sweepbatcher: fixes for presigned mode #952

starius · 2025-06-12T18:52:30Z

Fixed TODO left from #891 : re-add sweeps to new batches only after fully confirmed.
In case of a reorg sweeps should not go to another batch but stay in the current batch until it is fully confirmed. Only after that the remaining sweeps are re-added to another batch. Field sweep.completed is now set to true only for fully-confirmed sweeps.

Fixed OnChainFeePortion values. There were two mistakes. In case of a swap with multiple sweeps only the fee of the first sweep of a swap was accounted. Another issue is that rounding diff (the remainder) was attributed to all the sweeps rather than to the first (primary) sweep of the batch. The sweep to attribute the remainder was chosen by comparing SignatureScript which is
always empty. New approach is to find the primary sweep and to compare its outpoint directly.

Pull Request Checklist

Update release_notes.md if your PR contains major features, breaking changes or bugfixes

hieblmi

Very nice work @starius, I did a first pass of the PR.

sweepbatcher/sweep_batch.go

lightninglabs-deploy · 2025-06-23T14:27:07Z

@bhandras: review reminder
@sputn1ck: review reminder
@starius, remember to re-request review from reviewers when ready

bhandras

LGTM! 🎉

bhandras · 2025-06-25T11:50:12Z

sweepbatcher/sweep_batch.go

+		go func() {
+			// Make sure this context doesn't expire so we
+			// successfully notify the caller.
+			ctx := context.WithoutCancel(ctx)


nit: alternatively we could just create a new context.Background() and use that.

I prefer context.WithoutCancel, because it inherits key-values of the parent context. This is used for new logging.

bhandras · 2025-06-25T11:57:46Z

sweepbatcher/sweep_batch.go

+		confirmedSet[txIn.PreviousOutPoint] = struct{}{}
+	}
+
+	// As a previous version of the batch transaction may get confirmed,


nit: I think this comment is a bit misleading as we don't really collect the non-confirmed sweeps here to feed them back to the batcher, we just notify those that confirmed and then start the monitoring for 3 conf after which we will do the collection and re-feeding.

The comment was outdated. I updated it:

// As a previous version of the batch transaction may get confirmed, // which does not contain the latest sweeps, we need to detect which // sweeps are in the transaction to correctly calculate fee portions // and notify proper sweeps.

bhandras · 2025-06-25T12:04:34Z

sweepbatcher/presigned.go

@@ -604,8 +605,9 @@ func CheckSignedTx(unsignedTx, signedTx *wire.MsgTx, inputAmt btcutil.Amount,
 	unsignedOut := unsignedTx.TxOut[0]
 	signedOut := signedTx.TxOut[0]
 	if !bytes.Equal(unsignedOut.PkScript, signedOut.PkScript) {
-		return fmt.Errorf("mismatch of output pkScript: %v, %v",
-			unsignedOut.PkScript, signedOut.PkScript)
+		return fmt.Errorf("mismatch of output pkScript: %s, %s",


nit: can just use %x

bhandras · 2025-06-25T12:04:54Z

sweepbatcher/sweep_batcher.go

+	infof("PresignSweepsGroup: nextBlockFeeRate is %v, inputs: %v, "+
+		"destAddress: %v, destPkscript: %v sweepTimeout: %d",
+		nextBlockFeeRate, inputs, destAddress,
+		hex.EncodeToString(destPkscript), sweepTimeout)


nit: can just use %x

This is needed to have multiple spending registrations running and to send a notification to a specific spending registration.

In case of a reorg sweeps should not go to another batch but stay in the current batch until it is fully confirmed. Only after that the remaining sweeps are re-added to another batch. Field sweep.completed is now set to true only for fully-confirmed sweeps. In handleConf we now use batch.persist() (i.e. store.UpdateSweepBatch) instead of ConfirmBatch, because we set not only Confirmed flag, but also batchTxid.

There were two mistakes. In case of a swap with multiple sweeps only the fee of the first sweep of a swap was accounted. Rounding diff (the remainder) was attributed to all the sweeps rather than to the first (primary) sweep of the batch. The sweep to attribute the remainder was chosen by comparing SignatureScript which is always empty. New approach is to find the primary sweep and to compare its outpoint directly.

This is needed because sweepbatcher can use this channel in multiple select statements to unblock itself if the caller cancels.

It doesn't need loopdb, so remove that code.

Make sure that broadcasted tx has feeRate >= minRelayFee. Make sure that feeRate of broadcasted tx doesn't decrease.

Method Presign is not as reliable as SignTx, because it checks transaction by txid and can miss for example if LockTime is different. SignTx can do everything Presign was used for.

