Allow holding ChannelMonitorUpdates until later, completing one

In the coming commits, we need to delay `ChannelMonitorUpdate`s
until future actions (specifically `Event` handling). However,
because we should only notify users once of a given
`ChannelMonitorUpdate` and they must be provided in-order, we need
to track which ones have or have not been given to users and, once
updating resumes, fly the ones that haven't already made it to
users.

To do this we simply add a `bool` in the `ChannelMonitorUpdate` set
stored in the `Channel` which indicates if an update flew and
decline to provide new updates back to the `ChannelManager` if any
updates have their flown bit unset.

Further, because we'll now by releasing `ChannelMonitorUpdate`s
which were already stored in the pending list, we now need to
support getting a `Completed` result for a monitor which isn't the
only pending monitor (or even out of order), thus we also rewrite
the way monitor updates are marked completed.

Author: TheBlueMatt

lightning/src/ln/chanmon_update_fail_tests.rs

@@ -144,7 +144,7 @@ fn test_monitor_and_persister_update_fail() {
 		let mut node_0_per_peer_lock;
 		let mut node_0_peer_state_lock;
 		let mut channel = get_channel_ref!(nodes[0], nodes[1], node_0_per_peer_lock, node_0_peer_state_lock, chan.2);
-		if let Ok(update) = channel.commitment_signed(&updates.commitment_signed, &node_cfgs[0].logger) {
+		if let Ok(Some(update)) = channel.commitment_signed(&updates.commitment_signed, &node_cfgs[0].logger) {
 			// Check that even though the persister is returning a InProgress,
 			// because the update is bogus, ultimately the error that's returned
 			// should be a PermanentFailure.

lightning/src/ln/channel.rs

@@ -400,7 +400,7 @@ pub enum UpdateFulfillCommitFetch<'a> {
 	/// previously placed in the holding cell (and has since been removed).
 	NewClaim {
 		/// The ChannelMonitorUpdate which places the new payment preimage in the channel monitor
-		monitor_update: &'a ChannelMonitorUpdate,
+		monitor_update_opt: Option<&'a ChannelMonitorUpdate>,
 		/// The value of the HTLC which was claimed, in msat.
 		htlc_value_msat: u64,
 	},
@@ -479,6 +479,17 @@ pub(crate) const MIN_AFFORDABLE_HTLC_COUNT: usize = 4;
 ///   * `EXPIRE_PREV_CONFIG_TICKS` = convergence_delay / tick_interval
 pub(crate) const EXPIRE_PREV_CONFIG_TICKS: usize = 5;
 
+struct PendingChannelMonitorUpdate {
+	update: ChannelMonitorUpdate,
+	/// In some cases we need to delay letting the [`ChannelMonitorUpdate`] fly until after an
+	/// `Event` is processed by the user. This bool indicates the [`ChannelMonitorUpdate`] has
+	/// flown and we're waiting to hear back, otherwise the update is waiting on some external
+	/// event and the [`ChannelManager`] will update us when we're ready.
+	///
+	/// [`ChannelManager`]: super::channelmanager::ChannelManager
+	flown: bool,
+}
+
 // TODO: We should refactor this to be an Inbound/OutboundChannel until initial setup handshaking
 // has been completed, and then turn into a Channel to get compiler-time enforcement of things like
 // calling channel_id() before we're set up or things like get_outbound_funding_signed on an
@@ -740,7 +751,7 @@ pub(super) struct Channel<Signer: ChannelSigner> {
 	/// If we then persist the [`channelmanager::ChannelManager`] and crash before the persistence
 	/// completes we still need to be able to complete the persistence. Thus, we have to keep a
 	/// copy of the [`ChannelMonitorUpdate`] here until it is complete.
-	pending_monitor_updates: Vec<ChannelMonitorUpdate>,
+	pending_monitor_updates: Vec<PendingChannelMonitorUpdate>,
 }
 
