2020-12-11 13:06:44 +03:00
|
|
|
// Copyright 2020 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 difficulty
|
|
|
|
|
|
|
|
import (
|
|
|
|
"bytes"
|
|
|
|
"encoding/binary"
|
|
|
|
"fmt"
|
|
|
|
"io"
|
|
|
|
"math/big"
|
|
|
|
|
|
|
|
"github.com/ethereum/go-ethereum/consensus/ethash"
|
|
|
|
"github.com/ethereum/go-ethereum/core/types"
|
|
|
|
)
|
|
|
|
|
|
|
|
type fuzzer struct {
|
|
|
|
input io.Reader
|
|
|
|
exhausted bool
|
|
|
|
}
|
|
|
|
|
|
|
|
func (f *fuzzer) read(size int) []byte {
|
|
|
|
out := make([]byte, size)
|
|
|
|
if _, err := f.input.Read(out); err != nil {
|
|
|
|
f.exhausted = true
|
|
|
|
}
|
|
|
|
return out
|
|
|
|
}
|
|
|
|
|
|
|
|
func (f *fuzzer) readSlice(min, max int) []byte {
|
|
|
|
var a uint16
|
|
|
|
binary.Read(f.input, binary.LittleEndian, &a)
|
|
|
|
size := min + int(a)%(max-min)
|
|
|
|
out := make([]byte, size)
|
|
|
|
if _, err := f.input.Read(out); err != nil {
|
|
|
|
f.exhausted = true
|
|
|
|
}
|
|
|
|
return out
|
|
|
|
}
|
|
|
|
|
|
|
|
func (f *fuzzer) readUint64(min, max uint64) uint64 {
|
|
|
|
if min == max {
|
|
|
|
return min
|
|
|
|
}
|
|
|
|
var a uint64
|
|
|
|
if err := binary.Read(f.input, binary.LittleEndian, &a); err != nil {
|
|
|
|
f.exhausted = true
|
|
|
|
}
|
|
|
|
a = min + a%(max-min)
|
|
|
|
return a
|
|
|
|
}
|
|
|
|
func (f *fuzzer) readBool() bool {
|
|
|
|
return f.read(1)[0]&0x1 == 0
|
|
|
|
}
|
|
|
|
|
|
|
|
// The function must return
|
2022-09-10 14:25:40 +03:00
|
|
|
//
|
|
|
|
// - 1 if the fuzzer should increase priority of the
|
|
|
|
// given input during subsequent fuzzing (for example, the input is lexically
|
|
|
|
// correct and was parsed successfully);
|
|
|
|
// - -1 if the input must not be added to corpus even if gives new coverage; and
|
|
|
|
// - 0 otherwise
|
|
|
|
//
|
2020-12-11 13:06:44 +03:00
|
|
|
// other values are reserved for future use.
|
|
|
|
func Fuzz(data []byte) int {
|
|
|
|
f := fuzzer{
|
|
|
|
input: bytes.NewReader(data),
|
|
|
|
exhausted: false,
|
|
|
|
}
|
|
|
|
return f.fuzz()
|
|
|
|
}
|
|
|
|
|
|
|
|
var minDifficulty = big.NewInt(0x2000)
|
|
|
|
|
|
|
|
type calculator func(time uint64, parent *types.Header) *big.Int
|
|
|
|
|
|
|
|
func (f *fuzzer) fuzz() int {
|
|
|
|
// A parent header
|
|
|
|
header := &types.Header{}
|
|
|
|
if f.readBool() {
|
|
|
|
header.UncleHash = types.EmptyUncleHash
|
|
|
|
}
|
|
|
|
// Difficulty can range between 0x2000 (2 bytes) and up to 32 bytes
|
|
|
|
{
|
|
|
|
diff := new(big.Int).SetBytes(f.readSlice(2, 32))
|
|
|
|
if diff.Cmp(minDifficulty) < 0 {
|
|
|
|
diff.Set(minDifficulty)
|
|
|
|
}
|
|
|
|
header.Difficulty = diff
|
|
|
|
}
|
|
|
|
// Number can range between 0 and up to 32 bytes (but not so that the child exceeds it)
|
|
|
|
{
|
|
|
|
// However, if we use astronomic numbers, then the bomb exp karatsuba calculation
|
|
|
|
// in the legacy methods)
|
|
|
|
// times out, so we limit it to fit within reasonable bounds
|
|
|
|
number := new(big.Int).SetBytes(f.readSlice(0, 4)) // 4 bytes: 32 bits: block num max 4 billion
|
|
|
|
header.Number = number
|
|
|
|
}
|
|
|
|
// Both parent and child time must fit within uint64
|
|
|
|
var time uint64
|
|
|
|
{
|
|
|
|
childTime := f.readUint64(1, 0xFFFFFFFFFFFFFFFF)
|
|
|
|
//fmt.Printf("childTime: %x\n",childTime)
|
|
|
|
delta := f.readUint64(1, childTime)
|
|
|
|
//fmt.Printf("delta: %v\n", delta)
|
|
|
|
pTime := childTime - delta
|
|
|
|
header.Time = pTime
|
|
|
|
time = childTime
|
|
|
|
}
|
|
|
|
// Bomb delay will never exceed uint64
|
|
|
|
bombDelay := new(big.Int).SetUint64(f.readUint64(1, 0xFFFFFFFFFFFFFFFe))
|
|
|
|
|
|
|
|
if f.exhausted {
|
|
|
|
return 0
|
|
|
|
}
|
|
|
|
|
|
|
|
for i, pair := range []struct {
|
|
|
|
bigFn calculator
|
|
|
|
u256Fn calculator
|
|
|
|
}{
|
2022-04-26 11:16:57 +03:00
|
|
|
{ethash.FrontierDifficultyCalculator, ethash.CalcDifficultyFrontierU256},
|
|
|
|
{ethash.HomesteadDifficultyCalculator, ethash.CalcDifficultyHomesteadU256},
|
2020-12-11 13:06:44 +03:00
|
|
|
{ethash.DynamicDifficultyCalculator(bombDelay), ethash.MakeDifficultyCalculatorU256(bombDelay)},
|
|
|
|
} {
|
|
|
|
want := pair.bigFn(time, header)
|
|
|
|
have := pair.u256Fn(time, header)
|
|
|
|
if want.Cmp(have) != 0 {
|
|
|
|
panic(fmt.Sprintf("pair %d: want %x have %x\nparent.Number: %x\np.Time: %x\nc.Time: %x\nBombdelay: %v\n", i, want, have,
|
|
|
|
header.Number, header.Time, time, bombDelay))
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 1
|
|
|
|
}
|