For presigned possible remaining groups, the destination address of the current batch was used instead of the destination address of an expected future batch. TODO: reproduce in unit test "purged". For this, each swap should have a separate destination address.

hieblmi

Great work @starius! The detailed commit messages make it very pleasant to review.

hieblmi · 2025-06-25T15:45:58Z

sweepbatcher/sweep_batch.go

+
+	// We are no longer able to accept new sweeps, so we mark the batch as
+	// closed and persist on storage.
+	b.state = Closed


Just checking, this couldn't this result in races because of b.scheduleNextCall(), right? E.g. in

loop/sweepbatcher/sweep_batch.go

Line 515 in 38bf764

if b.state != Open {

hieblmi · 2025-06-26T07:52:46Z

sweepbatcher/store.go

@@ -236,7 +236,7 @@ type dbSweep struct {
 	// Amount is the amount of the sweep.
 	Amount btcutil.Amount

-	// Completed indicates whether this sweep is completed.
+	// Completed indicates whether this sweep is fully-confirmed.


nit: to clarify maybe say "... this sweep is considered reorg-safely confirmed."

hieblmi · 2025-06-26T08:09:26Z

sweepbatcher/sweep_batcher.go

@@ -274,7 +274,7 @@ type addSweepsRequest struct {
 	notifier *SpendNotifier

 	// completed is set if the sweep is spent and the spending transaction
-	// is confirmed.
+	// is fully confirmed.


nit: to clarify maybe say "... the spending transaction is considered reorg-safely confirmed."

hieblmi · 2025-06-26T08:16:16Z

loopout.go

@@ -1147,7 +1147,7 @@ func (s *loopOutSwap) waitForHtlcSpendConfirmedV2(globalCtx context.Context,
 	quitChan := make(chan bool, 1)

 	defer func() {
-		quitChan <- true
+		close(quitChan)


so before this change, only one of multiple selects would have ended?

hieblmi · 2025-06-26T08:18:08Z

sweepbatcher/sweep_batcher.go

-type FeeRateProvider func(ctx context.Context,
-	swapHash lntypes.Hash) (chainfee.SatPerKWeight, error)
+type FeeRateProvider func(ctx context.Context, swapHash lntypes.Hash,
+	utxo wire.OutPoint) (chainfee.SatPerKWeight, error)


why is this needed?

starius marked this pull request as ready for review June 12, 2025 19:15

starius requested review from bhandras, sputn1ck and hieblmi June 13, 2025 03:17

hieblmi reviewed Jun 16, 2025

View reviewed changes

sweepbatcher/sweep_batch.go Show resolved Hide resolved

sweepbatcher/sweep_batch.go Show resolved Hide resolved

starius force-pushed the presigned-fixes2 branch from 239e288 to 70dd1a7 Compare June 24, 2025 15:58

starius requested a review from hieblmi June 24, 2025 16:08

bhandras removed the request for review from sputn1ck June 25, 2025 05:45

bhandras approved these changes Jun 25, 2025

View reviewed changes

starius added 15 commits June 25, 2025 22:53

test/chainnotifier: send to specific spend reg

d2c1689

This is needed to have multiple spending registrations running and to send a notification to a specific spending registration.

loopout: close sweepbatcher quitChan

9ded996

This is needed because sweepbatcher can use this channel in multiple select statements to unblock itself if the caller cancels.

sweepbatcher: pass utxo to fee provider

db5db40

sweepbatcher: make sure dest pkscript is filled

2b54684

sweepbatcher: simplify presigned/purging test

b76ab73

It doesn't need loopdb, so remove that code.

sweepbatcher: make sure HTLC.PkScript is filled

4ada859

sweepbatcher/presigned: minRelayFee edge cases

c4c0a34

Make sure that broadcasted tx has feeRate >= minRelayFee. Make sure that feeRate of broadcasted tx doesn't decrease.

sweepbatcher: remove method Presign

0b0cd1e

Method Presign is not as reliable as SignTx, because it checks transaction by txid and can miss for example if LockTime is different. SignTx can do everything Presign was used for.

sweepbatcher: format pkscript as hex

2dc2119

sweepbatcher: more logging in PresignSweepsGroup

0427df1

loopdb: remove unused method ConfirmBatch

1a2af26

sweepbatcher: check of conf notification has tx

1301904

starius force-pushed the presigned-fixes2 branch from 59da056 to 1301904 Compare June 26, 2025 02:05

hieblmi approved these changes Jun 26, 2025

View reviewed changes

levmi assigned starius Jun 26, 2025

sweepbatcher: fixes for presigned mode #952

Are you sure you want to change the base?

sweepbatcher: fixes for presigned mode #952

Conversation

starius commented Jun 12, 2025

Pull Request Checklist

Uh oh!

hieblmi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

lightninglabs-deploy commented Jun 23, 2025

Uh oh!

bhandras left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

hieblmi left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!