 #[cfg(any(test, fuzzing))]
@@ -1967,6 +1978,7 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 	}
 
 	pub fn get_update_fulfill_htlc_and_commit<L: Deref>(&mut self, htlc_id: u64, payment_preimage: PaymentPreimage, logger: &L) -> UpdateFulfillCommitFetch where L::Target: Logger {
+		let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
 		match self.get_update_fulfill_htlc(htlc_id, payment_preimage, logger) {
 			UpdateFulfillFetch::NewClaim { mut monitor_update, htlc_value_msat, msg: Some(_) } => {
 				let mut additional_update = self.build_commitment_no_status_check(logger);
@@ -1975,17 +1987,21 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 				self.latest_monitor_update_id = monitor_update.update_id;
 				monitor_update.updates.append(&mut additional_update.updates);
 				self.monitor_updating_paused(false, true, false, Vec::new(), Vec::new(), Vec::new());
-				self.pending_monitor_updates.push(monitor_update);
+				self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+					update: monitor_update, flown: fly_monitor,
+				});
 				UpdateFulfillCommitFetch::NewClaim {
-					monitor_update: self.pending_monitor_updates.last().unwrap(),
+					monitor_update_opt: if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None },
 					htlc_value_msat,
 				}
 			},
 			UpdateFulfillFetch::NewClaim { monitor_update, htlc_value_msat, msg: None } => {
 				self.monitor_updating_paused(false, false, false, Vec::new(), Vec::new(), Vec::new());
-				self.pending_monitor_updates.push(monitor_update);
+				self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+					update: monitor_update, flown: fly_monitor
+				});
 				UpdateFulfillCommitFetch::NewClaim {
-					monitor_update: self.pending_monitor_updates.last().unwrap(),
+					monitor_update_opt: if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None },
 					htlc_value_msat,
 				}
 			}
@@ -3054,7 +3070,7 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 		Ok(())
 	}
 
-	pub fn commitment_signed<L: Deref>(&mut self, msg: &msgs::CommitmentSigned, logger: &L) -> Result<&ChannelMonitorUpdate, ChannelError>
+	pub fn commitment_signed<L: Deref>(&mut self, msg: &msgs::CommitmentSigned, logger: &L) -> Result<Option<&ChannelMonitorUpdate>, ChannelError>
 		where L::Target: Logger
 	{
 		if (self.channel_state & (ChannelState::ChannelReady as u32)) != (ChannelState::ChannelReady as u32) {
@@ -3230,8 +3246,11 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 			}
 			log_debug!(logger, "Received valid commitment_signed from peer in channel {}, updated HTLC state but awaiting a monitor update resolution to reply.",
 				log_bytes!(self.channel_id));
-			self.pending_monitor_updates.push(monitor_update);
-			return Ok(self.pending_monitor_updates.last().unwrap());
+			let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+			self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+				update: monitor_update, flown: fly_monitor
+			});
+			return Ok(if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None });
 		}
 
 		let need_commitment_signed = if need_commitment && (self.channel_state & (ChannelState::AwaitingRemoteRevoke as u32)) == 0 {
@@ -3248,9 +3267,12 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 
 		log_debug!(logger, "Received valid commitment_signed from peer in channel {}, updating HTLC state and responding with{} a revoke_and_ack.",
 			log_bytes!(self.channel_id()), if need_commitment_signed { " our own commitment_signed and" } else { "" });
-		self.pending_monitor_updates.push(monitor_update);
+		let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+		self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+			update: monitor_update, flown: fly_monitor,
+		});
 		self.monitor_updating_paused(true, need_commitment_signed, false, Vec::new(), Vec::new(), Vec::new());
-		return Ok(self.pending_monitor_updates.last().unwrap());
+		return Ok(if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None });
 	}
 
