Skip to content

Commit c1125f0

Browse files
committed
Remove MonitorUpdateId from persist trait
MonitorUpdateId was an opaque abstraction for id's generated by UpdateOrigin:Offchain and UpdateOrigin::ChainSync monitor updates. It was mainly needed to map calls made to ChainMonitor::channel_monitor_updated. We no longer track UpdateOrigin::ChainSync MonitorUpdates and can directly use ChannelMonitor::get_latest_update_id() for tracking UpdateOrigin::Offchain monitor updates.
1 parent 7de602a commit c1125f0

File tree

5 files changed

+91
-146
lines changed

5 files changed

+91
-146
lines changed

fuzz/src/utils/test_persister.rs

Lines changed: 2 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,5 @@
11
use lightning::chain;
22
use lightning::chain::{chainmonitor, channelmonitor};
3-
use lightning::chain::chainmonitor::MonitorUpdateId;
43
use lightning::chain::transaction::OutPoint;
54
use lightning::util::test_channel_signer::TestChannelSigner;
65

@@ -10,11 +9,11 @@ pub struct TestPersister {
109
pub update_ret: Mutex<chain::ChannelMonitorUpdateStatus>,
1110
}
1211
impl chainmonitor::Persist<TestChannelSigner> for TestPersister {
13-
fn persist_new_channel(&self, _funding_txo: OutPoint, _data: &channelmonitor::ChannelMonitor<TestChannelSigner>, _update_id: MonitorUpdateId) -> chain::ChannelMonitorUpdateStatus {
12+
fn persist_new_channel(&self, _funding_txo: OutPoint, _data: &channelmonitor::ChannelMonitor<TestChannelSigner>) -> chain::ChannelMonitorUpdateStatus {
1413
self.update_ret.lock().unwrap().clone()
1514
}
1615

17-
fn update_persisted_channel(&self, _funding_txo: OutPoint, _update: Option<&channelmonitor::ChannelMonitorUpdate>, _data: &channelmonitor::ChannelMonitor<TestChannelSigner>, _update_id: MonitorUpdateId) -> chain::ChannelMonitorUpdateStatus {
16+
fn update_persisted_channel(&self, _funding_txo: OutPoint, _update: Option<&channelmonitor::ChannelMonitorUpdate>, _data: &channelmonitor::ChannelMonitor<TestChannelSigner>) -> chain::ChannelMonitorUpdateStatus {
1817
self.update_ret.lock().unwrap().clone()
1918
}
2019

lightning-persister/src/fs_store.rs

Lines changed: 1 addition & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -448,8 +448,6 @@ mod tests {
448448
nodes[1].node.force_close_broadcasting_latest_txn(&chan.2, &nodes[0].node.get_our_node_id()).unwrap();
449449
check_closed_event!(nodes[1], 1, ClosureReason::HolderForceClosed, [nodes[0].node.get_our_node_id()], 100000);
450450
let mut added_monitors = nodes[1].chain_monitor.added_monitors.lock().unwrap();
451-
let update_map = nodes[1].chain_monitor.latest_monitor_update_id.lock().unwrap();
452-
let update_id = update_map.get(&added_monitors[0].1.channel_id()).unwrap();
453451

454452
// Set the store's directory to read-only, which should result in
455453
// returning an unrecoverable failure when we then attempt to persist a
@@ -463,7 +461,7 @@ mod tests {
463461
txid: Txid::from_str("8984484a580b825b9972d7adb15050b3ab624ccd731946b3eeddb92f4e7ef6be").unwrap(),
464462
index: 0
465463
};
466-
match store.persist_new_channel(test_txo, &added_monitors[0].1, update_id.2) {
464+
match store.persist_new_channel(test_txo, &added_monitors[0].1) {
467465
ChannelMonitorUpdateStatus::UnrecoverableError => {},
468466
_ => panic!("unexpected result from persisting new channel")
469467
}

lightning/src/chain/chainmonitor.rs

Lines changed: 63 additions & 111 deletions
Original file line numberDiff line numberDiff line change
@@ -47,46 +47,6 @@ use core::ops::Deref;
4747
use core::sync::atomic::{AtomicUsize, Ordering};
4848
use bitcoin::secp256k1::PublicKey;
4949

