khenaidoo | ab1f7bd | 2019-11-14 14:00:27 -0500 | [diff] [blame] | 1 | package bbolt |
| 2 | |
| 3 | import ( |
| 4 | "bytes" |
| 5 | "fmt" |
| 6 | "unsafe" |
| 7 | ) |
| 8 | |
| 9 | const ( |
| 10 | // MaxKeySize is the maximum length of a key, in bytes. |
| 11 | MaxKeySize = 32768 |
| 12 | |
| 13 | // MaxValueSize is the maximum length of a value, in bytes. |
| 14 | MaxValueSize = (1 << 31) - 2 |
| 15 | ) |
| 16 | |
| 17 | const bucketHeaderSize = int(unsafe.Sizeof(bucket{})) |
| 18 | |
| 19 | const ( |
| 20 | minFillPercent = 0.1 |
| 21 | maxFillPercent = 1.0 |
| 22 | ) |
| 23 | |
| 24 | // DefaultFillPercent is the percentage that split pages are filled. |
| 25 | // This value can be changed by setting Bucket.FillPercent. |
| 26 | const DefaultFillPercent = 0.5 |
| 27 | |
| 28 | // Bucket represents a collection of key/value pairs inside the database. |
| 29 | type Bucket struct { |
| 30 | *bucket |
| 31 | tx *Tx // the associated transaction |
| 32 | buckets map[string]*Bucket // subbucket cache |
| 33 | page *page // inline page reference |
| 34 | rootNode *node // materialized node for the root page. |
| 35 | nodes map[pgid]*node // node cache |
| 36 | |
| 37 | // Sets the threshold for filling nodes when they split. By default, |
| 38 | // the bucket will fill to 50% but it can be useful to increase this |
| 39 | // amount if you know that your write workloads are mostly append-only. |
| 40 | // |
| 41 | // This is non-persisted across transactions so it must be set in every Tx. |
| 42 | FillPercent float64 |
| 43 | } |
| 44 | |
| 45 | // bucket represents the on-file representation of a bucket. |
| 46 | // This is stored as the "value" of a bucket key. If the bucket is small enough, |
| 47 | // then its root page can be stored inline in the "value", after the bucket |
| 48 | // header. In the case of inline buckets, the "root" will be 0. |
| 49 | type bucket struct { |
| 50 | root pgid // page id of the bucket's root-level page |
| 51 | sequence uint64 // monotonically incrementing, used by NextSequence() |
| 52 | } |
| 53 | |
| 54 | // newBucket returns a new bucket associated with a transaction. |
| 55 | func newBucket(tx *Tx) Bucket { |
| 56 | var b = Bucket{tx: tx, FillPercent: DefaultFillPercent} |
| 57 | if tx.writable { |
| 58 | b.buckets = make(map[string]*Bucket) |
| 59 | b.nodes = make(map[pgid]*node) |
| 60 | } |
| 61 | return b |
| 62 | } |
| 63 | |
| 64 | // Tx returns the tx of the bucket. |
| 65 | func (b *Bucket) Tx() *Tx { |
| 66 | return b.tx |
| 67 | } |
| 68 | |
| 69 | // Root returns the root of the bucket. |
| 70 | func (b *Bucket) Root() pgid { |
| 71 | return b.root |
| 72 | } |
| 73 | |
| 74 | // Writable returns whether the bucket is writable. |
| 75 | func (b *Bucket) Writable() bool { |
| 76 | return b.tx.writable |
| 77 | } |
| 78 | |
| 79 | // Cursor creates a cursor associated with the bucket. |
| 80 | // The cursor is only valid as long as the transaction is open. |
| 81 | // Do not use a cursor after the transaction is closed. |
| 82 | func (b *Bucket) Cursor() *Cursor { |
| 83 | // Update transaction statistics. |
| 84 | b.tx.stats.CursorCount++ |
| 85 | |
| 86 | // Allocate and return a cursor. |
| 87 | return &Cursor{ |
| 88 | bucket: b, |
| 89 | stack: make([]elemRef, 0), |
| 90 | } |
| 91 | } |
| 92 | |
| 93 | // Bucket retrieves a nested bucket by name. |
| 94 | // Returns nil if the bucket does not exist. |
| 95 | // The bucket instance is only valid for the lifetime of the transaction. |
