dan | 763afe6 | 2010-08-03 06:42:39 +0000 | [diff] [blame] | 2398 | ** Return the size of the database in pages (or zero, if unknown). |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2399 | */ |
dan | 763afe6 | 2010-08-03 06:42:39 +0000 | [diff] [blame] | 2400 | Pgno sqlite3WalDbsize(Wal *pWal){ |
drh | 7e9e70b | 2010-08-16 14:17:59 +0000 | [diff] [blame] | 2401 | if( pWal && ALWAYS(pWal->readLock>=0) ){ |
dan | 763afe6 | 2010-08-03 06:42:39 +0000 | [diff] [blame] | 2402 | return pWal->hdr.nPage; |
| 2403 | } |
| 2404 | return 0; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2405 | } |
| 2406 | |
dan | 30c8629 | 2010-04-30 16:24:46 +0000 | [diff] [blame] | 2407 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2408 | /* |
| 2409 | ** This function starts a write transaction on the WAL. |
| 2410 | ** |
| 2411 | ** A read transaction must have already been started by a prior call |
| 2412 | ** to sqlite3WalBeginReadTransaction(). |
| 2413 | ** |
| 2414 | ** If another thread or process has written into the database since |
| 2415 | ** the read transaction was started, then it is not possible for this |
| 2416 | ** thread to write as doing so would cause a fork. So this routine |
| 2417 | ** returns SQLITE_BUSY in that case and no write transaction is started. |
| 2418 | ** |
| 2419 | ** There can only be a single writer active at a time. |
| 2420 | */ |
| 2421 | int sqlite3WalBeginWriteTransaction(Wal *pWal){ |
| 2422 | int rc; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2423 | |
| 2424 | /* Cannot start a write transaction without first holding a read |
| 2425 | ** transaction. */ |
| 2426 | assert( pWal->readLock>=0 ); |
| 2427 | |
dan | 1e5de5a | 2010-07-15 18:20:53 +0000 | [diff] [blame] | 2428 | if( pWal->readOnly ){ |
| 2429 | return SQLITE_READONLY; |
| 2430 | } |
| 2431 | |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2432 | /* Only one writer allowed at a time. Get the write lock. Return |
| 2433 | ** SQLITE_BUSY if unable. |
| 2434 | */ |
| 2435 | rc = walLockExclusive(pWal, WAL_WRITE_LOCK, 1); |
| 2436 | if( rc ){ |
| 2437 | return rc; |
| 2438 | } |
drh | c99597c | 2010-05-31 01:41:15 +0000 | [diff] [blame] | 2439 | pWal->writeLock = 1; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2440 | |
| 2441 | /* If another connection has written to the database file since the |
| 2442 | ** time the read transaction on this connection was started, then |
| 2443 | ** the write is disallowed. |
| 2444 | */ |
dan | 4280eb3 | 2010-06-12 12:02:35 +0000 | [diff] [blame] | 2445 | if( memcmp(&pWal->hdr, (void *)walIndexHdr(pWal), sizeof(WalIndexHdr))!=0 ){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2446 | walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1); |
drh | c99597c | 2010-05-31 01:41:15 +0000 | [diff] [blame] | 2447 | pWal->writeLock = 0; |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2448 | rc = SQLITE_BUSY; |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2449 | } |
| 2450 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2451 | return rc; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2452 | } |
| 2453 | |
dan | 74d6cd8 | 2010-04-24 18:44:05 +0000 | [diff] [blame] | 2454 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2455 | ** End a write transaction. The commit has already been done. This |
| 2456 | ** routine merely releases the lock. |
| 2457 | */ |
| 2458 | int sqlite3WalEndWriteTransaction(Wal *pWal){ |
dan | da9fe0c | 2010-07-13 18:44:03 +0000 | [diff] [blame] | 2459 | if( pWal->writeLock ){ |
| 2460 | walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1); |
| 2461 | pWal->writeLock = 0; |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 2462 | pWal->truncateOnCommit = 0; |
dan | da9fe0c | 2010-07-13 18:44:03 +0000 | [diff] [blame] | 2463 | } |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2464 | return SQLITE_OK; |
| 2465 | } |
| 2466 | |
| 2467 | /* |
dan | 74d6cd8 | 2010-04-24 18:44:05 +0000 | [diff] [blame] | 2468 | ** If any data has been written (but not committed) to the log file, this |
| 2469 | ** function moves the write-pointer back to the start of the transaction. |
| 2470 | ** |
| 2471 | ** Additionally, the callback function is invoked for each frame written |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2472 | ** to the WAL since the start of the transaction. If the callback returns |
dan | 74d6cd8 | 2010-04-24 18:44:05 +0000 | [diff] [blame] | 2473 | ** other than SQLITE_OK, it is not invoked again and the error code is |
| 2474 | ** returned to the caller. |
| 2475 | ** |
| 2476 | ** Otherwise, if the callback function does not return an error, this |
| 2477 | ** function returns SQLITE_OK. |
| 2478 | */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2479 | int sqlite3WalUndo(Wal *pWal, int (*xUndo)(void *, Pgno), void *pUndoCtx){ |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2480 | int rc = SQLITE_OK; |
drh | 7e9e70b | 2010-08-16 14:17:59 +0000 | [diff] [blame] | 2481 | if( ALWAYS(pWal->writeLock) ){ |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 2482 | Pgno iMax = pWal->hdr.mxFrame; |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2483 | Pgno iFrame; |
| 2484 | |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2485 | /* Restore the clients cache of the wal-index header to the state it |
