trie: separate pruner and optimize liveness check

This commit is contained in:
Péter Szilágyi 2019-01-07 14:43:47 +02:00
parent 0d93fc76b2
commit e24c092d0f
No known key found for this signature in database
GPG key ID: E9AE538CEDF8293D
2 changed files with 251 additions and 147 deletions

View file

@ -19,7 +19,6 @@ package trie
import (
"fmt"
"io"
"math/big"
"sync"
"time"
@ -572,7 +571,7 @@ func (db *Database) Dereference(root common.Hash, prune bool) error {
nodes, storage, start := len(db.dirties), db.dirtiesSize, time.Now()
prunetime, prunenodes, prunesize := db.prunetime, db.prunenodes, db.prunesize
if err := db.dereference(common.Hash{}, root, common.Hash{}, common.Hash{}, prune, nil, make(map[string]node)); err != nil {
if err := db.dereference(common.Hash{}, root, common.Hash{}, common.Hash{}, prune, nil, db.newPruner(root)); err != nil {
return err
}
db.gcnodes += uint64(nodes - len(db.dirties))
@ -595,7 +594,7 @@ func (db *Database) Dereference(root common.Hash, prune bool) error {
}
// dereference is the private locked version of Dereference.
func (db *Database) dereference(childOwner common.Hash, childHash common.Hash, parentOwner common.Hash, parentHash common.Hash, prune bool, path []byte, cache map[string]node) error {
func (db *Database) dereference(childOwner common.Hash, childHash common.Hash, parentOwner common.Hash, parentHash common.Hash, prune bool, path []byte, pruner *pruner) error {
// Dereference the parent-child
parentKey := makeNodeKey(parentOwner, parentHash)
parent := db.dirties[parentKey]
@ -614,7 +613,7 @@ func (db *Database) dereference(childOwner common.Hash, childHash common.Hash, p
batch := db.diskdb.NewBatch()
start := time.Now()
db.prune(childOwner, childHash, path, batch, cache)
pruner.prune(childOwner, childHash, path, batch)
db.prunetime += time.Since(start)
if err := batch.Write(); err != nil {
@ -646,12 +645,12 @@ func (db *Database) dereference(childOwner common.Hash, childHash common.Hash, p
}
// Dereference all children and delete the node
child.iterateRefs(path, func(path []byte, hash common.Hash) error {
db.dereference(childOwner, hash, childOwner, childHash, prune, path, cache)
db.dereference(childOwner, hash, childOwner, childHash, prune, path, pruner)
return nil
})
for key := range child.children {
owner, hash := splitNodeKey(key)
db.dereference(owner, hash, childOwner, childHash, prune, nil, cache)
db.dereference(owner, hash, childOwner, childHash, prune, nil, pruner)
}
delete(db.dirties, childKey)
db.dirtiesSize -= common.StorageSize(common.HashLength + int(child.size))
@ -659,147 +658,6 @@ func (db *Database) dereference(childOwner common.Hash, childHash common.Hash, p
return nil
}
// prune deletes a trie node from disk if there are no more live references to
// it, cascading until all dangling nodes are removed.
func (db *Database) prune(owner common.Hash, hash common.Hash, path []byte, batch ethdb.Batch, cache map[string]node) {
// If the node is still live in the memory cache, it's still referenced so we
// can abort. This case is important when and old trie being pruned references
// a new node (maybe that node was recreted since), since currently live nodes
// are stored expanded, not as hashes.
