lbcd/wire/msgtx.go
Dave Collins a59ac5b18f
multi: Rework utxoset/view to use outpoints.
This modifies the utxoset in the database and related UtxoViewpoint to
store and work with unspent transaction outputs on a per-output basis
instead of at a transaction level.  This was inspired by similar recent
changes in Bitcoin Core.

The primary motivation is to simplify the code, pave the way for a
utxo cache, and generally focus on optimizing runtime performance.

The tradeoff is that this approach does somewhat increase the size of
the serialized utxoset since it means that the transaction hash is
duplicated for each output as a part of the key and some additional
details such as whether the containing transaction is a coinbase and the
block height it was a part of are duplicated in each output.

However, in practice, the size difference isn't all that large, disk
space is relatively cheap, certainly cheaper than memory, and it is much
more important to provide more efficient runtime operation since that is
the ultimate purpose of the daemon.

While performing this conversion, it also simplifies the code to remove
the transaction version information from the utxoset as well as the
spend journal.  The logic for only serializing it under certain
circumstances is complicated and it isn't actually used anywhere aside
from the gettxout RPC where it also isn't used by anything important
either.  Consequently, this also removes the version field of the
gettxout RPC result.

The utxos in the database are automatically migrated to the new format
with this commit and it is possible to interrupt and resume the
migration process.

Finally, it also updates the tests for the new format and adds a new
function to the tests to convert the old test data to the new format for
convenience.  The data has already been converted and updated in the
commit.

An overview of the changes are as follows:

- Remove transaction version from both spent and unspent output entries
  - Update utxo serialization format to exclude the version
  - Modify the spend journal serialization format
    - The old version field is now reserved and always stores zero and
      ignores it when reading
    - This allows old entries to be used by new code without having to
      migrate the entire spend journal
  - Remove version field from gettxout RPC result
- Convert UtxoEntry to represent a specific utxo instead of a
  transaction with all remaining utxos
  - Optimize for memory usage with an eye towards a utxo cache
    - Combine details such as whether the txout was contained in a
      coinbase, is spent, and is modified into a single packed field of
      bit flags
    - Align entry fields to eliminate extra padding since ultimately
      there will be a lot of these in memory
    - Introduce a free list for serializing an outpoint to the database
      key format to significantly reduce pressure on the GC
  - Update all related functions that previously dealt with transaction
    hashes to accept outpoints instead
  - Update all callers accordingly
  - Only add individually requested outputs from the mempool when
    constructing a mempool view
- Modify the spend journal to always store the block height and coinbase
  information with every spent txout
  - Introduce code to handle fetching the missing information from
    another utxo from the same transaction in the event an old style
    entry is encountered
    - Make use of a database cursor with seek to do this much more
      efficiently than testing every possible output
- Always decompress data loaded from the database now that a utxo entry
  only consists of a specific output
- Introduce upgrade code to migrate the utxo set to the new format
  - Store versions of the utxoset and spend journal buckets
  - Allow migration process to be interrupted and resumed
- Update all tests to expect the correct encodings, remove tests that no
  longer apply, and add new ones for the new expected behavior
  - Convert old tests for the legacy utxo format deserialization code to
    test the new function that is used during upgrade
  - Update the utxostore test data and add function that was used to
    convert it
- Introduce a few new functions on UtxoViewpoint
  - AddTxOut for adding an individual txout versus all of them
  - addTxOut to handle the common code between the new AddTxOut and
    existing AddTxOuts
  - RemoveEntry for removing an individual txout
  - fetchEntryByHash for fetching any remaining utxo for a given
    transaction hash
2018-05-27 03:07:41 -05:00

1027 lines
34 KiB
Go

// Copyright (c) 2013-2016 The btcsuite developers
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
package wire
import (
"bytes"
"fmt"
"io"
"strconv"
"github.com/btcsuite/btcd/chaincfg/chainhash"
)
const (
// TxVersion is the current latest supported transaction version.
TxVersion = 1
// MaxTxInSequenceNum is the maximum sequence number the sequence field
// of a transaction input can be.
MaxTxInSequenceNum uint32 = 0xffffffff
// MaxPrevOutIndex is the maximum index the index field of a previous
// outpoint can be.
MaxPrevOutIndex uint32 = 0xffffffff
// SequenceLockTimeDisabled is a flag that if set on a transaction
// input's sequence number, the sequence number will not be interpreted
// as a relative locktime.
SequenceLockTimeDisabled = 1 << 31
// SequenceLockTimeIsSeconds is a flag that if set on a transaction
// input's sequence number, the relative locktime has units of 512
// seconds.
SequenceLockTimeIsSeconds = 1 << 22
// SequenceLockTimeMask is a mask that extracts the relative locktime
// when masked against the transaction input sequence number.