| 2486 | ** was in before the client began writing to the database. |
| 2487 | */ |
dan | 067f316 | 2010-06-14 10:30:12 +0000 | [diff] [blame] | 2488 | memcpy(&pWal->hdr, (void *)walIndexHdr(pWal), sizeof(WalIndexHdr)); |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2489 | |
| 2490 | for(iFrame=pWal->hdr.mxFrame+1; |
| 2491 | ALWAYS(rc==SQLITE_OK) && iFrame<=iMax; |
| 2492 | iFrame++ |
| 2493 | ){ |
| 2494 | /* This call cannot fail. Unless the page for which the page number |
| 2495 | ** is passed as the second argument is (a) in the cache and |
| 2496 | ** (b) has an outstanding reference, then xUndo is either a no-op |
| 2497 | ** (if (a) is false) or simply expels the page from the cache (if (b) |
| 2498 | ** is false). |
| 2499 | ** |
| 2500 | ** If the upper layer is doing a rollback, it is guaranteed that there |
| 2501 | ** are no outstanding references to any page other than page 1. And |
| 2502 | ** page 1 is never written to the log until the transaction is |
| 2503 | ** committed. As a result, the call to xUndo may not fail. |
| 2504 | */ |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2505 | assert( walFramePgno(pWal, iFrame)!=1 ); |
| 2506 | rc = xUndo(pUndoCtx, walFramePgno(pWal, iFrame)); |
dan | 6f15014 | 2010-05-21 15:31:56 +0000 | [diff] [blame] | 2507 | } |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2508 | walCleanupHash(pWal); |
dan | 74d6cd8 | 2010-04-24 18:44:05 +0000 | [diff] [blame] | 2509 | } |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2510 | assert( rc==SQLITE_OK ); |
dan | 74d6cd8 | 2010-04-24 18:44:05 +0000 | [diff] [blame] | 2511 | return rc; |
| 2512 | } |
| 2513 | |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2514 | /* |
| 2515 | ** Argument aWalData must point to an array of WAL_SAVEPOINT_NDATA u32 |
| 2516 | ** values. This function populates the array with values required to |
| 2517 | ** "rollback" the write position of the WAL handle back to the current |
| 2518 | ** point in the event of a savepoint rollback (via WalSavepointUndo()). |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2519 | */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2520 | void sqlite3WalSavepoint(Wal *pWal, u32 *aWalData){ |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2521 | assert( pWal->writeLock ); |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2522 | aWalData[0] = pWal->hdr.mxFrame; |
| 2523 | aWalData[1] = pWal->hdr.aFrameCksum[0]; |
| 2524 | aWalData[2] = pWal->hdr.aFrameCksum[1]; |
dan | 6e6bd56 | 2010-06-02 18:59:03 +0000 | [diff] [blame] | 2525 | aWalData[3] = pWal->nCkpt; |
dan | 4cd78b4 | 2010-04-26 16:57:10 +0000 | [diff] [blame] | 2526 | } |
| 2527 | |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2528 | /* |
| 2529 | ** Move the write position of the WAL back to the point identified by |
| 2530 | ** the values in the aWalData[] array. aWalData must point to an array |
| 2531 | ** of WAL_SAVEPOINT_NDATA u32 values that has been previously populated |
| 2532 | ** by a call to WalSavepoint(). |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2533 | */ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2534 | int sqlite3WalSavepointUndo(Wal *pWal, u32 *aWalData){ |
dan | 4cd78b4 | 2010-04-26 16:57:10 +0000 | [diff] [blame] | 2535 | int rc = SQLITE_OK; |
dan | 4cd78b4 | 2010-04-26 16:57:10 +0000 | [diff] [blame] | 2536 | |
dan | 6e6bd56 | 2010-06-02 18:59:03 +0000 | [diff] [blame] | 2537 | assert( pWal->writeLock ); |
| 2538 | assert( aWalData[3]!=pWal->nCkpt || aWalData[0]<=pWal->hdr.mxFrame ); |
| 2539 | |
| 2540 | if( aWalData[3]!=pWal->nCkpt ){ |
| 2541 | /* This savepoint was opened immediately after the write-transaction |
| 2542 | ** was started. Right after that, the writer decided to wrap around |
| 2543 | ** to the start of the log. Update the savepoint values to match. |
| 2544 | */ |
| 2545 | aWalData[0] = 0; |
| 2546 | aWalData[3] = pWal->nCkpt; |
| 2547 | } |
| 2548 | |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2549 | if( aWalData[0]<pWal->hdr.mxFrame ){ |
dan | 71d8991 | 2010-05-24 13:57:42 +0000 | [diff] [blame] | 2550 | pWal->hdr.mxFrame = aWalData[0]; |
| 2551 | pWal->hdr.aFrameCksum[0] = aWalData[1]; |
| 2552 | pWal->hdr.aFrameCksum[1] = aWalData[2]; |
dan | 5d65685 | 2010-06-14 07:53:26 +0000 | [diff] [blame] | 2553 | walCleanupHash(pWal); |
dan | 6f15014 | 2010-05-21 15:31:56 +0000 | [diff] [blame] | 2554 | } |
dan | 6e6bd56 | 2010-06-02 18:59:03 +0000 | [diff] [blame] | 2555 | |
dan | 4cd78b4 | 2010-04-26 16:57:10 +0000 | [diff] [blame] | 2556 | return rc; |
| 2557 | } |
| 2558 | |
drh | 8dd4afa | 2011-12-08 19:50:32 +0000 | [diff] [blame] | 2559 | |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2560 | /* |
| 2561 | ** This function is called just before writing a set of frames to the log |
| 2562 | ** file (see sqlite3WalFrames()). It checks to see if, instead of appending |
| 2563 | ** to the current log file, it is possible to overwrite the start of the |
| 2564 | ** existing log file with the new frames (i.e. "reset" the log). If so, |
| 2565 | ** it sets pWal->hdr.mxFrame to 0. Otherwise, pWal->hdr.mxFrame is left |
| 2566 | ** unchanged. |
| 2567 | ** |
| 2568 | ** SQLITE_OK is returned if no error is encountered (regardless of whether |
