go-ethereum/trie/utils/verkle.go

// Copyright 2023 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 utils

import (
	"encoding/binary"
	"sync"

	"github.com/crate-crypto/go-ipa/bandersnatch/fr"
	"github.com/ethereum/go-ethereum/common/lru"
	"github.com/ethereum/go-ethereum/metrics"
	"github.com/ethereum/go-verkle"
	"github.com/holiman/uint256"
)

const (
	BasicDataLeafKey = 0
	CodeHashLeafKey  = 1

	BasicDataVersionOffset  = 0
	BasicDataCodeSizeOffset = 5
	BasicDataNonceOffset    = 8
	BasicDataBalanceOffset  = 16
)

var (
	zero                                = uint256.NewInt(0)
	verkleNodeWidthLog2                 = 8
	headerStorageOffset                 = uint256.NewInt(64)
	mainStorageOffsetLshVerkleNodeWidth = new(uint256.Int).Lsh(uint256.NewInt(256), 31-uint(verkleNodeWidthLog2))
	codeOffset                          = uint256.NewInt(128)
	verkleNodeWidth                     = uint256.NewInt(256)
	codeStorageDelta                    = uint256.NewInt(0).Sub(codeOffset, headerStorageOffset)

	index0Point *verkle.Point // pre-computed commitment of polynomial [2+256*64]

	// cacheHitGauge is the metric to track how many cache hit occurred.
	cacheHitGauge = metrics.NewRegisteredGauge("trie/verkle/cache/hit", nil)

	// cacheMissGauge is the metric to track how many cache miss occurred.
	cacheMissGauge = metrics.NewRegisteredGauge("trie/verkle/cache/miss", nil)
)

func init() {
	// The byte array is the Marshalled output of the point computed as such:
	//
	// 	var (
	//		config = verkle.GetConfig()
	//		fr     verkle.Fr
	//	)
	//	verkle.FromLEBytes(&fr, []byte{2, 64})
	//	point := config.CommitToPoly([]verkle.Fr{fr}, 1)
	index0Point = new(verkle.Point)
	err := index0Point.SetBytes([]byte{34, 25, 109, 242, 193, 5, 144, 224, 76, 52, 189, 92, 197, 126, 9, 145, 27, 152, 199, 130, 165, 3, 210, 27, 193, 131, 142, 28, 110, 26, 16, 191})
	if err != nil {
		panic(err)
	}
}

// PointCache is the LRU cache for storing evaluated address commitment.
type PointCache struct {
	lru  lru.BasicLRU[string, *verkle.Point]
	lock sync.RWMutex
}

// NewPointCache returns the cache with specified size.
func NewPointCache(maxItems int) *PointCache {
	return &PointCache{
		lru: lru.NewBasicLRU[string, *verkle.Point](maxItems),
	}
}

// Get returns the cached commitment for the specified address, or computing
// it on the flight.
func (c *PointCache) Get(addr []byte) *verkle.Point {
	c.lock.Lock()
	defer c.lock.Unlock()

	p, ok := c.lru.Get(string(addr))
	if ok {
		cacheHitGauge.Inc(1)
		return p
	}
	cacheMissGauge.Inc(1)
	p = evaluateAddressPoint(addr)
	c.lru.Add(string(addr), p)
	return p
}

// GetStem returns the first 31 bytes of the tree key as the tree stem. It only
// works for the account metadata whose treeIndex is 0.
func (c *PointCache) GetStem(addr []byte) []byte {
	p := c.Get(addr)
	return pointToHash(p, 0)[:31]
}

// GetTreeKey performs both the work of the spec's get_tree_key function, and that
// of pedersen_hash: it builds the polynomial in pedersen_hash without having to
// create a mostly zero-filled buffer and "type cast" it to a 128-long 16-byte
// array. Since at most the first 5 coefficients of the polynomial will be non-zero,
// these 5 coefficients are created directly.
func GetTreeKey(address []byte, treeIndex *uint256.Int, subIndex byte) []byte {
	if len(address) < 32 {
		var aligned [32]byte
		address = append(aligned[:32-len(address)], address...)
	}
	// poly = [2+256*64, address_le_low, address_le_high, tree_index_le_low, tree_index_le_high]
	var poly [5]fr.Element