SequenceLockTimeMask = 0x0000ffff
// SequenceLockTimeGranularity is the defined time based granularity
// for seconds-based relative time locks. When converting from seconds
// to a sequence number, the value is right shifted by this amount,
// therefore the granularity of relative time locks in 512 or 2^9
// seconds. Enforced relative lock times are multiples of 512 seconds.
SequenceLockTimeGranularity = 9
// defaultTxInOutAlloc is the default size used for the backing array for
// transaction inputs and outputs. The array will dynamically grow as needed,
// but this figure is intended to provide enough space for the number of
// inputs and outputs in a typical transaction without needing to grow the
// backing array multiple times.
defaultTxInOutAlloc = 15
// minTxInPayload is the minimum payload size for a transaction input.
// PreviousOutPoint.Hash + PreviousOutPoint.Index 4 bytes + Varint for
// SignatureScript length 1 byte + Sequence 4 bytes.
minTxInPayload = 9 + chainhash.HashSize
// maxTxInPerMessage is the maximum number of transactions inputs that
// a transaction which fits into a message could possibly have.
maxTxInPerMessage = (MaxMessagePayload / minTxInPayload) + 1
// MinTxOutPayload is the minimum payload size for a transaction output.
// Value 8 bytes + Varint for PkScript length 1 byte.
MinTxOutPayload = 9
// maxTxOutPerMessage is the maximum number of transactions outputs that
// a transaction which fits into a message could possibly have.
maxTxOutPerMessage = (MaxMessagePayload / MinTxOutPayload) + 1
// minTxPayload is the minimum payload size for a transaction. Note
// that any realistically usable transaction must have at least one
// input or output, but that is a rule enforced at a higher layer, so
// it is intentionally not included here.
// Version 4 bytes + Varint number of transaction inputs 1 byte + Varint
// number of transaction outputs 1 byte + LockTime 4 bytes + min input
// payload + min output payload.
minTxPayload = 10
// freeListMaxScriptSize is the size of each buffer in the free list
// that is used for deserializing scripts from the wire before they are
// concatenated into a single contiguous buffers. This value was chosen
// because it is slightly more than twice the size of the vast majority
// of all "standard" scripts. Larger scripts are still deserialized
// properly as the free list will simply be bypassed for them.
freeListMaxScriptSize = 512
// freeListMaxItems is the number of buffers to keep in the free list
// to use for script deserialization. This value allows up to 100
// scripts per transaction being simultaneously deserialized by 125
// peers. Thus, the peak usage of the free list is 12,500 * 512 =
// 6,400,000 bytes.
freeListMaxItems = 12500
// maxWitnessItemsPerInput is the maximum number of witness items to
// be read for the witness data for a single TxIn. This number is
// derived using a possble lower bound for the encoding of a witness
// item: 1 byte for length + 1 byte for the witness item itself, or two
// bytes. This value is then divided by the currently allowed maximum
// "cost" for a transaction.
maxWitnessItemsPerInput = 500000
// maxWitnessItemSize is the maximum allowed size for an item within
// an input's witness data. This number is derived from the fact that
// for script validation, each pushed item onto the stack must be less
// than 10k bytes.
maxWitnessItemSize = 11000
)
// witnessMarkerBytes are a pair of bytes specific to the witness encoding. If
// this sequence is encoutered, then it indicates a transaction has iwtness
// data. The first byte is an always 0x00 marker byte, which allows decoders to
// distinguish a serialized transaction with witnesses from a regular (legacy)
// one. The second byte is the Flag field, which at the moment is always 0x01,
// but may be extended in the future to accommodate auxiliary non-committed
// fields.
var witessMarkerBytes = []byte{0x00, 0x01}
// scriptFreeList defines a free list of byte slices (up to the maximum number
// defined by the freeListMaxItems constant) that have a cap according to the
// freeListMaxScriptSize constant. It is used to provide temporary buffers for
// deserializing scripts in order to greatly reduce the number of allocations
// required.
//
// The caller can obtain a buffer from the free list by calling the Borrow
// function and should return it via the Return function when done using it.
type scriptFreeList chan []byte
// Borrow returns a byte slice from the free list with a length according the
// provided size. A new buffer is allocated if there are any items available.