| 2569 | ** or not pWal->hdr.mxFrame is modified). An SQLite error code is returned |
drh | 4533cd0 | 2010-10-05 15:41:05 +0000 | [diff] [blame] | 2570 | ** if an error occurs. |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2571 | */ |
| 2572 | static int walRestartLog(Wal *pWal){ |
| 2573 | int rc = SQLITE_OK; |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2574 | int cnt; |
| 2575 | |
dan | 13a3cb8 | 2010-06-11 19:04:21 +0000 | [diff] [blame] | 2576 | if( pWal->readLock==0 ){ |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2577 | volatile WalCkptInfo *pInfo = walCkptInfo(pWal); |
| 2578 | assert( pInfo->nBackfill==pWal->hdr.mxFrame ); |
| 2579 | if( pInfo->nBackfill>0 ){ |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2580 | u32 salt1; |
| 2581 | sqlite3_randomness(4, &salt1); |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2582 | rc = walLockExclusive(pWal, WAL_READ_LOCK(1), WAL_NREADER-1); |
| 2583 | if( rc==SQLITE_OK ){ |
| 2584 | /* If all readers are using WAL_READ_LOCK(0) (in other words if no |
| 2585 | ** readers are currently using the WAL), then the transactions |
| 2586 | ** frames will overwrite the start of the existing log. Update the |
| 2587 | ** wal-index header to reflect this. |
| 2588 | ** |
| 2589 | ** In theory it would be Ok to update the cache of the header only |
| 2590 | ** at this point. But updating the actual wal-index header is also |
| 2591 | ** safe and means there is no special case for sqlite3WalUndo() |
| 2592 | ** to handle if this transaction is rolled back. |
| 2593 | */ |
dan | 199100e | 2010-06-09 16:58:49 +0000 | [diff] [blame] | 2594 | int i; /* Loop counter */ |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2595 | u32 *aSalt = pWal->hdr.aSalt; /* Big-endian salt values */ |
drh | 85a8375 | 2011-05-16 21:00:27 +0000 | [diff] [blame] | 2596 | |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2597 | pWal->nCkpt++; |
| 2598 | pWal->hdr.mxFrame = 0; |
| 2599 | sqlite3Put4byte((u8*)&aSalt[0], 1 + sqlite3Get4byte((u8*)&aSalt[0])); |
drh | 658d76c | 2011-02-19 15:22:14 +0000 | [diff] [blame] | 2600 | aSalt[1] = salt1; |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2601 | walIndexWriteHdr(pWal); |
dan | 199100e | 2010-06-09 16:58:49 +0000 | [diff] [blame] | 2602 | pInfo->nBackfill = 0; |
| 2603 | for(i=1; i<WAL_NREADER; i++) pInfo->aReadMark[i] = READMARK_NOT_USED; |
| 2604 | assert( pInfo->aReadMark[0]==0 ); |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2605 | walUnlockExclusive(pWal, WAL_READ_LOCK(1), WAL_NREADER-1); |
drh | 4533cd0 | 2010-10-05 15:41:05 +0000 | [diff] [blame] | 2606 | }else if( rc!=SQLITE_BUSY ){ |
| 2607 | return rc; |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2608 | } |
| 2609 | } |
| 2610 | walUnlockShared(pWal, WAL_READ_LOCK(0)); |
| 2611 | pWal->readLock = -1; |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2612 | cnt = 0; |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2613 | do{ |
| 2614 | int notUsed; |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2615 | rc = walTryBeginRead(pWal, ¬Used, 1, ++cnt); |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2616 | }while( rc==WAL_RETRY ); |
drh | c90e081 | 2011-02-19 17:02:44 +0000 | [diff] [blame] | 2617 | assert( (rc&0xff)!=SQLITE_BUSY ); /* BUSY not possible when useWal==1 */ |
drh | ab1cc74 | 2011-02-19 16:51:45 +0000 | [diff] [blame] | 2618 | testcase( (rc&0xff)==SQLITE_IOERR ); |
| 2619 | testcase( rc==SQLITE_PROTOCOL ); |
| 2620 | testcase( rc==SQLITE_OK ); |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2621 | } |
| 2622 | return rc; |
| 2623 | } |
| 2624 | |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2625 | /* |
| 2626 | ** Write iAmt bytes of content into the WAL file beginning at iOffset. |
| 2627 | ** |
| 2628 | ** When crossing the boundary between the first and second sectors of the |
| 2629 | ** file, first write all of the first sector content, then fsync(), then |
| 2630 | ** continue writing content for the second sector. This ensures that |
| 2631 | ** the WAL header is overwritten before the first commit mark. |
| 2632 | */ |
| 2633 | static int walWriteToLog( |
| 2634 | Wal *pWal, /* WAL to write to */ |
| 2635 | void *pContent, /* Content to be written */ |
| 2636 | int iAmt, /* Number of bytes to write */ |
| 2637 | sqlite3_int64 iOffset /* Start writing at this offset */ |
| 2638 | ){ |
| 2639 | int rc; |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 2640 | if( iOffset>=pWal->szFirstBlock |
| 2641 | || iOffset+iAmt<pWal->szFirstBlock |
| 2642 | || pWal->syncFlags==0 |
| 2643 | ){ |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2644 | /* The common and fast case. Just write the data. */ |
| 2645 | rc = sqlite3OsWrite(pWal->pWalFd, pContent, iAmt, iOffset); |
| 2646 | }else{ |
| 2647 | /* If this write will cross the first sector boundary, it has to |
| 2648 | ** be split it two with a sync in between. */ |
| 2649 | int iFirstAmt = pWal->szFirstBlock - iOffset; |
| 2650 | assert( iFirstAmt>0 && iFirstAmt<iAmt ); |
| 2651 | rc = sqlite3OsWrite(pWal->pWalFd, pContent, iFirstAmt, iOffset); |
| 2652 | if( rc ) return rc; |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 2653 | assert( pWal->syncFlags & (SQLITE_SYNC_NORMAL|SQLITE_SYNC_FULL) ); |
