pub struct ChannelManager<Signer: Sign, M: Deref, T: Deref, K: Deref, F: Deref, L: Deref> where
    M::Target: Watch<Signer>,
    T::Target: BroadcasterInterface,
    K::Target: KeysInterface<Signer = Signer>,
    F::Target: FeeEstimator,
    L::Target: Logger
{ /* private fields */ }
Expand description

Manager which keeps track of a number of channels and sends messages to the appropriate channel, also tracking HTLC preimages and forwarding onion packets appropriately.

Implements ChannelMessageHandler, handling the multi-channel parts and passing things through to individual Channels.

Implements Writeable to write out all channel state to disk. Implies peer_disconnected() for all peers during write/read (though does not modify this instance, only the instance being serialized). This will result in any channels which have not yet exchanged funding_created (ie called funding_transaction_generated for outbound channels).

Note that you can be a bit lazier about writing out ChannelManager than you can be with ChannelMonitors. With ChannelMonitors you MUST write each monitor update out to disk before returning from chain::Watch::watch_/update_channel, with ChannelManagers, writing updates happens out-of-band (and will prevent any other ChannelManager operations from occurring during the serialization process). If the deserialized version is out-of-date compared to the ChannelMonitors passed by reference to read(), those channels will be force-closed based on the ChannelMonitor state and no funds will be lost (mod on-chain transaction fees).

Note that the deserializer is only implemented for (BlockHash, ChannelManager), which tells you the last block hash which was block_connect()ed. You MUST rescan any blocks along the “reorg path” (ie call block_disconnected() until you get to a common block and then call block_connected() to step towards your best block) upon deserialization before using the object!

Note that ChannelManager is responsible for tracking liveness of its channels and generating ChannelUpdate messages informing peers that the channel is temporarily disabled. To avoid spam due to quick disconnection/reconnection, updates are not sent until the channel has been offline for a full minute. In order to track this, you must call timer_tick_occurred roughly once per minute, though it doesn’t have to be perfect.

Rather than using a plain ChannelManager, it is preferable to use either a SimpleArcChannelManager a SimpleRefChannelManager, for conciseness. See their documentation for more details, but essentially you should default to using a SimpleRefChannelManager, and use a SimpleArcChannelManager when you require a ChannelManager with a static lifetime, such as when you’re using lightning-net-tokio.

Implementations

Constructs a new ChannelManager to hold several channels and route between them.

This is the main “logic hub” for all channel-related actions, and implements ChannelMessageHandler.

Non-proportional fees are fixed according to our risk using the provided fee estimator.

panics if channel_value_satoshis is >= MAX_FUNDING_SATOSHIS!

Users need to notify the new ChannelManager when a new block is connected or disconnected using its block_connected and block_disconnected methods, starting from after params.latest_hash.

Gets the current configuration applied to all new channels, as

Creates a new outbound channel to the given remote node and with the given value.

user_channel_id will be provided back as in Event::FundingGenerationReady::user_channel_id to allow tracking of which events correspond with which create_channel call. Note that the user_channel_id defaults to 0 for inbound channels, so you may wish to avoid using 0 for user_channel_id here. user_channel_id has no meaning inside of LDK, it is simply copied to events and otherwise ignored.

Raises APIError::APIMisuseError when channel_value_satoshis > 2**24 or push_msat is greater than channel_value_satoshis * 1k or channel_value_satoshis < 1000.

Note that we do not check if you are currently connected to the given peer. If no connection is available, the outbound open_channel message may fail to send, resulting in the channel eventually being silently forgotten (dropped on reload).

Returns the new Channel’s temporary channel_id. This ID will appear as Event::FundingGenerationReady::temporary_channel_id and in ChannelDetails::channel_id until after ChannelManager::funding_transaction_generated is called, swapping the Channel’s ID for one derived from the funding transaction’s TXID. If the counterparty rejects the channel immediately, this temporary ID will appear in Event::ChannelClosed::channel_id.

Gets the list of open channels, in random order. See ChannelDetail field documentation for more information.

Gets the list of usable channels, in random order. Useful as an argument to get_route to ensure non-announced channels are used.

These are guaranteed to have their ChannelDetails::is_usable value set to true, see the documentation for ChannelDetails::is_usable for more info on exactly what the criteria are.

Begins the process of closing a channel. After this call (plus some timeout), no new HTLCs will be accepted on the given channel, and after additional timeout/the closing of all pending HTLCs, the channel will be closed on chain.

