quorum/trie/trie.go

444 lines
13 KiB
Go
Raw Normal View History

2015-07-06 17:54:22 -07:00
// Copyright 2014 The go-ethereum Authors
// This file is part of the go-ethereum library.
2015-07-06 17:54:22 -07:00
//
// The go-ethereum library is free software: you can redistribute it and/or modify
2015-07-06 17:54:22 -07:00
// 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,
2015-07-06 17:54:22 -07:00
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2015-07-06 17:54:22 -07:00
// 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/>.
2015-07-06 17:54:22 -07:00
2015-07-06 20:08:16 -07:00
// Package trie implements Merkle Patricia Tries.
2014-10-31 06:45:03 -07:00
package trie
2014-02-14 14:56:09 -08:00
import (
2014-07-02 08:47:18 -07:00
"bytes"
2015-07-05 16:19:48 -07:00
"errors"
2014-02-14 14:56:09 -08:00
"fmt"
2015-07-05 16:19:48 -07:00
"hash"
2014-08-04 01:38:18 -07:00
2015-03-16 03:27:38 -07:00
"github.com/ethereum/go-ethereum/common"
2015-07-05 16:19:48 -07:00
"github.com/ethereum/go-ethereum/crypto/sha3"
"github.com/ethereum/go-ethereum/logger"
"github.com/ethereum/go-ethereum/logger/glog"
"github.com/ethereum/go-ethereum/rlp"
2014-02-14 14:56:09 -08:00
)
2015-07-05 16:19:48 -07:00
const defaultCacheCapacity = 800
2014-07-02 08:47:18 -07:00
2015-07-05 16:19:48 -07:00
var (
// The global cache stores decoded trie nodes by hash as they get loaded.
globalCache = newARC(defaultCacheCapacity)
// This is the known root hash of an empty trie.
emptyRoot = common.HexToHash("56e81f171bcc55a6ff8345e692c0f86e5b48e01b996cadc001622fb5e363b421")
)
2014-02-14 14:56:09 -08:00
2015-07-05 16:19:48 -07:00
var ErrMissingRoot = errors.New("missing root node")
2014-02-14 14:56:09 -08:00
2015-07-05 16:19:48 -07:00
// Database must be implemented by backing stores for the trie.
type Database interface {
DatabaseWriter
// Get returns the value for key from the database.
Get(key []byte) (value []byte, err error)
2014-06-30 04:08:00 -07:00
}
2015-07-05 16:19:48 -07:00
// DatabaseWriter wraps the Put method of a backing store for the trie.
type DatabaseWriter interface {
// Put stores the mapping key->value in the database.
// Implementations must not hold onto the value bytes, the trie
// will reuse the slice across calls to Put.
Put(key, value []byte) error
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
// Trie is a Merkle Patricia Trie.
// The zero value is an empty trie with no database.
// Use New to create a trie that sits on top of a database.
//
// Trie is not safe for concurrent use.
type Trie struct {
root node
db Database
*hasher
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
// New creates a trie with an existing root node from db.
//
// If root is the zero hash or the sha3 hash of an empty string, the
// trie is initially empty and does not require a database. Otherwise,
// New will panics if db is nil or root does not exist in the
// database. Accessing the trie loads nodes from db on demand.
func New(root common.Hash, db Database) (*Trie, error) {
trie := &Trie{db: db}
if (root != common.Hash{}) && root != emptyRoot {
if db == nil {
panic("trie.New: cannot use existing root without a database")
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
if v, _ := trie.db.Get(root[:]); len(v) == 0 {
return nil, ErrMissingRoot
}
trie.root = hashNode(root.Bytes())
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
return trie, nil
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
// Iterator returns an iterator over all mappings in the trie.
func (t *Trie) Iterator() *Iterator {
return NewIterator(t)
2014-04-29 03:36:27 -07:00
}
2015-07-05 16:19:48 -07:00
// Get returns the value for key stored in the trie.