| 96 | func (b *Bucket) Bucket(name []byte) *Bucket { |
| 97 | if b.buckets != nil { |
| 98 | if child := b.buckets[string(name)]; child != nil { |
| 99 | return child |
| 100 | } |
| 101 | } |
| 102 | |
| 103 | // Move cursor to key. |
| 104 | c := b.Cursor() |
| 105 | k, v, flags := c.seek(name) |
| 106 | |
| 107 | // Return nil if the key doesn't exist or it is not a bucket. |
| 108 | if !bytes.Equal(name, k) || (flags&bucketLeafFlag) == 0 { |
| 109 | return nil |
| 110 | } |
| 111 | |
| 112 | // Otherwise create a bucket and cache it. |
| 113 | var child = b.openBucket(v) |
| 114 | if b.buckets != nil { |
| 115 | b.buckets[string(name)] = child |
| 116 | } |
| 117 | |
| 118 | return child |
| 119 | } |
| 120 | |
| 121 | // Helper method that re-interprets a sub-bucket value |
| 122 | // from a parent into a Bucket |
| 123 | func (b *Bucket) openBucket(value []byte) *Bucket { |
| 124 | var child = newBucket(b.tx) |
| 125 | |
| 126 | // If unaligned load/stores are broken on this arch and value is |
| 127 | // unaligned simply clone to an aligned byte array. |
| 128 | unaligned := brokenUnaligned && uintptr(unsafe.Pointer(&value[0]))&3 != 0 |
| 129 | |
| 130 | if unaligned { |
| 131 | value = cloneBytes(value) |
| 132 | } |
| 133 | |
| 134 | // If this is a writable transaction then we need to copy the bucket entry. |
| 135 | // Read-only transactions can point directly at the mmap entry. |
| 136 | if b.tx.writable && !unaligned { |
| 137 | child.bucket = &bucket{} |
| 138 | *child.bucket = *(*bucket)(unsafe.Pointer(&value[0])) |
| 139 | } else { |
| 140 | child.bucket = (*bucket)(unsafe.Pointer(&value[0])) |
| 141 | } |
| 142 | |
| 143 | // Save a reference to the inline page if the bucket is inline. |
| 144 | if child.root == 0 { |
| 145 | child.page = (*page)(unsafe.Pointer(&value[bucketHeaderSize])) |
| 146 | } |
| 147 | |
| 148 | return &child |
| 149 | } |
| 150 | |
| 151 | // CreateBucket creates a new bucket at the given key and returns the new bucket. |
| 152 | // Returns an error if the key already exists, if the bucket name is blank, or if the bucket name is too long. |
| 153 | // The bucket instance is only valid for the lifetime of the transaction. |
| 154 | func (b *Bucket) CreateBucket(key []byte) (*Bucket, error) { |
| 155 | if b.tx.db == nil { |
| 156 | return nil, ErrTxClosed |
| 157 | } else if !b.tx.writable { |
| 158 | return nil, ErrTxNotWritable |
| 159 | } else if len(key) == 0 { |
| 160 | return nil, ErrBucketNameRequired |
| 161 | } |
| 162 | |
| 163 | // Move cursor to correct position. |
| 164 | c := b.Cursor() |
| 165 | k, _, flags := c.seek(key) |
| 166 | |
| 167 | // Return an error if there is an existing key. |
| 168 | if bytes.Equal(key, k) { |
| 169 | if (flags & bucketLeafFlag) != 0 { |
| 170 | return nil, ErrBucketExists |
| 171 | } |
| 172 | return nil, ErrIncompatibleValue |
| 173 | } |
| 174 | |
| 175 | // Create empty, inline bucket. |
| 176 | var bucket = Bucket{ |
| 177 | bucket: &bucket{}, |
| 178 | rootNode: &node{isLeaf: true}, |
| 179 | FillPercent: DefaultFillPercent, |
| 180 | } |
| 181 | var value = bucket.write() |
| 182 | |
| 183 | // Insert into node. |
| 184 | key = cloneBytes(key) |
| 185 | c.node().put(key, key, value, 0, bucketLeafFlag) |
| 186 | |
| 187 | // Since subbuckets are not allowed on inline buckets, we need to |
| 188 | // dereference the inline page, if it exists. This will cause the bucket |
| 189 | // to be treated as a regular, non-inline bucket for the rest of the tx. |
