lbcd/txscript/engine.go
Dave Collins 67e2cbe374 txscript: Optimize IsMultisigScript.
This converts the IsMultisigScript function to make use of the new
tokenizer instead of the far less efficient parseScript thereby
significantly optimizing the function.

In order to accomplish this, it introduces two new functions.  The first
one is named extractMultisigScriptDetails and works with the raw script
bytes to simultaneously determine if the script is a multisignature
script, and in the case it is, extract and return the relevant details.
The second new function is named isMultisigScript and is defined in
terms of the former.

The extract function accepts the script version, raw script bytes, and a
flag to determine whether or not the public keys should also be
extracted.  The flag is provided because extracting pubkeys results in
an allocation that the caller might wish to avoid.

The extract function approach was chosen because it is common for
callers to want to only extract relevant details from a script if the
script is of the specific type.  Extracting those details requires
performing the exact same checks to ensure the script is of the correct
type, so it is more efficient to combine the two into one and define the
type determination in terms of the result so long as the extraction does
not require allocations.

It is important to note that this new implementation intentionally has a
semantic difference from the existing implementation in that it will now
correctly identify a multisig script with zero pubkeys whereas
previously it incorrectly required at least one pubkey.  This change is
acceptable because the function only deals with standardness rather than
consensus rules.

Finally, this also deprecates the isMultiSig function that requires
opcodes in favor of the new functions and deprecates the error return on
the export IsMultisigScript function since it really does not make sense
given the purpose of the function.

The following is a before and after comparison of analyzing both a large
script that is not a multisig script and a 1-of-2 multisig public key
script:

benchmark                            old ns/op     new ns/op     delta
BenchmarkIsMultisigScriptLarge-8     64166         5.52          -99.99%
BenchmarkIsMultisigScript-8          630           59.4          -90.57%

benchmark                            old allocs     new allocs     delta
BenchmarkIsMultisigScriptLarge-8     1              0              -100.00%
BenchmarkIsMultisigScript-8          1              0              -100.00%

benchmark                            old bytes     new bytes     delta
BenchmarkIsMultisigScriptLarge-8     311299        0             -100.00%
BenchmarkIsMultisigScript-8          2304          0             -100.00%
2022-05-23 21:46:21 -07:00

1026 lines
34 KiB
Go

// Copyright (c) 2013-2018 The btcsuite developers
// Copyright (c) 2015-2018 The Decred developers
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
package txscript
import (
"bytes"
"crypto/sha256"
"fmt"
"math/big"
"github.com/btcsuite/btcd/btcec"
"github.com/btcsuite/btcd/wire"
)
// ScriptFlags is a bitmask defining additional operations or tests that will be
// done when executing a script pair.
type ScriptFlags uint32
const (
// ScriptBip16 defines whether the bip16 threshold has passed and thus
// pay-to-script hash transactions will be fully validated.
ScriptBip16 ScriptFlags = 1 << iota
// ScriptStrictMultiSig defines whether to verify the stack item
// used by CHECKMULTISIG is zero length.
ScriptStrictMultiSig
// ScriptDiscourageUpgradableNops defines whether to verify that
// NOP1 through NOP10 are reserved for future soft-fork upgrades. This
// flag must not be used for consensus critical code nor applied to
// blocks as this flag is only for stricter standard transaction
// checks. This flag is only applied when the above opcodes are
// executed.
ScriptDiscourageUpgradableNops
// ScriptVerifyCheckLockTimeVerify defines whether to verify that
// a transaction output is spendable based on the locktime.
// This is BIP0065.
ScriptVerifyCheckLockTimeVerify
// ScriptVerifyCheckSequenceVerify defines whether to allow execution
// pathways of a script to be restricted based on the age of the output
// being spent. This is BIP0112.
ScriptVerifyCheckSequenceVerify
// ScriptVerifyCleanStack defines that the stack must contain only
// one stack element after evaluation and that the element must be
// true if interpreted as a boolean. This is rule 6 of BIP0062.
// This flag should never be used without the ScriptBip16 flag nor the
// ScriptVerifyWitness flag.
ScriptVerifyCleanStack
// ScriptVerifyDERSignatures defines that signatures are required
// to compily with the DER format.
ScriptVerifyDERSignatures
// ScriptVerifyLowS defines that signtures are required to comply with
// the DER format and whose S value is <= order / 2. This is rule 5
// of BIP0062.
ScriptVerifyLowS
// ScriptVerifyMinimalData defines that signatures must use the smallest
// push operator. This is both rules 3 and 4 of BIP0062.
ScriptVerifyMinimalData
// ScriptVerifyNullFail defines that signatures must be empty if
// a CHECKSIG or CHECKMULTISIG operation fails.