//
// When the size is larger than the max size allowed for items on the free list
// a new buffer of the appropriate size is allocated and returned. It is safe
// to attempt to return said buffer via the Return function as it will be
// ignored and allowed to go the garbage collector.
func (c scriptFreeList) Borrow(size uint64) []byte {
if size > freeListMaxScriptSize {
return make([]byte, size)
}
var buf []byte
select {
case buf = <-c:
default:
buf = make([]byte, freeListMaxScriptSize)
}
return buf[:size]
}
// Return puts the provided byte slice back on the free list when it has a cap
// of the expected length. The buffer is expected to have been obtained via
// the Borrow function. Any slices that are not of the appropriate size, such
// as those whose size is greater than the largest allowed free list item size
// are simply ignored so they can go to the garbage collector.
func (c scriptFreeList) Return(buf []byte) {
// Ignore any buffers returned that aren't the expected size for the
// free list.
if cap(buf) != freeListMaxScriptSize {
return
}
// Return the buffer to the free list when it's not full. Otherwise let
// it be garbage collected.
select {
case c <- buf:
default:
// Let it go to the garbage collector.
}
}
// Create the concurrent safe free list to use for script deserialization. As
// previously described, this free list is maintained to significantly reduce
// the number of allocations.
var scriptPool scriptFreeList = make(chan []byte, freeListMaxItems)
// OutPoint defines a bitcoin data type that is used to track previous
// transaction outputs.
type OutPoint struct {
Hash chainhash.Hash
Index uint32
}
// NewOutPoint returns a new bitcoin transaction outpoint point with the
// provided hash and index.
func NewOutPoint(hash *chainhash.Hash, index uint32) *OutPoint {
return &OutPoint{
Hash: *hash,
Index: index,
}
}
// String returns the OutPoint in the human-readable form "hash:index".
func (o OutPoint) String() string {
// Allocate enough for hash string, colon, and 10 digits. Although
// at the time of writing, the number of digits can be no greater than
// the length of the decimal representation of maxTxOutPerMessage, the
// maximum message payload may increase in the future and this
// optimization may go unnoticed, so allocate space for 10 decimal
// digits, which will fit any uint32.
buf := make([]byte, 2*chainhash.HashSize+1, 2*chainhash.HashSize+1+10)
copy(buf, o.Hash.String())
buf[2*chainhash.HashSize] = ':'
buf = strconv.AppendUint(buf, uint64(o.Index), 10)
return string(buf)
}
// TxIn defines a bitcoin transaction input.
type TxIn struct {
PreviousOutPoint OutPoint
SignatureScript []byte
Witness TxWitness
Sequence uint32
}
// SerializeSize returns the number of bytes it would take to serialize the
// the transaction input.
func (t *TxIn) SerializeSize() int {
// Outpoint Hash 32 bytes + Outpoint Index 4 bytes + Sequence 4 bytes +
// serialized varint size for the length of SignatureScript +
// SignatureScript bytes.
return 40 + VarIntSerializeSize(uint64(len(t.SignatureScript))) +
len(t.SignatureScript)
}
// NewTxIn returns a new bitcoin transaction input with the provided
// previous outpoint point and signature script with a default sequence of
// MaxTxInSequenceNum.
func NewTxIn(prevOut *OutPoint, signatureScript []byte, witness [][]byte) *TxIn {
return &TxIn{
PreviousOutPoint: *prevOut,
SignatureScript: signatureScript,
Witness: witness,
Sequence: MaxTxInSequenceNum,
}
}
// TxWitness defines the witness for a TxIn. A witness is to be interpreted as
// a slice of byte slices, or a stack with one or many elements.
type TxWitness [][]byte
// SerializeSize returns the number of bytes it would take to serialize the the
// transaction input's witness.
func (t TxWitness) SerializeSize() int {
// A varint to signal the number of elements the witness has.
n := VarIntSerializeSize(uint64(len(t)))
// For each element in the witness, we'll need a varint to signal the
// size of the element, then finally the number of bytes the element
// itself comprises.
for _, witItem := range t {
n += VarIntSerializeSize(uint64(len(witItem)))
n += len(witItem)
}
return n
}
// TxOut defines a bitcoin transaction output.
type TxOut struct {
Value int64
PkScript []byte
}
// SerializeSize returns the number of bytes it would take to serialize the
// the transaction output.
func (t *TxOut) SerializeSize() int {
// Value 8 bytes + serialized varint size for the length of PkScript +
// PkScript bytes.
return 8 + VarIntSerializeSize(uint64(len(t.PkScript))) + len(t.PkScript)
}
// NewTxOut returns a new bitcoin transaction output with the provided
// transaction value and public key script.
func NewTxOut(value int64, pkScript []byte) *TxOut {
return &TxOut{
Value: value,
PkScript: pkScript,
}
}
// MsgTx implements the Message interface and represents a bitcoin tx message.
// It is used to deliver transaction information in response to a getdata
// message (MsgGetData) for a given transaction.