  • If we are the channel initiator, we will pay between our Background and ChannelConfig::force_close_avoidance_max_fee_satoshis plus our Normal fee estimate.
  • If our counterparty is the channel initiator, we will require a channel closing transaction feerate of at least our Background feerate or the feerate which would appear on a force-closure transaction, whichever is lower. We will allow our counterparty to pay as much fee as they’d like, however.

May generate a SendShutdown message event on success, which should be relayed.

Begins the process of closing a channel. After this call (plus some timeout), no new HTLCs will be accepted on the given channel, and after additional timeout/the closing of all pending HTLCs, the channel will be closed on chain.

target_feerate_sat_per_1000_weight has different meanings depending on if we initiated the channel being closed or not:

  • If we are the channel initiator, we will pay at least this feerate on the closing transaction. The upper-bound is set by ChannelConfig::force_close_avoidance_max_fee_satoshis plus our Normal fee estimate (or target_feerate_sat_per_1000_weight, if it is greater).
  • If our counterparty is the channel initiator, we will refuse to accept a channel closure transaction feerate below target_feerate_sat_per_1000_weight (or the feerate which will appear on a force-closure transaction, whichever is lower).

May generate a SendShutdown message event on success, which should be relayed.

Force closes a channel, immediately broadcasting the latest local commitment transaction to the chain and rejecting new HTLCs on the given channel. Fails if channel_id is unknown to the manager.

Force close all channels, immediately broadcasting the latest local commitment transaction for each to the chain and rejecting new HTLCs on each.

Sends a payment along a given route.

Value parameters are provided via the last hop in route, see documentation for RouteHop fields for more info.

Note that if the payment_hash already exists elsewhere (eg you’re sending a duplicative payment), we don’t do anything to stop you! We always try to ensure that if the provided next hop knows the preimage to payment_hash they can claim an additional amount as specified in the last hop in the route! Thus, you should probably do your own payment_preimage tracking (which you should already be doing as they represent “proof of payment”) and prevent double-sends yourself.

May generate SendHTLCs message(s) event on success, which should be relayed.

Each path may have a different return value, and PaymentSendValue may return a Vec with each entry matching the corresponding-index entry in the route paths, see PaymentSendFailure for more info.

In general, a path may raise:

  • APIError::RouteError when an invalid route or forwarding parameter (cltv_delta, fee, node public key) is specified.
  • APIError::ChannelUnavailable if the next-hop channel is not available for updates (including due to previous monitor update failure or new permanent monitor update failure).
  • APIError::MonitorUpdateFailed if a new monitor update failure prevented sending the relevant updates.

Note that depending on the type of the PaymentSendFailure the HTLC may have been irrevocably committed to on our end. In such a case, do NOT retry the payment with a different route unless you intend to pay twice!

payment_secret is unrelated to payment_hash (or PaymentPreimage) and exists to authenticate the sender to the recipient and prevent payment-probing (deanonymization) attacks. For newer nodes, it will be provided to you in the invoice. If you do not have one, the Route must not contain multiple paths as multi-path payments require a recipient-provided payment_secret. If a payment_secret is provided, we assume that the invoice had the payment_secret feature bit set (either as required or as available). If multiple paths are present in the Route, we assume the invoice had the basic_mpp feature set.

Retries a payment along the given Route.

Errors returned are a superset of those returned from send_payment, so see send_payment documentation for more details on errors. This method will also error if the retry amount puts the payment more than 10% over the payment’s total amount, if the payment for the given payment_id cannot be found (likely due to timeout or success), or if further retries have been disabled with abandon_payment.

Signals that no further retries for the given payment will occur.

After this method returns, any future calls to retry_payment for the given payment_id will fail with PaymentSendFailure::ParameterError. If no such event has been generated, an Event::PaymentFailed event will be generated as soon as there are no remaining pending HTLCs for this payment.

Note that calling this method does not prevent a payment from succeeding. You must still wait until you receive either a Event::PaymentFailed or Event::PaymentSent event to determine the ultimate status of a payment.

Send a spontaneous payment, which is a payment that does not require the recipient to have generated an invoice. Optionally, you may specify the preimage. If you do choose to specify the preimage, it must be a cryptographically secure random value that no intermediate node would be able to guess – otherwise, an intermediate node may claim the payment and it will never reach the recipient.

See send_payment documentation for more details on the return value of this function.

Similar to regular payments, you MUST NOT reuse a payment_preimage value. See send_payment for more information about the risks of duplicate preimage usage.

Note that route must have exactly one path.

Call this upon creation of a funding transaction for the given channel.

Returns an APIError::APIMisuseError if the funding_transaction spent non-SegWit outputs or if no output was found which matches the parameters in Event::FundingGenerationReady.

Returns APIError::ChannelUnavailable if a funding transaction has already been provided for the channel or if the channel has been closed as indicated by Event::ChannelClosed.