50-
mod update_origin {
51-
#[derive(Debug, Clone, Copy, Hash, PartialEq, Eq)]
52-
/// A specific update's ID stored in a `MonitorUpdateId`, separated out to make the contents
53-
/// entirely opaque.
54-
pub(crate) enum UpdateOrigin {
55-
/// An update that was generated by the `ChannelManager` (via our [`crate::chain::Watch`]
56-
/// implementation). This corresponds to an actual [ChannelMonitorUpdate::update_id] field
57-
/// and [ChannelMonitor::get_latest_update_id].
58-
///
59-
/// [ChannelMonitor::get_latest_update_id]: crate::chain::channelmonitor::ChannelMonitor::get_latest_update_id
60-
/// [ChannelMonitorUpdate::update_id]: crate::chain::channelmonitor::ChannelMonitorUpdate::update_id
61-
OffChain(u64),
62-
/// An update that was generated during blockchain processing. The ID here is specific to the
63-
/// generating [ChannelMonitor] and does *not* correspond to any on-disk IDs.
64-
///
65-
/// [ChannelMonitor]: crate::chain::channelmonitor::ChannelMonitor
66-
ChainSync(u64),
67-
}
68-
}
69-
70-
#[cfg(any(feature = "_test_utils", test))]
71-
pub(crate) use update_origin::UpdateOrigin;
72-
#[cfg(not(any(feature = "_test_utils", test)))]
73-
use update_origin::UpdateOrigin;
74-
75-
/// An opaque identifier describing a specific [`Persist`] method call.
76-
#[derive(Debug, Clone, Copy, Hash, PartialEq, Eq)]
77-
pub struct MonitorUpdateId {
78-
pub(crate) contents: UpdateOrigin,
79-
}
80-
81-
impl MonitorUpdateId {
82-
pub(crate) fn from_monitor_update(update: &ChannelMonitorUpdate) -> Self {
83-
Self { contents: UpdateOrigin::OffChain(update.update_id) }
84-
}
85-
pub(crate) fn from_new_monitor<ChannelSigner: WriteableEcdsaChannelSigner>(monitor: &ChannelMonitor<ChannelSigner>) -> Self {
86-
Self { contents: UpdateOrigin::OffChain(monitor.get_latest_update_id()) }
87-
}
88-
}
89-
9050
/// `Persist` defines behavior for persisting channel monitors: this could mean
9151
/// writing once to disk, and/or uploading to one or more backup services.
9252
///
@@ -119,7 +79,7 @@ impl MonitorUpdateId {
11979
/// All calls should generally spawn a background task and immediately return
12080
/// [`ChannelMonitorUpdateStatus::InProgress`]. Once the update completes,
12181
/// [`ChainMonitor::channel_monitor_updated`] should be called with the corresponding
122-
/// [`MonitorUpdateId`].
82+
/// [`ChannelMonitor::get_latest_update_id`] or [`ChannelMonitorUpdate::update_id`].
12383
///
12484
/// Note that unlike the direct [`chain::Watch`] interface,
12585
/// [`ChainMonitor::channel_monitor_updated`] must be called once for *each* update which occurs.
@@ -150,15 +110,16 @@ pub trait Persist<ChannelSigner: WriteableEcdsaChannelSigner> {
150110
/// channel's outpoint (and it is up to you to maintain a correct mapping between the outpoint
151111
/// and the stored channel data). Note that you **must** persist every new monitor to disk.
152112
///
153-
/// The `update_id` is used to identify this call to [`ChainMonitor::channel_monitor_updated`],
154-
/// if you return [`ChannelMonitorUpdateStatus::InProgress`].
113+
/// The [`ChannelMonitor::get_latest_update_id`] uniquely links this call to [`ChainMonitor::channel_monitor_updated`].
114+
/// For [`Persist::persist_new_channel`], it is only necessary to call [`ChainMonitor::channel_monitor_updated`]
115+
/// when you return [`ChannelMonitorUpdateStatus::InProgress`].
155116
///
156117
/// See [`Writeable::write`] on [`ChannelMonitor`] for writing out a `ChannelMonitor`
157118
/// and [`ChannelMonitorUpdateStatus`] for requirements when returning errors.
158119
///
159120
/// [`ChannelManager`]: crate::ln::channelmanager::ChannelManager
160121
/// [`Writeable::write`]: crate::util::ser::Writeable::write
161-
fn persist_new_channel(&self, channel_funding_outpoint: OutPoint, data: &ChannelMonitor<ChannelSigner>, update_id: MonitorUpdateId) -> ChannelMonitorUpdateStatus;
122+
fn persist_new_channel(&self, channel_funding_outpoint: OutPoint, monitor: &ChannelMonitor<ChannelSigner>) -> ChannelMonitorUpdateStatus;
162123

