Update goleveldb

Godeps/Godeps.json

@@ -49,7 +49,7 @@
 		},
 		{
 			"ImportPath": "github.com/syndtr/goleveldb/leveldb",
-			"Rev": "59d87758aeaab5ab6ed289c773349500228a1557"
+			"Rev": "2b99e8d4757bf06eeab1b0485d80b8ae1c088874"
 		},
 		{
 			"ImportPath": "github.com/vitrun/qart/coding",

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/cache/cache.go

@@ -40,10 +40,21 @@ type Cache interface {
 	// Size returns entire alive cache objects size.
 	Size() int
 
+	// NumObjects returns number of alive objects.
+	NumObjects() int
+
 	// GetNamespace gets cache namespace with the given id.
 	// GetNamespace is never return nil.
 	GetNamespace(id uint64) Namespace
 
+	// PurgeNamespace purges cache namespace with the given id from this cache tree.
+	// Also read Namespace.Purge.
+	PurgeNamespace(id uint64, fin PurgeFin)
+
+	// ZapNamespace detaches cache namespace with the given id from this cache tree.
+	// Also read Namespace.Zap.
+	ZapNamespace(id uint64)
+
 	// Purge purges all cache namespace from this cache tree.
 	// This is behave the same as calling Namespace.Purge method on all cache namespace.
 	Purge(fin PurgeFin)

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/cache/lru_cache.go

@@ -19,7 +19,7 @@ type lruCache struct {
 	recent            lruNode
 	table             map[uint64]*lruNs
 	capacity          int
-	used, size int
+	used, size, alive int
 }
 
 // NewLRUCache creates a new initialized LRU cache with the given capacity.
@@ -51,6 +51,12 @@ func (c *lruCache) Size() int {
 	return c.size
 }
 
+func (c *lruCache) NumObjects() int {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return c.alive
+}
+
 // SetCapacity set cache capacity.
 func (c *lruCache) SetCapacity(capacity int) {
 	c.mu.Lock()
@@ -77,6 +83,23 @@ func (c *lruCache) GetNamespace(id uint64) Namespace {
 	return ns
 }
 
+func (c *lruCache) ZapNamespace(id uint64) {
+	c.mu.Lock()
+	if ns, exist := c.table[id]; exist {
+		ns.zapNB()
+		delete(c.table, id)
+	}
+	c.mu.Unlock()
+}
+
+func (c *lruCache) PurgeNamespace(id uint64, fin PurgeFin) {
+	c.mu.Lock()
+	if ns, exist := c.table[id]; exist {
+		ns.purgeNB(fin)
+	}
+	c.mu.Unlock()
+}
+
 // Purge purge entire cache.
 func (c *lruCache) Purge(fin PurgeFin) {
 	c.mu.Lock()
@@ -158,11 +181,12 @@ func (ns *lruNs) Get(key uint64, setf SetFunc) Handle {
 		}
 		ns.table[key] = node
 
+		ns.lru.size += charge
+		ns.lru.alive++
 		if charge > 0 {
 			node.ref++
 			node.rInsert(&ns.lru.recent)
 			ns.lru.used += charge
-			ns.lru.size += charge
 			ns.lru.evict()
 		}
 	}
@@ -322,8 +346,10 @@ func (n *lruNode) derefNB() {
 			// Remove elemement.
 			delete(n.ns.table, n.key)
 			n.ns.lru.size -= n.charge
+			n.ns.lru.alive--
 			n.fin()
 		}
+		n.value = nil
 	} else if n.ref < 0 {
 		panic("leveldb/cache: lruCache: negative node reference")
 	}

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db.go

@@ -14,6 +14,7 @@ import (
 	"runtime"
 	"strings"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/syndtr/goleveldb/leveldb/iterator"
@@ -35,7 +36,7 @@ type DB struct {
 
 	// MemDB.
 	memMu             sync.RWMutex
-	memPool           *util.Pool
+	memPool           chan *memdb.DB
 	mem, frozenMem    *memDB
 	journal           *journal.Writer
 	journalWriter     storage.Writer
@@ -47,6 +48,9 @@ type DB struct {
 	snapsMu   sync.Mutex
 	snapsRoot snapshotElement
 
+	// Stats.
+	aliveSnaps, aliveIters int32
+
 	// Write.
 	writeC       chan *Batch
 	writeMergedC chan bool
@@ -80,7 +84,7 @@ func openDB(s *session) (*DB, error) {
 		// Initial sequence
 		seq: s.stSeq,
 		// MemDB
-		memPool: util.NewPool(1),
+		memPool: make(chan *memdb.DB, 1),
 		// Write
 		writeC:       make(chan *Batch),
 		writeMergedC: make(chan bool),
@@ -122,6 +126,7 @@ func openDB(s *session) (*DB, error) {
 	go db.tCompaction()
 	go db.mCompaction()
 	go db.jWriter()
+	go db.mpoolDrain()
 
 	s.logf("db@open done T·%v", time.Since(start))
 
@@ -568,7 +573,7 @@ func (db *DB) get(key []byte, seq uint64, ro *opt.ReadOptions) (value []byte, er
 		}
 		defer m.decref()
 