// The value bytes must not be modified by the caller.
func (t *Trie) Get(key []byte) []byte {
key = compactHexDecode(key)
tn := t.root
for len(key) > 0 {
switch n := tn.(type) {
case shortNode:
if len(key) < len(n.Key) || !bytes.Equal(n.Key, key[:len(n.Key)]) {
return nil
}
tn = n.Val
key = key[len(n.Key):]
case fullNode:
tn = n[key[0]]
key = key[1:]
case nil:
return nil
case hashNode:
tn = t.resolveHash(n)
default:
panic(fmt.Sprintf("%T: invalid node: %v", tn, tn))
}
2014-10-29 06:20:42 -07:00
}
2015-07-05 16:19:48 -07:00
return tn.(valueNode)
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
// Update associates key with value in the trie. Subsequent calls to
// Get will return value. If value has length zero, any existing value
// is deleted from the trie and calls to Get will return nil.
//
// The value bytes must not be modified by the caller while they are
// stored in the trie.
func (t *Trie) Update(key, value []byte) {
k := compactHexDecode(key)
2015-01-08 02:47:04 -08:00
if len(value) != 0 {
2015-07-05 16:19:48 -07:00
t.root = t.insert(t.root, k, valueNode(value))
} else {
2015-07-05 16:19:48 -07:00
t.root = t.delete(t.root, k)
}
2014-07-02 04:40:02 -07:00
}
2015-07-05 16:19:48 -07:00
func (t *Trie) insert(n node, key []byte, value node) node {
2015-01-08 02:47:04 -08:00
if len(key) == 0 {
return value
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
switch n := n.(type) {
case shortNode:
matchlen := prefixLen(key, n.Key)
// If the whole key matches, keep this short node as is
// and only update the value.
if matchlen == len(n.Key) {
return shortNode{n.Key, t.insert(n.Val, key[matchlen:], value)}
2015-01-08 02:47:04 -08:00
}
2015-07-05 16:19:48 -07:00
// Otherwise branch out at the index where they differ.
var branch fullNode
branch[n.Key[matchlen]] = t.insert(nil, n.Key[matchlen+1:], n.Val)
branch[key[matchlen]] = t.insert(nil, key[matchlen+1:], value)
// Replace this shortNode with the branch if it occurs at index 0.
if matchlen == 0 {
return branch
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
// Otherwise, replace it with a short node leading up to the branch.
return shortNode{key[:matchlen], branch}
2014-02-14 14:56:09 -08:00
2015-07-05 16:19:48 -07:00
case fullNode:
n[key[0]] = t.insert(n[key[0]], key[1:], value)
return n
2014-02-14 14:56:09 -08:00
2015-07-05 16:19:48 -07:00
case nil:
return shortNode{key, value}
2014-02-14 14:56:09 -08:00
2015-07-05 16:19:48 -07:00
case hashNode:
// We've hit a part of the trie that isn't loaded yet. Load
// the node and insert into it. This leaves all child nodes on
// the path to the value in the trie.
//
// TODO: track whether insertion changed the value and keep
// n as a hash node if it didn't.
return t.insert(t.resolveHash(n), key, value)
2014-02-14 14:56:09 -08:00
2015-01-08 02:47:04 -08:00
default:
2015-07-05 16:19:48 -07:00
panic(fmt.Sprintf("%T: invalid node: %v", n, n))
2014-02-14 14:56:09 -08:00
}
}
2015-07-05 16:19:48 -07:00
// Delete removes any existing value for key from the trie.
func (t *Trie) Delete(key []byte) {
k := compactHexDecode(key)
t.root = t.delete(t.root, k)
2014-02-14 14:56:09 -08:00
}
2015-07-05 16:19:48 -07:00
// delete returns the new root of the trie with key deleted.
// It reduces the trie to minimal form by simplifying
// nodes on the way up after deleting recursively.
func (t *Trie) delete(n node, key []byte) node {
switch n := n.(type) {
case shortNode:
matchlen := prefixLen(key, n.Key)
if matchlen < len(n.Key) {
return n // don't replace n on mismatch
}
if matchlen == len(key) {
return nil // remove n entirely for whole matches
}
// The key is longer than n.Key. Remove the remaining suffix
// from the subtrie. Child can never be nil here since the
// subtrie must contain at least two other values with keys
// longer than n.Key.