ScriptVerifyNullFail
// ScriptVerifySigPushOnly defines that signature scripts must contain
// only pushed data. This is rule 2 of BIP0062.
ScriptVerifySigPushOnly
// ScriptVerifyStrictEncoding defines that signature scripts and
// public keys must follow the strict encoding requirements.
ScriptVerifyStrictEncoding
// ScriptVerifyWitness defines whether or not to verify a transaction
// output using a witness program template.
ScriptVerifyWitness
// ScriptVerifyDiscourageUpgradeableWitnessProgram makes witness
// program with versions 2-16 non-standard.
ScriptVerifyDiscourageUpgradeableWitnessProgram
// ScriptVerifyMinimalIf makes a script with an OP_IF/OP_NOTIF whose
// operand is anything other than empty vector or [0x01] non-standard.
ScriptVerifyMinimalIf
// ScriptVerifyWitnessPubKeyType makes a script within a check-sig
// operation whose public key isn't serialized in a compressed format
// non-standard.
ScriptVerifyWitnessPubKeyType
)
const (
// MaxStackSize is the maximum combined height of stack and alt stack
// during execution.
MaxStackSize = 1000
// MaxScriptSize is the maximum allowed length of a raw script.
MaxScriptSize = 10000
// payToWitnessPubKeyHashDataSize is the size of the witness program's
// data push for a pay-to-witness-pub-key-hash output.
payToWitnessPubKeyHashDataSize = 20
// payToWitnessScriptHashDataSize is the size of the witness program's
// data push for a pay-to-witness-script-hash output.
payToWitnessScriptHashDataSize = 32
)
// halforder is used to tame ECDSA malleability (see BIP0062).
var halfOrder = new(big.Int).Rsh(btcec.S256().N, 1)
// Engine is the virtual machine that executes scripts.
type Engine struct {
scripts [][]parsedOpcode
scriptIdx int
scriptOff int
lastCodeSep int
dstack stack // data stack
astack stack // alt stack
tx wire.MsgTx
txIdx int
condStack []int
numOps int
flags ScriptFlags
sigCache *SigCache
hashCache *TxSigHashes
bip16 bool // treat execution as pay-to-script-hash
savedFirstStack [][]byte // stack from first script for bip16 scripts
witnessVersion int
witnessProgram []byte
inputAmount int64
}
// hasFlag returns whether the script engine instance has the passed flag set.
func (vm *Engine) hasFlag(flag ScriptFlags) bool {
return vm.flags&flag == flag
}
// isBranchExecuting returns whether or not the current conditional branch is
// actively executing. For example, when the data stack has an OP_FALSE on it
// and an OP_IF is encountered, the branch is inactive until an OP_ELSE or
// OP_ENDIF is encountered. It properly handles nested conditionals.
func (vm *Engine) isBranchExecuting() bool {
if len(vm.condStack) == 0 {
return true
}
return vm.condStack[len(vm.condStack)-1] == OpCondTrue
}
// executeOpcode peforms execution on the passed opcode. It takes into account
// whether or not it is hidden by conditionals, but some rules still must be
// tested in this case.
func (vm *Engine) executeOpcode(pop *parsedOpcode) error {
// Disabled opcodes are fail on program counter.
if pop.isDisabled() {
str := fmt.Sprintf("attempt to execute disabled opcode %s",
pop.opcode.name)
return scriptError(ErrDisabledOpcode, str)
}
// Always-illegal opcodes are fail on program counter.
if pop.alwaysIllegal() {
str := fmt.Sprintf("attempt to execute reserved opcode %s",
pop.opcode.name)
return scriptError(ErrReservedOpcode, str)
}
// Note that this includes OP_RESERVED which counts as a push operation.
if pop.opcode.value > OP_16 {
vm.numOps++
if vm.numOps > MaxOpsPerScript {
str := fmt.Sprintf("exceeded max operation limit of %d",
MaxOpsPerScript)
return scriptError(ErrTooManyOperations, str)
}
} else if len(pop.data) > MaxScriptElementSize {
str := fmt.Sprintf("element size %d exceeds max allowed size %d",
len(pop.data), MaxScriptElementSize)
return scriptError(ErrElementTooBig, str)
}
// Nothing left to do when this is not a conditional opcode and it is
// not in an executing branch.
if !vm.isBranchExecuting() && !pop.isConditional() {
return nil
}
// Ensure all executed data push opcodes use the minimal encoding when
// the minimal data verification flag is set.
if vm.dstack.verifyMinimalData && vm.isBranchExecuting() &&
pop.opcode.value >= 0 && pop.opcode.value <= OP_PUSHDATA4 {
if err := pop.checkMinimalDataPush(); err != nil {
return err
}
}
return pop.opcode.opfunc(pop, vm)
}
// disasm is a helper function to produce the output for DisasmPC and
// DisasmScript. It produces the opcode prefixed by the program counter at the
// provided position in the script. It does no error checking and leaves that
// to the caller to provide a valid offset.