-		mk, mv, me := m.db.Find(ikey)
+		mk, mv, me := m.mdb.Find(ikey)
 		if me == nil {
 			ukey, _, t, ok := parseIkey(mk)
 			if ok && db.s.icmp.uCompare(ukey, key) == 0 {
@@ -657,6 +662,14 @@ func (db *DB) GetSnapshot() (*Snapshot, error) {
 //		Returns sstables list for each level.
 //	leveldb.blockpool
 //		Returns block pool stats.
+//	leveldb.cachedblock
+//		Returns size of cached block.
+//	leveldb.openedtables
+//		Returns number of opened tables.
+//	leveldb.alivesnaps
+//		Returns number of alive snapshots.
+//	leveldb.aliveiters
+//		Returns number of alive iterators.
 func (db *DB) GetProperty(name string) (value string, err error) {
 	err = db.ok()
 	if err != nil {
@@ -712,6 +725,10 @@ func (db *DB) GetProperty(name string) (value string, err error) {
 		}
 	case p == "openedtables":
 		value = fmt.Sprintf("%d", db.s.tops.cache.Size())
+	case p == "alivesnaps":
+		value = fmt.Sprintf("%d", atomic.LoadInt32(&db.aliveSnaps))
+	case p == "aliveiters":
+		value = fmt.Sprintf("%d", atomic.LoadInt32(&db.aliveIters))
 	default:
 		err = errors.New("leveldb: GetProperty: unknown property: " + name)
 	}

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_compaction.go

@@ -221,10 +221,10 @@ func (db *DB) memCompaction() {
 	c := newCMem(db.s)
 	stats := new(cStatsStaging)
 
-	db.logf("mem@flush N·%d S·%s", mem.db.Len(), shortenb(mem.db.Size()))
+	db.logf("mem@flush N·%d S·%s", mem.mdb.Len(), shortenb(mem.mdb.Size()))
 
 	// Don't compact empty memdb.
-	if mem.db.Len() == 0 {
+	if mem.mdb.Len() == 0 {
 		db.logf("mem@flush skipping")
 		// drop frozen mem
 		db.dropFrozenMem()
@@ -242,7 +242,7 @@ func (db *DB) memCompaction() {
 	db.compactionTransact("mem@flush", func(cnt *compactionTransactCounter) (err error) {
 		stats.startTimer()
 		defer stats.stopTimer()
-		return c.flush(mem.db, -1)
+		return c.flush(mem.mdb, -1)
 	}, func() error {
 		for _, r := range c.rec.addedTables {
 			db.logf("mem@flush rollback @%d", r.num)

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_iter.go

@@ -10,6 +10,7 @@ import (
 	"errors"
 	"runtime"
 	"sync"
+	"sync/atomic"
 
 	"github.com/syndtr/goleveldb/leveldb/iterator"
 	"github.com/syndtr/goleveldb/leveldb/opt"
@@ -38,11 +39,11 @@ func (db *DB) newRawIterator(slice *util.Range, ro *opt.ReadOptions) iterator.It
 	ti := v.getIterators(slice, ro)
 	n := len(ti) + 2
 	i := make([]iterator.Iterator, 0, n)
-	emi := em.db.NewIterator(slice)
+	emi := em.mdb.NewIterator(slice)
 	emi.SetReleaser(&memdbReleaser{m: em})
 	i = append(i, emi)
 	if fm != nil {
-		fmi := fm.db.NewIterator(slice)
+		fmi := fm.mdb.NewIterator(slice)
 		fmi.SetReleaser(&memdbReleaser{m: fm})
 		i = append(i, fmi)
 	}
@@ -66,6 +67,7 @@ func (db *DB) newIterator(seq uint64, slice *util.Range, ro *opt.ReadOptions) *d
 	}
 	rawIter := db.newRawIterator(islice, ro)
 	iter := &dbIter{
+		db:     db,
 		icmp:   db.s.icmp,
 		iter:   rawIter,
 		seq:    seq,
@@ -73,6 +75,7 @@ func (db *DB) newIterator(seq uint64, slice *util.Range, ro *opt.ReadOptions) *d
 		key:    make([]byte, 0),
 		value:  make([]byte, 0),
 	}
+	atomic.AddInt32(&db.aliveIters, 1)
 	runtime.SetFinalizer(iter, (*dbIter).Release)
 	return iter
 }
@@ -89,6 +92,7 @@ const (
 
 // dbIter represent an interator states over a database session.
 type dbIter struct {
+	db     *DB
 	icmp   *iComparer
 	iter   iterator.Iterator
 	seq    uint64
@@ -303,6 +307,7 @@ func (i *dbIter) Release() {
 
 		if i.releaser != nil {
 			i.releaser.Release()
+			i.releaser = nil
 		}
 
 		i.dir = dirReleased
@@ -310,6 +315,8 @@ func (i *dbIter) Release() {
 		i.value = nil
 		i.iter.Release()
 		i.iter = nil
+		atomic.AddInt32(&i.db.aliveIters, -1)
+		i.db = nil
 	}
 }
 

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_snapshot.go

@@ -9,6 +9,7 @@ package leveldb
 import (
 	"runtime"
 	"sync"
+	"sync/atomic"
 
 	"github.com/syndtr/goleveldb/leveldb/iterator"
 	"github.com/syndtr/goleveldb/leveldb/opt"
@@ -81,7 +82,7 @@ func (db *DB) minSeq() uint64 {
 type Snapshot struct {
 	db       *DB
 	elem     *snapshotElement
-	mu       sync.Mutex
+	mu       sync.RWMutex
 	released bool
 }
 