//
// Use the AddTxIn and AddTxOut functions to build up the list of transaction
// inputs and outputs.
type MsgTx struct {
Version int32
TxIn []*TxIn
TxOut []*TxOut
LockTime uint32
}
// AddTxIn adds a transaction input to the message.
func (msg *MsgTx) AddTxIn(ti *TxIn) {
msg.TxIn = append(msg.TxIn, ti)
}
// AddTxOut adds a transaction output to the message.
func (msg *MsgTx) AddTxOut(to *TxOut) {
msg.TxOut = append(msg.TxOut, to)
}
// TxHash generates the Hash for the transaction.
func (msg *MsgTx) TxHash() chainhash.Hash {
// Encode the transaction and calculate double sha256 on the result.
// Ignore the error returns since the only way the encode could fail
// is being out of memory or due to nil pointers, both of which would
// cause a run-time panic.
buf := bytes.NewBuffer(make([]byte, 0, msg.SerializeSizeStripped()))
_ = msg.SerializeNoWitness(buf)
return chainhash.DoubleHashH(buf.Bytes())
}
// WitnessHash generates the hash of the transaction serialized according to
// the new witness serialization defined in BIP0141 and BIP0144. The final
// output is used within the Segregated Witness commitment of all the witnesses
// within a block. If a transaction has no witness data, then the witness hash,
// is the same as its txid.
func (msg *MsgTx) WitnessHash() chainhash.Hash {
if msg.HasWitness() {
buf := bytes.NewBuffer(make([]byte, 0, msg.SerializeSize()))
_ = msg.Serialize(buf)
return chainhash.DoubleHashH(buf.Bytes())
}
return msg.TxHash()
}
// Copy creates a deep copy of a transaction so that the original does not get
// modified when the copy is manipulated.
func (msg *MsgTx) Copy() *MsgTx {
// Create new tx and start by copying primitive values and making space
// for the transaction inputs and outputs.
newTx := MsgTx{
Version: msg.Version,
TxIn: make([]*TxIn, 0, len(msg.TxIn)),
TxOut: make([]*TxOut, 0, len(msg.TxOut)),
LockTime: msg.LockTime,
}
// Deep copy the old TxIn data.
for _, oldTxIn := range msg.TxIn {
// Deep copy the old previous outpoint.
oldOutPoint := oldTxIn.PreviousOutPoint
newOutPoint := OutPoint{}
newOutPoint.Hash.SetBytes(oldOutPoint.Hash[:])
newOutPoint.Index = oldOutPoint.Index
// Deep copy the old signature script.
var newScript []byte
oldScript := oldTxIn.SignatureScript
oldScriptLen := len(oldScript)
if oldScriptLen > 0 {
newScript = make([]byte, oldScriptLen)
copy(newScript, oldScript[:oldScriptLen])
}
// Create new txIn with the deep copied data.
newTxIn := TxIn{
PreviousOutPoint: newOutPoint,
SignatureScript: newScript,
Sequence: oldTxIn.Sequence,
}
// If the transaction is witnessy, then also copy the
// witnesses.
if len(oldTxIn.Witness) != 0 {
// Deep copy the old witness data.
newTxIn.Witness = make([][]byte, len(oldTxIn.Witness))
for i, oldItem := range oldTxIn.Witness {
newItem := make([]byte, len(oldItem))
copy(newItem, oldItem)
newTxIn.Witness[i] = newItem
}
}
// Finally, append this fully copied txin.
newTx.TxIn = append(newTx.TxIn, &newTxIn)
}
// Deep copy the old TxOut data.
for _, oldTxOut := range msg.TxOut {
// Deep copy the old PkScript
var newScript []byte
oldScript := oldTxOut.PkScript
oldScriptLen := len(oldScript)
if oldScriptLen > 0 {
newScript = make([]byte, oldScriptLen)
copy(newScript, oldScript[:oldScriptLen])
}
// Create new txOut with the deep copied data and append it to
// new Tx.
newTxOut := TxOut{
Value: oldTxOut.Value,
PkScript: newScript,
}
newTx.TxOut = append(newTx.TxOut, &newTxOut)
}
return &newTx
}
// BtcDecode decodes r using the bitcoin protocol encoding into the receiver.
// This is part of the Message interface implementation.
// See Deserialize for decoding transactions stored to disk, such as in a
// database, as opposed to decoding transactions from the wire.
func (msg *MsgTx) BtcDecode(r io.Reader, pver uint32, enc MessageEncoding) error {
version, err := binarySerializer.Uint32(r, littleEndian)
if err != nil {
return err
}
msg.Version = int32(version)
count, err := ReadVarInt(r, pver)
if err != nil {
return err
}
// A count of zero (meaning no TxIn's to the uninitiated) indicates
// this is a transaction with witness data.
var flag [1]byte
if count == 0 && enc == WitnessEncoding {
// Next, we need to read the flag, which is a single byte.
if _, err = io.ReadFull(r, flag[:]); err != nil {
return err
}
// At the moment, the flag MUST be 0x01. In the future other
// flag types may be supported.