 	/// Public version of the below, checking relevant preconditions first.
@@ -3365,8 +3387,12 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 				update_add_htlcs.len(), update_fulfill_htlcs.len(), update_fail_htlcs.len());
 
 			self.monitor_updating_paused(false, true, false, Vec::new(), Vec::new(), Vec::new());
-			self.pending_monitor_updates.push(monitor_update);
-			(Some(self.pending_monitor_updates.last().unwrap()), htlcs_to_fail)
+			let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+			self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+				update: monitor_update, flown: fly_monitor,
+			});
+			(if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None },
+				htlcs_to_fail)
 		} else {
 			(None, Vec::new())
 		}
@@ -3377,7 +3403,7 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 	/// waiting on this revoke_and_ack. The generation of this new commitment_signed may also fail,
 	/// generating an appropriate error *after* the channel state has been updated based on the
 	/// revoke_and_ack message.
-	pub fn revoke_and_ack<L: Deref>(&mut self, msg: &msgs::RevokeAndACK, logger: &L) -> Result<(Vec<(HTLCSource, PaymentHash)>, &ChannelMonitorUpdate), ChannelError>
+	pub fn revoke_and_ack<L: Deref>(&mut self, msg: &msgs::RevokeAndACK, logger: &L) -> Result<(Vec<(HTLCSource, PaymentHash)>, Option<&ChannelMonitorUpdate>), ChannelError>
 		where L::Target: Logger,
 	{
 		if (self.channel_state & (ChannelState::ChannelReady as u32)) != (ChannelState::ChannelReady as u32) {
@@ -3574,21 +3600,29 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 			self.monitor_pending_failures.append(&mut revoked_htlcs);
 			self.monitor_pending_finalized_fulfills.append(&mut finalized_claimed_htlcs);
 			log_debug!(logger, "Received a valid revoke_and_ack for channel {} but awaiting a monitor update resolution to reply.", log_bytes!(self.channel_id()));
-			self.pending_monitor_updates.push(monitor_update);
-			return Ok((Vec::new(), self.pending_monitor_updates.last().unwrap()));
+			let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+			self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+				update: monitor_update, flown: fly_monitor,
+			});
+			return Ok((Vec::new(),
+				if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }));
 		}
 
 		match self.free_holding_cell_htlcs(logger) {
 			(Some(_), htlcs_to_fail) => {
-				let mut additional_update = self.pending_monitor_updates.pop().unwrap();
+				let mut additional_update = self.pending_monitor_updates.pop().unwrap().update;
 				// free_holding_cell_htlcs may bump latest_monitor_id multiple times but we want them to be
 				// strictly increasing by one, so decrement it here.
 				self.latest_monitor_update_id = monitor_update.update_id;
 				monitor_update.updates.append(&mut additional_update.updates);
 
 				self.monitor_updating_paused(false, true, false, to_forward_infos, revoked_htlcs, finalized_claimed_htlcs);
-				self.pending_monitor_updates.push(monitor_update);
-				Ok((htlcs_to_fail, self.pending_monitor_updates.last().unwrap()))
+				let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+				self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+					update: monitor_update, flown: fly_monitor,
+				});
+				Ok((htlcs_to_fail,
+					if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }))
 			},
 			(None, htlcs_to_fail) => {
 				if require_commitment {
@@ -3602,13 +3636,21 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 					log_debug!(logger, "Received a valid revoke_and_ack for channel {}. Responding with a commitment update with {} HTLCs failed.",
 						log_bytes!(self.channel_id()), update_fail_htlcs.len() + update_fail_malformed_htlcs.len());
 					self.monitor_updating_paused(false, true, false, to_forward_infos, revoked_htlcs, finalized_claimed_htlcs);
-					self.pending_monitor_updates.push(monitor_update);
-					Ok((htlcs_to_fail, self.pending_monitor_updates.last().unwrap()))
+					let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+					self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+						update: monitor_update, flown: fly_monitor,
+					});
+					Ok((htlcs_to_fail,
+						if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }))
 				} else {
 					log_debug!(logger, "Received a valid revoke_and_ack for channel {} with no reply necessary.", log_bytes!(self.channel_id()));
 					self.monitor_updating_paused(false, false, false, to_forward_infos, revoked_htlcs, finalized_claimed_htlcs);
-					self.pending_monitor_updates.push(monitor_update);
-					Ok((htlcs_to_fail, self.pending_monitor_updates.last().unwrap()))
+					let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+					self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+						update: monitor_update, flown: fly_monitor,
+					});
+					Ok((htlcs_to_fail,
+						if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }))
 				}
 			}
 		}
@@ -3797,7 +3839,12 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 	{
 		assert_eq!(self.channel_state & ChannelState::MonitorUpdateInProgress as u32, ChannelState::MonitorUpdateInProgress as u32);
 		self.channel_state &= !(ChannelState::MonitorUpdateInProgress as u32);
-		self.pending_monitor_updates.clear();
+		let mut found_unflown = false;
+		self.pending_monitor_updates.retain(|upd| {
+			if found_unflown { debug_assert!(!upd.flown, "No mons may fly after one is paused"); }
+			if !upd.flown { found_unflown = true; }
+			!upd.flown
+		});
 