@@ -91,6 +92,7 @@ func (db *DB) newSnapshot() *Snapshot {
 		db:   db,
 		elem: db.acquireSnapshot(),
 	}
+	atomic.AddInt32(&db.aliveSnaps, 1)
 	runtime.SetFinalizer(snap, (*Snapshot).Release)
 	return snap
 }
@@ -105,8 +107,8 @@ func (snap *Snapshot) Get(key []byte, ro *opt.ReadOptions) (value []byte, err er
 	if err != nil {
 		return
 	}
-	snap.mu.Lock()
-	defer snap.mu.Unlock()
+	snap.mu.RLock()
+	defer snap.mu.RUnlock()
 	if snap.released {
 		err = ErrSnapshotReleased
 		return
@@ -160,6 +162,7 @@ func (snap *Snapshot) Release() {
 
 		snap.released = true
 		snap.db.releaseSnapshot(snap.elem)
+		atomic.AddInt32(&snap.db.aliveSnaps, -1)
 		snap.db = nil
 		snap.elem = nil
 	}

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_state.go

@@ -8,15 +8,15 @@ package leveldb
 
 import (
 	"sync/atomic"
+	"time"
 
 	"github.com/syndtr/goleveldb/leveldb/journal"
 	"github.com/syndtr/goleveldb/leveldb/memdb"
-	"github.com/syndtr/goleveldb/leveldb/util"
 )
 
 type memDB struct {
-	pool *util.Pool
-	db   *memdb.DB
+	db  *DB
+	mdb *memdb.DB
 	ref int32
 }
 
@@ -26,7 +26,13 @@ func (m *memDB) incref() {
 
 func (m *memDB) decref() {
 	if ref := atomic.AddInt32(&m.ref, -1); ref == 0 {
-		m.pool.Put(m)
+		// Only put back memdb with std capacity.
+		if m.mdb.Capacity() == m.db.s.o.GetWriteBuffer() {
+			m.mdb.Reset()
+			m.db.mpoolPut(m.mdb)
+		}
+		m.db = nil
+		m.mdb = nil
 	} else if ref < 0 {
 		panic("negative memdb ref")
 	}
@@ -42,6 +48,41 @@ func (db *DB) addSeq(delta uint64) {
 	atomic.AddUint64(&db.seq, delta)
 }
 
+func (db *DB) mpoolPut(mem *memdb.DB) {
+	defer func() {
+		recover()
+	}()
+	select {
+	case db.memPool <- mem:
+	default:
+	}
+}
+
+func (db *DB) mpoolGet() *memdb.DB {
+	select {
+	case mem := <-db.memPool:
+		return mem
+	default:
+		return nil
+	}
+}
+
+func (db *DB) mpoolDrain() {
+	ticker := time.NewTicker(30 * time.Second)
+	for {
+		select {
+		case <-ticker.C:
+			select {
+			case <-db.memPool:
+			default:
+			}
+		case _, _ = <-db.closeC:
+			close(db.memPool)
+			return
+		}
+	}
+}
+
 // Create new memdb and froze the old one; need external synchronization.
 // newMem only called synchronously by the writer.
 func (db *DB) newMem(n int) (mem *memDB, err error) {
@@ -70,18 +111,15 @@ func (db *DB) newMem(n int) (mem *memDB, err error) {
 	db.journalWriter = w
 	db.journalFile = file
 	db.frozenMem = db.mem
-	mem, ok := db.memPool.Get().(*memDB)
-	if ok && mem.db.Capacity() >= n {
-		mem.db.Reset()
-		mem.incref()
-	} else {
+	mdb := db.mpoolGet()
+	if mdb == nil || mdb.Capacity() < n {
+		mdb = memdb.New(db.s.icmp, maxInt(db.s.o.GetWriteBuffer(), n))
+	}
 	mem = &memDB{
-			pool: db.memPool,
-			db:   memdb.New(db.s.icmp, maxInt(db.s.o.GetWriteBuffer(), n)),
-			ref:  1,
+		db:  db,
+		mdb: mdb,
+		ref: 2,
 	}
-	}
-	mem.incref()
 	db.mem = mem
 	// The seq only incremented by the writer. And whoever called newMem
 	// should hold write lock, so no need additional synchronization here.

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_test.go

@@ -1577,7 +1577,11 @@ func TestDb_BloomFilter(t *testing.T) {
 		return fmt.Sprintf("key%06d", i)
 	}
 
-	n := 10000
+	const (
+		n              = 10000
+		indexOverheat  = 19898
+		filterOverheat = 19799
+	)
 
 	// Populate multiple layers
 	for i := 0; i < n; i++ {
@@ -1601,7 +1605,7 @@ func TestDb_BloomFilter(t *testing.T) {
 	cnt := int(h.stor.ReadCounter())
 	t.Logf("lookup of %d present keys yield %d sstable I/O reads", n, cnt)
 
-	if min, max := n, n+2*n/100; cnt < min || cnt > max {
+	if min, max := n+indexOverheat+filterOverheat, n+indexOverheat+filterOverheat+2*n/100; cnt < min || cnt > max {
 		t.Errorf("num of sstable I/O reads of present keys not in range of %d - %d, got %d", min, max, cnt)
 	}
 
@@ -1612,7 +1616,7 @@ func TestDb_BloomFilter(t *testing.T) {
 	}
 	cnt = int(h.stor.ReadCounter())
 	t.Logf("lookup of %d missing keys yield %d sstable I/O reads", n, cnt)
-	if max := 3 * n / 100; cnt > max {
+	if max := 3*n/100 + indexOverheat + filterOverheat; cnt > max {
 		t.Errorf("num of sstable I/O reads of missing keys was more than %d, got %d", max, cnt)
 	}
 