| 190 | b.page = nil |
| 191 | |
| 192 | return b.Bucket(key), nil |
| 193 | } |
| 194 | |
| 195 | // CreateBucketIfNotExists creates a new bucket if it doesn't already exist and returns a reference to it. |
| 196 | // Returns an error if the bucket name is blank, or if the bucket name is too long. |
| 197 | // The bucket instance is only valid for the lifetime of the transaction. |
| 198 | func (b *Bucket) CreateBucketIfNotExists(key []byte) (*Bucket, error) { |
| 199 | child, err := b.CreateBucket(key) |
| 200 | if err == ErrBucketExists { |
| 201 | return b.Bucket(key), nil |
| 202 | } else if err != nil { |
| 203 | return nil, err |
| 204 | } |
| 205 | return child, nil |
| 206 | } |
| 207 | |
| 208 | // DeleteBucket deletes a bucket at the given key. |
| 209 | // Returns an error if the bucket does not exists, or if the key represents a non-bucket value. |
| 210 | func (b *Bucket) DeleteBucket(key []byte) error { |
| 211 | if b.tx.db == nil { |
| 212 | return ErrTxClosed |
| 213 | } else if !b.Writable() { |
| 214 | return ErrTxNotWritable |
| 215 | } |
| 216 | |
| 217 | // Move cursor to correct position. |
| 218 | c := b.Cursor() |
| 219 | k, _, flags := c.seek(key) |
| 220 | |
| 221 | // Return an error if bucket doesn't exist or is not a bucket. |
| 222 | if !bytes.Equal(key, k) { |
| 223 | return ErrBucketNotFound |
| 224 | } else if (flags & bucketLeafFlag) == 0 { |
| 225 | return ErrIncompatibleValue |
| 226 | } |
| 227 | |
| 228 | // Recursively delete all child buckets. |
| 229 | child := b.Bucket(key) |
| 230 | err := child.ForEach(func(k, v []byte) error { |
| 231 | if v == nil { |
| 232 | if err := child.DeleteBucket(k); err != nil { |
| 233 | return fmt.Errorf("delete bucket: %s", err) |
| 234 | } |
| 235 | } |
| 236 | return nil |
| 237 | }) |
| 238 | if err != nil { |
| 239 | return err |
| 240 | } |
| 241 | |
| 242 | // Remove cached copy. |
| 243 | delete(b.buckets, string(key)) |
| 244 | |
| 245 | // Release all bucket pages to freelist. |
| 246 | child.nodes = nil |
| 247 | child.rootNode = nil |
| 248 | child.free() |
| 249 | |
| 250 | // Delete the node if we have a matching key. |
| 251 | c.node().del(key) |
| 252 | |
| 253 | return nil |
| 254 | } |
| 255 | |
| 256 | // Get retrieves the value for a key in the bucket. |
| 257 | // Returns a nil value if the key does not exist or if the key is a nested bucket. |
| 258 | // The returned value is only valid for the life of the transaction. |
| 259 | func (b *Bucket) Get(key []byte) []byte { |
| 260 | k, v, flags := b.Cursor().seek(key) |
| 261 | |
| 262 | // Return nil if this is a bucket. |
| 263 | if (flags & bucketLeafFlag) != 0 { |
| 264 | return nil |
| 265 | } |
| 266 | |
| 267 | // If our target node isn't the same key as what's passed in then return nil. |
| 268 | if !bytes.Equal(key, k) { |
| 269 | return nil |
| 270 | } |
| 271 | return v |
| 272 | } |
| 273 | |
| 274 | // Put sets the value for a key in the bucket. |
| 275 | // If the key exist then its previous value will be overwritten. |
| 276 | // Supplied value must remain valid for the life of the transaction. |
| 277 | // Returns an error if the bucket was created from a read-only transaction, if the key is blank, if the key is too large, or if the value is too large. |
| 278 | func (b *Bucket) Put(key []byte, value []byte) error { |
| 279 | if b.tx.db == nil { |
| 280 | return ErrTxClosed |
| 281 | } else if !b.Writable() { |
| 282 | return ErrTxNotWritable |
