2015-12-21 02:10:24 +03:00
|
|
|
package lnwallet
|
|
|
|
|
|
|
|
import (
|
|
|
|
"bytes"
|
2016-06-30 22:08:27 +03:00
|
|
|
"crypto/sha256"
|
2015-12-21 02:10:24 +03:00
|
|
|
"fmt"
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
"math/big"
|
2015-12-21 02:10:24 +03:00
|
|
|
|
2016-06-30 22:08:27 +03:00
|
|
|
"golang.org/x/crypto/hkdf"
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
"github.com/btcsuite/fastsha256"
|
2016-05-15 17:17:44 +03:00
|
|
|
"github.com/roasbeef/btcd/btcec"
|
|
|
|
"github.com/roasbeef/btcd/txscript"
|
|
|
|
"github.com/roasbeef/btcd/wire"
|
|
|
|
"github.com/roasbeef/btcutil"
|
2015-12-21 02:10:24 +03:00
|
|
|
)
|
|
|
|
|
2015-12-31 09:32:09 +03:00
|
|
|
var (
|
|
|
|
// TODO(roasbeef): remove these and use the one's defined in txscript
|
|
|
|
// within testnet-L.
|
|
|
|
SequenceLockTimeSeconds = uint32(1 << 22)
|
|
|
|
SequenceLockTimeMask = uint32(0x0000ffff)
|
|
|
|
OP_CHECKSEQUENCEVERIFY byte = txscript.OP_NOP3
|
|
|
|
)
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// witnessScriptHash generates a pay-to-witness-script-hash public key script
|
|
|
|
// paying to a version 0 witness program paying to the passed redeem script.
|
|
|
|
func witnessScriptHash(redeemScript []byte) ([]byte, error) {
|
2015-12-21 02:10:24 +03:00
|
|
|
bldr := txscript.NewScriptBuilder()
|
2016-05-04 05:45:32 +03:00
|
|
|
|
|
|
|
bldr.AddOp(txscript.OP_0)
|
|
|
|
scriptHash := fastsha256.Sum256(redeemScript)
|
|
|
|
bldr.AddData(scriptHash[:])
|
2015-12-21 02:10:24 +03:00
|
|
|
return bldr.Script()
|
|
|
|
}
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// genMultiSigScript generates the non-p2sh'd multisig script for 2 of 2
|
2015-12-21 02:10:24 +03:00
|
|
|
// pubkeys.
|
2016-05-04 05:45:32 +03:00
|
|
|
func genMultiSigScript(aPub, bPub []byte) ([]byte, error) {
|
2015-12-21 02:10:24 +03:00
|
|
|
if len(aPub) != 33 || len(bPub) != 33 {
|
|
|
|
return nil, fmt.Errorf("Pubkey size error. Compressed pubkeys only")
|
|
|
|
}
|
|
|
|
|
2015-12-29 08:55:36 +03:00
|
|
|
// Swap to sort pubkeys if needed. Keys are sorted in lexicographical
|
|
|
|
// order. The signatures within the scriptSig must also adhere to the
|
|
|
|
// order, ensuring that the signatures for each public key appears
|
|
|
|
// in the proper order on the stack.
|
|
|
|
if bytes.Compare(aPub, bPub) == -1 {
|
2015-12-21 02:10:24 +03:00
|
|
|
aPub, bPub = bPub, aPub
|
|
|
|
}
|
|
|
|
|
|
|
|
bldr := txscript.NewScriptBuilder()
|
|
|
|
bldr.AddOp(txscript.OP_2)
|
2015-12-29 08:55:36 +03:00
|
|
|
bldr.AddData(aPub) // Add both pubkeys (sorted).
|
2015-12-21 02:10:24 +03:00
|
|
|
bldr.AddData(bPub)
|
|
|
|
bldr.AddOp(txscript.OP_2)
|
|
|
|
bldr.AddOp(txscript.OP_CHECKMULTISIG)
|
|
|
|
return bldr.Script()
|
|
|
|
}
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// genFundingPkScript creates a redeem script, and its matching p2wsh
|
|
|
|
// output for the funding transaction.
|
|
|
|
func genFundingPkScript(aPub, bPub []byte, amt int64) ([]byte, *wire.TxOut, error) {
|
|
|
|
// As a sanity check, ensure that the passed amount is above zero.
|
|
|
|
if amt <= 0 {
|
2015-12-29 08:55:36 +03:00
|
|
|
return nil, nil, fmt.Errorf("can't create FundTx script with " +
|
2016-05-04 05:45:32 +03:00
|
|
|
"zero, or negative coins")
|
2015-12-21 02:10:24 +03:00
|
|
|
}
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// First, create the 2-of-2 multi-sig script itself.
|
|
|
|
redeemScript, err := genMultiSigScript(aPub, bPub)
|
2015-12-21 02:10:24 +03:00
|
|
|
if err != nil {
|
|
|
|
return nil, nil, err
|
|
|
|
}
|
2016-02-05 23:33:22 +03:00
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// With the 2-of-2 script in had, generate a p2wsh script which pays
|
|
|
|
// to the funding script.