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/db_write.go

@@ -75,7 +75,7 @@ func (db *DB) flush(n int) (mem *memDB, nn int, err error) {
 				mem = nil
 			}
 		}()
-		nn = mem.db.Free()
+		nn = mem.mdb.Free()
 		switch {
 		case v.tLen(0) >= kL0_SlowdownWritesTrigger && !delayed:
 			delayed = true
@@ -90,13 +90,13 @@ func (db *DB) flush(n int) (mem *memDB, nn int, err error) {
 			}
 		default:
 			// Allow memdb to grow if it has no entry.
-			if mem.db.Len() == 0 {
+			if mem.mdb.Len() == 0 {
 				nn = n
 			} else {
 				mem.decref()
 				mem, err = db.rotateMem(n)
 				if err == nil {
-					nn = mem.db.Free()
+					nn = mem.mdb.Free()
 				} else {
 					nn = 0
 				}
@@ -190,7 +190,7 @@ drain:
 			return
 		case db.journalC <- b:
 			// Write into memdb
-			b.memReplay(mem.db)
+			b.memReplay(mem.mdb)
 		}
 		// Wait for journal writer
 		select {
@@ -200,7 +200,7 @@ drain:
 		case err = <-db.journalAckC:
 			if err != nil {
 				// Revert memdb if error detected
-				b.revertMemReplay(mem.db)
+				b.revertMemReplay(mem.mdb)
 				return
 			}
 		}
@@ -209,7 +209,7 @@ drain:
 		if err != nil {
 			return
 		}
-		b.memReplay(mem.db)
+		b.memReplay(mem.mdb)
 	}
 
 	// Set last seq number.
@@ -271,7 +271,7 @@ func (db *DB) CompactRange(r util.Range) error {
 	// Check for overlaps in memdb.
 	mem := db.getEffectiveMem()
 	defer mem.decref()
-	if isMemOverlaps(db.s.icmp, mem.db, r.Start, r.Limit) {
+	if isMemOverlaps(db.s.icmp, mem.mdb, r.Start, r.Limit) {
 		// Memdb compaction.
 		if _, err := db.rotateMem(0); err != nil {
 			<-db.writeLockC

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/opt/options.go

@@ -34,7 +34,10 @@ func (noCache) SetCapacity(capacity int)               {}
 func (noCache) Capacity() int                                { return 0 }
 func (noCache) Used() int                                    { return 0 }
 func (noCache) Size() int                                    { return 0 }
+func (noCache) NumObjects() int                              { return 0 }
 func (noCache) GetNamespace(id uint64) cache.Namespace       { return nil }
+func (noCache) PurgeNamespace(id uint64, fin cache.PurgeFin) {}
+func (noCache) ZapNamespace(id uint64)                       {}
 func (noCache) Purge(fin cache.PurgeFin)                     {}
 func (noCache) Zap()                                         {}
 

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/table.go

@@ -7,7 +7,6 @@
 package leveldb
 
 import (
-	"io"
 	"sort"
 	"sync/atomic"
 
@@ -323,15 +322,6 @@ func (t *tOps) createFrom(src iterator.Iterator) (f *tFile, n int, err error) {
 	return
 }
 
-type tableWrapper struct {
-	*table.Reader
-	closer io.Closer
-}
-
-func (tr tableWrapper) Release() {
-	tr.closer.Close()
-}
-
 // Opens table. It returns a cache handle, which should
 // be released after use.
 func (t *tOps) open(f *tFile) (ch cache.Handle, err error) {
@@ -347,7 +337,7 @@ func (t *tOps) open(f *tFile) (ch cache.Handle, err error) {
 		if bc := t.s.o.GetBlockCache(); bc != nil {
 			bcacheNS = bc.GetNamespace(num)
 		}
-		return 1, tableWrapper{table.NewReader(r, int64(f.size), bcacheNS, t.bpool, t.s.o), r}
+		return 1, table.NewReader(r, int64(f.size), bcacheNS, t.bpool, t.s.o)
 	})
 	if ch == nil && err == nil {
 		err = ErrClosed
@@ -363,7 +353,7 @@ func (t *tOps) find(f *tFile, key []byte, ro *opt.ReadOptions) (rkey, rvalue []b
 		return nil, nil, err
 	}
 	defer ch.Release()
-	return ch.Value().(tableWrapper).Find(key, ro)
+	return ch.Value().(*table.Reader).Find(key, ro)
 }
 
 // Returns approximate offset of the given key.
@@ -372,10 +362,9 @@ func (t *tOps) offsetOf(f *tFile, key []byte) (offset uint64, err error) {
 	if err != nil {
 		return
 	}
-	_offset, err := ch.Value().(tableWrapper).OffsetOf(key)
-	offset = uint64(_offset)
-	ch.Release()
-	return
+	defer ch.Release()
+	offset_, err := ch.Value().(*table.Reader).OffsetOf(key)
+	return uint64(offset_), err
 }
 
 // Creates an iterator from the given table.
@@ -384,7 +373,7 @@ func (t *tOps) newIterator(f *tFile, slice *util.Range, ro *opt.ReadOptions) ite
 	if err != nil {
 		return iterator.NewEmptyIterator(err)
 	}
-	iter := ch.Value().(tableWrapper).NewIterator(slice, ro)
+	iter := ch.Value().(*table.Reader).NewIterator(slice, ro)
 	iter.SetReleaser(ch)
 	return iter
 }
@@ -401,7 +390,7 @@ func (t *tOps) remove(f *tFile) {
 				t.s.logf("table@remove removed @%d", num)
 			}
 			if bc := t.s.o.GetBlockCache(); bc != nil {
-				bc.GetNamespace(num).Zap()
+				bc.ZapNamespace(num)
 			}
 		}
 	})
@@ -411,6 +400,7 @@ func (t *tOps) remove(f *tFile) {
 // regadless still used or not.
 func (t *tOps) close() {
 	t.cache.Zap()
+	t.bpool.Close()
 }
 
 // Creates new initialized table ops instance.