func (vm *Engine) disasm(scriptIdx int, scriptOff int) string {
return fmt.Sprintf("%02x:%04x: %s", scriptIdx, scriptOff,
vm.scripts[scriptIdx][scriptOff].print(false))
}
// validPC returns an error if the current script position is valid for
// execution, nil otherwise.
func (vm *Engine) validPC() error {
if vm.scriptIdx >= len(vm.scripts) {
str := fmt.Sprintf("past input scripts %v:%v %v:xxxx",
vm.scriptIdx, vm.scriptOff, len(vm.scripts))
return scriptError(ErrInvalidProgramCounter, str)
}
if vm.scriptOff >= len(vm.scripts[vm.scriptIdx]) {
str := fmt.Sprintf("past input scripts %v:%v %v:%04d",
vm.scriptIdx, vm.scriptOff, vm.scriptIdx,
len(vm.scripts[vm.scriptIdx]))
return scriptError(ErrInvalidProgramCounter, str)
}
return nil
}
// curPC returns either the current script and offset, or an error if the
// position isn't valid.
func (vm *Engine) curPC() (script int, off int, err error) {
err = vm.validPC()
if err != nil {
return 0, 0, err
}
return vm.scriptIdx, vm.scriptOff, nil
}
// isWitnessVersionActive returns true if a witness program was extracted
// during the initialization of the Engine, and the program's version matches
// the specified version.
func (vm *Engine) isWitnessVersionActive(version uint) bool {
return vm.witnessProgram != nil && uint(vm.witnessVersion) == version
}
// verifyWitnessProgram validates the stored witness program using the passed
// witness as input.
func (vm *Engine) verifyWitnessProgram(witness [][]byte) error {
if vm.isWitnessVersionActive(0) {
switch len(vm.witnessProgram) {
case payToWitnessPubKeyHashDataSize: // P2WKH
// The witness stack should consist of exactly two
// items: the signature, and the pubkey.
if len(witness) != 2 {
err := fmt.Sprintf("should have exactly two "+
"items in witness, instead have %v", len(witness))
return scriptError(ErrWitnessProgramMismatch, err)
}
// Now we'll resume execution as if it were a regular
// p2pkh transaction.
pkScript, err := payToPubKeyHashScript(vm.witnessProgram)
if err != nil {
return err
}
pops, err := parseScript(pkScript)
if err != nil {
return err
}
// Set the stack to the provided witness stack, then
// append the pkScript generated above as the next
// script to execute.
vm.scripts = append(vm.scripts, pops)
vm.SetStack(witness)
case payToWitnessScriptHashDataSize: // P2WSH
// Additionally, The witness stack MUST NOT be empty at
// this point.
if len(witness) == 0 {
return scriptError(ErrWitnessProgramEmpty, "witness "+
"program empty passed empty witness")
}
// Obtain the witness script which should be the last
// element in the passed stack. The size of the script
// MUST NOT exceed the max script size.
witnessScript := witness[len(witness)-1]
if len(witnessScript) > MaxScriptSize {
str := fmt.Sprintf("witnessScript size %d "+
"is larger than max allowed size %d",
len(witnessScript), MaxScriptSize)
return scriptError(ErrScriptTooBig, str)
}
// Ensure that the serialized pkScript at the end of
// the witness stack matches the witness program.
witnessHash := sha256.Sum256(witnessScript)
if !bytes.Equal(witnessHash[:], vm.witnessProgram) {
return scriptError(ErrWitnessProgramMismatch,
"witness program hash mismatch")
}
// With all the validity checks passed, parse the
// script into individual op-codes so w can execute it
// as the next script.
pops, err := parseScript(witnessScript)
if err != nil {
return err
}
// The hash matched successfully, so use the witness as
// the stack, and set the witnessScript to be the next
// script executed.
vm.scripts = append(vm.scripts, pops)
vm.SetStack(witness[:len(witness)-1])
default:
errStr := fmt.Sprintf("length of witness program "+
"must either be %v or %v bytes, instead is %v bytes",
payToWitnessPubKeyHashDataSize,
payToWitnessScriptHashDataSize,
len(vm.witnessProgram))
return scriptError(ErrWitnessProgramWrongLength, errStr)
}
} else if vm.hasFlag(ScriptVerifyDiscourageUpgradeableWitnessProgram) {
errStr := fmt.Sprintf("new witness program versions "+
"invalid: %v", vm.witnessProgram)
return scriptError(ErrDiscourageUpgradableWitnessProgram, errStr)
} else {
// If we encounter an unknown witness program version and we
// aren't discouraging future unknown witness based soft-forks,
// then we de-activate the segwit behavior within the VM for
// the remainder of execution.