| 2654 | rc = sqlite3OsSync(pWal->pWalFd, pWal->syncFlags); |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2655 | if( rc ) return rc; |
| 2656 | pContent = (void*)(iFirstAmt + (char*)pContent); |
| 2657 | rc = sqlite3OsWrite(pWal->pWalFd, pContent, |
| 2658 | iAmt-iFirstAmt, iOffset+iFirstAmt); |
| 2659 | } |
| 2660 | return rc; |
| 2661 | } |
| 2662 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2663 | /* |
dan | 4cd78b4 | 2010-04-26 16:57:10 +0000 | [diff] [blame] | 2664 | ** Write a set of frames to the log. The caller must hold the write-lock |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2665 | ** on the log file (obtained using sqlite3WalBeginWriteTransaction()). |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2666 | */ |
drh | c438efd | 2010-04-26 00:19:45 +0000 | [diff] [blame] | 2667 | int sqlite3WalFrames( |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2668 | Wal *pWal, /* Wal handle to write to */ |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 2669 | int szPage, /* Database page-size in bytes */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2670 | PgHdr *pList, /* List of dirty pages to write */ |
| 2671 | Pgno nTruncate, /* Database size after this commit */ |
| 2672 | int isCommit, /* True if this is a commit */ |
dan | c511878 | 2010-04-17 17:34:41 +0000 | [diff] [blame] | 2673 | int sync_flags /* Flags to pass to OsSync() (or 0) */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2674 | ){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2675 | int rc; /* Used to catch return codes */ |
| 2676 | u32 iFrame; /* Next frame address */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2677 | u8 aFrame[WAL_FRAME_HDRSIZE]; /* Buffer to assemble frame-header in */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2678 | PgHdr *p; /* Iterator to run through pList with. */ |
drh | e874d9e | 2010-05-07 20:02:23 +0000 | [diff] [blame] | 2679 | PgHdr *pLast = 0; /* Last frame in list */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2680 | int nLast = 0; /* Number of extra copies of last page */ |
| 2681 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2682 | assert( pList ); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2683 | assert( pWal->writeLock ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2684 | |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 2685 | #if defined(SQLITE_TEST) && defined(SQLITE_DEBUG) |
| 2686 | { int cnt; for(cnt=0, p=pList; p; p=p->pDirty, cnt++){} |
| 2687 | WALTRACE(("WAL%p: frame write begin. %d frames. mxFrame=%d. %s\n", |
| 2688 | pWal, cnt, pWal->hdr.mxFrame, isCommit ? "Commit" : "Spill")); |
| 2689 | } |
| 2690 | #endif |
| 2691 | |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2692 | /* See if it is possible to write these frames into the start of the |
| 2693 | ** log file, instead of appending to it at pWal->hdr.mxFrame. |
| 2694 | */ |
| 2695 | if( SQLITE_OK!=(rc = walRestartLog(pWal)) ){ |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2696 | return rc; |
| 2697 | } |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2698 | |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 2699 | /* If this is the first frame written into the log, write the WAL |
| 2700 | ** header to the start of the WAL file. See comments at the top of |
| 2701 | ** this source file for a description of the WAL header format. |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 2702 | */ |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 2703 | iFrame = pWal->hdr.mxFrame; |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 2704 | if( iFrame==0 ){ |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 2705 | u8 aWalHdr[WAL_HDRSIZE]; /* Buffer to assemble wal-header in */ |
| 2706 | u32 aCksum[2]; /* Checksum for wal-header */ |
| 2707 | |
dan | b8fd6c2 | 2010-05-24 10:39:36 +0000 | [diff] [blame] | 2708 | sqlite3Put4byte(&aWalHdr[0], (WAL_MAGIC | SQLITE_BIGENDIAN)); |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 2709 | sqlite3Put4byte(&aWalHdr[4], WAL_MAX_VERSION); |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 2710 | sqlite3Put4byte(&aWalHdr[8], szPage); |
| 2711 | sqlite3Put4byte(&aWalHdr[12], pWal->nCkpt); |
drh | 2327f5a | 2010-07-07 21:06:48 +0000 | [diff] [blame] | 2712 | sqlite3_randomness(8, pWal->hdr.aSalt); |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 2713 | memcpy(&aWalHdr[16], pWal->hdr.aSalt, 8); |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 2714 | walChecksumBytes(1, aWalHdr, WAL_HDRSIZE-2*4, 0, aCksum); |
| 2715 | sqlite3Put4byte(&aWalHdr[24], aCksum[0]); |
| 2716 | sqlite3Put4byte(&aWalHdr[28], aCksum[1]); |
| 2717 | |
drh | b2eced5 | 2010-08-12 02:41:12 +0000 | [diff] [blame] | 2718 | pWal->szPage = szPage; |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 2719 | pWal->hdr.bigEndCksum = SQLITE_BIGENDIAN; |
| 2720 | pWal->hdr.aFrameCksum[0] = aCksum[0]; |
| 2721 | pWal->hdr.aFrameCksum[1] = aCksum[1]; |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 2722 | pWal->truncateOnCommit = 1; |
dan | 10f5a50 | 2010-06-23 15:55:43 +0000 | [diff] [blame] | 2723 | |