May panic if the output found in the funding transaction is duplicative with some other channel (note that this should be trivially prevented by using unique funding transaction keys per-channel).

Do NOT broadcast the funding transaction yourself. When we have safely received our counterparty’s signature the funding transaction will automatically be broadcast via the BroadcasterInterface provided when this ChannelManager was constructed.

Note that this includes RBF or similar transaction replacement strategies - lightning does not currently support replacing a funding transaction on an existing channel. Instead, create a new channel with a conflicting funding transaction.

Regenerates channel_announcements and generates a signed node_announcement from the given arguments, providing them in corresponding events via get_and_clear_pending_msg_events, if at least one public channel has been confirmed on-chain. This effectively re-broadcasts all channel announcements and sends our node announcement to ensure that the lightning P2P network is aware of the channels we have and our network addresses.

rgb is a node “color” and alias is a printable human-readable string to describe this node to humans. They carry no in-protocol meaning.

addresses represent the set (possibly empty) of socket addresses on which this node accepts incoming connections. These will be included in the node_announcement, publicly tying these addresses together and to this node. If you wish to preserve user privacy, addresses should likely contain only Tor Onion addresses.

Panics if addresses is absurdly large (more than 500).

Processes HTLCs which are pending waiting on random forward delay.

Should only really ever be called in response to a PendingHTLCsForwardable event. Will likely generate further events.

Performs actions which should happen on startup and roughly once per minute thereafter.

This currently includes:

  • Increasing or decreasing the on-chain feerate estimates for our outbound channels,
  • Broadcasting ChannelUpdate messages if we’ve been disconnected from our peer for more than a minute, informing the network that they should no longer attempt to route over the channel.

Note that this may cause reentrancy through chain::Watch::update_channel calls or feerate estimate fetches.

Indicates that the preimage for payment_hash is unknown or the received amount is incorrect after a PaymentReceived event, failing the HTLC back to its origin and freeing resources along the path (including in our own channel on which we received it). Returns false if no payment was found to fail backwards, true if the process of failing the HTLC backwards has been started.

Provides a payment preimage in response to Event::PaymentReceived, generating any MessageSendEvents needed to claim the payment.

Note that if you did not set an amount_msat when calling create_inbound_payment or create_inbound_payment_for_hash you must check that the amount in the PaymentReceived event matches your expectation. If you fail to do so and call this method, you may provide the sender “proof-of-payment” when they did not fulfill the full expected payment.

Returns whether any HTLCs were claimed, and thus if any new MessageSendEvents are now pending for processing via get_and_clear_pending_msg_events.

Gets the node_id held by this ChannelManager

Called to accept a request to open a channel after Event::OpenChannelRequest has been triggered.

The temporary_channel_id parameter indicates which inbound channel should be accepted.

Gets a payment secret and payment hash for use in an invoice given to a third party wishing to pay us.

This differs from create_inbound_payment_for_hash only in that it generates the PaymentHash and PaymentPreimage for you.

The PaymentPreimage will ultimately be returned to you in the PaymentReceived, which will have the PaymentReceived::payment_preimage field filled in. That should then be passed directly to claim_funds.

See create_inbound_payment_for_hash for detailed documentation on behavior and requirements.

Note that a malicious eavesdropper can intuit whether an inbound payment was created by create_inbound_payment or create_inbound_payment_for_hash based on runtime.

Note

If you register an inbound payment with this method, then serialize the ChannelManager, then deserialize it with a node running 0.0.103 and earlier, the payment will fail to be received.

Errors if min_value_msat is greater than total bitcoin supply.

👎 Deprecated

Legacy version of create_inbound_payment. Use this method if you wish to share serialized state with LDK node(s) running 0.0.103 and earlier.

Note

This method is deprecated and will be removed soon.

Gets a PaymentSecret for a given PaymentHash, for which the payment preimage is stored external to LDK.

A PaymentReceived event will only be generated if the PaymentSecret matches a payment secret fetched via this method or create_inbound_payment, and which is at least the min_value_msat provided here, if one is provided.

The PaymentHash (and corresponding PaymentPreimage) should be globally unique, though note that LDK will not stop you from registering duplicate payment hashes for inbound payments.

min_value_msat should be set if the invoice being generated contains a value. Any payment received for the returned PaymentHash will be required to be at least min_value_msat before a PaymentReceived event will be generated, ensuring that we do not provide the sender “proof-of-payment” unless they have paid the required amount.

invoice_expiry_delta_secs describes the number of seconds that the invoice is valid for in excess of the current time. This should roughly match the expiry time set in the invoice. After this many seconds, we will remove the inbound payment, resulting in any attempts to pay the invoice failing. The BOLT spec suggests 3,600 secs as a default validity time for invoices when no timeout is set.