vm.witnessProgram = nil
}
if vm.isWitnessVersionActive(0) {
// All elements within the witness stack must not be greater
// than the maximum bytes which are allowed to be pushed onto
// the stack.
for _, witElement := range vm.GetStack() {
if len(witElement) > MaxScriptElementSize {
str := fmt.Sprintf("element size %d exceeds "+
"max allowed size %d", len(witElement),
MaxScriptElementSize)
return scriptError(ErrElementTooBig, str)
}
}
}
return nil
}
// DisasmPC returns the string for the disassembly of the opcode that will be
// next to execute when Step() is called.
func (vm *Engine) DisasmPC() (string, error) {
scriptIdx, scriptOff, err := vm.curPC()
if err != nil {
return "", err
}
return vm.disasm(scriptIdx, scriptOff), nil
}
// DisasmScript returns the disassembly string for the script at the requested
// offset index. Index 0 is the signature script and 1 is the public key
// script.
func (vm *Engine) DisasmScript(idx int) (string, error) {
if idx >= len(vm.scripts) {
str := fmt.Sprintf("script index %d >= total scripts %d", idx,
len(vm.scripts))
return "", scriptError(ErrInvalidIndex, str)
}
var disstr string
for i := range vm.scripts[idx] {
disstr = disstr + vm.disasm(idx, i) + "\n"
}
return disstr, nil
}
// CheckErrorCondition returns nil if the running script has ended and was
// successful, leaving a a true boolean on the stack. An error otherwise,
// including if the script has not finished.
func (vm *Engine) CheckErrorCondition(finalScript bool) error {
// Check execution is actually done. When pc is past the end of script
// array there are no more scripts to run.
if vm.scriptIdx < len(vm.scripts) {
return scriptError(ErrScriptUnfinished,
"error check when script unfinished")
}
// If we're in version zero witness execution mode, and this was the
// final script, then the stack MUST be clean in order to maintain
// compatibility with BIP16.
if finalScript && vm.isWitnessVersionActive(0) && vm.dstack.Depth() != 1 {
return scriptError(ErrEvalFalse, "witness program must "+
"have clean stack")
}
if finalScript && vm.hasFlag(ScriptVerifyCleanStack) &&
vm.dstack.Depth() != 1 {
str := fmt.Sprintf("stack contains %d unexpected items",
vm.dstack.Depth()-1)
return scriptError(ErrCleanStack, str)
} else if vm.dstack.Depth() < 1 {
return scriptError(ErrEmptyStack,
"stack empty at end of script execution")
}
v, err := vm.dstack.PopBool()
if err != nil {
return err
}
if !v {
// Log interesting data.
log.Tracef("%v", newLogClosure(func() string {
dis0, _ := vm.DisasmScript(0)
dis1, _ := vm.DisasmScript(1)
return fmt.Sprintf("scripts failed: script0: %s\n"+
"script1: %s", dis0, dis1)
}))
return scriptError(ErrEvalFalse,
"false stack entry at end of script execution")
}
return nil
}
// Step will execute the next instruction and move the program counter to the
// next opcode in the script, or the next script if the current has ended. Step
// will return true in the case that the last opcode was successfully executed.
//
// The result of calling Step or any other method is undefined if an error is
// returned.
func (vm *Engine) Step() (done bool, err error) {
// Verify that it is pointing to a valid script address.
err = vm.validPC()
if err != nil {
return true, err
}
opcode := &vm.scripts[vm.scriptIdx][vm.scriptOff]
vm.scriptOff++
// Execute the opcode while taking into account several things such as
// disabled opcodes, illegal opcodes, maximum allowed operations per
// script, maximum script element sizes, and conditionals.
err = vm.executeOpcode(opcode)
if err != nil {
return true, err
}
// The number of elements in the combination of the data and alt stacks
// must not exceed the maximum number of stack elements allowed.
combinedStackSize := vm.dstack.Depth() + vm.astack.Depth()
if combinedStackSize > MaxStackSize {
str := fmt.Sprintf("combined stack size %d > max allowed %d",
combinedStackSize, MaxStackSize)
return false, scriptError(ErrStackOverflow, str)
}
// Prepare for next instruction.
if vm.scriptOff >= len(vm.scripts[vm.scriptIdx]) {
// Illegal to have an `if' that straddles two scripts.
if err == nil && len(vm.condStack) != 0 {
return false, scriptError(ErrUnbalancedConditional,
"end of script reached in conditional execution")
}
// Alt stack doesn't persist.