drh | 23ea97b | 2010-05-20 16:45:58 +0000 | [diff] [blame] | 2724 | rc = sqlite3OsWrite(pWal->pWalFd, aWalHdr, sizeof(aWalHdr), 0); |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 2725 | WALTRACE(("WAL%p: wal-header write %s\n", pWal, rc ? "failed" : "ok")); |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 2726 | if( rc!=SQLITE_OK ){ |
| 2727 | return rc; |
| 2728 | } |
| 2729 | } |
shaneh | bd2aaf9 | 2010-09-01 02:38:21 +0000 | [diff] [blame] | 2730 | assert( (int)pWal->szPage==szPage ); |
dan | 97a3135 | 2010-04-16 13:59:31 +0000 | [diff] [blame] | 2731 | |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 2732 | /* Setup information needed to do the WAL header sync */ |
| 2733 | if( pWal->noSyncHeader ){ |
| 2734 | assert( pWal->szFirstBlock==0 ); |
| 2735 | assert( pWal->syncFlags==0 ); |
| 2736 | }else{ |
| 2737 | pWal->szFirstBlock = sqlite3OsSectorSize(pWal->pWalFd); |
| 2738 | if( szPage>pWal->szFirstBlock ) pWal->szFirstBlock = szPage; |
| 2739 | pWal->syncFlags = sync_flags & SQLITE_SYNC_MASK; |
| 2740 | } |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2741 | |
dan | 9971e71 | 2010-06-01 15:44:57 +0000 | [diff] [blame] | 2742 | /* Write the log file. */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2743 | for(p=pList; p; p=p->pDirty){ |
| 2744 | u32 nDbsize; /* Db-size field for frame header */ |
| 2745 | i64 iOffset; /* Write offset in log file */ |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2746 | void *pData; |
| 2747 | |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 2748 | iOffset = walFrameOffset(++iFrame, szPage); |
drh | e9187b4 | 2010-07-07 14:39:59 +0000 | [diff] [blame] | 2749 | /* testcase( IS_BIG_INT(iOffset) ); // requires a 4GiB WAL */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2750 | |
| 2751 | /* Populate and write the frame header */ |
| 2752 | nDbsize = (isCommit && p->pDirty==0) ? nTruncate : 0; |
drh | a715211 | 2010-06-22 21:15:49 +0000 | [diff] [blame] | 2753 | #if defined(SQLITE_HAS_CODEC) |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2754 | if( (pData = sqlite3PagerCodec(p))==0 ) return SQLITE_NOMEM; |
drh | a715211 | 2010-06-22 21:15:49 +0000 | [diff] [blame] | 2755 | #else |
| 2756 | pData = p->pData; |
| 2757 | #endif |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2758 | walEncodeFrame(pWal, p->pgno, nDbsize, pData, aFrame); |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2759 | rc = walWriteToLog(pWal, aFrame, sizeof(aFrame), iOffset); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2760 | if( rc!=SQLITE_OK ){ |
| 2761 | return rc; |
| 2762 | } |
| 2763 | |
| 2764 | /* Write the page data */ |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2765 | rc = walWriteToLog(pWal, pData, szPage, iOffset+sizeof(aFrame)); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2766 | if( rc!=SQLITE_OK ){ |
| 2767 | return rc; |
| 2768 | } |
| 2769 | pLast = p; |
| 2770 | } |
| 2771 | |
| 2772 | /* Sync the log file if the 'isSync' flag was specified. */ |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 2773 | if( isCommit && (sync_flags & WAL_SYNC_TRANSACTIONS)!=0 ){ |
drh | d9e5c4f | 2010-05-12 18:01:39 +0000 | [diff] [blame] | 2774 | i64 iSegment = sqlite3OsSectorSize(pWal->pWalFd); |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 2775 | i64 iOffset = walFrameOffset(iFrame+1, szPage); |
dan | 6703239 | 2010-04-17 15:42:43 +0000 | [diff] [blame] | 2776 | |
drh | 69c4696 | 2010-05-17 20:16:50 +0000 | [diff] [blame] | 2777 | assert( iSegment>0 ); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2778 | |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2779 | iSegment = (((iOffset+iSegment-1)/iSegment) * iSegment); |
| 2780 | while( iOffset<iSegment ){ |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2781 | void *pData; |
drh | a715211 | 2010-06-22 21:15:49 +0000 | [diff] [blame] | 2782 | #if defined(SQLITE_HAS_CODEC) |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2783 | if( (pData = sqlite3PagerCodec(pLast))==0 ) return SQLITE_NOMEM; |
drh | a715211 | 2010-06-22 21:15:49 +0000 | [diff] [blame] | 2784 | #else |
| 2785 | pData = pLast->pData; |
| 2786 | #endif |
dan | 47ee386 | 2010-06-22 15:18:44 +0000 | [diff] [blame] | 2787 | walEncodeFrame(pWal, pLast->pgno, nTruncate, pData, aFrame); |
drh | e9187b4 | 2010-07-07 14:39:59 +0000 | [diff] [blame] | 2788 | /* testcase( IS_BIG_INT(iOffset) ); // requires a 4GiB WAL */ |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2789 | rc = walWriteToLog(pWal, aFrame, sizeof(aFrame), iOffset); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2790 | if( rc!=SQLITE_OK ){ |
| 2791 | return rc; |
| 2792 | } |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2793 | iOffset += WAL_FRAME_HDRSIZE; |
drh | 88f975a | 2011-12-16 19:34:36 +0000 | [diff] [blame] | 2794 | rc = walWriteToLog(pWal, pData, szPage, iOffset); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2795 | if( rc!=SQLITE_OK ){ |
| 2796 | return rc; |
| 2797 | } |
| 2798 | nLast++; |
drh | 6e81096 | 2010-05-19 17:49:50 +0000 | [diff] [blame] | 2799 | iOffset += szPage; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2800 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2801 | |
drh | 4eb02a4 | 2011-12-16 21:26:26 +0000 | [diff] [blame] | 2802 | rc = sqlite3OsSync(pWal->pWalFd, sync_flags & SQLITE_SYNC_MASK); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2803 | } |
| 2804 | |