163124
/// Update one channel's data. The provided [`ChannelMonitor`] has already applied the given
164125
/// update.
@@ -185,15 +146,17 @@ pub trait Persist<ChannelSigner: WriteableEcdsaChannelSigner> {
185146
/// them in batches. The size of each monitor grows `O(number of state updates)`
186147
/// whereas updates are small and `O(1)`.
187148
///
188-
/// The `update_id` is used to identify this call to [`ChainMonitor::channel_monitor_updated`],
189-
/// if you return [`ChannelMonitorUpdateStatus::InProgress`].
149+
/// The [`ChannelMonitorUpdate::update_id`] or [`ChannelMonitor::get_latest_update_id`] uniquely
150+
/// links this call to [`ChainMonitor::channel_monitor_updated`].
151+
/// For [`Persist::update_persisted_channel`], it is only necessary to call [`ChainMonitor::channel_monitor_updated`]
152+
/// when an [`ChannelMonitorUpdate`] is provided and when you return [`ChannelMonitorUpdateStatus::InProgress`].
190153
///
191154
/// See [`Writeable::write`] on [`ChannelMonitor`] for writing out a `ChannelMonitor`,
192155
/// [`Writeable::write`] on [`ChannelMonitorUpdate`] for writing out an update, and
193156
/// [`ChannelMonitorUpdateStatus`] for requirements when returning errors.
194157
///
195158
/// [`Writeable::write`]: crate::util::ser::Writeable::write
196-
fn update_persisted_channel(&self, channel_funding_outpoint: OutPoint, update: Option<&ChannelMonitorUpdate>, data: &ChannelMonitor<ChannelSigner>, update_id: MonitorUpdateId) -> ChannelMonitorUpdateStatus;
159+
fn update_persisted_channel(&self, channel_funding_outpoint: OutPoint, monitor_update: Option<&ChannelMonitorUpdate>, monitor: &ChannelMonitor<ChannelSigner>) -> ChannelMonitorUpdateStatus;
197160
/// Prevents the channel monitor from being loaded on startup.
198161
///
199162
/// Archiving the data in a backup location (rather than deleting it fully) is useful for
@@ -209,13 +172,12 @@ struct MonitorHolder<ChannelSigner: WriteableEcdsaChannelSigner> {
209172
/// update_persisted_channel, the user returns a
210173
/// [`ChannelMonitorUpdateStatus::InProgress`], and then calls channel_monitor_updated
211174
/// immediately, racing our insertion of the pending update into the contained Vec.
212-
pending_monitor_updates: Mutex<Vec<MonitorUpdateId>>,
175+
pending_monitor_updates: Mutex<Vec<u64>>,
213176
}
214177

215178
impl<ChannelSigner: WriteableEcdsaChannelSigner> MonitorHolder<ChannelSigner> {
216-
fn has_pending_offchain_updates(&self, pending_monitor_updates_lock: &MutexGuard<Vec<MonitorUpdateId>>) -> bool {
217-
pending_monitor_updates_lock.iter().any(|update_id|
218-
if let UpdateOrigin::OffChain(_) = update_id.contents { true } else { false })
179+
fn has_pending_updates(&self, pending_monitor_updates_lock: &MutexGuard<Vec<u64>>) -> bool {
180+
!pending_monitor_updates_lock.is_empty()
219181
}
220182
}
221183

