Skip to content

Allow counterparty pending monitor update within quiescence handshake #3806

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jun 4, 2025
+60 −2
Merged

Allow counterparty pending monitor update within quiescence handshake #3806

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 6 additions & 2 deletions lightning/src/ln/channel.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9675,10 +9675,14 @@ impl<SP: Deref> FundedChannel<SP> where
self.mark_response_received();

if self.context.is_waiting_on_peer_pending_channel_update()
|| self.context.is_monitor_or_signer_pending_channel_update()
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What happens if we receive a commitment_signed, causing us to go monitor_update_in_progress and prep to send an RAA in response? I guess in theory they're not supposed to send their stfu after they sent their CS because they're waiting on a response from us, but also probably we want to reject that because trying to handle splicing when we have two pending commitments seems annoying?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess in theory they're not supposed to send their stfu after they sent their CS because they're waiting on a response from us

This is allowed as long as the commitment_signed is not sent as a result of a local update. In any case, our implementation will accept an inbound stfu at any point, but it will hold back sending its own as a response until both is_waiting_on_peer_pending_channel_update and is_monitor_or_signer_pending_channel_update are false.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I was thinking in the context of where we'd already sent our stfu. AFAIU if we send a stfu, then they send a CS followed by an stfu (for a local update, because if it was in response to our update we shouldn't have sent our stfu). They may violate the protocol, but we really should be rejecting it explicitly because otherwise we're gonna respond with an RAA+CS and mark ourselves quiescent even though there are two valid commitment transactions pending (and have to deal with that when splicing, which we shouldn't have to).

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe this case is already covered. We can receive commitment_signed if:

  1. a local update was made, and we now owe them a final revoke_and_ack or
  2. a remote update was made, and we now owe them a revoke_and_ack + commitment_signed, and they owe us a final revoke_and_ack

In case 1, we cannot send stfu until the local update is no longer pending, so receiving stfu here doesn't change anything.
In case 2, we'll disconnect them if they send stfu immediately after commitment_signed because we're awaiting_remote_revoke (covered by is_waiting_on_peer_pending_channel_update).

|| self.context.signer_pending_revoke_and_ack
|| self.context.signer_pending_commitment_update
{
// Since we've already sent `stfu`, it should not be possible for one of our updates to
// be pending, so anything pending currently must be from a counterparty update.
// be pending, so anything pending currently must be from a counterparty update. We may
// have a monitor update pending if we've processed a message from the counterparty, but
// we don't consider this when becoming quiescent since the states are not mutually
// exclusive.
return Err(ChannelError::WarnAndDisconnect(
"Received counterparty stfu while having pending counterparty updates".to_owned()
));
Expand Down
54 changes: 54 additions & 0 deletions lightning/src/ln/quiescence_tests.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -276,6 +276,60 @@ fn test_quiescence_waits_for_async_signer_and_monitor_update() {
send_payment(&nodes[0], &[&nodes[1]], payment_amount);
}

#[test]
fn test_quiescence_on_final_revoke_and_ack_pending_monitor_update() {
// Test that we do not let a pending monitor update for a final `revoke_and_ack` prevent us from
// entering quiescence. This was caught by the fuzzer, reported as #3805.
let chanmon_cfgs = create_chanmon_cfgs(2);
let node_cfgs = create_node_cfgs(2, &chanmon_cfgs);
let node_chanmgrs = create_node_chanmgrs(2, &node_cfgs, &[None, None]);
let nodes = create_network(2, &node_cfgs, &node_chanmgrs);
let chan_id = create_announced_chan_between_nodes(&nodes, 0, 1).2;

let node_id_0 = nodes[0].node.get_our_node_id();
let node_id_1 = nodes[1].node.get_our_node_id();

let payment_amount = 1_000_000;
let (route, payment_hash, _, payment_secret) =
get_route_and_payment_hash!(&nodes[0], &nodes[1], payment_amount);
let onion = RecipientOnionFields::secret_only(payment_secret);
let payment_id = PaymentId(payment_hash.0);
nodes[0].node.send_payment_with_route(route, payment_hash, onion, payment_id).unwrap();
check_added_monitors(&nodes[0], 1);

nodes[1].node.maybe_propose_quiescence(&node_id_0, &chan_id).unwrap();
let stfu = get_event_msg!(&nodes[1], MessageSendEvent::SendStfu, node_id_0);
nodes[0].node.handle_stfu(node_id_1, &stfu);

let update_add = get_htlc_update_msgs!(&nodes[0], node_id_1);
nodes[1].node.handle_update_add_htlc(node_id_0, &update_add.update_add_htlcs[0]);
nodes[1].node.handle_commitment_signed_batch_test(node_id_0, &update_add.commitment_signed);
check_added_monitors(&nodes[1], 1);

let (revoke_and_ack, commit_sig) = get_revoke_commit_msgs!(&nodes[1], node_id_0);
nodes[0].node.handle_revoke_and_ack(node_id_1, &revoke_and_ack);
check_added_monitors(&nodes[0], 1);
nodes[0].node.handle_commitment_signed_batch_test(node_id_1, &commit_sig);
check_added_monitors(&nodes[0], 1);

chanmon_cfgs[1].persister.set_update_ret(ChannelMonitorUpdateStatus::InProgress);
let msgs = nodes[0].node.get_and_clear_pending_msg_events();
if let MessageSendEvent::SendRevokeAndACK { msg, .. } = &msgs[0] {
nodes[1].node.handle_revoke_and_ack(node_id_0, &msg);
check_added_monitors(&nodes[1], 1);
} else {
panic!();
}
if let MessageSendEvent::SendStfu { msg, .. } = &msgs[1] {
nodes[1].node.handle_stfu(node_id_0, &msg);
} else {
panic!();
}

nodes[0].node.exit_quiescence(&node_id_1, &chan_id).unwrap();
nodes[1].node.exit_quiescence(&node_id_0, &chan_id).unwrap();
}

#[test]
fn test_quiescence_updates_go_to_holding_cell() {
quiescence_updates_go_to_holding_cell(false);
Expand Down
Loading