dan | f60b7f3 | 2011-12-16 13:24:27 +0000 | [diff] [blame] | 2805 | if( isCommit && pWal->truncateOnCommit && pWal->mxWalSize>=0 ){ |
| 2806 | i64 sz = pWal->mxWalSize; |
| 2807 | if( walFrameOffset(iFrame+nLast+1, szPage)>pWal->mxWalSize ){ |
| 2808 | sz = walFrameOffset(iFrame+nLast+1, szPage); |
| 2809 | } |
| 2810 | walLimitSize(pWal, sz); |
| 2811 | pWal->truncateOnCommit = 0; |
| 2812 | } |
| 2813 | |
drh | e730fec | 2010-05-18 12:56:50 +0000 | [diff] [blame] | 2814 | /* Append data to the wal-index. It is not necessary to lock the |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 2815 | ** wal-index to do this as the SQLITE_SHM_WRITE lock held on the wal-index |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2816 | ** guarantees that there are no other writers, and no data that may |
| 2817 | ** be in use by existing readers is being overwritten. |
| 2818 | */ |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 2819 | iFrame = pWal->hdr.mxFrame; |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2820 | for(p=pList; p && rc==SQLITE_OK; p=p->pDirty){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2821 | iFrame++; |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2822 | rc = walIndexAppend(pWal, iFrame, p->pgno); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2823 | } |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2824 | while( nLast>0 && rc==SQLITE_OK ){ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2825 | iFrame++; |
| 2826 | nLast--; |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2827 | rc = walIndexAppend(pWal, iFrame, pLast->pgno); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2828 | } |
| 2829 | |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2830 | if( rc==SQLITE_OK ){ |
| 2831 | /* Update the private copy of the header. */ |
shaneh | 1df2db7 | 2010-08-18 02:28:48 +0000 | [diff] [blame] | 2832 | pWal->hdr.szPage = (u16)((szPage&0xff00) | (szPage>>16)); |
drh | 9b78f79 | 2010-08-14 21:21:24 +0000 | [diff] [blame] | 2833 | testcase( szPage<=32768 ); |
| 2834 | testcase( szPage>=65536 ); |
drh | 027a128 | 2010-05-19 01:53:53 +0000 | [diff] [blame] | 2835 | pWal->hdr.mxFrame = iFrame; |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2836 | if( isCommit ){ |
| 2837 | pWal->hdr.iChange++; |
| 2838 | pWal->hdr.nPage = nTruncate; |
| 2839 | } |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2840 | /* If this is a commit, update the wal-index header too. */ |
| 2841 | if( isCommit ){ |
drh | 7e26372 | 2010-05-20 21:21:09 +0000 | [diff] [blame] | 2842 | walIndexWriteHdr(pWal); |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2843 | pWal->iCallback = iFrame; |
| 2844 | } |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2845 | } |
dan | c7991bd | 2010-05-05 19:04:59 +0000 | [diff] [blame] | 2846 | |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 2847 | WALTRACE(("WAL%p: frame write %s\n", pWal, rc ? "failed" : "ok")); |
dan | 8d22a17 | 2010-04-19 18:03:51 +0000 | [diff] [blame] | 2848 | return rc; |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2849 | } |
| 2850 | |
| 2851 | /* |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2852 | ** This routine is called to implement sqlite3_wal_checkpoint() and |
| 2853 | ** related interfaces. |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2854 | ** |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2855 | ** Obtain a CHECKPOINT lock and then backfill as much information as |
| 2856 | ** we can from WAL into the database. |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2857 | ** |
| 2858 | ** If parameter xBusy is not NULL, it is a pointer to a busy-handler |
| 2859 | ** callback. In this case this function runs a blocking checkpoint. |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2860 | */ |
drh | c438efd | 2010-04-26 00:19:45 +0000 | [diff] [blame] | 2861 | int sqlite3WalCheckpoint( |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2862 | Wal *pWal, /* Wal connection */ |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 2863 | int eMode, /* PASSIVE, FULL or RESTART */ |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2864 | int (*xBusy)(void*), /* Function to call when busy */ |
| 2865 | void *pBusyArg, /* Context argument for xBusyHandler */ |
dan | c511878 | 2010-04-17 17:34:41 +0000 | [diff] [blame] | 2866 | int sync_flags, /* Flags to sync db file with (or 0) */ |
dan | b6e099a | 2010-05-04 14:47:39 +0000 | [diff] [blame] | 2867 | int nBuf, /* Size of temporary buffer */ |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 2868 | u8 *zBuf, /* Temporary buffer to use */ |
| 2869 | int *pnLog, /* OUT: Number of frames in WAL */ |
| 2870 | int *pnCkpt /* OUT: Number of backfilled frames in WAL */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2871 | ){ |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2872 | int rc; /* Return code */ |
dan | 31c0390 | 2010-04-29 14:51:33 +0000 | [diff] [blame] | 2873 | int isChanged = 0; /* True if a new wal-index header is loaded */ |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2874 | int eMode2 = eMode; /* Mode to pass to walCheckpoint() */ |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2875 | |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 2876 | assert( pWal->ckptLock==0 ); |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2877 | assert( pWal->writeLock==0 ); |