| 283 | } else if len(key) == 0 { |
| 284 | return ErrKeyRequired |
| 285 | } else if len(key) > MaxKeySize { |
| 286 | return ErrKeyTooLarge |
| 287 | } else if int64(len(value)) > MaxValueSize { |
| 288 | return ErrValueTooLarge |
| 289 | } |
| 290 | |
| 291 | // Move cursor to correct position. |
| 292 | c := b.Cursor() |
| 293 | k, _, flags := c.seek(key) |
| 294 | |
| 295 | // Return an error if there is an existing key with a bucket value. |
| 296 | if bytes.Equal(key, k) && (flags&bucketLeafFlag) != 0 { |
| 297 | return ErrIncompatibleValue |
| 298 | } |
| 299 | |
| 300 | // Insert into node. |
| 301 | key = cloneBytes(key) |
| 302 | c.node().put(key, key, value, 0, 0) |
| 303 | |
| 304 | return nil |
| 305 | } |
| 306 | |
| 307 | // Delete removes a key from the bucket. |
| 308 | // If the key does not exist then nothing is done and a nil error is returned. |
| 309 | // Returns an error if the bucket was created from a read-only transaction. |
| 310 | func (b *Bucket) Delete(key []byte) error { |
| 311 | if b.tx.db == nil { |
| 312 | return ErrTxClosed |
| 313 | } else if !b.Writable() { |
| 314 | return ErrTxNotWritable |
| 315 | } |
| 316 | |
| 317 | // Move cursor to correct position. |
| 318 | c := b.Cursor() |
| 319 | k, _, flags := c.seek(key) |
| 320 | |
| 321 | // Return nil if the key doesn't exist. |
| 322 | if !bytes.Equal(key, k) { |
| 323 | return nil |
| 324 | } |
| 325 | |
| 326 | // Return an error if there is already existing bucket value. |
| 327 | if (flags & bucketLeafFlag) != 0 { |
| 328 | return ErrIncompatibleValue |
| 329 | } |
| 330 | |
| 331 | // Delete the node if we have a matching key. |
| 332 | c.node().del(key) |
| 333 | |
| 334 | return nil |
| 335 | } |
| 336 | |
| 337 | // Sequence returns the current integer for the bucket without incrementing it. |
| 338 | func (b *Bucket) Sequence() uint64 { return b.bucket.sequence } |
| 339 | |
| 340 | // SetSequence updates the sequence number for the bucket. |
| 341 | func (b *Bucket) SetSequence(v uint64) error { |
| 342 | if b.tx.db == nil { |
| 343 | return ErrTxClosed |
| 344 | } else if !b.Writable() { |
| 345 | return ErrTxNotWritable |
| 346 | } |
| 347 | |
| 348 | // Materialize the root node if it hasn't been already so that the |
| 349 | // bucket will be saved during commit. |
| 350 | if b.rootNode == nil { |
| 351 | _ = b.node(b.root, nil) |
| 352 | } |
| 353 | |
| 354 | // Increment and return the sequence. |
| 355 | b.bucket.sequence = v |
| 356 | return nil |
| 357 | } |
| 358 | |
| 359 | // NextSequence returns an autoincrementing integer for the bucket. |
| 360 | func (b *Bucket) NextSequence() (uint64, error) { |
| 361 | if b.tx.db == nil { |
| 362 | return 0, ErrTxClosed |
| 363 | } else if !b.Writable() { |
| 364 | return 0, ErrTxNotWritable |
| 365 | } |
| 366 | |
| 367 | // Materialize the root node if it hasn't been already so that the |
| 368 | // bucket will be saved during commit. |
| 369 | if b.rootNode == nil { |
| 370 | _ = b.node(b.root, nil) |
| 371 | } |
| 372 | |
| 373 | // Increment and return the sequence. |
| 374 | b.bucket.sequence++ |
| 375 | return b.bucket.sequence, nil |
| 376 | } |
| 377 | |
| 378 | // ForEach executes a function for each key/value pair in a bucket. |
| 379 | // If the provided function returns an error then the iteration is stopped and |
| 380 | // the error is returned to the caller. The provided function must not modify |
| 381 | // the bucket; this will result in undefined behavior. |