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/table/block_test.go

@@ -40,7 +40,7 @@ var _ = testutil.Defer(func() {
 			data := bw.buf.Bytes()
 			restartsLen := int(binary.LittleEndian.Uint32(data[len(data)-4:]))
 			return &block{
-				cmp:            comparer.DefaultComparer,
+				tr:             &Reader{cmp: comparer.DefaultComparer},
 				data:           data,
 				restartsLen:    restartsLen,
 				restartsOffset: len(data) - (restartsLen+1)*4,

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/table/reader.go

@@ -37,8 +37,7 @@ func max(x, y int) int {
 }
 
 type block struct {
-	bpool          *util.BufferPool
-	cmp            comparer.BasicComparer
+	tr             *Reader
 	data           []byte
 	restartsLen    int
 	restartsOffset int
@@ -47,31 +46,25 @@ type block struct {
 }
 
 func (b *block) seek(rstart, rlimit int, key []byte) (index, offset int, err error) {
-	n := b.restartsOffset
-	data := b.data
-	cmp := b.cmp
-
 	index = sort.Search(b.restartsLen-rstart-(b.restartsLen-rlimit), func(i int) bool {
-		offset := int(binary.LittleEndian.Uint32(data[n+4*(rstart+i):]))
+		offset := int(binary.LittleEndian.Uint32(b.data[b.restartsOffset+4*(rstart+i):]))
 		offset += 1                                 // shared always zero, since this is a restart point
-		v1, n1 := binary.Uvarint(data[offset:])   // key length
-		_, n2 := binary.Uvarint(data[offset+n1:]) // value length
+		v1, n1 := binary.Uvarint(b.data[offset:])   // key length
+		_, n2 := binary.Uvarint(b.data[offset+n1:]) // value length
 		m := offset + n1 + n2
-		return cmp.Compare(data[m:m+int(v1)], key) > 0
+		return b.tr.cmp.Compare(b.data[m:m+int(v1)], key) > 0
 	}) + rstart - 1
 	if index < rstart {
 		// The smallest key is greater-than key sought.
 		index = rstart
 	}
-	offset = int(binary.LittleEndian.Uint32(data[n+4*index:]))
+	offset = int(binary.LittleEndian.Uint32(b.data[b.restartsOffset+4*index:]))
 	return
 }
 
 func (b *block) restartIndex(rstart, rlimit, offset int) int {
-	n := b.restartsOffset
-	data := b.data
 	return sort.Search(b.restartsLen-rstart-(b.restartsLen-rlimit), func(i int) bool {
-		return int(binary.LittleEndian.Uint32(data[n+4*(rstart+i):])) > offset
+		return int(binary.LittleEndian.Uint32(b.data[b.restartsOffset+4*(rstart+i):])) > offset
 	}) + rstart - 1
 }
 
@@ -141,10 +134,10 @@ func (b *block) newIterator(slice *util.Range, inclLimit bool, cache util.Releas
 }
 
 func (b *block) Release() {
-	if b.bpool != nil {
-		b.bpool.Put(b.data)
-		b.bpool = nil
+	if b.tr.bpool != nil {
+		b.tr.bpool.Put(b.data)
 	}
+	b.tr = nil
 	b.data = nil
 }
 
@@ -270,7 +263,7 @@ func (i *blockIter) Seek(key []byte) bool {
 		i.dir = dirForward
 	}
 	for i.Next() {
-		if i.block.cmp.Compare(i.key, key) >= 0 {
+		if i.block.tr.cmp.Compare(i.key, key) >= 0 {
 			return true
 		}
 	}
@@ -479,7 +472,7 @@ func (i *blockIter) Error() error {
 }
 
 type filterBlock struct {
-	filter     filter.Filter
+	tr         *Reader
 	data       []byte
 	oOffset    int
 	baseLg     uint
@@ -493,7 +486,7 @@ func (b *filterBlock) contains(offset uint64, key []byte) bool {
 		n := int(binary.LittleEndian.Uint32(o))
 		m := int(binary.LittleEndian.Uint32(o[4:]))
 		if n < m && m <= b.oOffset {
-			return b.filter.Contains(b.data[n:m], key)
+			return b.tr.filter.Contains(b.data[n:m], key)
 		} else if n == m {
 			return false
 		}
@@ -501,9 +494,16 @@ func (b *filterBlock) contains(offset uint64, key []byte) bool {
 	return true
 }
 
+func (b *filterBlock) Release() {
+	if b.tr.bpool != nil {
+		b.tr.bpool.Put(b.data)
+	}
+	b.tr = nil
+	b.data = nil
+}
+
 type indexIter struct {
-	blockIter
-	tableReader *Reader
+	*blockIter
 	slice *util.Range
 	// Options
 	checksum  bool
@@ -523,7 +523,7 @@ func (i *indexIter) Get() iterator.Iterator {
 	if i.slice != nil && (i.blockIter.isFirst() || i.blockIter.isLast()) {
 		slice = i.slice
 	}
-	return i.tableReader.getDataIter(dataBH, slice, i.checksum, i.fillCache)
+	return i.blockIter.block.tr.getDataIter(dataBH, slice, i.checksum, i.fillCache)
 }
 