dan | 39c79f5 | 2010-04-15 10:58:51 +0000 | [diff] [blame] | 2878 | |
drh | 66dfec8b | 2011-06-01 20:01:49 +0000 | [diff] [blame] | 2879 | if( pWal->readOnly ) return SQLITE_READONLY; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 2880 | WALTRACE(("WAL%p: checkpoint begins\n", pWal)); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2881 | rc = walLockExclusive(pWal, WAL_CKPT_LOCK, 1); |
| 2882 | if( rc ){ |
| 2883 | /* Usually this is SQLITE_BUSY meaning that another thread or process |
| 2884 | ** is already running a checkpoint, or maybe a recovery. But it might |
| 2885 | ** also be SQLITE_IOERR. */ |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2886 | return rc; |
| 2887 | } |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 2888 | pWal->ckptLock = 1; |
dan | 64d039e | 2010-04-13 19:27:31 +0000 | [diff] [blame] | 2889 | |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2890 | /* If this is a blocking-checkpoint, then obtain the write-lock as well |
| 2891 | ** to prevent any writers from running while the checkpoint is underway. |
| 2892 | ** This has to be done before the call to walIndexReadHdr() below. |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2893 | ** |
| 2894 | ** If the writer lock cannot be obtained, then a passive checkpoint is |
| 2895 | ** run instead. Since the checkpointer is not holding the writer lock, |
| 2896 | ** there is no point in blocking waiting for any readers. Assuming no |
| 2897 | ** other error occurs, this function will return SQLITE_BUSY to the caller. |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2898 | */ |
dan | cdc1f04 | 2010-11-18 12:11:05 +0000 | [diff] [blame] | 2899 | if( eMode!=SQLITE_CHECKPOINT_PASSIVE ){ |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2900 | rc = walBusyLock(pWal, xBusy, pBusyArg, WAL_WRITE_LOCK, 1); |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2901 | if( rc==SQLITE_OK ){ |
| 2902 | pWal->writeLock = 1; |
| 2903 | }else if( rc==SQLITE_BUSY ){ |
| 2904 | eMode2 = SQLITE_CHECKPOINT_PASSIVE; |
| 2905 | rc = SQLITE_OK; |
| 2906 | } |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2907 | } |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2908 | |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2909 | /* Read the wal-index header. */ |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2910 | if( rc==SQLITE_OK ){ |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2911 | rc = walIndexReadHdr(pWal, &isChanged); |
| 2912 | } |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2913 | |
| 2914 | /* Copy data from the log to the database file. */ |
dan | 9c5e368 | 2011-02-07 15:12:12 +0000 | [diff] [blame] | 2915 | if( rc==SQLITE_OK ){ |
| 2916 | if( pWal->hdr.mxFrame && walPagesize(pWal)!=nBuf ){ |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2917 | rc = SQLITE_CORRUPT_BKPT; |
| 2918 | }else{ |
dan | 9c5e368 | 2011-02-07 15:12:12 +0000 | [diff] [blame] | 2919 | rc = walCheckpoint(pWal, eMode2, xBusy, pBusyArg, sync_flags, zBuf); |
| 2920 | } |
| 2921 | |
| 2922 | /* If no error occurred, set the output variables. */ |
| 2923 | if( rc==SQLITE_OK || rc==SQLITE_BUSY ){ |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2924 | if( pnLog ) *pnLog = (int)pWal->hdr.mxFrame; |
dan | 9c5e368 | 2011-02-07 15:12:12 +0000 | [diff] [blame] | 2925 | if( pnCkpt ) *pnCkpt = (int)(walCkptInfo(pWal)->nBackfill); |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2926 | } |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2927 | } |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2928 | |
dan | 31c0390 | 2010-04-29 14:51:33 +0000 | [diff] [blame] | 2929 | if( isChanged ){ |
| 2930 | /* If a new wal-index header was loaded before the checkpoint was |
drh | a2a4201 | 2010-05-18 18:01:08 +0000 | [diff] [blame] | 2931 | ** performed, then the pager-cache associated with pWal is now |
dan | 31c0390 | 2010-04-29 14:51:33 +0000 | [diff] [blame] | 2932 | ** out of date. So zero the cached wal-index header to ensure that |
| 2933 | ** next time the pager opens a snapshot on this database it knows that |
| 2934 | ** the cache needs to be reset. |
| 2935 | */ |
| 2936 | memset(&pWal->hdr, 0, sizeof(WalIndexHdr)); |
| 2937 | } |
dan | b9bf16b | 2010-04-14 11:23:30 +0000 | [diff] [blame] | 2938 | |
| 2939 | /* Release the locks. */ |
dan | a58f26f | 2010-11-16 18:56:51 +0000 | [diff] [blame] | 2940 | sqlite3WalEndWriteTransaction(pWal); |
drh | 73b64e4 | 2010-05-30 19:55:15 +0000 | [diff] [blame] | 2941 | walUnlockExclusive(pWal, WAL_CKPT_LOCK, 1); |
dan | d54ff60 | 2010-05-31 11:16:30 +0000 | [diff] [blame] | 2942 | pWal->ckptLock = 0; |
drh | c74c333 | 2010-05-31 12:15:19 +0000 | [diff] [blame] | 2943 | WALTRACE(("WAL%p: checkpoint %s\n", pWal, rc ? "failed" : "ok")); |
dan | f2b8dd5 | 2010-11-18 19:28:01 +0000 | [diff] [blame] | 2944 | return (rc==SQLITE_OK && eMode!=eMode2 ? SQLITE_BUSY : rc); |
dan | 7c24610 | 2010-04-12 19:00:29 +0000 | [diff] [blame] | 2945 | } |
| 2946 | |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2947 | /* Return the value to pass to a sqlite3_wal_hook callback, the |
| 2948 | ** number of frames in the WAL at the point of the last commit since |