| 382 | func (b *Bucket) ForEach(fn func(k, v []byte) error) error { |
| 383 | if b.tx.db == nil { |
| 384 | return ErrTxClosed |
| 385 | } |
| 386 | c := b.Cursor() |
| 387 | for k, v := c.First(); k != nil; k, v = c.Next() { |
| 388 | if err := fn(k, v); err != nil { |
| 389 | return err |
| 390 | } |
| 391 | } |
| 392 | return nil |
| 393 | } |
| 394 | |
| 395 | // Stat returns stats on a bucket. |
| 396 | func (b *Bucket) Stats() BucketStats { |
| 397 | var s, subStats BucketStats |
| 398 | pageSize := b.tx.db.pageSize |
| 399 | s.BucketN += 1 |
| 400 | if b.root == 0 { |
| 401 | s.InlineBucketN += 1 |
| 402 | } |
| 403 | b.forEachPage(func(p *page, depth int) { |
| 404 | if (p.flags & leafPageFlag) != 0 { |
| 405 | s.KeyN += int(p.count) |
| 406 | |
| 407 | // used totals the used bytes for the page |
| 408 | used := pageHeaderSize |
| 409 | |
| 410 | if p.count != 0 { |
| 411 | // If page has any elements, add all element headers. |
| 412 | used += leafPageElementSize * int(p.count-1) |
| 413 | |
| 414 | // Add all element key, value sizes. |
| 415 | // The computation takes advantage of the fact that the position |
| 416 | // of the last element's key/value equals to the total of the sizes |
| 417 | // of all previous elements' keys and values. |
| 418 | // It also includes the last element's header. |
| 419 | lastElement := p.leafPageElement(p.count - 1) |
| 420 | used += int(lastElement.pos + lastElement.ksize + lastElement.vsize) |
| 421 | } |
| 422 | |
| 423 | if b.root == 0 { |
| 424 | // For inlined bucket just update the inline stats |
| 425 | s.InlineBucketInuse += used |
| 426 | } else { |
| 427 | // For non-inlined bucket update all the leaf stats |
| 428 | s.LeafPageN++ |
| 429 | s.LeafInuse += used |
| 430 | s.LeafOverflowN += int(p.overflow) |
| 431 | |
| 432 | // Collect stats from sub-buckets. |
| 433 | // Do that by iterating over all element headers |
| 434 | // looking for the ones with the bucketLeafFlag. |
| 435 | for i := uint16(0); i < p.count; i++ { |
| 436 | e := p.leafPageElement(i) |
| 437 | if (e.flags & bucketLeafFlag) != 0 { |
| 438 | // For any bucket element, open the element value |
| 439 | // and recursively call Stats on the contained bucket. |
| 440 | subStats.Add(b.openBucket(e.value()).Stats()) |
| 441 | } |
| 442 | } |
| 443 | } |
| 444 | } else if (p.flags & branchPageFlag) != 0 { |
| 445 | s.BranchPageN++ |
| 446 | lastElement := p.branchPageElement(p.count - 1) |
| 447 | |
| 448 | // used totals the used bytes for the page |
| 449 | // Add header and all element headers. |
| 450 | used := pageHeaderSize + (branchPageElementSize * int(p.count-1)) |
| 451 | |
| 452 | // Add size of all keys and values. |
| 453 | // Again, use the fact that last element's position equals to |
| 454 | // the total of key, value sizes of all previous elements. |
| 455 | used += int(lastElement.pos + lastElement.ksize) |
| 456 | s.BranchInuse += used |
| 457 | s.BranchOverflowN += int(p.overflow) |
| 458 | } |
| 459 | |
| 460 | // Keep track of maximum page depth. |
| 461 | if depth+1 > s.Depth { |
| 462 | s.Depth = (depth + 1) |
| 463 | } |
| 464 | }) |
| 465 | |
| 466 | // Alloc stats can be computed from page counts and pageSize. |
| 467 | s.BranchAlloc = (s.BranchPageN + s.BranchOverflowN) * pageSize |
| 468 | s.LeafAlloc = (s.LeafPageN + s.LeafOverflowN) * pageSize |
| 469 | |
| 470 | // Add the max depth of sub-buckets to get total nested depth. |
| 471 | s.Depth += subStats.Depth |
| 472 | // Add the stats for all sub-buckets |