if flag[0] != 0x01 {
str := fmt.Sprintf("witness tx but flag byte is %x", flag)
return messageError("MsgTx.BtcDecode", str)
}
// With the Segregated Witness specific fields decoded, we can
// now read in the actual txin count.
count, err = ReadVarInt(r, pver)
if err != nil {
return err
}
}
// Prevent more input transactions than could possibly fit into a
// message. It would be possible to cause memory exhaustion and panics
// without a sane upper bound on this count.
if count > uint64(maxTxInPerMessage) {
str := fmt.Sprintf("too many input transactions to fit into "+
"max message size [count %d, max %d]", count,
maxTxInPerMessage)
return messageError("MsgTx.BtcDecode", str)
}
// returnScriptBuffers is a closure that returns any script buffers that
// were borrowed from the pool when there are any deserialization
// errors. This is only valid to call before the final step which
// replaces the scripts with the location in a contiguous buffer and
// returns them.
returnScriptBuffers := func() {
for _, txIn := range msg.TxIn {
if txIn == nil {
continue
}
if txIn.SignatureScript != nil {
scriptPool.Return(txIn.SignatureScript)
}
for _, witnessElem := range txIn.Witness {
if witnessElem != nil {
scriptPool.Return(witnessElem)
}
}
}
for _, txOut := range msg.TxOut {
if txOut == nil || txOut.PkScript == nil {
continue
}
scriptPool.Return(txOut.PkScript)
}
}
// Deserialize the inputs.
var totalScriptSize uint64
txIns := make([]TxIn, count)
msg.TxIn = make([]*TxIn, count)
for i := uint64(0); i < count; i++ {
// The pointer is set now in case a script buffer is borrowed
// and needs to be returned to the pool on error.
ti := &txIns[i]
msg.TxIn[i] = ti
err = readTxIn(r, pver, msg.Version, ti)
if err != nil {
returnScriptBuffers()
return err
}
totalScriptSize += uint64(len(ti.SignatureScript))
}
count, err = ReadVarInt(r, pver)
if err != nil {
returnScriptBuffers()
return err
}
// Prevent more output transactions than could possibly fit into a
// message. It would be possible to cause memory exhaustion and panics
// without a sane upper bound on this count.
if count > uint64(maxTxOutPerMessage) {
returnScriptBuffers()
str := fmt.Sprintf("too many output transactions to fit into "+
"max message size [count %d, max %d]", count,
maxTxOutPerMessage)
return messageError("MsgTx.BtcDecode", str)
}
// Deserialize the outputs.
txOuts := make([]TxOut, count)
msg.TxOut = make([]*TxOut, count)
for i := uint64(0); i < count; i++ {
// The pointer is set now in case a script buffer is borrowed
// and needs to be returned to the pool on error.
to := &txOuts[i]
msg.TxOut[i] = to
err = readTxOut(r, pver, msg.Version, to)
if err != nil {
returnScriptBuffers()
return err
}
totalScriptSize += uint64(len(to.PkScript))
}
// If the transaction's flag byte isn't 0x00 at this point, then one or
// more of its inputs has accompanying witness data.
if flag[0] != 0 && enc == WitnessEncoding {
for _, txin := range msg.TxIn {
// For each input, the witness is encoded as a stack
// with one or more items. Therefore, we first read a
// varint which encodes the number of stack items.
witCount, err := ReadVarInt(r, pver)
if err != nil {
returnScriptBuffers()
return err
}
// Prevent a possible memory exhaustion attack by
// limiting the witCount value to a sane upper bound.
if witCount > maxWitnessItemsPerInput {
returnScriptBuffers()
str := fmt.Sprintf("too many witness items to fit "+
"into max message size [count %d, max %d]",
witCount, maxWitnessItemsPerInput)
return messageError("MsgTx.BtcDecode", str)
}
// Then for witCount number of stack items, each item
// has a varint length prefix, followed by the witness
// item itself.
txin.Witness = make([][]byte, witCount)
for j := uint64(0); j < witCount; j++ {
txin.Witness[j], err = readScript(r, pver,
maxWitnessItemSize, "script witness item")
if err != nil {
returnScriptBuffers()
return err
}
totalScriptSize += uint64(len(txin.Witness[j]))
}
}
}
msg.LockTime, err = binarySerializer.Uint32(r, littleEndian)
if err != nil {
returnScriptBuffers()
return err
}
// Create a single allocation to house all of the scripts and set each
// input signature script and output public key script to the
// appropriate subslice of the overall contiguous buffer. Then, return
// each individual script buffer back to the pool so they can be reused
// for future deserializations. This is done because it significantly
// reduces the number of allocations the garbage collector needs to
// track, which in turn improves performance and drastically reduces the
// amount of runtime overhead that would otherwise be needed to keep
// track of millions of small allocations.