key := makeNodeKey(owner, hash)
if db.dirties[key] != nil {
return
}
// Iterate over all the live tries in the cache and check node liveliness
for key := range db.dirties[metaRoot].children {
_, root := splitNodeKey(key)
var paths [][]byte
if owner != (common.Hash{}) {
paths = [][]byte{keybytesToHex(owner[:])}
}
if db.live(hashNode(root[:]), owner, hash, append(paths, path), cache) {
return
}
}
// Dead node found, delete it from the database
dead := []byte(makeNodeKey(owner, hash))
blob, err := db.diskdb.Get(dead)
if blob == nil || err != nil {
log.Error("Missing prune target", "owner", owner, "hash", hash, "path", fmt.Sprintf("%x", path))
return
}
node := mustDecodeNode(hash[:], blob, 0)
// Prune the node and its children if it's not a bytecode blob
db.cleans.Delete(key)
batch.Delete(dead)
db.prunenodes++
db.prunesize += common.StorageSize(len(blob))
iterateRefs(node, path, func(path []byte, hash common.Hash) error {
db.prune(owner, hash, path, batch, cache)
return nil
})
}
// live descends in the trie and returns whether the given hash is part of the
// trie or not.
func (db *Database) live(root node, owner common.Hash, hash common.Hash, paths [][]byte, cache map[string]node) bool {
// If we reached the end of our path, it should be a hash node
if len(paths) == 1 && len(paths[0]) == 0 {
if have, ok := root.(hashNode); ok {
return common.BytesToHash(have) == hash
}
// Not a hash node? It rarely happens that a 32+ byte leaf short node gets
// turned into a 31- byte one, converting if from a hash node to an embedded
// one. Allow this case, but reject as a wrong path.
// TODO(karalabe): get rid of this warning, only curiosity
log.Warn("Liveness check terminated on non-hash", "type", fmt.Sprintf("%T", root), "node", root.fstring(""))
return false
}
// If we're at a hash node, expand before continuing
if n, ok := root.(hashNode); ok {
var (
key string
hash = common.BytesToHash(n)
)
if len(paths) > 1 {
key = makeNodeKey(common.Hash{}, hash)
} else {
key = makeNodeKey(owner, hash)
}
if enc, err := db.cleans.Get(key); err == nil && enc != nil {
root = mustDecodeNode(hash[:], enc, 0)
cache[key] = root
} else if node := db.dirties[key]; node != nil {
root = node.node
} else if node := cache[key]; node != nil {
root = node
} else {
blob, err := db.diskdb.Get([]byte(key))
if blob == nil || err != nil {
panic(fmt.Sprintf("missing referenced node %x (searching for %x:%x at %x)", key, owner, hash, paths))
}
root = mustDecodeNode(hash[:], blob, 0)
cache[key] = root
}
}
// If we reached an account node, extract the storage trie root to continue on
if len(paths) == 2 && len(paths[0]) == 0 {
if have, ok := root.(valueNode); ok {
var account struct {
Nonce uint64
Balance *big.Int
Root common.Hash
CodeHash []byte
}
if err := rlp.DecodeBytes(have, &account); err != nil {
panic(err)
}
if account.Root == emptyRoot {
return false
}
return db.live(hashNode(account.Root[:]), owner, hash, paths[1:], cache)
}
panic(fmt.Sprintf("liveness check path swap terminated on non value node: %T", root))
}
// Descend into the trie following the specified path. This code segment must
// be able to handle both simplified raw nodes kept in this cache as well as
// cold nodes loaded directly from disk.
switch n := root.(type) {
case *rawShortNode:
if prefixLen(n.Key, paths[0]) == len(n.Key) {
return db.live(n.Val, owner, hash, append([][]byte{paths[0][len(n.Key):]}, paths[1:]...), cache)
}
return false
case *shortNode:
if prefixLen(n.Key, paths[0]) == len(n.Key) {
return db.live(n.Val, owner, hash, append([][]byte{paths[0][len(n.Key):]}, paths[1:]...), cache)
}
return false
case rawFullNode:
if child := n[paths[0][0]]; child != nil {
return db.live(child, owner, hash, append([][]byte{paths[0][1:]}, paths[1:]...), cache)
}
return false
case *fullNode:
if child := n.Children[paths[0][0]]; child != nil {
return db.live(child, owner, hash, append([][]byte{paths[0][1:]}, paths[1:]...), cache)
}
return false
default:
panic(fmt.Sprintf("unknown node type: %T", n))
}
}
// Cap iteratively flushes old but still referenced trie nodes until the total
// memory usage goes below the given threshold.
func (db *Database) Cap(limit common.StorageSize) error {

246
trie/database_pruning.go Normal file
View file

@ -0,0 +1,246 @@
// Copyright 2019 The go-ethereum Authors
// This file is part of the go-ethereum library.