| 473 | s.Add(subStats) |
| 474 | return s |
| 475 | } |
| 476 | |
| 477 | // forEachPage iterates over every page in a bucket, including inline pages. |
| 478 | func (b *Bucket) forEachPage(fn func(*page, int)) { |
| 479 | // If we have an inline page then just use that. |
| 480 | if b.page != nil { |
| 481 | fn(b.page, 0) |
| 482 | return |
| 483 | } |
| 484 | |
| 485 | // Otherwise traverse the page hierarchy. |
| 486 | b.tx.forEachPage(b.root, 0, fn) |
| 487 | } |
| 488 | |
| 489 | // forEachPageNode iterates over every page (or node) in a bucket. |
| 490 | // This also includes inline pages. |
| 491 | func (b *Bucket) forEachPageNode(fn func(*page, *node, int)) { |
| 492 | // If we have an inline page or root node then just use that. |
| 493 | if b.page != nil { |
| 494 | fn(b.page, nil, 0) |
| 495 | return |
| 496 | } |
| 497 | b._forEachPageNode(b.root, 0, fn) |
| 498 | } |
| 499 | |
| 500 | func (b *Bucket) _forEachPageNode(pgid pgid, depth int, fn func(*page, *node, int)) { |
| 501 | var p, n = b.pageNode(pgid) |
| 502 | |
| 503 | // Execute function. |
| 504 | fn(p, n, depth) |
| 505 | |
| 506 | // Recursively loop over children. |
| 507 | if p != nil { |
| 508 | if (p.flags & branchPageFlag) != 0 { |
| 509 | for i := 0; i < int(p.count); i++ { |
| 510 | elem := p.branchPageElement(uint16(i)) |
| 511 | b._forEachPageNode(elem.pgid, depth+1, fn) |
| 512 | } |
| 513 | } |
| 514 | } else { |
| 515 | if !n.isLeaf { |
| 516 | for _, inode := range n.inodes { |
| 517 | b._forEachPageNode(inode.pgid, depth+1, fn) |
| 518 | } |
| 519 | } |
| 520 | } |
| 521 | } |
| 522 | |
| 523 | // spill writes all the nodes for this bucket to dirty pages. |
| 524 | func (b *Bucket) spill() error { |
| 525 | // Spill all child buckets first. |
| 526 | for name, child := range b.buckets { |
| 527 | // If the child bucket is small enough and it has no child buckets then |
| 528 | // write it inline into the parent bucket's page. Otherwise spill it |
| 529 | // like a normal bucket and make the parent value a pointer to the page. |
| 530 | var value []byte |
| 531 | if child.inlineable() { |
| 532 | child.free() |
| 533 | value = child.write() |
| 534 | } else { |
| 535 | if err := child.spill(); err != nil { |
| 536 | return err |
| 537 | } |
| 538 | |
| 539 | // Update the child bucket header in this bucket. |
| 540 | value = make([]byte, unsafe.Sizeof(bucket{})) |
| 541 | var bucket = (*bucket)(unsafe.Pointer(&value[0])) |
| 542 | *bucket = *child.bucket |
| 543 | } |
| 544 | |
| 545 | // Skip writing the bucket if there are no materialized nodes. |
| 546 | if child.rootNode == nil { |
| 547 | continue |
| 548 | } |
| 549 | |
| 550 | // Update parent node. |
| 551 | var c = b.Cursor() |
| 552 | k, _, flags := c.seek([]byte(name)) |
| 553 | if !bytes.Equal([]byte(name), k) { |
| 554 | panic(fmt.Sprintf("misplaced bucket header: %x -> %x", []byte(name), k)) |
| 555 | } |
| 556 | if flags&bucketLeafFlag == 0 { |
| 557 | panic(fmt.Sprintf("unexpected bucket header flag: %x", flags)) |
| 558 | } |
| 559 | c.node().put([]byte(name), []byte(name), value, 0, bucketLeafFlag) |
| 560 | } |
| 561 | |
| 562 | // Ignore if there's not a materialized root node. |
| 563 | if b.rootNode == nil { |
| 564 | return nil |
| 565 | } |
| 566 | |
| 567 | // Spill nodes. |
| 568 | if err := b.rootNode.spill(); err != nil { |
| 569 | return err |
| 570 | } |
| 571 | b.rootNode = b.rootNode.root() |
| 572 | |
| 573 | // Update the root node for this bucket. |