| 2949 | ** sqlite3WalCallback() was called. If no commits have occurred since |
| 2950 | ** the last call, then return 0. |
| 2951 | */ |
| 2952 | int sqlite3WalCallback(Wal *pWal){ |
dan | 8d22a17 | 2010-04-19 18:03:51 +0000 | [diff] [blame] | 2953 | u32 ret = 0; |
drh | 7ed91f2 | 2010-04-29 22:34:07 +0000 | [diff] [blame] | 2954 | if( pWal ){ |
| 2955 | ret = pWal->iCallback; |
| 2956 | pWal->iCallback = 0; |
dan | 8d22a17 | 2010-04-19 18:03:51 +0000 | [diff] [blame] | 2957 | } |
| 2958 | return (int)ret; |
| 2959 | } |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2960 | |
| 2961 | /* |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2962 | ** This function is called to change the WAL subsystem into or out |
| 2963 | ** of locking_mode=EXCLUSIVE. |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2964 | ** |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2965 | ** If op is zero, then attempt to change from locking_mode=EXCLUSIVE |
| 2966 | ** into locking_mode=NORMAL. This means that we must acquire a lock |
| 2967 | ** on the pWal->readLock byte. If the WAL is already in locking_mode=NORMAL |
| 2968 | ** or if the acquisition of the lock fails, then return 0. If the |
| 2969 | ** transition out of exclusive-mode is successful, return 1. This |
| 2970 | ** operation must occur while the pager is still holding the exclusive |
| 2971 | ** lock on the main database file. |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2972 | ** |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2973 | ** If op is one, then change from locking_mode=NORMAL into |
| 2974 | ** locking_mode=EXCLUSIVE. This means that the pWal->readLock must |
| 2975 | ** be released. Return 1 if the transition is made and 0 if the |
| 2976 | ** WAL is already in exclusive-locking mode - meaning that this |
| 2977 | ** routine is a no-op. The pager must already hold the exclusive lock |
| 2978 | ** on the main database file before invoking this operation. |
| 2979 | ** |
| 2980 | ** If op is negative, then do a dry-run of the op==1 case but do |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 2981 | ** not actually change anything. The pager uses this to see if it |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2982 | ** should acquire the database exclusive lock prior to invoking |
| 2983 | ** the op==1 case. |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 2984 | */ |
| 2985 | int sqlite3WalExclusiveMode(Wal *pWal, int op){ |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2986 | int rc; |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2987 | assert( pWal->writeLock==0 ); |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 2988 | assert( pWal->exclusiveMode!=WAL_HEAPMEMORY_MODE || op==-1 ); |
dan | 3cac5dc | 2010-06-04 18:37:59 +0000 | [diff] [blame] | 2989 | |
| 2990 | /* pWal->readLock is usually set, but might be -1 if there was a |
| 2991 | ** prior error while attempting to acquire are read-lock. This cannot |
| 2992 | ** happen if the connection is actually in exclusive mode (as no xShmLock |
| 2993 | ** locks are taken in this case). Nor should the pager attempt to |
| 2994 | ** upgrade to exclusive-mode following such an error. |
| 2995 | */ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 2996 | assert( pWal->readLock>=0 || pWal->lockError ); |
dan | 3cac5dc | 2010-06-04 18:37:59 +0000 | [diff] [blame] | 2997 | assert( pWal->readLock>=0 || (op<=0 && pWal->exclusiveMode==0) ); |
| 2998 | |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 2999 | if( op==0 ){ |
| 3000 | if( pWal->exclusiveMode ){ |
| 3001 | pWal->exclusiveMode = 0; |
dan | 3cac5dc | 2010-06-04 18:37:59 +0000 | [diff] [blame] | 3002 | if( walLockShared(pWal, WAL_READ_LOCK(pWal->readLock))!=SQLITE_OK ){ |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 3003 | pWal->exclusiveMode = 1; |
| 3004 | } |
| 3005 | rc = pWal->exclusiveMode==0; |
| 3006 | }else{ |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 3007 | /* Already in locking_mode=NORMAL */ |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 3008 | rc = 0; |
| 3009 | } |
| 3010 | }else if( op>0 ){ |
| 3011 | assert( pWal->exclusiveMode==0 ); |
drh | aab4c02 | 2010-06-02 14:45:51 +0000 | [diff] [blame] | 3012 | assert( pWal->readLock>=0 ); |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 3013 | walUnlockShared(pWal, WAL_READ_LOCK(pWal->readLock)); |
| 3014 | pWal->exclusiveMode = 1; |
| 3015 | rc = 1; |
| 3016 | }else{ |
| 3017 | rc = pWal->exclusiveMode==0; |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 3018 | } |
drh | 61e4ace | 2010-05-31 20:28:37 +0000 | [diff] [blame] | 3019 | return rc; |
dan | 5543759 | 2010-05-11 12:19:26 +0000 | [diff] [blame] | 3020 | } |
| 3021 | |
dan | 8c40800 | 2010-11-01 17:38:24 +0000 | [diff] [blame] | 3022 | /* |
| 3023 | ** Return true if the argument is non-NULL and the WAL module is using |
| 3024 | ** heap-memory for the wal-index. Otherwise, if the argument is NULL or the |
| 3025 | ** WAL module is using shared-memory, return false. |
| 3026 | */ |
| 3027 | int sqlite3WalHeapMemory(Wal *pWal){ |
| 3028 | return (pWal && pWal->exclusiveMode==WAL_HEAPMEMORY_MODE ); |
| 3029 | } |
| 3030 | |
dan | 5cf5353 | 2010-05-01 16:40:20 +0000 | [diff] [blame] | 3031 | #endif /* #ifndef SQLITE_OMIT_WAL */ |