_ = vm.astack.DropN(vm.astack.Depth())
vm.numOps = 0 // number of ops is per script.
vm.scriptOff = 0
if vm.scriptIdx == 0 && vm.bip16 {
vm.scriptIdx++
vm.savedFirstStack = vm.GetStack()
} else if vm.scriptIdx == 1 && vm.bip16 {
// Put us past the end for CheckErrorCondition()
vm.scriptIdx++
// Check script ran successfully and pull the script
// out of the first stack and execute that.
err := vm.CheckErrorCondition(false)
if err != nil {
return false, err
}
script := vm.savedFirstStack[len(vm.savedFirstStack)-1]
pops, err := parseScript(script)
if err != nil {
return false, err
}
vm.scripts = append(vm.scripts, pops)
// Set stack to be the stack from first script minus the
// script itself
vm.SetStack(vm.savedFirstStack[:len(vm.savedFirstStack)-1])
} else if (vm.scriptIdx == 1 && vm.witnessProgram != nil) ||
(vm.scriptIdx == 2 && vm.witnessProgram != nil && vm.bip16) { // Nested P2SH.
vm.scriptIdx++
witness := vm.tx.TxIn[vm.txIdx].Witness
if err := vm.verifyWitnessProgram(witness); err != nil {
return false, err
}
} else {
vm.scriptIdx++
}
// there are zero length scripts in the wild
if vm.scriptIdx < len(vm.scripts) && vm.scriptOff >= len(vm.scripts[vm.scriptIdx]) {
vm.scriptIdx++
}
vm.lastCodeSep = 0
if vm.scriptIdx >= len(vm.scripts) {
return true, nil
}
}
return false, nil
}
// Execute will execute all scripts in the script engine and return either nil
// for successful validation or an error if one occurred.
func (vm *Engine) Execute() (err error) {
done := false
for !done {
log.Tracef("%v", newLogClosure(func() string {
dis, err := vm.DisasmPC()
if err != nil {
return fmt.Sprintf("stepping (%v)", err)
}
return fmt.Sprintf("stepping %v", dis)
}))
done, err = vm.Step()
if err != nil {
return err
}
log.Tracef("%v", newLogClosure(func() string {
var dstr, astr string
// if we're tracing, dump the stacks.
if vm.dstack.Depth() != 0 {
dstr = "Stack:\n" + vm.dstack.String()
}
if vm.astack.Depth() != 0 {
astr = "AltStack:\n" + vm.astack.String()
}
return dstr + astr
}))
}
return vm.CheckErrorCondition(true)
}
// subScript returns the script since the last OP_CODESEPARATOR.
func (vm *Engine) subScript() []parsedOpcode {
return vm.scripts[vm.scriptIdx][vm.lastCodeSep:]
}
// checkHashTypeEncoding returns whether or not the passed hashtype adheres to
// the strict encoding requirements if enabled.
func (vm *Engine) checkHashTypeEncoding(hashType SigHashType) error {
if !vm.hasFlag(ScriptVerifyStrictEncoding) {
return nil
}
sigHashType := hashType & ^SigHashAnyOneCanPay
if sigHashType < SigHashAll || sigHashType > SigHashSingle {
str := fmt.Sprintf("invalid hash type 0x%x", hashType)
return scriptError(ErrInvalidSigHashType, str)
}
return nil
}
// isStrictPubKeyEncoding returns whether or not the passed public key adheres
// to the strict encoding requirements.
func isStrictPubKeyEncoding(pubKey []byte) bool {
if len(pubKey) == 33 && (pubKey[0] == 0x02 || pubKey[0] == 0x03) {
// Compressed
return true
}
if len(pubKey) == 65 {
switch pubKey[0] {
case 0x04:
// Uncompressed
return true
case 0x06, 0x07:
// Hybrid
return true
}
}
return false
}
// checkPubKeyEncoding returns whether or not the passed public key adheres to
// the strict encoding requirements if enabled.
func (vm *Engine) checkPubKeyEncoding(pubKey []byte) error {
if vm.hasFlag(ScriptVerifyWitnessPubKeyType) &&
vm.isWitnessVersionActive(0) && !btcec.IsCompressedPubKey(pubKey) {
str := "only compressed keys are accepted post-segwit"
return scriptError(ErrWitnessPubKeyType, str)
}
if !vm.hasFlag(ScriptVerifyStrictEncoding) {
return nil
}
if len(pubKey) == 33 && (pubKey[0] == 0x02 || pubKey[0] == 0x03) {
// Compressed
return nil
}
if len(pubKey) == 65 && pubKey[0] == 0x04 {
// Uncompressed
return nil
}
return scriptError(ErrPubKeyType, "unsupported public key type")
}
// checkSignatureEncoding returns whether or not the passed signature adheres to
// the strict encoding requirements if enabled.