|
|
|
|
pkScript, err := witnessScriptHash(redeemScript)
|
2015-12-21 02:10:24 +03:00
|
|
|
if err != nil {
|
|
|
|
return nil, nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
return redeemScript, wire.NewTxOut(amt, pkScript), nil
|
|
|
|
}
|
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// spendMultiSig generates the witness stack required to redeem the 2-of-2 p2wsh
|
2015-12-29 08:55:36 +03:00
|
|
|
// multi-sig output.
|
2016-05-04 05:45:32 +03:00
|
|
|
func spendMultiSig(redeemScript, pubA, sigA, pubB, sigB []byte) [][]byte {
|
|
|
|
witness := make([][]byte, 4)
|
2015-12-21 02:10:24 +03:00
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// When spending a p2wsh multi-sig script, rather than an OP_0, we add
|
|
|
|
// a nil stack element to eat the extra pop.
|
|
|
|
witness[0] = nil
|
2015-12-21 02:10:24 +03:00
|
|
|
|
2016-02-05 23:33:22 +03:00
|
|
|
// When initially generating the redeemScript, we sorted the serialized
|
|
|
|
// public keys in descending order. So we do a quick comparison in order
|
|
|
|
// ensure the signatures appear on the Script Virual Machine stack in
|
|
|
|
// the correct order.
|
|
|
|
if bytes.Compare(pubA, pubB) == -1 {
|
2016-05-04 05:45:32 +03:00
|
|
|
witness[1] = sigB
|
|
|
|
witness[2] = sigA
|
2016-02-05 23:33:22 +03:00
|
|
|
} else {
|
2016-05-04 05:45:32 +03:00
|
|
|
witness[1] = sigA
|
|
|
|
witness[2] = sigB
|
2016-02-05 23:33:22 +03:00
|
|
|
}
|
2015-12-21 02:10:24 +03:00
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
// Finally, add the pre-image as the last witness element.
|
|
|
|
witness[3] = redeemScript
|
2015-12-21 02:10:24 +03:00
|
|
|
|
2016-05-04 05:45:32 +03:00
|
|
|
return witness
|
2015-12-21 02:10:24 +03:00
|
|
|
}
|
2015-12-21 06:47:29 +03:00
|
|
|
|
2015-12-29 08:55:36 +03:00
|
|
|
// findScriptOutputIndex finds the index of the public key script output
|
|
|
|
// matching 'script'. Additionally, a boolean is returned indicating if
|
|
|
|
// a matching output was found at all.
|
|
|
|
// NOTE: The search stops after the first matching script is found.
|
2015-12-21 06:47:29 +03:00
|
|
|
func findScriptOutputIndex(tx *wire.MsgTx, script []byte) (bool, uint32) {
|
|
|
|
found := false
|
|
|
|
index := uint32(0)
|
|
|
|
for i, txOut := range tx.TxOut {
|
|
|
|
if bytes.Equal(txOut.PkScript, script) {
|
|
|
|
found = true
|
|
|
|
index = uint32(i)
|
|
|
|
break
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return found, index
|
|
|
|
}
|
2015-12-31 09:32:09 +03:00
|
|
|
|
2016-01-02 23:10:14 +03:00
|
|
|
// senderHTLCScript constructs the public key script for an outgoing HTLC
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// output payment for the sender's version of the commitment transaction:
|
|
|
|
//
|
|
|
|
// Possible Input Scripts:
|
|
|
|
// SENDR: <sig> 0
|
|
|
|
// RECVR: <sig> <preimage> 0 1
|
|
|
|
// REVOK: <sig <preimage> 1 1
|
|
|
|
// * reciever revoke
|
|
|
|
//
|
|
|
|
// OP_IF
|
|
|
|
// //Receiver
|
|
|
|
// OP_IF
|
|
|
|
// //Revoke
|
|
|
|
// <revocation hash>
|
|
|
|
// OP_ELSE
|
|
|
|
// //Receive
|
|
|
|
// OP_SIZE 32 OP_EQUALVERIFY
|
|
|
|
// <payment hash>
|
|
|
|
// OP_ENDIF
|
|
|
|
// OP_SWAP
|
|
|
|
// OP_SHA256 OP_EQUALVERIFY
|
|
|
|
// <recv key> OP_CHECKSIG
|
|
|
|
// OP_ELSE
|
|
|
|
// //Sender
|
|
|
|
// <absolute blockheight> OP_CHECKLOCKTIMEVERIFY
|
|
|
|
// <relative blockheight> OP_CHECKSEQUENCEVERIFY
|
|
|
|
// OP_2DROP
|
|
|
|
// <sendr key> OP_CHECKSIG
|
|
|
|
// OP_ENDIF
|
2015-12-31 09:32:09 +03:00
|
|
|
func senderHTLCScript(absoluteTimeout, relativeTimeout uint32, senderKey,
|
|
|
|
receiverKey *btcec.PublicKey, revokeHash, paymentHash []byte) ([]byte, error) {
|
|
|
|
|
|
|
|
builder := txscript.NewScriptBuilder()
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// The receiver of the HTLC places a 1 as the first item in the witness
|
|
|
|
// stack, forcing Script execution to enter the "if" clause within the
|
|
|
|
// main body of the script.