| 574 | if b.rootNode.pgid >= b.tx.meta.pgid { |
| 575 | panic(fmt.Sprintf("pgid (%d) above high water mark (%d)", b.rootNode.pgid, b.tx.meta.pgid)) |
| 576 | } |
| 577 | b.root = b.rootNode.pgid |
| 578 | |
| 579 | return nil |
| 580 | } |
| 581 | |
| 582 | // inlineable returns true if a bucket is small enough to be written inline |
| 583 | // and if it contains no subbuckets. Otherwise returns false. |
| 584 | func (b *Bucket) inlineable() bool { |
| 585 | var n = b.rootNode |
| 586 | |
| 587 | // Bucket must only contain a single leaf node. |
| 588 | if n == nil || !n.isLeaf { |
| 589 | return false |
| 590 | } |
| 591 | |
| 592 | // Bucket is not inlineable if it contains subbuckets or if it goes beyond |
| 593 | // our threshold for inline bucket size. |
| 594 | var size = pageHeaderSize |
| 595 | for _, inode := range n.inodes { |
| 596 | size += leafPageElementSize + len(inode.key) + len(inode.value) |
| 597 | |
| 598 | if inode.flags&bucketLeafFlag != 0 { |
| 599 | return false |
| 600 | } else if size > b.maxInlineBucketSize() { |
| 601 | return false |
| 602 | } |
| 603 | } |
| 604 | |
| 605 | return true |
| 606 | } |
| 607 | |
| 608 | // Returns the maximum total size of a bucket to make it a candidate for inlining. |
| 609 | func (b *Bucket) maxInlineBucketSize() int { |
| 610 | return b.tx.db.pageSize / 4 |
| 611 | } |
| 612 | |
| 613 | // write allocates and writes a bucket to a byte slice. |
| 614 | func (b *Bucket) write() []byte { |
| 615 | // Allocate the appropriate size. |
| 616 | var n = b.rootNode |
| 617 | var value = make([]byte, bucketHeaderSize+n.size()) |
| 618 | |
| 619 | // Write a bucket header. |
| 620 | var bucket = (*bucket)(unsafe.Pointer(&value[0])) |
| 621 | *bucket = *b.bucket |
| 622 | |
| 623 | // Convert byte slice to a fake page and write the root node. |
| 624 | var p = (*page)(unsafe.Pointer(&value[bucketHeaderSize])) |
| 625 | n.write(p) |
| 626 | |
| 627 | return value |
| 628 | } |
| 629 | |
| 630 | // rebalance attempts to balance all nodes. |
| 631 | func (b *Bucket) rebalance() { |
| 632 | for _, n := range b.nodes { |
| 633 | n.rebalance() |
| 634 | } |
| 635 | for _, child := range b.buckets { |
| 636 | child.rebalance() |
| 637 | } |
| 638 | } |
| 639 | |
| 640 | // node creates a node from a page and associates it with a given parent. |
| 641 | func (b *Bucket) node(pgid pgid, parent *node) *node { |
| 642 | _assert(b.nodes != nil, "nodes map expected") |
| 643 | |
| 644 | // Retrieve node if it's already been created. |
| 645 | if n := b.nodes[pgid]; n != nil { |
| 646 | return n |
| 647 | } |
| 648 | |
| 649 | // Otherwise create a node and cache it. |
| 650 | n := &node{bucket: b, parent: parent} |
| 651 | if parent == nil { |
| 652 | b.rootNode = n |
| 653 | } else { |
| 654 | parent.children = append(parent.children, n) |
| 655 | } |
| 656 | |
| 657 | // Use the inline page if this is an inline bucket. |
| 658 | var p = b.page |
| 659 | if p == nil { |
| 660 | p = b.tx.page(pgid) |
| 661 | } |
| 662 | |
| 663 | // Read the page into the node and cache it. |
| 664 | n.read(p) |
| 665 | b.nodes[pgid] = n |
| 666 | |
| 667 | // Update statistics. |
| 668 | b.tx.stats.NodeCount++ |
| 669 | |
| 670 | return n |
| 671 | } |
| 672 | |
| 673 | // free recursively frees all pages in the bucket. |
| 674 | func (b *Bucket) free() { |
| 675 | if b.root == 0 { |
| 676 | return |
| 677 | } |
| 678 | |
| 679 | var tx = b.tx |
| 680 | b.forEachPageNode(func(p *page, n *node, _ int) { |