//
// NOTE: It is no longer valid to call the returnScriptBuffers closure
// after these blocks of code run because it is already done and the
// scripts in the transaction inputs and outputs no longer point to the
// buffers.
var offset uint64
scripts := make([]byte, totalScriptSize)
for i := 0; i < len(msg.TxIn); i++ {
// Copy the signature script into the contiguous buffer at the
// appropriate offset.
signatureScript := msg.TxIn[i].SignatureScript
copy(scripts[offset:], signatureScript)
// Reset the signature script of the transaction input to the
// slice of the contiguous buffer where the script lives.
scriptSize := uint64(len(signatureScript))
end := offset + scriptSize
msg.TxIn[i].SignatureScript = scripts[offset:end:end]
offset += scriptSize
// Return the temporary script buffer to the pool.
scriptPool.Return(signatureScript)
for j := 0; j < len(msg.TxIn[i].Witness); j++ {
// Copy each item within the witness stack for this
// input into the contiguous buffer at the appropriate
// offset.
witnessElem := msg.TxIn[i].Witness[j]
copy(scripts[offset:], witnessElem)
// Reset the witness item within the stack to the slice
// of the contiguous buffer where the witness lives.
witnessElemSize := uint64(len(witnessElem))
end := offset + witnessElemSize
msg.TxIn[i].Witness[j] = scripts[offset:end:end]
offset += witnessElemSize
// Return the temporary buffer used for the witness stack
// item to the pool.
scriptPool.Return(witnessElem)
}
}
for i := 0; i < len(msg.TxOut); i++ {
// Copy the public key script into the contiguous buffer at the
// appropriate offset.
pkScript := msg.TxOut[i].PkScript
copy(scripts[offset:], pkScript)
// Reset the public key script of the transaction output to the
// slice of the contiguous buffer where the script lives.
scriptSize := uint64(len(pkScript))
end := offset + scriptSize
msg.TxOut[i].PkScript = scripts[offset:end:end]
offset += scriptSize
// Return the temporary script buffer to the pool.
scriptPool.Return(pkScript)
}
return nil
}
// Deserialize decodes a transaction from r into the receiver using a format
// that is suitable for long-term storage such as a database while respecting
// the Version field in the transaction. This function differs from BtcDecode
// in that BtcDecode decodes from the bitcoin wire protocol as it was sent
// across the network. The wire encoding can technically differ depending on
// the protocol version and doesn't even really need to match the format of a
// stored transaction at all. As of the time this comment was written, the
// encoded transaction is the same in both instances, but there is a distinct
// difference and separating the two allows the API to be flexible enough to
// deal with changes.
func (msg *MsgTx) Deserialize(r io.Reader) error {
// At the current time, there is no difference between the wire encoding
// at protocol version 0 and the stable long-term storage format. As
// a result, make use of BtcDecode.
return msg.BtcDecode(r, 0, WitnessEncoding)
}
// DeserializeNoWitness decodes a transaction from r into the receiver, where
// the transaction encoding format within r MUST NOT utilize the new
// serialization format created to encode transaction bearing witness data
// within inputs.
func (msg *MsgTx) DeserializeNoWitness(r io.Reader) error {
return msg.BtcDecode(r, 0, BaseEncoding)
}
// BtcEncode encodes the receiver to w using the bitcoin protocol encoding.
// This is part of the Message interface implementation.
// See Serialize for encoding transactions to be stored to disk, such as in a
// database, as opposed to encoding transactions for the wire.
func (msg *MsgTx) BtcEncode(w io.Writer, pver uint32, enc MessageEncoding) error {
err := binarySerializer.PutUint32(w, littleEndian, uint32(msg.Version))
if err != nil {
return err
}
// If the encoding version is set to WitnessEncoding, and the Flags
// field for the MsgTx aren't 0x00, then this indicates the transaction
// is to be encoded using the new witness inclusionary structure
// defined in BIP0144.
doWitness := enc == WitnessEncoding && msg.HasWitness()
if doWitness {
// After the txn's Version field, we include two additional
// bytes specific to the witness encoding. The first byte is an
// always 0x00 marker byte, which allows decoders to
// distinguish a serialized transaction with witnesses from a
// regular (legacy) one. The second byte is the Flag field,
// which at the moment is always 0x01, but may be extended in
// the future to accommodate auxiliary non-committed fields.
if _, err := w.Write(witessMarkerBytes); err != nil {
return err
}
}
count := uint64(len(msg.TxIn))
err = WriteVarInt(w, pver, count)
if err != nil {
return err
}
for _, ti := range msg.TxIn {
err = writeTxIn(w, pver, msg.Version, ti)
if err != nil {
return err
}
}
count = uint64(len(msg.TxOut))
err = WriteVarInt(w, pver, count)
if err != nil {
return err
}
for _, to := range msg.TxOut {
err = WriteTxOut(w, pver, msg.Version, to)
if err != nil {
return err
}
}
// If this transaction is a witness transaction, and the witness
// encoded is desired, then encode the witness for each of the inputs
// within the transaction.