|
|
|
|
builder.AddOp(txscript.OP_IF)
|
2015-12-31 09:32:09 +03:00
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// The receiver will place a 1 as the second item of the witness stack
|
|
|
|
// in the case the sender broadcasts a revoked commitment transaction.
|
|
|
|
// Executing this branch allows the receiver to claim the sender's
|
|
|
|
// funds as a result of their contract violation.
|
|
|
|
builder.AddOp(txscript.OP_IF)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(revokeHash)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
|
|
|
// Alternatively, the receiver can place a 0 as the second item of the
|
|
|
|
// witness stack if they wish to claim the HTLC with the proper
|
|
|
|
// pre-image as normal. In order to prevent an over-sized pre-image
|
|
|
|
// attack (which can create undesirable redemption asymmerties, we
|
|
|
|
// strongly require that all HTLC pre-images are exactly 32 bytes.
|
|
|
|
builder.AddOp(txscript.OP_ELSE)
|
|
|
|
builder.AddOp(txscript.OP_SIZE)
|
|
|
|
builder.AddInt64(32)
|
|
|
|
builder.AddOp(txscript.OP_EQUALVERIFY)
|
|
|
|
builder.AddData(paymentHash)
|
|
|
|
builder.AddOp(txscript.OP_ENDIF)
|
|
|
|
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_SWAP)
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_SHA256)
|
|
|
|
builder.AddOp(txscript.OP_EQUALVERIFY)
|
|
|
|
|
|
|
|
// In either case, we require a valid signature by the receiver.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(receiverKey.SerializeCompressed())
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_CHECKSIG)
|
|
|
|
|
|
|
|
// Otherwise, the sender of the HTLC will place a 0 as the first item
|
|
|
|
// of the witness stack in order to sweep the funds back after the HTLC
|
|
|
|
// times out.
|
|
|
|
builder.AddOp(txscript.OP_ELSE)
|
2015-12-31 09:32:09 +03:00
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// In this case, the sender will need to wait for an absolute HTLC
|
2015-12-31 09:32:09 +03:00
|
|
|
// timeout, then afterwards a relative timeout before we claim re-claim
|
2016-01-02 23:10:14 +03:00
|
|
|
// the unsettled funds. This delay gives the other party a chance to
|
|
|
|
// present the pre-image to the revocation hash in the event that the
|
|
|
|
// sender (at this time) broadcasts this commitment transaction after
|
|
|
|
// it has been revoked.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddInt64(int64(absoluteTimeout))
|
|
|
|
builder.AddOp(txscript.OP_CHECKLOCKTIMEVERIFY)
|
|
|
|
builder.AddInt64(int64(relativeTimeout))
|
|
|
|
builder.AddOp(OP_CHECKSEQUENCEVERIFY)
|
|
|
|
builder.AddOp(txscript.OP_2DROP)
|
|
|
|
builder.AddData(senderKey.SerializeCompressed())
|
|
|
|
builder.AddOp(txscript.OP_CHECKSIG)
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_ENDIF)
|
|
|
|
|
2015-12-31 09:32:09 +03:00
|
|
|
return builder.Script()
|
|
|
|
}
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// senderHtlcSpendRevoke constructs a valid witness allowing the reciever of an
|
|
|
|
// HTLC to claim the output with knowledge of the revocation preimage in the
|
|
|
|
// scenario that the sender of the HTLC broadcasts a previously revoked
|
|
|
|
// commitment transaction. A valid spend requires knowledge of the pre-image to
|
|
|
|
// the commitment transaction's revocation hash, and a valid signature under
|
|
|
|
// the receiver's public key.
|
|
|
|
func senderHtlcSpendRevoke(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
reciverKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
revokePreimage []byte) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, reciverKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// In order to force script execution to enter the revocation clause,
|
|
|
|
// we place two one's as the first items in the final evalulated
|
|
|
|
// witness stack.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 5))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = revokePreimage
|
|
|
|
witnessStack[2] = []byte{1}
|
|
|
|
witnessStack[3] = []byte{1}
|
|
|
|
witnessStack[4] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// senderHtlcSpendRedeem constructs a valid witness allowing the receiver of an
|
|
|
|
// HTLC to redeem the pending output in the scenario that the sender broadcasts
|
|
|
|
// their version of the commitment transaction. A valid spend requires
|
|
|
|
// knowledge of the payment pre-image, and a valid signature under the
|
|
|
|
// receivers public key.