Note that we use block header time to time-out pending inbound payments (with some margin to compensate for the inaccuracy of block header timestamps). Thus, in practice we will accept a payment and generate a PaymentReceived event for some time after the expiry. If you need exact expiry semantics, you should enforce them upon receipt of PaymentReceived.

May panic if invoice_expiry_delta_secs is greater than one year.

Note that invoices generated for inbound payments should have their min_final_cltv_expiry set to at least MIN_FINAL_CLTV_EXPIRY.

Note that a malicious eavesdropper can intuit whether an inbound payment was created by create_inbound_payment or create_inbound_payment_for_hash based on runtime.

Note

If you register an inbound payment with this method, then serialize the ChannelManager, then deserialize it with a node running 0.0.103 and earlier, the payment will fail to be received.

Errors if min_value_msat is greater than total bitcoin supply.

👎 Deprecated

Legacy version of create_inbound_payment_for_hash. Use this method if you wish to share serialized state with LDK node(s) running 0.0.103 and earlier.

Note

This method is deprecated and will be removed soon.

Gets an LDK-generated payment preimage from a payment hash and payment secret that were previously returned from create_inbound_payment.

Gets a fake short channel id for use in receiving phantom node payments. These fake scids are used when constructing the phantom invoice’s route hints.

Gets route hints for use in receiving phantom node payments.

This is supported on crate feature std only.

Blocks until ChannelManager needs to be persisted or a timeout is reached. It returns a bool indicating whether persistence is necessary. Only one listener on await_persistable_update or await_persistable_update_timeout is guaranteed to be woken up.

Note that this method is not available with the no-std feature.

Blocks until ChannelManager needs to be persisted. Only one listener on await_persistable_update or await_persistable_update_timeout is guaranteed to be woken up.

Gets the latest best block which was connected either via the chain::Listen or chain::Confirm interfaces.

Trait Implementations

Handle an incoming open_channel message from the given peer.

Handle an incoming accept_channel message from the given peer.

Handle an incoming funding_created message from the given peer.

Handle an incoming funding_signed message from the given peer.

Handle an incoming funding_locked message from the given peer.

Handle an incoming shutdown message from the given peer.

Handle an incoming closing_signed message from the given peer.

Handle an incoming update_add_htlc message from the given peer.

Handle an incoming update_fulfill_htlc message from the given peer.

Handle an incoming update_fail_htlc message from the given peer.

Handle an incoming update_fail_malformed_htlc message from the given peer.

Handle an incoming commitment_signed message from the given peer.

Handle an incoming revoke_and_ack message from the given peer.

Handle an incoming update_fee message from the given peer.

Handle an incoming announcement_signatures message from the given peer.

Handle an incoming channel update from the given peer.

Handle an incoming channel_reestablish message from the given peer.

Indicates a connection to the peer failed/an existing connection was lost. If no connection is believed to be possible in the future (eg they’re sending us messages we don’t understand or indicate they require unknown feature bits), no_connection_possible is set and any outstanding channels should be failed. Read more

Handle a peer reconnecting, possibly generating channel_reestablish message(s).

Handle an incoming error message from the given peer.

Processes transactions confirmed in a block with a given header and height. Read more

Processes an update to the best header connected at the given height. Read more

Returns transactions that should be monitored for reorganization out of the chain. Read more

Processes a transaction that is no longer confirmed as result of a chain reorganization. Read more

Processes events that must be periodically handled.

An EventHandler may safely call back to the provider in order to handle an event. However, it must not call Writeable::write as doing so would result in a deadlock.

Pending events are persisted as part of ChannelManager. While these events are cleared when processed, an EventHandler must be able to handle previously seen events when restarting from an old state.

Notifies the listener that a block was added at the given height.

Notifies the listener that a block was removed at the given height.

Gets the list of pending events which were generated by previous actions, clearing the list in the process. Read more

Writes self out to the given Writer

Writes self out to a Vec

Writes self out to a Vec

Gets the length of this object after it has been serialized. This can be overridden to optimize cases where we prepend an object with its length. Read more

Auto Trait Implementations

Blanket Implementations

Gets the TypeId of self. Read more

Immutably borrows from an owned value. Read more

Mutably borrows from an owned value. Read more

Returns the argument unchanged.

Calls U::from(self).

That is, this conversion is whatever the implementation of From<T> for U chooses to do.

The type returned in the event of a conversion error.

Performs the conversion.

The type returned in the event of a conversion error.

Performs the conversion.