This commit properly removes any/all closed channels from the routing
table. In the current implementation individual links (channels)
between nodes are treated sparely from the PoV of the routing table. In
the future, this behavior should be modified such that, the routing
table views all the links between nodes as a single channel. Such a
change will simplify the task of path finding as the links can simply
be viewed as a channel with the sum of their capacities. The link layer
(htlcSwitch) will handle the details of fragmentation on a local basis.
This commit adds full support for multi-hop onion routed payments
within the daemon.
The switch has been greatly extended in order to gain the functionality
required to manage Sphinx payment circuits amongst active links. A
payment circuit is initiated when a link sends an HTLC add to the
downstream htlcSwitch it received from the upstream peer. The switch
then examines the parsed sphinx packet to set up the clear/settle ends
of the circuit. Created circuits can be re-used amongst HTLC payments
which share the same RHash.
All bandwidth updates within a link’s internal state are now managed
with atomic increments/decrements in order to avoid race conditions
amongst the two goroutines the switch currently uses.
Each channel’s htlcManager has also been extended to parse out the
next-hop contained within Sphinx packets, and construct a proper
htlcPkt such that the htlcSwitch can initiate then manage the payment
circuit.
This commit alters the send/receive HTLC pipe line a bit in order to
fully integrate onion routing into the daemon.
The server now stores the global Sphinx router which all active
htlcManagers will used when processing upstream HTLC add messages.
Currently the onion routing private key is static, and identical to the
node’s current identity public key. In the future this key will be
rotated daily the node based on the current block hash.
When sending a payment via the SendPayment RPC, the routing manager is
now queried for the existence of a route before the payment request is
sent to the HTLC switch. If a path is found, then a Sphinx onion packet
encoding the route is created, then populated within the HTLC add
message.
Finally, when processing an upstream HTLC add request, the sphinx
packet is decoded, then processed by the target peer. If the peer is
indicated as the exit node, then the HTLC is queue’d to be settled
within the next state update.
This commit extends the existing invoiceRegistry functionality to wrap
the on-disk invoices available via the channeldb with an in-memory
cache on invoices. Currently the in-memory cache is only reserved for
the storage of special “debug” invoices which all nodes are able to
settle immediately.
This commit includes some slight refactoring to properly execute force
closures which are initiated by RPC clients.
The CloseLink method within the htlcSwitch has been extended to take an
additional parameter which indicates if the link should be closed
forcefully. If so, then the channelManager which dispatches the request
executes a force closure using the target channel state machine. Once
the closing transaction has been broadcast, the summary is sent to the
utxoNursery so the outputs can be swept once they’re mature.
This commit introduces the concept of a manually initiated “force”
closer within the channel state machine. A force closure is a closure
initiated by a local subsystem which broadcasts the current commitment
state directly on-chain rather than attempting to cooperatively
negotiate a closure with the remote party.
A force closure returns a ForceCloseSummary which includes all the
details required for claiming all rightfully owned outputs within the
broadcast commitment transaction.
Additionally two new publicly exported channels are introduced, one
which is closed due a locally initiated force closure, and the other
which is closed once we detect that the remote party has executed a
unilateral closure by broadcasting their version of the commitment
transaction.
LIGHT-138, LIGHT-141. Due to some issues in sending/receiving parts of lnd,
messages with zero length are not sent. So added some mock content to
NeighborAck. Moved sender/receiver from routing message to wrap message
which contains lnwire routing message.
This commit modifies the daemon’s initialization within the `lndMain`
method to create an instance of the current default ChainNotifier
outside of the LightningWallet.
At this point, since there are no other implementations of the
ChainNotifier, the current concrete implementation BtcdNotifier is used
by default. In the future, once other ChainNotifier implementations are
in place, config parsing should be fed into a factory function which
creates the proper ChainNotifier implementation.
Finally, several imports have been updated to reflect the change in
package name.
This commit modifies the internal workflow for opening or closing a
channel in order to create a path in which RPC clients can receive
updates. Updates are now communicated via channels from the goroutines
spawned by the RPC server to process the request, and the sub-system
within the daemon that actually executes the request.
With this change clients can now receive updates that the request is
pending (final message has been sent to the target client), or that the
request has been completed. Confirmation related updates have not yet
been implemented as that will require some changes to the ChainNotifier
interface.
This commit fixes an omission within the htlcSwitch. With this commit,
a channels bandwidth is now properly updated once an incoming HTLC is
settled.
This also fixes a bug where if a node received a payment, it wouldn’t
be able to then utilize the newly available bandwidth to send further
payments.