|
|
|
|
func senderHtlcSpendRedeem(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
reciverKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
paymentPreimage []byte) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, reciverKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// We force script execution into the HTLC redemption clause by placing
|
|
|
|
// a one, then a zero as the first items in the final evalulated
|
|
|
|
// witness stack.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 5))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = paymentPreimage
|
|
|
|
witnessStack[2] = []byte{0}
|
|
|
|
witnessStack[3] = []byte{1}
|
|
|
|
witnessStack[4] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// htlcSpendTimeout constructs a valid witness allowing the sender of an HTLC
|
|
|
|
// to recover the pending funds after an absolute, then relative locktime
|
|
|
|
// period.
|
|
|
|
func senderHtlcSpendTimeout(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
senderKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
absoluteTimeout, relativeTimeout uint32) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// Since the HTLC output has an absolute timeout before we're permitted
|
|
|
|
// to sweep the output, we need to set the locktime of this sweepign
|
|
|
|
// transaction to that aboslute value in order to pass Script
|
|
|
|
// verification.
|
|
|
|
sweepTx.LockTime = absoluteTimeout
|
|
|
|
|
|
|
|
// Additionally, we're required to wait a relative period of time
|
|
|
|
// before we can sweep the output in order to allow the other party to
|
|
|
|
// contest our claim of validity to this version of the commitment
|
|
|
|
// transaction.
|
|
|
|
sweepTx.TxIn[0].Sequence = lockTimeToSequence(false, relativeTimeout)
|
|
|
|
|
|
|
|
// Finally, OP_CSV requires that the version of the transaction
|
|
|
|
// spending a pkscript with OP_CSV within it *must* be >= 2.
|
|
|
|
sweepTx.Version = 2
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, senderKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// We place a zero as the first item of the evaluated witness stack in
|
|
|
|
// order to force Script execution to the HTLC timeout clause.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 3))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = []byte{0}
|
|
|
|
witnessStack[2] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// receiverHTLCScript constructs the public key script for an incoming HTLC
|
|
|
|
// output payment for the receiver's version of the commitment transaction:
|
|
|
|
//
|
|
|
|
// Possible Input Scripts:
|
|
|
|
// RECVR: <sig> <preimage> 1
|
|
|
|
// REVOK: <sig> <preimage> 1 0
|
|
|
|
// SENDR: <sig> 0 0
|
|
|
|
//
|
|
|
|
// OP_IF
|
|
|
|
// //Receiver
|
|
|
|
// OP_SIZE 32 OP_EQUALVERIFY
|
|
|
|
// OP_SHA256
|
|
|
|
// <payment hash> OP_EQUALVERIFY
|
|
|
|
// <relative blockheight> OP_CHECKSEQUENCEVERIFY OP_DROP
|
|
|
|
// <receiver key> OP_CHECKSIG
|
|
|
|
// OP_ELSE
|
|
|
|
// //Sender
|
|
|
|
// OP_IF
|
|
|
|
// //Revocation
|
|
|
|
// OP_SHA256
|
|
|
|
// <revoke hash> OP_EQUALVERIFY
|
|
|
|
// OP_ELSE
|
|
|
|
// //Refund
|
|
|
|
// <absolute blockehight> OP_CHECKLOCKTIMEVERIFY OP_DROP
|
|
|
|
// OP_ENDIF
|
|
|
|
// <sender key> OP_CHECKSIG
|
|
|
|
// OP_ENDIF
|
2016-06-30 22:08:27 +03:00
|
|
|
// TODO(roasbeef): rename these to sender vs receiver?
|
2015-12-31 09:32:09 +03:00
|
|
|
func receiverHTLCScript(absoluteTimeout, relativeTimeout uint32, senderKey,
|
|
|
|
receiverKey *btcec.PublicKey, revokeHash, paymentHash []byte) ([]byte, error) {
|
|
|
|
|
|
|
|
builder := txscript.NewScriptBuilder()
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// The receiver of the script will place a 1 as the first item of the
|
|
|
|
// witness stack forcing Script execution to enter the "if" clause of
|
|
|
|
// the main body of the script.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_IF)
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// In this clause, the receiver can redeem the HTLC after a relative timeout.
|
|
|
|
// This added delay gives the sender (at this time) an opportunity to
|
|
|
|
// re-claim the pending HTLC in the event that the receiver
|
|
|
|
// (at this time) broadcasts this old commitment transaction after it
|
|
|
|
// has been revoked. Additionally, we require that the pre-image is
|
|
|
|
// exactly 32-bytes in order to avoid undesirable redemption
|
|
|
|
// asymmerties in the multi-hop scenario.
|
|
|
|
builder.AddOp(txscript.OP_SIZE)
|
|
|
|
builder.AddInt64(32)
|
|
|
|
builder.AddOp(txscript.OP_EQUALVERIFY)
|
|
|
|
builder.AddOp(txscript.OP_SHA256)
|
|
|
|
builder.AddData(paymentHash)
|
|
|
|
builder.AddOp(txscript.OP_EQUALVERIFY)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddInt64(int64(relativeTimeout))
|
|
|
|
builder.AddOp(OP_CHECKSEQUENCEVERIFY)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_DROP)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(receiverKey.SerializeCompressed())
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_CHECKSIG)
|
2015-12-31 09:32:09 +03:00
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// Otherwise, the sender will place a 0 as the first item of the
|
|
|
|
// witness stack forcing exeuction to enter the "else" clause of the
|
|
|
|
// main body of the script.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_ELSE)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
|
|
|
// The sender will place a 1 as the second item of the witness stack
|
|
|
|
// in the scenario that the receiver broadcasts an invalidated
|
|
|
|
// commitment transaction, allowing the sender to sweep all the
|
|
|
|
// receiver's funds.