 // Reader is a table reader.
@@ -539,8 +539,7 @@ type Reader struct {
 	strictIter bool
 
 	dataEnd           int64
-	indexBlock  *block
-	filterBlock *filterBlock
+	indexBH, filterBH blockHandle
 }
 
 func verifyChecksum(data []byte) bool {
@@ -557,6 +556,7 @@ func (r *Reader) readRawBlock(bh blockHandle, checksum bool) ([]byte, error) {
 	}
 	if checksum || r.checksum {
 		if !verifyChecksum(data) {
+			r.bpool.Put(data)
 			return nil, errors.New("leveldb/table: Reader: invalid block (checksum mismatch)")
 		}
 	}
@@ -575,6 +575,7 @@ func (r *Reader) readRawBlock(bh blockHandle, checksum bool) ([]byte, error) {
 			return nil, err
 		}
 	default:
+		r.bpool.Put(data)
 		return nil, fmt.Errorf("leveldb/table: Reader: unknown block compression type: %d", data[bh.length])
 	}
 	return data, nil
@@ -587,7 +588,7 @@ func (r *Reader) readBlock(bh blockHandle, checksum bool) (*block, error) {
 	}
 	restartsLen := int(binary.LittleEndian.Uint32(data[len(data)-4:]))
 	b := &block{
-		cmp:            r.cmp,
+		tr:             r,
 		data:           data,
 		restartsLen:    restartsLen,
 		restartsOffset: len(data) - (restartsLen+1)*4,
@@ -596,7 +597,44 @@ func (r *Reader) readBlock(bh blockHandle, checksum bool) (*block, error) {
 	return b, nil
 }
 
-func (r *Reader) readFilterBlock(bh blockHandle, filter filter.Filter) (*filterBlock, error) {
+func (r *Reader) readBlockCached(bh blockHandle, checksum, fillCache bool) (*block, util.Releaser, error) {
+	if r.cache != nil {
+		var err error
+		ch := r.cache.Get(bh.offset, func() (charge int, value interface{}) {
+			if !fillCache {
+				return 0, nil
+			}
+			var b *block
+			b, err = r.readBlock(bh, checksum)
+			if err != nil {
+				return 0, nil
+			}
+			return cap(b.data), b
+		})
+		if ch != nil {
+			b, ok := ch.Value().(*block)
+			if !ok {
+				ch.Release()
+				return nil, nil, errors.New("leveldb/table: Reader: inconsistent block type")
+			}
+			if !b.checksum && (r.checksum || checksum) {
+				if !verifyChecksum(b.data) {
+					ch.Release()
+					return nil, nil, errors.New("leveldb/table: Reader: invalid block (checksum mismatch)")
+				}
+				b.checksum = true
+			}
+			return b, ch, err
+		} else if err != nil {
+			return nil, nil, err
+		}
+	}
+
+	b, err := r.readBlock(bh, checksum)
+	return b, b, err
+}
+
+func (r *Reader) readFilterBlock(bh blockHandle) (*filterBlock, error) {
 	data, err := r.readRawBlock(bh, true)
 	if err != nil {
 		return nil, err
@@ -611,7 +649,7 @@ func (r *Reader) readFilterBlock(bh blockHandle, filter filter.Filter) (*filterB
 		return nil, errors.New("leveldb/table: Reader: invalid filter block (invalid offset)")
 	}
 	b := &filterBlock{
-		filter:     filter,
+		tr:         r,
 		data:       data,
 		oOffset:    oOffset,
 		baseLg:     uint(data[n-1]),
@@ -620,42 +658,42 @@ func (r *Reader) readFilterBlock(bh blockHandle, filter filter.Filter) (*filterB
 	return b, nil
 }
 
-func (r *Reader) getDataIter(dataBH blockHandle, slice *util.Range, checksum, fillCache bool) iterator.Iterator {
+func (r *Reader) readFilterBlockCached(bh blockHandle, fillCache bool) (*filterBlock, util.Releaser, error) {
 	if r.cache != nil {
-		// Get/set block cache.
 		var err error
-		cache := r.cache.Get(dataBH.offset, func() (charge int, value interface{}) {
+		ch := r.cache.Get(bh.offset, func() (charge int, value interface{}) {
 			if !fillCache {
 				return 0, nil
 			}
-			var dataBlock *block
-			dataBlock, err = r.readBlock(dataBH, checksum)
+			var b *filterBlock
+			b, err = r.readFilterBlock(bh)
 			if err != nil {
 				return 0, nil
 			}
-			return int(dataBH.length), dataBlock
+			return cap(b.data), b
 		})
+		if ch != nil {
+			b, ok := ch.Value().(*filterBlock)
+			if !ok {
+				ch.Release()
+				return nil, nil, errors.New("leveldb/table: Reader: inconsistent block type")
+			}
+			return b, ch, err
+		} else if err != nil {
+			return nil, nil, err
+		}
+	}
+
+	b, err := r.readFilterBlock(bh)
+	return b, b, err
+}
+
+func (r *Reader) getDataIter(dataBH blockHandle, slice *util.Range, checksum, fillCache bool) iterator.Iterator {
+	b, rel, err := r.readBlockCached(dataBH, checksum, fillCache)
 	if err != nil {
 		return iterator.NewEmptyIterator(err)
 	}
-		if cache != nil {
-			dataBlock := cache.Value().(*block)
-			if !dataBlock.checksum && (r.checksum || checksum) {
-				if !verifyChecksum(dataBlock.data) {
-					return iterator.NewEmptyIterator(errors.New("leveldb/table: Reader: invalid block (checksum mismatch)"))
-				}
-				dataBlock.checksum = true
-			}
-			iter := dataBlock.newIterator(slice, false, cache)
-			return iter
-		}
-	}
-	dataBlock, err := r.readBlock(dataBH, checksum)
-	if err != nil {
-		return iterator.NewEmptyIterator(err)
-	}
-	iter := dataBlock.newIterator(slice, false, dataBlock)
-	return iter
+	return b.newIterator(slice, false, rel)
 }
 