	// 32-byte address, interpreted as two little endian
	// 16-byte numbers.
	verkle.FromLEBytes(&poly[1], address[:16])
	verkle.FromLEBytes(&poly[2], address[16:])

	// treeIndex must be interpreted as a 32-byte aligned little-endian integer.
	// e.g: if treeIndex is 0xAABBCC, we need the byte representation to be 0xCCBBAA00...00.
	// poly[3] = LE({CC,BB,AA,00...0}) (16 bytes), poly[4]=LE({00,00,...}) (16 bytes).
	//
	// To avoid unnecessary endianness conversions for go-ipa, we do some trick:
	// - poly[3]'s byte representation is the same as the *top* 16 bytes (trieIndexBytes[16:]) of
	//   32-byte aligned big-endian representation (BE({00,...,AA,BB,CC})).
	// - poly[4]'s byte representation is the same as the *low* 16 bytes (trieIndexBytes[:16]) of
	//   the 32-byte aligned big-endian representation (BE({00,00,...}).
	trieIndexBytes := treeIndex.Bytes32()
	verkle.FromBytes(&poly[3], trieIndexBytes[16:])
	verkle.FromBytes(&poly[4], trieIndexBytes[:16])

	cfg := verkle.GetConfig()
	ret := cfg.CommitToPoly(poly[:], 0)

	// add a constant point corresponding to poly[0]=[2+256*64].
	ret.Add(ret, index0Point)

	return pointToHash(ret, subIndex)
}

// GetTreeKeyWithEvaluatedAddress is basically identical to GetTreeKey, the only
// difference is a part of polynomial is already evaluated.
//
// Specifically, poly = [2+256*64, address_le_low, address_le_high] is already
// evaluated.
func GetTreeKeyWithEvaluatedAddress(evaluated *verkle.Point, treeIndex *uint256.Int, subIndex byte) []byte {
	var poly [5]fr.Element

	// little-endian, 32-byte aligned treeIndex
	var index [32]byte
	for i := 0; i < len(treeIndex); i++ {
		binary.LittleEndian.PutUint64(index[i*8:(i+1)*8], treeIndex[i])
	}
	verkle.FromLEBytes(&poly[3], index[:16])
	verkle.FromLEBytes(&poly[4], index[16:])

	cfg := verkle.GetConfig()
	ret := cfg.CommitToPoly(poly[:], 0)

	// add the pre-evaluated address
	ret.Add(ret, evaluated)

	return pointToHash(ret, subIndex)
}

// BasicDataKey returns the verkle tree key of the basic data field for
// the specified account.
func BasicDataKey(address []byte) []byte {
	return GetTreeKey(address, zero, BasicDataLeafKey)
}

// CodeHashKey returns the verkle tree key of the code hash field for
// the specified account.
func CodeHashKey(address []byte) []byte {
	return GetTreeKey(address, zero, CodeHashLeafKey)
}

func codeChunkIndex(chunk *uint256.Int) (*uint256.Int, byte) {
	var (
		chunkOffset            = new(uint256.Int).Add(codeOffset, chunk)
		treeIndex, subIndexMod = new(uint256.Int).DivMod(chunkOffset, verkleNodeWidth, new(uint256.Int))
	)
	return treeIndex, byte(subIndexMod.Uint64())
}

// CodeChunkKey returns the verkle tree key of the code chunk for the
// specified account.
func CodeChunkKey(address []byte, chunk *uint256.Int) []byte {
	treeIndex, subIndex := codeChunkIndex(chunk)
	return GetTreeKey(address, treeIndex, subIndex)
}

func StorageIndex(storageKey []byte) (*uint256.Int, byte) {
	// If the storage slot is in the header, we need to add the header offset.
	var key uint256.Int
	key.SetBytes(storageKey)
	if key.Cmp(codeStorageDelta) < 0 {
		// This addition is always safe; it can't ever overflow since pos<codeStorageDelta.
		key.Add(headerStorageOffset, &key)

		// In this branch, the tree-index is zero since we're in the account header,
		// and the sub-index is the LSB of the modified storage key.
		return zero, byte(key[0] & 0xFF)
	}
	// If the storage slot is in the main storage, we need to add the main storage offset.