This commit integrates BitFury's current routing functionality into lnd. The
primary ochestration point for the routing sub-system in the routingMgr. The
routingMgr manages all persistent and volatile state related to routing within
the network.
Newly opened channels, either when the initiator or responder are inserted into
the routing table once the channel is fully open. Once new links are inserted
the routingMgr can then perform path selection in order to locate an "optimal"
path to a target destination.
This commit optimizes the previous deadlock bug-fix within the peer’s
channelManager which handles driving the LCP state machine with
additional context-specific state.
Rather than forwarding to the HTLC switch within the primary loop which
handles fully locked-in HTLCs, we now launch a distinct goroutine which
is responsible for properly forwarding lock-in HTLC’s to the
htlcSwitch.
This commit *significantly increases* the payment throughput per-core,
per-channel of the daemon.
With this commit updates are properly pipelined respecting the current
revocation window, htlc updates are batched, a timer is checked to push
chain convergence, and htlc update below the batch size are
periodically flushed to the remote chain.
The current pending update timer, trickle timer, and batch size have
been arbitrarily chosen based on my local tests. In the future these
parameters should be chosen to optimize response-time and throughput
after measurements are gathered.
With this commit, calls to htlcSwitch.SendHTLC() are now synchronous,
only returning after the payment has been fully settled. This will
allow one to accurately measure the commitment update speed with the
current state machine implementation which is missing a number of
low-hanging optimizations.
The htlcManager for each channel now keeps a map of cleared HTLC’s
keyed by the index number of the add entry within the state machine’s
HTLC log. This map of HTLC’s will later be used to properly implement
time outs
Additionally, a slight refactoring has been executed w.r.t handling
upstream/downstream messages. This cleans up the main htlcManager loop,
freeing it up for the addition of future logic to properly observe
timeouts as well as, proper batching+trickling of HTLC updates, and a
commitment signature ticker.
With this commitment, the daemon is now able to properly send+redeem
single-hop HTLC’s with another daemon running on the same network.
The htlcManager gains an additional channel which is reads from
receiving updates from either downstream peers, or the rpc server. An
htlcManager is spawned for each active channel with the remote peer. As
a result, the readHandler must now de-multiplex any messages which
update a known channel to the proper htlcManager.
Batching HTLC add updates with a trickle timer has not yet been
implemented, but will be in the near future along with several other
optimizations.
Each active channel now gains its a dedicated htlcManager goroutine
which currently accepts to two golang channels, and a lightning
channel. The “downstream” channel will be used for dispatched multi-hop
payments sent from the htlcSwitch, while the “upstream” channel will be
used once the readHandler de-multiplexes hltc add/timeout/settle
messages.
Each time a new channel is fully created after N confirmations, the
peer’s channelManager registers the new link with the htlcSwitch. Once
the channel is closed either cooperatively or non-cooperatively, then
the link is unregistered.
This commit switches the implementation of the open/close channel RPC’s
from a fully blocking synchronous model to one that’s async by default,
allowing callers to add a sync wrapper.
The new proto specs also allow for “updates” for the pending channels
in the form of new confirmations which progress the pending status of
the channel. At this point, only the final open/close updates have been
implemented. Obtaining confirmation notifications requires a bit of
re-working within the current ChainNotifier interface, thus this has
been deferred to a later time.
This commit fixes a bug introduced within a prior commit. The prior
commit failed to drollery reverse the txid string taken in as user
input, therefore in order to properly close a channel, the user needed
to manually reverse the txid themselves.
With this change, `wire.NewShaHashFromStr` is used which properly
reverses the string within the constructor. This allows the string
reported not be directly used to the close an active channel.
This commit also corrects a few logging messages.
This commit adds the necessary plumbing within the server, peer, and
rpcServer to handle opening and cooperatively closing a channel with a
remote peer.
Many new data structures have been added to the peer in order to allow
it to efficiently manage opening+.losing new/existing lightning
channels. Additional documentation has been added to several methods
within the peer struct, with some minor renaming along with way. The
peer has also gained a dedicated goroutine whose job it is to manage
any requests pertaining to opening, or closing any channels with the
remote peer.
The messages have been added to lnrpc define the requests and responses
to channel open+close messages. Additional channel logic has been added
between the rpcServer, peer, and server in order to properly manage the
necessary synchronization.
* With this commit, then initial draft of the peer struct is finished.
Items to still complete include the interaction between the peer and
internal wallet, version handshake, pings, etc.