 // NewIterator creates an iterator from the table.
@@ -669,15 +707,18 @@ func (r *Reader) getDataIter(dataBH blockHandle, slice *util.Range, checksum, fi
 // when not used.
 //
 // Also read Iterator documentation of the leveldb/iterator package.
-
 func (r *Reader) NewIterator(slice *util.Range, ro *opt.ReadOptions) iterator.Iterator {
 	if r.err != nil {
 		return iterator.NewEmptyIterator(r.err)
 	}
 
+	fillCache := !ro.GetDontFillCache()
+	b, rel, err := r.readBlockCached(r.indexBH, true, fillCache)
+	if err != nil {
+		return iterator.NewEmptyIterator(err)
+	}
 	index := &indexIter{
-		blockIter:   *r.indexBlock.newIterator(slice, true, nil),
-		tableReader: r,
+		blockIter: b.newIterator(slice, true, rel),
 		slice:     slice,
 		checksum:  ro.GetStrict(opt.StrictBlockChecksum),
 		fillCache: !ro.GetDontFillCache(),
@@ -697,7 +738,13 @@ func (r *Reader) Find(key []byte, ro *opt.ReadOptions) (rkey, value []byte, err
 		return
 	}
 
-	index := r.indexBlock.newIterator(nil, true, nil)
+	indexBlock, rel, err := r.readBlockCached(r.indexBH, true, true)
+	if err != nil {
+		return
+	}
+	defer rel.Release()
+
+	index := indexBlock.newIterator(nil, true, nil)
 	defer index.Release()
 	if !index.Seek(key) {
 		err = index.Error()
@@ -711,9 +758,15 @@ func (r *Reader) Find(key []byte, ro *opt.ReadOptions) (rkey, value []byte, err
 		err = errors.New("leveldb/table: Reader: invalid table (bad data block handle)")
 		return
 	}
-	if r.filterBlock != nil && !r.filterBlock.contains(dataBH.offset, key) {
-		err = ErrNotFound
-		return
+	if r.filter != nil {
+		filterBlock, rel, ferr := r.readFilterBlockCached(r.filterBH, true)
+		if ferr == nil {
+			if !filterBlock.contains(dataBH.offset, key) {
+				rel.Release()
+				return nil, nil, ErrNotFound
+			}
+			rel.Release()
+		}
 	}
 	data := r.getDataIter(dataBH, nil, ro.GetStrict(opt.StrictBlockChecksum), !ro.GetDontFillCache())
 	defer data.Release()
@@ -760,7 +813,13 @@ func (r *Reader) OffsetOf(key []byte) (offset int64, err error) {
 		return
 	}
 
-	index := r.indexBlock.newIterator(nil, true, nil)
+	indexBlock, rel, err := r.readBlockCached(r.indexBH, true, true)
+	if err != nil {
+		return
+	}
+	defer rel.Release()
+
+	index := indexBlock.newIterator(nil, true, nil)
 	defer index.Release()
 	if index.Seek(key) {
 		dataBH, n := decodeBlockHandle(index.Value())
@@ -778,6 +837,17 @@ func (r *Reader) OffsetOf(key []byte) (offset int64, err error) {
 	return
 }
 
+// Release implements util.Releaser.
+// It also close the file if it is an io.Closer.
+func (r *Reader) Release() {
+	if closer, ok := r.reader.(io.Closer); ok {
+		closer.Close()
+	}
+	r.reader = nil
+	r.cache = nil
+	r.bpool = nil
+}
+
 // NewReader creates a new initialized table reader for the file.
 // The cache and bpool is optional and can be nil.
 //
@@ -817,16 +887,11 @@ func NewReader(f io.ReaderAt, size int64, cache cache.Namespace, bpool *util.Buf
 		return r
 	}
 	// Decode the index block handle.
-	indexBH, n := decodeBlockHandle(footer[n:])
+	r.indexBH, n = decodeBlockHandle(footer[n:])
 	if n == 0 {
 		r.err = errors.New("leveldb/table: Reader: invalid table (bad index block handle)")
 		return r
 	}
-	// Read index block.
-	r.indexBlock, r.err = r.readBlock(indexBH, true)
-	if r.err != nil {
-		return r
-	}
 	// Read metaindex block.
 	metaBlock, err := r.readBlock(metaBH, true)
 	if err != nil {
@@ -842,32 +907,28 @@ func NewReader(f io.ReaderAt, size int64, cache cache.Namespace, bpool *util.Buf
 			continue
 		}
 		fn := key[7:]
-		var filter filter.Filter
 		if f0 := o.GetFilter(); f0 != nil && f0.Name() == fn {
-			filter = f0
+			r.filter = f0
 		} else {
 			for _, f0 := range o.GetAltFilters() {
 				if f0.Name() == fn {
-					filter = f0
+					r.filter = f0
 					break
 				}
 			}
 		}
-		if filter != nil {
+		if r.filter != nil {
 			filterBH, n := decodeBlockHandle(metaIter.Value())
 			if n == 0 {
 				continue
 			}
+			r.filterBH = filterBH
 			// Update data end.
 			r.dataEnd = int64(filterBH.offset)
-			filterBlock, err := r.readFilterBlock(filterBH, filter)
-			if err != nil {
-				continue
-			}
-			r.filterBlock = filterBlock
 			break
 		}
 	}
 	metaIter.Release()
+	metaBlock.Release()
 	return r
 }