|
|
|
|
builder.AddOp(txscript.OP_IF)
|
|
|
|
builder.AddOp(txscript.OP_SHA256)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(revokeHash)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_EQUALVERIFY)
|
2015-12-31 09:32:09 +03:00
|
|
|
|
2016-01-02 23:10:14 +03:00
|
|
|
// If not, then the sender needs to wait for the HTLC timeout. This
|
|
|
|
// clause may be executed if the receiver fails to present the r-value
|
|
|
|
// in time. This prevents the pending funds from being locked up
|
|
|
|
// indefinately.
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
|
|
|
// The sender will place a 0 as the second item of the witness stack if
|
|
|
|
// they wish to sweep the HTLC after an absolute refund timeout. This
|
|
|
|
// time out clause prevents the pending funds from being locked up
|
|
|
|
// indefinately.
|
|
|
|
builder.AddOp(txscript.OP_ELSE)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddInt64(int64(absoluteTimeout))
|
2016-01-19 11:19:16 +03:00
|
|
|
builder.AddOp(txscript.OP_CHECKLOCKTIMEVERIFY)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_DROP)
|
|
|
|
builder.AddOp(txscript.OP_ENDIF)
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// In either case, we also require a valid signature with the sender's
|
|
|
|
// commitment private key.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(senderKey.SerializeCompressed())
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_CHECKSIG)
|
2015-12-31 09:32:09 +03:00
|
|
|
|
|
|
|
builder.AddOp(txscript.OP_ENDIF)
|
|
|
|
|
|
|
|
return builder.Script()
|
|
|
|
}
|
|
|
|
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// receiverHtlcSpendRedeem constructs a valid witness allowing the receiver of
|
|
|
|
// an HTLC to redeem the conditional payment in the event that their commitment
|
|
|
|
// transaction is broadcast. Since this is a pay out to the receiving party as
|
|
|
|
// an output on their commitment transaction, a relative time delay is required
|
|
|
|
// before the output can be spent.
|
|
|
|
func receiverHtlcSpendRedeem(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
reciverKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
paymentPreimage []byte, relativeTimeout uint32) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// In order to properly spend the transaction, we need to set the
|
|
|
|
// sequence number. We do this by convering the relative block delay
|
|
|
|
// into a sequence number value able to be interpeted by
|
|
|
|
// OP_CHECKSEQUENCEVERIFY.
|
|
|
|
sweepTx.TxIn[0].Sequence = lockTimeToSequence(false, relativeTimeout)
|
|
|
|
|
|
|
|
// Additionally, OP_CSV requires that the version of the transaction
|
|
|
|
// spending a pkscript with OP_CSV within it *must* be >= 2.
|
|
|
|
sweepTx.Version = 2
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, reciverKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// Place a one as the first item in the evaluated witness stack to
|
|
|
|
// force script execution to the HTLC redemption clause.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 4))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = paymentPreimage
|
|
|
|
witnessStack[2] = []byte{1}
|
|
|
|
witnessStack[3] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// receiverHtlcSpendRevoke constructs a valid witness allowing the sender of an
|
|
|
|
// HTLC within a previously revoked commitment transaction to re-claim the
|
|
|
|
// pending funds in the case that the receiver broadcasts this revoked
|
|
|
|
// commitment transaction.
|
|
|
|
func receiverHtlcSpendRevoke(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
senderKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
revokePreimage []byte) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// TODO(roasbeef): move sig generate outside func, or just factor out?
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, senderKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// We place a zero, then one as the first items in the evaluated
|
|
|
|
// witness stack in order to force script execution to the HTLC
|
|
|
|
// revocation clause.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 5))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = revokePreimage
|
|
|
|
witnessStack[2] = []byte{1}
|
|
|
|
witnessStack[3] = []byte{0}
|
|
|
|
witnessStack[4] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// receiverHtlcSpendTimeout constructs a valid witness allowing the sender of
|
|
|
|
// an HTLC to recover the pending funds after an absolute timeout in the
|
|
|
|
// scenario that the receiver of the HTLC broadcasts their version of the
|
|
|
|
// commitment transaction.
|
|
|
|
func receiverHtlcSpendTimeout(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
senderKey *btcec.PrivateKey, sweepTx *wire.MsgTx,
|
|
|
|
absoluteTimeout uint32) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// The HTLC output has an absolute time period before we are permitted
|
|
|
|
// to recover the pending funds. Therefore we need to set the locktime
|
|
|
|
// on this sweeping transaction in order to pass Script verification.