func (vm *Engine) checkSignatureEncoding(sig []byte) error {
if !vm.hasFlag(ScriptVerifyDERSignatures) &&
!vm.hasFlag(ScriptVerifyLowS) &&
!vm.hasFlag(ScriptVerifyStrictEncoding) {
return nil
}
// The format of a DER encoded signature is as follows:
//
// 0x30 <total length> 0x02 <length of R> <R> 0x02 <length of S> <S>
// - 0x30 is the ASN.1 identifier for a sequence
// - Total length is 1 byte and specifies length of all remaining data
// - 0x02 is the ASN.1 identifier that specifies an integer follows
// - Length of R is 1 byte and specifies how many bytes R occupies
// - R is the arbitrary length big-endian encoded number which
// represents the R value of the signature. DER encoding dictates
// that the value must be encoded using the minimum possible number
// of bytes. This implies the first byte can only be null if the
// highest bit of the next byte is set in order to prevent it from
// being interpreted as a negative number.
// - 0x02 is once again the ASN.1 integer identifier
// - Length of S is 1 byte and specifies how many bytes S occupies
// - S is the arbitrary length big-endian encoded number which
// represents the S value of the signature. The encoding rules are
// identical as those for R.
const (
asn1SequenceID = 0x30
asn1IntegerID = 0x02
// minSigLen is the minimum length of a DER encoded signature and is
// when both R and S are 1 byte each.
//
// 0x30 + <1-byte> + 0x02 + 0x01 + <byte> + 0x2 + 0x01 + <byte>
minSigLen = 8
// maxSigLen is the maximum length of a DER encoded signature and is
// when both R and S are 33 bytes each. It is 33 bytes because a
// 256-bit integer requires 32 bytes and an additional leading null byte
// might required if the high bit is set in the value.
//
// 0x30 + <1-byte> + 0x02 + 0x21 + <33 bytes> + 0x2 + 0x21 + <33 bytes>
maxSigLen = 72
// sequenceOffset is the byte offset within the signature of the
// expected ASN.1 sequence identifier.
sequenceOffset = 0
// dataLenOffset is the byte offset within the signature of the expected
// total length of all remaining data in the signature.
dataLenOffset = 1
// rTypeOffset is the byte offset within the signature of the ASN.1
// identifier for R and is expected to indicate an ASN.1 integer.
rTypeOffset = 2
// rLenOffset is the byte offset within the signature of the length of
// R.
rLenOffset = 3
// rOffset is the byte offset within the signature of R.
rOffset = 4
)
// The signature must adhere to the minimum and maximum allowed length.
sigLen := len(sig)
if sigLen < minSigLen {
str := fmt.Sprintf("malformed signature: too short: %d < %d", sigLen,
minSigLen)
return scriptError(ErrSigTooShort, str)
}
if sigLen > maxSigLen {
str := fmt.Sprintf("malformed signature: too long: %d > %d", sigLen,
maxSigLen)
return scriptError(ErrSigTooLong, str)
}
// The signature must start with the ASN.1 sequence identifier.
if sig[sequenceOffset] != asn1SequenceID {
str := fmt.Sprintf("malformed signature: format has wrong type: %#x",
sig[sequenceOffset])
return scriptError(ErrSigInvalidSeqID, str)
}
// The signature must indicate the correct amount of data for all elements
// related to R and S.
if int(sig[dataLenOffset]) != sigLen-2 {
str := fmt.Sprintf("malformed signature: bad length: %d != %d",
sig[dataLenOffset], sigLen-2)
return scriptError(ErrSigInvalidDataLen, str)
}
// Calculate the offsets of the elements related to S and ensure S is inside
// the signature.
//
// rLen specifies the length of the big-endian encoded number which
// represents the R value of the signature.
//
// sTypeOffset is the offset of the ASN.1 identifier for S and, like its R
// counterpart, is expected to indicate an ASN.1 integer.
//
// sLenOffset and sOffset are the byte offsets within the signature of the
// length of S and S itself, respectively.
rLen := int(sig[rLenOffset])
sTypeOffset := rOffset + rLen
sLenOffset := sTypeOffset + 1
if sTypeOffset >= sigLen {
str := "malformed signature: S type indicator missing"
return scriptError(ErrSigMissingSTypeID, str)
}
if sLenOffset >= sigLen {
str := "malformed signature: S length missing"
return scriptError(ErrSigMissingSLen, str)
}
// The lengths of R and S must match the overall length of the signature.