if doWitness {
for _, ti := range msg.TxIn {
err = writeTxWitness(w, pver, msg.Version, ti.Witness)
if err != nil {
return err
}
}
}
return binarySerializer.PutUint32(w, littleEndian, msg.LockTime)
}
// HasWitness returns false if none of the inputs within the transaction
// contain witness data, true false otherwise.
func (msg *MsgTx) HasWitness() bool {
for _, txIn := range msg.TxIn {
if len(txIn.Witness) != 0 {
return true
}
}
return false
}
// Serialize encodes the transaction to w using a format that suitable for
// long-term storage such as a database while respecting the Version field in
// the transaction. This function differs from BtcEncode in that BtcEncode
// encodes the transaction to the bitcoin wire protocol in order to be sent
// across the network. The wire encoding can technically differ depending on
// the protocol version and doesn't even really need to match the format of a
// stored transaction at all. As of the time this comment was written, the
// encoded transaction is the same in both instances, but there is a distinct
// difference and separating the two allows the API to be flexible enough to
// deal with changes.
func (msg *MsgTx) Serialize(w io.Writer) error {
// At the current time, there is no difference between the wire encoding
// at protocol version 0 and the stable long-term storage format. As
// a result, make use of BtcEncode.
//
// Passing a encoding type of WitnessEncoding to BtcEncode for MsgTx
// indicates that the transaction's witnesses (if any) should be
// serialized according to the new serialization structure defined in
// BIP0144.
return msg.BtcEncode(w, 0, WitnessEncoding)
}
// SerializeNoWitness encodes the transaction to w in an identical manner to
// Serialize, however even if the source transaction has inputs with witness
// data, the old serialization format will still be used.
func (msg *MsgTx) SerializeNoWitness(w io.Writer) error {
return msg.BtcEncode(w, 0, BaseEncoding)
}
// baseSize returns the serialized size of the transaction without accounting
// for any witness data.
func (msg *MsgTx) baseSize() int {
// Version 4 bytes + LockTime 4 bytes + Serialized varint size for the
// number of transaction inputs and outputs.
n := 8 + VarIntSerializeSize(uint64(len(msg.TxIn))) +
VarIntSerializeSize(uint64(len(msg.TxOut)))
for _, txIn := range msg.TxIn {
n += txIn.SerializeSize()
}
for _, txOut := range msg.TxOut {
n += txOut.SerializeSize()
}
return n
}
// SerializeSize returns the number of bytes it would take to serialize the
// the transaction.
func (msg *MsgTx) SerializeSize() int {
n := msg.baseSize()
if msg.HasWitness() {
// The marker, and flag fields take up two additional bytes.
n += 2
// Additionally, factor in the serialized size of each of the
// witnesses for each txin.
for _, txin := range msg.TxIn {
n += txin.Witness.SerializeSize()
}
}
return n
}
// SerializeSizeStripped returns the number of bytes it would take to serialize
// the transaction, excluding any included witness data.
func (msg *MsgTx) SerializeSizeStripped() int {
return msg.baseSize()
}
// Command returns the protocol command string for the message. This is part
// of the Message interface implementation.
func (msg *MsgTx) Command() string {
return CmdTx
}
// MaxPayloadLength returns the maximum length the payload can be for the
// receiver. This is part of the Message interface implementation.
func (msg *MsgTx) MaxPayloadLength(pver uint32) uint32 {
return MaxBlockPayload
}
// PkScriptLocs returns a slice containing the start of each public key script
// within the raw serialized transaction. The caller can easily obtain the
// length of each script by using len on the script available via the
// appropriate transaction output entry.
func (msg *MsgTx) PkScriptLocs() []int {
numTxOut := len(msg.TxOut)
if numTxOut == 0 {
return nil
}
// The starting offset in the serialized transaction of the first
// transaction output is:
//
// Version 4 bytes + serialized varint size for the number of
// transaction inputs and outputs + serialized size of each transaction
// input.
n := 4 + VarIntSerializeSize(uint64(len(msg.TxIn))) +
VarIntSerializeSize(uint64(numTxOut))
// If this transaction has a witness input, the an additional two bytes
// for the marker, and flag byte need to be taken into account.
if len(msg.TxIn) > 0 && msg.TxIn[0].Witness != nil {
n += 2
}
for _, txIn := range msg.TxIn {
n += txIn.SerializeSize()
}
// Calculate and set the appropriate offset for each public key script.