//
// The go-ethereum library is free software: you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
package trie
import (
"bytes"
"fmt"
"math/big"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/ethdb"
"github.com/ethereum/go-ethereum/log"
"github.com/ethereum/go-ethereum/rlp"
)
// pruner is responsible for pruning the state trie based on liveness checks
// whenever the in-memory garbage collector attempt to dereference a node from
// disk.
type pruner struct {
db *Database // Trie database for accessing dirty and clean data
tries []*traverser // Individual stateful trie traversers for fast liveness checks
}
// newPruner creates a new trie pruner tied to the liveness of all the currently
// referenced in-memory nodes, except the specified one (currently being pruned).
func (db *Database) newPruner(skip common.Hash) *pruner {
// Create the set of traversers based on the live tries
var traversers []*traverser
for key := range db.dirties[metaRoot].children {
if _, root := splitNodeKey(key); root != skip {
traversers = append(traversers, &traverser{
db: db,
state: &tranverserState{
node: hashNode(root[:]),
},
})
}
}
// Assemble and return the pruner
return &pruner{
db: db,
tries: traversers,
}
}
// prune deletes a trie node from disk if there are no more live references to
// it, cascading until all dangling nodes are removed.
func (p *pruner) prune(owner common.Hash, hash common.Hash, path []byte, batch ethdb.Batch) {
// If the node is still live in the memory cache, it's still referenced so we
// can abort. This case is important when and old trie being pruned references
// a new node (maybe that node was recreted since), since currently live nodes
// are stored expanded, not as hashes.
key := makeNodeKey(owner, hash)
if p.db.dirties[key] != nil {
return
}
// Iterate over all the live tries and check node liveliness
crosspath := path
if owner != (common.Hash{}) {
crosspath = append(append(keybytesToHex(owner[:]), 0xff), crosspath...)
}
for _, trie := range p.tries {
if trie.live(owner, hash, crosspath) {
return
}
}
// Dead node found, delete it from the database
dead := []byte(makeNodeKey(owner, hash))
blob, err := p.db.diskdb.Get(dead)
if blob == nil || err != nil {
log.Error("Missing prune target", "owner", owner, "hash", hash, "path", fmt.Sprintf("%x", path))
return
}
node := mustDecodeNode(hash[:], blob, 0)
// Prune the node and its children if it's not a bytecode blob
p.db.cleans.Delete(key)
batch.Delete(dead)
p.db.prunenodes++
p.db.prunesize += common.StorageSize(len(blob))
iterateRefs(node, path, func(path []byte, hash common.Hash) error {
p.prune(owner, hash, path, batch)
return nil
})
}
// traverser is a stateful trie traversal data structure used by the pruner to
// verify the liveness of a node within a specific trie. The reason for having
// a separate data structure is to allow reusing previous traversals to check
// the liveness of nested nodes (i.e. entire subtried during pruning).
type traverser struct {
db *Database // Trie database for accessing dirty and clean data
state *tranverserState // Leftover state from the previous traversals
}
// tranverserState is the internal state of a trie traverser.
type tranverserState struct {
parent *tranverserState // Parent traverser to allow backtracking
prefix []byte // Path leading up to the root of this traverser
node node // Trie node where this traverser is currently at
}
// live checks whether the trie iterated by this traverser contains the hashnode
// at the given path, minimizing data access and processing by reusing previous
// state instead of starting fresh.