|
|
|
|
sweepTx.LockTime = absoluteTimeout
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, senderKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 4))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = []byte{0}
|
|
|
|
witnessStack[2] = []byte{0}
|
|
|
|
witnessStack[3] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
2016-02-06 07:24:17 +03:00
|
|
|
// lockTimeToSequence converts the passed relative locktime to a sequence
|
|
|
|
// number in accordance to BIP-68.
|
|
|
|
// See: https://github.com/bitcoin/bips/blob/master/bip-0068.mediawiki
|
|
|
|
// * (Compatibility)
|
|
|
|
func lockTimeToSequence(isSeconds bool, locktime uint32) uint32 {
|
|
|
|
if !isSeconds {
|
|
|
|
// The locktime is to be expressed in confirmations. Apply the
|
|
|
|
// mask to restrict the number of confirmations to 65,535 or
|
|
|
|
// 1.25 years.
|
|
|
|
return SequenceLockTimeMask & locktime
|
|
|
|
}
|
|
|
|
|
|
|
|
// Set the 22nd bit which indicates the lock time is in seconds, then
|
|
|
|
// shift the locktime over by 9 since the time granularity is in
|
|
|
|
// 512-second intervals (2^9). This results in a max lock-time of
|
|
|
|
// 33,554,431 seconds, or 1.06 years.
|
|
|
|
return SequenceLockTimeSeconds | (locktime >> 9)
|
|
|
|
}
|
|
|
|
|
2016-01-02 23:10:14 +03:00
|
|
|
// commitScriptToSelf constructs the public key script for the output on the
|
|
|
|
// commitment transaction paying to the "owner" of said commitment transaction.
|
|
|
|
// If the other party learns of the pre-image to the revocation hash, then they
|
|
|
|
// can claim all the settled funds in the channel, plus the unsettled funds.
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
//
|
|
|
|
// Possible Input Scripts:
|
|
|
|
// REVOKE: <sig> 1
|
|
|
|
// SENDRSWEEP: <sig> 0
|
|
|
|
//
|
|
|
|
// Output Script:
|
|
|
|
// OP_IF
|
|
|
|
// <revokeKey> OP_CHECKSIG
|
|
|
|
// OP_ELSE
|
|
|
|
// <timeKey> OP_CHECKSIGVERIFY
|
|
|
|
// <numRelativeBlocks> OP_CHECKSEQUENCEVERIFY
|
|
|
|
// OP_ENDIF
|
|
|
|
func commitScriptToSelf(csvTimeout uint32, selfKey, revokeKey *btcec.PublicKey) ([]byte, error) {
|
2015-12-31 09:32:09 +03:00
|
|
|
// This script is spendable under two conditions: either the 'csvTimeout'
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// has passed and we can redeem our funds, or they can produce a valid
|
|
|
|
// signature with the revocation public key. The revocation public key
|
|
|
|
// will *only* be known to the other party if we have divulged the
|
|
|
|
// revocation hash, allowing them to homomorphically derive the proper
|
|
|
|
// private key which coresponds to the revoke public key.
|
2015-12-31 09:32:09 +03:00
|
|
|
builder := txscript.NewScriptBuilder()
|
|
|
|
|
|
|
|
builder.AddOp(txscript.OP_IF)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
|
|
|
// If a valid signature using the revocation key is presented, then
|
|
|
|
// allow an immediate spend provided the proper signature.
|
|
|
|
builder.AddData(revokeKey.SerializeCompressed())
|
|
|
|
builder.AddOp(txscript.OP_CHECKSIG)
|
|
|
|
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_ELSE)
|
|
|
|
|
|
|
|
// Otherwise, we can re-claim our funds after a CSV delay of
|
|
|
|
// 'csvTimeout' timeout blocks, and a valid signature.
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddData(selfKey.SerializeCompressed())
|
|
|
|
builder.AddOp(txscript.OP_CHECKSIGVERIFY)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddInt64(int64(csvTimeout))
|
|
|
|
builder.AddOp(OP_CHECKSEQUENCEVERIFY)
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddOp(txscript.OP_ENDIF)
|
|
|
|
|
|
|
|
return builder.Script()
|
|
|
|
}
|
|
|
|
|
2016-01-02 23:10:14 +03:00
|
|
|
// commitScriptUnencumbered constructs the public key script on the commitment
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
// transaction paying to the "other" party. The constructed output is a normal
|
|
|
|
// p2wkh output spendable immediately, requiring no contestation period.
|
2015-12-31 09:32:09 +03:00
|
|
|
func commitScriptUnencumbered(key *btcec.PublicKey) ([]byte, error) {
|
|
|
|
// This script goes to the "other" party, and it spendable immediately.
|
|
|
|
builder := txscript.NewScriptBuilder()
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
builder.AddOp(txscript.OP_0)
|
2015-12-31 09:32:09 +03:00
|
|
|
builder.AddData(btcutil.Hash160(key.SerializeCompressed()))
|
|
|
|
|
|
|
|
return builder.Script()
|
|
|
|
}
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
|
|
|
|
// commitSpendTimeout constructs a valid witness allowing the owner of a
|
|
|
|
// particular commitment transaction to spend the output returning settled
|
|
|
|
// funds back to themselves after an absolute block timeout.