child := t.delete(n.Val, key[len(n.Key):])
switch child := child.(type) {
case shortNode:
// Deleting from the subtrie reduced it to another
// short node. Merge the nodes to avoid creating a
// shortNode{..., shortNode{...}}. Use concat (which
// always creates a new slice) instead of append to
// avoid modifying n.Key since it might be shared with
// other nodes.
return shortNode{concat(n.Key, child.Key...), child.Val}
default:
return shortNode{n.Key, child}
2014-02-20 05:40:00 -08:00
}
2015-07-05 16:19:48 -07:00
case fullNode:
n[key[0]] = t.delete(n[key[0]], key[1:])
// Check how many non-nil entries are left after deleting and
// reduce the full node to a short node if only one entry is
// left. Since n must've contained at least two children
// before deletion (otherwise it would not be a full node) n
// can never be reduced to nil.
//
// When the loop is done, pos contains the index of the single
// value that is left in n or -2 if n contains at least two
// values.
2015-01-08 02:47:04 -08:00
pos := -1
2015-07-05 16:19:48 -07:00
for i, cld := range n {
if cld != nil {
2015-01-08 02:47:04 -08:00
if pos == -1 {
pos = i
2014-02-20 05:40:00 -08:00
} else {
2015-01-08 02:47:04 -08:00
pos = -2
2015-07-05 16:19:48 -07:00
break
2014-02-20 05:40:00 -08:00
}
}
}
2015-07-05 16:19:48 -07:00
if pos >= 0 {
if pos != 16 {
// If the remaining entry is a short node, it replaces
// n and its key gets the missing nibble tacked to the
// front. This avoids creating an invalid
// shortNode{..., shortNode{...}}. Since the entry
// might not be loaded yet, resolve it just for this
// check.
cnode := t.resolve(n[pos])
if cnode, ok := cnode.(shortNode); ok {
k := append([]byte{byte(pos)}, cnode.Key...)
return shortNode{k, cnode.Val}
}
2015-01-08 02:47:04 -08:00
}
2015-07-05 16:19:48 -07:00
// Otherwise, n is replaced by a one-nibble short node
// containing the child.
return shortNode{[]byte{byte(pos)}, n[pos]}
2014-02-20 05:40:00 -08:00
}
2015-07-05 16:19:48 -07:00
// n still contains at least two values and cannot be reduced.
return n
2014-02-20 05:40:00 -08:00
2015-01-08 02:47:04 -08:00
case nil:
return nil
2015-07-05 16:19:48 -07:00
case hashNode:
// We've hit a part of the trie that isn't loaded yet. Load
// the node and delete from it. This leaves all child nodes on
// the path to the value in the trie.
//
// TODO: track whether deletion actually hit a key and keep
// n as a hash node if it didn't.
return t.delete(t.resolveHash(n), key)
2015-01-08 02:47:04 -08:00
default:
2015-07-05 16:19:48 -07:00
panic(fmt.Sprintf("%T: invalid node: %v (%v)", n, n, key))
2014-02-20 05:40:00 -08:00
}
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
func concat(s1 []byte, s2 ...byte) []byte {
r := make([]byte, len(s1)+len(s2))
copy(r, s1)
copy(r[len(s1):], s2)
return r
}
func (t *Trie) resolve(n node) node {
if n, ok := n.(hashNode); ok {
return t.resolveHash(n)
}
return n
}
func (t *Trie) resolveHash(n hashNode) node {
if v, ok := globalCache.Get(n); ok {
return v
}
enc, err := t.db.Get(n)
if err != nil || enc == nil {
// TODO: This needs to be improved to properly distinguish errors.
// Disk I/O errors shouldn't produce nil (and cause a
// consensus failure or weird crash), but it is unclear how
// they could be handled because the entire stack above the trie isn't
// prepared to cope with missing state nodes.
if glog.V(logger.Error) {
glog.Errorf("Dangling hash node ref %x: %v", n, err)
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
return nil
}
dec := mustDecodeNode(n, enc)
if dec != nil {
globalCache.Put(n, dec)
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
return dec
}
// Root returns the root hash of the trie.