 		// If we're past (or at) the FundingSent stage on an outbound channel, try to
 		// (re-)broadcast the funding transaction as we may have declined to broadcast it when we
@@ -4338,8 +4385,11 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 				}],
 			};
 			self.monitor_updating_paused(false, false, false, Vec::new(), Vec::new(), Vec::new());
-			self.pending_monitor_updates.push(monitor_update);
-			Some(self.pending_monitor_updates.last().unwrap())
+			let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+			self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+				update: monitor_update, flown: fly_monitor,
+			});
+			if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }
 		} else { None };
 		let shutdown = if send_shutdown {
 			Some(msgs::Shutdown {
@@ -4889,8 +4939,25 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 		(self.channel_state & ChannelState::MonitorUpdateInProgress as u32) != 0
 	}
 
-	pub fn get_next_monitor_update(&self) -> Option<&ChannelMonitorUpdate> {
-		self.pending_monitor_updates.first()
+	/// Returns the next unflown monitor update, if one exists, and a bool which indicates a
+	/// further unflown monitor update exists after the next.
+	pub fn fly_next_unflown_monitor_update(&mut self) -> Option<(&ChannelMonitorUpdate, bool)> {
+		for i in 0..self.pending_monitor_updates.len() {
+			if !self.pending_monitor_updates[i].flown {
+				self.pending_monitor_updates[i].flown = true;
+				return Some((&self.pending_monitor_updates[i].update,
+					self.pending_monitor_updates.len() > i + 1));
+			}
+		}
+		None
+	}
+
+	pub fn no_monitor_updates_pending(&self) -> bool {
+		self.pending_monitor_updates.is_empty()
+	}
+
+	pub fn complete_one_mon_update(&mut self, update_id: u64) {
+		self.pending_monitor_updates.retain(|upd| upd.update.update_id != update_id);
 	}
 
 	/// Returns true if funding_created was sent/received.
@@ -5930,8 +5997,12 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 			Some(_) => {
 				let monitor_update = self.build_commitment_no_status_check(logger);
 				self.monitor_updating_paused(false, true, false, Vec::new(), Vec::new(), Vec::new());
-				self.pending_monitor_updates.push(monitor_update);
-				Ok(Some(self.pending_monitor_updates.last().unwrap()))
+
+				let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+				self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+					update: monitor_update, flown: fly_monitor,
+				});
+				Ok(if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None })
 			},
 			None => Ok(None)
 		}
@@ -6020,8 +6091,11 @@ impl<Signer: WriteableEcdsaChannelSigner> Channel<Signer> {
 				}],
 			};
 			self.monitor_updating_paused(false, false, false, Vec::new(), Vec::new(), Vec::new());
-			self.pending_monitor_updates.push(monitor_update);
-			Some(self.pending_monitor_updates.last().unwrap())
+			let fly_monitor = self.pending_monitor_updates.iter().all(|upd| upd.flown);
+			self.pending_monitor_updates.push(PendingChannelMonitorUpdate {
+				update: monitor_update, flown: fly_monitor,
+			});
+			if fly_monitor { self.pending_monitor_updates.last().map(|upd| &upd.update) } else { None }
 		} else { None };
 		let shutdown = msgs::Shutdown {
 			channel_id: self.channel_id,