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/table/table_test.go

@@ -111,7 +111,9 @@ var _ = testutil.Defer(func() {
 			testutil.AllKeyValueTesting(nil, Build)
 			Describe("with one key per block", Test(testutil.KeyValue_Generate(nil, 9, 1, 10, 512, 512), func(r *Reader) {
 				It("should have correct blocks number", func() {
-					Expect(r.indexBlock.restartsLen).Should(Equal(9))
+					indexBlock, err := r.readBlock(r.indexBH, true)
+					Expect(err).To(BeNil())
+					Expect(indexBlock.restartsLen).Should(Equal(9))
 				})
 			}))
 		})

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/util/buffer_pool.go

@@ -19,15 +19,21 @@ type buffer struct {
 
 // BufferPool is a 'buffer pool'.
 type BufferPool struct {
-	pool      [4]chan []byte
-	size      [3]uint32
-	sizeMiss  [3]uint32
+	pool       [6]chan []byte
+	size       [5]uint32
+	sizeMiss   [5]uint32
+	sizeHalf   [5]uint32
+	baseline   [4]int
+	baselinex0 int
+	baselinex1 int
 	baseline0  int
 	baseline1  int
 	baseline2  int
+	close      chan struct{}
 
 	get     uint32
 	put     uint32
+	half    uint32
 	less    uint32
 	equal   uint32
 	greater uint32
@@ -35,16 +41,15 @@ type BufferPool struct {
 }
 
 func (p *BufferPool) poolNum(n int) int {
-	switch {
-	case n <= p.baseline0:
+	if n <= p.baseline0 && n > p.baseline0/2 {
 		return 0
-	case n <= p.baseline1:
-		return 1
-	case n <= p.baseline2:
-		return 2
-	default:
-		return 3
 	}
+	for i, x := range p.baseline {
+		if n <= x {
+			return i + 1
+		}
+	}
+	return len(p.baseline) + 1
 }
 
 // Get returns buffer with length of n.
@@ -59,13 +64,22 @@ func (p *BufferPool) Get(n int) []byte {
 		case b := <-pool:
 			switch {
 			case cap(b) > n:
+				if cap(b)-n >= n {
+					atomic.AddUint32(&p.half, 1)
+					select {
+					case pool <- b:
+					default:
+					}
+					return make([]byte, n)
+				} else {
 					atomic.AddUint32(&p.less, 1)
 					return b[:n]
+				}
 			case cap(b) == n:
 				atomic.AddUint32(&p.equal, 1)
 				return b[:n]
 			default:
-				panic("not reached")
+				atomic.AddUint32(&p.greater, 1)
 			}
 		default:
 			atomic.AddUint32(&p.miss, 1)
@@ -79,8 +93,23 @@ func (p *BufferPool) Get(n int) []byte {
 		case b := <-pool:
 			switch {
 			case cap(b) > n:
+				if cap(b)-n >= n {
+					atomic.AddUint32(&p.half, 1)
+					sizeHalfPtr := &p.sizeHalf[poolNum-1]
+					if atomic.AddUint32(sizeHalfPtr, 1) == 20 {
+						atomic.StoreUint32(sizePtr, uint32(cap(b)/2))
+						atomic.StoreUint32(sizeHalfPtr, 0)
+					} else {
+						select {
+						case pool <- b:
+						default:
+						}
+					}
+					return make([]byte, n)
+				} else {
 					atomic.AddUint32(&p.less, 1)
 					return b[:n]
+				}
 			case cap(b) == n:
 				atomic.AddUint32(&p.equal, 1)
 				return b[:n]
@@ -126,22 +155,36 @@ func (p *BufferPool) Put(b []byte) {
 
 }
 
+func (p *BufferPool) Close() {
+	select {
+	case p.close <- struct{}{}:
+	default:
+	}
+}
+
 func (p *BufferPool) String() string {
-	return fmt.Sprintf("BufferPool{B·%d Z·%v Zm·%v G·%d P·%d <·%d =·%d >·%d M·%d}",
-		p.baseline0, p.size, p.sizeMiss, p.get, p.put, p.less, p.equal, p.greater, p.miss)
+	return fmt.Sprintf("BufferPool{B·%d Z·%v Zm·%v Zh·%v G·%d P·%d H·%d <·%d =·%d >·%d M·%d}",
+		p.baseline0, p.size, p.sizeMiss, p.sizeHalf, p.get, p.put, p.half, p.less, p.equal, p.greater, p.miss)
 }
 
 func (p *BufferPool) drain() {
+	ticker := time.NewTicker(2 * time.Second)
 	for {
-		time.Sleep(1 * time.Second)
 		select {
-		case <-p.pool[0]:
-		case <-p.pool[1]:
-		case <-p.pool[2]:
-		case <-p.pool[3]:
+		case <-ticker.C:
+			for _, ch := range p.pool {
+				select {
+				case <-ch:
 				default:
 				}
 			}
+		case <-p.close:
+			for _, ch := range p.pool {
+				close(ch)
+			}
+			return
+		}
+	}
 }
 
 // NewBufferPool creates a new initialized 'buffer pool'.
@@ -151,10 +194,10 @@ func NewBufferPool(baseline int) *BufferPool {
 	}
 	p := &BufferPool{
 		baseline0: baseline,
-		baseline1: baseline * 2,
-		baseline2: baseline * 4,
+		baseline:  [...]int{baseline / 4, baseline / 2, baseline * 2, baseline * 4},
+		close:     make(chan struct{}, 1),
 	}
-	for i, cap := range []int{6, 6, 3, 1} {
+	for i, cap := range []int{2, 2, 4, 4, 2, 1} {
 		p.pool[i] = make(chan []byte, cap)
 	}
 	go p.drain()