|
|
|
|
func commitSpendTimeout(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
blockTimeout uint32, selfKey *btcec.PrivateKey,
|
|
|
|
sweepTx *wire.MsgTx) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// In order to properly spend the transaction, we need to set the
|
|
|
|
// sequence number. We do this by convering the relative block delay
|
|
|
|
// into a sequence number value able to be interpeted by
|
|
|
|
// OP_CHECKSEQUENCEVERIFY.
|
|
|
|
sweepTx.TxIn[0].Sequence = lockTimeToSequence(false, blockTimeout)
|
|
|
|
|
|
|
|
// Additionally, OP_CSV requires that the version of the transaction
|
|
|
|
// spending a pkscript with OP_CSV within it *must* be >= 2.
|
|
|
|
sweepTx.Version = 2
|
|
|
|
|
|
|
|
// With the sequence number in place, we're now able to properly sign
|
|
|
|
// off on the sweep transaction.
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, selfKey)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// Place a zero as the first item in the evaluated witness stack to
|
|
|
|
// force script execution to the timeout spend clause.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 3))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = []byte{0}
|
|
|
|
witnessStack[2] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// commitSpendRevoke constructs a valid witness allowing a node to sweep the
|
|
|
|
// settled output of a malicious counter-party who broadcasts a revoked
|
|
|
|
// commitment trransaction.
|
|
|
|
func commitSpendRevoke(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
revocationPriv *btcec.PrivateKey, sweepTx *wire.MsgTx) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
sweepSig, err := txscript.RawTxInWitnessSignature(
|
|
|
|
sweepTx, hashCache, 0, int64(outputAmt), commitScript,
|
|
|
|
txscript.SigHashAll, revocationPriv)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
// Place a 1 as the first item in the evaluated witness stack to
|
|
|
|
// force script execution to the revocation clause.
|
|
|
|
witnessStack := wire.TxWitness(make([][]byte, 3))
|
|
|
|
witnessStack[0] = sweepSig
|
|
|
|
witnessStack[1] = []byte{1}
|
|
|
|
witnessStack[2] = commitScript
|
|
|
|
|
|
|
|
return witnessStack, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// commitSpendNoDelay constructs a valid witness allowing a node to spend their
|
|
|
|
// settled no-delay output on the counter-party's commitment transaction.
|
|
|
|
func commitSpendNoDelay(commitScript []byte, outputAmt btcutil.Amount,
|
|
|
|
commitPriv *btcec.PrivateKey, sweepTx *wire.MsgTx) (wire.TxWitness, error) {
|
|
|
|
|
|
|
|
// This is just a regular p2wkh spend which looks something like:
|
|
|
|
// * witness: <sig> <pubkey>
|
|
|
|
hashCache := txscript.NewTxSigHashes(sweepTx)
|
|
|
|
witness, err := txscript.WitnessScript(sweepTx, hashCache, 0,
|
|
|
|
int64(outputAmt), commitScript, txscript.SigHashAll,
|
|
|
|
commitPriv, true)
|
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
return wire.TxWitness(witness), nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// deriveRevocationPubkey derives the revocation public key given the
|
|
|
|
// counter-party's commitment key, and revocation pre-image derived via a
|
|
|
|
// pseudo-random-function. In the event that we (for some reason) broadcast a
|
|
|
|
// revoked commitment transaction, then if the other party knows the revocation
|
|
|
|
// pre-image, then they'll be able to derive the corresponding private key to
|
|
|
|
// this private key by exploting the homomorphism in the elliptic curve group:
|
|
|
|
// * https://en.wikipedia.org/wiki/Group_homomorphism#Homomorphisms_of_abelian_groups
|
|
|
|
//
|
|
|
|
// The derivation is performed as follows:
|
|
|
|
//
|
|
|
|
// revokeKey := commitKey + revokePoint
|
|
|
|
// := G*k + G+h
|
|
|
|
// := G * (k+h)
|
|
|
|
//
|
|
|
|
// Therefore, once we divulge the revocation pre-image, the remote peer is able to
|
|
|
|
// compute the proper private key for the revokeKey by computing:
|
|
|
|
// revokePriv := commitPriv + revokePreimge mod N
|
|
|
|
//
|
|
|
|
// Where N is the order of the sub-group.
|
|
|
|
func deriveRevocationPubkey(commitPubKey *btcec.PublicKey,
|
|
|
|
revokePreimage []byte) *btcec.PublicKey {
|
|
|
|
|
|
|
|
// First we need to convert the revocation hash into a point on the
|
|
|
|
// elliptic curve.
|
|
|
|
revokePointX, revokePointY := btcec.S256().ScalarBaseMult(revokePreimage)
|
|
|
|
|
|
|
|
// Now that we have the revocation point, we add this to their commitment
|
|
|
|
// public key in order to obtain the revocation public key.