@@ -259,7 +221,7 @@ pub struct ChainMonitor<ChannelSigner: WriteableEcdsaChannelSigner, C: Deref, T:
259221
P::Target: Persist<ChannelSigner>,
260222
{
261223
monitors: RwLock<HashMap<OutPoint, MonitorHolder<ChannelSigner>>>,
262-
/// When we generate a [`MonitorUpdateId`] for a chain-event monitor persistence, we need a
224+
/// When we generate a monitor update for a chain-event monitor persistence, we need a
263225
/// unique ID, which we calculate by simply getting the next value from this counter. Note that
264226
/// the ID is never persisted so it's ok that they reset on restart.
265227
sync_persistence_id: AtomicCounter,
@@ -346,20 +308,11 @@ where C::Target: chain::Filter,
346308
let mut txn_outputs;
347309
{
348310
txn_outputs = process(monitor, txdata);
349-
let chain_sync_update_id = self.sync_persistence_id.get_increment();
350-
let update_id = MonitorUpdateId {
351-
contents: UpdateOrigin::ChainSync(chain_sync_update_id),
352-
};
353-
354-
log_trace!(logger, "Syncing Channel Monitor for channel {} for block-data update_id {}",
355-
log_funding_info!(monitor),
356-
chain_sync_update_id
357-
);
358-
match self.persister.update_persisted_channel(*funding_outpoint, None, monitor, update_id) {
311+
log_trace!(logger, "Syncing Channel Monitor for channel {}", log_funding_info!(monitor));
312+
match self.persister.update_persisted_channel(*funding_outpoint, None, monitor) {
359313
ChannelMonitorUpdateStatus::Completed =>
360-
log_trace!(logger, "Finished syncing Channel Monitor for channel {} for block-data update_id {}",
361-
log_funding_info!(monitor),
362-
chain_sync_update_id
314+
log_trace!(logger, "Finished syncing Channel Monitor for channel {} for block-data",
315+
log_funding_info!(monitor)
363316
),
364317
ChannelMonitorUpdateStatus::InProgress => {
365318
log_debug!(logger, "Channel Monitor sync for channel {} in progress.", log_funding_info!(monitor));
@@ -464,15 +417,21 @@ where C::Target: chain::Filter,
464417

465418
#[cfg(not(c_bindings))]
466419
/// Lists the pending updates for each [`ChannelMonitor`] (by `OutPoint` being monitored).
467-
pub fn list_pending_monitor_updates(&self) -> HashMap<OutPoint, Vec<MonitorUpdateId>> {
420+
/// Each `Vec<u64>` contains `update_id`s from [`ChannelMonitor::get_latest_update_id`] for updates
421+
/// that have not yet been fully persisted. Note that if a full monitor is persisted all the pending
422+
/// monitor updates must be individually marked completed by calling [`ChainMonitor::channel_monitor_updated`].
423+
pub fn list_pending_monitor_updates(&self) -> HashMap<OutPoint, Vec<u64>> {
468424
hash_map_from_iter(self.monitors.read().unwrap().iter().map(|(outpoint, holder)| {
469425
(*outpoint, holder.pending_monitor_updates.lock().unwrap().clone())
470426
}))
471427
}
472428