pkScriptLocs := make([]int, numTxOut)
for i, txOut := range msg.TxOut {
// The offset of the script in the transaction output is:
//
// Value 8 bytes + serialized varint size for the length of
// PkScript.
n += 8 + VarIntSerializeSize(uint64(len(txOut.PkScript)))
pkScriptLocs[i] = n
n += len(txOut.PkScript)
}
return pkScriptLocs
}
// NewMsgTx returns a new bitcoin tx message that conforms to the Message
// interface. The return instance has a default version of TxVersion and there
// are no transaction inputs or outputs. Also, the lock time is set to zero
// to indicate the transaction is valid immediately as opposed to some time in
// future.
func NewMsgTx(version int32) *MsgTx {
return &MsgTx{
Version: version,
TxIn: make([]*TxIn, 0, defaultTxInOutAlloc),
TxOut: make([]*TxOut, 0, defaultTxInOutAlloc),
}
}
// readOutPoint reads the next sequence of bytes from r as an OutPoint.
func readOutPoint(r io.Reader, pver uint32, version int32, op *OutPoint) error {
_, err := io.ReadFull(r, op.Hash[:])
if err != nil {
return err
}
op.Index, err = binarySerializer.Uint32(r, littleEndian)
return err
}
// writeOutPoint encodes op to the bitcoin protocol encoding for an OutPoint
// to w.
func writeOutPoint(w io.Writer, pver uint32, version int32, op *OutPoint) error {
_, err := w.Write(op.Hash[:])
if err != nil {
return err
}
return binarySerializer.PutUint32(w, littleEndian, op.Index)
}
// readScript reads a variable length byte array that represents a transaction
// script. It is encoded as a varInt containing the length of the array
// followed by the bytes themselves. An error is returned if the length is
// greater than the passed maxAllowed parameter which helps protect against
// memory exhaustion attacks and forced panics through malformed messages. The
// fieldName parameter is only used for the error message so it provides more
// context in the error.
func readScript(r io.Reader, pver uint32, maxAllowed uint32, fieldName string) ([]byte, error) {
count, err := ReadVarInt(r, pver)
if err != nil {
return nil, err
}
// Prevent byte array larger than the max message size. It would
// be possible to cause memory exhaustion and panics without a sane
// upper bound on this count.
if count > uint64(maxAllowed) {
str := fmt.Sprintf("%s is larger than the max allowed size "+
"[count %d, max %d]", fieldName, count, maxAllowed)
return nil, messageError("readScript", str)
}
b := scriptPool.Borrow(count)
_, err = io.ReadFull(r, b)
if err != nil {
scriptPool.Return(b)
return nil, err
}
return b, nil
}
// readTxIn reads the next sequence of bytes from r as a transaction input
// (TxIn).
func readTxIn(r io.Reader, pver uint32, version int32, ti *TxIn) error {
err := readOutPoint(r, pver, version, &ti.PreviousOutPoint)
if err != nil {
return err
}
ti.SignatureScript, err = readScript(r, pver, MaxMessagePayload,
"transaction input signature script")
if err != nil {
return err
}
return readElement(r, &ti.Sequence)
}
// writeTxIn encodes ti to the bitcoin protocol encoding for a transaction
// input (TxIn) to w.
func writeTxIn(w io.Writer, pver uint32, version int32, ti *TxIn) error {
err := writeOutPoint(w, pver, version, &ti.PreviousOutPoint)
if err != nil {
return err
}
err = WriteVarBytes(w, pver, ti.SignatureScript)
if err != nil {
return err
}
return binarySerializer.PutUint32(w, littleEndian, ti.Sequence)
}
// readTxOut reads the next sequence of bytes from r as a transaction output
// (TxOut).
func readTxOut(r io.Reader, pver uint32, version int32, to *TxOut) error {
err := readElement(r, &to.Value)
if err != nil {
return err
}
to.PkScript, err = readScript(r, pver, MaxMessagePayload,
"transaction output public key script")
return err
}
// WriteTxOut encodes to into the bitcoin protocol encoding for a transaction
// output (TxOut) to w.
//
// NOTE: This function is exported in order to allow txscript to compute the
// new sighashes for witness transactions (BIP0143).
func WriteTxOut(w io.Writer, pver uint32, version int32, to *TxOut) error {
err := binarySerializer.PutUint64(w, littleEndian, uint64(to.Value))
if err != nil {
return err
}
return WriteVarBytes(w, pver, to.PkScript)
}
// writeTxWitness encodes the bitcoin protocol encoding for a transaction
// input's witness into to w.
func writeTxWitness(w io.Writer, pver uint32, version int32, wit [][]byte) error {
err := WriteVarInt(w, pver, uint64(len(wit)))
if err != nil {
return err
}
for _, item := range wit {
err = WriteVarBytes(w, pver, item)
if err != nil {
return err
}
}
return nil
}