	// The first MAIN_STORAGE_OFFSET group will see its
	// first 64 slots unreachable. This is either a typo in the
	// spec or intended to conserve the 256-u256
	// aligment. If we decide to ever access these 64
	// slots, uncomment this.
	// // Get the new offset since we now know that we are above 64.
	// pos.Sub(&pos, codeStorageDelta)
	// suffix := byte(pos[0] & 0xFF)
	suffix := storageKey[len(storageKey)-1]

	// We first divide by VerkleNodeWidth to create room to avoid an overflow next.
	key.Rsh(&key, uint(verkleNodeWidthLog2))

	// We add mainStorageOffset/VerkleNodeWidth which can't overflow.
	key.Add(&key, mainStorageOffsetLshVerkleNodeWidth)

	// The sub-index is the LSB of the original storage key, since mainStorageOffset
	// doesn't affect this byte, so we can avoid masks or shifts.
	return &key, suffix
}

// StorageSlotKey returns the verkle tree key of the storage slot for the
// specified account.
func StorageSlotKey(address []byte, storageKey []byte) []byte {
	treeIndex, subIndex := StorageIndex(storageKey)
	return GetTreeKey(address, treeIndex, subIndex)
}

// BasicDataKeyWithEvaluatedAddress returns the verkle tree key of the basic data
// field for the specified account. The difference between BasicDataKey is the
// address evaluation is already computed to minimize the computational overhead.
func BasicDataKeyWithEvaluatedAddress(evaluated *verkle.Point) []byte {
	return GetTreeKeyWithEvaluatedAddress(evaluated, zero, BasicDataLeafKey)
}

// CodeHashKeyWithEvaluatedAddress returns the verkle tree key of the code
// hash for the specified account. The difference between CodeHashKey is the
// address evaluation is already computed to minimize the computational overhead.
func CodeHashKeyWithEvaluatedAddress(evaluated *verkle.Point) []byte {
	return GetTreeKeyWithEvaluatedAddress(evaluated, zero, CodeHashLeafKey)
}

// CodeChunkKeyWithEvaluatedAddress returns the verkle tree key of the code
// chunk for the specified account. The difference between CodeChunkKey is the
// address evaluation is already computed to minimize the computational overhead.
func CodeChunkKeyWithEvaluatedAddress(addressPoint *verkle.Point, chunk *uint256.Int) []byte {
	treeIndex, subIndex := codeChunkIndex(chunk)
	return GetTreeKeyWithEvaluatedAddress(addressPoint, treeIndex, subIndex)
}

// StorageSlotKeyWithEvaluatedAddress returns the verkle tree key of the storage
// slot for the specified account. The difference between StorageSlotKey is the
// address evaluation is already computed to minimize the computational overhead.
func StorageSlotKeyWithEvaluatedAddress(evaluated *verkle.Point, storageKey []byte) []byte {
	treeIndex, subIndex := StorageIndex(storageKey)
	return GetTreeKeyWithEvaluatedAddress(evaluated, treeIndex, subIndex)
}

func pointToHash(evaluated *verkle.Point, suffix byte) []byte {
	// The output of Byte() is big endian for banderwagon. This
	// introduces an imbalance in the tree, because hashes are
	// elements of a 253-bit field. This means more than half the
	// tree would be empty. To avoid this problem, use a little
	// endian commitment and chop the MSB.
	bytes := evaluated.Bytes()
	for i := 0; i < 16; i++ {
		bytes[31-i], bytes[i] = bytes[i], bytes[31-i]
	}
	bytes[31] = suffix
	return bytes[:]
}

func evaluateAddressPoint(address []byte) *verkle.Point {
	if len(address) < 32 {
		var aligned [32]byte
		address = append(aligned[:32-len(address)], address...)
	}
	var poly [3]fr.Element

	// 32-byte address, interpreted as two little endian
	// 16-byte numbers.
	verkle.FromLEBytes(&poly[1], address[:16])
	verkle.FromLEBytes(&poly[2], address[16:])

	cfg := verkle.GetConfig()
	ret := cfg.CommitToPoly(poly[:], 0)

	// add a constant point
	ret.Add(ret, index0Point)
	return ret
}