473429
#[cfg(c_bindings)]
474430
/// Lists the pending updates for each [`ChannelMonitor`] (by `OutPoint` being monitored).
475-
pub fn list_pending_monitor_updates(&self) -> Vec<(OutPoint, Vec<MonitorUpdateId>)> {
431+
/// Each `Vec<u64>` contains `update_id`s from [`ChannelMonitor::get_latest_update_id`] for updates
432+
/// that have not yet been fully persisted. Note that if a full monitor is persisted all the pending
433+
/// monitor updates must be individually marked completed by calling [`ChainMonitor::channel_monitor_updated`].
434+
pub fn list_pending_monitor_updates(&self) -> Vec<(OutPoint, Vec<u64>)> {
476435
self.monitors.read().unwrap().iter().map(|(outpoint, holder)| {
477436
(*outpoint, holder.pending_monitor_updates.lock().unwrap().clone())
478437
}).collect()
@@ -491,56 +450,49 @@ where C::Target: chain::Filter,
491450
/// 1) This [`ChainMonitor`] calls [`Persist::update_persisted_channel`] which stores the
492451
/// update to disk and begins updating any remote (e.g. watchtower/backup) copies,
493452
/// returning [`ChannelMonitorUpdateStatus::InProgress`],
494-
/// 2) once all remote copies are updated, you call this function with the
495-
/// `completed_update_id` that completed, and once all pending updates have completed the
496-
/// channel will be re-enabled.
497-
// Note that we re-enable only after `UpdateOrigin::OffChain` updates complete, we don't
498-
// care about `UpdateOrigin::ChainSync` updates for the channel state being updated. We
499-
// only care about `UpdateOrigin::ChainSync` for returning `MonitorEvent`s.
453+
/// 2) once all remote copies are updated, you call this function with [`ChannelMonitor::get_latest_update_id`]
454+
/// or [`ChannelMonitorUpdate::update_id`] as the `completed_update_id`, and once all pending
455+
/// updates have completed the channel will be re-enabled.
456+
///
457+
/// It is only necessary to call [`ChainMonitor::channel_monitor_updated`] when you return [`ChannelMonitorUpdateStatus::InProgress`]
458+
/// from [`Persist`] and either:
459+
/// 1. A new [`ChannelMonitor`] was added in [`Persist::persist_new_channel`], or
460+
/// 2. A [`ChannelMonitorUpdate`] was provided as part of [`Persist::update_persisted_channel`].
461+
/// Note that we don't care about calls to [`Persist::update_persisted_channel`] where no
462+
/// [`ChannelMonitorUpdate`] was provided.
500463
///
501464
/// Returns an [`APIError::APIMisuseError`] if `funding_txo` does not match any currently
502465
/// registered [`ChannelMonitor`]s.
503-
pub fn channel_monitor_updated(&self, funding_txo: OutPoint, completed_update_id: MonitorUpdateId) -> Result<(), APIError> {
466+
pub fn channel_monitor_updated(&self, funding_txo: OutPoint, completed_update_id: u64) -> Result<(), APIError> {
504467
let monitors = self.monitors.read().unwrap();
505468
let monitor_data = if let Some(mon) = monitors.get(&funding_txo) { mon } else {
506469
return Err(APIError::APIMisuseError { err: format!("No ChannelMonitor matching funding outpoint {:?} found", funding_txo) });
507470
};
508471
let mut pending_monitor_updates = monitor_data.pending_monitor_updates.lock().unwrap();
509472
pending_monitor_updates.retain(|update_id| *update_id != completed_update_id);
510473