| 681 | if p != nil { |
| 682 | tx.db.freelist.free(tx.meta.txid, p) |
| 683 | } else { |
| 684 | n.free() |
| 685 | } |
| 686 | }) |
| 687 | b.root = 0 |
| 688 | } |
| 689 | |
| 690 | // dereference removes all references to the old mmap. |
| 691 | func (b *Bucket) dereference() { |
| 692 | if b.rootNode != nil { |
| 693 | b.rootNode.root().dereference() |
| 694 | } |
| 695 | |
| 696 | for _, child := range b.buckets { |
| 697 | child.dereference() |
| 698 | } |
| 699 | } |
| 700 | |
| 701 | // pageNode returns the in-memory node, if it exists. |
| 702 | // Otherwise returns the underlying page. |
| 703 | func (b *Bucket) pageNode(id pgid) (*page, *node) { |
| 704 | // Inline buckets have a fake page embedded in their value so treat them |
| 705 | // differently. We'll return the rootNode (if available) or the fake page. |
| 706 | if b.root == 0 { |
| 707 | if id != 0 { |
| 708 | panic(fmt.Sprintf("inline bucket non-zero page access(2): %d != 0", id)) |
| 709 | } |
| 710 | if b.rootNode != nil { |
| 711 | return nil, b.rootNode |
| 712 | } |
| 713 | return b.page, nil |
| 714 | } |
| 715 | |
| 716 | // Check the node cache for non-inline buckets. |
| 717 | if b.nodes != nil { |
| 718 | if n := b.nodes[id]; n != nil { |
| 719 | return nil, n |
| 720 | } |
| 721 | } |
| 722 | |
| 723 | // Finally lookup the page from the transaction if no node is materialized. |
| 724 | return b.tx.page(id), nil |
| 725 | } |
| 726 | |
| 727 | // BucketStats records statistics about resources used by a bucket. |
| 728 | type BucketStats struct { |
| 729 | // Page count statistics. |
| 730 | BranchPageN int // number of logical branch pages |
| 731 | BranchOverflowN int // number of physical branch overflow pages |
| 732 | LeafPageN int // number of logical leaf pages |
| 733 | LeafOverflowN int // number of physical leaf overflow pages |
| 734 | |
| 735 | // Tree statistics. |
| 736 | KeyN int // number of keys/value pairs |
| 737 | Depth int // number of levels in B+tree |
| 738 | |
| 739 | // Page size utilization. |
| 740 | BranchAlloc int // bytes allocated for physical branch pages |
| 741 | BranchInuse int // bytes actually used for branch data |
| 742 | LeafAlloc int // bytes allocated for physical leaf pages |
| 743 | LeafInuse int // bytes actually used for leaf data |
| 744 | |
| 745 | // Bucket statistics |
| 746 | BucketN int // total number of buckets including the top bucket |
| 747 | InlineBucketN int // total number on inlined buckets |
| 748 | InlineBucketInuse int // bytes used for inlined buckets (also accounted for in LeafInuse) |
| 749 | } |
| 750 | |
| 751 | func (s *BucketStats) Add(other BucketStats) { |
| 752 | s.BranchPageN += other.BranchPageN |
| 753 | s.BranchOverflowN += other.BranchOverflowN |
| 754 | s.LeafPageN += other.LeafPageN |
| 755 | s.LeafOverflowN += other.LeafOverflowN |
| 756 | s.KeyN += other.KeyN |
| 757 | if s.Depth < other.Depth { |
| 758 | s.Depth = other.Depth |
| 759 | } |
| 760 | s.BranchAlloc += other.BranchAlloc |
| 761 | s.BranchInuse += other.BranchInuse |
| 762 | s.LeafAlloc += other.LeafAlloc |
| 763 | s.LeafInuse += other.LeafInuse |
| 764 | |
| 765 | s.BucketN += other.BucketN |
| 766 | s.InlineBucketN += other.InlineBucketN |
| 767 | s.InlineBucketInuse += other.InlineBucketInuse |
| 768 | } |
| 769 | |
| 770 | // cloneBytes returns a copy of a given slice. |
| 771 | func cloneBytes(v []byte) []byte { |
| 772 | var clone = make([]byte, len(v)) |
| 773 | copy(clone, v) |
| 774 | return clone |
| 775 | } |