//
// sLen specifies the length of the big-endian encoded number which
// represents the S value of the signature.
sOffset := sLenOffset + 1
sLen := int(sig[sLenOffset])
if sOffset+sLen != sigLen {
str := "malformed signature: invalid S length"
return scriptError(ErrSigInvalidSLen, str)
}
// R elements must be ASN.1 integers.
if sig[rTypeOffset] != asn1IntegerID {
str := fmt.Sprintf("malformed signature: R integer marker: %#x != %#x",
sig[rTypeOffset], asn1IntegerID)
return scriptError(ErrSigInvalidRIntID, str)
}
// Zero-length integers are not allowed for R.
if rLen == 0 {
str := "malformed signature: R length is zero"
return scriptError(ErrSigZeroRLen, str)
}
// R must not be negative.
if sig[rOffset]&0x80 != 0 {
str := "malformed signature: R is negative"
return scriptError(ErrSigNegativeR, str)
}
// Null bytes at the start of R are not allowed, unless R would otherwise be
// interpreted as a negative number.
if rLen > 1 && sig[rOffset] == 0x00 && sig[rOffset+1]&0x80 == 0 {
str := "malformed signature: R value has too much padding"
return scriptError(ErrSigTooMuchRPadding, str)
}
// S elements must be ASN.1 integers.
if sig[sTypeOffset] != asn1IntegerID {
str := fmt.Sprintf("malformed signature: S integer marker: %#x != %#x",
sig[sTypeOffset], asn1IntegerID)
return scriptError(ErrSigInvalidSIntID, str)
}
// Zero-length integers are not allowed for S.
if sLen == 0 {
str := "malformed signature: S length is zero"
return scriptError(ErrSigZeroSLen, str)
}
// S must not be negative.
if sig[sOffset]&0x80 != 0 {
str := "malformed signature: S is negative"
return scriptError(ErrSigNegativeS, str)
}
// Null bytes at the start of S are not allowed, unless S would otherwise be
// interpreted as a negative number.
if sLen > 1 && sig[sOffset] == 0x00 && sig[sOffset+1]&0x80 == 0 {
str := "malformed signature: S value has too much padding"
return scriptError(ErrSigTooMuchSPadding, str)
}
// Verify the S value is <= half the order of the curve. This check is done
// because when it is higher, the complement modulo the order can be used
// instead which is a shorter encoding by 1 byte. Further, without
// enforcing this, it is possible to replace a signature in a valid
// transaction with the complement while still being a valid signature that
// verifies. This would result in changing the transaction hash and thus is
// a source of malleability.
if vm.hasFlag(ScriptVerifyLowS) {
sValue := new(big.Int).SetBytes(sig[sOffset : sOffset+sLen])
if sValue.Cmp(halfOrder) > 0 {
return scriptError(ErrSigHighS, "signature is not canonical due "+
"to unnecessarily high S value")
}
}
return nil
}
// getStack returns the contents of stack as a byte array bottom up
func getStack(stack *stack) [][]byte {
array := make([][]byte, stack.Depth())
for i := range array {
// PeekByteArry can't fail due to overflow, already checked
array[len(array)-i-1], _ = stack.PeekByteArray(int32(i))
}
return array
}
// setStack sets the stack to the contents of the array where the last item in
// the array is the top item in the stack.
func setStack(stack *stack, data [][]byte) {
// This can not error. Only errors are for invalid arguments.
_ = stack.DropN(stack.Depth())
for i := range data {
stack.PushByteArray(data[i])
}
}
// GetStack returns the contents of the primary stack as an array. where the
// last item in the array is the top of the stack.
func (vm *Engine) GetStack() [][]byte {
return getStack(&vm.dstack)
}
// SetStack sets the contents of the primary stack to the contents of the
// provided array where the last item in the array will be the top of the stack.
func (vm *Engine) SetStack(data [][]byte) {
setStack(&vm.dstack, data)
}
// GetAltStack returns the contents of the alternate stack as an array where the
// last item in the array is the top of the stack.
func (vm *Engine) GetAltStack() [][]byte {
return getStack(&vm.astack)
}
// SetAltStack sets the contents of the alternate stack to the contents of the
// provided array where the last item in the array will be the top of the stack.
func (vm *Engine) SetAltStack(data [][]byte) {
setStack(&vm.astack, data)
}
// NewEngine returns a new script engine for the provided public key script,
// transaction, and input index. The flags modify the behavior of the script
// engine according to the description provided by each flag.
func NewEngine(scriptPubKey []byte, tx *wire.MsgTx, txIdx int, flags ScriptFlags,
sigCache *SigCache, hashCache *TxSigHashes, inputAmount int64) (*Engine, error) {
// The provided transaction input index must refer to a valid input.
if txIdx < 0 || txIdx >= len(tx.TxIn) {
str := fmt.Sprintf("transaction input index %d is negative or "+
">= %d", txIdx, len(tx.TxIn))
return nil, scriptError(ErrInvalidIndex, str)
}
scriptSig := tx.TxIn[txIdx].SignatureScript
// When both the signature script and public key script are empty the
// result is necessarily an error since the stack would end up being
// empty which is equivalent to a false top element. Thus, just return
// the relevant error now as an optimization.
if len(scriptSig) == 0 && len(scriptPubKey) == 0 {
return nil, scriptError(ErrEvalFalse,
"false stack entry at end of script execution")
}
// The clean stack flag (ScriptVerifyCleanStack) is not allowed without
// either the pay-to-script-hash (P2SH) evaluation (ScriptBip16)
// flag or the Segregated Witness (ScriptVerifyWitness) flag.