511-
match completed_update_id {
512-
MonitorUpdateId { contents: UpdateOrigin::OffChain(completed_update_id) } => {
513-
// Note that we only check for `UpdateOrigin::OffChain` failures here - if
514-
// we're being told that a `UpdateOrigin::OffChain` monitor update completed,
515-
// we only care about ensuring we don't tell the `ChannelManager` to restore
516-
// the channel to normal operation until all `UpdateOrigin::OffChain` updates
517-
// complete.
518-
// If there's some `UpdateOrigin::ChainSync` update still pending that's okay
519-
// - we can still update our channel state, just as long as we don't return
520-
// `MonitorEvent`s from the monitor back to the `ChannelManager` until they
521-
// complete.
522-
let monitor_is_pending_updates = monitor_data.has_pending_offchain_updates(&pending_monitor_updates);
523-
log_debug!(self.logger, "Completed off-chain monitor update {} for channel with funding outpoint {:?}, {}",
524-
completed_update_id,
525-
funding_txo,
526-
if monitor_is_pending_updates {
527-
"still have pending off-chain updates"
528-
} else {
529-
"all off-chain updates complete, returning a MonitorEvent"
530-
});
531-
if monitor_is_pending_updates {
532-
// If there are still monitor updates pending, we cannot yet construct a
533-
// Completed event.
534-
return Ok(());
535-
}
536-
let channel_id = monitor_data.monitor.channel_id();
537-
self.pending_monitor_events.lock().unwrap().push((funding_txo, channel_id, vec![MonitorEvent::Completed {
538-
funding_txo, channel_id,
539-
monitor_update_id: monitor_data.monitor.get_latest_update_id(),
540-
}], monitor_data.monitor.get_counterparty_node_id()));
541-
},
542-
MonitorUpdateId { contents: UpdateOrigin::ChainSync(_) } => {},
474+
// Note that we only check for pending non-chainsync monitor updates and we don't track monitor
475+
// updates resulting from chainsync in `pending_monitor_updates`.
476+
let monitor_is_pending_updates = monitor_data.has_pending_updates(&pending_monitor_updates);
477+
log_debug!(self.logger, "Completed off-chain monitor update {} for channel with funding outpoint {:?}, {}",
478+
completed_update_id,
479+
funding_txo,
480+
if monitor_is_pending_updates {
481+
"still have pending off-chain updates"
482+
} else {
483+
"all off-chain updates complete, returning a MonitorEvent"
484+
});
485+
if monitor_is_pending_updates {
486+
// If there are still monitor updates pending, we cannot yet construct a
487+
// Completed event.
488+
return Ok(());
543489
}
490+
let channel_id = monitor_data.monitor.channel_id();
491+
self.pending_monitor_events.lock().unwrap().push((funding_txo, channel_id, vec![MonitorEvent::Completed {
492+
funding_txo, channel_id,
493+
monitor_update_id: monitor_data.monitor.get_latest_update_id(),
494+
}], monitor_data.monitor.get_counterparty_node_id()));
495+
544496
self.event_notifier.notify();
545497
Ok(())
546498
}
@@ -771,9 +723,9 @@ where C::Target: chain::Filter,
771723
hash_map::Entry::Vacant(e) => e,
772724
};
773725
log_trace!(logger, "Got new ChannelMonitor for channel {}", log_funding_info!(monitor));
774-
let update_id = MonitorUpdateId::from_new_monitor(&monitor);
726+
let update_id = monitor.get_latest_update_id();
775727
let mut pending_monitor_updates = Vec::new();
776-
let persist_res = self.persister.persist_new_channel(funding_outpoint, &monitor, update_id);
728+
let persist_res = self.persister.persist_new_channel(funding_outpoint, &monitor);
777729
match persist_res {
778730
ChannelMonitorUpdateStatus::InProgress => {
779731
log_info!(logger, "Persistence of new ChannelMonitor for channel {} in progress", log_funding_info!(monitor));
@@ -823,7 +775,7 @@ where C::Target: chain::Filter,
823775
log_trace!(logger, "Updating ChannelMonitor to id {} for channel {}", update.update_id, log_funding_info!(monitor));
824776
let update_res = monitor.update_monitor(update, &self.broadcaster, &self.fee_estimator, &self.logger);
825777

826-
let update_id = MonitorUpdateId::from_monitor_update(update);
778+
let update_id = update.update_id;
827779
let mut pending_monitor_updates = monitor_state.pending_monitor_updates.lock().unwrap();
828780
let persist_res = if update_res.is_err() {
829781
// Even if updating the monitor returns an error, the monitor's state will
@@ -832,9 +784,9 @@ where C::Target: chain::Filter,
832784
// while reading `channel_monitor` with updates from storage. Instead, we should persist
833785
// the entire `channel_monitor` here.
834786
log_warn!(logger, "Failed to update ChannelMonitor for channel {}. Going ahead and persisting the entire ChannelMonitor", log_funding_info!(monitor));
835-
self.persister.update_persisted_channel(funding_txo, None, monitor, update_id)
787+
self.persister.update_persisted_channel(funding_txo, None, monitor)
836788
} else {
837-
self.persister.update_persisted_channel(funding_txo, Some(update), monitor, update_id)
789+
self.persister.update_persisted_channel(funding_txo, Some(update), monitor)
838790
};
839791
match persist_res {
840792
ChannelMonitorUpdateStatus::InProgress => {

0 commit comments

Comments
 (0)