|
|
|
|
revokeX, revokeY := btcec.S256().Add(commitPubKey.X, commitPubKey.Y,
|
|
|
|
revokePointX, revokePointY)
|
|
|
|
return &btcec.PublicKey{X: revokeX, Y: revokeY}
|
|
|
|
}
|
|
|
|
|
|
|
|
// deriveRevocationPrivKey derives the revocation private key given a node's
|
|
|
|
// commitment private key, and the pre-image to a previously seen revocation
|
|
|
|
// hash. Using this derived private key, a node is able to claim the output
|
|
|
|
// within the commitment transaction of a node in the case that they broadcast
|
|
|
|
// a previously revoked commitment transaction.
|
|
|
|
//
|
|
|
|
// The private key is derived as follwos:
|
2016-06-30 22:08:27 +03:00
|
|
|
// revokePriv := commitPriv + revokePreimage mod N
|
lnwallet: update HTLC+commitment scripts
This commit updates the previous HTLC and commitment scripts to their
current latest evolution.
The HTLC scripts have been optimized for space savings, the
functionality itself has remained relatively unchanged. A trade off was
made to add additional bytes into the sigScript in order to avoid
extraneous CHECKSIG’s. The rationale is that an extra 1-2 bytes in the
sigScript to guide execution, are worthwhile since they’re in the
witness, and witness data may be pruned in the near future.
The primary change is within the commitment transaction itself. Instead
of using revocation hashes, we now use signature based revocation. This
saves space in the Script, and optimizes away an extra hashing
operation. Elkrem/shachain is still used but, we now use the pre-images
to homomorphically derive a public key which the other party will be
able to sign with, once we disclose the pre-image itself.
Finally, we have switched to using SHA-256 everywhere uniformly for
both revocation hashes, and payment hashes. The rationale is that the
output of ripemd160 is too small for modern security margins, and that
other coins/chains are more likely to have SHA-256 implemented, than
ripemd160.
A set of tests has also been included which contain (mostly) exhaustive
tests of all possible redemption paths for both commitment and HTLC.
2016-06-27 21:33:59 +03:00
|
|
|
//
|
|
|
|
// Where N is the order of the sub-group.
|
|
|
|
func deriveRevocationPrivKey(commitPrivKey *btcec.PrivateKey,
|
|
|
|
revokePreimage []byte) *btcec.PrivateKey {
|
|
|
|
|
|
|
|
// Convert the revocation pre-image into a scalar value so we can
|
|
|
|
// manipulate it within the curve's defined finite field.
|
|
|
|
revokeScalar := new(big.Int).SetBytes(revokePreimage)
|
|
|
|
|
|
|
|
// To derive the revocation private key, we simply add the revocation
|
|
|
|
// pre-image to the commitment private key.
|
|
|
|
//
|
|
|
|
// This works since:
|
|
|
|
// P = G*a + G*b
|
|
|
|
// = G*(a+b)
|
|
|
|
// = G*p
|
|
|
|
revokePriv := revokeScalar.Add(revokeScalar, commitPrivKey.D)
|
|
|
|
revokePriv = revokePriv.Mod(revokePriv, btcec.S256().N)
|
|
|
|
|
|
|
|
privRevoke, _ := btcec.PrivKeyFromBytes(btcec.S256(), revokePriv.Bytes())
|
|
|
|
return privRevoke
|
|
|
|
}
|
2016-06-30 22:08:27 +03:00
|
|
|
|
|
|
|
// deriveElkremRoot derives an elkrem root unique to a channel given the
|
|
|
|
// private key for our public key in the 2-of-2 multi-sig, and the remote
|
|
|
|
// node's multi-sig public key. The root is derived using the HKDF[1][2]
|
|
|
|
// instantiated with sha-256. The secret data used is our multi-sig private
|
|
|
|
// key, with the salt being the remote node's public key.
|
|
|
|
//
|
|
|
|
// [1]: https://eprint.iacr.org/2010/264.pdf
|
|
|
|
// [2]: https://tools.ietf.org/html/rfc5869
|
|
|
|
func deriveElkremRoot(localMultiSigKey *btcec.PrivateKey,
|
|
|
|
remoteMultiSigKey *btcec.PublicKey) wire.ShaHash {
|
|
|
|
|
|
|
|
secret := localMultiSigKey.Serialize()
|
|
|
|
salt := remoteMultiSigKey.SerializeCompressed()
|
|
|
|
info := []byte("elkrem")
|
|
|
|
|
|
|
|
rootReader := hkdf.New(sha256.New, secret, salt, info)
|
|
|
|
|
|
|
|
// It's safe to ignore the error her as we know for sure that we won't
|
|
|
|
// be draining the HKDF past its available entropy horizon.
|
|
|
|
// TODO(roasbeef): revisit...
|
|
|
|
var elkremRoot wire.ShaHash
|
|
|
|
rootReader.Read(elkremRoot[:])
|
|
|
|
|
|
|
|
return elkremRoot
|
|
|
|
}
|