//
// Recall that evaluating a P2SH script without the flag set results in
// non-P2SH evaluation which leaves the P2SH inputs on the stack.
// Thus, allowing the clean stack flag without the P2SH flag would make
// it possible to have a situation where P2SH would not be a soft fork
// when it should be. The same goes for segwit which will pull in
// additional scripts for execution from the witness stack.
vm := Engine{flags: flags, sigCache: sigCache, hashCache: hashCache,
inputAmount: inputAmount}
if vm.hasFlag(ScriptVerifyCleanStack) && (!vm.hasFlag(ScriptBip16) &&
!vm.hasFlag(ScriptVerifyWitness)) {
return nil, scriptError(ErrInvalidFlags,
"invalid flags combination")
}
// The signature script must only contain data pushes when the
// associated flag is set.
if vm.hasFlag(ScriptVerifySigPushOnly) && !IsPushOnlyScript(scriptSig) {
return nil, scriptError(ErrNotPushOnly,
"signature script is not push only")
}
// The engine stores the scripts in parsed form using a slice. This
// allows multiple scripts to be executed in sequence. For example,
// with a pay-to-script-hash transaction, there will be ultimately be
// a third script to execute.
scripts := [][]byte{scriptSig, scriptPubKey}
vm.scripts = make([][]parsedOpcode, len(scripts))
for i, scr := range scripts {
if len(scr) > MaxScriptSize {
str := fmt.Sprintf("script size %d is larger than max "+
"allowed size %d", len(scr), MaxScriptSize)
return nil, scriptError(ErrScriptTooBig, str)
}
var err error
vm.scripts[i], err = parseScript(scr)
if err != nil {
return nil, err
}
}
// Advance the program counter to the public key script if the signature
// script is empty since there is nothing to execute for it in that
// case.
if len(scripts[0]) == 0 {
vm.scriptIdx++
}
if vm.hasFlag(ScriptBip16) && isScriptHash(vm.scripts[1]) {
// Only accept input scripts that push data for P2SH.
if !isPushOnly(vm.scripts[0]) {
return nil, scriptError(ErrNotPushOnly,
"pay to script hash is not push only")
}
vm.bip16 = true
}
if vm.hasFlag(ScriptVerifyMinimalData) {
vm.dstack.verifyMinimalData = true
vm.astack.verifyMinimalData = true
}
// Check to see if we should execute in witness verification mode
// according to the set flags. We check both the pkScript, and sigScript
// here since in the case of nested p2sh, the scriptSig will be a valid
// witness program. For nested p2sh, all the bytes after the first data
// push should *exactly* match the witness program template.
if vm.hasFlag(ScriptVerifyWitness) {
// If witness evaluation is enabled, then P2SH MUST also be
// active.
if !vm.hasFlag(ScriptBip16) {
errStr := "P2SH must be enabled to do witness verification"
return nil, scriptError(ErrInvalidFlags, errStr)
}
var witProgram []byte
switch {
case isWitnessProgram(vm.scripts[1]):
// The scriptSig must be *empty* for all native witness
// programs, otherwise we introduce malleability.
if len(scriptSig) != 0 {
errStr := "native witness program cannot " +
"also have a signature script"
return nil, scriptError(ErrWitnessMalleated, errStr)
}
witProgram = scriptPubKey
case len(tx.TxIn[txIdx].Witness) != 0 && vm.bip16:
// The sigScript MUST be *exactly* a single canonical
// data push of the witness program, otherwise we
// reintroduce malleability.
sigPops := vm.scripts[0]
if len(sigPops) == 1 && canonicalPush(sigPops[0]) &&
IsWitnessProgram(sigPops[0].data) {
witProgram = sigPops[0].data
} else {
errStr := "signature script for witness " +
"nested p2sh is not canonical"
return nil, scriptError(ErrWitnessMalleatedP2SH, errStr)
}
}
if witProgram != nil {
var err error
vm.witnessVersion, vm.witnessProgram, err = ExtractWitnessProgramInfo(witProgram)
if err != nil {
return nil, err
}
} else {
// If we didn't find a witness program in either the
// pkScript or as a datapush within the sigScript, then
// there MUST NOT be any witness data associated with
// the input being validated.
if vm.witnessProgram == nil && len(tx.TxIn[txIdx].Witness) != 0 {
errStr := "non-witness inputs cannot have a witness"
return nil, scriptError(ErrWitnessUnexpected, errStr)
}
}
}
vm.tx = *tx
vm.txIdx = txIdx
return &vm, nil
}