//
// The path is a full canonical path from the account trie root down to the node
// potentially crossing over into a storage trie. The account and storage trie
// paths are separated by a 0xff byte (nibbles range from 0x00-0x10). This byte
// is needed to differentiate between the leaf of the account trie and the root
// of a storage trie (which otherwise would have the same traversal path).
func (t *traverser) live(owner common.Hash, hash common.Hash, path []byte) bool {
// Rewind the traverser until it's prefix is actually a prefix of the path
for !bytes.HasPrefix(path, t.state.prefix) {
t.state = t.state.parent
}
// Traverse downward until the prefix matches the path completely
path = path[len(t.state.prefix):]
for len(path) > 0 {
// If we're at a hash node, expand before continuing
if n, ok := t.state.node.(hashNode); ok {
// Generate the database key for this hash node
var (
key string
hash = common.BytesToHash(n)
)
if len(t.state.prefix) < 2*common.HashLength {
key = makeNodeKey(common.Hash{}, hash)
} else {
key = makeNodeKey(owner, hash)
}
// Replace the node in the traverser with the expanded one
if enc, err := t.db.cleans.Get(key); err == nil && enc != nil {
t.state.node = mustDecodeNode(hash[:], enc, 0)
} else if node := t.db.dirties[key]; node != nil {
t.state.node = node.node
} else {
blob, err := t.db.diskdb.Get([]byte(key))
if blob == nil || err != nil {
panic(fmt.Sprintf("missing referenced node %x (searching for %x:%x at %x%x)", key, owner, hash, t.state.prefix, path))
}
t.state.node = mustDecodeNode(hash[:], blob, 0)
}
}
// If we reached an account node, extract the storage trie root to continue on
if path[0] == 0xff {
// Retrieve the storage trie root and abort if empty
if have, ok := t.state.node.(valueNode); ok {
var account struct {
Nonce uint64
Balance *big.Int
Root common.Hash
CodeHash []byte
}
if err := rlp.DecodeBytes(have, &account); err != nil {
panic(err)
}
if account.Root == emptyRoot {
return false
}
// Create a new nesting in the traversal and continue on that depth
t.state, path = &tranverserState{
parent: t.state,
prefix: append(t.state.prefix, 0xff),
node: hashNode(account.Root[:]),
}, path[1:]
continue
}
panic(fmt.Sprintf("liveness check path swap terminated on non value node: %T", t.state.node))
}
// Descend into the trie following the specified path. This code segment must
// be able to handle both simplified raw nodes kept in this cache as well as
// cold nodes loaded directly from disk.
switch n := t.state.node.(type) {
case *rawShortNode:
if prefixLen(n.Key, path) == len(n.Key) {
t.state, path = &tranverserState{
parent: t.state,
prefix: append(t.state.prefix, path[:len(n.Key)]...),
node: n.Val,
}, path[len(n.Key):]
continue
}
return false
case *shortNode:
if prefixLen(n.Key, path) == len(n.Key) {
t.state, path = &tranverserState{
parent: t.state,
prefix: append(t.state.prefix, path[:len(n.Key)]...),
node: n.Val,
}, path[len(n.Key):]
continue
}
return false
case rawFullNode:
if child := n[path[0]]; child != nil {
t.state, path = &tranverserState{
parent: t.state,
prefix: append(t.state.prefix, path[0]),
node: child,
}, path[1:]
continue
}
return false
case *fullNode:
if child := n.Children[path[0]]; child != nil {
t.state, path = &tranverserState{
parent: t.state,
prefix: append(t.state.prefix, path[0]),
node: child,
}, path[1:]
continue
}
return false
default:
panic(fmt.Sprintf("unknown node type: %T", n))
}
}
// The prefix should match perfectly here, check if the hashes matches
if have, ok := t.state.node.(hashNode); ok {
return common.BytesToHash(have) == hash
}
if have, _ := t.state.node.cache(); have != nil {
return common.BytesToHash(have) == hash
}
return false
}