// Deprecated: use Hash instead.
func (t *Trie) Root() []byte { return t.Hash().Bytes() }
2014-02-24 03:11:00 -08:00
2015-07-05 16:19:48 -07:00
// Hash returns the root hash of the trie. It does not write to the
// database and can be used even if the trie doesn't have one.
func (t *Trie) Hash() common.Hash {
root, _ := t.hashRoot(nil)
return common.BytesToHash(root.(hashNode))
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
// Commit writes all nodes to the trie's database.
// Nodes are stored with their sha3 hash as the key.
//
// Committing flushes nodes from memory.
// Subsequent Get calls will load nodes from the database.
func (t *Trie) Commit() (root common.Hash, err error) {
if t.db == nil {
panic("Commit called on trie with nil database")
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
return t.CommitTo(t.db)
2014-02-24 03:11:00 -08:00
}
2015-07-05 16:19:48 -07:00
// CommitTo writes all nodes to the given database.
// Nodes are stored with their sha3 hash as the key.
//
// Committing flushes nodes from memory. Subsequent Get calls will
// load nodes from the trie's database. Calling code must ensure that
// the changes made to db are written back to the trie's attached
// database before using the trie.
func (t *Trie) CommitTo(db DatabaseWriter) (root common.Hash, err error) {
n, err := t.hashRoot(db)
if err != nil {
return (common.Hash{}), err
}
t.root = n
return common.BytesToHash(n.(hashNode)), nil
}
2014-05-26 16:08:51 -07:00
2015-07-05 16:19:48 -07:00
func (t *Trie) hashRoot(db DatabaseWriter) (node, error) {
if t.root == nil {
return hashNode(emptyRoot.Bytes()), nil
}
if t.hasher == nil {
t.hasher = newHasher()
2015-01-08 02:47:04 -08:00
}
2015-07-05 16:19:48 -07:00
return t.hasher.hash(t.root, db, true)
}
2014-05-26 16:08:51 -07:00
2015-07-05 16:19:48 -07:00
type hasher struct {
tmp *bytes.Buffer
sha hash.Hash
2014-05-26 16:08:51 -07:00
}
2015-07-05 16:19:48 -07:00
func newHasher() *hasher {
return &hasher{tmp: new(bytes.Buffer), sha: sha3.NewKeccak256()}
}
func (h *hasher) hash(n node, db DatabaseWriter, force bool) (node, error) {
hashed, err := h.replaceChildren(n, db)
if err != nil {
return hashNode{}, err
}
if n, err = h.store(hashed, db, force); err != nil {
return hashNode{}, err
}
return n, nil
}
// hashChildren replaces child nodes of n with their hashes if the encoded
// size of the child is larger than a hash.
func (h *hasher) replaceChildren(n node, db DatabaseWriter) (node, error) {
var err error
switch n := n.(type) {
case shortNode:
n.Key = compactEncode(n.Key)
if _, ok := n.Val.(valueNode); !ok {
if n.Val, err = h.hash(n.Val, db, false); err != nil {
return n, err
}
}
if n.Val == nil {
// Ensure that nil children are encoded as empty strings.
n.Val = valueNode(nil)
}
return n, nil
case fullNode:
for i := 0; i < 16; i++ {
if n[i] != nil {
if n[i], err = h.hash(n[i], db, false); err != nil {
return n, err
}
} else {
// Ensure that nil children are encoded as empty strings.
n[i] = valueNode(nil)
}
}
if n[16] == nil {
n[16] = valueNode(nil)
}
return n, nil
default:
return n, nil
}
}
func (h *hasher) store(n node, db DatabaseWriter, force bool) (node, error) {
// Don't store hashes or empty nodes.
if _, isHash := n.(hashNode); n == nil || isHash {
return n, nil
}
h.tmp.Reset()
if err := rlp.Encode(h.tmp, n); err != nil {
panic("encode error: " + err.Error())
}
if h.tmp.Len() < 32 && !force {
// Nodes smaller than 32 bytes are stored inside their parent.
return n, nil
}
// Larger nodes are replaced by their hash and stored in the database.
h.sha.Reset()
h.sha.Write(h.tmp.Bytes())
key := hashNode(h.sha.Sum(nil))
if db != nil {
err := db.Put(key, h.tmp.Bytes())
return key, err
}
return key, nil
2014-05-26 16:08:51 -07:00
}